Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science: INCREASE 2022, 22–23 Nov, Indonesia 9811997675, 9789811997679

Table of contents :
Preface
Contents
Contributors
Indonesian Coastlines Controlling the Whole Earth’s Atmosphere
1 Introduction
2 Data and Evidence of Coastal Diurnal Cycle (CDC)
3 Result on Gravity Waves and Zonal Wind
3.1 Coastal Diurnal Cycle (CDC)-Induced Gravity Waves
3.2 Mean Zonal Wind and QBO
3.3 Discussions and Comparison with Theoretical Studies
4 Conclusions
References
EAR Construction Motivation Revisited: Indonesian Coastline Representing Earth
1 Introduction
2 Result and Discussion
2.1 Middle/Upper-Atmospheric Motivation of EAR Before Construction
2.2 Climatological Importance of IMC Recognized in These Two Decades
2.3 Momentum Budget and Vertical Coupling
2.4 Recent Whole IMC Tropospheric Observations
3 Conclusion
References
Study on Diurnal Variation of Rainfall Observed by X-band Polarimetric Radar in Peatlands Over Bengkalis Island, Eastern Sumatra, Indonesia
1 Introduction
2 Instruments and Methodology
2.1 Furuno Radar
2.2 Radar Preprocessing and Rainfall Estimation Algorithms
2.3 Average Hourly Rainfall
3 Observation and Data
4 Results and Discussion
4.1 Evaluation of Rainfall Estimation by the Furuno Radar
4.2 Regional Characteristics of the Diurnal Rainfall Cycle
5 Conclusions
References
Detrended Fluctuation Analysis (DFA) of Gunungsitoli Geomagnetic Station to Assess the Possibility of the Earthquake Precursor
1 Introduction
2 Data and Method
3 Results and Discussion
4 Conclusion
References
Low-Latitude Fluctuation of Ionospheric Magnetic Field Measured by LAPAN-A3 Satellite
1 Introduction
2 Data and Methods
3 Result and Discussions
4 Conclusion
References
Study of the Low Latitude Ionosphere Irregularities Using Multi-instrument Observations
1 Introduction
2 Data and Method
3 Result and Discussion
3.1 Seasonal Variability of Field Aligned Irregularities (FAI)
3.2 Ionospheric Scintillation Occurrences During March 2011
3.3 Relationship Between Field Aligned Irregularities (FAI), Scintillation, and Equatorial Spread F (ESF)
4 Conclusion
References
Distribution of Peat Soil Carbon Under Different Land Uses in Tidal Swampland
1 Introduction
2 Methods
2.1 Study Site
2.2 Carbon Deposit Measurement
2.3 Data Analysis
3 Results and Discussion
4 Conclusion
References
Study of Air Quality and Pollutant Distribution Patterns in Balikpapan Using the WRF-Chem Model
1 Introduction
2 Data and Method
3 Results and Discussion
3.1 Concentration of Air Pollutants in the City of Balikpapan
3.2 Pollutant Distribution in Dry Season, Transition Season, and Wet Season
3.3 Wind Characteristics of Balikpapan City in Balikpapan Baru and Plaza Balikpapan
4 Summary and Conclusion
References
Automatic Processing for Aerosol, Snow/Ice, Cloud, and Volcanic Ash Imagery (ASCI) Products Based on NOAA-JPSS Satellites Data
1 Introduction
2 Data and Method
3 Results
4 Discussion
5 Conclusion
References
The Seasonal Composition of Inorganic Aerosol in an Urban Region of Bandung, Indonesia
1 Introduction
2 Methodology
2.1 Sampling Site
2.2 Data Used
2.3 Data Analysis
3 Result and Discussion
3.1 Seasonal Variabilities
3.2 Meteorological Factors Versus Inorganic Aerosol Concentration
4 Conclusion
References
The Development of DSS SRIKANDI Fifth Version
1 Introduction
2 The DSS SRIKANDI Fifth Version Architecture’s
3 The Interface of DSS SRIKANDI Fifth Version
4 Further Development
5 Summary
References
Characteristics of PM2.5 Concentration at Bandung and Palembang from December 2019 to November 2021 Measured by Low-Cost Sensor
1 Introduction
2 Data and Method
3 Results and Discussions
4 Conclusions
References
Air Pollution Impact During Forest Fire 2019 Over Sumatra, Indonesia
1 Introduction
2 Data and Method
3 Result
3.1 The Strong Positive IOD and Forest Fire 2019
3.2 Environment Atmosphere Impact
4 Conclusion
References
Influence Impregnation Method in the Structure of Bimetallic Ni-Zn/ZrO2 Catalyst
1 Introduction
2 Materials and Methods
2.1 Materials
2.2 Catalyst Preparations
2.3 Catalyst Characterizations
3 Result and Discussion
4 Conclusion
References
Assessment of Pollution and Sources of Metals in the Brantas River in East Java, Indonesia
1 Introduction
2 Method
2.1 Research Site
2.2 Heavy Metal Contamination Assessments
2.3 Multivariate Analysis
3 Result and Discussion
3.1 Heavy Metal Concentration in the Brantas River’s Water
3.2 Heavy Metal Contamination Assessments
4 Conclusion
References
Utilization EDGAR Fifth Version as Input Emission of WRF-Chem Model for Simulating Ozone and PM2.5 Over Jakarta and Its Surroundings Area
1 Introduction
2 Data and Method
2.1 Model WRF-Chem Configuration
2.2 Model Evaluation
3 Result
3.1 PM2.5 Parameter
3.2 Ozone Parameter
4 Conclusion
References
GEMS Satellite Identification on Volcanic Ash Distribution of Mount Dukono Eruption in April 2022
1 Introduction
2 Data and Method
2.1 Research Location
2.2 Data
2.3 Method
3 Results and Discussion
4 Conclusion
References
Emission Inventory, Investigation of EKC Presence, Handling of SO2 Emission, and PBL SO2 Column Problems in Jakarta
1 Introduction
2 Methods and Data
3 Results and Discussion
4 Conclusion
References
Impact of Vortex on Rainfall as a Trigger for Tropical Cyclones (TC) in Maritime Continent of Southern Indonesia (Case Studies: Victoria, Ernie, and Seroja TC)
1 Introduction
2 Data and Method
3 Results and Discussion
3.1 Vortex Events in Maritime Continent of Southern Indonesia
3.2 Single Vortex on Victoria Cyclone Event (2013)
3.3 Double Vortex on Ernie Cyclone Event (2017)
3.4 Triple Vortex on Seroja Cyclone Event (2021)
3.5 Differences Impact of Rainfall on the Number of Vortex Events
4 Conclusions
References
Study of the Inter-Tropical Convergence Zone (ITCZ) Movement Over the Maritime Continent Region
1 Introduction
2 Data and Methodology
3 Results and Discussion
3.1 Bimodality of the ITCZ
3.2 Seasonal Migration of the ITCZ
3.3 ITCZ Jump and Monsoon Onset
4 Conclusions
References
Weather Condition Identification Using Edge Detection Method for Early Warning System
1 Introduction
2 Data and Method
2.1 Data
2.2 Method
3 Result and Analysis
3.1 Preprocessing
3.2 Processing
4 Conclusion
References
Numerical Simulation of MCC Evolution Over Borneo Island Using a High-Resolution Model, Case Study: April 14–15, 2012
1 Introduction
2 Data and Method
3 Results and Discussion
3.1 Influence of Topography in the Initial Stage of MCC
4 Conclusion
References
Tropopause Height Variation Toward Different Land-Sea Convection Activities in Java Using GNSS-RO Data
1 Introduction
2 Data and Method
3 Result
3.1 Seasonal Tropopause Altitude Diurnal Pattern
3.2 Effect of Convective Activity on Tropopause Altitude
4 Discussion
5 Conclusion
References
Sensitivity Analysis of Constructed Analogue Statistical Downscaling Method for Extreme Rainfall Prediction
1 Introduction
2 Data and Method
3 Result and Discussion
4 Summary
References
Meteorological Factors Influencing Coastal Flooding in Semarang, Central Java, Indonesia, on 23 May 2022
1 Introduction
2 Data and Methods
3 Result and Discussion
3.1 The MCS Evolution Ahead of Coastal Flooding
3.2 Evolution of Convection and Near-Surface Easterly Wind
3.3 Ekman Mass Transport (EMT) and Ekman Pumping Velocity (EPV)
4 Conclusion
References
Assessment of Predictability of Convective-Induced Turbulence Event Using High-Resolution Model, Case Study: Hong Kong Airlines Incident on 6 May 2016
1 Introduction
2 Data and Method
3 Results and Discussion
4 Conclusion
References
The Effects of the Cross Equatorial Northerly Surge (CENS) on Atmospheric Convection and Convergence Over Jakarta and the Surrounding Area
1 Introduction
2 Data and Methods
2.1 Data
2.2 Methods
3 Results and Discussion
3.1 The Effects of CENS Over the Land Area
3.2 The Effects of CENS Over the Sea Area
4 Conclusions
References
Comparison Influencing of El-Nino Southern Oscillation and Indian Ocean Dipole on Rainfall Variability During the Asian Winter and Summer Monsoon Over Indonesian Maritime Continent
1 Introduction
2 Data and Method
3 Result and Discussion
3.1 Statistical Data of ENSO and IOD
3.2 Interaction ENSO with IOD
3.3 Interaction of ENSO and IOD If One is Neutral
3.4 SST and Wind Anomaly Analysis
4 Conclusion
References
The Distribution and Characteristics of Mesoscale Convective Complex (MCC) and Its Relation with Rainfall During Madden-Julian Oscillation (MJO) Conditions Over Indonesia
1 Introduction
2 Data and Method
3 Result and Discussion
3.1 MJO Climatologies
3.2 MCC Distribution During MJO Condition
3.3 Characteristics of MCC During MJO Condition
4 Conclusion
References
Diurnal Rainfall Pattern in Riau Islands as Observed by Rain Gauge and IMERG Data
1 Introduction
2 Data and Methods
3 Result and Discussions
4 Conclusions
References
Comparison of Statistical Properties of Rainfall Extremes Between Megacity Jakarta and New Capital City Nusantara
1 Introduction
2 Data and Method
3 Results
3.1 Characteristics of Rainfall Events
3.2 Large-Scale Conditions
4 Conclusion
References
Analysis of Mesoscale Convective Complex (MCC) that Often Appears on the Northern Coast of the Borneo Island
1 Introduction
2 Data and Method
3 Result and Discussion
3.1 MCC Characteristics
3.2 MCC Evolution Using Composite Method
3.3 Cold Pool and Convergent Wind Support the MCC Development
4 Conclusion
References
Impact of Cold Surge Based on Its Strength on Rainfall Distribution in Western Indonesia
1 Introduction
2 Data and Method
3 Result and Discussion
3.1 Rainfall Distribution During CS Event
3.2 Impact of CS During MJO
3.3 Moisture Transport and Case Study Analysis
4 Conclusion
References
Numerical Simulation of Low-Level Wind Shear at Soekarno-Hatta Airport Associated with Landward Propagation of Mesoscale Convective System
1 Introduction
2 Data and Method
3 Results and Discussion
3.1 Mesoscale Convective System
3.2 WRF Simulation
4 Conclusion
References
Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia Using the VECM Model
1 Introduction
2 Data and Methodology
2.1 CENS, MJO, and Rainfall Data
2.2 Building of Vector Error Correction Model
2.3 Rain Cluster Division
3 Result and Discussion
3.1 Time Series Data Plot
3.2 Data Transformation
3.3 Determining Optimum Lag
3.4 Cointegration Test
3.5 VECM
3.6 Test of Impulse Response Functions (IRFs)
3.7 Comparison of the Original Data and the Estimated Data Using the VAR Model
3.8 VECM Modeling Accuracy Values
3.9 Forecasting Results for the Following Period
4 Conclusions
References
Implementation of Zero Runoff to Reduce Runoff Discharges in Timbang Langsa Village, Langsa City
1 Introduction
2 Methods
2.1 Study Area
2.2 Data Collection and Processing
3 Results and Discussion
3.1 Runoff Coefficient
3.2 Time Concentration (Tc)
3.3 Runoff Discharge (Qt)
3.4 Clean Water Needs
3.5 Reduction of Flood Discharge
4 Conclusion
References
Monthly Rainfall Prediction Using Vector Autoregressive Models Based on ENSO and IOD Phenomena in Cilacap
1 Introduction
2 Methodology
2.1 Data
2.2 Method
3 Result and Discussion
3.1 Time Series of Rainfall Data and the ENSO Index, and the IOD
3.2 Determining Optimum Lag
4 Conclusion
References
Extreme Rainfall Clusters in Borneo and Their Synoptic Climate Causes
1 Introduction
2 Methodology
2.1 Data
2.2 Methods
3 Results
3.1 Temporal Distribution of Extreme Rainfall
3.2 Spatial Distribution
3.3 The Synoptic Patterns of Extreme Rainfall Events
4 Discussion
5 Conclusion
References
Analysis of Wind Variations and Differences on the Airport Runways: a Case Study at I Gusti Ngurah Rai Airport, Bali
1 Introduction
2 Data and Methods
3 Result and Analysis
3.1 Wind Pattern in Runways 09 and 27
3.2 Significant Wind Difference Between Runway 09 and 27
4 Conclusion
References
Diurnal Variation of Rainfall Over Bangka Belitung Islands Determined from Rain Gauge and IMERG Observations
1 Introduction
2 Data and Methods
3 Result and Discussion
4 Conclusions
References
Analysis of Upwelling Variations Caused by ENSO Intensification in the Southern Makassar Strait
1 Introduction
2 Study Area and Methods
3 Results and Discussion
3.1 Distribution Pattern of Chlorophyll-A During Weak El Niño and Normal IOD
3.2 Distribution Pattern of Sea Surface Temperature During Weak El Niño and Normal IOD
3.3 Distribution Pattern of Chlorophyll-A During Strong El Niño and Normal IOD
3.4 Distribution Pattern of Sea Surface Temperature During Strong El Niño and Normal IOD
4 Conclusion
References
The Influence of Climatic-Oceanographic Factors on Triggering Sea Level Variations in the Equatorial Malacca Strait, Indonesia
1 Introduction
2 Materials and Methods
2.1 Study Site
2.2 Data Acquisition and Analysis
3 Results and Discussion
3.1 Sea Level Trends Over 27 Years of Observation
3.2 Correlation Analysis of Sea Level Variations Versus Climatic-Oceanographic Parameters
3.3 Possible Influence of Madden Julian Oscillation on Triggering Sea Level Anomaly
4 Conclusion
References
Zooplankton Response to Harmful Algae Blooms (HABs) Species Phytoplankton
1 Introduction
2 Materials and Methods
3 HABs Species Phytoplankton
4 Zooplankton Response to HABs Species
5 Conclusion
References
Verification of Significant Wave Height of the INA-WAVES Against Sentinel 3 Altimetry Data: Case Study for Bali Waters
1 Introduction
2 Data and Methods
2.1 Data Collection
2.2 Data Processing
2.3 Analysis Method
3 Results and Discussion
3.1 Visual Analysis
3.2 Quantitative Analysis
4 Conclusion
References
Warming of the Upper Ocean in the Indonesian Maritime Continent
1 Introduction
2 Material and Methods
3 Results and Discussion
4 Conclusions
References
Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) During Storm Intensification Episodes in the Early Dry Season Period
1 Introduction
2 Methods
3 Results
3.1 Background Conditions
3.2 Ocean–Atmosphere Interaction in Sub-Daily Cases
4 Conclusion
References
Anomalous Sea Surface Temperature and Chlorophyll-a Induced by Mesoscale Cyclonic Eddies in the Southeastern Tropical Indian Ocean During the 2019 Extreme Positive Indian Ocean Dipole
1 Introduction
2 Data and Methods
2.1 Data
2.2 Eddy Detection and Tracking Methods
3 Results and Discussion
3.1 The Characteristics of Mesoscale Eddies in the Southeastern Tropical Indian Ocean
3.2 The Variations of Eddy-Induced Sea Surface Temperature Anomalies and Chlorophyll-a Concentration
4 Conclusion
References
Modeling of Surface Current-Driven Water Pollution and ASEAN Countries Index Water Quality Assessment in the Rupat Strait, Riau Province, Indonesia
1 Introduction
2 Materials and Methods
2.1 Study Site
2.2 Hydrodynamic Model Simulation
2.3 Water Sampling and Analysis
3 Results and Discussion
3.1 Tidal Current Patterns in the Rupat Strait
3.2 The Spatial Distribution of Environmental Parameters
4 Conclusion
References
Identification of Seasonal Water Mass Characteristics in West Sumatra Waters
1 Introduction
2 Study Area and Methods
3 Results and Discussion
3.1 Water Mass Characteristics
3.2 Seasonal Variability of Water Mass
3.3 Seasonal Variability of Horizontal Temperature and Salinity
4 Conclusion
References
The Climate Comfort and Risk Assessment for Tourism in Bali, Indonesia
1 Introduction
2 Data and Methods
3 Result and Analysis
3.1 Climate Parameters in Bali
3.2 Holiday Climate Index (HCI) and Humidity Index (Humidex)
3.3 Relation Between HCI, Humidex, and International Visitor
4 Conclusion
References
Wind Effect on Spectral Ratio Analyses of Acoustic Waves Excited by Volcanic Explosion: Preliminary Result of Application at Sakurajima Volcano, Japan
1 Introduction
2 Method
3 Result
4 Discussion
5 Conclusion
References
Spatial Assessment Impact of Tsunami Hazard on the Transportation Infrastructure in Phuket South of Thailand
1 Introduction
2 Material and Method
2.1 Material
2.2 Method
3 Result and Discussion
4 Conclusion
References
Potential Application of Machine Learning on Agriculture and Capture Fisheries
1 Introduction
2 Theoretical Framework
3 Application of ML
3.1 Application in Agriculture
3.2 Application in Fishery Management of Fisheries Resources
3.3 Monitoring Fishing and Fish Catch
4 Conclusion
References
Climate Change Risk Assessment Toward Agriculture and Food Security in Sumedang Regency
1 Introduction
2 Method
3 Results and Discussion
4 Conclusion
5 Recommendation
References
Measuring the Impact of Climate Resilience Actions in Agriculture: A Preliminary Study
1 Introduction
2 Study Area
3 Data and Method
3.1 Data
3.2 Method
4 Result and Discussion
4.1 Climate Resilient Actions
5 Conclusion
References
Soil CO2 Emissions Through Peat Decomposition in a Strong El Niño Year Were Higher Than in a Normal Year
1 Introduction
2 Methods
2.1 Study Site
2.2 Groundwater Level Measurement
2.3 Soil CO2 Emissions Through Peat Decomposition Measurement
2.4 Data Analysis
3 Results and Discussion
3.1 Groundwater Level
3.2 Soil CO2 Emissions Through Peat Decomposition
4 Conclusion
References
The Level of Paddy Rice Farmer's Adoption on Sustainable Agricultural Practices in Ciamis Regency, West Java
1 Introduction
2 Method
3 Results and Discussion
3.1 Sustainable Agricultural Practices
3.2 The Level of Paddy Rice Farmer’s Adoption on Sustainable Agricultural Practices
4 Conclusion
References
Index-Based Insurance for Climate Risk Management in Indonesia Agriculture
1 Introduction
2 Methodology
3 Result and Discussion
4 Conclusion
References
Near-Future Projections of Rainfall, Temperature, and Solar Radiation in Sumatra Island Under Climate Change Scenarios
1 Introduction
2 Data and Methods
2.1 Study Area
2.2 Data
2.3 Methodology
3 Results
3.1 Projection of Rainfall, Minimum Temperature, Maximum Temperature, and Solar Radiation Under RCP4.5 and RCP8.5 Scenarios
3.2 Spatial Distribution Projected Change of Rainfall
3.3 Spatial Distribution Projected Change of Minimum Temperature
3.4 Spatial Distribution Projected Change of Maximum Temperature
3.5 Spatial Distribution Projected Change of Solar Radiation
4 Conclusion
References
An Assessment of Groundwater Quality in the Southeastern Part of Rajasthan, India
1 Introduction
2 Material and Methods
2.1 The Study Area
2.2 Methodology
3 Results and Discussion
3.1 Water Quality Assessment Using GIS
4 Water Quality Index Analysis
5 Conclusion
References
Application of Remote Sensing Data to Assess Tidal Inundation and Its Potential for Food Crop Cultivation
1 Introduction
2 Data and Method
2.1 Data
2.2 Method
3 Results and Discussion
4 Conclusion
References
Rice Projection in Sumatra by 2045 Regarding Climate Projection and Crop Model
1 Introduction
2 Materials and Methods
2.1 Scope of Activities
2.2 Research Data
2.3 APSIM Model Setup
2.4 Plant Phenology Data
2.5 Parameterization
3 Results and Discussion
3.1 Projected Rice Productivity in Sumatra
4 Conclusions
References
Multiyear La Niña Events and Poor Harvest of Sea Salt in Madura Island
1 Introduction
2 Data and Methods
3 Result and Discussion
4 Conclusion
References
Climate Indicators Triggering Attacks of Rice Stem Borer as Early Detection Information
1 Introduction
2 Material and Methods
3 Result and Discussion
3.1 Result Complement Data with NASA POWER
3.2 The Distribution of Additional Monthly Rice Stem Borer (RSB) Attack
3.3 The Relationship of Area Affected by Rice Stem Borer (RSB) Attack with Climatic Parameters
4 Conclusions
References
Improvement of the Cropping Index and Farmers’ Resilience in Rainfed Fields Through the Application of Climate Smart Agriculture
1 Introduction
2 Methodology
2.1 Study Location
2.2 Rainfall Analysis for Improving the Cropping Calendar Recommendations and the Cropping Patterns
2.3 Capacity Building of Farmers and Improvement of Land Quality
3 Results
3.1 Rainfall Pattern
3.2 Improvement of the Crop Planting and Cropping Pattern Management
3.3 Improvement of Land Quality and Farmer Capacity
3.4 The Impact of CSA Implementation
4 Conclusion
References
Applications of Soil Conditioner Polyacrylamide to Suppress Runoff and P (Phosphorus) Nutrients Loss at the Sweet Corn Cultivation Under Climate Change Issue
1 Introduction
2 Methods
3 Methods
3.1 General Condition of Research Land
3.2 Effect of Polyacrylamide on Surface Runoff
3.3 Effect of Polyacrylamide on P Nutrients Loss
3.4 Effect of Polyacrylamide on Sweet Corn Production
4 Conclusion
References
Elemental Analysis of Breadnut Seed Biochar and Its Potential Application as a Soil Amendment
1 Introduction
2 Methods
2.1 Material
2.2 Biochar Preparation and Characterization
3 Results and Discussions
3.1 Elemental Analysis of Biochar Product
3.2 Potential Application of Biochar as Soil Amendment
3.3 Potential Application of Biochar as Climate Change Mitigation
4 Conclusion
References
The Impact of Climate Change on Meteorological Drought in Yogyakarta
1 Introduction
2 Data and Method
2.1 Data
2.2 Method
3 Results and Discussion
4 Conclusion
References
Foraminiferal Paleoproductivity During Holocene in Sunda Strait
1 Introduction
2 Study Area
3 Analytical Methods
4 Results and Discussion
4.1 Geomorphology
4.2 The Core’s Physical Attributes
4.3 Spectrophotometry Analysis
4.4 Magnetic Susceptibility
4.5 Foraminifera Analysis
5 Conclusion
References
Microplastic Formation from Weathered Single-Use Plastic Straw in Panjang Island Beach, Banten Bay: Preliminary Result
1 Introduction
2 Study Area
3 Result and Discussion
4 Conclusion
References
Comparison of CHAMP GPS Radio Occultation Dry Temperature Profiles Among Different Product Retrievals
1 Introduction
2 Data and Method
3 Results
3.1 Individual Comparison Among Three Products
3.2 Synoptic and Planetary Scale Waves
3.3 Zonal Mean Temperature Fluctuations in the Stratosphere
4 Summary
References
Evaluation of ERA5 Precipitation Reanalysis Data in Indonesia
1 Introduction
2 Data and Methodology
2.1 Study Area
2.2 Dataset
2.3 Evaluation Method and Criteria
3 Results
3.1 Comparison of Precipitation Products
3.2 Basic Statistics and Distribution of the Precipitation Products
4 Conclusion
References
Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis of Atmospheric Conditions During Tropical Cyclone Dahlia
1 Introduction
2 Data and Method
2.1 Time and Location
2.2 Data
2.3 Method
3 Result and Discussion
3.1 Tropical Cyclone Dahlia Movement
3.2 Sea Surface Temperatures Analysis
3.3 Streamline Analysis
3.4 Geopotential Height and Air Temperature Analysis
3.5 Divergence Analysis
3.6 Water Vapor Transport Analysis
3.7 Influence of Tropical Waves
4 Conclusion
References
Evaluation of Detecting and Tracking Algorithms of Reflectivity Area Based on Rain Scanner Observation Data
1 Introduction
2 Methodology
3 Results and Discussion
4 Conclusion
References
Total Suspended Solids Concentration Estimation in Coastal Wasters Using Remote Sensing Data and Machine Learning Approach
1 Introduction
2 Data and Method
2.1 In Situ TSS Data
2.2 Landsat Data Collections and Pre-processing
2.3 Model Development
2.4 Accuracy Assessment
3 Result and Discussion
3.1 The Machine Learning Models for Estimating the TSS from Landsat-8 OLI
3.2 The Reliability of the Estimated TSS Concentration
3.3 TSS Concentration Map Using the Selected Model
4 Conclusion
References
Practical Approach to Designing Radar Linear and Nonlinear Frequency Modulation Chirp Waveforms
1 Introduction
2 Problem Definitions
3 Main Results and Discussion
3.1 Template Signal
3.2 GNU Radio-Compatible Format Data
3.3 Simulation on GNU Radio
3.4 Matched Filtering
3.5 Discussion: Peak Sidelobe Comparison
4 Concluding Remarks
References
Ensembles Simulation on the Seasonal Rainfall Characteristics Over Indonesia Maritime Continent
1 Introduction
2 Data and Methods
2.1 Data
2.2 Methods
3 Results and Discussion
4 Summary and Conclusions
References
Application and Analysis of Remote Sensing for the Initial Baseline of the Quantitative Comfort Index in Ibukota Nusantara (IKN) and Its Surroundings
1 Introduction
2 Data and Method
3 Results and Discussion
4 Conclusion
References
A Model to Describe Elastic Light Scattering of a Single Sphere in a Non-homogeneous Illuminating Light Measured in an Optical Particle Counter
1 Introduction
2 Design and Calibration
3 Model Description
4 Results and Discussion
5 Conclusion
References
Analysis of Atmospheric Conditions on Hail Events in Bandung on March 8, 2022
1 Introduction
2 Data and Method
2.1 Precipitation Analysis
2.2 Cloud Identification and Analysis
3 Result and Discussion
3.1 Temporal Evolution of Precipitation Based on Rain Scanner Observation
3.2 Wind Analysis
3.3 Cloud Identification and Vertical Analysis
4 Conclusion
References
Microphysical Features During Rainfall Events in Bandung, West Java. Case Study: Weather Modification Technology in Citarum Basin, November 2021
1 Introduction
2 Data and Methodology
2.1 Automatic Weather Station (AWS)
2.2 Dual-Polarization Doppler Radar
2.3 Laser Precipitation Monitor
3 Result and Discussion
3.1 AWS Data
3.2 Radar Data 
4 Conclusion
References
The Role of Self-Organization Convective Clouds Resulting in Heavy Rainfall Over the Western Part of Java Island on July 15–16, 2022
1 Introduction
2 Data and Methodology
2.1 Cloud Identification and Analysis
2.2 Precipitation Analysis
2.3 Temperature and Moisture Analysis
3 Result and Discussion
3.1 Cloud Evolution and Propagation
3.2 Precipitation Analysis
3.3 Temperature and Moisture Analysis
4 Conclusions
References
Characteristics of Marine Aerosol Deposition and Its Relation to Wind Condition in Three Locations of West Java
1 Introduction
2 Data and Method
2.1 Research Locations
2.2 Data
2.3 Method
3 Results and Discussion
4 Conclusion
References

Citation preview

Springer Proceedings in Physics 290

Abdul Basit · Erma Yulihastin · Sri Yudawati Cahyarini · Heru Santoso · Widodo S. Pranowo · Lilik Slamet S. · Halda Aditya Belgaman   Editors

Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science INCREASE 2022, 22–23 Nov, Indonesia

Springer Proceedings in Physics Volume 290

Indexed by Scopus The series Springer Proceedings in Physics, founded in 1984, is devoted to timely reports of state-of-the-art developments in physics and related sciences. Typically based on material presented at conferences, workshops and similar scientific meetings, volumes published in this series will constitute a comprehensive up to date source of reference on a field or subfield of relevance in contemporary physics. Proposals must include the following: – – – – –

Name, place and date of the scientific meeting A link to the committees (local organization, international advisors etc.) Scientific description of the meeting List of invited/plenary speakers An estimate of the planned proceedings book parameters (number of pages/articles, requested number of bulk copies, submission deadline).

Please contact: For Americas and Europe: Dr. Zachary Evenson; [email protected] For Asia, Australia and New Zealand: Dr. Loyola DSilva; loyola.dsilva@springer. com

Abdul Basit · Erma Yulihastin · Sri Yudawati Cahyarini · Heru Santoso · Widodo S. Pranowo · Lilik Slamet S. · Halda Aditya Belgaman Editors

Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science INCREASE 2022, 22–23 Nov, Indonesia

Editors Abdul Basit National Research and Innovation Agency Bandung, Indonesia

Erma Yulihastin National Research and Innovation Agency Bandung, Indonesia

Sri Yudawati Cahyarini National Research and Innovation Agency Bandung, Indonesia

Heru Santoso National Research and Innovation Agency Bandung, Indonesia

Widodo S. Pranowo The Indonesian Naval Postgraduate Military Service School (STTAL) Jakarta, Indonesia

Lilik Slamet S. National Research and Innovation Agency Bandung, Indonesia

Halda Aditya Belgaman National Research and Innovation Agency Bandung, Indonesia

ISSN 0930-8989 ISSN 1867-4941 (electronic) Springer Proceedings in Physics ISBN 978-981-19-9767-9 ISBN 978-981-19-9768-6 (eBook) https://doi.org/10.1007/978-981-19-9768-6 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Singapore Pte Ltd. The registered company address is: 152 Beach Road, #21-01/04 Gateway East, Singapore 189721, Singapore

Preface

This book highlights research on climate and atmospheric dynamics over the maritime continent and its impact on sectors. It is part of the 2nd International Conference on Radioscience, Equatorial Atmospheric Science and Environment (INCREASE) organized by the Research Center for Climate and Atmosphere (PRIMA) of National Research and Innovation Agency (BRIN). The symposium aims to provide a scientific platform for university students, school teachers and scientists to discuss ideas and current issues as well as to design the solutions. The symposium was held online via Zoom and ran over two days from November 22 to 23, 2022. The first day of the symposium was opened by presentations with Keynote speakers, Dr. Albertus Sulaiman, Prof. Edvin Aldrian, Prof. Greg McFarquhar, Prof. Manabu Yamanaka and followed by ten invited speakers as climate weather experts from Indonesia, Japan, Philippine and USA. On the second day, the presentations were delivered by selected participants (86 people) who were divided into five groups. The time allocation for each presenter was about 10 minutes followed by the question and answer sessions after all presenters finished delivering their talks. Along with the symposium, we have also facilitated the presenters to submit their manuscripts to be able to publish in a peer-reviewed journal. The review process includes (1) the first screen of manuscripts by the chief editor and the decision for full peer review, (2) the review of manuscripts by one or two reviewers who are experts in the field, (3) final decision of the editors based on reviewers’ comments, (4) decision for accepted or rejected. We have received 135 manuscripts, and only 83 manuscripts were finally selected for the further scientific publication process. High appreciation would be dedicated to the participants and speakers who have attended and supported the 2nd INCREASE. We also would like to thank the scientific committee and reviewers for their contributed time and thorough comments

v

vi

Preface

to improve the manuscripts. Lastly, many thanks should be addressed to the organizing committee for their hard work so that the symposium can be accomplished as planned. See you all in the next INCREASE. Bandung, Indonesia

Dr.rer.nat. Abdul Basit Editor in Chief of 2nd INCREASE Proceeding 2022

Contents

Indonesian Coastlines Controlling the Whole Earth’s Atmosphere . . . . . Manabu D. Yamanaka and Shin-Ya Ogino EAR Construction Motivation Revisited: Indonesian Coastline Representing Earth . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Manabu D. Yamanaka Study on Diurnal Variation of Rainfall Observed by X-band Polarimetric Radar in Peatlands Over Bengkalis Island, Eastern Sumatra, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mariko Ogawa, Manabu D. Yamanaka, Awaluddin, Arief Darmawan, Albertus Sulaiman, Reni Sulistyowati, I. Dewa Gede Arya Putra, and Osamu Kozan Detrended Fluctuation Analysis (DFA) of Gunungsitoli Geomagnetic Station to Assess the Possibility of the Earthquake Precursor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Febty Febriani, Cinantya Nirmala Dewi, Suaidi Ahadi, Titi Anggono, Syuhada, Mohammad Hasib, and Aditya Dwi Prasetio Low-Latitude Fluctuation of Ionospheric Magnetic Field Measured by LAPAN-A3 Satellite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Fitri Nuraeni, La Ode M. Musafar Kilowasid, Clara Y. Yatini, Visca Wellyanita, Satriya Utama, Yoga Andrian, Teti Zubaidah, Wahyudi Hasbi, B. Moh. Andi Aris, Setyanto C. Pranoto, Harry Bangkit, Muzirwan, Ega A. Anggari, Erlansyah, Nata Miharja, Dwi Ratnasari, Prita Ayuningtyas, and Angga Yolanda Putra Study of the Low Latitude Ionosphere Irregularities Using Multi-instrument Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dyah Rahayu Martiningrum, Sri Ekawati, Prayitno Abadi, and Bambang Suhandi

1

13

33

45

55

63

vii

viii

Contents

Distribution of Peat Soil Carbon Under Different Land Uses in Tidal Swampland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nur Wakhid and Siti Nurzakiah Study of Air Quality and Pollutant Distribution Patterns in Balikpapan Using the WRF-Chem Model . . . . . . . . . . . . . . . . . . . . . . . . . Dessy Gusnita, Prawira Yuda Kumbara, Waluyo Eko Cahyono, Angga Yolanda Putra, Fahmi Rahmatia, and Jen Supriyanto Automatic Processing for Aerosol, Snow/Ice, Cloud, and Volcanic Ash Imagery (ASCI) Products Based on NOAA-JPSS Satellites Data . . . Olivia Maftukhaturrizqoh, Andy Indradjad, Tri Astuti Pandansari, Hidayat Gunawan, and Karunika Diwyacitta

71

79

89

The Seasonal Composition of Inorganic Aerosol in an Urban Region of Bandung, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 Wiwiek Setyawati, Dyah Aries Tanti, Saipul Hamdi, Asri Indrawati, Atep Radiana, Sumaryati, Risyanto, and Retno Puji Lestari The Development of DSS SRIKANDI Fifth Version . . . . . . . . . . . . . . . . . . . 111 Emmanuel Adetya, Nani Cholianawati, Prawira Yudha Kombara, Sumaryati, and Ninong Komala Characteristics of PM2.5 Concentration at Bandung and Palembang from December 2019 to November 2021 Measured by Low-Cost Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119 Saipul Hamdi, Sumaryati, Asri Indrawati, Atep Radiana, Syahril Rizal, Ridho Pratama, Fahmi Rahmatia, Yutaka Matsumi, and Takashi Shibata Air Pollution Impact During Forest Fire 2019 Over Sumatra, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129 Prawira Yudha Kombara, Waluyo Eko Cahyono, Wiwiek Setyawati, Hana Listi Fitriana, Emmanuel Adetya, and Alvin Pratama Influence Impregnation Method in the Structure of Bimetallic Ni-Zn/ZrO2 Catalyst . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141 Fildzah ‘Adany, Kiky Corneliasari Sembiring, Mustofa Amirullah, and Reva Edra Nugraha Assessment of Pollution and Sources of Metals in the Brantas River in East Java, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Cicik Oktasari Handayani, Hidayatuz Zu’amah, and Sukarjo Utilization EDGAR Fifth Version as Input Emission of WRF-Chem Model for Simulating Ozone and PM2.5 Over Jakarta and Its Surroundings Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Prawira Yudha Kombara and Ninong Komala

Contents

ix

GEMS Satellite Identification on Volcanic Ash Distribution of Mount Dukono Eruption in April 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173 Amalia Nurlatifah, Emmanuel Adetya, Asri Indrawati, Risyanto, Sumaryati, Ninong Komala, Nani Cholianawati, and Prawira Yudha Kombara Emission Inventory, Investigation of EKC Presence, Handling of SO2 Emission, and PBL SO2 Column Problems in Jakarta . . . . . . . . . . 185 Toni Samiaji Impact of Vortex on Rainfall as a Trigger for Tropical Cyclones (TC) in Maritime Continent of Southern Indonesia (Case Studies: Victoria, Ernie, and Seroja TC) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199 Luthfiyah Jannatunnisa, Nurjanna Joko Trilaksono, and Muhammad Ridho Syahputra Study of the Inter-Tropical Convergence Zone (ITCZ) Movement Over the Maritime Continent Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 Didi Satiadi, Ibnu Fathrio, and Anis Purwaningsih Weather Condition Identification Using Edge Detection Method for Early Warning System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221 Aisya Nafiisyanti, Farid Lasmono, Ibnu Fathrio, Risyanto, Teguh Harjana, Didi Satiadi, and Acep Catur Nugraha Numerical Simulation of MCC Evolution Over Borneo Island Using a High-Resolution Model, Case Study: April 14–15, 2012 . . . . . . . . 231 Ibnu Fathrio and Trismidianto Tropopause Height Variation Toward Different Land-Sea Convection Activities in Java Using GNSS-RO Data . . . . . . . . . . . . . . . . . . 241 Khanifah Afifi and Nurjanna Joko Trilaksono Sensitivity Analysis of Constructed Analogue Statistical Downscaling Method for Extreme Rainfall Prediction . . . . . . . . . . . . . . . . . 251 Trinah Wati, Tri Wahyu Hadi, Faiz Rohman Fajary, Ardhasena Sopaheluwakan, and Lambok M. Hutasoit Meteorological Factors Influencing Coastal Flooding in Semarang, Central Java, Indonesia, on 23 May 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 259 Teguh Harjana, Eddy Hermawan, Risyanto, Anis Purwaningsih, Dita Fatria Andarini, Ainur Ridho, Dian Nur Ratri, and Akas Pinaringan Sujalu

x

Contents

Assessment of Predictability of Convective-Induced Turbulence Event Using High-Resolution Model, Case Study: Hong Kong Airlines Incident on 6 May 2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269 Ibnu Fathrio, Aisya Nafiisyanti, Ina Juaeni, Muhammad Arif Munandar, Dita Fatria, Anis Purwaningsih, Fadli Nauval, Alfan Sukmana Praja, Elfira Saufina, Didi Satiadi, Teguh Harjana, Wendi Harjupa, and Risyanto The Effects of the Cross Equatorial Northerly Surge (CENS) on Atmospheric Convection and Convergence Over Jakarta and the Surrounding Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279 Didi Satiadi, Anis Purwaningsih, Wendi Harjupa, Trismidianto, Dita Fatria Andarini, Fadli Nauval, Elfira Saufina, Fahmi Rahmatia, Ridho Pratama, Teguh Harjana, Risyanto, Ibnu Fathrio, Eddy Hermawan, Mutia Yollanda, and Dodi Devianto Comparison Influencing of El-Nino Southern Oscillation and Indian Ocean Dipole on Rainfall Variability During the Asian Winter and Summer Monsoon Over Indonesian Maritime Continent . . . 291 Trisni Hadiningrum, Deni Okta Lestari, and Trismidianto The Distribution and Characteristics of Mesoscale Convective Complex (MCC) and Its Relation with Rainfall During Madden-Julian Oscillation (MJO) Conditions Over Indonesia . . . . . . . . . 303 Trismidianto Diurnal Rainfall Pattern in Riau Islands as Observed by Rain Gauge and IMERG Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 317 Ravidho Ramadhan, Helmi Yusnaini, Marzuki Marzuki, Zahwa Vieny Adha, Mutya Vonnisa, and Robi Muharsyah Comparison of Statistical Properties of Rainfall Extremes Between Megacity Jakarta and New Capital City Nusantara . . . . . . . . . . . . . . . . . . . 325 Sopia Lestari, Fadli Syamsudin, Teguh A. Pianto, Reni Sulistyowati, Erma Yulihastin, Dwiyoga Nugroho, Rahaden B. Hatmaja, Dava Amrina, Muhammad N. Habibi, and Namira N. Perdani Analysis of Mesoscale Convective Complex (MCC) that Often Appears on the Northern Coast of the Borneo Island . . . . . . . . . . . . . . . . . . 335 Trismidianto Impact of Cold Surge Based on Its Strength on Rainfall Distribution in Western Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 349 Alfan Sukmana Praja and Trismidianto Numerical Simulation of Low-Level Wind Shear at Soekarno-Hatta Airport Associated with Landward Propagation of Mesoscale Convective System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 359 Dita Fatria Andarini, Muhammad Arif Munandar, and Ibnu Fathrio

Contents

xi

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia Using the VECM Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367 Mutia Yollanda, Wendi Harjupa, Dodi Devianto, Dita Fatria Andarini, Fadli Nauval, Elfira Saufina, Anis Purwaningsih, Wendi Harjupa, Trismidianto, Teguh Harjana, Risyanto, Fahmi Rahmatia, Ridho Pratama, and Didi Satiadi Implementation of Zero Runoff to Reduce Runoff Discharges in Timbang Langsa Village, Langsa City . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 385 Eka Mutia, Ellida Novita Lydia, Wan Alamsyah, and Danil Rahmad Priatna Monthly Rainfall Prediction Using Vector Autoregressive Models Based on ENSO and IOD Phenomena in Cilacap . . . . . . . . . . . . . . . . . . . . . 395 Fadli Nauval, Mutia Yollanda, Dodi Devianto, Wendi Harjupa, Dita Fatria Andarini, Elfira Saufina, Anis Purwaningsih, Fahmi Rahmatia, Ridho Pratama, Trismidianto, Teguh Harjana, Risyanto, and Didi Satiadi Extreme Rainfall Clusters in Borneo and Their Synoptic Climate Causes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 407 Narizka Nanda Purwadani, Mohamad Rahman Djuwansah, Muhammad Rais Abdillah, Faiz Rohman Fajary, and Ida Narulita Analysis of Wind Variations and Differences on the Airport Runways: a Case Study at I Gusti Ngurah Rai Airport, Bali . . . . . . . . . . . 417 Kadek Sumaja and Amanda Pasa Kencana Diurnal Variation of Rainfall Over Bangka Belitung Islands Determined from Rain Gauge and IMERG Observations . . . . . . . . . . . . . . 427 Helmi Yusnaini, Zahwa Vieny Adha, Ravidho Ramadhan, Marzuki Marzuki, and Robi Muharsyah Analysis of Upwelling Variations Caused by ENSO Intensification in the Southern Makassar Strait . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437 Gandhi Napitupulu, Naffisa Adyan Fekranie, Susanna Nurdjaman, Totok Suprijo, and Luki Subehi The Influence of Climatic-Oceanographic Factors on Triggering Sea Level Variations in the Equatorial Malacca Strait, Indonesia . . . . . . . 449 Ulung Jantama Wisha, Yusuf Jati Wijaya, and Yukiharu Hisaki Zooplankton Response to Harmful Algae Blooms (HABs) Species Phytoplankton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461 Hanung Agus Mulyadi, Arief Rachman, and Nurul Fitriya

xii

Contents

Verification of Significant Wave Height of the INA-WAVES Against Sentinel 3 Altimetry Data: Case Study for Bali Waters . . . . . . . . . . . . . . . . 477 Timothy Kenoly, Tirtha Wijaya, Andri Ramdhani, Mahardika Jalu Pradana, and Subekti Mujiasih Warming of the Upper Ocean in the Indonesian Maritime Continent . . . 489 Mochamad Furqon Azis Ismail, Asep Sandra Budiman, Abdul Basit, Erma Yulihastin, Herlina Ika Ratnawati, Dewi Surinati, Adi Purwandana, Widodo Setiyo Pranowo, Subekti Mujiasih, Rahaden Bagas Hatmaja, and Praditya Avianto Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) During Storm Intensification Episodes in the Early Dry Season Period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 497 Namira Nasywa Perdani, Ankiq Taofiqorahman, Erma Yulihastin, Rahaden Bagas Hatmaja, Gammamerdianti, Eka Putri Wulandari, Noersomadi, and Haries Satyawardhana Anomalous Sea Surface Temperature and Chlorophyll-a Induced by Mesoscale Cyclonic Eddies in the Southeastern Tropical Indian Ocean During the 2019 Extreme Positive Indian Ocean Dipole . . . . . . . . . 509 Rahaden Bagas Hatmaja, M. Rizqi Ramadhan, Sigit Kurniawan Jati Wicaksana, and Suaydhi Modeling of Surface Current-Driven Water Pollution and ASEAN Countries Index Water Quality Assessment in the Rupat Strait, Riau Province, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 519 Koko Ondara and Ulung Jantama Wisha Identification of Seasonal Water Mass Characteristics in West Sumatra Waters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531 Gandhi Napitupulu, Ivonne M. Radjawane, Nabila Afifah Azuga, Khafid Rizki Pratama, Naffisa Adyan Fekranie, and Hansan Park The Climate Comfort and Risk Assessment for Tourism in Bali, Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 545 Kadek Sumaja, I Kadek Mas Satriyabawa, Sindy Maharani, and Weny Anggi Mustika Wind Effect on Spectral Ratio Analyses of Acoustic Waves Excited by Volcanic Explosion: Preliminary Result of Application at Sakurajima Volcano, Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555 Mohammad Hasib, Takeshi Nishimura, Albertus Sulaiman, Titi Anggono, Syuhada, Febty Febriani, Cinantya Nirmala Dewi, Aditya Dwi Prasetio, and Trinugroho

Contents

xiii

Spatial Assessment Impact of Tsunami Hazard on the Transportation Infrastructure in Phuket South of Thailand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 567 Sirapop Rukhajee, Thanadon Brahmasakha Na Sakolnagara, Jakkrapong Srisuwan, Wasana Putklang, and Muhammad Hanif Potential Application of Machine Learning on Agriculture and Capture Fisheries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 577 Bernadetta Rina Hastiestari and Dewi Syahidah Climate Change Risk Assessment Toward Agriculture and Food Security in Sumedang Regency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585 Hadi Ferdiansyah, Nugrahana Fitria Ruhyana, and Erti Nurfindarti Measuring the Impact of Climate Resilience Actions in Agriculture: A Preliminary Study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595 Elza Surmaini, Woro Estiningtyas, and Yayan Apriyana Soil CO2 Emissions Through Peat Decomposition in a Strong El Niño Year Were Higher Than in a Normal Year . . . . . . . . . . . . . . . . . . . . . . 605 Hidayatuz Zu’amah, Cicik Oktasari Handayani, and Nur Wakhid The Level of Paddy Rice Farmer’s Adoption on Sustainable Agricultural Practices in Ciamis Regency, West Java . . . . . . . . . . . . . . . . . 613 Abdul Mutolib, Candra Nuraini, Indah Listiana, Helvi Yanfika, Raden Ajeng Diana Widyastuti, Ali Rahmat, and Rinaldi Bursan Index-Based Insurance for Climate Risk Management in Indonesia Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 621 Woro Estiningtyas, Kiki Kartikasari, Perdinan, and Saktyanu Kristyantoadi Dermoredjo Near-Future Projections of Rainfall, Temperature, and Solar Radiation in Sumatra Island Under Climate Change Scenarios . . . . . . . . 631 Misnawati, E. Susanti, E. Surmaini, Y. R. Fanggidae, E. R. Dewi, Suciantini, M. R. Syahputra, U. A. Linarka, and A. Sopaheluwakan An Assessment of Groundwater Quality in the Southeastern Part of Rajasthan, India . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 645 Sushmi Nimje, Niruti Gupta, and Yash Kumar Mittal Application of Remote Sensing Data to Assess Tidal Inundation and Its Potential for Food Crop Cultivation . . . . . . . . . . . . . . . . . . . . . . . . . . 655 Lilik Slamet Supriatin, Indah Susanti, Sinta Berliana Sipayung, and Martono

xiv

Contents

Rice Projection in Sumatra by 2045 Regarding Climate Projection and Crop Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 667 E. R. Dewi, E. Susanti, E. Surmaini, Suciantini, Misnawati, U. A. Linarka, F. Ramadhani, A. Dariah, A. Sopaheluwakan, and M. R. Syahputra Multiyear La Niña Events and Poor Harvest of Sea Salt in Madura Island . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 687 Rikha Bramawanto and Suaydhi Climate Indicators Triggering Attacks of Rice Stem Borer as Early Detection Information . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 697 Suciantini, Erni Susanti, Elza Surmaini, Misnawati, and Yudi Riadi Fanggidae Improvement of the Cropping Index and Farmers’ Resilience in Rainfed Fields Through the Application of Climate Smart Agriculture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 707 Aris Pramudia, Abriani Fensionita, Yunita Fauziah Rahim, Asis Purwoko, Andriarti Kusumawardani, and Muhammad Takdir Mulyadi Applications of Soil Conditioner Polyacrylamide to Suppress Runoff and P (Phosphorus) Nutrients Loss at the Sweet Corn Cultivation Under Climate Change Issue . . . . . . . . . . . . . . . . . . . . . . . . . . . . 717 Niken Ida Lovita, Ali Rahmat, Yuliana Eva Agasi, Sendi Purnama Hidayat, Yogina Lestari Ayu Situmorang, and Dwi Rustam Kendarto Elemental Analysis of Breadnut Seed Biochar and Its Potential Application as a Soil Amendment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 727 Sukamto and Ali Rahmat The Impact of Climate Change on Meteorological Drought in Yogyakarta . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 735 Sinta Berliana Sipayung, Indah Susanti, Edy Maryadi, Amalia Nurlatifah, Adi Witono, Lilik Slamet, and Laras Tursilowati Foraminiferal Paleoproductivity During Holocene in Sunda Strait . . . . . 745 Eldian Yosua Budiarto, Marfasran Hendrizan, Rachmad Setijadi, Rainer Arief Troa, Rina Zuraida, and Eko Triarso Microplastic Formation from Weathered Single-Use Plastic Straw in Panjang Island Beach, Banten Bay: Preliminary Result . . . . . . . . . . . . . 757 Dwi Amanda Utami, Sri Yudawati Cahyarini, Ayu Utami Nurhidayati, Tubagus Solihuddin, and Marfasran Hendrizan

Contents

xv

Comparison of CHAMP GPS Radio Occultation Dry Temperature Profiles Among Different Product Retrievals . . . . . . . . . . . . . . . . . . . . . . . . . 765 Noersomadi and Nani Cholianawati Evaluation of ERA5 Precipitation Reanalysis Data in Indonesia . . . . . . . . 781 Sigit Kurniawan Jati Wicaksana and Iis Sofiati Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis of Atmospheric Conditions During Tropical Cyclone Dahlia . . . . . . . . . . . 793 Dendi Rona Purnama, I. Nyoman Agus Astina Putra, Dewangga Palguna, and Gandhi Mahendra Evaluation of Detecting and Tracking Algorithms of Reflectivity Area Based on Rain Scanner Observation Data . . . . . . . . . . . . . . . . . . . . . . . 809 Tiin Sinatra, Edy Maryadi, Syahrul, Syukri Darmawan, Ginaldi Ari Nugroho, and Asif Awaludin Total Suspended Solids Concentration Estimation in Coastal Wasters Using Remote Sensing Data and Machine Learning Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 819 D. N. Lintangsasi, A. Rahmadya, I. Ridwansyah, and F. Setiawan Practical Approach to Designing Radar Linear and Nonlinear Frequency Modulation Chirp Waveforms . . . . . . . . . . . . . . . . . . . . . . . . . . . . 829 Harry Septanto and Djoko Suprijanto Ensembles Simulation on the Seasonal Rainfall Characteristics Over Indonesia Maritime Continent . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 837 Haries Satyawardhana, M. Arif Setyo Aji, Risyanto, Erma Yulihastin, Gammamerdianti, Candra N. Ihsan, Eka P. Wulandari, Lely Q. Avia, and Iis Sofiati Application and Analysis of Remote Sensing for the Initial Baseline of the Quantitative Comfort Index in Ibukota Nusantara (IKN) and Its Surroundings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 849 Indah Susanti, Lilik S. Supriatin, Sinta B. Sipayung, Edy Maryadi, Adi Witono, Martono, Laras Toersilowati, and Amalia Nurlatifah A Model to Describe Elastic Light Scattering of a Single Sphere in a Non-homogeneous Illuminating Light Measured in an Optical Particle Counter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 861 Moch S. Romadhon Analysis of Atmospheric Conditions on Hail Events in Bandung on March 8, 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 869 Elfira Saufina, Tiin Sinatra, Anis Purwaningsih, Dita Fatria Andarini, Fahmi Rahmatia, Fadli Nauval, Ina Juaeni, Asif Awaludin, Aisya Nafiisyanti, Farid Lasmono, Adi Witono, Arief Suryantoro, Eddy Hermawan, and Acep Catur Nugraha

xvi

Contents

Microphysical Features During Rainfall Events in Bandung, West Java. Case Study: Weather Modification Technology in Citarum Basin, November 2021 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 881 Halda A. Belgaman, Sholehhudin A. Ayubi, Saraswati Dewi, Sopia Lestari, Findy Renggono, Edi Riawan, Neneng S. Juariah, and R. D. Goenawan The Role of Self-Organization Convective Clouds Resulting in Heavy Rainfall Over the Western Part of Java Island on July 15–16, 2022 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 893 Anis Purwaningsih, Trismidianto, Dita Fatria Andarini, Noersomadi, Teguh Harjana, Didi Satiadi, Fahmi Rahmatia, Elfira Saufina, Wendi Harjupa, Erma Yulihastin, Fadli Nauval, Ibnu Fathrio, Alfan Sukmana Praja, and Risyanto Characteristics of Marine Aerosol Deposition and Its Relation to Wind Condition in Three Locations of West Java . . . . . . . . . . . . . . . . . . . 907 Dyah Aries Tanti, Arif Rachman, Opik Taopik, Amalia Nurlatifah, Asri Indrawati, Sumaryati, Wiwiek Setyawati, Nani Cholianawati, and Wilin Julian Sari

Contributors

Fildzah ‘Adany Research Center for Chemistry, National Research and Innovation Agency (BRIN), South Tangerang, Banten, Indonesia Prayitno Abadi Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia Muhammad Rais Abdillah Atmospheric Sciences Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Emmanuel Adetya Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia ; Research Center of Climate and Atmosphere, National Research and Innovation Agency, Central Jakarta, Indonesia Zahwa Vieny Adha Department of Physics, Universitas Andalas, Padang, Indonesia Khanifah Afifi Master Program in Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Yuliana Eva Agasi Department Agriculture Engineering, Faculty of Agricultural Industrial Technology, Padjadjaran University, Sumedang, Indonesia Suaidi Ahadi Indonesian Meteorological, Climatological and Geophysical Agency (BMKG), Kemayoran, DKI Jakarta, Indonesia Wan Alamsyah Universitas Samudra, Langsa-Aceh, Indonesia Mustofa Amirullah Research Center for Smart Mechatronics, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Dava Amrina Meteorology Climatology and Geophysics Agency, Jakarta Pusat, Indonesia

xvii

xviii

Contributors

Dita Fatria Andarini Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; National Research and Innovation Agency, Jakarta, Indonesia B. Moh. Andi Aris Research Center for Space, Bandung, West Java, Indonesia Yoga Andrian Research Center for Space, Bandung, West Java, Indonesia Ega A. Anggari Research Center for Satellite Technology, Bogor, West Java, Indonesia Titi Anggono Research Center for Geological Disasters, National Agency of Research and Innovation (BRIN), Jawa Barat, Indonesia Yayan Apriyana Research Centre for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia M. Arif Setyo Aji Bandung Institute of Technology, Bandung, Indonesia Lely Q. Avia Research Center for Atmospheric Science and Technology, Indonesia National Agency of Research and Innovation, Bandung, Indonesia Praditya Avianto National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia Awaluddin Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Jawa Barat, Indonesia Asif Awaludin BRIN—Research Organization for Earth Sciences and Maritime, Bandung, Indonesia; Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Sholehhudin A. Ayubi National Research and Innovation Agency, Jakarta, Indonesia Prita Ayuningtyas Pontianak Observatory for Space and Atmosphere, Pontianak, West Kalimantan, Indonesia Nabila Afifah Azuga Department of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Harry Bangkit Research Center for Smart Mechatronics, Bandung, Indonesia Abdul Basit National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia Halda A. Belgaman National Research and Innovation Agency, Jakarta, Indonesia Rikha Bramawanto Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, Indonesia Eldian Yosua Budiarto Faculty of Earth Sciences and Technology-Institut Teknologi Bandung, Bandung, Indonesia

Contributors

xix

Asep Sandra Budiman National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia Rinaldi Bursan Faculty of Economics, University of Lampung, Bandar Lampung, Indonesia Sri Yudawati Cahyarini Research Center of Climate and Atmosphere, Research Group of Paleoclimate and Paleoenvironment, National Research and Innovation Agency, Jakarta, Indonesia Waluyo Eko Cahyono Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia Nani Cholianawati Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, Indonesia; Research Center of Climate and Atmosphere, National Research and Innovation Agency, Central Jakarta, Indonesia A. Dariah National Research and Innovation Agency, Jakarta, Indonesia Arief Darmawan Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Jawa Barat, Indonesia Syukri Darmawan BRIN—Research Organization for Electronics and Informatics, Bandung, Indonesia Saktyanu Kristyantoadi Dermoredjo Research Centre Industry, Services, and Trade, BRIN, Jakarta, Indonesia

for

Economics

of

Dodi Devianto Department of Mathematics and Data Science, Andalas University, Padang, West Sumatera, Indonesia; Department of Mathematics, UNAND, West Sumatera, Indonesia Cinantya Nirmala Dewi Research Center for Geological Disasters, National Agency of Research and Innovation (BRIN), Jawa Barat, Indonesia E. R. Dewi National Research and Innovation Agency, Jakarta, Indonesia Saraswati Dewi National Research and Innovation Agency, Jakarta, Indonesia Karunika Diwyacitta Research Center for Remote Sensing, National Research and Innovation Agency, Jakarta, Indonesia Mohamad Rahman Djuwansah Research Centre for Environment and Clean Technology, National Agency for Research and Innovation, Bandung, Indonesia Sri Ekawati Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia Erlansyah Research Center for Space, Bandung, West Java, Indonesia

xx

Contributors

Woro Estiningtyas Research Centre for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia; Research Center for Climate and Atmosphere, BRIN, Jakarta, Indonesia Faiz Rohman Fajary Atmospheric Sciences Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia; Graduate Program of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Y. R. Fanggidae National Research and Innovation Agency, Jakarta, Indonesia Yudi Riadi Fanggidae The Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia Ibnu Fathrio Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; National Research and Innovation Agency, Jakarta, Indonesia; Indonesian National Research and Innovation Agency, Jakarta, Indonesia Dita Fatria National Research and Innovation Agency, Bandung, Indonesia Febty Febriani Research Center for Geological Disasters, National Agency of Research and Innovation (BRIN), Jawa Barat, Indonesia; Research Center for Geological Disaster, BRIN, Bandung, Indonesia Naffisa Adyan Fekranie Department of Oceanography, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Abriani Fensionita Directorate of Food Crops Protection, Directorate General of Food Crops, Indonesia Ministry of Agriculture, Jakarta, Indonesia Hadi Ferdiansyah Regional Research and Development Planning Agency of Sumedang Regency, Sumedang, Indonesia Hana Listi Fitriana Research Center for Remote Sensing, National Research and Innovation Agency, Jakarta, Indonesia Nurul Fitriya Research Center for Oceanography, National Research and Innovation Agency, Jakarta Utara, Indonesia Gammamerdianti National Research and Innovation Agency, Jakarta, Indonesia; Research Center for Atmospheric Science and Technology, Indonesia National Agency of Research and Innovation, Bandung, Indonesia R. D. Goenawan National Research and Innovation Agency, Jakarta, Indonesia Hidayat Gunawan Research Center for Remote Sensing, National Research and Innovation Agency, Jakarta, Indonesia Niruti Gupta Malaviya National Institute of Technology, Jaipur, India Dessy Gusnita National Research and Innovation Agency, Bandung, Indonesia

Contributors

xxi

Muhammad N. Habibi Meteorology Climatology and Geophysics Agency, Jakarta Pusat, Indonesia Tri Wahyu Hadi Graduate Program of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Trisni Hadiningrum Department of Atmospheric and Planetary Science, Sumatera Institute of Technology, Lampung, Indonesia Saipul Hamdi Research Center for Climate and Atmosphere, National Research and Innovation Agency-BRIN, Bandung, Indonesia Cicik Oktasari Handayani Research Center for Horticultural and Estate Crops, National Research and Innovation Agency, Cibinong Science Center, Cibinong, Bogor, Indonesia Muhammad Hanif Geoinformatics, Department of Computer Science, College of Computing, Khon Kaen University, Khon Kaen, Thailand Teguh Harjana Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; Indonesian National Research and Innovation Agency, Jakarta, Indonesia Wendi Harjupa Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Wahyudi Hasbi Research Center for Satellite Technology, Bogor, West Java, Indonesia Mohammad Hasib Research Center for Geological Disasters, National Agency of Research and Innovation (BRIN), Jawa Barat, Indonesia Bernadetta Rina Hastiestari Research Centre for Genetic Engineering, National Research and Innovation Agency, Indonesia (BRIN), Jakarta, Indonesia Rahaden Bagas Hatmaja National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia; Research Centre for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia Marfasran Hendrizan Research Center for Climate and Atmosphere-National Research and Innovation Agency, Bandung, Indonesia; Research Center of Climate and Atmosphere, Research Group of Paleoclimate and Paleoenvironment, National Research and Innovation Agency, Jakarta, Indonesia Eddy Hermawan Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Sendi Purnama Hidayat Department Agriculture Engineering, Faculty of Agricultural Industrial Technology, Padjadjaran University, Sumedang, Indonesia

xxii

Contributors

Yukiharu Hisaki Department of Physics and Earth Sciences, Graduate School of Engineering and Science, University of the Ryukyus, Nishihara, Japan Lambok M. Hutasoit Graduate Program of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Candra N. Ihsan Research Center for Atmospheric Science and Technology, Indonesia National Agency of Research and Innovation, Bandung, Indonesia Andy Indradjad Research Center for Remote Sensing, National Research and Innovation Agency, Jakarta, Indonesia Asri Indrawati Research Center for Climate and Atmosphere, National Research and Innovation Agency-BRIN, Bandung, Indonesia; Laboratory Management, Research Facilities and Science, and Technology Park, National Research and Innovation Agency, Central Jakarta, Indonesia; Directorate of Laboratory Management, Research Facilities, and Science and Technology Park, National Research and Innovation Agency, Jakarta, Indonesia Mochamad Furqon Azis Ismail National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia; Humboldt Fellow, GEOMAR Helmholtz Centre for Ocean Research Kiel, Kiel, Germany Luthfiyah Jannatunnisa Master Program in Earth Sciences, Faculty of Earth Science and Technology, Institut Teknologi Bandung, Bandung, Indonesia Ina Juaeni Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Neneng S. Juariah National Research and Innovation Agency, Jakarta, Indonesia Kiki Kartikasari Carbon and Environmental Research Indonesia, Bogor, Indonesia Amanda Pasa Kencana Bandung Institute of Technology, Bandung, Indonesia Dwi Rustam Kendarto Department Agriculture Engineering, Faculty of Agricultural Industrial Technology, Padjadjaran University, Sumedang, Indonesia Timothy Kenoly Marine Science Program, Faculty of Marine Sciences and Fisheries, Udayana University, Bali, Indonesia Ninong Komala Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia; Research Center of Climate and Atmosphere, National Research and Innovation Agency, Central Jakarta, Indonesia Prawira Yudha Kombara Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia; Research Center of Climate and Atmosphere, National Research and Innovation Agency, Central Jakarta, Indonesia

Contributors

xxiii

Osamu Kozan Center for Southeast Asian Studies (CSEAS), Kyoto University, Kyoto, Japan; Research Institute for Humanity and Nature (RIHN), Kyoto, Japan Prawira Yuda Kumbara National Research and Innovation Agency, Bandung, Indonesia Andriarti Kusumawardani Directorate of Food Crops Protection, Directorate General of Food Crops, Indonesia Ministry of Agriculture, Jakarta, Indonesia Farid Lasmono Center for Data and Information, National Research and Innovation Agency (BRIN), Jakarta, Indonesia; Indonesian National Research and Innovation Agency, Jakarta, Indonesia Deni Okta Lestari Department of Atmospheric and Planetary Science, Sumatera Institute of Technology, Lampung, Indonesia Retno Puji Lestari The Center for Standardization of Environmental Quality Instruments, The Ministry of Environment and Forestry, Jakarta, Indonesia Sopia Lestari Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, West Java, Indonesia; National Research and Innovation Agency, Jakarta, Indonesia U. A. Linarka Meteorology, Climatology, and Geophysical Agency, Jakarta, Indonesia; National Research and Innovation Agency, Jakarta, Indonesia D. N. Lintangsasi Undergraduate School of Faculty of Geography, Gadjah Mada University, Yogyakarta, Indonesia Indah Listiana Faculty of Agriculture, University of Lampung, Bandar Lampung, Indonesia Niken Ida Lovita Department Agriculture Engineering, Faculty of Agricultural Industrial Technology, Padjadjaran University, Sumedang, Indonesia Ellida Novita Lydia Universitas Samudra, Langsa-Aceh, Indonesia Olivia Maftukhaturrizqoh Research Center for Remote Sensing, National Research and Innovation Agency, Jakarta, Indonesia Sindy Maharani I Gusti Ngurah Rai Meteorological Office, Indonesian Agency for Meteorology Climatology and Geophysics, Bali, Indonesia Gandhi Mahendra Sentani Meteorological Station, Indonesian Agency for Meteorology Climatology and Geophysics, Sentani, Jayapura Region, Indonesia Dyah Rahayu Martiningrum Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia Martono Research Center for Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia

xxiv

Contributors

Edy Maryadi Research Center for Data Science and Information, National Research and Innovation Agency, Jakarta Pusat, Indonesia; BRIN—Research Organization for Electronics and Informatics, Bandung, Indonesia; Research Center for Data and Information Sciences, National Research and Innovation Agency, West Java, Indonesia Marzuki Marzuki Department of Physics, Universitas Andalas, Padang, Indonesia Yutaka Matsumi Nagoya University, Nagoya, Japan Nata Miharja Research Center for Space, Bandung, West Java, Indonesia Misnawati The Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia; National Research and Innovation Agency, Jakarta, Indonesia Yash Kumar Mittal Malaviya National Institute of Technology, Jaipur, India Robi Muharsyah Agency for Meteorology, Climatology and Geophysics of Republic Indonesia, Jakarta, Indonesia Subekti Mujiasih Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; GeoHydrodynamics and Environment Research (GHER), Department of Astrophysics, Geophysics and Oceanography (AGO), University of Liège, Liège, Belgium; Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Jakarta, Indonesia Hanung Agus Mulyadi Research Center for Oceanography, National Research and Innovation Agency, Jakarta Utara, Indonesia; Department of Marine Science and Technology, Faculty of Fisheries and Marine Science, IPB University, Bogor, Indonesia Muhammad Arif Munandar Indonesian Agency for Meteorological, Climatological, and Geophysics, Jakarta, Indonesia La Ode M. Musafar Kilowasid Pontianak Observatory for Space and Atmosphere, Pontianak, West Kalimantan, Indonesia Weny Anggi Mustika Regional III Office, Indonesian Agency for Meteorology Climatology and Geophysics, Bali, Indonesia Eka Mutia Universitas Samudra, Langsa-Aceh, Indonesia Abdul Mutolib Graduate Program, University of Siliwangi, Tasikmalaya, Indonesia Muzirwan Research Center for Climate and Atmosphere, Bandung, Indonesia Aisya Nafiisyanti Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; Indonesian National Research and Innovation Agency, Jakarta, Indonesia

Contributors

xxv

Gandhi Napitupulu Department of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Ida Narulita Research Centre for Climate and Atmospheric Science, National Agency for Research and Innovation, Bandung, Indonesia; Faculty of Engineering, University of Indonesia, Depok, Indonesia Fadli Nauval Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Sushmi Nimje Malaviya National Institute of Technology, Jaipur, India Takeshi Nishimura Department of Geophysics, Tohoku University, Sendai, Japan Noersomadi National Research and Innovation Agency, Jakarta, Indonesia; Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, Indonesia Acep Catur Nugraha Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; Indonesian National Research and Innovation Agency, Jakarta, Indonesia Reva Edra Nugraha Department of Chemical Engineering, Faculty of Engineering, Universitas Pembangunan Nasional “Veteran” Jawa Timur, Surabaya, East Java, Indonesia Dwiyoga Nugroho Ministry of Marine Affairs and Fisheries Republic of Indonesia, Jakarta Pusat, Indonesia Ginaldi Ari Nugroho BRIN—Research Organization for Earth Sciences and Maritime, Bandung, Indonesia Fitri Nuraeni Research Center for Space, Bandung, West Java, Indonesia Candra Nuraini Faculty of Agriculture, University of Siliwangi, Tasikmalaya, Indonesia Susanna Nurdjaman Oceanography Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Erti Nurfindarti Regional Development Planning Agency of Serang City, Serang, Indonesia Ayu Utami Nurhidayati Research Center of Climate and Atmosphere, Research Group of Paleoclimate and Paleoenvironment, National Research and Innovation Agency, Jakarta, Indonesia Amalia Nurlatifah Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta Pusat, Indonesia; Research Center of Climate and Atmosphere, National Research and Innovation Agency, Central Jakarta, Indonesia; Research Center of Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia;

xxvi

Contributors

Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta, Indonesia Siti Nurzakiah Research Center for Food Crops, National Research and Innovation Agency, Cibinong, Bogor, Indonesia Mariko Ogawa Center for Southeast Asian Studies (CSEAS), Kyoto University, Kyoto, Japan Shin-Ya Ogino Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan Koko Ondara Research Center for Oceanography, National Research and Innovation Agency, Jakarta, Indonesia Dewangga Palguna Komodo Meteorological Station, Indonesian Agency for Meteorology Climatology and Geophysics, West Manggarai, Indonesia Tri Astuti Pandansari Research Center for Remote Sensing, National Research and Innovation Agency, Jakarta, Indonesia Hansan Park Korea-Indonesia Marine Technology Cooperation Research Center, Bandung Institute of Technology, Bandung, Indonesia Namira N. Perdani Faculty of Fishery and Marine Science, Universitas Padjadjaran, West Java, Indonesia Namira Nasywa Perdani Marine Sciences Department, Faculty of Fisheries and Marine Sciences, Padjajaran University, Bandung, Indonesia Perdinan IPB University, Bogor, Indonesia Teguh A. Pianto Research Center for Geospatial, National Research and Innovation Agency, Geostech Building, PUSPIPTEK, South Tangerang, Banten, Indonesia Mahardika Jalu Pradana Agency for Meteorology, Climatology, and Geophysics of the Republic of Indonesia (BMKG), Jakarta, Indonesia Alfan Sukmana Praja Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Aris Pramudia Research Center for Climate and Atmospheric, National Research and Innovation Agency, Kota, Bandung, West Java, Indonesia; Innovation Center for Tropical Science, Bogor, Indonesia Setyanto C. Pranoto Research Center for Space, Bandung, West Java, Indonesia Widodo Setiyo Pranowo National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia Aditya Dwi Prasetio Research Center for Geological Disasters, National Agency of Research and Innovation (BRIN), Jawa Barat, Indonesia

Contributors

xxvii

Alvin Pratama Department of Atmospheric and Planetary Science, Sumatera Institute of Technology, Lampung, Indonesia Khafid Rizki Pratama Department of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Ridho Pratama Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Danil Rahmad Priatna Universitas Samudra, Langsa-Aceh, Indonesia Dendi Rona Purnama Public Weather Services, Indonesian Agency for Meteorology Climatology and Geophysics, Kemayoran, Central Jakarta, Indonesia Narizka Nanda Purwadani Master Program in Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Adi Purwandana National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia Anis Purwaningsih Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Asis Purwoko Directorate of Food Crops Protection, Directorate General of Food Crops, Indonesia Ministry of Agriculture, Jakarta, Indonesia Wasana Putklang Geoinformatics, Department of Computer Science, College of Computing, Khon Kaen University, Khon Kaen, Thailand Angga Yolanda Putra Pontianak Observatory for Space and Atmosphere, Pontianak, West Kalimantan, Indonesia; National Research and Innovation Agency, Bandung, Indonesia I. Dewa Gede Arya Putra Center for Research and Development, Meteorological, Climatological, and Geophysical Agency (BMKG), Jakarta Pusat, Indonesia I. Nyoman Agus Astina Putra Kendari Marine Meteorological Station, Indonesian Agency for Meteorology Climatology and Geophysics, Kendari, Indonesia Arif Rachman Directorate of Laboratory Management, Research Facilities, and Science and Technology Park, National Research and Innovation Agency, Jakarta, Indonesia Arief Rachman Research Center for Oceanography, National Research and Innovation Agency, Jakarta Utara, Indonesia; Directorate of Laboratory Management, Research Facilities, and Science and Technology Park, National Research and Innovation Agency, Jakarta, Indonesia Atep Radiana Research Center for Climate and Atmosphere, National Research and Innovation Agency-BRIN, Bandung, Indonesia

xxviii

Contributors

Ivonne M. Radjawane Oceanography Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia; Korea-Indonesia Marine Technology Cooperation Research Center, Bandung Institute of Technology, Bandung, Indonesia Yunita Fauziah Rahim Directorate of Food Crops Protection, Directorate General of Food Crops, Indonesia Ministry of Agriculture, Jakarta, Indonesia A. Rahmadya Research Center for Limnology and Water Resources, National Research and Innovation Agency (BRIN), Jakarta Pusat, Indonesia Ali Rahmat Research Center for Limnology and Water Resources, National Research and Innovation Agency, Bogor, Indonesia; Research Center for Limnology and Water Resources, National Research and Innovation Agency, Jakarta Pusat, Indonesia Fahmi Rahmatia Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; Agam Space and Atmospheric Observation Station, National Research and Innovation Agency, West Sumatera, Indonesia M. Rizqi Ramadhan Oceanography Department, Faculty of Earth Sciences and Technology, Bandung Institute of Technology, Bandung, Indonesia Ravidho Ramadhan Department of Physics, Universitas Andalas, Padang, Indonesia F. Ramadhani National Research and Innovation Agency, Jakarta, Indonesia Andri Ramdhani Agency for Meteorology, Climatology, and Geophysics of the Republic of Indonesia (BMKG), Jakarta, Indonesia Dwi Ratnasari Mataram University, Lombok, Indonesia Herlina Ika Ratnawati National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia Dian Nur Ratri Meteorological, Climatological, and Geophysical Agency (BMKG), Jakarta, Indonesia; Droevendaalsesteeg, Wageningen University and Research, Wageningen, The Netherlands Findy Renggono National Research and Innovation Agency, Jakarta, Indonesia Edi Riawan Bandung Institute of Technology, Bandung, Indonesia Ainur Ridho Cerdas Antisipasi Risiko Bencana Indonesia (CARI), Bandung, Indonesia I. Ridwansyah Research Center for Limnology and Water Resources, National Research and Innovation Agency (BRIN), Jakarta Pusat, Indonesia

Contributors

xxix

Risyanto Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; Laboratory Management, Research Facilities and Science, and Technology Park, National Research and Innovation Agency, Central Jakarta, Indonesia Syahril Rizal Bina Darma University, Palembang, Indonesia Moch S. Romadhon Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta Pusat, Indonesia Nugrahana Fitria Ruhyana Regional Research and Development Planning Agency of Sumedang Regency, Sumedang, Indonesia; Universitas Sebelas April, Sumedang, Indonesia Sirapop Rukhajee Geoinformatics, Department of Computer Science, College of Computing, Khon Kaen University, Khon Kaen, Thailand Thanadon Brahmasakha Na Sakolnagara Geoinformatics, Department of Computer Science, College of Computing, Khon Kaen University, Khon Kaen, Thailand Toni Samiaji Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, West Java, Indonesia Wilin Julian Sari Research Center for Quantum Physics, National Research and Innovation Agency, Jakarta, Indonesia Didi Satiadi Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; Indonesian National Research and Innovation Agency, Jakarta, Indonesia I Kadek Mas Satriyabawa I Gusti Ngurah Rai Meteorological Office, Indonesian Agency for Meteorology Climatology and Geophysics, Bali, Indonesia Haries Satyawardhana National Research and Innovation Agency, Jakarta, Indonesia; Research Center for Atmospheric Science and Technology, Indonesia National Agency of Research and Innovation, Bandung, Indonesia Elfira Saufina Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Kiky Corneliasari Sembiring Research Center for Chemistry, National Research and Innovation Agency (BRIN), South Tangerang, Banten, Indonesia Harry Septanto Research Center for Smart Mechatronics, National Research and Innovation Agency, KST Samaun Samadikun, Bandung, Indonesia F. Setiawan Research Center for Limnology and Water Resources, National Research and Innovation Agency (BRIN), Jakarta Pusat, Indonesia

xxx

Contributors

Rachmad Setijadi Department of Geological Engineering, Universitas Jenderal Soedirman, Purbalingga, Indonesia Wiwiek Setyawati Research Center for Climate and Atmosphere, National Research and Innovation Agency-BRIN, Bandung, Indonesia; Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta, Indonesia Takashi Shibata Nagoya University, Nagoya, Japan Tiin Sinatra BRIN—Research Organization for Earth Sciences and Maritime, Bandung, Indonesia; Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Sinta B. Sipayung Research Center of Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia Sinta Berliana Sipayung Research Center for Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia; Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta Pusat, Indonesia Yogina Lestari Ayu Situmorang Department Agriculture Engineering, Faculty of Agricultural Industrial Technology, Padjadjaran University, Sumedang, Indonesia Lilik Slamet Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta Pusat, Indonesia Iis Sofiati Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia; Research Center for Atmospheric Science and Technology, Indonesia National Agency of Research and Innovation, Bandung, Indonesia Tubagus Solihuddin Research Center of Climate and Atmosphere, Research Group of Paleoclimate and Paleoenvironment, National Research and Innovation Agency, Jakarta, Indonesia A. Sopaheluwakan Meteorology, Climatology, and Geophysical Agency, Jakarta, Indonesia Ardhasena Sopaheluwakan Indonesia Agency for Meteorology, Climatology, and Geophysics, Jakarta, Indonesia Jakkrapong Srisuwan Geoinformatics, Department of Computer Science, College of Computing, Khon Kaen University, Khon Kaen, Thailand Suaydhi Research Centre for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, Indonesia Luki Subehi Research Center for Limnology and Water Resources, National Research and Innovation Agency, Bogor, Indonesia

Contributors

xxxi

Suciantini The Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia; National Research and Innovation Agency, Jakarta, Indonesia Bambang Suhandi Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia Akas Pinaringan Sujalu Universitas, Samarinda, Indonesia Sukamto Hokkaido University, Sapporo, Japan Sukarjo Research Center for Horticultural and Estate Crop, National Research and Innovation Agency, Cibinong Science Center, Cibinong, Indonesia Albertus Sulaiman Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Jawa Barat, Indonesia; Research Center for Climate and Atmosphere, BRIN, Bandung, Indonesia Reni Sulistyowati Research Center for Geospatial, National Research and Innovation Agency, Geostech Building, PUSPIPTEK, South Tangerang, Banten, Indonesia; Research Center for Geospatial, National Research and Innovation Agency (BRIN), Gedung GEOSTECH, Serpong, Tangerang Selatan, Indonesia Kadek Sumaja I Gusti Ngurah Rai Meteorological Office, Indonesian Agency for Meteorology Climatology and Geophysics, Bali, Indonesia Sumaryati Research Center for Climate and Atmosphere, National Research and Innovation Agency-BRIN, Bandung, Indonesia; Research Center of Climate and Atmosphere, National Research and Innovation Agency, Central Jakarta, Indonesia Lilik S. Supriatin Research Center of Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia Lilik Slamet Supriatin Research Center for Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia Djoko Suprijanto Faculty of Mathematics and Natural Sciences, Combinatorial Mathematics Research Group, Institut Teknologi Bandung, Bandung, Indonesia Totok Suprijo Oceanography Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia Jen Supriyanto Environmental Agency, Balikpapan, Indonesia Dewi Surinati National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia E. Surmaini National Research and Innovation Agency, Jakarta, Indonesia Elza Surmaini Research Centre for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia

xxxii

Contributors

Arief Suryantoro Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia E. Susanti National Research and Innovation Agency, Jakarta, Indonesia Erni Susanti The Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia Indah Susanti Directorate of Repository, Multi Media, and Scientific Publishing, National Research and Innovation Agency, West Java, Indonesia; Directorate of Repository, Multimedia, and Scientific Publishing, National Research and Innovation Agency, Jakarta Pusat, Indonesia Dewi Syahidah Research Centre for Veterinary, National Research and Innovation Agency, Indonesia (BRIN), Jakarta, Indonesia M. R. Syahputra Bandung Institute of Technology, Bandung, Indonesia Muhammad Ridho Syahputra Atmospheric Sciences Research Group, Faculty of Earth Science and Technology, Institut Teknologi Bandung, Bandung, Indonesia Syahrul BRIN—Research Organization for Electronics and Informatics, Bandung, Indonesia Fadli Syamsudin Faculty of Fishery and Marine Science, Universitas Padjadjaran, West Java, Indonesia Syuhada Research Center for Geological Disasters, National Agency of Research and Innovation (BRIN), Jawa Barat, Indonesia Muhammad Takdir Mulyadi Directorate of Food Crops Protection, Directorate General of Food Crops, Indonesia Ministry of Agriculture, Jakarta, Indonesia Dyah Aries Tanti Research Center for Climate and Atmosphere, National Research and Innovation Agency-BRIN, Bandung, Indonesia; Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta, Indonesia Ankiq Taofiqorahman Marine Sciences Department, Faculty of Fisheries and Marine Sciences, Padjajaran University, Bandung, Indonesia Opik Taopik Directorate of Laboratory Management, Research Facilities, and Science and Technology Park, National Research and Innovation Agency, Jakarta, Indonesia Laras Toersilowati Research Center of Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia Eko Triarso Ministry of Marine Affairs and Fisheries, Jakarta, Indonesia Nurjanna Joko Trilaksono Atmospheric Sciences Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia

Contributors

xxxiii

Trinugroho Research Center for Geological Disaster, BRIN, Bandung, Indonesia Trismidianto Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia Rainer Arief Troa Ministry of Marine Affairs and Fisheries, Jakarta, Indonesia Laras Tursilowati Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta Pusat, Indonesia Satriya Utama Research Center for Satellite Technology, Bogor, West Java, Indonesia Dwi Amanda Utami Research Center of Climate and Atmosphere, Research Group of Paleoclimate and Paleoenvironment, National Research and Innovation Agency, Jakarta, Indonesia Mutya Vonnisa Department of Physics, Universitas Andalas, Padang, Indonesia Nur Wakhid Research Center for Ecology and Ethnobiology, National Research and Innovation Agency, Cibinong, Bogor, Indonesia Trinah Wati Graduate Program of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia; Indonesia Agency for Meteorology, Climatology, and Geophysics, Jakarta, Indonesia Visca Wellyanita Research Center for Space, Bandung, West Java, Indonesia Sigit Kurniawan Jati Wicaksana Research Centre for Oceanography, National Research and Innovation Agency, Jakarta, Indonesia Raden Ajeng Diana Widyastuti Faculty of Agriculture, University of Lampung, Bandar Lampung, Indonesia Tirtha Wijaya Agency for Meteorology, Climatology, and Geophysics of the Republic of Indonesia (BMKG), Badung, Bali, Indonesia Yusuf Jati Wijaya Department of Oceanography, Faculty of Fisheries and Marine Science, Diponegoro University, Semarang, Indonesia Ulung Jantama Wisha Department of Physics and Earth Sciences, Graduate School of Engineering and Science, University of the Ryukyus, Nishihara, Japan; Research Center for Oceanography, National Research and Innovation Agency (BRIN), Jakarta, Indonesia Adi Witono Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta Pusat, Indonesia; Research Center of Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia; Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia

xxxiv

Contributors

Eka P. Wulandari Research Center for Atmospheric Science and Technology, Indonesia National Agency of Research and Innovation, Bandung, Indonesia Eka Putri Wulandari National Research and Innovation Agency, Jakarta, Indonesia Manabu D. Yamanaka Research Institute for Humanity and Nature, Kyoto, Japan; Professor Emeritus, Kobe University, Kobe, Japan; Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan Helvi Yanfika Faculty of Agriculture, University of Lampung, Bandar Lampung, Indonesia Clara Y. Yatini Research Center for Space, Bandung, West Java, Indonesia Mutia Yollanda Department of Mathematics and Data Science, Andalas University, Padang, West Sumatera, Indonesia; Department of Mathematics, UNAND, West Sumatera, Indonesia Erma Yulihastin Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia; National Research and Innovation Agency of Indonesia (BRIN), Jakarta, Indonesia; Research Center for Atmospheric Science and Technology, Indonesia National Agency of Research and Innovation, Bandung, Indonesia Helmi Yusnaini Department of Physics, Universitas Andalas, Padang, Indonesia Teti Zubaidah Mataram University, Lombok, Indonesia Rina Zuraida Research Center for Climate and Atmosphere-National Research and Innovation Agency, Bandung, Indonesia Hidayatuz Zu’amah Research Center for Horticultural and Estate Crops, National Research and Innovation Agency, Cibinong Science Center, Cibinong, Bogor, Indonesia

Indonesian Coastlines Controlling the Whole Earth’s Atmosphere Manabu D. Yamanaka

and Shin-Ya Ogino

Abstract A coastline is the triple boundary among land, ocean, and atmosphere constructing the Earth system. We have revealed that the diurnal cycle (sea-land breeze circulation) along the world’s longest coastline surrounding major islands of the Indonesian maritime continent (IMC) is the most robust mode of cloud-rainfall generation. This process produces latent heat compensating the global radiationaldynamical energy imbalance, and any modifications of the IMC land surface may change the local diurnal cycle and finally the global climate. The diurnal cycle circulation consists of equi-amplitude, bidirectional (sea- and land-ward propagating), internal gravity waves, and our recent analysis (Yamanaka and Ogino in Tropical coastlines robustizing the stratospheric quasi-biennial oscillation. (in preparation) (2022)) shows that they propagate also upward and maintain the stratospheric most robust mode: quasi-biennial oscillation (QBO). As so far shown by many studies, QBO affects the whole middle and upper atmospheric processes, and therefore the IMC coastlines control the whole Earth’s atmosphere.

1 Introduction Robust periodic phenomena in the atmosphere should be forced by fixed (e.g., astronomical) sources. However, any such sources have been known for the quasi-biennial oscillation (QBO; discovered by [4, 28] reviews by e.g., [1, 2, 8, 9] maintained robustly with around 2.5 years periodicity in the equatorial lower stratosphere. Almost zonally-uniform zonal-wind amplitude and downward phase progression are around ± 20 m/s and (20–35 km)/(27 months/2) ≈ −1.1 km/month, respectively, M. D. Yamanaka (B) Research Institute for Humanity and Nature, Kyoto, Japan e-mail: [email protected]; [email protected] Kobe University, Kobe, Japan M. D. Yamanaka · S.-Y. Ogino Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_1

1

2

M. D. Yamanaka and S.-Y. Ogino

and the easterly phase is somewhat stronger and slower than the westerly phase. QBO has been explained by the wave-mean flow interaction (e.g., [19, 25, 26], but requested bidirectional (east- and west-ward propagating) waves generated in the troposphere have not been identified. Another robust phenomenon confirmed recently in the equatorial lower troposphere is the coastal diurnal cycle (CDC) of sea-land breeze circulations (SLBC) (e.g., reviews by [34, 37]). The diurnal cycle is more dominant than seasonal (annual and semiannual) cycles and on land also than the intraseasonal/interannual variations, and the latter phenomena appear as amplitude modulations of the former. Those climatological modulations (as well as anthropogenic land modifications) cause the sea-land temperature contrast that forces the CDC-SLBC. SLBC is a sum of up- and downward propagating internal gravity waves in latitudes lower than about 30º ([29], Section 6.5 of [17]). Dominance of CDC is the reason why the Indonesian maritime continent (IMC) with the longest coastlines has the most active convection on Earth [27, 34]. These two robust phenomena appear dominantly in the latitudes lower than about 15º, which is narrower than the internal range of diurnal waves, and both of them are still not accurately simulated in an operational numerical model. These suggest that they are related to each other. The equatorial SLBC is somewhat larger horizontally and vertically than the other (higher) latitude regions, which corresponds to horizontal and vertical wavelengths of 103 km and 10 km, respectively, of internal gravity waves with 1-day periodicity of 102 times of Väisälä-Brunt period (as expected from the dispersion relation). In this paper, we analyze a global objective reanalysis dataset of horizontal wind and pressure velocity and obtain components of 1-day periodicity and zonal scale shorter than 30° in longitude. They are concentrated along coastlines of IMC and African and South American continents at the surface, and their local-time behaviors correspond clearly to SLBC. Upward fluxes of their zonal momenta are distributed in zonal-wind vertical shears of the same direction, as expected from the QBO theory.

2 Data and Evidence of Coastal Diurnal Cycle (CDC) Major features of CDC-SLBC analyzed from campaign observations over major islands of IMC have been shown in the INCREASE symposium last year [35]. SLBC corresponds to internal gravity waves, with sea-/land-ward phase velocity c = ω/k, where the frequency ω = 2π/1 day, and the horizontal wavenumber k ~ 2π/(102 −103 km). Usually, c = (1−3) × 102 km/10 h = 3−7 m/s, which is faster on sea than on land and is much slower than cumulonimbus gust (~20 m/s). The amplitude is proportional to the sea-land temperature contrast kΔT, where ΔT is the temperature difference determined mainly by the land surface temperature with daytime heating and nighttime cooling. Inland sea propagations between major islands and re-excitation at small islands on the way have been observed.

Indonesian Coastlines Controlling the Whole Earth’s Atmosphere

3

In small islands, propagation directions are unclear, and the periodicities are often semi-diurnal (with a secondary peak in the seaside of a major island). In order to study CDC-SLBC over the whole equatorial region, we [36] used a 1.25° latitude/longitude grid dataset of the Japanese 55-year Reanalysis (JRA-55; [14]) for 20-years during 00 UTC 1 January 2000–18 UTC 31 December 2019. In the vertical direction and time, data are given for the surface and 37 isobaric surfaces (1000, 975, 950, 925, 900, 875, 850, 825, 800, 775, 750, 700, 650, 600, 550, 500, 450, 400, 350, 300, 250, 225, 200, 175, 150, 125, 100, 70, 50, 30, 20, 10, 7, 5, 3, 2, and 1 hPa) every six hours (00, 06, 12, and 18 UTC). We have used the same dataset of 1981–2010 for analyzing the coastal concentration along coastlines mainly due to CDC [23]. The JRA-55 dataset has been used in QBO studies (e.g., [1]). The 20-year mean of surface horizontal wind anomaly at each UTC from the diurnal mean and its divergence are shown in Fig. 1 for IMC and African and South American continents. We confirm CDC-SLBC is distributed along all the coastlines, and is most active over IMC with the longest coastlines [34, 37].

Fig. 1 Diurnal cycles analyzed from JRA-55 (2000–2019) for equatorial regions of Africa (left), IMC (middle), and South America (right) [36]. The 20-year mean of surface horizontal wind anomaly (vectors) at each UTC from the daily mean and its divergence (contours) are shown. Results for Africa and South America are displaced by +6 and −12 h, respectively, from those for IMC (LT = UTC + (7−9) h), showing the morning (top), afternoon (second row), evening (third row), and midnight (bottom)

4

M. D. Yamanaka and S.-Y. Ogino

3 Result on Gravity Waves and Zonal Wind 3.1 Coastal Diurnal Cycle (CDC)-Induced Gravity Waves The diurnal cycle consists of local CDC-SLBC and global tide (see discussions later). An example of zonal wind anomaly from daily mean is shown in Fig. 2. A westward super-sonic (40,000 km/day = 463 m/s) propagation of the diurnal tide is found, but it is too fast to interact with the mean flow at least up to the middle atmosphere. The other disturbances are also diurnally reversed, but much shorter, vertical, and almost stationary (in such a global plot) in the troposphere below 100 hPa (about 17 km) level, which is reasonable as SLBCs. Between 100 and 30 hPa levels, where the zonal mean wind was westerly for the time of Fig. 3 (January 2000), contours seem rather horizontal, suggesting typical stratospheric gravity waves (with diurnal periodicity in this case), which is consistent to their interactions with the mean zonal flow and generation of QBO shown later in this paper. In order to study internal gravity waves generated as the CDC-SLBC seen clearly in Fig. 1, we make a monthly mean and a high-pass filter of 30° in longitude, and obtain 20-year mean zonal-vertical plots along the equator of zonal-vertical flow and vertical flux of zonal momentum (divided by atmospheric density) as shown in Fig. 3, in which p-velocity in the p-coordinate are converted into vertical velocity in the lnp coordinate. Westward (eastward) waves are emitted dominantly from the west (east) coast to the troposphere. This implies that seaward waves are stronger than landward waves, and the bidirectional waves are balanced in the both coasts of each continent, rather than the day and night at each coast. Because of 20-year mean, stratospheric values of Fig. 3 are very small.

3.2 Mean Zonal Wind and QBO Figure 4 shows the interannual-vertical variations of the zonal-mean zonal wind and longitudinally 30°-high passed vertical flux of zonal momentum (roughly multiplied by gravity acceleration). In the lower troposphere the zonal flow is very weak easterly, and an annual cycle of northern winter-peaked nearly 850 hPa eastward and summerto-winter ascending westward momentum fluxes is found. In the upper troposphere an annual cycle of northern summer easterly and winter westerly is found, but the momentum flux is unclear. The stratospheric part is enlarged in Fig. 5. QBO is dominant as known well so far, and upward fluxes of zonal momenta of CDCSLBC-waves distributed upward-decreasingly in zonal-wind vertical shears of the same directions are known newly in this paper. We have also confirmed that zonal momentum fluxes with lower cutoff and no filters show much worse correlations with QBO than 30°-high passed results shown in Figs. 4 and 5. Therefore, the CDC-SLBC waves make the largest contribution (at least among the whole 1-day period waves) to interaction with QBO. It has been reported that QBO was disrupted around 2016

Indonesian Coastlines Controlling the Whole Earth’s Atmosphere

5

Fig. 2 Example of monthly-mean zonal-vertical plot along the equator at each UTC of zonal wind anomaly from daily mean of JRA-55 data in January 2000 [36]. In addition to local diurnal cycles, a fast westward propagating diurnal tide with a peak at 06 LT (indicated by a green arrow)

(e.g., [24]) (and also 2020 but is not seen well in the data until 2019). During this period, the CDC-SLBC wave momentum flux is also somewhat unusual.

3.3 Discussions and Comparison with Theoretical Studies We have shown observationally that the both robust phenomena CDC-SLBC and QBO are related with each other. Theoretically, the CDC is forced by sea and land

6

M. D. Yamanaka and S.-Y. Ogino

Fig. 3 Same as Fig. 1 but for zonal-vertical plots along the equator of longitudinally 30°-high passed monthly-mean zonal-vertical wind anomaly (vectors) from the daily mean and their product (contours) [36]

surface temperature difference as shown in Fig. 6. Coastal sea-surface temperature is changed by ~± 1 °C/year or weaker for interannual or intraseasonal variations, whereas the land surface temperature is changed by ~± 3 °C/half day or larger every day. Therefore, the bottom boundary condition of CDC-SLBC for temperature θ is: θ = θ0 − θ1 (sin kx + 1) sin ωt θ1 = θ0 − [cos(kx − ωt) − cos(kx + ωt) + 2 sin ωt] at z = 0, 2

(1)

where x, z, and t are the distance from coast, altitude, and time, k ~ 2π/(102 −103 ) km, and ω ~ 2π/1 day are horizontal wavenumber (a component perpendicular to the coastline) and frequency, θ0 and θ0 ± 2θ1 are sea-side and land-side (highest/lowest) temperature, respectively. The first two terms of the bracket of (1) excite bidirectional waves, and the last term corresponds to the diurnal tide with the circum-equatorial wavelength (= 4 × 104 km). The bidirectional waves make a standing wave in the troposphere, which is SLBC, and are separated by the mean zonal flow in the stratosphere: one wave with the same propagation direction as the vertical shear of the mean zonal flow is amplified and absorbed, and the other wave is propagated into the upper layer with the opposite shear of the mean zonal flow. The diurnal tide

Indonesian Coastlines Controlling the Whole Earth’s Atmosphere

7

Fig. 4 Temporal-vertical sections along the equator of the zonal-daily-mean zonal flow (top) and product of longitudinally 30°-high passed monthly-mean zonal wind and –p-velocity anomalies from the daily means (bottom) analyzed from the 20-year JRA-55 data [36]

has a supersonic propagation velocity, which is far from and has no interaction with the mean flow (except the top of the atmosphere just below the turbopause; see, e.g., [22]). Note that shorter waves generated with individual clouds cannot propagate upward (i.e., trapped and ducting near the ground), as discussed by Scorer [30]. It should be noted that Kawatani et al. [12] studied contributions of various waves in a high-resolution general circulation model, but the diurnal component was filtered out. The situations mentioned above are quite similar to Plumb’s (1977) model and Plumb and McEwan’s (1978) experiment which simulated QBO with standing waves forced at the bottom. The wave with the phase velocity (either ω/k or −ω/k) with the same direction of vertical shear of the mean zonal flow is separated just below the level of the phase velocity approaching (but somewhat slower than) the mean zonal velocity, because vertical wavelength and group velocity become much smaller than the other (opposite direction) wave and small enough to cause local instability (“breaking”) and/or radiative dissipation (“damping”). Thus, the mean zonal shear flow at somewhat lower level (somewhat slower than the wave) absorbs the wave momentum, and the mean zonal shear flow descends: 1 ∂ ∂U =− (wave momentum flux) ∂t ρ ∂z

8

M. D. Yamanaka and S.-Y. Ogino

Fig. 5 Same as Fig. 4 but enlarged for the stratosphere [36]

Fig. 6 Schematic plots of the surface temperature (black) for each local time (LT) near a coast (left: sea; right: land), consisting of seaward (red) and landward (blue) gravity waves and global tide (gray). The sea-surface temperature is almost constant (actually varying slightly with interannual variations). The land-surface temperature varies largely by daytime solar heating and nighttime cooling (mainly by sprinkler-like effect of rainfall; see, e.g., [34])

Indonesian Coastlines Controlling the Whole Earth’s Atmosphere

=−

∂U 1 ∂ ∂U ≡ −W , (wave momentum flux) ρ ∂U ∂z ∂z

9

(2)

where U is the mean zonal velocity, ρ is the mean density, and the wave momentum flux is given by ρ × product of zonal and vertical velocity anomalies and is ∝ 1/(zonal phase velocity−U) (see e.g., Section 4.6 of Andrews et al., 1987). The advection ∂ velocity W = ρ1 ∂U (wave momentum flux) = (wave momentum flux)/(zonal phase velocity−U) is always negative, which gives descending velocity of the QBO pattern. The zonal phase speed |ω/k|/cos α (dependent on the coastline direction α oblique to the meridian) restricts the zonal mean flow amplitude of QBO, and the wave momentum flux controls |W| and the shortness of QBO period (as so far known). Interannual and intraseasonal variations may modulate CDC-SLBC amplitudes, wave momentum flux and then QBO. Waves associated with CDC-SLBC also have the meridional and vertical components, which may be related to the Brewer–Dobson circulation (e.g., [6, 11]). The other (opposite-direction) wave is propagated much higher up to the oppositedirection shear region of the mean zonal flow, otherwise up to a self-breaking level due to amplification compensating atmospheric density decrease, and is absorbed in a similar way. The diurnal tide should be self-breaking just below the turbopause [22], and this should determine the turbopause (homopause for the atmospheric constituents) level [16, 33]. Some results using Plumb’s (1977) model have been presented in the INCREASE symposium last year [35]. The general dispersion relation of CDC-SLBC waves in viscous atmosphere on the surface for each of bidirectional modes is a cubic equation for vertical wavenumber squared, of which a steady case with large eddy viscosity/diffusivity corresponds to a local circulation convection such as a heat island [13, 31]. When the atmospheric density stratification (scale height) is stronger than any damping, atmospheric motions (circulation layers or waves) appear also above the circulation at the bottom. In case of weaker viscosity/diffusivity/damping with a mean flow apart from the surface, three roots forced at the top (amplifying upward) are omitted, and two roots are degenerated as amplitude damping of the remaining one root under so-called WKB-approximation, which is the solution used by Plumb (1977). The period and structure of calculated “QBO” may be dependent on the number and location of coastlines, that is, the distribution of lands. For example, if coastlines are less (more), the period of “QBO” becomes longer (shorter). IMC with the longest coastline is most responsible to produce QBO. Because CDC-SLBC is quite robust as seen in Fig. 1, QBO generated by bidirectional waves separated from SLBC must be robust, which is different from foregoing discussions on generation of QBO with broadband waves (e.g., [3]).

10

M. D. Yamanaka and S.-Y. Ogino

4 Conclusions In this paper, we show observational evidence that the equatorial tropospheric CDCSLBC plays a principal role to maintain the stratospheric QBO robustly and explain it from theories of SLBC and the wave-mean-flow interaction. The coastlines are the triple boundary among land, sea, and air constructing the Earth system, and our present results suggest even the QBO is not occurred incidentally but is controlled under the Earth system. As in the distributions of surface rainfall and tropospheric convective cloud activity, the strongest CDC along the world’s longest coastline surrounding major islands of IMC is most responsible in the generation of bidirectional (sea- and land-ward) gravity waves interacting with QBO. The longer equatorial (Kelvin and mixed Rossby-gravity) waves induced by a larger-scale oceancontinent or open-ocean heating contrast [5, 20] may modulate (cluster) the bidirectional gravity waves in the stratosphere, just as land distributions and some intraseasonal variations do in the troposphere [7]. This Matsuno-Gill heating is equivalent to meridional gradient of zonal forcing, and zonally uniform case makes Hadley circulation and subtropical trade wind zones [15]. The other non-annual variations such as semiannual oscillations (SAOs) and mesospheric QBO may be considered by CDC-SLBC bidirectional waves similarly. In the extratropics the orographic (stationary) internal gravity waves robustize the middle-stratospheric weak wind layers [32], which permits both east- and westward traveling gravity waves propagating upward and producing the reversal of the middleatmospheric zonal circulation and winter-poleward meridional circulation (balanced with annually-reversed both hemispheric difference of radiational budget) at the mesopause [10, 18, 21]. Both the extratropical orography and the equatorial coastlines have been changed in the long history of Earth, and the paleoclimate may be discussed as applications of those gravity-wave processes. The robust 1-day periodicity of equatorial CDC is another effect of Earth’s rotation like the extratropical Coriolis force. As Himalaya-Tibet in the mid-latitudes, IMC in the tropics is the geologically most important product of plate tectonics in Mesozoic-Cenozoic (recent 100 million years). They govern the global atmosphere, through the process mentioned in this paper for the equatorial lower-to-upper atmosphere. Therefore, more observational and theoretical understanding of CDC-SLBC may contribute to further development of atmospheric sciences in Indonesia. Acknowledgements This study has been supported partly by JSPS KAKENHI Grant Number JP19H04248.

Indonesian Coastlines Controlling the Whole Earth’s Atmosphere

11

References 1. Anstey, J.A., Osprey, S.M., Alexander, J., Baldwin, M.P., Butchart, N., Gray, L., Kawatani, Y., Newman, P.A., Richter, J.H.: Impacts, processes and projections of the quasi-biennial oscillation. Nat. Rev. Earth Environ. 3, 588–603 (2022). https://doi.org/10.1038/s43017-022-003 23-7 2. Baldwin, M.P., Gray, L.J., Dunkerton, T.J., Hamilton, K., Haynes, P.H., Randel, W.J., Holton, J.R., Alexander, M.J., Hirota, I., Horinouchi, T., Jones, D.B.A., Kinnersley, J.S., Marquardt, C., Sato, K., Takahashi, M.: The quasi-biennial oscillation. Rev. Geophys. 39, 179–229 (2001). https://doi.org/10.1029/1999RG000073 3. Dunkerton, T.J.: The role of gravity waves in the quasi-biennial oscillation. J. Geophys. Res. 102, 26053–26076 (1997). https://doi.org/10.1029/96JD02999 4. Ebdon, R.A.: Notes on the wind flow at 50 mb in tropical and sub-tropical regions in January 1957 and January 1958. Quart. J. Roy. Meteor. Soc. 86, 540–542 (1960). https://doi.org/10. 1002/qj.49708637011 5. Gill, A.E.: Some simple solutions for heat-induced tropical circulation. Quart. J. Roy. Meteor. Soc. 106, 447–462 (1980). https://doi.org/10.1002/qj.49710644905 6. Hasebe, F., Aoki, S., Morimoto, S., Inai, Y., Nakazawa, T., Sugawara, S., Ikeda, C., Honda, H., Yamazaki, H., Komala, N., Putri, F.A., Budiyono, A., Soedjarwo, M., Ishidoya, S., Toyoda, S., Shibata, T., Hayashi, M., Eguchi, N., Nishi, N., Fujiwara, M., Ogino, S.-Y., Shiotani, M., Sugidachi, T.: Coordinated upper-troposphere-to-stratosphere balloon experiment in Biak. Bull. Amer. Meteor. Soc. 99, 1213–1230 (2018). https://doi.org/10.1175/BAMS-D-16-0289.1 7. Hayashi, Y.-Y., Sumi, A.: The 30–40 day oscillations simulated in an “aqua planet” model. J. Meteor. Soc. Jpn. 64, 451–467 (1986). https://doi.org/10.2151/jmsj1965.64.4_451 8. Haynes, P., Hitchcock, P., Hitchman, M., Yoden, S., Hendon, H., Kiladis, G., Kodera, K., Simpson, I.: The influence of the stratosphere on the tropical troposphere. J. Meteor. Soc. Jpn. 99, 803–845 (2021). https://doi.org/10.2151/jmsj.2021-040 9. Hitchman, M.H., Yoden, S., Haynes, P.H., Kumar, V., Tegtmeier, S.: An observational history of the direct influence of the stratospheric quasi-biennial oscillation on the tropical and subtropical upper troposphere and lower stratosphere. J. Meteor. Soc. Jpn. 99, 239–267 (2021). https://doi. org/10.2151/jmsj.2021-012 10. Holton, J.R.: The role of gravity wave induced drag and diffusion in the momentum budget of the mesosphere. J. Atmos. Sci. 39, 791–799 (1982). https://doi.org/10.1175/1520-0469(198 2)039%3c0791:TROGWI%3e2.0.CO;2 11. Kawatani, Y., Hamilton, K.: Weakened stratospheric quasibiennial oscillation driven by increased tropical mean upwelling. Nature 497, 478–481 (2013). https://doi.org/10.1038/nat ure12140 12. Kawatani, Y., Sato, K., Dunkerton, T.J., Watanabe, S., Miyahara, S., Takahashi, M.: The roles of equatorial trapped waves and internal inertia-gravity waves in driving the quasi-biennial oscillation, Part I: zonal mean wave forcing. J. Atmos. Sci. 67, 963–980 (2010). https://doi. org/10.1175/2009JAS3222.1 13. Kimura, R.: Dynamics of steady convections over heat and cool islands. J. Meteor. Soc. Jpn. 53, 440–457 (1975). https://doi.org/10.2151/jmsj1965.53.6_440 14. Kobayashi, S., Ota, Y., Harada, Y., Ebita, A., Moriya, M., Onoda, H., Onogi, K., Kamahori, H., Kobayashi, C., Endo, H., Miyaoka, K., Takahashi, K.: The JRA-55 reanalysis: general specifications and basic characteristics. J. Meteor. Soc. Jpn. 93, 5–48 (2015). https://doi.org/ 10.2151/jmsj.2015-001 15. Kosaka, Y., Matsuda, Y.: Roles of Rossby and gravity waves on circulation associated with tropical and subtropical heating. J. Meteor. Soc. Jpn. 83, 481–498 (2005). https://doi.org/10. 2151/jmsj.83.481 16. Leovy, C.B.: Control of the homopause level. Icarus 50, 311–321 (1982). https://doi.org/10. 1016/0019-1035(82)90128-2 17. Lin, Y.-L.: Mesoscale dynamics. Cambridge University Press, pp. 630, ISBN-978-0-52180875-0 (2007)

12

M. D. Yamanaka and S.-Y. Ogino

18. Lindzen, R.S.: Turbulence and stress owing to gravity wave and tidal breakdown. J. Geophys. Res. 86, 9707–9714 (1981). https://doi.org/10.1029/JC086iC10p09707 19. Lindzen, R.S., Holton, J.R.: A theory of the quasi-biennial oscillation. J. Atmos. Sci. 25, 1095–1107 (1968). https://doi.org/10.1175/1520-0469(1968)025%3c1095:ATOTQB%3e2.0. CO;2 20. Matsuno, T.: Quasi-geostrophic motions in the equatorial area. J. Meteor. Soc. Jpn. 44, 25–43 (1966). https://doi.org/10.2151/jmsj1965.44.1_25 21. Matsuno, T.: A quasi one-dimensional model of the middle atmosphere circulation interacting with internal gravity waves. J. Meteor. Soc. Jpn. 60, 215–226 (1982). https://doi.org/10.2151/ jmsj1965.60.1_215 22. Miyahara, S.: Zonal mean winds induced by vertically propagating atmospheric tidal waves in the lower thermosphere. J. Meteor. Soc. Jpn. 56, 86–97 (1978). https://doi.org/10.2151/jms j1965.56.2_86 23. Ogino, S.Y., Yamanaka, M.D., Mori, S., Matsumoto, J.: Tropical coastal dehydrator in global atmospheric water circulation. Geophys. Res. Lett. 44(1), 636–11643 (2017). https://doi.org/ 10.1002/2017GL075760 24. Osprey, S.M., Butchart, N., Knight, J.R., Scaife, A.A., Hamilton, K., Anstey, J.A., Schenzinger, V., Zhang, C.: An unexpected disruption of the atmospheric quasi-biennial oscillation. Science 353, 1424–1427 (2016). https://doi.org/10.1126/science.aah4156 25. Plumb, R.A.: The interaction of two internal waves with the mean flow: implications for the theory of the quasi-biennial oscillation. J. Atmos. Sci. 34, 1847–1858 (1977). https://doi.org/ 10.1175/1520-0469(1977)0342.0.CO;2 26. Plumb, R.A., McEwan, A.D.: The instability of a forced standing wave in a viscous stratified fluid: a laboratory analogue of the quasi-biennial oscillation. J. Atmos. Sci. 35, 1827–1839 (1978). https://doi.org/10.1175/1520-0469(1978)035%3c1827:TIOAFS%3e2.0.CO;2 27. Qian, J.H.: Why precipitation is mostly concentrated over islands in the maritime continent. J. Atmos. Sci. 65, 1428–1441 (2008). https://doi.org/10.1175/2007JAS2422.1 28. Reed, R.J., Campbell, W.J., Rasmussen, L.A., Rogers, D.G.: Evidence of a downwardpropagating, annual wind reversal in the equatorial stratosphere. J. Geophys. Res. 66, 813–818 (1961). https://doi.org/10.1029/JZ066i003p00813 29. Rotunno, R.: On the linear theory of the land and sea breeze. J. Atmos. Sci. 40, 1999–2009 (1983). https://doi.org/10.1175/1520-0469(1983)040%3c1999:OTLTOT%3e2.0.CO;2 30. Scorer, R.: Theory of waves in the lee of mountains. Quart. J. Roy. Meteor. Soc. 75, 41–56 (1949). https://doi.org/10.1002/qj.49707532308 31. Stommel, H., Veronis, G.: Steady convective motion in a horizontal layer of fluid heated uniformly from above and cooled non-uniformly from below. Tellus 9, 401–407 (1957). https:// doi.org/10.3402/tellusa.v9i3.9100 32. Tanaka, H., Yamanaka, M.D.: Atmospheric circulation in the lower stratosphere induced by the mesoscale mountain wave breakdown. J. Meteor. Soc. Jpn. 63, 1047–1054 (1985). https:// doi.org/10.2151/jmsj1965.63.6_1047 33. Yamanaka, M.D.: Homopause control by gravity wave breaking in the planetary atmospheres. Adv. Space Res. 15(4), 47–50 (1995). https://doi.org/10.1016/0273-1177(94)00063-7 34. Yamanaka, M.D.: Physical climatology of Indonesian maritime continent: an outline to comprehend observational studies. Atmos. Res. 178–179, 231–259 (2016). https://doi.org/10.1016/j. atmosres.2016.03.017 35. Yamanaka, M.D.: EAR construction motivation revisited: Indonesian coastline representing earth. An invited talk at EAR 20-Year Anniversary—INCREASE 2021, 20–21 Sept 2021. (to be printed in this volume) (2021) 36. Yamanaka, M.D., Ogino, S.-Y.: Tropical coastlines robustizing the stratospheric quasi-biennial oscillation. (in preparation) (2023) 37. Yamanaka, M.D., Ogino, S.-Y., Wu, P.-M., Hamada, J.-I., Mori, S., Matsumoto, J., Syamsudin, F.: Maritime continent coastlines controlling Earth’s climate. Prog. Earth Planet Sci. 5(21), 1–28 (2018). https://doi.org/10.1186/s40645-018-0174-9

EAR Construction Motivation Revisited: Indonesian Coastline Representing Earth Manabu D. Yamanaka

Abstract EAR was constructed to target the vertical coupling at a single station. The equatorial region with maximum insolation and moisture in the lower atmosphere was less observed and less understood because of its ageostrophic, sub-synoptic clouddominant features. For the middle and upper atmosphere, vertically-temporally highresolution radar sounding was much anticipated to clarify waves generating stratospheric quasi-biennial oscillation (QBO), for example. During these two decades after the construction of EAR, the significant role of the Indonesian maritime continent (IMC) for the global lower-atmospheric climate has been recognized. Many “mini-EARs” (wind profilers and weather radars), as well as satellites and cloudresolving model, have revealed that the diurnal cycle (sea-land breeze circulation) along the world’s longest coastline surrounding major islands of IMC is the most robust mode of cloud-rainfall generation. This process produces latent heat (amount of roughly 10–20% of the greenhouse effect) compensating the global radiationaldynamical energy imbalance, and causes floods in stronger land-sea temperature gradients during remarkable monsoon (cold surge) and/or La Niña periods. Any anthropogenic modification of the land surface-hydrologic conditions such as urbanization and cultivation also may change the diurnal cycle, local weather, and also the global climate. The diurnal-cycle circulation consists of equi-amplitude bidirectional (sea- and land-ward) internal gravity waves likely to cause QBO. Indeed, the coastline of IMC is the triple boundary maintained among land, sea, and atmosphere of the earth, and the EAR and many mini-EARs listen to sounds of the whole earth system.

A paper presented at INCREASE 2021 M. D. Yamanaka (B) Research Institute for Humanity and Nature, Kyoto, Japan e-mail: [email protected]; [email protected] Professor Emeritus, Kobe University, Kobe, Japan © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_2

13

14

M. D. Yamanaka

1 Introduction It was 32 years ago when Prof. Susumu Kato and Prof. Shoichiro Fukao of Kyoto University, Pak Alex Sudibyo of LAPAN, Pak Chandra Manan Mangan and Bu Tien Sribimawati of BPPT, and others including myself climbed Bukit Kototabang for the first time. It was 4 years after Kyoto University and LAPAN started site surveys for “Equatorial Radar” in Biak, followed in Pontianak, and was among more than ten locations in Sumatera Barat (Fig. 1). Considering ground conditions for construction of a large VHF radar, infrastructure, and support for operation, and accessibility from Jakarta and overseas, Kototabang was selected [25]. After very quick permission of Prof. Bacharuddin Jusuf Habibie, followed soon by BMKG’s GAW station, we needed 12 years until constructing the EAR in 2001. On the first proposal of “Equatorial Radar” construction [25], establishment of “International Center for Equatorial Research (ICEAR)” was also planned for (i) effective management of international observational activities and data archives, (ii) development and integration of ground-based profiling techniques including various types of radars, and (iii) construction of a prototype of new-generation atmospheric science. Prof. Fukao considered (iii) based on (ii), that is, “radar atmosphere physics” [5–7]. Atmospheric radars providing temporally-vertically high-resolution observations almost automatically obtained innovative results for the extratropical middle and upper atmosphere in 1980s and were much anticipated to do so in the tropics with more special features and less observations (Fig. 2). This dream came true by EAR for these two decades, as many other papers presented in this symposium. In this paper, the background before the EAR construction and the progress of tropical loweratmospheric observations in parallel with the EAR observations are overviewed.

Fig. 1 a Professor Susumu Kato (1928–2020) and Professor Shoichiro Fukao (1943–2014) on a site survey at Bukit Kototabang in June 1989. Candidate site maps b during and c after site surveys. d Map of Bukit Kototabang [25]

EAR Construction Motivation Revisited: Indonesian Coastline …

15

Fig. 2 Schematic a meridional-vertical cross-section, and equatorial b vertical and c horizontal structures (Overview by Fukao, in [25])

2 Result and Discussion 2.1 Middle/Upper-Atmospheric Motivation of EAR Before Construction The first use of the name Sumatera occurred in 1017 as Samudra, which meant “sealand” [31], that is, the maritime continent! “Great voyage” re-discovered trade wind, monsoon, and diurnal cycles [49], reprinted report of de Houtman’s first arrival in Enggano and Banten in 1596, [9] which were discovered originally by ancient Greek and Arab sailors. Before WW2, the IMC was one of the densest meteorological observation regions over the world (Fig. 3). Nederlands-Indië (Hindia Belanda) installed more than 2000 rain gauge stations in 19C [22, 26, 30] which were used by the first climatological classification by Köppen [27]. Telegram network was first established by Lord Kelvin during 1866–70, which communicated the 1883 eruption of Krakatau (misspelled as “Krakatoa”) over the world within a couple of days [65]. The volcanic ashes went around the equatorial stratosphere westward for two weeks, suggesting easterly of about 30 m/s [13, 14]. At Batavia (Jakarta), van Bemmelen [61, 62] did many balloon observations to confirm high tropopause, both easterly and westerly in the stratosphere, tropospheric monsoons, and diurnal cycles of sea-land breeze circulations. Operational meteorological observations were continued under Japanese occupation (1942–45) [69] and were partly supported by Japanese war reparation (1958– 70) after independence of Indonesia. They were organized as a national institute (Lembaga MG) in 1955, an agency (Badan MG) in 1980, and the present BMKG in 2008. Middle/upper-atmospheric (large balloon and rocket) and space (ionospheric and satellite) observations were started by LAPAN established in 1963, which

16

M. D. Yamanaka

Fig. 3 Parts of first global maps of a surface wind distribution [9] and b climatological classification [27]. c Annual rain amount map for Sumatera [30], and vertical distributions of d local time-mean wind [61, 62] and e wind direction histograms for rainy and dry seasons [61, 62] at Batavia

becomes an organization under BRIN in 2021, together with other agencies such as BPPT and LIPI. The middle atmosphere (stratosphere and mesosphere) is interdisciplinary between meteorology and upper atmospheric/space physics, and its dynamics have been understood as the wave-mean flow interaction during 1970–80s. The extratropical mesopausal zonal-wind reversal is explained by east/westward gravity waves propagating from the troposphere [29, 32], which results from stationary (orographic) gravity waves breaking completely in the middle-stratospheric weak wind [58] (Fig. 4). The equatorial-stratospheric quasi-biennial oscillation (QBO) has been explained by equi-amplitude zonally-bidirectional equatorial waves (e.g., [21, 44], but what waves and how they are generated are still controversial [3, 20]. Probably, gravity waves contribute greatly, and are observed by sufficiently high-resolution observations. Balloons have been operated by LAPAN at Watukosek, Jawa Timur (Fig. 4c), but their time resolution (every three hours, i.e., eight times per day at maximum) and vertical coverage (up to the stratosphere and often below the tropopause) are limited (e.g., [59, 60]). Therefore, the atmospheric radars are keenly anticipated to clarify these unsolved wave issues at least, as succeeded already for the extratropical middle-atmosphere.

EAR Construction Motivation Revisited: Indonesian Coastline …

17

Fig. 4 a Meridional-vertical distribution of zonal wind in January (CIRA, 1986) [47], and b breaking levels of internal gravity waves [58]. c Balloon observations and d zonal wind vertical profiles at Watukosek, Jawa Timur [72]

2.2 Climatological Importance of IMC Recognized in These Two Decades During these two decades, the significant role of the Indonesian maritime continent (IMC) for the global lower-atmospheric climate has been recognized. Interannual (ENSO, IOD), annual (monsoons, rainy/dry seasons), and intraseasonal (MJO) variations have been known (e.g., [4, 10–12, 17–19, 28, 37, 38, 42, 54, 63, 64, 71, 74], but the most dominant mode is the diurnal cycle [71, 74]. GMS cloud data clearly show that intraseasonal variations dominant over the ocean are replaced by the diurnal cycle near the coastline (left two panels of Fig. 5). Diurnal cycle as the tropospheric coastal most dominant mode The diurnal cycle along the world’s longest coastline surrounding many islands of IMC is the most robust feature of cloud-rainfall generation [1, 2, 17, 18, 23, 33–37, 50–52, 55–57, 66–68]. It is associated with the updraft (surface convergence) zone of a sea-land breeze circulation. Figure 6 shows a long-time average of morning/evening surface wind obtained from so-called “objective analysis” (a numerical model calculation with integrating dynamical equations based on available observational data) calculated by Japan Meteorological Agency (JMA) almost two decades ago. At that time both observational and modeling capabilities were rather limited, but the diurnal cycle of sea-land breeze was clearly seen, because of the diurnal cycle itself is a very basic mode of atmosphere. Recent analysis is done much more accurately, and the diurnal cycle appears as observed without averaging (see, e.g., [46]).

18

M. D. Yamanaka

Fig. 5 Intraseasonal and diurnal dominance over open ocean and near coast, respectively, (left two local time-pentad diagrams; [11], and diurnal march along five cross-sections (right-hand side five Hovmöller (local time-longitude) diagrams, [52] of cloud activity over Sumatera

Fig. 6 Examples of long-term (one year of 2001) average of objective analysis data of Japan Meteorological Agency (unpublished MS Thesis work in 2003)

EAR Construction Motivation Revisited: Indonesian Coastline …

19

Radar and satellite observations as well as numerical models (Figs. 5, 6 and 7) have shown that the characteristic horizontal scale of the diurnal cycle zone around a coastline is of the order of some hundreds of kilometers, which is somewhat broader than in the extratropics. In Sumatera (Fig. 5) active rainclouds are generated near the central (Barisan) mountains around sunset, migrated both coastward, and arrived offshore before sunrise. In the northern (inland sea) coast of Jawa (Fig. 7), morning seaside to evening mountainside and opposite routes have been observed. A smaller island in the open-ocean side of a large island has two (evening and morning) rainfall peaks [12, 57, 68] and those in the inland (Karimata) sea re-excite the diurnal-cycle migration between two large (Kalimantan and Sumatera) islands [2]. Observational evidence of the diurnal cycle mentioned above is obtained mainly by weather radars as horizontal raindrop distributions. Profilers including EAR and WPRs may provide vertical wind profiles including vertical components and clear areas without raindrops. Complementary uses of both types of radars, as well as collaborations with high-resolution numerical models, are anticipated. Energy/Water cycles controlled by coastal diurnal cycle The diurnal cycle appears also in the river water [55], and may cause floods in stronger land-sea temperature gradients during remarkable monsoon (cold surge) and/or La Niña periods. Such applications are fruitful for disaster prevention.

Fig. 7 a Diurnal cycle zones along IMC coastlines are shown as AM–PM differences (worm and cold colors indicate positive and negative) of TRMM rainfall [35]. b Surface/radiosonde network (●) during 16 Jan–15 Feb 2010 around a C-band radar (CDR) (+) providing c Hovmöller (latitude-local time) diagram of diurnal-cycle rainfall over west Jawa [33]

20

M. D. Yamanaka

Fig. 8 a Brewer-Dobson circulation [16] and b surface water budget [15] in the meridional plot. c Vertical velocity and d precipitation (and population) referenced to coastline (modified from [40, 41])

The annual rainfall amount in the IMC coastal regions is about 2500 mm/year, which is about 2.5 times of the global mean amount 1000 mm/year corresponding to latent heat release 80 W/m2 (about 1/3 of the greenhouse effect). Thus, the diurnal cycle (producing about 2/3 of the total rainfall amount, or roughly 10–20% of the greenhouse effect) compensates the global radiational-dynamical energy imbalance [74]. Concerning the water budget, rainwater of 1 × 1014 m3 falling annually over the IMC coastal region is about 1/3 of the tropical total and about 1/6 of the global total [74]. Thus the global atmospheric water transport (so-called Brewer-Dobson circulation or “cold trap” mechanism) concentrated in the tropics in the zonal-mean image (Fig. 8a, b) is more localized near the coastal regions (Fig. 8c, d) in particular of IMC [40, 41]. EAR in the mountains of Sumatera is included in the coastal region mentioned here, and may observe just the global water and atmosphere circulations.

2.3 Momentum Budget and Vertical Coupling The diurnal-cycle sea-land breeze circulation consists of horizontally (sea- and landward) and vertically (upward) propagating internal gravity waves (Fig. 9) (e.g., [48]), which is a different category from the global tides [24]. In the lower troposphere,

EAR Construction Motivation Revisited: Indonesian Coastline …

21

both up/downward propagating modes make a circulation cell satisfying zero vertical velocity at the ground, and large viscosity makes it almost steady and reversed between day and night. Above the cell, waves are almost inviscid and propagating sea- or land-ward, and the top boundary (radiative) condition selects only upward mode which is amplified with upward-decreasing density to satisfy energy (action) conservation. Note that shorter waves corresponding to individual clouds cannot propagate upward (i.e., trapped and ducting near the ground), which is just the same as Scorer’s (1949) [53] cutoff for stationary (mountain) waves. The internal gravity waves generated above the diurnal-cycle sea-land breeze circulation cell are equi-amplitude horizontally-bidirectional waves, which are likely to cause the quasi-biennial oscillation (QBO) [43–45] as the most robust mode in the stratosphere. It is reasonable that the meridional range of QBO is narrower than the internal (vertical propagation) region of inertio-gravity waves because the solar diurnal cycle is larger than the annual cycle within 10º-latitudes similar to the meridional range of the IMC. The horizontal propagation direction are from mountains in the evening with destination to the maritime destination in the morning, or from seas before sunrise to the mountainous destination in the evening. The horizontal phase velocity is 100 km/half a day = 10 km/h = 3 m/s on land, and faster off the coast, and its zonal component becomes much master if the coastline is inclined from the meridional direction.

Fig. 9 Gravity wave solutions for a stratosphere with the radiative top boundary condition and b troposphere with the rigid bottom boundary condition

22

M. D. Yamanaka

The wave amplitude is dependent on the sea-land temperature contrast, which is modified by climatological sea-surface variations (ENSO, IOD, and MJO) and anthropogenic land-surface modifications. The period of QBO decreases with an increase of the wave amplitudes, and becomes biperiodic for too strong amplitudes (Fig. 9). The QBO period becomes longer for a weaker diurnal cycle, and decayed if too weak. Longitudinal locations of Africa (−90°) and South America (+180°) and rapid rotation of the earth homogenize the local solar heating zonally. The sea-land breeze circulations have been observed successfully by singlestationed lower-atmospheric UHF radars (WPRs) [1, 8, 17, 18, 56]. Connecting them with upper-tropospheric/stratospheric internal gravity waves observed by larger VHF radars such as EAR is strongly expected (Fig. 10).

2.4 Recent Whole IMC Tropospheric Observations Many “mini-EARs” (wind profilers and weather radars at BMKG stations) have been installed and operated successfully over IMC (Figs. 11, 12 and 13). Scientific projects become multi-national, multi-stationed, multi-instrumented and/or multi-disciplinary, and the BMKG operational radar network covers almost all the Indonesian territory. Collaborations and contributions of G20 (G7 + EU + China, India, Brazil, Russia, Korea, Australia, Mexico, Indonesia, Saudi Arabia, Turkey, Argentina, and South Africa) covering our planet more densely are strongly anticipated. In Indonesia National Research and Innovation Agency (BRIN) has been established in 2020, and integrates various scientific activities covered by BPPT, LIPI, and LAPAN, which is noted internationally. As an example of multi-disciplinary approach, any modification of the land surface-hydrologic conditions such as urbanization and plantationization may change the diurnal cycle, local weather, and also the global climate. Radars have been developed for long years in the extratropical countries with cooler and drier climate and cleaner environment. This is one reason why radar operations are not so easy in the tropics such as IMC. For example, the HARIMAU C-band radar was damaged by a connector shorting due to humid dust in 2014, and finally destroyed by circuit damages due to thunderbolt current in 2015. Establishment of Indonesian radar makers and development of Indonesian radars are strongly anticipated. New applications of radars in particular for special targets in the tropics are also expected. For example mixed layer top detection initiated by an L-band radar at Serpong [17, 18] may be applied also by a C-band weather radar for watching smog caused by wildfire [46]. The future (Equatorial MU Radar) project as the next step of EAR [70] is also expected in the common directions mentioned here.

EAR Construction Motivation Revisited: Indonesian Coastline …

23

Fig. 10 Plumb’s [44] QBO analogue applied for internal gravity waves generated by diurnalcycle sea-land breeze circulations

24

M. D. Yamanaka

Fig. 11 The JEPP/HARIMAU-SATREPS/MCCOE radar-profiler-buoy networks (see [73, 74] for details)

3 Conclusion Successful twenty years of EAR proved correctness of the initial motivation of EAR construction. The geographical blank of the middle atmosphere, in particular the equatorial region, before construction is covered now vertically by EAR, and horizontally by satellites. Operational meteorological observations by BMKG have been promoted greatly (more than 40 weather radars, more than 20 radiosonde stations, etc.) and their data are opened on the internet. Those observational progresses lead to improvement of the global objective analyses, and they may complement the vertical observation by EAR. However, the dynamical mystery on equi-amplitude zonallybidirectional wave sources in particular for the QBO generation is still controversial. Similarly, relationships or interactions among interannual (tropospheric ENSO, IOD, global warming, stratospheric QBO, etc.), seasonal (monsoons), and intraseasonal (MJO) variations have not yet been solved completely. This is serious also for disaster prevention, because superimposition of these phenomena may induce extreme phenomena (flood or drought). The dominance of diurnal cycles along coastlines of IMC may be a key also for the unsolved issues mentioned above [71, 74]. A sea or land breeze circulation cell (of which the updraft portion is associated with active convective clouds) is equivalent to the sum of a pair of sea- and land-ward propagating internal gravity waves satisfying zero velocity at the ground. Those waves resemble the simple two-wave model of

EAR Construction Motivation Revisited: Indonesian Coastline …

25

Fig. 12 YMC [75], RIHN Tropical Peatland Society Project [39], and BMKG C-band weather radar network

QBO by Plumb (1977) and propagate also vertically satisfying the energy conservation through the upward decreasing density and the radiative condition at the top. The cell intensity and wave amplitude are proportional to the land-sea surface temperature gradient, modified with the interannual, annual, and intraseasonal variations, as observed. Indeed, the coastline of IMC is the triple boundary maintained among land, sea, and atmosphere of the earth, and the EAR and many mini-EARs listen to sounds of the whole earth system. The coastal region has a large gradient of physical quantities such as temperature, ground viscosity, and mountain orography vs flat sea surface. The biosphere and its diversity are maintained to fit/relax this gradient, and the anthroposphere (including megacities, peatland plantations, etc.) is made inside this biosphere. Therefore, more utilization/application of radars in particular coastal regions (some of which are not covered by the BMKG radars installed mainly near the major airports) may be important for the future of Indonesia.

26

M. D. Yamanaka

Fig. 13 Radar reflectivity-frequency diagram for raindrops (labeled by Z) and turbulence layers (by Cn2 ) (top) [7], radars used by our collaborative studies in Indonesia (middle), and the mixed layer top detected by Palangkaraya BMKG C-band weather radar (bottom) [46]

EAR Construction Motivation Revisited: Indonesian Coastline …

27

Acknowledgements This study has been supported partly by JSPS KAKENHI Grant Number JP19H04248.

References 1. Araki, R., Yamanaka, M.D., Murata, F., Hashiguchi, H., Oku, Y., Sribimawati, T., Kudsy, M., Renggono, F.: Seasonal and interannual variations of diurnal cycles of local circulation and cloud activity observed at Serpong, West Jawa, Indonesia. J. Meteor. Soc. Jpn. 84A, 171–194 (2006). https://doi.org/10.2151/jmsj.84A.171 2. As-syakur, A.R., Imaoka, K., Ogawara, K., Yamanaka, M.D., Tanaka, T., Kashino, Y., Nuarsa, I.W., Osawa, T.: Analysis of spatial and seasonal differences in the diurnal rainfall cycle over Sumatera revealed by 17-year TRMM 3B42 dataset. SOLA, 15, 216–221 (2019). https://doi. org/10.2151/sola.2019-039 3. Baldwin, M.P., Gray, L.J., Dunkerton, T.J., Hamilton, K., Haynes, P.H., Randel, W.J., Holton, J.R., Alexander, M.J., Hirota, I., Horinouchi, T., Jones, D.B.A., Kinnersley, J.S., Marquardt, C., Sato, K., Takahashi, M.: The quasi-biennial oscillation. Rev. Geophys. 39, 179–229 (2001). https://doi.org/10.1029/1999RG000073 4. Fujiwara, M., Tomikawa, Y., Komala, N., Saraspriya, S., Manik, T., Suripto, A., Kita, K., Kondo, Y., Kawakami, S., Ogawa, T., Kelana, E., Suhardi, B., Sribimawati, T., Kudsy, M., Harijono, S.W.B., Yamanaka, M.D.: Ozonesonde soundings in the Indonesian maritime continent in September-October 1998 and in August-September 1999. Atmos. Environ. 37, 353–362 (2003). https://doi.org/10.1016/S1352-2310(02)00891-9 5. Fukao, S.: Coupling processes in the equatorial atmosphere (CPEA): a project overview. J. Meteor. Soc. Jpn. 84A, 1–18 (2006). https://doi.org/10.2151/jmsj.84A.1 6. Fukao, S.: Recent advances in atmospheric radar study. J. Meteor. Soc. Jpn. 85B, 215–239 (2007). https://doi.org/10.2151/jmsj.85B.215 7. Fukao, S., Hamazu, K.: Radar for Meteorological and Atmospheric Observations, pp. 537. Springer (2014) 8. Hadi, T.W., Horinouchi, T., Tsuda, T., Hashiguchi, H., Fukao, S.: Sea-breeze circulation over Jakarta, Indonesia: a climatology based on boundary layer radar observations. Mon. Wea. Rev. 130, 2153–2166 (2002). https://doi.org/10.1175/1520-0493(2002)1302. 0.CO;2 9. Halley, E.: An historical account of the trade winds, and monsoons, observable in the seas between and near the Tropicks, with an attempt to assign the physical cause of the said winds. Phil. Trans. 16, 153–168 (1686). https://doi.org/10.1098/rstl.1686.0026 10. Hamada, J.-I., Yamanaka, M.D., Matsumoto, J., Fukao, S., Winarso, P.A., Sribimawati, T.: Spatial and temporal variations of the rainy season over Indonesia and their link to ENSO. J. Meteor. Soc. Jpn. 80, 285–310 (2002). https://doi.org/10.2151/jmsj.80.285 11. Hamada J.-I., Yamanaka, M.D., Mori, S., Tauhid, Y.I., Sribimawati, T.: Differences of rainfall characteristics between coastal and interior areas of central western Sumatera, Indonesia. J. Meteor. Soc. Jpn. 86, 593–611 (2008). https://doi.org/10.2151/jmsj.86.593 12. Hamada J.-I., Mori, S., Yamanaka, M.D., Haryoko, U., Lestari, S., Sulistyowati, R., Syamsudin, F.: Interannual rainfall variability over northwestern Jawa and its relation to the Indian Ocean dipole and El Niño southern-oscillation events. SOLA, 8, 69–72 (2012). https://doi.org/10. 2151/sola.2012-018 13. Hamilton, K.: Observations of tropical stratospheric winds before World War II. Bull. Amer. Meteor. Soc. 79, 1367–1371 (1998). https://doi.org/10.1175/1520-0477(1998)0792.0.CO;2 14. Hamilton, K.: Sereno Bishop, Rollo Russell, Bishop’s ring and the discovery of the “Krakatoa easterlies”. Atmos.-Ocean, 50, 169–175 (2012). https://doi.org/10.1080/07055900. 2011.639736

28

M. D. Yamanaka

15. Hartmann, D.L.: Global Physical Climatology, pp. 411. Academic (1994) 16. Hasebe, F., Aoki, S., Morimoto, S., Inai, Y., Nakazawa, T., Sugawara, S., Ikeda, C., Honda, H., Yamazaki, H., Komala, N., Putri, F.A., Budiyono, A., Soedjarwo, M., Ishidoya, S., Toyoda, S., Shibata, T., Hayashi, M., Eguchi, N., Nishi, N., Fujiwara, M., Ogino, S.-Y., Shiotani, M., Sugidachi, T.: Coordinated upper-troposphere-to-stratosphere balloon experiment in Biak. Bull. Amer. Meteor. Soc. 99, 1213–1230 (2018). https://doi.org/10.1175/BAMS-D-16-0289.1 17. Hashiguchi, H., Fukao, S., Tsuda, T., Yamanaka, M.D., Tobing, D.L., Sribimawati, T., Harijono, S.W.B., Wiryosumarto, H.: Observations of the planetary boundary layer over equatorial Indonesia with an L-band clear-air Doppler radar: initial results. Radio Sci. 30, 1043–1054 (1995a). https://doi.org/10.1029/95RS00653 18. Hashiguchi, H., Tsuda, T., Fukao, S., Yamanaka, M.D., Harijono, S.W.B., Wiryosumarto, H.: Boundary layer radar observations of the passage of the convection center over Serpong, Indonesia (6°S, 107°E) during the TOGA-COARE intensive observation period. J. Meteor. Soc. Jpn. 73, 535–548 (1995b). https://doi.org/10.2151/jmsj1965.73.2B_535 19. Hashiguchi, N.O., Yamanaka, M.D., Ogino, S.-Y., Shiotani, M., Sribimawati, T.: Seasonal and interannual variations of temperature in tropical tropopause layer (TTL) over Indonesia based on operational rawinsonde data during 1992–1999. J. Geophys. Res. 111, D15110 (2006). https://doi.org/10.1029/2005JD006501 20. Hitchman, M.H., Yoden, S., Haynes, P.H., Kumar, V., Tegtmeier, S.: An observational history of the direct influence of the stratospheric quasi-biennial oscillation on the tropical and subtropical upper troposphere and lower stratosphere. J. Meteor. Soc. Jpn. 99, 239–267 (2021). https://doi. org/10.2151/jmsj.2021-012 21. Holton, J.R., Lindzen, R.S.: An updated theory for the quasi-biennial cycle of the tropical stratosphere. J. Atmos. Sci. 29, 1076–1080 (1972). https://doi.org/10.1175/1520-0469(1972)029< 1076:AUTFTQ>2.0.CO;2 22. Kajita, R., Yamanaka, M.D., Kozan, O.: Reconstruction of rainfall records at 24 observation stations in Sumatera, Colonial Indonesia, from 1879 to 1900. J. Hydrometeor. 23, 1627–1643 (2022). https://doi.org/10.1175/JHM-D-20-0245.1 23. Kamimera, H., Mori, S., Yamanaka, M.D., Syamsudin, F.: Modulation of diurnal rainfall cycle by the Madden-Julian oscillation based on one-year continuous observation with a meteorological radar in west Sumatera. SOLA, 8, 111–114 (2012). https://doi.org/10.2151/sola.201 2-028 24. Kato, S.: Diurnal atmospheric oscillation, 1. Eigenvalues and Hough functions. J. Geophys. Res. 71, 3201–3209 (1966). https://doi.org/10.1029/JZ071i013p03201 25. Kato, S. (ed.): Report on “Feasibility Study on the Equatorial Radar in Indonesia”. STP SubCommittee, Committee on Geoelectromagnetism, Science Council of Japan (1989). https:// www.researchgate.net/publication/359861088 26. Konnen, ¨ G.P., Jones, P.D., Kaltofen, M.H., Allan, R.J.: Pre-1866 extensions of the southern oscillation index using early Indonesian and Tahitian meteorological readings. J. Clim. 11, 2325–2339 (1998). https://doi.org/10.1175/1520-0442(1998)0112.0.CO;2 27. Köppen, W.: Klassifikation der Klimate nach Temperatur, Niederschlag und Jahresablauf. Petermanns Geogr. Mitt., 64, 193–203, 243−248 (1918). http://koeppen-geiger.vu-wien.ac.at/pdf/ Koppen_1918.pdf 28. Lestari, S., Hamada J.-I., Syamsudin, F., Sunaryo, Matsumoto, J., Yamanaka, M.D.: ENSO influences on rainfall extremes around Sulawesi and Maluku Islands in the eastern Indonesian maritime continent. SOLA, 12, 37–41 (2016). https://doi.org/10.2151/sola.2016-008 29. Lindzen, R.S.: Turbulence and stress owing to gravity wave and tidal breakdown. J. Geophys. Res. 86, 9707–9714 (1981). https://doi.org/10.1029/JC086iC10p09707 30. LMG (Lembaga Meteorological dan Geofisika): Rainfall atlas of Indonesia, Vol. II: Sumatra, Kalimantan, Sulawesi, Nusatenggara, Maluku, Irian Jaya (1911–1940). In: Meteorological Note No. 9, Department of Communications, Meteorological and Geophysical Institute Republic of Indonesia, 47pp. (1973) 31. Loeb, E.M.: Sumatra; its history and people, Universitat Wien (reprinted at Oxford University Press in 1972), 350pp. (1935)

EAR Construction Motivation Revisited: Indonesian Coastline …

29

32. Matsuno, T.: A quasi one-dimensional model of the middle atmosphere circulation interacting with internal gravity waves. J. Meteor. Soc. Jpn. 60, 215–226 (1982). https://doi.org/10.2151/ jmsj1965.60.1_215 33. Mori, S., Hamada J.-I., Hattori, M., Wu, P.-M., Katsumata, M., Endo, N., Ichiyanagi, K., Hashiguchi, H., Arbain, A.A., Sulistyowati, R., Lestari, S., Syamsudin, F., Manik, T., Yamanaka, M.D.: Meridional march of diurnal rainfall over Jakarta, Indonesia, observed with a C-band Doppler radar: an overview of the HARIMAU2010 campaign. Prog. Earth Planet Sci. 5(47), 1–23 (2018). https://doi.org/10.1186/s40645-018-0202-9 34. Mori, S., Hamada J.-I., Sakurai, N., Fudeyasu, H., Kawashima, M., Hashiguchi, H., Syamsudin, F., Arbain, A.A., Sulistyowati, R., Matsumoto, J., Yamanaka, M.D.: Convective systems developed along the coastline of Sumatera Island, Indonesia observed with an X-band Doppler radar during the HARIMAU2006 campaign. J. Meteor. Soc. Jpn. 89A, 61–81 (2011). https:// doi.org/10.2151/jmsj.2011-A04 35. Mori, S., Hamada J.-I., Tauhid, Y.I., Yamanaka, M.D., Okamoto, N., Murata, F., Sakurai, N., Sribimawati, T.: Diurnal rainfall peak migrations around Sumatera Island, Indonesian maritime continent observed by TRMM satellite and intensive rawinsonde soundings. Mon. Wea. Rev. 132, 2021–2039 (2004). https://doi.org/10.1175/1520-0493(2004)1322.0. CO;2 36. Mori, S., Hamada J.-I., Yamanaka, M.D., Kodama, Y.-M., Kawashima, M., Shimomai, T., Shibagaki, Y., Hashiguchi, H., Sribimawati, T.: Vertical wind characteristics in precipitating cloud systems over west Sumatera, Indonesia, observed with Equatorial Atmosphere Radar: case study on 23-24 April 2004 during the first CPEA Campaign Period. J. Meteor. Soc. Jpn. 84A, 113–131 (2006). https://doi.org/10.1029/2005JD006501 37. Murata, F., Yamanaka, M.D., Fujiwara, M., Ogino, S.-Y., Hashiguchi, H., Fukao, S., Kudsy, M., Sribimawati, T., Harijono, S.W.B., Kelana, E.: Relationship between wind and precipitation observed with a UHF radar, GPS rawinsonde and surface meteorological instruments at Kototabang, West Sumatera during September−October 1998. J. Meteor. Soc. Jpn. 80, 347–360 (2002). https://doi.org/10.2151/jmsj.80.347 38. Murata, F., Yamanaka, M.D., Hashiguchi, H., Mori, S., Kudsy, M., Sribimawati, T., Suhardi, B., Emrizal: Dry intrusions following eastward-propagating synoptic-scale cloud systems over Sumatera Island. J. Meteor. Soc. Jpn. 84, 277–294 (2006). https://doi.org/10.2151/jmsj.84.277 39. Ogawa, M., Yamanaka, M.D., Awaluddin, Darmawan, A., Sulistyowati, R., Sulaiman, A., Kozan, O.: X-band weather radar observation over Eastern Sumatra coastal area: statistical analysis of diurnal cycle of rainfall. (in this issue) (2023) 40. Ogino, S.-Y., Yamanaka, M.D., Mori, S., Matsumoto, J.: How much is the precipitation amount over the tropical coastal region? J. Clim. 29, 1231–1236 (2016). https://doi.org/10.1175/JCLID-15-0484.1 41. Ogino, S.Y., Yamanaka, M.D., Mori, S., Matsumoto, J.: Tropical coastal dehydrator in global atmospheric water circulation. Geophys. Res. Lett. 44, 11636–11643 (2017). https://doi.org/ 10.1002/2017GL075760 42. Okamoto, N., Yamanaka, M.D., Ogino, S.-Y., Hashiguchi, H., Nishi, N., Sribimawati, T., Numaguti, A.: Seasonal variation of tropospheric wind over Indonesia: comparison between collected operational rawinsonde data and NCEP reanalysis for 1992–99. J. Meteor. Soc. Jpn. 81, 829–850 (2003). https://doi.org/10.2151/jmsj.81.829 43. Otobe, N., Sakai, S., Yoden, S., Shiotani, M.: Visualization and WKB analysis of the internal gravity wave in the QBO experiment. Nagare: Jpn. Soc. Fluid Mech. 17 (1998). http://www2. nagare.or.jp/mm/98/otobe/index.htm 44. Plumb, R.A.: The interaction of two internal waves with the mean flow: implications for the theory of the quasi-biennial oscillation. J. Atmos. Sci. 34, 1847–1858 (1977). https://doi.org/ 10.1175/1520-0469(1977)0342.0.CO;2 45. Plumb, R.A., McEwan, A.D.: The instability of a forced standing wave in a viscous stratified fluid: a laboratory analogue of the quasi-biennial oscillation. J. Atmos. Sci. 35, 1827–1839 (1978). https://doi.org/10.1175/1520-0469(1978)0352.0.CO;2

30

M. D. Yamanaka

46. Rahman, M.A., Nugroho, D.S., Yamanaka, M.D., Kawasaki, M., Kozan, O., Ohashi, M., Hashiguchi, H., Mori, S.: Weather radar detection of planetary boundary layer and smoke layer top of peatland fire in Central Kalimantan, Indonesia. Sci. Rep. 11, 367 (2021). https:// doi.org/10.1038/s41598-020-79486-6 47. Rees, D., Barnett, J.J., Labitzke, K. (eds.): COSPAR international reference atmosphere (CIRA) 1986, Part II: middle atmosphere models. Adv. Space Res. 10(12), 1–525 (1990). https://www. sciencedirect.com/journal/advances-in-space-research/vol/10/issue/12 (Some data are freely available at https://ccmc.gsfc.nasa.gov/modelweb/atmos/cospar1.html) 48. Rotunno, R.: On the linear theory of the land and sea breeze. J. Atmos. Sci. 40, 1999–2009 (1983). https://doi.org/10.1175/1520-0469(1983)0402.0.CO;2 49. Rouffaer, C.P., en IJzerman (uitg. en toeg.): De Eerste Schipvaart der Nederlanders naar OostIndië onder Cornelis de Houtman, 1595–1597. Martinus Nijhoff, Bladz. 248 (1915). https:// archive.org/details/deeersteschipvaa01rouf 50. Sakurai, N., Kawashima, M., Fujiyoshi, Y., Hashiguchi, H., Shimomai, T., Mori, S., Hamada J.-I., Murata, F., Yamanaka, M.D., Tauhid, Y.I., Sribimawati, T., Suhardi, B.: Internal structure of precipitating cloud system migrating with diurnal cycle over Sumatera Island during CPEA-I campaign observations. J. Meteor. Soc. Jpn. 86, 157–170 (2009). https://doi.org/10.2151/jmsj. 87.157 51. Sakurai, N., Mori, S., Kawashima, M., Fujiyoshi, Y., Hamada J.-I., Shimizu, S., Fudeyasu, H., Tabata, Y., Harjupa, W., Hashiguchi, H., Yamanaka, M.D., Matsumoto, J., Emrizal, Syamsudin, F.: Migration process and 3D wind field of precipitation systems associated with a diurnal cycle in west Sumatera: Dual Doppler radar analysis during the HARIMAU 2006 campaign. J. Meteor. Soc. Jpn. 89, 341–361 (2011). https://doi.org/10.2151/jmsj.2011-404 52. Sakurai, N., Murata, F., Yamanaka, M.D., Hashiguchi, H., Mori, S., Hamada J.-I., Tauhid, Y.I., Sribimawati, T., Suhardi, B.: Diurnal cycle of migration of convective cloud systems over Sumatera Island. J. Meteor. Soc. Jpn. 83, 835–850 (2005). https://doi.org/10.2151/jmsj.83.835 53. Scorer, R.: Theory of waves in the lee of mountains. Quart. J. Roy. Meteor. Soc. 75, 41–56 (1947). https://doi.org/10.1002/qj.49707532308 54. Shibagaki, Y., Shimomai, T., Kozu, T., Mori, S., Fujiyoshi, Y., Hashiguchi, H., Yamamoto, M.K., Fukao, S., Yamanaka, M.D.: Multiscale aspects of convective systems associated with an intraseasonal oscillation over the Indonesian maritime continent. Mon. Weather Rev. 134, 1682–1696 (2006). https://doi.org/10.1175/MWR3152.1 55. Sulistyowati, R., Hapsari, R.I., Syamsudin, F., Mori, S., Oishi, S.T., Yamanaka, M.D.: Rainfalldriven diurnal variations of water level in the Ciliwung River, West Jawa, Indonesia. SOLA, 10, 141–144 (2014). https://doi.org/10.2151/sola.2014-029 56. Tabata, Y., Hashiguchi, H., Yamamoto, M.K., Yamamoto, M., Yamanaka, M.D, Mori, S., Syamsudin, F., Manik, T.: Lower tropospheric horizontal wind over Indonesia: a comparison of windprofiler network observations with global reanalyses. J. Atmos. Solar Terr. Phys. 73, 986–995 (2011a). https://doi.org/10.1016/j.jastp.2010.09.016 57. Tabata, Y., Hashiguchi, H., Yamamoto, M.K., Yamamoto, M., Yamanaka, M.D, Mori, S., Syamsudin, F., Manik, T.: Observational study on diurnal precipitation cycle in equatorial Indonesia using 1.3-GHz wind profiling radar network and TRMM precipitation radar. J. Atmos. Solar Terr. Phys. 73, 1031–1042 (2011b). https://doi.org/10.1016/j.jastp.2010.10.003 58. Tanaka, H., Yamanaka, M.D.: Atmospheric circulation in the lower stratosphere induced by the mesoscale mountain wave breakdown. J. Meteor. Soc. Jpn. 63, 1047–1054 (1985). https:// doi.org/10.2151/jmsj1965.63.6_1047 59. Tsuda, T., Murayama, Y., Wiryosumarto, H., Harijono, S.W.B., Kato, S.: Radiosonde observations of equatorial atmosphere dynamics over Indonesia: 1. Equatorial waves and diurnal tides. J. Geophys. Res. 99, 10491–10505 (1994a). https://doi.org/10.1029/94JD00355 60. Tsuda, T., Murayama, Y., Wiryosumarto, H., Harijono, S.W.B., Kato, S.: Radiosonde observations of equatorial atmosphere dynamics over Indonesia: 2. Characteristics of gravity waves. J. Geophys. Res. 99, 10507–10516 (1994b). https://doi.org/10.1029/94JD00354 61. van Bemmelen, W.: Die Erforschungdes tropischen Luftoszeans in Niederländische-OstIndien. Luftfahrt und Wissenschaft 5, 1–50 (1913)

EAR Construction Motivation Revisited: Indonesian Coastline …

31

62. van Bemmelen, W.: Land- und Seebrise in Batavia. Beitr. Phys. Frei. Atmos. 10, 169–177 (1922) 63. Widiyatmi, I., Yamanaka, M.D., Hashiguchi, H., Fukao, S., Tsuda, T., Ogino, S.-Y., Harijono, S.W.B., Wiryosumarto, H.: Quasi-four-day mode observed by the boundary layer radar at Serpong (6°S, 107°E), Indonesia. J. Meteor. Soc. Jpn. 77, 1177–1184 (1999). https://doi.org/ 10.2151/jmsj1965.77.6_1177 64. Widiyatmi, I., Hashiguchi, H., Fukao, S., Yamanaka, M.D., Ogino, S.-Y., Gage, K.S., Harijono, S.W.B., Diharto, S., Djojodihardjo, H.: Examination of 3–6 day disturbances over equatorial Indonesia based on boundary layer radar observations during 1996–1999 at Serpong, Biak and Bukittinggi. J. Meteor. Soc. Jpn. 79, 317–331 (2001). https://doi.org/10.2151/jmsj.79.317 65. Winchester, S.: Krakatoa: The Day the World Exploded: August 27, 1883. Harper Collins Pub., New York, 416pp. (2003) 66. Wu, P.-M., Hamada J.-I., Mori, S., Tauhid, Y.I., Yamanaka, M.D., Kimura, F.: Diurnal variation of precipitable water over a mountainous area in Sumatera Island. J. Appl. Meteorol. 42, 1107–1115 (2003). https://doi.org/10.1175/1520-0450(2003)0422.0.CO;2 67. Wu, P.-M., Hara, M., Hamada J.-I., Yamanaka, M.D., Kimura, F.: Why a large amount of rain falls over the sea in the vicinity of western Sumatera Island during nighttime. J. Appl. Meteor. Climatol. 48, 1345–1361 (2009). https://doi.org/10.1175/2009JAMC2052.1 68. Wu, P.-M., Mori, S., Hamada J.-I., Yamanaka, M.D., Matsumoto, Kimura, F.: Diurnal variation of rainfall and precipitable water over Siberut Island off the western coast of Sumatera Island. SOLA, 4, 125–128 (2008). https://doi.org/10.2151/sola.2008-032 69. Yamamoto, H.: Weather Observations at Empire of Japan, II: The Army Weather Part. Agriculture and Forestry Statistics Publishing Inc., 531pp. (in Japanese) (2015) 70. Yamamoto, M.: Future project. (presented at INCREASE 2021) (2021) 71. Yamanaka, M.D.: Physical climatology of Indonesian maritime continent: an outline to comprehend observational studies. Atmos. Res. 178–179, 231–259 (2016). https://doi.org/10.1016/j. atmosres.2016.03.017 72. Yamanaka, M.D., Akiyama, H., Koma, Y., Okabe, Y., Nishimura, J., Soegijo, J.: Vertical microstructure of equatorial stratospheric wind on the zonal wind reversal of QBO. Abstracts of Papers Presented at the Meteorological Society of Japan, 44, 8. (in Japanese) (1983) 73. Yamanaka, M.D., Mori, S., Wu, P.-M., Hamada J.-I., Sakurai, N., Hashiguchi, H., Yamamoto, M.K., Shibagaki, Y., Kawashima, M., Fujiyoshi, Y., Shimomai, T., Manik, T., Erlansyah, Setiawan, W., Tejasukmana, B., Syamsudin, F., Djajadihardja, Y.S., Anggadiredja, J.T.: HARIMAU radar-profiler network over Indonesian maritime continent: a GEOSS early achievement for hydrological cycle and disaster prevention. J. Disaster Res. 3, 78–88 (2008). https://doi.org/10. 20965/jdr.2008.p0078 74. Yamanaka, M.D., Ogino, S.-Y., Wu, P.-M., Hamada J.-I., Mori, S., Matsumoto, J., Syamsudin, F.: Maritime continent coastlines controlling Earth’s climate. Prog. Earth Planet Sci. 5(21), 1–28 (2018). https://doi.org/10.1186/s40645-018-0174-9 75. Yoneyama, K., Zhang, C.: Years of the maritime continent. Geophys. Res. Lett. 47, e2020GL087182 (2020). https://doi.org/10.1029/2020GL087182

Study on Diurnal Variation of Rainfall Observed by X-band Polarimetric Radar in Peatlands Over Bengkalis Island, Eastern Sumatra, Indonesia Mariko Ogawa, Manabu D. Yamanaka, Awaluddin, Arief Darmawan, Albertus Sulaiman, Reni Sulistyowati, I. Dewa Gede Arya Putra, and Osamu Kozan Abstract Understanding the precipitation systems over tropical peatlands is essential for flood and forest fire risk management. The diurnal variation of rainfall over the Indonesian maritime continent has been analyzed previously by satellite, and more recently, by analyzing data from rain gauge network across Sumatra. In the complicated topography of Bengkalis Island located between Sumatra Island and the Malay Peninsula, the ground observation network is unevenly distributed. Consequently, analyzing the regional characteristics of diurnal variations of rainfall over the region is difficult. This paper presents preliminary results from observations for five rain gauges over Bengkalis Island and high spatiotemporal resolution compact X-band polarimetric radar whose data were collected from February 2020 to June 2021. The average hourly rainfall estimates by radar was compared with rain gauge data. The average hourly rainfall mentioned above during the observation period were re-averaged for each time zone. Rainfall occurred over both sea and land areas along the coast from 20:00 to 06:00 local time (LT), and heavy rain (≥50 mm/h) occurred across Bengkalis Island from west to east after midnight from 00:00 to M. Ogawa (B) · O. Kozan Center for Southeast Asian Studies (CSEAS), Kyoto University, 46 Shimoadachi-Cho, Yoshida Sakyo-Ku, Kyoto 606-8501, Japan e-mail: [email protected] M. D. Yamanaka · O. Kozan Research Institute for Humanity and Nature (RIHN), 457-4 Motoyama, Kamigamo, Kita-Ku, Kyoto 603-8047, Japan Awaluddin · A. Darmawan · A. Sulaiman Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Jalan Dr. Djunjunan No.133, Jawa Barat 40173, Indonesia R. Sulistyowati Research Center for Geospatial, National Research and Innovation Agency (BRIN), Gedung GEOSTECH, Jalan Kawasan Puspitek, Serpong, Tangerang Selatan 15314, Indonesia I. D. G. A. Putra Center for Research and Development, Meteorological, Climatological, and Geophysical Agency (BMKG), Jalan Angkasa I, No.2 Kemayoran, Jakarta Pusat 10610, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_3

33

34

M. Ogawa et al.

03:00 LT. On the other hand, rainfall was distributed over land areas in the daytime from 10:00 to 16:00 LT. These findings showed that sea-land breeze circulation is the dominant wind system in this area.

1 Introduction Tropical peatlands are distributed in lowlands along the coast of Southeast Asia. It is essential for understanding the water balance and fire hazards in the peatlands to obtain accurate real-time rainfall data for each region. Such information can only be collected by radar network in Indonesia (e.g. [16]) because the remoteness area makes rain gages maintenance difficult. Compact X-band polarimetric radars have been installed in Indonesia for disaster prevention against heavy rains and volcanic debris flows [2, 14]. These compact X-band polarimetric radar systems cover a smaller area than Global Satellite Mapping of Precipitation (GSMaP) [6] and the Indonesian Meteorological, Climatological, and Geophysical Agency (BMKG) C-band radar (single polarization) network. The compact X-band radar systems have temporal and spatial resolutions of approximately 1 min and 100 m, respectively. Consequently, more accurate rainfall with a high spatiotemporal resolution can be expected through observations of the compact X-band polarimetric radar. Tropical coastal rainfall has a regular diurnal cycle which is typical in Indonesia [15, 17]. Satellite observations of diurnal rainfall cycles in Indonesia have been performed as part of the tropical rainfall measuring mission (TRMM) [9]. The findings of the TRMM have shown that more rain falls in the evening than in the morning over the relatively large islands and that there is more rainfall in the morning than in the evening over coastal sea regions surrounding the islands. Furthermore, rainfall moves from the central mountains to the coastal areas on either side of Sumatra Island. On the other hand, the east coast of Sumatra, where peatlands are widespread, has a complex topography and the Malay Peninsula and numerous small islands lie to the north. Mori et al. [9] showed almost no difference in the annual mean rainfall between the mornings and evenings in this region. Kozan [5] investigated the diurnal cycle of rainfall and the maximum hourly rainfall after midnight using one rain gauge installed in Bukit Batu village on the east coast of Sumatra. More recently, Marzuki et al. [8] studied the regional characteristics of the diurnal cycle in rainfall over Sumatra. They used 229 BMKG rain gauges deployed throughout Sumatra to perform cluster analyses of rainfall amount, intensity, duration, and frequency. However, since the density of the BMKG rain gauge network in eastern Sumatra is lower than that in the western part, the regional characteristics of the diurnal cycle of rainfall in the east region contain uncertainties. The purpose of this study is to examine the regional characteristics of the diurnal cycle of rainfall and the movement of rain by using a compact X-band polarimetric radar system from February 2020 to June 2021 over Bengkalis Island, Riau Province. This is the first report on weather radar observations from the east coast of Sumatra.

Study on Diurnal Variation of Rainfall Observed by X-band Polarimetric …

35

Firstly, this study verifies the accuracy of radar-based rainfall estimates by comparing the average hourly rainfall calculated from the radar data with accumulated rainfall measured by five rain gauges. Secondly, to clarify the regional characteristics of the diurnal rainfall cycle, the average hourly rainfall mentioned above was re-averaged to each local time zone from 00:00 to 23:00.

2 Instruments and Methodology 2.1 Furuno Radar A compact type X-band polarimetric radar system (WR-2100, Furuno Electric Co., Ltd), hereafter referred to as the Furuno radar, was installed at Bengkalis State Islamic College (STAIN Bengkalis) on Bengkalis Island, Riau Province (1° 29’ N, 102° 07’ E) in February 2020 (Fig. 1). Figure 2 shows the study area. Table 1 shows the specifications of the Furuno radar system.

Fig. 1 Furuno radar was installed at STAIN Bengkalis in February 2020

36

M. Ogawa et al.

Fig. 2 Study area. The line indicates the location of Furuno radar installed at STAIN Bengkalis on Bengkalis Island, Riau Province

2.2 Radar Preprocessing and Rainfall Estimation Algorithms Radar data were converted from polar to cartesian coordinates. The rainfall data within a 100 m grid cell is determined by calculating the weighted average using all data values within a search radius of 1 km. Estimates of radar rainfall data in this study used radar parameters, Z H and K DP , and the method of Park et al. [11], where Z H is the horizontal reflectivity factor and K DP is the specific differential phase. Before calculating K DP , the observed phase information was smoothed using an iterative filter method [3] to obtain Φ DP , which is the cross-polarization difference phase. In addition, attenuation correction using the method of Bringi and Chandrasekar [1] was performed before estimating rainfall using the radar data.

2.3 Average Hourly Rainfall Scanning radar data obtained at 10 min intervals were used to calculate the average hourly rainfall only when six radar scans could be performed per hour. These values may differ from accumulated hourly rainfall measured using a rain gauge because tropical rainfall intensity can change at scales of 0.9 between the amount of rain measured by rain gauge and by radar estimated rainfall using the method of Park et al. [11]. They interpolated Furuno radar data using the scanning mode with a constant altitude plan position indicator (CAPPI) composed of multiple PPI data with different elevation angles. They then extracted the radar grid at a constant altitude of 500 m, and confirmed the existence of a precipitation system and melting layer through radiosonde data and the radar parameters, Z DR and ρ HV , where Z DR is the differential reflectivity and ρ HV is the copolar correlation coefficient.

Study on Diurnal Variation of Rainfall Observed by X-band Polarimetric …

41

Fig. 6 Scatterplots of data from radar and five rain gauges during the observation period from February 15, 2020 to June 3, 2021. The numbers in the titles of scatterplots correspond to the station numbers shown in Table 2. The X-axis shows 1 h rainfall accumulation (mm) measured by rain gauges. Y-axis shows average hourly rainfall (mm) calculated by the Furuno radar using the method of Park et al. [11]. The coefficient of determination varied from 0.4 to 0.9, excluding two stations (No. 6 and 7) where heavy rainfalls of 30 mm/hr or more were frequently observed

4.2 Regional Characteristics of the Diurnal Rainfall Cycle Figure 7 shows the average hourly rainfall during the observation period in each time zone from 00:00 to 23:00 LT. Rainfall occurred over both sea and land areas along the coast from nighttime to sunrise (i.e., 20:00–06:00 LT). Heavy rains approximately ≥50 mm/h occurred across Bengkalis Island from west to east after midnight (i.e., 00:00–03:00 LT). The timing of heavy rainfalls was similar to that observed by Kozan [5]. On the other hand, rainfall was distributed over land before noon to before sunset (i.e., 10:00–16:00 LT), with heavy rains ≥50 mm/h moving from north to south across Bengkalis Island during that time. These findings show that land-sea breeze circulation is dominant over Bengkalis Island and the surrounding area. Here, it should be noted that the rainfall estimated by the Furuno radar can overestimate compared to the rain gauge. According to Fig. 6, when the rainfall estimated by the Furuno radar was about 50 mm/hr at station No. 3, the rain gauge was about 23 mm/hr. Although previous studies using rain gauges have reported that the rainfall in eastern Sumatra has a diurnal cycle [5, 8], this study confirmed the regional characteristics of this diurnal rainfall cycle in eastern Sumatra using radar data.

42

M. Ogawa et al.

Fig. 7 Average hourly rainfall derived from Furuno radar during the observation period from February 15, 2020 to June 3, 2021 in each time zone from 00:00 to 23:00 LT (modified from Ogawa et al. [10]). Rainfall data were estimated using the method of Park et al. [11]. Rainfall occurred over both sea and land areas along the coast from nighttime to sunrise (the area surrounded by the blue dotted line from 20:00 to 6:00 LT). Heavy rains approximately ≥50 mm/h occurred across Bengkalis Island from west to east after midnight (the area was surrounded by the blue dotted line from 00:00 to 03:00 LT). On the other hand, rainfall was distributed over land before noon to before sunset, with heavy rains ≥50 mm/h (the area surrounded by the red dotted line from 10:00 to 16:00 LT) moving from north to south across Bengkalis Island during that time (down arrow)

5 Conclusions This paper aims to clarify the regional characteristics of the diurnal cycle of rainfall and the movement of rain over the east coast of Sumatra, Indonesia. Furuno radar, a compact X-band polarimetric radar with a high spatiotemporal resolution, was newly installed and operated from February 2020, to June 2021 on Bengkalis Island, Riau Province. Firstly, the average hourly rainfall observed by Furuno radar every 10 min was compared to accumulated rainfall data obtained from five rain gauge stations. The coefficient of determination varied from 0.4 to 0.9, excluding two stations where heavy rainfalls of 30 mm/hr or more were frequently observed. These results include uncertainties related to the time resolution of the data and the altitude of the extracted radar data, which are associated with the precipitation system. On the other hand, rainfall measurements at two stations were considerably underestimated. This is because the rainfall moves faster and the heavy rain of about 30 mm/hr or more is localized, so that the radar data analyzed every 10 min might miss the heavy rain.

Study on Diurnal Variation of Rainfall Observed by X-band Polarimetric …

43

Secondly, this study qualitatively clarified the regional characteristics of the diurnal cycle of rainfall and the rainfall movement over the island. Using average hourly rainfall data obtained from the Furuno radar during the observation period in each time zone (i.e., 00:00–23:00 LT), the land-sea rainfall and the movement of rainfall were explained in terms of their diurnal characteristics, i.e., whether they occurred at night time (20:00–06:00 LT) or daytime (10:00–16:00 LT) and in which time zone. The findings indicated that land-sea breeze circulation is dominant in the Bengkalis Island area. Radar can also be applied to monitor the upper bounds of smoke layers when there is no precipitation [12], such as the smoke associated with peatland fires which have caused extensive environmental issues both locally and globally [5]. Acknowledgements This study was supported in part by a grant from JSPS KAKENHI (No. 19KK0106) and RIHN research project entitled “Toward the Regeneration of Tropical Peatland Societies: Building an International Research Network on Paludiculture and Sustainable Peatland Management.” Rain gauge data from stations at Perapat Tunggal and Selatbaru were provided by Assoc. Prof. Koichi Yamamoto, Yamaguchi University.

References 1. Bringi, V.N., Chandrasekar, V.: Polarimetric Doppler Weather Radar: Principles and Applications. Cambridge University Press, United Kingdom (2001) 2. Hapsari, R.I., Oishi, S., Syarifuddin, M., Asmara, R.A., Legono, D.: X-MP Radar for developing a lahar rainfall threshold for the Merapi Volcano using a Bayesian approach. J. Disaster Res. 14(5), 811–828 (2019). https://doi.org/10.20965/jdr.2019.p0811 3. Hubbert, J., Bringi, V.N.: An iterative filtering technique for the analysis of copolar differetial phase and dual-frequency radar measurements. J. Atmos. Oceanic Technol. 12(3), 643–648 (1995). https://doi.org/10.1175/1520-0426(1995)012%3c0643:AIFTFT%3e2.0.CO;2 4. Kiguchi, M., Oki, T.: Point precipitation observation extremes in the world and Japan. J. Jpn. Soc. Hydrol. Water Resour. 23(3), 231–247 (2010). https://doi.org/10.3178/jjshwr.23.231 5. Kozan, O.: Precipitation and groundwater variations in peat swamps. Stud. Sustain. Humanosphere 4, 271–287 (2012). (in Japanese) 6. Kubota, T., Aonashi, K., Ushio, T., Shige, S., Takayabu, Y.N., Kachi, M., Arai, Y., Tashima, T., Masaki, T., Kawamoto, N., Mega, T., Yamamoto, M. K., Hamada, A., Yamaji, M., Liu, G., Oki, R.: Global Satellite Mapping of Precipitation (GSMaP) products in the GPM era. In: Satellite Precipitation Measurement, Springer (2020). https://doi.org/10.1007/978-3-030-24568-9_20 7. Marshall, J.S., Palmer, W.M.: The distribution of raindrops with size. J. Meteor. 5(4), 154–166 (1948). https://doi.org/10.1175/1520-0469(1948)005%3c0165:TDORWS%3e2.0.CO;2 8. Marzuki, M., Suryanti, K., Yusnaini, H., Tangang, F., Muharsyah, R., Vonnisa, M., Devianto, D.: Diurnal variation of precipitation from the perspectives of precipitation amount, intensity and duration over Sumatra from rain gauge observations. Int. J. Climatol. 41(8), 4386–4397 (2021). https://doi.org/10.1002/joc.7078 9. Mori, S., Jun-Ichi, H., Tauhid, Y.I., Yamanaka, M.D., Okamoto, N., Murata, F., Sakurai, N., Hashiguchi, H., Sribimawati, T.: Diurnal land-sea rainfall peak migration over Sumatera Island, Indonesian maritime continent observed by TRMM satellite and intensive rawinsonde soundings. Mon. Wea. Rev. 132(8), 2021–2039 (2004). https://doi.org/10.1175/1520-0493(2004)132 2.0.CO;2

44

M. Ogawa et al.

10. Ogawa, M., Yamanaka, M. D., Awaluddin, A., Darmawan, A., Sulistyowati, R., Sulaiman, A., Kozan, O.: X-band weather radar observations in the east coast of Sumatra: statistical analysis of diurnal cycle of rainfall. In: Proceedings of the 15th Symposium on MU Radar and Equatorial Atmosphere Radar (2021) 11. Park, S.G., Bringi, V.N., Chandrasekar, V., Maki, M., Iwanami, K.: Correction of radar reflectivity and differential reflectivity for rain attenuation at X band. Part I: theoretical and empirical basis. J. Atmos. Oceanic Technol. 22(11), 1621–1631 (2005). https://doi.org/10.1175/JTECH1 803.1 12. Rahman, M.A., Nugroho, D.S., Yamanaka, M.D., Kawasaki, M., Kozan, O., Ohashi, M., Hashiguchi, H., Mori, S.: Weather radar detection of planetary boundary layer and smoke layer top of peatland fire in Central Kalimantan Indonesia. Sci. Rep. 11(367), 1–9 (2021). https://doi.org/10.1038/s41598-020-79486-6 13. Syarifuddin, M., Ogawa, M., Nakamichi, H., Oishi, S., Iguchi, M.: Performance of a smallcompact X-MP radar to monitor extreme rainfall event on 7 July 2018. In: Proceedings of the 22nd IAHR-APD Congress 2020, pp. 1–8 (2020) 14. Syarifuddin, M., Oishi, S., Hapsari, R.I., Shiokawa, J., Mawandha, H.G., Iguchi, M.: Estimating the volcanic ash fall rate from the mount sinabung eruption on february 19, 2018 using weather radar. J. Disaster Res. 14(1), 135–150 (2019). https://doi.org/10.20965/jdr.2019.p0135 15. Yamanaka, M.D.: Physical climatology of Indonesian maritime continent: an outline to comprehend observational studies. Atmos. Res. 178–179, 231–259 (2016). https://doi.org/10.1016/j. atmosres.2016.03.017 16. Yamanaka, M.D., Hashiguchi, H., Mori, S., Wu, P.M., Syamsudin, F., Manik, T., Hamada, J.I., Yamamoto, M.K., Kawashima, M., Fujiyoshi, Y., Sakurai, N., Ohi, M., Shirooka, R., Katsumata, M., Shibagaki, Y., Shimomai, T., Erlansyah, Setiawan, W., Tejasukmana, B., Djajadihardja, Y.S., Anggadiredja, J.T.: HARIMAU radar-profiler network over the Indonesian maritime continent: a GEOSS early achievement for hydrological cycle and disaster prevention. J. Disaster Res. 3(1), 78–88 (2008). https://doi.org/10.20965/jdr.2008.p0078 17. Yamanaka, M.D., Ogino, S.-Y., Wu, P.-M., Hamada, J.-I., Mori, S., Matsumoto, J., Syamsudin, F.: Maritime continent coastlines controlling Earth’s climate. Prog. Earth Planet Sci. 5(21), 1–28 (2018)

Detrended Fluctuation Analysis (DFA) of Gunungsitoli Geomagnetic Station to Assess the Possibility of the Earthquake Precursor Febty Febriani, Cinantya Nirmala Dewi, Suaidi Ahadi, Titi Anggono, Syuhada, Mohammad Hasib, and Aditya Dwi Prasetio Abstract Nias Island is one of the active seismic regions in Indonesia. We analyzed three components of geomagnetic data recorded at the Gunungsitoli (GSI) station to identify possible geomagnetic anomalies before earthquakes on Nias Island in 2018. We selected earthquakes with a magnitude > 5 and an epicenter ≤ 100 km from the GSI. We calculated the local seismicity index (KLS ) and the daily local earthquake energy (Es). We applied detrended fluctuation analysis (DFA) to investigate the possibility of the earthquake precursor related to the analyzed earthquake. We also used the disturbance storm time (Dst) index to analyze geomagnetic storm activity. The result presents a decrease in the DFA value, which exceeds the DFA’s value threshold during the analyzed period. The decrease occurred about two months and one month prior to the Mw5.4 (06/09/2018) and Mw5.2 (20/08/2018) earthquakes. The Dst index also shows the quiet period where the DFA value decreases. It confirms that the anomalies were not related to geomagnetic storms. The Mw5.4 earthquake had a larger magnitude and the shallowest depth. It can be the main contributor to the appearance of geomagnetic anomalies. Therefore, we suggest that the decrease in the DFA value is possibly related to the Mw5.4 earthquake. Keywords Geomagnetic anomaly · Detrended fluctuation analysis · Earthquake forecast

F. Febriani (B) · C. N. Dewi · T. Anggono · Syuhada · M. Hasib · A. D. Prasetio Research Center for Geological Disasters, National Agency of Research and Innovation (BRIN), Jl. Sangkuriang, Kompleks BRIN, Jawa Barat 40135, Indonesia e-mail: [email protected] C. N. Dewi e-mail: [email protected] S. Ahadi Indonesian Meteorological, Climatological and Geophysical Agency (BMKG), Jl. Angkasa I No. 2, Kemayoran, DKI Jakarta 10720, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_4

45

46

F. Febriani et al.

1 Introduction Nias is a small Island on the western coast of Sumatra, Indonesia, and is geologically located in the Sunda megathrust. The collision between the Indo-Australian Plate affects the geological structure of Nias Island and causes the Island as one of the most active seismic regions in Indonesia. As a result, some earthquakes occurred in the surrounding Nias Island in the past. For example, the M8.6 Nias earthquake occurred on March 28, 2005, and caused many casualties, loss, and damaged buildings [1]. Many studies have observed that geomagnetic precursors may appear before moderate-large earthquakes worldwide e.g., [2–8]. In Indonesia, the research related to the earthquake precursor by investigating anomalies in the geomagnetic signals has evolved. Furthermore, the geomagnetic anomalies before the earthquakes occurred on Sumatra Island have been reported e.g., [3, 9–13]. Many methods have been applied to investigate geomagnetic anomalies prior to an earthquake. One of them is detrended fluctuation analysis (DFA) e.g., [7, 14–17]. Therefore, it is essential to assess the precursor associated with earthquakes based on the geomagnetic data as a tool for earthquake hazards and risk mitigation on Sumatra Island. In this study, we apply DFA as the analysis method. We also calculated the certain earthquake epicenter distance and local earthquake energy released by earthquakes in observing geomagnetic anomalies to investigate the anomalies related to the earthquakes.

2 Data and Method Figure 1 presents the location of the Gunungsitoli (GSI) station and all Mw5 earthquake epicenters during 2018 based on the United States Geological Survey (USGS) catalog. The magnetometer system is operated by the Indonesian Meteorological, Climatological, and Geophysical Agency (BMKG) with an averaged sampling frequency of 1 Hz. It was developed by Magnetic Data Acquisition System (MAGDAS) project belonging to Space Environment Research Center (SERC), Kyushu University. It recorded H, D, Z, and total (F) components with the observation range ± 70,000 nT. In this study, we chose an earthquake based on criteria M > 5 and the distance between the epicenter and the station within 100 km. The analyzed data period in this study is from January 25, 2018, to December 31, 2018. There is no data from January 1, 2018, to January 24, 2018. We use X, Y, and Z components in this study. Therefore, we convert the H, D, and Z to X, Y, and Z components. We also calculated the local seismicity index (K LS ) and the daily sum of the local earthquake energy (E s ) by applying Eqs. 1 and 2 [18, 19]. K LS =

100.75M R + 100

(1)

Detrended Fluctuation Analysis (DFA) of Gunungsitoli Geomagnetic …

47

Fig. 1 The red square in Indonesia’s map inset indicates the location of the study area. The gray circle is the Gunung Sitoli (GSI) station. The stars are the epicenter of the earthquakes, which is indicated in Table 1. The color scale describes the depth of the earthquakes



Es =

104.8+1.5M r2

(2)

R is the distance between the earthquake epicenter and the station, r is the hypocenter distance, and M is the earthquake’s magnitude. The local seismicity index (K LS ) represents the correlation between the magnitude and distance of an earthquake [18]. It can be applied to investigate the primary source of the detected precursor in an earthquake series [20]. The E s parameter considers the hypocenter distance and the magnitude of the selected earthquake events [19]. The detailed parameters of all earthquakes analyzed in this study are listed in Table 1. In this study, we use the night data (17:00–21:00 UTC) since they have less contamination from artificial and natural noises [7, 21, 22]. Then, we apply detrended fractal analysis (DFA) for further analysis. This analysis is carried out to define the data scaling behavior. Even though we do not know the detail of the data origin and shape, we can investigate the possible trends of the data by applying DFA [23]. We assume y(k) is a time series of interevent x(i) and x ave indicates the mean interevent time. Then, we can calculate y(k)as follows:

Lat (N)

1.041

0.821

1.077

No

1

2

3

97.338

96.895

97.376

Long (E)

25.91

10

25.69

Depth (km)

5.2

5.4

5.2

Mag (Mw)

29/12/2018

06/09/2018

20/08/2018

Date

08:15:27

00:13:50

10:51:05

Time (UTC)

36.085

92.424

36.403

Epicenter distance R (km)

44.424

92.963

44.555

Hypocenter distance r (km)

58.37

58.31

58.23

Local seismicity index K LS

Table 1 A list of earthquakes that occurred during the analyzed data period based on the USGS global earthquake catalog

2.2E + 09

9.2E + 08

2.0E + 09

Daily sum of the local earthquake energy E s

48 F. Febriani et al.

Detrended Fluctuation Analysis (DFA) of Gunungsitoli Geomagnetic …

y(k) =

k 

[x(i) − xave ]

49

(3)

i=

with k = 1 … N. Then, the y(k) is divided into boxes with equal length, n, and a least-square polynomial yn (k) of degree p is applied in each box. It will determine the trend of the order p. In the next step, we find F(n) through Eq. (4):  F(n) =

N 

1/2 [y(k) − yn (k)]

2

(4)

k=1

If we calculated F(n) for all box sizes, the relationship between F(n) and n could be determined as follows: F(n) ∝ n α

(5)

According to the above formula, we define the α value. In the last step, we calculated the mean ± 2σ of all α values during the analyzed data period as the threshold with σ is the standard deviation. We determined the anomaly if there is data that exceeds the threshold. This assumption refers to the normal distribution of data that can be assumed as the peculiarity of the data trend if data surpasses the threshold.

3 Results and Discussion Figure 2 shows the X, Y, and Z components of the geomagnetic data recorded at the GSI station. The data was recorded on November 25, 2018. The figure shows that the daytime data of the X, Y, and Z components are higher than that of nighttime data. It presents typical geomagnetic data. Figure 3 indicates the correlation between F ~ n for the X, Y, and Z components of November 25, 2018, geomagnetic data. We calculated the correlation for all available data by applying Eq. (5) to obtain the α value of each component during the analyzed data period. We present them in Fig. 4. The first, second, and third panels in Fig. 4 indicate the α value during the analyzed data period for the X, Y, and Z components. The last panel in Fig. 4 is the Dst index during the analyzed data period. It describes the geomagnetic activity, especially in the low latitude region [24]. We used the Dst index to confirm that the geomagnetic anomalies which appeared prior to the earthquake does not origin from external disturbance or solar-terrestrial interaction (solar wind, interplanetary magnetic field). The Dst ≤ −100 nT, −100 nT < Dst ≤ −50 nT, and −50 nT < Dst ≤ -30 nT indicates intense, moderate, and small storms, respectively [25].

50

F. Febriani et al.

Fig. 2 The X, Y, and Z components of the November 25, 2018, geomagnetic data were recorded at the GSI geomagnetic station. The X-axis indicates the recorded data time in UTC. The Y axes are magnetic fields for each component in nano Tesla (nT)

Figure 4 presents global geomagnetic storms from January 25, 2018, to December 31, 2018. They are indicated by the green circle in the last panel of Fig. 4. The highest storm occurred before the Mw5.4 earthquake, as indicated by number 9 in the last panel of Fig. 4. However, there is no α value of the three components that surpassed the threshold when the ninth storm occurred. The storm prior to an earthquake is an exciting character since a recent publication suggested that the probability of a geomagnetic storm prior to a shallow earthquake is high [26]. In addition, Fig. 4 also shows that the decrease of αvalue exceeded the threshold when the second, third, fourth, fifth, tenth, and eleventh geomagnetic storms appeared. The anomalies related to a geomagnetic storm are very clear in the α value of the Y component. The decrease of the αvalue surpassing the threshold also appeared about 30 and 50 days prior to the Mw5.2 (20/08/2018) and Mw5.4 (06/09/2018) earthquakes. The anomaly period ends approximately 10 and 25 days before the Mw5.2 (20/08/2018) and Mw5.4 (06/09/2018) earthquakes. The anomalies emerge when the quiet period of the Dst index. The previous studies revealed that the α value could be increased or decreased before an earthquake [26, 27]. Figure 4 also presents there is no anomaly prior to the Mw5.2 (29/12/2018).

Detrended Fluctuation Analysis (DFA) of Gunungsitoli Geomagnetic …

51

Fig. 3 An example of the correlation between F ~ n for the X, Y, and Z components of the November 25, 2018, geomagnetic data

During the analyzed data period, the Mw5.4 earthquake (06/09/2018) is the earthquake that has the smallest depth. The K LS value of the Mw5.4 (06/09/2018) earthquake is larger than that of the Mw5.2 (20/08/2018) earthquake. Previous research has reported that an earthquake with a high K LS value in a series of earthquakes is considered the primary source of every detected precursor [21]. Therefore, we assume the anomalies are possibly related to the Mw5.4 earthquake. It also indicated that the critical factors related to the geomagnetic precursor before an earthquake is possibly the depth and magnitude of the earthquake. This study supports the recent research that revealed electromagnetic/geomagnetic anomaly could be one candidate of the short-term precursor before an earthquake [28]. Therefore, since Indonesia is an active seismic country in the world, studying the geomagnetic precursor related to an earthquake must be carried out intensively in the future.

52

F. Febriani et al.

Fig. 4 The result of the DFA analysis of the geomagnetic data and Dst index during the analyzed data period. The reverse triangles indicate that all M5 earthquakes occurred during 2018. The numbers in the last panel indicate the geomagnetic storm that occurred in 2018. The blue lines indicate the threshold mean ± 2σ, with σ as the standard deviation

4 Conclusion We analyzed the geomagnetic data recorded in the Guningsitoli station to assess the possibility of the geomagnetic anomalies associated with the Mw ≥ 5 earthquake in 2018. We considered the distance between the station and the epicenter less than 100 km. We applied detrended fluctuation analysis (DFA) in this study. The results present that the α value decreased, surpassing from 30–10 days and 50–25 days prior to the Mw5.2 (20/08/2018) and Mw5.4 (06/09/2018) earthquakes, respectively. Since the Mw5.4 (06/09/2018) earthquake has the smallest depth and bigger K LS than that

Detrended Fluctuation Analysis (DFA) of Gunungsitoli Geomagnetic …

53

of Mw5.2 (20/08/2018), the anomalies are possibly related to the Mw5.4 earthquake. This study also revealed that the possibility of the earthquake’s depth and magnitude are important factors for the geomagnetic precursor to appear before an earthquake. Funding We thank you for funding support from L’oreal-UNESCO for Women in Science Program and the Disaster Research Program (Grant No. SP DIPA-124.01.1.690501/2022) of the National Research and Innovation Agency (BRIN) for this research. We also thank the U.S. Geological Survey (USGS) for the earthquake database and World Data Center for Geomagnetism, Kyoto University, for the Dst index (http://wdc.kugi.kyoto-u.ac.jp/dstdir/). The Generic Mapping Tools 7 is applied in this paper to make some figures (http://www.soest.hawaii.edu/gmt/; Wessel and Smith 1998).

References 1. Setiyono, U., Gunawan, I., Proyobudi, Yatimantoro, T., Imananta, R.T., Ramdhan, M., Hidayanti, Anggraini, S., Rahayu, R.H., Hawati, P., Yogaswara, D.S., Julius, A.M., Apriani, Harvan, M., Simangunsong, G., Kriswinarso, T.: Significant and destructive earthquakes catalog 1821–2018. Indonesian Meteorological, Climatological, and Geophysical Agency (BMKG), Jakarta (2019) (in Indonesia) 2. Febriani, F., Han, P., Yoshino, C., Hattori, K., Nudiyanto, B., Effendi, N., Maulana, I., Suhardjono, Gaffar, E.: Ultra low frequency (ULF) electromagnetic anomalies associated with large earthquakes in Java Island, Indonesia by using wavelet transform and detrended fluctuation analysis. Nat. Hazards Earth Syst. Sci. 14, 789–798(2014) 3. Ahad, S., Puspito, N.T., Ibrahim, G., Saroso, S., Yumoto, K., Yoshikawa, A.: Muzli: anomalous ULF emissions and their possible association with the strong earthquakes in Sumatra, Indonesia, during 2007–2012. J. Math. Fundam. Sci. 47, 84–103 (2015) 4. Han, P., Hattori, K., Huang, Q., Hirooka, S., Yoshino, C.: Spatiotemporal characteristic of the geomagnetic diurnal variations prior to the 2011 off the Pacific coast of Tohuku earthquake (Mw9.0). J. Asian Earth Sci. 129, 13–21 (2016) 5. Han, P., Zhuang, J., Hattori, K., Chen, C.-H., Febriani, F., Chen, H., Yoshino, C., Yoshida, S.: Assesing the potential earthquake precursory information in ULF magnetic data recorded in Kanto, Japan, during 2000–2010: Distance and Magnitude Response. Entropy 22(8), 859 (2020) 6. Swati, Singh, B., Pundhir, D., Sinha, A.K., Rao, K.M., Guha, A., Hobara, Y.: Ultra-low frequency (ULF) magnetic field emissions associated with some major earthquake occurred in Indian Subcontinent. J. Atmos. Solar Terr. Phys. 211, 105469 (2020) 7. Febriani, F., Ahadi, S., Anggono, T., Dewi, C.N., Prasetio, A.D.: Applying wavelet analysis to assess the ultra low frequency (ULF) geomagnetic anomalies prior to M6.1 Banten earthquake (2018). In: Subehi, L., Sugeha, H.Y. (eds), International Conference on the Ocean and Earth Sciences, vol. 789, 012064. IOP Publishing Ltd (2021) 8. Yusof, K.A., Abdullah, M., Hamid, N.S.A., Ahadi, S., Yoshikawa, A.: Correlations between earthquake properties and characteristic of possible ULF geomagnetic precursor over multiple earthquake. Universe 7(1), 20 (2021) 9. Ahyar, A.S., Sunardi, B.: Correlation of Z/H magnetic polarization for earthquake precursor identification around Pelabuhan Ratu. Spektra: Jurnal Fisika dan Aplikasinya 2, 179–186 (2017) (in Indonesia) 10. Takla, E.M., Yumoto, K., Liu, J.Y., Kakinami, Y., Uozumi, T., Abe. S., Ikeda, A.: Anomalous geomagnetic variations possibly linked with the Taiwan earthquake (Mw= 6.4) on 19 December 2009. Int. J. Geophys. 2011, 848467 (2011)

54

F. Febriani et al.

11. Ramadhani, A.F., Syafriani, Triyono, R., Ubaya, T.: Analysis of magnetic field anomalies before earthquakes based on ULF (Ultra Low Frequency) method using magnetic sensor data in Sumatra. IOP Conf. Ser. J. Phys. Conf. Ser. 1317, 012044 (2019) 12. Sokacana, I., Rosid, M.S., Ahadi, S.: Optimum frequency identification of anomalous geomagnetic signals related to earthquake precursor in Sumatra Island. IOP Conf. Ser. Earth Environ. Sci. 406, 012021 (2019) 13. Marzuki, M., Hamidi, M., Ahadi, S., Putra, A., Afdal, A., Harmadi, H., Karnawati, D., Suprihatin, H.S., Syirojudin, M., Marsyam, I.: ULF geomagnetic anomaly associated with the Sumatra-Pagai Island earthquake swarm during 2020. Contribution Geophys. Geodesy 52(2), 185–207 (2022) 14. Ida, Y., Hayakawa, M.: Fractal Analysis for the ULF data during 1993 Guam earthquake to study prefracture critically. Nonlinear Process. Geophys. 13, 409–412 (2006) 15. Telesca, L., Hattori, K.: Non-uniform scaling behavior in ultra-low frequency (ULF) earthquake-related geomagnetic signals. Phys. A 384, 522–528 (2007) 16. Telesca, L., Lapenma, V., Macchiato, M., Hattori, K.: Earth Planet. Sci. Lett. 268, 219–224 (2008) 17. Hayakawa, M., Ida, Y.: Fractal (mono- and multi-) analysis for the ULF data during the 1993 Guam earthquake for study of prefracture criticality. Curr. Dev. Theory Appl. Wavelets 2, 159–174 (2008) 18. Molchanov, O.A., Hayakawa, M.: Seismo-Electromagnetics and Related Phenomena: History and Latest Results, vol. 189. Terrapub, Tokyo (2008) 19. Hattori, K., Serita, A., Yoshino, C., Hayakawa, M., Isezaki, N.: Singular spectral analysis and principal component analysis for signal discrimination of ULF geomagnetic data associated with 2000 Izu Island Earthquake Swarm. Phys. Chem. Earth 31, 281–291 (2006). https://doi. org/10.1016/j.pce.2006.02.034 20. Yusof, K.A., Hamid, N.S.A., Abdullah, M., Ahadi, S., Yoshikawa, A.: Assessment of signal processing methods for geomagnetic precursor of the 2012 M6.9 Visayas, Philippines earthquake. Acta Geophysica 67, 1297–1306 (2019) 21. Potirakis, S.M., Hayakawa, M., Schekotov, A.: (2017) Fractal analysis of the ground-recorded ULF magnetic fields prior to the 11 March 2011 Tohoku earthquake (Mw = 9): discriminating possible earthquake precursors from space-sourced disturbances. Nat. Hazards 85, 59–86 (2017) 22. Peng, C.K., Buldyrev, S.V., Havlin, S., Simons, M., Stanley, H.E., Goldberger, A.L.: Mosaic organization of DNA nucleotides. Phys. Rev. E 49, 1685–1689 (1994) 23. Nose, M., Iyemori, T., Sugiura, M., Kamei, T.: Geomagnetic Dst index world data center for geomagnetism Kyoto (2015).https://doi.org/10.17593/14515-74000 24. Gonzalez, W.D., Joselyn, J.A., Kamide, Y., Kroehl, H.W., Rostoker, G., Tsurutani, B.T., Vasyliunas, V.M.: What is a geomagnetic storm? J. Geophys. Res. 99, 5771 (1994) 25. Chen, H., Wang, R., Miao, M., Liu, X., Ma, Y., Hattori, K., Han, P.: A statistical study of the correlation between geomagnetic storms and M≥7.0 global earthquake during 1957–2020. Entropy 22(11), 1270 (2020) 26. Ida, Y., Yang, D., Li, Q., Sunvand, H., Hayakawa, M.: Fractal analysis of ULF electromagnetic emissions in possible association with earthquakes in China. Nonlinear Process. Geophys. 19, 577–583 (2012) 27. Potirakis, M., Hayakawa, M., Schekotov, A.: Fractal analysis of the ground-recorded ULF magneticfields prior to the 11 March 2011 Tohoku earthquake (Mw = 9): discriminating possible earthquake precursors from space-sourced disturbances. Nat. Hazards 85, 59–86 (2017) 28. Wang, J.H.: Piezoelectricity as a mechanism on generation of electromagnetic precursors before earthquakes. Geophys. J. Int. 224(1), 682–700 (2021)

Low-Latitude Fluctuation of Ionospheric Magnetic Field Measured by LAPAN-A3 Satellite Fitri Nuraeni, La Ode M. Musafar Kilowasid, Clara Y. Yatini, Visca Wellyanita, Satriya Utama, Yoga Andrian, Teti Zubaidah, Wahyudi Hasbi, B. Moh. Andi Aris, Setyanto C. Pranoto, Harry Bangkit, Muzirwan, Ega A. Anggari, Erlansyah, Nata Miharja, Dwi Ratnasari, Prita Ayuningtyas, and Angga Yolanda Putra Abstract The fluctuation of the magnetic field measured by satellite LAPAN-A3 is analyzed for low-latitude data. The low-latitude magnetic field characteristic has been long well understood from the ground-based observation that the magnetic variation shows a peak around local time and it is strongly related to the amount of ionospheric ionization, which affects current flow in the ionosphere. The LAPAN-A3 satellite orbits at an ionospheric altitude where the low-latitude geomagnetic field lines lie almost horizontally in the ionosphere so that variations in the ionospheric magnetic field at low latitudes will show a pattern similar to local time variations at groundbased stations at low latitudes. To examine the magnetic field fluctuation measured by LAPAN-A3 satellite, we are analyzing the local time ionospheric magnetic variation under quiet conditions.

F. Nuraeni · C. Y. Yatini · V. Wellyanita · Y. Andrian · B. Moh. Andi Aris · S. C. Pranoto · Erlansyah · N. Miharja Research Center for Space, Bandung, West Java, Indonesia L. O. M. Musafar Kilowasid (B) · P. Ayuningtyas · A. Y. Putra Pontianak Observatory for Space and Atmosphere, Pontianak, West Kalimantan, Indonesia e-mail: [email protected] S. Utama · W. Hasbi · E. A. Anggari Research Center for Satellite Technology, Bogor, West Java, Indonesia T. Zubaidah · D. Ratnasari Mataram University, Lombok, Indonesia H. Bangkit Research Center for Smart Mechatronics, Bandung, Indonesia Muzirwan Research Center for Climate and Atmosphere, Bandung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_5

55

56

F. Nuraeni et al.

1 Introduction The study of geomagnetic fields and related phenomena in space can be carried out using the measurements of ground-based magnetometers and satellites. Measurement of the geomagnetic field on the earth’s surface has been carried out for a long time, and several magnetometer networks have been established to study and understand dynamic processes in space [1]. From the measurements on the ground, various phenomena related to the magnetic field in space have been identified and characterized that can describe the dynamics and processes that occur in the magnetosphere and/or ionosphere [2]. In addition, several satellites have also been launched to carry out in-situ measurements to confirm the gains through measurements of the ground-based magnetometer. One of the earliest observations regarding the characteristics of the geomagnetic field was that geomagnetic variation is the variation in solar quiet that depends on solar activity [3]. Thus, an increase in solar radiation will cause an increase in the amount of ionization of the upper atmosphere and change the flow of ionospheric currents whose strength depends on the conductivity known as the ionospheric dynamo effect. Ionospheric currents at geomagnetically quiet conditions are commonly referred to as Sq currents. At low latitudes, this current system results in a magnetic field variation pattern having a maximum value at noon local time. LAPAN-A3 satellite is the first experimental satellite launched by Indonesia to measure magnetic fields in space. The satellite has a polar orbit at an altitude of 500– 550 km. However, the LAPAN-A3 satellite magnetic field measurement experiment does not carry out magnetic field measurements continuously for 24 h because the satellite is not specifically intended to perform magnetic field measurements. The longest continuous measurement time span was 6 h. This paper aims to validate the magnetic field measured by LAPAN-A3 satellite. This is carried out by analyzing the pattern of magnetic field variations at low latitudes where in the morning local time the magnetic field variation increases and reaches a maximum around midday local time and then decreases around the afternoon.

2 Data and Methods In this paper, we used the magnetic data from magnetometer measurements carried by the LAPAN-A3 satellite where the satellite orbits at the height of the ionosphere. The magnetometer mounted on the satellite records three components of the magnetic field with a resolution of 128 Hz in each direction of the sensor that is mounted orthogonal to each other. Because the attitude of the satellite is always changing, the three components of the magnetic field recorded by the satellite are representations of the magnetic field vector components in an arbitrary orthogonal coordinate system. The satellite only carries one magnetometer, therefore the magnetic field correction related to the satellite’s attitude is carried out using the IGRF model, and in this

Low-Latitude Fluctuation of Ionospheric Magnetic Field Measured …

57

Fig. 1 Fit of the measured magnetic field by LAPAN-A3 satellite with IGRF model

case, the IGRF-13th model has been used. The component of the magnetic field vector is calculated at each time and position of the satellite [4]. Furthermore, the correction results are transformed into the NEC coordinate system using the GEOPACK software package [5–7]. Examples of magnetic field correction results from LAPAN-A3 satellite measurements and the IGRF model are given in Fig. 1. The panels in Fig. 1 represent the poloidal, toroidal, and compressional components of the magnetic field in the satellite’s local coordinate system, respectively. In this paper, the magnetometer data of LAPAN-A3 satellite during 2021 were analyzed. The positions of the satellite are expressed in a magnetic coordinate system, and the data are selected when the satellite is around θ = 10°N and θ = 10°S and latitudinal width of θ = ± 2°. So, we analyzed the data in the latitudinal range between 8°N to 12°N and 8°S to 12°S. During 2021, in that range of latitude, the altitude of LAPAN-A3 is between 500 and 520 km, as shown in Figs. 2 and 3 for the northern and southern hemisphere, respectively. The figures can be seen that the altitude distribution of LAPAN-A3 is almost similar. The magnetic field variation is then calculated as the difference between the measured magnetic field component and the corresponding magnetic field component which is calculated using the 13th IGRF model for 2015. In addition, the position of the satellite in the magnetic coordinate system is expressed in local magnetic time where the calculation is done by utilizing the GEOPACK module written in Python.

58

F. Nuraeni et al.

Fig. 2 Distribution of LAPAN-A3 altitude in the northern hemisphere of 8◦ to 12◦ latitude

Fig. 3 Distribution of LAPAN-A3 altitude in the southern hemisphere of 8◦ to 12◦ latitude

3 Result and Discussions The H (north–south) component of the low-latitude magnetic field increases at midday local time. This increase in the H component occurs due to ionospheric currents flowing at the height of the ionosphere. An example of such an increase in the H component is shown in Fig. 4. The data in the figure are selected on an international geomagnetically quiet day, K p ≤ 1. In addition to these properties, the D component (east–west) has anti-symmetric properties to the equatorial plane, while the Z component has symmetry properties to the equatorial plane. Because the LAPAN-A3 satellite orbits at the height of the ionosphere, the magnetic field measured by the magnetometer on the satellite experiences the effect of ionospheric currents. Magnetometer data are presented in poloidal, toroidal, and

Low-Latitude Fluctuation of Ionospheric Magnetic Field Measured …

59

Fig. 4 H component of magnetic field recorded by ground-based magnetometer at Tanjungsari (blue) and Pontianak (red) stations

compressional representations. At low latitudes, the magnetic field lines are almost parallel to the ionosphere plane so that the poloidal component of the magnetic field almost has a direction parallel to the H component of the ground magnetic field. Therefore, the properties of the low-latitude ionospheric magnetic field can be considered similar to the character of H component of the ground magnetic field. The toroidal and compressional components can also consider having a similar character with the D and Z components of the geomagnetic field. Figures 5, 6 and 7 show the behavior of the poloidal, toroidal, and compressional components as a result of magnetic field measurements by the LAPAN-A3 satellite at low latitudes during 2021. The top panel of each image represents the magnetic field in the northern hemisphere, while the bottom panel represents the magnetic field in the northern hemisphere for the southern hemisphere. Due to the limited measurement, the LAPAN-A3 satellite only recorded data in a fairly short period namely at 07–09 MLT and 20–22 MLT. So the complete daily behavior of each component of the geomagnetic field is difficult to characterize. However, the poloidal components of the magnetic field in the northern and southern hemispheres have opposite trends, as shown in Fig. 5 for data on 07–09 MLT. That is, in the northern hemisphere, there is an increase toward local noon (MLT), while in the southern hemisphere, there is a decrease. These may reflect the magnetic field anti-symmetry of low-latitude poloidal components. The toroidal component also showed the same thing. Meanwhile, in every component around 20–22 MLT, there was no clear trend showed. Figure 7 shows compressional components for the same measurement range as Figs. 5 and 6. At 07–09 MLT, there was an increase of this component in both hemispheres. Data at 20–22 MLT also showed the same thing. This behavior shows evidence that the compressional component is symmetrical about the equatorial plane. From the results mentioned above, we can state that although it is difficult to observe clearly, measurements of LAPAN-A3 satellite are still consistent with the behavior of the geomagnetic field.

60

F. Nuraeni et al.

Fig. 5 Average poloidal component of the magnetic field was measured by the LAPAN-A3 satellite at geomagnetically quiet conditions during 2021

Fig. 6 Average toroidal component of the magnetic field was measured by the LAPAN-A3 satellite at geomagnetically quiet conditions during 2021

The asymmetry of the poloidal and toroidal of ionospheric magnetic field component has an important relationship with the direction of the ionospheric current flow. In the low latitude of northern hemisphere, the ionospheric currents flow eastward during the daytime, and based on asymmetry of magnetic field, the current flows eastward in the southern hemisphere. As a result, asymmetry of poloidal and toroidal components is strongly related to the two vortices ionospheric current systems at low-to mid-latitude in the northern and southern hemispheres. And, the symmetry of compressional components is related to the radial component of the magnetic field and can have an effect on ionospheric plasma transport. However, this statement requires further confirmation.

Low-Latitude Fluctuation of Ionospheric Magnetic Field Measured …

61

Fig. 7 Average compressional component of the magnetic field measured by the LAPAN-A3 satellite at geomagnetic quiet conditions during 2021

4 Conclusion The LAPAN-A3 satellite orbits at an ionospheric altitude where the low-latitude geomagnetic field lines lie almost horizontally in the ionosphere so that variations in the ionospheric magnetic field at low latitudes will show a pattern similar to local time variations at ground-based stations at low latitudes. Based on the result, the poloidal and toroidal components of the magnetic field in the northern and southern hemispheres have opposite trends that may reflect the magnetic field anti-symmetry of low-latitude poloidal and toroidal components, while the compressional component is symmetrical about the equatorial plane. Therefore, measurements of LAPANA3 satellite are still consistent with the behavior of the low-latitude ionospheric geomagnetic field. Acknowledgements This research is funded by Space research center BRIN of 2022 fiscal year.

References 1. Kamide, Y., Balan, N.: The importance of ground magnetic data in specifying the state of magnetosphere–ionosphere coupling: a personal view. Geosci. Lett. 3(1), 10 (2016) 2. Motoba, T., Kikuchi, T., Okuzawa, T., Yumoto, K.: Dynamical response of the magnetosphereionosphere system to a solar wind dynamic pressure oscillation. J. Geophys. Res. Space Phys. 108, A5 (2003) 3. Yamazaki, Y.: Solar and lunar daily geomagnetic variations and their equivalent current systems observed by Swarm. Earth, Planets Space 74, 99 (2022) 4. Musafar, L.M., Yatini, Y., Nuraeni, F., Wellyanita, V., Winarko, A., Juangsih, M., Ratnasari, E.A., Hasbi, W., Utama, S.: Pre-processing data magnetometer satelit LAPAN-A3. Jurnal Sains Dirgantara 19(1), 1–12 (2021)

62

F. Nuraeni et al.

5. Hapgood, M.A.: Space physics coordinate transformations: a user guide. Planet. Space Sci. 40(5), 711–717 (1992) 6. Tsyganenko, N.A.: A new data-based model of the near magnetosphere magnetic field: 1. Mathematical structure. 2. Parameterization and fitting to observations (2001) 7. Tsyganenko, N.A., Sitnov, M.I.: Modeling the dynamics of the inner magnetosphere during strong geomagnetic storms. J. Geophys. Res. 110(A3) (2005)

Study of the Low Latitude Ionosphere Irregularities Using Multi-instrument Observations Dyah Rahayu Martiningrum , Sri Ekawati , Prayitno Abadi , and Bambang Suhandi

Abstract Equatorial plasma bubbles (EPB) or in other term known as equatorial spread F(ESF) are extended zones of depleted F-region plasma density that grow from irregularities caused by the generalized Rayleigh–Taylor instability mechanism in the post-sunset equatorial sector. Study of equatorial plasma bubbles is interesting because plasma density depletions generate rapid changes in both the amplitude and phase of radio signals, which propagate through the bubble, producing ionospheric scintillations that degrade communications and navigation signals. Scintillations come from many sources. Therefore, this study was conducted to understand characteristics of the EPB/ESF and to confirm the source of interference in the case study scintillation events at any given time. In this research, scintillations have analyzed using scintillation index (S4) from GPS Ionospheric Scintillation and TEC Monitor (GISTM) at Kototabang (0.20o S, 100.32o E; 10.36o S dip lat) and Pontianak (0.05o N, 109.34o E; 8.9o S dip lat) for the case March 1 and March 8, 2011. The field aligned irregularities (FAI) obtained from Equatorial Atmosphere Radar (EAR) data. Analysis of ionosonde data from CADI Pontianak and FMCW Kototabang also showed that spread F occurred on March 1 and March 8, 2011. Scintillation index showed strong value at the appropriate time. Further, analysis of EAR and other data will help understanding of equatorial plasma bubbles (EPBs) and scintillation.

1 Introduction Equatorial plasma bubbles (EPBs) or equatorial spread F (ESF) are plasma density irregularities in nighttime low latitude ionosphere with a large range of scale sizes and amplitudes at almost all altitudes along the latitude and longitude sectors. The ESF predominantly occurs at nighttime. The complex dynamical phenomena and D. R. Martiningrum (B) · S. Ekawati · P. Abadi · B. Suhandi Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_6

63

64

D. R. Martiningrum et al.

plasma instability play an important role in development of the formation of EPB/ESF included the background ionospheric conditions [1]. The horizontal geomagnetic field lines at the magnetic equator perpendicular to gravity, and also prevailing natural wind and background electric field, compose unique phenomena to develop the plasma irregularities. These plasma irregularities are generally magnetic field aligned. They have zonal widths of typically a few tens of km and extend along the magnetic field lines for hundreds to thousands of km depending on the peak altitude of the irregularity (bubble) development [2], while their vertical heights range from a few tens of km to several hundred km [3]. Vertical plasma density gradient at the bottom side of ionosphere F-region also controls the growth rate of EPB/ESF. Some previous studies have reported occurrence of EPB/ESF under different geophysical conditions which potentially change the ionospheric parameters [4, 5]. When radio signals propagate through these disturbed regions, they cause scintillation [6]. This results in a fade in received signal power, meaning a loss of signal. The region of equatorial scintillations extends 30° on either side of the earth’s magnetic equator, and the strongest effects are found at approximately 10° N and S [7]. Another ionospheric irregularities can also find in the nighttime. Under certain condition, we can find the irregularity of electron density which is aligned along the earth’s magnetic field. It is known as field aligned irregularity (FAI) that related to EPB/ESF. Another research pointed out that frequent and stronger scintillations associated with EPB during solar maximum period and a drastic reduction in frequency and intensity of scintillation during solar minimum period [8]. This work has been studied the characteristics of the equatorial spread F in associated with the scintillation and field aligned irregularities (FAI) occurrences over Kototabang by using GPS Ionospheric Scintillation and TEC Monitor (GISTM), Equatorial Atmosphere Radar (EAR), and ionosonde data. This study also has been done to understand variability of plasma irregularities of ionosphere, and therefore, it gives general pattern of the relationship between scintillation and plasma irregularities of ionospheric layer.

2 Data and Method Equatorial Atmosphere Radar (EAR) has operated at Kototabang, West Sumatra, and Indonesia (0.20o S, 100.32o E; 10.36o S dip lat). EAR is the large monostatic radar which operate at 47.0 MHz. This instrument developed to study dynamic of lower and upper atmosphere [9] (Fig. 1). Under certain condition, amplitude fading and phase scintillation effects can cause loss of carrier lock and intermittent GPS receiver operation [10]. GPS scintillations generally occur shortly after sunset and may persist until just after local midnight. GISTM used to obtain scintillation index (S 4 ) which indicated occurrences of scintillation. Ionosonde is another instrument was used in this study to confirm the equatorial spread F (ESF) occurrences over Kototabang and Pontianak (Fig. 2).

Study of the Low Latitude Ionosphere Irregularities Using …

65

Fig. 1 Equatorial atmosphere radar specification and location

Fig. 2 GPS ionospheric scintillation and TEC monitor (left) and ionosonde (right)

The Equatorial Atmosphere Radar (EAR) unable to observe the ionosphere region in directions perpendicular to the geomagnetic field for studying F-region FAI associated with equatorial spread F (ESF). Intense FAI echoes have been observed in the nighttime F-region over Kototabang during 2011–2013 and classified in post-sunset and post-midnight occurrences. Furthermore, we have taken two cases of ionosphere irregularity and analyzed possibility relationship between ESF and scintillation between Kototabang and Pontianak.

3 Result and Discussion 3.1 Seasonal Variability of Field Aligned Irregularities (FAI) Characteristics of the EPB/ESF have already been observed as daily percentage of occurrence of field aligned irregularities (FAI). We have used a threshold SNR value of −5 dB. Percentage obtained from ratio between number of signal of particular range and time to total number of signal. Figure 3 is result of radar data analysis

66

D. R. Martiningrum et al.

during 2011–2013 shows that the higher percentage of F-region post-sunset FAI occurred around March (~30%) and September–October (~40%), while F-region post-midnight FAI occurs around June–July (~35–50%). The result clearly confirms dependency of ionosphere irregularities to solar activity.

Fig. 3 Daily percentage of occurrence of F-region echoes

Study of the Low Latitude Ionosphere Irregularities Using …

67

Fig. 4 Scintillation index (S4) of Kototabang (left) and Pontianak (right)

3.2 Ionospheric Scintillation Occurrences During March 2011 The corrected scintillation index (S 4 ) obtained from amplitude scintillation data of GISTM which has already been reduced multipath effects and ambient noise described as  S4 = S42 tot − S42 corr where the S4 is a metric for indicating the amount of variation in the amplitude of a signal; S4 tot is total value of S4 and S4 corr is correction value of S4 . Figure 4 shows scintillation occurrences during March 2011 over Kototabang and Pontianak. Those results have been compared with FAI occurrences from EAR (Kototabang), shows the strong correlation of both data.

3.3 Relationship Between Field Aligned Irregularities (FAI), Scintillation, and Equatorial Spread F (ESF) During March 2011, we founded strong relationship between FAI and scintillation events over Kototabang. We also observed scintillation event which possible associated with ionospheric irregularities over Pontianak. Pontianak placed about 1058.17 km eastern of Kototabang. For further analysis, we took two cases, March 1, 2011, and March 8, 2011. The echoes scatter were analyzed to obtain zonal distance by using distance approximation of two location at the certain height over the earth’s surface. Each of bubbles of SNR plot named as A, B, C, D, and E (Fig. 5). Drift velocity determined from

68

D. R. Martiningrum et al.

zonal distance of each bubble and gradient of time. Possibility of period of bubbles propagate from Kototabang to Pontianak was obtained from simple relation between distance and drift velocity. Table 1 shows the result of all bubbles. The echoes scatter of Bubble A shows the fastest drift velocity and Bubble B is the slowest. By using the similar relation between distance, velocity, and time, will be obtained period/time which needed all of bubbles for propagating from Kototabang to Pontianak (Table 2). The virtual height over Kototabang and Pontianak from ionosonde data also increased at that time. It means that electron density of F-region of ionosphere layer over Kototabang and Pontianak fluctuated or changed mainly for March 1 and March 8, 2011 (red line in Fig. 6).

Fig. 5 Zonal distance from SNR plot of echoes scatter EAR for March 1, 2011, (left) and March 8, 2011 (right)

Table 1 Date and time of FAI from EAR

Study of the Low Latitude Ionosphere Irregularities Using …

69

Table 2 Characteristic of each bubble No.

Date

Bubble’s name

Drift velocity (m/s)

Period/time to propagate (h:m)

1

March 1, 2011

A

208.3

1:24

2

March 1, 2011

B

72.46

4:36

3

March 8, 2011

C

166.67

1:48

4

March 8, 2011

D

138.89

2:06

5

March 8, 2011

E

98.04

3:00

Fig. 6 Virtual height of F-region for Kototabang and Pontianak

4 Conclusion We studied characteristic of ionospheric irregularity by using three different instruments that is Equatorial Atmosphere Radar Kototabang, GPS Ionospheric Scintillation and TEC Monitor (GISTM) at Kototabang and Pontianak, and also ionosonde at Kototabang and Pontianak. The results obtained from present study suggest that.

70

D. R. Martiningrum et al.

• Radar data analysis during 2011–2013 shows that the higher percentage of Fregion post-sunset FAI occurred around March (~30%) and September–October (~40%), while F-region post-midnight FAI occurred around June–July (~35– 50%). The result clearly confirms dependency of ionosphere irregularities to solar activity. • The relationship between the scintillation and ionospheric irregularities (ESF) found from analysis GISTM and ionosonde data over Kototabang and Pontianak. Variability of virtual height also shows increasing of F-region’s virtual height when equatorial spread F and scintillation occurred. • The further analysis of echoes scatter of EAR for certain case shows characteristics of bubbles at certain height (350 km), their zonal distance and possibility the relationship of ionospheric irregularity between Kototabang and Pontianak that have longitudinally spacing about 1058.17 km. Acknowledgements The authors thank Research Institute for Sustainable Humanosphere and Yamamoto Laboratory, Kyoto University for supporting EAR data.

References 1. Kelley, M.C.: The Earth’s Ionosphere: Plasma Physics and Electrodynamics, 2nd edn. Academic Press, Amsterdam (2009) 2. Sobral, J.H.A., Abdu, M.A., Takahashi, H., Taylor, M.J., de Paula, E.R., Zamlutti, C.J., de Aquino, M.G., Borba., G.L.: Ionospheric plasma bubble climatology over Brazil based on 22 Years (1977–1998) of 630 nm airglow observations. J. Atmos. Terr. Phys. 64, 1517–1524 (2002) 3. Labelle, J., Jahn, J.M., Pfaff, R.F., Swartz, W.E., Sobral, J.H.A., Abdu, M.A., Muralikrishna, P., de Paula., R.: The Brazil/guara equatorial spread F campaign: results of the large-scale measurements. Geophys. Res. Lett. 24(13), 1691–1694 (1997) 4. Abdu, M.A.: Day to day and short-term variabilities in the equatorial plasma bubble/spread F irregularity seeding and development. Prog. Earth Planet. Sci. (2019). https://doi.org/10.1186/ s40645-019-0258-1 5. Yamamoto, M., Otsuka, Y., Jin, H., Miyoshi, Y.: Relationship between day-to-day variability of equatorial plasma bubble activity from GPS scintillation and atmospheric properties from ground-to-topside model of atmosphere and ionosphere for aeronomy (GAIA) assimilation. Prog. Earth Planet. Sci. (2018). https://doi.org/10.1186/s40645-018-0184-7 6. Huang, N., Cheng, K., Huang, W.T.: Seasonal and solar cycle variation of spread F at the equator anomaly crest zone. J. Geomag. Geoelectr. 39, 639–657 (1987) 7. Wanninger, L.: Effects of equatorial ionosphere on GPS. GPS World 4(7), 48–52, 54 (1993). ISSN 1048-5104 8. Basu, S., Basu, S., Stubbe, P., Kopka, H., Waaramaa, J.: Daytime scintillations induced by highpower HF Waves at Tromso, Norway. J. Geophys. Res. Space Phys. 92(A10), 11149–11157 (1978) 9. Fukao, S., Hashiguchi, H., Yamamoto, M., Tsuda, T., Nakamura, T., Yamamoto, M.K., Sato, T., Hagio, M., Yabugaki, Y.: Equatorial atmosphere radar (EAR): system description and first results. Radio Sci. 38(3), 1053 (2003). https://doi.org/10.1029/2002RS002767 10. Klobuchar, J.A.: Ionospheric effects on GPS. GPS World 2, 48–51 (1991)

Distribution of Peat Soil Carbon Under Different Land Uses in Tidal Swampland Nur Wakhid and Siti Nurzakiah

Abstract Indonesia has the largest area of tropical peatlands in Southeast Asia and is an essential component of the global carbon (C) pool through its large C deposit inside the peat. Peat soil in Indonesia was usually associated with swampland both in freshwater and tidal swampland. However, the assessment of peat C deposit on swampland remains uncertain. Therefore, this study aimed to assess the C deposit of peat soil established on tidal swampland. The study was conducted in 2012 at Pulang Pisau, Central Kalimantan, under three different land uses, intercropping of rubber and pineapple, rubber plantation mixed with shrubland, and shrubland. The depth of peat soil was in the range of 4 to 8 m, dominated by sapric and hemic. The mean of C deposit was 3963 ± 561, 6666 ± 1065, and 4325 ± 605 t ha−1 , respectively, in intercropping of rubber and pineapple, rubber plantation mixed with shrubland, and shrubland. C deposit in a rubber plantation mixed with shrubland was high because the peat depth was deeper than other land uses. We expect this finding could be used as a baseline information for sustainable peatland management cultivated for agricultural plantation.

1 Introduction Tropical peatlands are an important component of the global carbon (C) deposit. They have a huge C deposit both above–ground through plant biomass and below-ground underlying in the soil. It has been estimated that the tropical peatlands store up to 104.7 Petagrams of C below the soil, even though only about 17% of global peatland areas [1, 2]. More than 50% of the tropical peatland area is located in Southeast N. Wakhid (B) Research Center for Ecology and Ethnobiology, National Research and Innovation Agency, Cibinong, Bogor 16911, Indonesia e-mail: [email protected] S. Nurzakiah Research Center for Food Crops, National Research and Innovation Agency, Cibinong, Bogor 16911, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_7

71

72

N. Wakhid and S. Nurzakiah

Asia, with the vast majority in Indonesia and Malaysia [2]. The total peatland area in Indonesia was estimated to be 13.4 million ha distributed in Sumatra, Kalimantan, Sulawesi, and Papua islands [3]. In Indonesia, peat soil is usually associated with swampland under water-logged conditions, both in fresh water and or tidal swampland. The logged conditions bring through the slow decomposition of organic matter under an anaerobic situation. Today’s C deposit on tropical peat soil is the result of organic matter accumulation over thousands of years from the forest that grows on the top of soil under these logged conditions. However, the huge C deposit underlying peat soil is now vulnerable due to land conversion. During the last decades, many areas of tropical peatlands have been converted mainly for plantations [4]. The land conversion is usually accompanied by drainage construction to shift the water-logged conditions to low groundwater levels. As a result, the peat soil characteristic would be changed [5]. Also, the peat soil C deposit is potentially decreased drastically. Therefore, it is crucial to quantify the C deposit of peat soil to determine the effects of land conversion. The objective of this study was to quantify the distribution of peat soil C deposits under different land uses in tidal swampland.

2 Methods 2.1 Study Site The study was conducted on three different land uses (1) an intercropping of rubber and pineapple, (2) rubber mixed with shrubland, and (3) shrubland, on the same area established on peat soil in Pulang Pisau, Central Kalimantan, Indonesia (see Figs. 1 and 2). The distance between each land use is about 100 m. The study site originally was a peat swamp forest that was degraded due to the failure of the mega rice project in the late 1990s.

Fig. 1 Illustration of peat soil sampling plots

Distribution of Peat Soil Carbon Under Different Land Uses in Tidal …

73

Fig. 2 Study plots for peat soil sampling, a intercropping of rubber and pineapple, b rubber mixed with shrubland, and c shrubland

2.2 Carbon Deposit Measurement C deposit was estimated from soil sampling that collected in 2012 within five plots on each land use, using a peat auger (Eijkelkamp). The plots were on the same perpendicular transect at 75 m intervals between each plot to cover the different depths of the peat (see Fig. 1). In each plot, the samples were collected deeply until the peat auger reached the substratum. The samples were stratified based on the maturity of the peat. Thus, the depth interval of the samples was different between each plot. Also, the groundwater level in each plot was measured simultaneously during peat soil sampling. Peat maturity was estimated directly in the field using three indicators: decomposition stage, color, and squeezing the peat soil in the palm [6]. The peat maturity was divided into sapric (mature), hemic (medium), and fibric (immature). After stratification, each soil sample at different depth interval was separated, tagged, and stored in plastic bags for the next analysis in laboratory. The C deposit was quantified from C content, peat bulk density, and peat depth. The peat depth was determined directly in the field during peat soil sampling. Whereas C content and BD were estimated at laboratory for each soil sample at different depth intervals. The C content was determined using a conversion factor of 0.58 from the percent of organic matter to organic C [6]. The percent of organic matter

74

N. Wakhid and S. Nurzakiah

was measured using the loss on ignition (LOI) method at 550 °C in a muffle furnace. BD was estimated using gravimetric methods by weighing the oven dry weight of peat [6]. The C deposit (CS, t ha−1 ) was calculated following the equation [6]: CS = CD × PV

(1)

CD = BD × (CC/100)

(2)

PV = PD × 10000

(3)

where CD is C density (t m−3 ), PV is peat volume in each layer (m3 ha−1 ), BD is bulk density (g cm−3 ), CC is C content (% in mass), and PD is peat depth or peat thickness (m).

2.3 Data Analysis The difference between each plot was compared using Tukey’s HSD following analysis of variance (ANOVA). Furthermore, the linear regression and Pearson correlation methods were used to analyze the relationship between peat depth and C deposit. All data analysis was conducted using Microsoft Excel and Real Statistics Resource Pack software (Charles Zaiontz, www.real-statistics.com).

3 Results and Discussion The peat soil in all study plots was dominated by sapric and hemic. The peat depth varied from 5.23–5.74, 6.16–8.50, and 4.95–5.46 m, respectively, in intercropping of rubber and pineapple, rubber mixed with shrubland, and shrubland (see Fig. 3a). The deepest peat soil was recorded in rubber mixed with shrubland up to 8.50 m. Mean of peat depth was 5.50, 7.12, and 5.16 m in intercropping of rubber and pineapple, rubber mixed with shrubland, and shrubland, respectively. The peat depth in rubber mixed with shrubland was significantly different with intercropping of rubber and pineapple, and shrubland (p < 0.05). The lowest average of bulk density (BD) was found in rubber and pineapple with a range of BD from 0.12 to 0.18 g cm−3 (see Fig. 3b). The BD in rubber and pineapple was lower than that of BD of peat soil in a nearby rubber plantation that measured from soil samples collected by soil ring sample [7]. In this study, BD was measured from soil samples collected using a peat soil auger. Soil sampling method is likely to affect the result of BD measurement. The mean of BD was 0.16, 0.19, and 0.20 g cm−3 , respectively, in intercropping of rubber and pineapple, rubber mixed with shrubland, and shrubland. BD of peat soil in this study was in the range of

Distribution of Peat Soil Carbon Under Different Land Uses in Tidal …

75

Fig. 3 Variations of a Peat depth, b Bulk density (BD), and c Carbon (C) content, at different plots in a intercropping of rubber and pineapple (gray square with solid line), b rubber plantation mixed with shrubland (black circle with solid line), and c shrubland (black triangle with dashed line). Vertical bars in (b) and (c) denote 1 standard deviation

BD in Sumatra and Kalimantan of 0.02–0.03 g cm−3 [6]. BD in intercropping of rubber and pineapple was significantly different from BD in shrubland (p < 0.05), but not for rubber mixed with shrubland. In intercropping of rubber and pineapple and rubber mixed with shrubland, the highest BD was measured near the canal, but not for shrubland. BD in a degraded tropical peatland was reported to increase near the canal [8]. In shrubland, as an abandoned peatland, the drainage canal is likely not intense as in intercropping of rubber and pineapple and rubber mixed with shrubland. Thus, the effect of drainage on BD was less. Mean of C content in all plots was similar in the range of 46, 49, and 43%, in intercropping of rubber and pineapple, rubber mixed with shrubland, and shrubland, respectively. The mean C content was in the range of peat C content across Southeast Asian peatlands (41.6–62%) [2, 9]. The C content varied from 40 to 51% in all plots (see Fig. 3c). The C content in rubber mixed with shrubland showed a significant difference with shrubland (p < 0.05), but in intercropping of rubber and pineapple was not significantly different. Shrubland has the highest BD with the lowest C content (see Table 1). The highest value of BD is usually associated with low C content [2]. Also, high BD with low C content may indicate a more advanced degree of peat decomposition at the site [10].

76

N. Wakhid and S. Nurzakiah

Table 1 C deposit distribution under different land uses (mean ± 1 SD) Plot

Peat depth (m)

Sample number

BD (g cm−3 )

Carbon content (%)

C deposit (t/ha)

Rubber and pineapple 1

5.50

n = 11

0.18 ± 0.09

47 ± 8

4494

2

5.64

n = 13

0.16 ± 0.06

49 ± 4

4465

3

5.38

n = 11

0.15 ± 0.03

44 ± 6

3424

4

5.74

n = 12

0.12 ± 0.06

47 ± 7

3319

5

5.23

n = 11

0.17 ± 0.05

47 ± 5

4113 3963 ± 561

Rubber with shrubland n = 14

0.21 ± 0.08

51 ± 3

7288

8.5

n = 12

0.19 ± 0.07

47 ± 10

7540

7.89

n = 11

0.18 ± 0.08

50 ± 6

7225

4

6.32

n = 12

0.20 ± 0.07

49 ± 5

6333

5

6.16

n = 12

0.17 ± 0.05

47 ± 6

4945

1

6.75

2 3

6666 ± 1065 Shrubland 1

5.09

n = 10

0.17 ± 0.06

48 ± 5

4092

2

5.21

n = 11

0.18 ± 0.06

42 ± 9

3978

3

5.09

n=5

0.23 ± 009

41 ± 8

4744

4

5.46

n=4

0.24 ± 0.09

40 ± 6

5149

5

4.95

n = 10

0.17 ± 0.06

43 ± 10

3664 4325 ± 605

Mean of C deposit was estimated to be 3963 ± 561, 6666 ± 1065, and 4325 ± 605 t ha−1 (mean ± 1 standard deviation (SD)), respectively, in intercropping of rubber and pineapple, rubber mixed with shrubland, and shrubland (see Table 1). The largest C deposit was estimated in rubber mixed with shrubland. The C deposit in rubber mixed with shrubland was significantly different with intercropping of rubber and pineapple and shrubland. Deep peat soil would result in a higher peat C deposit (see Fig. 3a and Table 1). The relationship between peat depth and C deposit was significant if using all data (n = 15, p < 0.05) from three plots. However, if only using data from each plot (n = 5), the relationship between peat depth and C deposit was not significant (p > 0.05). The peat depth is probably the most factor that affected the amount of peat C deposit estimations on tropical peatland. However, the peat depth is likely also the most contributor to uncertainty on peat C deposit estimations since the peat depth measurements are often estimated from a few samples that may not represent the heterogeneity of the peat ecosystem [11]. Peat C deposit estimations may also be overestimated, as the assessment has not yet considered the mineral layer that may

Distribution of Peat Soil Carbon Under Different Land Uses in Tidal …

77

exist under the peat layers. Peat C deposit in this study (3963 ± 561, 6666 ± 1065, 4325 ± 605 t ha−1 ) was higher than the peat C deposit of rubber plantations on freshwater swampland in South Kalimantan (2025.21 ± 132.77 t ha−1 , peat depth: 3.1–4.7 m) [12]. The lower C deposit in that study was probably due to the shallow peat depth compared to our study.

4 Conclusion Peat carbon (C) deposit was estimated in a tidal swampland under three different land uses, intercropping of rubber and pineapple, rubber mixed with shrubland, and shrubland. Rubber mixed with shrubland has the deepest peat soil with the highest peat C deposit up to 6666 ± 1065 t ha−1 . The peat C deposit estimations on swampland are facing some uncertainties as the heterogeneity of peatland was high. Also, the peat depth measurement as the most factor for peat C estimations required a lot of effort and was expensive. Therefore, further study is needed to obtain a new method or model for inexpensive and fast peat C deposit assessment.

References 1. Dargie, G.C., Lewis, S.L., Lawson, I.T., Mitchard, E.T.A., Page, S.E., Bocko, Y.E., Ifo, S.A.: Age, extent and carbon storage of the central Congo Basin peatland complex. Nature 542, 86–90 (2017) 2. Page, S.E., Rieley, J.O., Banks, C.J.: Global and regional importance of the tropical peatland carbon pool. Glob. Change Biol. 17, 798–818 (2011) 3. Anda, M., Ritung, S., Suryani, E., Sukarman, S., Hikmat, M., Yatno, E., Mulyani, A., Subandiono, R.E., Suratman, S., Husnain, H.: Revisiting tropical peatlands in Indonesia: semi-detailed mapping, extent and depth distribution assessment. Geoderma 402(115235), 1–14 (2021) 4. Gaveau, D.L.A., Sheil, D., Husnayaen, Salim, M.A., Arjasakusuma, S., Ancrenaz, M., Pacheco, P., Meijaard, E.: Rapid conversions and avoided deforestation: examining four decades of industrial plantation expansion in Borneo. Sci. Rep. 6, 32017 (2016) 5. Furukawa, Y., Inubushi, K., Ali, M., Itang, A.M., Tsuruta, H.: Effect of changing groundwater levels caused by land-use changes on greenhouse gas fluxes from tropical peatlands. Nutr. Cycl. Agroecosyst. 71, 81–91 (2005) 6. Agus, F., Hairiah, K., Mulyani, A.: Measuring Carbon Stock in Peat Soils: Practical Guidelines. World Agroforestry Centre (ICRAF) Southeast Asia Regional Program, Indonesian Centre for Agricultural Land Resources Research and Development, Bogor, Indonesia, 60 pp. (2011) 7. Wakhid, N., Hirano, T., Okimoto, Y., Nurzakiah, S., Nursyamsi, D.: Soil carbon dioxide emissions from a rubber plantation on tropical peat. Sci. Total Environ. 581–582, 857–865 (2017) 8. Sinclair, A.L., Graham, L.L.B., Putra, E.I., Sahaijo, B.H., Applegate, G., Grover, S.P., Cochrane, M.A.: Effects of distance from canal and degradation history on peat bulk density in a degraded tropical peatland. Sci. Total Environ. 699, 134199 (2020) 9. Page, S.E., Wüst, R.A.J., Weiss, D., Rieley, J.O., Shotyk, W., Limin, S.H.: A record of Late Pleistocene and Holocene carbon accumulation and climate change from an equatorial peat bog (Kalimantan, Indonesia): implications for past, present and future carbon dynamics. J. Q. Sci. 19, 625–635 (2004)

78

N. Wakhid and S. Nurzakiah

10. Anshari, G.Z., Afifudin, M., Nuriman, M., Gusmayanti, E., Arianie, L., Susana, R., Nusantara, R.W., Sugardjito, J., Rafiastanto, A.: Drainage and land use impacts on changes in selected peat properties and peat degradation in West Kalimantan Province, Indonesia. Biogeosciences 7, 3403–3419 (2010) 11. Warren, M.W., Kauffman, J.B., Murdiyarso, D., Anshari, G., Hergoualc’h, K., Kurnianto, S., Purbopuspito, J., Gusmayanti, E., Afifudin, M., Rahajoe, J., Alhamd, L., Limin, S., Iswandi, A.: A cost-efficient method to assess carbon stocks in tropical peat soil. Biogeosciences 9, 4477–4485 (2012) 12. Nurzakiah, S., Noor, M., Nursyamsi, D.: Carbon stock stratification of peat soils in South Kalimantan, Indonesia. J. Wetlands Environ. Manag. 2(2), 55–59 (2014)

Study of Air Quality and Pollutant Distribution Patterns in Balikpapan Using the WRF-Chem Model Dessy Gusnita, Prawira Yuda Kumbara, Waluyo Eko Cahyono, Angga Yolanda Putra, Fahmi Rahmatia, and Jen Supriyanto

Abstract Indonesia faces serious problems related to air quality, in line with the plan to move the capital to East Kalimantan, so it is necessary to study of air quality in the area around the new capital. This paper aims to study air quality using insitu data from the Balikpapan Environmental Service (DLH Balikpapan) and pollutant distribution using the Weather Research Forecasting/Chemistry (WRF-Chem) model. This model combines topographic, atmospheric, and emission data that can produce a simulation of the distribution and concentration of pollutants in an area at a time. Simulations were carried out in January (14–16), April (21–23), July (16– 18), and October (5–7) 2020, with spatial resolution: 5 km (domain 1) and 1 km (domain 2) and hourly time. In-situ DLH Balikpapan data was obtained from AQMS tools installed at 2 points, namely Balikpapan Baru (BB-1.2422744, 116.8596525) and Plaza Balikpapan (PB-1.2771924, 116.8382858) covering pollutant parameters PM2.5, NO2 , and SO2 using hourly data. The results of the 2020 AQMS monitoring at BB station, indicated that PM2.5 concentration reached a peak concentration at 22.00 at 21.2 µg/m3 in April. At the PB station, the PM2.5 peak concentration was lower at 16.0 µg/m3 in January. The peak of SO2 in the PB occurred in July at 07.30 Central Indonesia time with a concentration of 229.3 µg/m3 and in the BB, the peak concentration of SO2 was at 48.9 µg/m3 , which occurred in October at 14:30 Central Indonesia time. Meanwhile, the peak of NO2 occurred at 07:00 Central Indonesia time in PB on July 2020, with concentrations reaching 85.18 µg/m3 . The results of the WRF-Chem model show that the distribution of pollutants with the highest concentration is found in the northern region, which is around Tenggarong, both for PM2.5, SO2, NO2 .

D. Gusnita (B) · P. Y. Kumbara · W. E. Cahyono · A. Y. Putra · F. Rahmatia National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected] J. Supriyanto Environmental Agency, Balikpapan, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_8

79

80

D. Gusnita et al.

1 Introduction Nowadays, the city of Balikpapan is growing quite rapidly, both in the construction of infrastructure, buildings, companies, the number of residents, the number of immigrants, the number of motorized vehicles circulating on the highway every day, especially private cars [1]. As the population grows, the demand increases, giving rise to an increase in urban activity. This increase causes the movement of people and goods which is called transportation [2]. According to the Directorate General of Traffic of the East Kalimantan Police, the number of vehicles in Balikpapan has increased by an average of 6.8%, so that in 2017 the number of vehicles in Balikpapan was 572,976 vehicles [3]. The high consumption of fuel oil in the transportation sector results in a higher potential for air pollution because the air will be polluted by exhaust gases resulting from combustion [4]. The distribution of atmospheric pollutants depends on the direction and speed of the wind, wind speed, air temperature, and humidity are part of the meteorological parameters that can affect the concentration of pollutant gases in the air [5]. One of the models commonly used to model ambient air conditions is Weather Research and Forecasting with Chemical (WRF-Chem). WRF-Chem is a regional scale weather model that models meteorological conditions and their interactions with chemical compounds in the atmosphere spatially and temporally. WRF-Chem simulation using regional emission inventories can show better results than global scale models [6]. To get the distribution of pollutant emissions in an area at a time, a dynamic model can be used. Weather Research Forecasting Chemistry (WRF-Chem) is a combined model of atmospheric modelling with atmospheric chemistry. The distribution of pollutant emissions is influenced by various factors, one of which is meteorological factors such as wind, temperature, atmospheric stability, and so on. Several studies in Indonesia using the WRF-Chem model to analyse the dispersion pattern of particulates and Sulphur dioxide using the WRF-Chem model around the industrial areas of Tangerang and Jakarta [7] and research on the analysis of the effect of seasonal variations on NO2 dispersion in the city of Tangerang [8]. Many studies use the WRF-Chem model to simulate and predict the distribution of pollutant emissions in an area over a certain period of time [9–11] WRF-Chem is a mesoscale dynamic model that combines atmospheric models with chemical models that include calculations of atmospheric chemical processes, photolysis, deposition, dispersion and mixing of emissions, aerosol interactions, and radiation [9, 10, 12, 13]. This model combines topographic, atmospheric, and emission data to produce a simulation of the distribution and concentration of emissions in an area at a time. This paper aims to analyse in-situ pollutant data in the city of Balikpapan which consists of PM2.5, SO2, and NO2 based on seasons, then to determine the pattern of pollutant distribution using the WRF-Chem model. It is hoped that the results of this study will provide input for the local government in supporting sustainable development.

Study of Air Quality and Pollutant Distribution Patterns in Balikpapan …

81

2 Data and Method This study uses in-situ data from the Balikpapan Environmental Service (DLH Balikpapan) from measurements of Air Quality Monitoring System (AQMS) consisting of: PM2.5, SO2 , and NO2 in the city of Balikpapan in 2020. AQMS is installed at two measurement points, namely Balikpapan Baru (BB-1.242744, 116.8596525) and Plaza Balikpapan. (PB-1.2771924, 116.8382588). In-situ data were analysed based on seasonal data, namely the rainy season, transition season, and dry season. Furthermore, to complete the in-situ data, a pollutant distribution model was run using the WRF-Chem model to see the direction of the distribution of pollutants from the city of Balikpapan. Simulation Setup This study did a simulation by using dynamical modelling to learn chemical emission concentration and dispersion. We use the WRF-Chem model to run the simulation. WRF-Chem is a mesoscale dynamic model that combines atmospheric models with chemical models that include calculations of atmospheric chemical processes, photolysis, deposition, emission dispersion and mixing, aerosol interactions, and radiation [9, 10, 12]. This model combines topographic, atmospheric, and emission data so that it can produce a simulation of the distribution and concentration of emissions in an area at a time. This study used two nested domains for the simulation and take study area in East Borneo. The spatial resolution for both domains is 5 and 1 km with temporal resolution is hourly. The domain design can be seen at Fig. 1. For meteorological input, this study used Final Reanalysis (FNL) data that can be downloaded at: http://rda.ucar.edu/datasets/ds083.2. The spatial resolution FNL data is 0.25 degree by 0.25 degree temporal resolution is six hourly. Then, for the emission data input, this study used two kinds of data, both of EDGAR-HTAP for anthropogenic emission and MEGAN for biogenic emission. Furthermore, this study used some parameterization schemes for simulation setup that can be seen in Table 1. Fig. 1 The domain design for simulation study

82 Table 1 The parameterization schemes for simulation setup

D. Gusnita et al. Parameterization

Domain 1

Domain 2

Cumulus

Grell 3

–

Microphysics

Purdue Lin

Purdue Lin

Boundary layer

YSU

YSU

Radiation (shortwave and longwave)

RRTMG

RRTMG

Surface

NOAH MP

NOAH MP

Chemical

RADM2 and MADE/SORGAM aerosols

RADM2 and MADE/SORGAM aerosols

The simulation was conducted in January (14–16), April (21–23), July (16–18), and October (5–7) 2020.

3 Results and Discussion 3.1 Concentration of Air Pollutants in the City of Balikpapan Balikpapan City DLH AQMS equipment is installed in 2 separate places, namely in the Balikpapan Baru area (BB-1.242744, 116.8596525) and in the Plaza Balikpapan area (PB-1.2771924, 1168382858). AQMS BB station is installed in residential areas, while AQMS PB is located on the Balikpapan city road, adjacent to shopping centres. The air pollutant parameters measured by AQMS DLH Balikpapan are PM2.5, SO2, and NO2 . PM2.5 Concentration Pattern in the observation area of Balikpapan Baru (BB) and Plaza Balikpapan (PB) Based on seasonal data, namely: rainy season (JF), transitional season (MAM), dry season (JJA), and transitional season (SON), analysis of three pollutant parameters, namely PM2.5, SO2, and NO2 based on hourly data, was carried out. This pollutant data analysis intends to find out the pattern of daily pollutant concentrations in each season. The results of monitoring PM2.5 concentrations from AQMS equipment at two points (Balikpapan Baru/BB and Plaza Balikpapan /PB) are shown in Fig. 2. At BB observation station the maximum PM 2.5 concentration occurred in January, concentrations peaked at 07.00 to 08.00 central Indonesia time in the morning, then peaked again at 16.00–17.00 central Indonesia time in the afternoon. The high concentration of PM2.5 in January is thought to have come from transportation activities in the city of Balikpapan in the morning and evening. In addition, the peak of PM2.5 concentration occurs again at 21.00–22.00 central Indonesia time at night.

30

January April July October

20

PB

10

0 00:00

5:00

10:00 15:00 Time

Concentration (µg/m3)

Concentration (µg/m3)

Study of Air Quality and Pollutant Distribution Patterns in Balikpapan … 30

January April July October

20

BB

10

0 00:00

20:00

83

5:00

10:00 15:00 Time

20:00

Fig. 2 PM2.5 concentration at the Balikpapan Baru (BB) and Balikpapan Plaza (PB) observation stations. Source Data from the Balikpapan Environmental Service

At the PB observation station, the maximum PM2.5 concentration occurred in April, peaking at 07.00–08.00 central Indonesia time in the morning, peaked again at 16.00–17.00 central Indonesia time in the afternoon. This is presumably, due to the transportation activities of the people of Balikpapan in the morning and evening. In addition, the peak of PM2.5 concentration occurred again at 19.00 central Indonesia time at night. The Pattern of NO2 Concentration in the observation area of Balikpapan Baru (BB) and Plaza Balikpapan (PB) The concentration of NO2 in PB is much higher than at the BB observation station, this is presumably, because PB’s location is in a fairly busy shopping centre, so the mobility of people who emit NO2 pollutants is much higher than BB’s located in residential areas (Fig. 3). The pattern of SO2 concentration in the observation area of Balikpapan Baru (BB) and Plaza Balikpapan (PB)

July October

90

PB

80

70 0:00

5:00

10:00 15:00 Time

20:00

Concentration (µg/m³)

Concentration (µg/m³)

In Fig. 4, the concentration of SO2 reaches 229.3 µg/m3 in the dry season, it is suspected that this is due to the lack of cleaning of pollutants during the dry season. 90 80 70 60 50 40 30 20 10 0

July October

00:00

5:00

10:00 15:00 Time

BB

20:00

Fig. 3 NO2 concentration at the Balikpapan Baru (BB) and Balikpapan Plaza (PB) Observation stations. Source data from the Balikpapan Environmental Service

400

January April July October

300

PB

200 100

Concentration (µg/m³)

D. Gusnita et al. Concentration (µg/m³)

84 100

July October

BB

50

0

0 00:00

5:00

10:00 15:00 Time

20:00

00:00

5:00

10:00 15:00 Time

20:00

Fig. 4 SO2 concentration at the Balikpapan Baru (BB) and Balikpapan Plaza (PB) Observation stations. Source data from the Balikpapan Environmental Service

As a result, there is an increase in pollutant concentrations. Meanwhile, in the rainy season (January) there was a significant decrease in SO2 concentration due to the washing out process of rainwater. However, in general, both BB and PB stations peak at 07.00 central Indonesia time in the morning.

3.2 Pollutant Distribution in Dry Season, Transition Season, and Wet Season The WRF-Chem running model was carried out on 14 Jan, 21 April, 16 July, and 05 October 2020 at 07.00, 13.00, and 19.00 Central Indonesia time for three pollutant parameters PM10, SO2, and NO2 . The results of running the WRF-Chem model to analyse the distribution of pollutants in the city of Balikpapan are presented in Fig. 5. In April, the highest PM2.5 concentration occurred in the morning when, atmospheric conditions tended to be stable, and during the day the concentration reaches 18 µg/m3 . In July, the concentration of PM2.5 was still quite high, especially in the Tenggarong area towards the north and its surroundings. The highest concentrations of SO2 and NO2 in the morning are in the northern area of Balikpapan, especially the Tenggarong area, where there is a coal mining centre which emits high levels of NO2 , SO2, and PM10 pollutants from their activities. During the day, NO2 and SO2 did not show high concentrations, and were more vertically dispersed due to unstable atmospheric conditions. The height of the stable layer during the day is higher than at night due to turbulence in the mixed layer, which causes vertical transport in the atmosphere [8].

Study of Air Quality and Pollutant Distribution Patterns in Balikpapan … PM 2.5 (µg/m3), 14 January 2020

PM 2.5 µg/m3), 21 April 2020

PM 2.5 (µg/m3), 16 Juli 2020

PM 2.5 (µg/m3), 05 October 2020

SO2 (µg/m3), 14 January 2020

SO2 (µg/m3), 16 July 2020

NO2 (µg/m3), 14 January 2020

NO2 (µg/m3), 16 Juli 2020

85

SO2 (µg/m3), 21 April 2020

SO2 (µg/m3), 05 October 2020

NO2 (µg/m3), 21 April

NO2 (µg/m3), 05 October 2020

Fig. 5 Pollutant distribution patterns in Balikpapan City and surrounding areas result of running WRF-Chem model

3.3 Wind Characteristics of Balikpapan City in Balikpapan Baru and Plaza Balikpapan Balikpapan Baru (BB) Wind direction conditions on 14–16 January 2020 (rainy season) show that the dominant wind comes from the north. The wind speed is relatively low because to the north of Balikpapan there are hills and mountains. 21–23 Apr 2020 (transition season). As in January, April 21–23, the dominant wind comes from the north. In the dry season,

86

D. Gusnita et al.

Balikpapan Baru (a)

Plaza Balikpapan (b) Fig. 6 Balikpapan City wind characteristics in BB (a) and PB (b) the results of running the WRFChem model

July 16–18, the direction of the wind in July is more varied. The most dominant winds are from the North-northwest and southeast. On 05–07 October 2020, the dominant wind direction comes from the south-southwest. The wind speed is greater because the wind directly comes from the sea area (Fig. 6). Balikpapan Plaza (PB) Wind conditions on 14–16 Jan 2020, dominant wind comes from the north, indicating the influence of the Asian monsoon. The wind speed is relatively low because to the north there are hills and mountains. 21–23 April 2020. As in January, April 21–23, the dominant wind comes from the north. 16–18 July 2020, the direction of the wind in July is more varied. And the most dominant comes from the North-northwest and southeast. On 05–07 Oct 2020. The dominant wind direction is from the southsouthwest. The wind speed is significant because the wind direction comes from the sea. The rate is higher than the BB point because the BP point is closer to the sea.

4 Summary and Conclusion Air quality simulations in the dry season, transition season, and wet season in the city of Balikpapan show that the distribution of PM10, NO2, and SO2 pollutants is strongly influenced by meteorological parameters, including wind direction and speed, atmospheric stability, and rainfall. In general, the dispersion of pollutants in the city of Balikpapan follows the pattern of local wind direction movements that affect the wet season, transition season, and dry season. Based on the distribution simulation from WRF-Chem, it shows that the Tenggarong and surrounding areas have the potential to have high pollutant concentrations. This information on air

Study of Air Quality and Pollutant Distribution Patterns in Balikpapan …

87

quality and pollutant distribution is expected to be an input for local governments to mitigate air quality in their area. The conclusion of the in-situ AQMS data analysis at the BB and PB Balikpapan stations is that, the pollutant concentrations of SO2 and NO2 at the Balikpapan PB station have higher concentrations than those at the Balikpapan Baru station, while the highest PM2.5 concentration is at the Balikpapan Baru observation point.

References 1. Winarni, Sistem Transportasi Massal, Kebutuhan Mutlak. Kaltimpost, hal. 25 [Online] tersedia di http://kaltim.prokal.co/read/news/298347-sistem-transportasi-massal-kebutuhanmutlak.html diakses pada tanggal 26 September 2021 (2017) 2. Al Hadar, A.: Analisis Kinerja Jalan dalam Upaya Mengatasi Kemacetan Lalu Lintas pada Ruas Simpang Bersinyal di Kota Palu. SMARTek. 9(4), 327–366 (2011) 3. BPS Kota Balikpapan.: Kota Balikpapan dalam Angka 2018. Balikpapan. Badan Pusat Statistik Kota Balikpapan. [Online] bisa diakses di: Desain Rute Jaringan Moda Bus Kota sebagai Antisipasi Kemacetan di Kota Balikpapan59 https://balikpapankota.bps.go.id/publication/2018/ 08/16/8df63ffd73378cb97b4fc3eb/kota-balikpapan-dalam-angka-2018.html. diakses pada 10 Sept 2018 (2018) 4. Ismiati., et al.: Pencemaran Udara Akibat Emisi Gas Buang Kendaraan Bermotor. Jurnal Manajemen Transportasi & Logistik (JMTransLog) 1(3) (2014) 5. Neiburger, M.: Understanding our Atmospheric environment, diterjemahkan Ardina Purbo. Memahami Lingkungan Atmosfer Kita, Edisi II, Bandung: ITB Bandung (1995) 6. Powers, J., Klemp, J., Skamarock, W., Davis, C., Dudhia, J., Gill, D., Coen, J., Gochis, D., Ahmadov, R., Peckham, S., Grell G., Michalakes, J., Trahan, S., Benjamin, S., Alexander, C., Diego, G., Wang, W., Schwartz, C., Romine, G., Liu, Z., Snyder, C., Chen, F., Barlage, M., Yu, W., dan Duda, M.: The weather research american meteorological society and forecasting (WRF) model: overview, system efforts, and future directions. Bull. Am. Meteor. Soc. https:// doi.org/10.1175/BAMS-D-15-00308.1, in press (2017) 7. Turyanti, A., et al.: Analisis Pola Dispersi Partikulat dan sulfurdioksida menggunakan model WRF Chem di sekitar wilayah industri Tangerang dan Jakarta. Manusia dan lingkungan 23(2), 169–178 (2016) 8. Faisal, I., Sofyan, A.: Analisis pengaruh variasi musiman terhadap dispersi NO2 di kota Tangerang dengan menggunakan model WRF-CHEM. Jurnal Teknik Lingkungan 25(1), Hal 1–14 (2019) 9. Fast, J.D., Gustafson, W.I., Easter, R.C., Zaveri, R.A., Barnard, J.C., Chapman, 46. E.G., et al.: Evolution of ozone, particulates, and aerosol direct radiative forcing in the vicinity of Houston using a fully coupled meteorology-chemistry-aerosol model. J. Geophys. Res.: Atmos. 111(D21) (2006). 10. Grell, G.A., Peckham, S.E., Schmitz, R., McKeen, S.A., Frost, G., Skamarock, W.C., et al.: Fully coupled “online” chemistry within the WRF Model. Atmos. Environ. 39(37), 6957–6975 (2005) 11. Zhang, Q., Tong, P., Liu, M., Lin, H., et al.: A WRF-chem model-based future vehicle emission control policy simulation and assessment for the Beijing-Tianjin-Hebei region, China. J. Environ. Manag. (2020)

88

D. Gusnita et al.

12. Baklanov, A., et al.: Online coupled regional meteorology chemistry models in Europe: current status and prospects. J. Atmos. Chem. Phys., 14, 317–398 (2014). www.atmos-chem-phys.net/ 14/317/2014/, https://doi.org/10.5194/acp-14-317 13. Gusnita, D.: Effect of meteorology parameters on air pollutant standard index in the urban area (case study in Jakarta). In: Proceedings of the International Conference on Radio science, Equatorial Atmospheric Science and Environment and Humanosphere Science, vol. 275, pp. 705–716 (2021)

Automatic Processing for Aerosol, Snow/Ice, Cloud, and Volcanic Ash Imagery (ASCI) Products Based on NOAA-JPSS Satellites Data Olivia Maftukhaturrizqoh, Andy Indradjad, Tri Astuti Pandansari, Hidayat Gunawan, and Karunika Diwyacitta Abstract Routine air quality monitoring is important to anticipate potential climate changes and other consequences including human health implications. Remote sensing observations can measure air quality estimation indirectly and complement temporally and spatially limited ground-based observation. Aerosol, Snow/Ice, Cloud, and Volcanic Ash (ASCI) is one of the satellite imagery products from the National Oceanic and Atmospheric Administration (NOAA) Joint Polar Satellite System (JPSS) series satellite data. Research Center for Remote Sensing BRIN acquired NOAA-JPSS satellite data via direct receive from the BRIN ground station. The satellite passes above Indonesia four times per day, thus can provide daily observations. An automation system to generate ASCI data products is developed to provide and ensure the continuity and availability of the data. Inputs of the system are from Sensor Data Records (SDR).h5 file format of Visible Infrared Imaging Radiometer Suite (VIIRS) instrument carried by NOAA-JPSS Satellite. The inputs are processed using an open-source module, Community Satellite Processing Package (CSPP) ASCI version 1.2. The output will be in NetCDF (.nc) file format including Aerosol Optical Depth (AOD), Aerosol detection product (ADP), Cloud Mask, Cloud Phase, Cloud Cover Layer, Volcanic Ash, Surface Albedo, Ice concentration (if any), and also generate Quicklook imageries in PNG file format. The automation system consists of several functions to crawl files and look for the newest input data available, process, and save it to the storage. The automation system has been implemented and can automatically produce ASCI imageries. Measurement of processing time was conducted and discovered that it corresponds to the input folder size. The average processing time was 19.82 and 15.31 minutes for daytime and nighttime data, respectively. The near-real-time ASCI products generated from this automation system are expected to be utilized by users to assist their research, forecast, and policy decisions related to air quality, climate, or other corresponding research. O. Maftukhaturrizqoh (B) · A. Indradjad · T. A. Pandansari · H. Gunawan · K. Diwyacitta Research Center for Remote Sensing, National Research and Innovation Agency, Jakarta, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_9

89

90

O. Maftukhaturrizqoh et al.

1 Introduction Aerosols are suspended particles in the air that can be either liquid or solid and semisolid, the main component of industrial smog and pollutant in natural disasters such as volcanic eruptions, dust, biomass burning, and forest fires [1, 2]. Aerosols leads to deteriorated air quality [2], and exposure especially to those smaller than 2.5 µm can cause health consequences including heart disease, strokes [1], respiratory problem [3], acute cardiovascular morbidity and mortality, lung cancer, and others that can shorten life expectancy [4] and also have harmful impacts for the environment [2]. Therefore, monitoring of aerosol distributions and studying how they are changing is important, especially for aerosols with large spatial and temporal variability, such as smoke, sand storms, and dust [5, 6]. Studies of aerosol characteristics allow for the anticipation of prospective climate changes, long-term ecological effects [7], and other consequences. Those valuable data can assist forecasters, meteorologists, and other stakeholders in decision-making related to air quality, long-term climate change assessment, and others. Currently, in Indonesia, there are already in-situ stations for air quality monitoring using Air Quality Monitoring System (AQMS) [8]. However, the time–space dynamics air assessment are constrained by the monitoring’s limited network [9]. A large spatial view and resolution of aerosol and air quality estimation [9, 10] can be achieved by spaceborne or remote sensing observation that complements temporally and spatially limited ground-based or in-situ sampling [11] and airborne observations [12]. Recently, remote sensing has become frequently used to monitor the worldwide spatial and temporal distributions of aerosols and offer long-term and continuous coverage [7]. Electromagnetic radiation from the earth’s surface is captured and recorded in satellite imageries [9] while aerosol directly scatter and absorb solar and thermal infrared radiation and indirectly change the Earth-leaving radiation [1]. Related to it, Aerosol optical depth (AOD) is a measure of the radiation extinction caused by aerosols due to absorption and scattering [13], the retrieval of AOD is from visible and near IR bands solar radiation reflection [1]. Retrieval of AOD and other aerosol products measured by satellite began in the 1970s that was made initially with sensor in Landsat satellite series, then followed by a variety sensors [1]. Aerosol data more recently obtained by the Visible Infrared Imaging Radiometer Suite (VIIRS) sensor carried by the Suomi National Polarorbiting Partnership (SNPP) satellite and the retrievals continued to the next generation Joint Polar Satellite System (JPSS) from National Oceanic and Atmospheric Administration (NOAA) [1]. NOAA-JPSS satellite provides continuous access to environmental data worldwide by carrying four instruments, one of those is the VIIRS instrument [14]. Aerosol, Snow/Ice, Cloud, and Volcanic Ash (ASCI) is one of the satellite imagery products from the VIIRS sensor of NOAA-JPSS satellite data. The VIIRS Atmospheric composition aerosol products are the aerosol optical depth (or aerosol optical thickness, AOD, AOT) and the aerosol detection product (ADP). The AOD is a quantitative measure of aerosol loading that requires spectral reflectance measurements from a

Automatic Processing for Aerosol, Snow/Ice, Cloud, and Volcanic Ash …

91

variety of VIS and NIR bands and many ancillary data sets. The ADP is information on the aerosol type in a qualitative form that identifies volcanic ash, dust, and smoke [14]. In terms of that, Research Center for Remote Sensing BRIN has been acquired satellite data of VIIRS sensor data from SNPP and NOAA-JPSS via direct receive from BRIN ground station since early 2018 [15]. The NOAA-JPSS satellite passes above Indonesia four times per day [16], two times in the daytime and 2 times other are passed in the nighttime, thus it can provide daily observations with national coverage. To provide and ensure the continuity and availability of the data, this paper proposed a near-real-time automation system to generate ASCI imagery data products. The automation system generates the ASCI imagery automatically right after the new input data from the ground station acquisition is detected. The system can reduce the workload of the human workforce and also can avoid human errors. From the automation system, the ASCI imagery data product will available in near-real-time. The automation system also can be used to process NOAA-JPSS older data from the beginning of input data available in BRIN storage (2018), so that multitemporal or historical data can be provided for long-term assessment, analysis or research. The near-real-time ASCI products generated from this automation system are expected to be utilized by users to assist their research, forecast, and policy decisions related to air quality, climate, or other corresponding researches.

2 Data and Method The automation system design consists of four major functions. Those functions are the files crawling function, comparing function, processing function, and storing function as shown in Fig. 1a. To be able to run the automation system requires three basic parts, those are input, process, and output. The inputs are from Sensor Data Records (SDR) of the Visible Infrared Imaging Radiometer Suite (VIIRS) instrument carried by NOAA-JPSS Satellite. NOAA-JPSS Satellite data is acquired via direct receive by BRIN ground station in Parepare, South Sulawesi. All the raw data from the direct receive are transferred further and stored in the BRIN cloud. The ASCI automation system gets the SDR input data to generate ASCI products from the BRIN cloud through BRIN internal network and also all data acquisition and transfer work under the network as shown in Fig. 1b. The Inputs data of the automation system are all VIIRS M-bands (SVM*) and I-bands (SVI*), the I-Band (GITCO*), and M-band geolocation files (GMTCO*) and Day/Night Band (SVDNB*, GDNBO*) files as shown in Fig. 2. The inputs are then processed in the next part, the process part. In the process part, the CSPP ASCI module from the Space Science and Engineering Center (SSEC) at University of Wisconsin-Madison [17] is utilized. It is an opensource module as an outcome of collaboration between Space Science and Engineering Center (SSEC), Cooperative Institute for Meteorological Satellite Studies

92

O. Maftukhaturrizqoh et al.

Fig. 1 a Four major functions of the automation system design. b The process of Satellite data acquisition, transfer, and processing in the Remote Sensing Research Center, National Research and Innovation Agency (BRIN)

(CIMSS), and NOAA Community Satellite Processing Package (CSPP) [18]. The inputs will be processed by the CSPP ASCI version 1.2 to generate several types of output in NetCDF (.nc) extension format including JRR-ADP*.nc, JRRAOD*.nc, JRR-CloudBase*.nc, JRR-CloudMask*.nc, JRR-CloudPhase*.nc, JRRCloudCoverLayers.nc, JRR-VolcanicAsh*.nc, SURFALB*.nc, etc. as mentioned in the Fig. 2 output list. The processing module will also generate Quicklook imageries in PNG extension format, those are JRR-AOD*AOD550.png, JRRCloudMask*.png, JRR-VolcanicAsh*.png, JRR-IceConcentration*.png (if any), and SURFALB*ALBEDO_EDR.png. JRR-AOD and JRR-ADP imagery could only be generated by using SDR VIIRS daytime input data, otherwise, the other product can be generated in the day and nighttime data. The four major functions as shown in Fig. 1 have a purpose as follows: The first one is a function to perform files crawling so that the system scan and search for the Fig. 2 Three basic parts of the automation system

Automatic Processing for Aerosol, Snow/Ice, Cloud, and Volcanic Ash …

93

unprocessed input files. The second function compares the file crawling results to the processed directory log to check whether an input directory has been processed. When there is a directory not processed yet, the system activates the next function to process the input directory. After processing the directory, the last storing function store the output in the corresponding directory based on the date and time of the data acquisition. The module was implemented by using Python programming language. Python programming language is one of the most popular ones that have various libraries to support automation functions. Some automation implementations that use the python programming language as their core program are automatic modeling [19, 20] automatic analysis [21, 22] automatic computation [23], and automatic classification [24]. The flowchart of the automatic system is shown in Fig. 3. below. First, the system will open and read a log file and store it as an array. The log file has contained a list of file directories which already been processed. The files crawling function then scan all files in directories. To determine whether an input directory had already been processed or still has to be processed further, the comparing function performs a comparison between the scanned filename to the log list. When the file name is not in the log yet, the processing function will process the input directories file further using the CSPP ASCI module. The module consists of two main processing, the first will produce ASCI products in .nc format. Afterward, the system run the next processing which will generate a Quicklook of ASCI product in PNG format. After finishing processing, the storing function check whether the storage directory path based on year/month/date and product name of the input data acquisition

Fig. 3 Flowchart of the automatic system

94

O. Maftukhaturrizqoh et al.

exists. If the storage path has the not existed yet, the system will create the appropriate storage directory according to acquisition time of the input data and store all the output files in. The directory that has finally been processed will then be written in the processed directory log. The system also computes the time required to process each input directory in generating ASCI products to find out the performance of this automation system. The log file also stores this processing time information of each data. The processes in the automation system will be run each time Sensor Data Records (SDR) from the NOAA-JPSS satellite VIIRS sensor is available from new acquisition activity. The system reduces the workload of the human workforce and also can avoid human errors.

3 Results As mentioned in the previous section, the ASCI output product that will be generated from the system are VIIRS Aerosol Optical Depth, VIIRS CloudMask, VIIRS Volcanic Ash, and VIIRS Surface Albedo in both.nc and Quicklook.png format. The ASCI output products are at VIIRS M-Band Spatial Resolution which is 750 m. Figure 4 shows four types of Quicklook output in PNG format generate from the automation system using the CSPP ASCI version 1.2 module, the satellite data acquired on May 15th 2022 at 05:31 (UTC). The Quicklook already contains a legend or color indication for each product. Figure 4a is the Quicklook PNG image of Aerosol Optical Depth (AOD) at 550 nm shown in colors ranging from dark violet to red that indicates the highest value. Figure 4b is the Quicklook PNG image of Cloud Mask shown in four kinds of color indicator, green color indicates clear, light blue indicates probable clear, red indicates probable cloudy, and white indicates cloudy. Figure 4c is the Quicklook of Volcanic Ash confidence product shown ranging from high to not ash. Figure 4d is the Quicklook of Surface Albedo shown in colors ranging from dark violet to red that indicates the highest ratio albedo. The automation system has been included in operational activities and integrated into the remote sensing satellite data processing system in the BRIN Research Center of Remote Sensing and shows satisfactory and stable performances. The system has been able to automatically perform file crawling, compared to the log, processing, and storing functions. The automation system has already been run for three months from May to July 2022 (Fig. 5). Table 1 gives a summary of the processing time data on May and June 2022 product data. The measurement of the processing time of each product generation was conducted to know the performance of this automation system. From Table 1 can be discovered that the time required to process data is corresponds to the input folder size. The larger the size of the input data, the longer time is required to process the data. All data in Table 1 are plotted in Fig. 6 graphics which show the correlation of input data size to processing time.

Automatic Processing for Aerosol, Snow/Ice, Cloud, and Volcanic Ash …

95

Fig. 4 As a sample of ASCI product Quicklook in.png format acquired on May 15th 2022 at 05:31 (UTC). a VIIRS Aerosol Optical Depth (JRR-AOD*550.png). b VIIRS CloudMask (JRRCloudMask*.png). c VIIRS Volcanic Ash (JRR-VolcanicAsh*.png). d VIIRS Surface Albedo (JRRSURFALB*.png)

Processing Time Required Processing Time (s)

Fig. 5 Scatter plot of correlation between input file size and processing time required

10000 8000 6000 4000 2000 0 0

2

4

6

8

10

12

Input File Size (Gb)

14

16

18

96

O. Maftukhaturrizqoh et al.

Table 1 Processing time data on May and June 2022 product data No.

Date

Acquisition time (UTC)

Input file size (Gb)

1

01/05/2022

04:54

15

1379.62

2

01/05/2022

06:37

8.9

818.05

3

01/05/2022

16:03

5.1

569.24

4

01/05/2022

17:37

13

1057.26

5

02/05/2022

04:37

12

1142.86

6

02/05/2022

06:18

12

1043.20

7

03/05/2022

04:18

12.0

1113.25

: :

: :

: :

: :

: :

219

29/06/2022

04:49

14

1412.15

220

29/06/2022

06:30

8.9

998.56

221

29/06/2022

15:56

5.1

859.60

222

29/06/2022

17:30

13

1253.72

223

30/06/2022

04:30

12

1617.93

224

30/06/2022

06:11

12

1276.91

225

30/06/2022

17:12

13

1324.17

Daytime Data Processing Time

Nighttime Data Processing Time

7000

12000

Processing Time (s)

Processing Time (s)

Processing time required (seconds)

6000 5000 4000 3000 2000 1000 0 0

2

4

6

8

10

12

14

Input data Size (Gb)

(a)

16

18

10000 8000 6000 4000 2000 0

0

2

4

6

8

10

12

14

Input data size (Gb) (b)

Fig. 6 Scatter plot of correlation between input file size and processing time required. a Daytime data. b Nighttime data

4 Discussion Daytime and Nighttime data then were plotted separately in Fig. 7a and b graphics because as mentioned before the product of Aerosol Optical Depth which was JRRAOD and JRR-ADP imagery could only be generated by using SDR VIIRS daytime input data. Therefore, daytime and nighttime data processing generate different

Automatic Processing for Aerosol, Snow/Ice, Cloud, and Volcanic Ash …

97

amounts of output data and may take different average processing times. The average processing time for daytime data was 1333.79 s or 22.23 min. While the average processing time for nighttime data was 1284.66 or 21.41 min. However, in addition to the input requirement, the processing module also required ancillary data. The ancillary data that match the date and time of the input will automatically be downloaded as needed at module runtime. This ancillary data takes time to download before the input data can be processed further by the module. Therefore, the Internet connection network while downloading the ancillary also affects the processing time. Accordingly, further analysis of the time-processing data is in the next section. In Fig. 7a and b there are several data points which are can be considered outliers. Those could be because of the slow Internet connection during ancillary data downloading at runtime. Figure 7a and d are shown the scattered plot result without the outliers. The removed outliers are interpreted using statistics as the data point which has a processing time longer than the upper fence of processing time data. The upper fence threshold of daytime processing time data was 2016.75 s, thus the data longer than that were removed and plotted into Fig. 7a. Likewise, nighttime processing data were plotted in Fig. 8b graphics after removing outliers that were longer than the upper fence of the data, 1609.46 s. The trendline of Fig. 7a and b confirm the results that input file size corresponds to the processing time. The processing time has a linear correlation to the input data size. The larger the size of input data, the longer time is required to process the data. While from the filtered processing time data above the average processing time were 1189.14 and 918.43 s or 19.82 and 15.31 min for daytime and nighttime data, respectively. Based on the calculation, this automation system has fulfilled the requirement to produce near-real-time ASCI products because the products can be generated no longer than 30 min after acquisition time. Most of the NOAA-JPSS imagery including ASCI products also available on the NOAA Comprehensive Large Array-Data Stewardship System (CLASS) website, but can be publicly accessed only after 6 h from the acquisition time [25], therefore this automation implementation,

Filtered Nighttime Data Processing Time

Filtered Daytime Data Processing Time

1600 1400

2000

Processing Time (s)

Processing Time (s)

2400

1600 1200 800 400

1200 1000 800 600 400 200

0

0

0

2

4

6

8

10 12 14 16 18

Input Data Size (Gb)

(a)

0

2

4

6

8

10

12

14

Input Data Size (Gb)

(b)

Fig. 7 Scatter plot of correlation between input file size and processing time required after removing outliers. a Daytime data. b Nighttime data

98

O. Maftukhaturrizqoh et al.

can provide ASCI products faster with national daily coverage. It is also confirmed that daytime data processing has a longer average time than nighttime data processing because they generate different amounts of output data.

5 Conclusion A near-real-time automation system for ASCI imagery products has been successfully designed and implemented in the remote sensing satellite data processing system in the BRIN Research Center of Remote Sensing and shows satisfactory and stable performances. The system has been executed to run automatically to generate ASCI imagery products from the NOAA-JPSS satellite data. The processes in the automation system will be run each time Sensor Data Records (SDR) from the NOAA-JPSS satellite VIIRS sensor are available from new acquisition activity. The system can reduce the workload of the human workforce and also can avoid human errors. This automation system can provide ASCI imagery for daily observations with national coverage and ensure the continuity and availability of the data in near-realtime data. Measurement of processing time was conducted and discovered that it corresponds to the input folder size. The average processing time was 19.82 and 15.31 min for daytime and nighttime data, respectively. The near-real-time ASCI products generated from this automation system are expected to be utilized by users to assist their research, forecast, and policy decisions related to air quality, climate, or other corresponding researches.

References 1. Nesdis, N.: NOAA NESDIS Center for Satellite Applications and EPS Aerosol Optical Depth (AOD) Algorithm Theoretical Basis Document (2016) 2. GOES-R Aerosols/Air Quality Applications fact sheet 3. Talib, N., et al.: Relationship between particulate matter (PM2.5) concentration and aerosol optical depth (AOD) throughout Peninsular Malaysia. IOP Conf. Ser. Earth Environ. Sci. 1019 (2022) 4. Zhang, H., Kondragunta, S.: Daily and hourly surface PM2.5 estimation from satellite AOD. Earth Sp. Sci. 8 (2021) 5. Morgan, M.R.: Climate change 2007. Weather 59 (2007) 6. Star, N.N.: Center for satellite applications and research algorithm theoretical basis document. In: JPSS Enterprise Processing System Aerosol Detection Product Pubu Ciren and Shobha Kondragunta, pp. 1–101 (2022) 7. Filonchyk, M. et al.: Author correction: combined use of satellite and surface observations to study aerosol optical depth in different regions of China (Scientific Reports, (2019), 9, 1, (6174), https://doi.org/10.1038/s41598-019-42466-6). Sci. Rep. 9, 1–15 (2019) 8. Zheng, H.: Design and application for environmental air quality automatic monitoring system. In: ICEMI 2009—Proceeding 9th International Conference Electron Measured Instruments, pp. 2218–2221 (2009). https://doi.org/10.1109/ICEMI.2009.5274611 9. Kumara, N., Chub, A., A.F.: An empirical relationship between PM2.5 and aerosol optical depth in Delhi Metropolitan. C. Model. Eng. Sci. 1, 119–131 (2007)

Automatic Processing for Aerosol, Snow/Ice, Cloud, and Volcanic Ash …

99

10. Chu, D.A., et al.: Global monitoring of air pollution over land from the Earth observing systemterra moderate resolution imaging spectroradiometer (MODIS). J. Geophys. Res. Atmos. 108 (2003) 11. Kaufman, Y.J., Tanré, D.: Satellite remote sensing|aerosol measurements. Encycl. Atmos. Sci. 1941–1956 (2003). https://doi.org/10.1016/b0-12-227090-8/00347-x 12. JGR Atmospheres—2013—Suomi-NPP VIIRS aerosol algorithms and data products.pdf. 13. List, A., et al.: Visible Infrared Imaging Radiometer Suite (VIIRS) Enterprise Aerosol Optical Depth and Aerosol Particle Size Products User’s Guide Version 1, February 2020 (2020) 14. Zhou, L., Divakarla, M., Liu, X., Layns, A., Goldberg, M.: An overview of the science performances and calibration/validation of joint polar satellite system operational products. Remote Sens. 11 (2019) 15. Gustiandi, B., Monica, D., Indradjad, A.: Sistem Pengolahan Data Satelit Seri Noaa Jpss Untuk Produksi Informasi Titik Panas Secara Otomatis (Automatic NOAA JPSS Satellite Series Data Processing System To Produce Active Fires Information) 17, 43–55 (2020) 16. Soleh, M., Suprijanto, A., Mahatmanto, B.P.A.: Preliminary design of remote sensing ground station system for The JPSS-1 (joint polar satellite system) data acquisition and processing. In: Proceedings of The 2nd International Conference of Indonesian Society for Remote Sensing, pp. 62–77. Indonesian Society for Remote Sensing (2016) 17. Space Science and Engineering Center 18. Webmaster, C.: Community Satellite Processing Package. Retrieved from Cooperative Institute for Meteorogical Satellite Studies (2021) 19. K., V.S., K., R.: A python automation script that generates UVM_RAL (Register Abstraction Layer) register model. Mater. Today Proc. 37 (2020) 20. Straatsma, M.W., Kleinhans, M.G.: Flood hazard reduction from automatically applied landscaping measures in RiverScape, a Python package coupled to a two-dimensional flow model. Environ. Model. Softw. 101, 102–116 (2018) 21. Loganathan, K., Sarath Kumar, R., Nagaraj, V., John, T.J.: CNN & LSTM using python for automatic image captioning. Mater. Today Proc. (2020). https://doi.org/10.1016/j.matpr.2020. 10.624 22. Stuckner, J., Frei, K., McCue, I., Demkowicz, M.J., Murayama, M.: AQUAMI: an open source Python package and GUI for the automatic quantitative analysis of morphologically complex multiphase materials. Comput. Mater. Sci. 139, 320–329 (2017) 23. Estévez Schwarz, D., Lamour, R.: InitDAE: computation of consistent values, index determination and diagnosis of singularities of DAEs using automatic differentiation in Python. J. Comput. Appl. Math. 387, 112486 (2021) 24. Lamy, J.-B.: Owlready: ontology-oriented programming in Python with automatic classification and high level constructs for biomedical ontologies. Artif. Intell. Med. 80, 11–28 (2017) 25. List, A., et al.: Suomi National Polar-Orbiting Partnership (SNPP) Visible Infrared Imaging Radiometer Suite (VIIRS) Aerosol Products User’s Guide, pp. 1–34 (2013)

The Seasonal Composition of Inorganic Aerosol in an Urban Region of Bandung, Indonesia Wiwiek Setyawati, Dyah Aries Tanti, Saipul Hamdi, Asri Indrawati, Atep Radiana, Sumaryati, Risyanto, and Retno Puji Lestari

Abstract Bandung is a basin-shaped region surrounded by high mountains with a dense population vulnerable to air quality degradation. The haze can potentially occur in a supportive meteorological condition. The study investigates the influence of local meteorology on the inorganic aerosol composition in an urban area of Bandung in 2020–2021. The ambient air sample is collected weekly using a filter pack and analyzed by ion Chromatography (IC). Local meteorological data is obtained from the Campbell weather station installed on the site. The data analysis is classified into seasons: November to March (NDJFM) for wet and May to September (MJJAS) for dry. The sulfate, nitrate, and ammonium (SNA) contributed more than 75% of the inorganic aerosol composition, and their concentrations were higher in the dry than in the wet season. Non-parametric Spearman’s rho correlation (r) shows that the relative humidity promoted NH4 + stronger in the dry than in the wet season. Therefore, NH4 + had a more significant possibility to increase under higher relative humidity in the dry season, thus increasing the possibility of haze formation. The sources of Na+ , Cl− , and NH4 + aerosols are long-distance transport, but K+ is local. Heterogeneous oxidation under high relative humidity was more critical for NH4 + generation than photochemical oxidation was, but for Cl− it was vice versa.

1 Introduction The tiny particles suspended in the air, called aerosol, are formed in the atmosphere by direct release from sources (primary aerosol) or indirectly through chemical processes involving precursor gases (secondary aerosol). Aerosol is essential in W. Setyawati (B) · D. A. Tanti · S. Hamdi · A. Indrawati · A. Radiana · Sumaryati · Risyanto Research Center for Climate and Atmosphere, National Research and Innovation Agency-BRIN, Jl. Cisitu, Sangkuriang, Bandung 40175, Indonesia e-mail: [email protected] R. P. Lestari The Center for Standardization of Environmental Quality Instruments, The Ministry of Environment and Forestry, Puspiptek Serpong, 15310 Jakarta, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_10

101

102

W. Setyawati et al.

climate, health, and ecosystem [1, 2]. Their direct impacts on climate are by scattering and absorbing solar radiation and indirect by altering clouds’ properties. The health impact of aerosols is their ability to penetrate deep into the human respiratory system, resulting in the dysfunction of human organs. The aerosols’ removal from the atmosphere to the earth’s surface can alter the air, water bodies, and soil properties. Secondary inorganic aerosols (SIA) of sulfate, nitrate, and ammonium (SNA) are the dominant components in particulate matter with sizes less than 2.5 microns (PM2.5) [2, 3]. They are responsible for haze formation and a decrease in visibility during heavily polluted events in urban regions [1, 2, 4]. Furthermore, the increased inorganic fraction in the aerosol particles led to frequent haze days. Meteorological factors have important roles in inorganic aerosol formation. Lower relative humidity favors aerosol hygroscopic formation due to more ammonium nitrate aerosol and the enhancement of the light-scattering ability of atmospheric aerosol. A high concentration of NOx can promote the conversion of SO2 to form sulfate aerosol via aqueous-phase oxidation during intensive pollution periods [3]. Indonesia lies in the equatorial region, where sunlight is available all year round. Bandung is one of the megapolitan cities in Indonesia, located in a basin area surrounded by high mountains. The unique climatic and topographic conditions and the magnitude of the emission sources can cause dangerous haze, such as those in China [3, 4] and India [2]. Therefore, it is necessary to study the meteorological factors dominant in influencing the inorganic aerosol formation in the atmosphere. This study aims to investigate the role of local meteorology on the inorganic aerosol composition and their possible source in an urban area of Bandung.

2 Methodology 2.1 Sampling Site The sampling site in Bandung city representing urban areas is located on the 5th floor of the Climate and Atmospheric Research Center (PRIMA), BRIN building at Jl. Dr Djundjunan 133, Bandung, West Java, Indonesia (6.8951710 S, 107.588370 E) as shown in Fig. 1. Bandung has a cool climate, high humidity, large wind speed variabilities, low sunshine duration, and high precipitation [5].

2.2 Data Used We used weekly dry deposition monitoring data from 2020–2021 carried out by the Research Center for Climate and Atmosphere, BRIN (previously the Center for Atmospheric Science and Technology, LAPAN). There was no sampling from March 9th–June 29th, 2020, June 14th–August 7th, 2021, Aug 14th–7th, 2021, and

The Seasonal Composition of Inorganic Aerosol in an Urban Region …

103

Fig. 1 The Sampling site is marked by a yellow broken circle (Left). The broken red line is the Bandung city administrative border (right)

September 27th–October 4th, 2021, because of the lockdown during the COVID19 pandemic. The dry deposition monitoring is a part of the East Asia Network (EANET) for acid deposition monitoring collaboration. The ambient air is collected using a filter pack. The sampling duration is one week, and the airflow rate is 1.5 L/min. The filter pack sampling method refers to the manual guide published by Asia Center for Air Pollution Research (ACAP) (EANET, 2010). We focus on the aerosol data: Sulfate (SO4 2− ), nitrate (NO3 − ), chloride (Cl− ), ammonium (NH4 + ), calcium (Ca2+ ), sodium (Na+ ), potassium (K+ ), and magnesium (Mg2+ ) ions analyzed by Ion Chromatography (DIONEX ICS-1600/09071855 for anions and DIONEX ICS-1600/09071855 for cation analysis). The 10-min synoptic data is obtained from the Campbell weather station adjacent to the filter pack. The data is then averaged according to the weekly filter pack sampling duration. The meteorological data used for further analysis are temperature (T ) in degrees Celsius, relative humidity (RH) in %, wind speed (WS) in m/s, wind direction (WD) in degrees, precipitation amount (PR) in mm, and solar radiation (SR) in W/m2 .

2.3 Data Analysis We grouped data into wet and dry seasons to study seasonal variabilities of inorganic aerosol concentrations and their related meteorological condition. In Java Island, the dominant northwest monsoon from November to March (NDJFM) leads to the wet season, and the southeast monsoon from May to September (MJJAS) to the dry season [6]. Spearman’s rho correlation coefficient with a 95% significance level (α = 0.05) is used to study the relationship between inorganic aerosol concentrations and corresponding selected meteorological factors. The formulae for calculating Spearman’s rho correlation coefficient are as follows:

104

W. Setyawati et al.

 6 di 2   r =1− n n2 − 1

(1)

where r is Spearman’s rho correlation coefficient, d is the difference between ranks, and n is the number of observations. We draw wind rose diagrams to get information about the wind direction and speed in Bandung during wet (NDJFM) and dry (MJJAS) seasons using a freeware named WRPLOT developed by Lakes Software (https://www.weblakes.com/software/fre eware/wrplot-view/). The software uses wind speed and direction as inputs. The output is a wind rose figure that illustrates the frequency of wind occurrence in the specified wind direction sectors and wind speed classes for a given location and time. We investigate the long-distance source of particles to Bandung during wet and dry seasons by utilizing The HYSPLIT trajectory model developed by Air Resources Laboratory, NOAA (https://www.ready.noaa.gov/hypub-bin/trajtype.pl?runtype=arc hive). The model vertical trajectories velocity was run for three heights: 100, 300, and 500 m above ground level by selecting GFS meteorological data having 0.25° spatial resolution.

3 Result and Discussion 3.1 Seasonal Variabilities We report the descriptive statistics of seasonal inorganic aerosol in ambient air by averaging the weekly concentrations for corresponding months in 2020–2021 (see Table 1 left). The average concentration of inorganic aerosols in the dry (MJJAS) was higher than in the wet (NDJFM) season, except for Mg2+ . The SO4 2− , NO3 − , and NH4 + (SNA) had the three highest concentrations with average ± standard deviation in wet were 5.8 ± 2.3 μg/m3 (51.0%), 1.8 ± 0.98 μg/m3 (15.5%), and 1.5 ± 0.8 μg/m3 (13.2%), respectively and in dry were 6.9 ± 2.7 μg/m3 (50.6%), 2.1 ± 0.7 μg/m3 (15.3%), and 1.8 ± 0.8 μg/m3 (13.3%), respectively. SNA’s compositions made up more than 75% of total inorganic aerosols. The SNA also dominated the haze pollution in Beijing, China, with NH4 + as the highest, followed by NO3 − and SO4 2− [7]. The secondary aerosols of SNA were produced by the chemical process of their precursors in the atmosphere [8] as follows: 2− 2NH3(g) + H2 SO4(g)  2NH+ 4 + SO4

(2)

− NH3(g) + HNO3(g)  NH+ 4 + NO3

(3)

The restricted mobilization imposed in Indonesia in 2020–2021 due to the COVID-19 pandemic led to fewer vehicles on the road and cutting down on most

The Seasonal Composition of Inorganic Aerosol in an Urban Region …

105

Table 1 The descriptive statistics of inorganic aerosol concentrations in an urban region of Bandung Inorganic aerosols Wet

Dry

Meteorology Wet

Dry

SO4 2− (μg/m3 )

5.8a ± 2.3b 6.9a ± 2.7b T (°C) 1.1c − 11.6d 0.4c − 13.1d 40e 28e

23.5a ± 0.5b 22.4c − 24.9d 40e

23.8a ± 0.4b 22.9c − 24.5d 27e

NO3 − (μg/m3 )

1.8a ± 0.98b 2.1a ± 0.7b 0.3c − 5.6d 0.1c − 3.3d 40e 28e

82.9a ± 3.8b 73.7c − 89.4d 40e

76.0a ± 6.2b 66.4c − 90.8d 27d

Cl− (μg/m3 )

0.3a ± 0.1b 0.1c − 0.7d 40e

0.4a ± 0.2b WS (m/s) 0.02c − 0.7d 28e

2.2a ± 0.6b 1.1a − 3.5b 40e

1.9a ± 0.3b 1.4c − 2.6d 27e

NH4 + (μg/m3 )

1.5a ± 0.8b 0.4c − 3.5d 40e

1.8a ± 0.8b 0.2c − 4.1d 28e

0.04a ± 0.03b 0.0c − 0.12d 40e

0.03a ± 0.04b 0.0c − 0.2d 27e

21.7bNa+ (μg/m3 ) 0.4a ± 0.3b 0.1c − 1.4d 40e

RH (%)

PR (mm)

0.6a ± 0.2b SR (W/m2 ) 0.01c − 1.1d 28e

K+ (μg/m3 )

0.3a ± 0.1b 0.1c − 0.6d 40e

Mg2+ (μg/m3 )

0.1a ± 0.1b 0.1a ± 0.04b 0.02c − 0.4d 0.02c − 0.2d 40e 28e

Ca2+ (μg/m3 )

1.1a ± 0.4b 0.1c − 2.4d 40e

146.3a ± 25.7b 135.6a ± 21.7b 93.4c − 220.7d 84.5c − 175.7d 40e 27e

0.5a ± 0.2b 0.04c − 1.2d 28e

1.2a ± 0.4b 0.2c − 1.97d 28e

Note superscript a = average, b = standard deviation, c = minimum, d = maximum, e = number of data

industrial activities. The SNA is expected to be lower than the years before the pandemic because their precursor emissions are also lower. Further study is needed to justify this statement. The seasonal meteorological data was calculated by averaging the data of the corresponding months in 2020–2021, and the results are in Table 1 right. We reported that the wet season in urban Bandung was slightly more humid, windier, wetter, and sunnier than the dry season. However, the dry season was slightly hotter than the wet season. Indonesia experienced a weak to moderate La Nina in 2020–2021, resulting in a wetter dry season.

106

W. Setyawati et al.

Fig. 2 Wind rose for wind speed and wind direction in a wet and b dry seasons

3.2 Meteorological Factors Versus Inorganic Aerosol Concentration Fig. 2 shows the wind rose plot of wind direction and speed in wet and dry seasons. During the wet season, the dominant winds, about 28.4% of the time, came from western Bandung, where a toll road connecting Bandung–Jakarta and very dense industrial and residential areas are located (see Fig. 2a). The wind blew from the west dominantly at speeds 0.5–2.10 m/s (8.18% of the time), 2.10–3.60 m/s (8.22% of the time), and 3.60–5.70 m/s (8.1% of the time). On the contrary, in the dry season, the dominant winds, about 25% of the time, came from eastern Bandung, where also there are heavy transportation, industrial, and residential areas (see Fig. 2b). The wind blew from the east dominantly at speed 0.5–2.10 m/s (14.1% of the time), 2.10–3.60 m/s (7.1% of the time), and 3.60–5.70 m/s (3.0% of the time). Table 2 presents the non-parametric Spearman’s rho correlation coefficient (r) values between seasonal inorganic aerosol concentrations and the corresponding average of temperature (T ), relative humidity (RH), wind speed (WS), precipitation (PR), and solar radiation (SR). T significantly inhibited Cl− during the wet (r = − 0.33, p < 0.05) but enhanced K+ during the dry (r = − 0.4, p < 0.01) seasons. RH significantly promoted SO4 2− (r = 0.37, p < 0.05), NO3 − (r = 0.31, p < 0.05), K+ (r = 0,38, p < 0.05), and Ca2+ (r = 0.41, p < 0.05) in the wet season, but inhibit Cl− (r = −0.44, p < 0.05), Na+ (r = −0.42, p < 0.01), Mg2+ (r = −0.28, p < 0.05) in dry seasons. RH promoted NH4 + stronger in dry (r = 0.52, p < 0.01) than in wet (r = 0.44, p < 0.01) and led to a higher concentration of NH4 + in dry than in wet seasons. Low T and high RH are more favorable to forming ammonium nitrate (NH4 NO3 ) [9, 10]. PR enhanced NH4 + (r = 0.43, p < 0.01) but reduced Na+ (r = −0.38, p < 0.05) in the dry season. Significant negative correlation between NH4 + and SR (r = − 0.31, p < 0.05), but a positive between NH4 + and RH (r = 0.52, p < 0.01) in the dry season showed that heterogeneous oxidation under high relative humidity was more critical for NH4 + generation than photochemical oxidation was [10]. On the

The Seasonal Composition of Inorganic Aerosol in an Urban Region …

107

Table 2 Spearman rho’s correlation coefficient values (r) between inorganic aerosol concentrations and selected meteorological variables Ion concentration

Meteorological data T (°C)

RH (%)

WS (m/s)

PR (mm)

SR (W/m2 )

0.09(1)

0.37*(1)

0.24(1)

0.11(2)

0.24(2)

−0.25(1) −0.22(2)

0.23(2)

0.09(1) −0.19(2)

NO3 − (μg/m3 )

−0.08(1) 0.09(2)

0.31*(1) 0.1(2)

0.14(1) −0.18(2)

−0.06(1) 0.09(2)

0.17(1) 0.00(2)

Cl− (μg/m3 )

−0.33*(1) −0.15(2)

0.02(1) −0.44**(2)

0.42**(1) 0.48**(2)

−0.3(1) −0.25(2)

−0.25(1) 0.40**(2)

NH4 + (μg/m3 )

0.01(1) 0.05(2)

0.44**(1) 0.52**(2)

−0.25(1) 0.43*(2)

0.19(1) 0.43**(2)

0.10(1) −0.31*(2)

Na+ (μg/m3 )

0.07(1) 0.1(2)

−0.12(1) −0.42**(2)

0.48**(1) 0.34*(2)

−0.24(1) −0.38*(2)

0.21(1) 0.26(2)

K+ (μg/m3 )

−0.14(1) 0.40**(2)

0.38*(1) −0.13(2)

−0.5**(1) 0.04(2)

0.25(1) −0.11(2)

−0.26(1) 0.24(2)

Mg2+ (μg/m3 )

0.1(1) 0.2(2)

0.06(1) −0.28*(2)

0.26(1) 0.23(2)

−0.13(1) −0.27(2)

0.3(1) 0.16(2)

Ca2+ (μg/m3 )

−0.07(1) 0.06(2)

0.41*(1) 0.08(2)

−0.31(1) −0.09(2)

0.01(1) −0.01(2)

−0.10(1) −0.15(2)

SO4

2−

(μg/m3 )

Note * correlation is significant at the 0.05 level (2-tailed) **correlation is significant at the 0.01 level (2-tailed) Superscript (1) = wet, (2) = dry

contrary, the significant positive correlation between Cl− and SR (r = 0.40, p < 0.01) but negative between Cl− and RH (r = −0.44, p < 0.01) in the dry season showed that photochemical oxidation was more critical to Cl− formation than heterogeneous oxidation under high relative humidity [4]. The reduced ventilation associated with the reduction in SR amount and lower temperature increased the SNA concentrations at the surface [11]. WS significantly supported Cl− in wet (r = 0.42, p < 0.01) and dry (r = 0.48, p < 0.01) seasons. WS also enhanced Na+ in the wet (r = 0.48, p < 0.01) and dry season (r = 0.34, p < 0.05). The source of Na+ and Cl− are mainly sea-salt aerosol. The strong wind can easily transport sea-salt aerosol hundreds of kilometers away from the sea to Bandung. The wet northwest monsoon strongly influenced the wet season and the southeast monsoon during the dry season [6]. We run backward particle trajectories using HYSPLIT by selecting January 1st, 2020, and July 1st, 2020, to represent the wet and dry seasons, respectively. As shown in Fig. 3, the result indicated that sea salt origin in the wet season was the northwestern part of the Java Sea (Fig. 3a) and, in the dry season, the southeastern part of the Indian Ocean (Fig. 3b). WS inhibited the wet season’s K+ (r = −0.5, p < 0.05) and supported the dry season’s NH4 + (r = 0.43, p < 0.05). The negative correlation between WS and K+ could indicate the local source. On the contrary, the positive correlation between WS and NH4 + could indicate long-distance transport.

108

W. Setyawati et al.

Fig. 3 The output of the backward trajectory model running by using HYSPLIT to simulate the long-distance source of Na+ and Cl− to Bandung in a wet and b dry seasons. A black star indicates the Bandung site

4 Conclusion Sulfate, nitrate, and ammonium (SNA) dominated the inorganic aerosol composition collected during the wet and dry seasons in 2020–2021. The SNA contributed more than 75% of secondary inorganic aerosol, and their concentrations were higher in the dry than in the wet season. Relative humidity had an essential role in forming SNA in the dry season and promoted NH4 + stronger in the dry than in the wet season. Therefore, NH4 + had a higher possibility to increase under higher relative humidity in the dry season, thus increasing the possibility of haze formation. Heterogeneous oxidation under high relative humidity was more critical for NH4 + generation than photochemical oxidation was, but for Cl− it was vice versa. The sources of Na+ , Cl− , and NH4 + aerosols are long-distance transport, but K+ is local. Sea salt originated from the northwestern part of the Java Sea in the wet season and, in the dry season, the southeastern part of the Indian Ocean. Acknowledgements We thank the chemistry and observation laboratories—BRIN, for providing this study’s acid deposition and surface meteorological data. Author Contributions Wiwiek Setyawati performed the data processing and analysis and wrote the manuscript; Dyah Aries Tanti and Asri Indrawati conducted sampling in the field and testing in the laboratory; Saipul Hamdi, Atep Radiana and Risyanto contributed to meteorological data observation and processing, Sumaryati and Retno Puji Lestari assisted in data quality assurance.

The Seasonal Composition of Inorganic Aerosol in an Urban Region …

109

References 1. Allen, S.A., Ana Godson, R.E.E., Micheal Ayodeji, S., Deborah, S.E., Ejike, O.M.: Secondary inorganic aerosols: impacts on the global climate system and human health. Biodiv. Int. J. 3(6), 245–259 (2019). https://doi.org/10.15406/bij.2019.03.00152 2. Chen, Y., Wang, Y., Nenes, A., Wild, O., Song, S., Hu, D., Liu, D., He, J., Ruiz, L.H., Apte, J.S., Gunthe, S.S., Liu, P.: Ammonium chloride associated aerosol liquid water enhances haze in Delhi, India. Environ. Sci. Technol. 56, 7163−7173 (2022). https://doi.org/10.1021/acs.est. 2c00650 3. Wang, X., Wang, W., Yang, L., Gao, X., Nie, W., Yu, Y., Xu, P., Zhou, Y., Wang, Z.: The secondary formation of inorganic aerosols in the droplet mode through heterogeneous aqueous reactions under haze conditions. Atmos. Environ. 6, 68–76 (2012). https://doi.org/10.1016/j. atmosenv.2012.09.029 4. Wang, Y., Wang, Y., Wang, L., Petäjä, T., Zha, Q., Gong, C., Li, S., Pan, Y., Hu, B., Xin, J., Kulmala, M.: Increased inorganic aerosol fraction contributes to air pollution and haze in China. Atmos. Chem. Phys. 19, 5881–5888 (2019). https://doi.org/10.5194/acp-19-5881-2019 5. Setyawati, W., Aries Tanti, D., Indrawati, A. (2022). Air quality in the bandung basin of Indonesia as measured by Passive Sampler. In: Yulihastin, E., Abadi, P., Sitompul, P., Harjupa, W. (eds.) Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, 2021. Springer Proceedings in Physics, vol. 275. Springer, Singapore. https://doi.org/10.1007/978-981-19-0308-3_36 6. Aldrian, E., Dwi Susanto, R.: Identification of three dominant rainfall regions within Indonesia and their relationship to sea surface temperature. Int. J. Climatol. 23, 1435–1452 (2003). https:// doi.org/10.1002/joc.950 7. Wang, Y., Zhuang, G., Tang, A., Yuan, H., Sun, Y., Chen, S., Zheng, A.: The ion chemistry and the source of PM2.5 aerosol in Beijing. Atmos. Environ. 39, 3771–3784 (2005). https:// doi.org/10.1016/j.atmosenv.2005.03.013 8. Pitts, F., Pitts, J.N.: Atmospheric chemistry: fundamentals and experimental techniques. A Willey—Interscience publication, John Willey & Sons, Inc. USA (1986) 9. Fu, X., Wang, S., Chang, X., Cai, S., Xing, J., Hao, J.: Modeling analysis of secondary inorganic aerosols over China: pollution characteristics, and meteorological and dust impacts. Sci. Rep. 6, 35992 (2016). https://doi.org/10.1038/srep35992 10. Squizzato, S., Masiol, M., Brunelli, A., Pistollato, M., Tarabotti, E., Rampazzo, G., Pavoni, B.: Factors determining the formation of secondary inorganic aerosol: a case study in the Po Valley (Italy). Atmos. Chem. Phys. 13, 1927–1939 (2013). https://doi.org/10.5194/acp-13-1927-2013 11. Jung, J., Souri, A.H., Wong, D.C., Lee, S., Jeon, W., Kim, J., Choi, Y.: The impact of the direct effect of aerosols on meteorology and air quality using aerosol optical depth assimilation during the KORUS-AQ campaign. Geophys Res Atmos. 124(14), 8303–8319 (2019). https://doi.org/ 10.1029/2019JD030641

The Development of DSS SRIKANDI Fifth Version Emmanuel Adetya, Nani Cholianawati, Prawira Yudha Kombara, Sumaryati, and Ninong Komala

Abstract The Research Center for Climate and Atmosphere (PRIMA) of the National Research and Innovation Agency (BRIN) is developing an information system based on a website application for the composition and air quality of Indonesia’s atmosphere. This application is called the Indonesia Atmospheric Composition Information System (DSS SRIKANDI). The version development of the DSS SRIKANDI that is being carried out is the fifth version. In this fifth version, the appearance of the web will be changed entirely with the addition of several new features. Some of these features are the prediction menu for atmospheric chemical parameters (O3 , CO, SO2 , PM2.5 , PM10 , and NO2 ) derived from the WRF-Chem model has been run independently by PRIMA. Furthermore, in the observation menu present the aerosol optical depth (AOD) information. The AOD was obtained from the GEMS instrument. The GEMS instrument is the South Korean satellite, which provides aerosol optical depth measurement.

1 Introduction In today’s digital era, a piece of information spreads very rapidly. There are many platforms can be used to disseminate information. One of these platforms is a websitebased application. Currently, many website-based applications that have been developed by various organizations to present information systems [1–3]. As one of the research organizations, the Research Center for Climate and Atmospheric (PRIMA) has developed an information system based on a website application called the Decision Supporting System of the Indonesian Atmospheric Composition Information E. Adetya · N. Cholianawati · P. Y. Kombara (B) · Sumaryati · N. Komala Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected] E. Adetya e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_11

111

112

E. Adetya et al.

System (DSS SRIKANDI). This DSS SRIKANDI displays information about the composition of the atmosphere in Indonesia. The atmosphere is composed of several chemical elements. The most dominant component is nitrogen and oxygen, then the rest are several other components including pollutant gases in it [4]. Pollutant gases are usually dominant in the atmosphere near the surface, especially in large cities. Compounds in gaseous form that include pollutants include Carbon Monoxide (CO), Ozone (O3 ), Sulfur Dioxide (SO2 ), and Nitrogen Dioxide (NO2 ). In addition to these gaseous compounds, there are also micro-sized solid particles such as PM2.5 and PM10 . Pollutant gases in the atmosphere will pollute the air usually inhaled by living things every day. If you continue to inhale it, it will accumulate in the body and can cause diseases, especially respiratory system diseases [5, 6]. Therefore, PRIMA developed the DSS SRIKANDI to monitor and present information on the concentration of these pollutant gases in the Indonesian atmosphere. Currently, the development of the DSS SRIKANDI has reached the fifth version. The development of the DSS SRIKANDI involves the Ministry of Environment and Forestry to share information and data on air quality observations for making decisions related to forest fires. The development of the fifth version was carried out to update some of the features that existed in the previous version. In this paper, we describe the updates that are in the fifth version. In this fifth version, an overhaul is implemented in the website’s appearance. In addition, updating and adding several features such as predictive information on pollutant concentrations sourced from the results of the operational running of the WRF-Chem model run by PRIMA and visualization of aerosol optical depth (AOD) parameters from South Korea’s GEMS satellite imagery. In previous version, prediction information was obtained from operationally downloaded CAMS-Chem global model data. However, the spatial resolution of the CAMS-Chem prediction data is still roughly about 70 km, while the prediction data of the WRF-Chem model independently run by PRIMA has a higher spatial resolution of about 15 km. The update of several features is expected to improve the quality of this DSS SRIKANDI.

2 The DSS SRIKANDI Fifth Version Architecture’s The DSS SRIKANDI is built through several parts, as shown in Fig. 1. The first part is the database that organizes the files and data will be displayed on the website. The database application used in this fifth version is MySQL which is installed on the XAMPP application. The next part is the service module or website application. The website application used is Lokomedia CMS. Lokomedia CMS is an open-source CMS created and developed by an Indonesian named Lukmanul Hakim and his team. Initially, this CMS was intended to practice creating websites using a CMS. However, because it is in great demand, Lokomedia finally becomes a CMS officially that can be used to create websites with various needs such as e-commerce, agencies, personal websites, and so on. The Lokomedia CMS can

The Development of DSS SRIKANDI Fifth Version

113

be downloaded on the page: https://members.phpmu.com/contribution/detail/dow nload-cms-lokomedia-161-versi-codeigniter. Lokomedia CMS is in great demand by website developers because it is PHP-based, easy to install on several OS, easy to use, and there is an online forum to discuss if you have problems during website development. The next section is the Application Programming Interface (API). The fifth version of SRIKANDI DSS also uses an API to transfer files and data from the database to the interface for display. Then the last part is the frontend or the face of the fifth version of the DSS SRIKANDI website. The website’s appearance has changed entirely compared to the previous version. A comparison of the appearance of versions four and five are shown in Fig. 2.

Fig. 1 The architecture of DSS SRIKANDI fifth version

Fig. 2 The comparison of the DSS SRIKANDI version fourth and fifth

114

E. Adetya et al.

3 The Interface of DSS SRIKANDI Fifth Version The fifth version of the DSS SRIKANDI interface change is shown at a glance in Fig. 2. For more details about the new interface, shown in Fig. 3. The DSS SRIKANDI version five is still in the prototype stage, so it cannot be accessed by the public. For a display, as shown in Fig. 3, DSS SRIKANDI version 5 has several main menus consist of Home, Forecast, Observation, Karhutla, Knowledge Sharing, and Tentang Srikandi. The main display or when the Home menu is accessed, will display the forecast for the PM2.5 parameter, as shown in Fig. 3a. The Forecast menu has several sub-menus, which are the parameters of the WRFChem model output as shown in Fig. 3b. The WRF-Chem model is a mesoscale numerical model that combines the weather and chemical components of the atmosphere [6–9]. The parameters of WRF-Chem output are PM2.5 , PM10 , CO, O3 , NO, NO2 , and SO2 . The forecasts of these parameters displayed are the results of predictions for the next 24 h for all parts of Indonesia with a spatial resolution of about 15 km and time intervals per hour. In previous version, the forecast data was obtained from the downloaded CAMS-Chem model. The purpose of replacing the data source for forecasting with PRIMA’s self-running WRF-Chem model is because the CAMSChem model has a lower spatial resolution. In addition, other goals are to reduce dependence on external models and increase independence. Then, the observation menu also has a sub-menu in the form of GEMS (Fig. 3c). The fifth version of the SRIKANDI DSS displays observation data from South Korea’s GEMS satellite imagery for AOD parameters. The satellite image displayed has a spatial resolution of 7 km × 8 km and an hourly time interval from 07.45 to 14.45 in Indonesian time. Actually, the GEMS satellite imagery has other parameters besides AOD, such as SO2 , O3 , cloud cover, and so on. However, this GEMS satellite is still relatively new and is still being developed. So, for the beginning, the AOD parameters are presented first. The reason of GEMS satellite imagery was chosen to be presented because the GEMS satellite is the first geostationary satellite to measure atmospheric chemistry. Therefore, it can produce images in the tropical area, especially in Indonesia with a higher time resolution. Furthermore, the menu about DSS Srikandi (Fig. 3d) has several sub-menus, such as DSS Srikandi, Tim Srikandi, and Contact Us.

4 Further Development The fifth version of the DSS SRIKANDI website can still be developed further. For further plans, in the observation menu section, especially in the GEMS satellite section, other parameters will also be displayed, such as O3 , SO2 , and cloud cover. In addition to the GEMS satellite, observation data from other instruments, such as the AQMS station owned by KLHK, and the surface observations from the PANDORA instrument will also be displayed. In addition, the Karhutla menu will also display

The Development of DSS SRIKANDI Fifth Version

Fig. 3 The interface of DSS SRIKANDI fifth version

115

116

E. Adetya et al.

the predicted distribution of PM2.5 and CO gas obtained from the WRF-Chem model when a forest fire occurs. In the initial development, it will display predictions for the Sumatra area first.

5 Summary The DSS SRIKANDI website-based application has been developed up to the fifth version. In this fifth version, a comprehensive display update was carried out using the Lokomedia CMS. Several new data are presented in the fifth version of the SRIKANDI DSS, such as prediction of pollutant parameters such as PM2.5 , PM10 , CO, O3 , NO, NO2 , and SO2 obtained from the WRF-Chem model. In addition, observation data from GEMS satellite imagery for AOD parameter is also displayed on the observation menu. This DSS SRIKANDI can still be developed further. The next development plan will add some observation data from other instruments and predict the distribution of PM2.5 and CO when forest fires occur.

References 1. Elekné Fodor, V., Pájer, J.: Application of environmental information systems in environmental impact assessment (in Hungary). Acta Silv. Lignaria Hungarica. 13, 55–67 (2017). https://doi. org/10.1515/aslh-2017-0004 2. Karayannis, V.G., Kagawa, T., Motosugi, G., Takeda, Y., Ichimura, M., Sasaki, A., Heterostructures, M., Saku, T., Horikoshi, Y., Tokura, Y.: Open Access Proceedings Journal of Physics_Conference Series_Enhanced Reader.pdf (2018) 3. Tudela, J., Martínez, M., Valdivia, R., Romo, J., Portillo, M., Rangel, R.: Enhanced Reader.pdf (2010) 4. Lutgens, F.K., Tarbuck, E.J., Herman, R.L.: The atmosphere: an introduction to meteorology. Pearson Education, New York (2018) 5. Fitri, D.W., Afifah, N., Anggarani, S.M.D., Chamidah, N.: Prediction concentration of PM2.5 in Surabaya using ordinary Kriging method. AIP Conf. Proc. 2329 (2021). https://doi.org/10. 1063/5.0042284 6. Ismiyati, Marlita, D., Saidah, D.: Pencemaran Udara Akibat Emisi Gas Buang Kendaraan Bermotor. J. Manaj. Transp. Logistik. 1, 241–248 (2014) 7. Baklanov, A., Schlünzen, K., Suppan, P., Baldasano, J., Brunner, D., Aksoyoglu, S., Carmichael, G., Douros, J., Flemming, J., Forkel, R., Galmarini, S., Gauss, M., Grell, G., Hirtl, M., Joffre, S., Jorba, O., Kaas, E., Kaasik, M., Kallos, G., Kong, X., Korsholm, U., Kurganskiy, A., Kushta, J., Lohmann, U., Mahura, A., Manders-Groot, A., Maurizi, A., Moussiopoulos, N., Rao, S.T., Savage, N., Seigneur, C., Sokhi, R.S., Solazzo, E., Solomos, S., Sørensen, B., Tsegas, G., Vignati, E., Vogel, B., Zhang, Y.: Online coupled regional meteorology chemistry models in Europe: current status and prospects. Atmos. Chem. Phys. 14, 317–398 (2014). https://doi.org/10.5194/ acp-14-317-2014

The Development of DSS SRIKANDI Fifth Version

117

8. Fast, J.D., Gustafson, W.I., Easter, R.C., Zaveri, R.A., Barnard, J.C., Chapman, E.G., Grell, G.A., Peckham, S.E.: Evolution of ozone, particulates, and aerosol direct radiative forcing in the vicinity of Houston using a fully coupled meteorology-chemistry-aerosol model. J. Geophys. Res. Atmos. 111, 1–29 (2006). https://doi.org/10.1029/2005JD006721 9. Grell, G.A., Peckham, S.E., Schmitz, R., McKeen, S.A., Frost, G., Skamarock, W.C., Eder, B.: Fully coupled “online” chemistry within the WRF model. Atmos. Environ. 39, 6957–6975 (2005). https://doi.org/10.1016/j.atmosenv.2005.04.027

Characteristics of PM2.5 Concentration at Bandung and Palembang from December 2019 to November 2021 Measured by Low-Cost Sensor Saipul Hamdi, Sumaryati, Asri Indrawati, Atep Radiana, Syahril Rizal, Ridho Pratama, Fahmi Rahmatia, Yutaka Matsumi, and Takashi Shibata Abstract This paper discusses the daily characteristics of PM2.5 at Bandung and Palembang measured by using low-cost sensor. Two low-cost sensors developed by Nagoya University has been being operated in Bandung and Palembang continuously and simultaneously. The data recorded every minute and has been grouped considering the season of DJF, MAM, JJA, and SON. To get the potential dispersion of PM2.5 concentration in Bandung and Palembang, the surface meteorological data have been used. It is found the different primary peak at both of locations. In Bandung, the primary peak appeared in the morning whereas the primary peak in Palembang appeared in the evening. In general, it is also found the annual concentration of PM2.5 at Bandung higher than in Palembang and both are overvalued of upper limit of Indonesia Government rule.

1 Introduction We use the term of particulate to describe the dispersion of solid and liquid particles at the atmosphere in normal condition. This particulate has dimension of more than a molecule size (± 0.0002 µm) and less than 500 µm. The particle in this range of size has residence time in suspension from several seconds to several months. The source of atmospheric particulate is natural source like volcanic eruption, forest fire, sea spray, and anthropogenic sources like industrial activities and transportation.

S. Hamdi (B) · Sumaryati · A. Indrawati · A. Radiana · R. Pratama · F. Rahmatia National Research and Innovation Agency, Jl DR Djundjunan No 133, Bandung, Indonesia e-mail: [email protected] S. Rizal Bina Darma University, Jl Jendral Ahmad Yani No 3, Palembang, Indonesia Y. Matsumi · T. Shibata Nagoya University, Nagoya, Japan © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_12

119

120

S. Hamdi et al.

About 50% of total emission of dust in the atmosphere is PM10. This emission contributes significantly for atmospheric warming. This emission not only disperses the atmosphere but also absorbs the solar radiation [6]. The dispersion of PM10 in the atmosphere is determined by physical condition of the source and physical and chemical processes of the source in the atmosphere. The contaminants follow the wind and accumulate at the destination. When the particulate enters the human respirator for longer time, it will damage the functions of human respiratory system and possibility make severe damage. The raising of PM10 concentration can correspond to season or weather. Muhaimin [13] reported a higher total concentration of particulate during the dry season than the rainy season. It is suspected by removal processes of contaminant or by washing out by rainwater [14]. Chaloulakou et al. [3] say that PM10 negatively correlates to wind speed with a value of −0.43 that value shows an inverse correlation between the average wind speed and the average of PM10 concentration, which means the fastest wind speed will cause a decrease in the concentration of PM10. This value also agrees with the results of Maraziotis et al. [12], who found a correlation between particulate concentration and wind speed that correlation value can be determined to enough correlation, which means the level concentration of PM10 is not only dependent on wind speed but also the other factors, i.e., residence time, temperature, and intensity of rain occurred on the observed days. Jallad et al. [8] say that a positive correlation between PM10 and surface temperature shows dependence of surface temperature on the processes of formation of new particle from gas to be particle. Generally, the concentration of particulates is higher on a hot day cause of the high activity of photochemical reactions on the day with high intensity of solar radiation [3]. Forest and land fires events which are causing an increase in the air pollution index, dominantly produces CO or carbon dioxide [7], CO2 , hydrocarbon, and particulate. As amount as 85% of greenhouse gas emission in Indonesia was produced in peatland fires [9]. Most of the area in Indonesia is forest, and potential of forest fire events in this area is inversely proportional to the area of forest. Sumatera Selatan is one of the ten provinces with the most hotspot for forest fire events in 2013 [2]. In the forest fire events in 2015 at South Sumatera, it was found that the longest distance of dispersion of PM10 is about 100 km for the highland natural primary forest, 60 km for secondary lowland natural forest, 170 km for community forests, and 50 km for secondary natural peatland [17]. Besides that, it is known that the height of plume has strong and inversely correlation to wind speed, which means that the low speed of wind will cause higher plume. Increasing wind speed will cause the atmospheric condition became more neutrally [1]. The monitoring of PM2.5 at Bandung basin using low-cost sensor shows the result that decreasing PBL at nighttime will cause local pollutants and transboundary to have mixed and tend to go down to the ground surface. This phenomenon can be correlated to higher concentration of PM2.5 at nighttime [16]. From sample test of black carbon content (BC) in Bandung and Lembang shows concentration of BC in Bandung (urban area) is higher than 60–70% compared to Lembang (suburban area), and the peak of BC concentration occurred in the rainy season [10]. This research aims to study about different PM2.5 characteristics at Bandung and Palembang.

Characteristics of PM2.5 Concentration at Bandung and Palembang …

121

Table 1 Classification of wind speed, Beaufort scale [15] No 1 2 3 4 5 6

Windspeed (m/s) 0 – 0.3 0.3 – 1.5 1.5 – 3.3 3.3 – 5.5 5.5 – 8 ≥8

Category Calm Light air Light breeze Gentle breeze Moderate breeze Fresh breeze

Color

2 Data and Method Two PM2.5 palm-sized optical sensors [18] have been being operated at Bandung (6°53 35.52 S 107°35 12.16 E) and Palembang (2°59 59.71 S 104°47 38.75 E) continuously and simultaneously. In Bandung City, the sensor is located at near road site which has heavy traffic during office hours. In Palembang City, the sensor is located inside the residence area of Perumahan Srimas Buana Rindang, about 1 km southern side of Musi River, or around 2300 m from Bina Darma University. Both sensors are doing the recording of data every ten minutes. We calculate the daily average of PM2.5, then separate it into four groups correlated to the season in Indonesia: DJF (rainy season), MAM, JJA (dry season), and SON. The period of data is taken from December 2019 to November 2021. To support the primary data, the wind speed and wind direction for Palembang were taken from the URL https://www.ogimet.com, whereas Bandung was taken from in-situ measurement by AWS. The wind speed classification used in this study has shown in Table 1 refers to Beaufort scale. Wind speed and wind direction data were analyzed by using freeware WRPLOT to find the frequency and dominant speed.

3 Results and Discussions The surface PM2.5 daily concentration at Bandung and Palembang in the period of 2019–2021 is shown in Fig. 1. The horizontal axis is the Day Number (DN) or Julian Day and started from December 1, 2019, (DN = 335) to November 30, 2021 (DN = 334). In general, it is shown that the PM2.5 concentration at Bandung is higher than in Palembang in two years monitoring but have similar pattern. Statistically, the average concentration of PM2.5 is shown in Table 2. The annual PM2.5 concentration in Bandung is higher than the PM2.5 concentration in Palembang. With the upper limit of PM2.5 as high as 15 µg/m3 then the PM2.5 concentration for both locations have over the limit by Indonesian Government rule of PP no 22/2021 about maintenance of protection and management of environment. Meanwhile, it was found that six days of event in 2020 had a daily concentration of more than 55 µg/m3 in August,

122

S. Hamdi et al.

one day in July, also one day in April. It was not found that the daily concentration at Palembang over the upper limit of concentration. Marsidi et al. [11] reported the average concentration of PM2.5 was valued as 10.77 µg/m3 at Plaju (Palembang) in fourth week of 2018 that result seems lower compared to the results of this study. The differences can appear correlated to corresponding factors such as relative humidity, surface temperature, and other artificial factors. The increasing temperature of environment will cause decreasing in particulate exposure and increasing of relative humidity will decreasing the particulate exposure. The effect of temperature and relative humidity correlating to changes in particulate concentration has been reported by Jallad et al. [8] and Csavina et al. [4]. Figure 2 shows the PM2.5 daily average concentration at Bandung for four seasons in two years of monitoring. It can be seen the appearance of sharp primary peak in the morning about 8 AM, with a variation in height. The secondary peak appears at the evening/nighttime, and this shape is very slope; also, the value is not as high as the primary peak. At both of MAM and JJA seasons, the concentration of primary peak

Fig. 1 Time series of daily concentration of PM2.5 at Bandung and Palembang from Dec 1, 2019, to November 30, 2021. The horizontal axis is day number (December 1 = day number 335, and November 30 = day number 334)

Table 2 Statistical of PM2.5 daily concentration measured by using low-cost sensor

2020

2021

Bandung

Palembang

Average

27.7

19.7

Maximum

66.4

50.9

Minimum

4.4

3.5

Count

351

331

Average

28.8

19.9

Maximum

51.3

43.3

Minimum

4.9

6.3

Count

257

255

Characteristics of PM2.5 Concentration at Bandung and Palembang …

123

60

PM2.5 Conc, μg/m3

50 40 30 20 10

0:00

23:00

22:00

21:00

19:00

20:00

17:00

18:00

16:00

15:00

14:00

12:00

13:00

11:00

10:00

8:00

9:00

7:00

6:00

5:00

4:00

3:00

1:00

2:00

0

Time DJF

MAM

JJA

SON

Fig. 2 Characteristics of PM2.5 concentration in Bandung at four seasons (DJF, MAM, JJA, and SON), calculated from daily concentration of December 2019 to November 2021

is relatively higher than the concentration in rest of the others season. MAM means a transition season from rainy season to dry season. The primary peak of PM2.5 at JJA season in Bandung is like MAM season, but PM2.5 concentration at JJA season is relatively higher than MAM season especially after mid-day. The temperature profile in March and July has similar characteristics, and maximum temperature differences at noon as amount as 1 °C. This difference can explain the difference of primary daily average (on 8–9 AM) not corresponding to temperature increment and not corresponding directly to increasing of the planetary boundary layer. Figure 3 shows the PM2.5 daily average concentration at Palembang for four seasons in two years of monitoring. The profiles were found by calculating PM2.5 concentration from December 2019 to November 2021. Different from concentration of PM2.5 in Bandung, the primary peak in Palembang is seen and sharp after sunset, or around 7 PM. In each seasons, the primary peak of PM2.5 has higher value compared to the secondary peak, and the primary peak appears in the morning around 7 AM. Referring to result about study for surface temperature profile in Palembang [5] says that the highest temperature at Palembang generally does not happen when sun reach culminates. However, it is delayed for around 2 h, then it is suspected that has a good correlation between PM2.5 concentration and temperature of the environment. The characteristics of wind speed and wind direction in Bandung are shown in Fig. 4, and the wind speed in Fig. 4 can be referred to as color in Table 1. The data on wind have been grouped into four seasons (DJF, MAM, JJA, and SON). For Fig. 4, wind data were measured from December 2019 to November 2021 using automatic weather station operated in the exact location. Generally, wind direction in Bandung for DJF and MAM flows from the west-northwest, so the surround area for this direction must have a dominant role in dispersing PM2.5 from and heading to Bandung. The dominant wind speed in DJF season is 1.5–3.3 m/s which amount to

124

S. Hamdi et al. 60

PM2.5 Conc, μg/m3

50 40 30 20 10

0:00

23:00

22:00

20:00

21:00

19:00

18:00

16:00

17:00

14:00

15:00

13:00

12:00

11:00

10:00

8:00

9:00

6:00

7:00

5:00

4:00

2:00

3:00

1:00

0

Time DJF

MAM

JJA

SON

Fig. 3 Characteristics of PM2.5 concentration at Palembang for four seasons (DJF, MAM, JJA, and SON), calculated from daily concentration of December 2019 to November 2021

38.6%, and the dominant wind direction is flowing from the west-northwest, amount 43.4%. The dominant wind direction in MAM is looks like DJF, and wind speed is 0.3–3.3 m/s. The opposite condition has occurred in JJA and SON seasons. In these seasons, the wind is distributed randomly from western and eastern and wind class 1.5–3.3 m/s dominantly. It was not found that the direction of wind heading blew from the North and South direction in JJA and SON. Tangkuban Parahu Mountain at northern side of Bandung does not contribute to PM2.5 concentration in Bandung. As well as southern site of Bandung, there are some industrial area and rice fields area do not contribute to PM2.5 concentration in Bandung. Figure 5 shows the distribution of wind speed and direction in Palembang for four seasons. Generally, it can be seen that the wind blows from northwest and southeast. These areas which have potentially sent particulate from southeast area are Kayu Agung, Pampangan, and Tulung Selapan. These areas are known as frequent forest fire especially in JJA and SON seasons. Consequently, if these areas get forest fired events in JJA and SON seasons, then these areas have high potential risk to disperse the pollutants reach Palembang City and contribute to increasing the PM2.5 concentration in Palembang. Meanwhile, in DJF season, which is marked as the rainy season, the wind blows from northwest Palembang. The areas located in northwest of Palembang are Sungai Lilin and Banyung Lencir, also neighboring country Singapore and Malaysia. However, Zhang [19] said wind shear seems more critical than wind speed for modulating ground surface PM2.5 concentration.

Characteristics of PM2.5 Concentration at Bandung and Palembang …

125

Fig. 4 Pattern of wind speed and direction in Bandung for four seasons calculated using data from December 2019 to November 2021. Source of data LAPAN

The wind speed in Palembang is commonly in range of 1.5–3.3 m/s or category of a light breeze, and a small concentration of wind speed is categorized as a gentle breeze. In light breeze category, wind fell on the face, leaves rustled, wind vane moved by wind, and small wavelets on sea. In gentle breeze category, some leaves and small twigs move constantly, light flags extended, and large wavelets on the sea.

4 Conclusions By the monitoring of PM2.5 concentration at Bandung and Palembang, it has been found different diurnal pattern of PM2.5 concentration for these two locations. It can be seen the primary peak in each locations are contrarily. The primary peak of PM2.5 concentration at Bandung occurred in the morning after sun rises. In contrast, the

126

S. Hamdi et al.

Fig. 5 Pattern of wind speed and wind direction at Palembang for four seasons calculated using data from December 2019 to November 2021. Source of data https://www.ogimet.com

primary PM2.5 concentration at Palembang occurred in the evening after sun sets. Meanwhile, the long-time series of monitoring shows that the daily average of PM2.5 concentration (24 h) at Palembang is lower compared to the daily average of PM2.5 concentration at Bandung. The yearly average concentration of PM2.5 in Bandung and Palembang has over the threshold value 15 µg/m3 (determined by Indonesian Government). The characteristics of seasonal concentration of PM2.5 does not show significant differences for these two locations. Primary peak of seasonal variation of PM2.5 concentration in Bandung is higher concentration than in Palembang. Acknowledgements Center for Atmospheric Research and Technology of The Indonesia National Research and Innovation Agency (BRIN) have fully supported this research. The appreciation is delivered to Nagoya University which has provided two PM2.5 palm-size optical sensors.

Characteristics of PM2.5 Concentration at Bandung and Palembang …

127

References 1. Alex, D.V.: Air dispersion modeling. John & Sons Inc., New Jersey (2014) 2. Badan Nasional Penanggulangan Bencana: Rencana Kontinjensi Nasional Menghadapi Ancaman Bencana Asap Akibat Kebakaran Hutan dan Lahan. Badan Nasional Penanggulangan Bencana, Jakarta (2013) 3. Chaloulakou, A., Kassomenos, P., Spyrellis, N., Demokritou, P., Koutrakis, P.: Measurements of PM10 and PM2,5 particle concentrations in Athens, Greece. J. Atmos. Environ. 37, 649–660 (2003) 4. Csavina, J., Field, J., Félix, O., Corral-Avitiac, A.Y., Sáeza, A.E., Bettertond, E.A.: Effect of wind speed and relative humidity on atmospheric dust concentrations in semi-arid climates. Sci. Total Environ. 15(487), 82–90 (2013) 5. Hamdi, S.: Karakter suhu udara kota Palembang berdasarkan pengukuran di Kampus Universitas Bina Darma. Sains Atmosfer Teknologi dan Aplikasinya, CV Andira Bandung, halaman, pp 13–25. ISBN 978-979-1458-68-9 (2013) 6. Haywood, J., Boucher, O.: Estimates of the direct and indirect radiative forcing due to tropospheric aerosols: a review. Rev. Geophys. 38(4), 513–543 (2000) 7. Heriyanto, E., Syaufina, L., Sobri, M.: Forecasting simulation of smoke dispersion from forest and land fires in Indonesia. Procedia Environ. Sci. 24, 111–119 (2015) 8. Jallad, F.A., Katheeri, E.A., Omar, M.A.: Concentrations of particulate matter and their relationships with meteorological variables. Sustain. Environ. Res. 23(3), 191–198 (2013) 9. Kementerian Lingkungan Hidup: Pengkajian Baku Mutu Kualitas Udara Ambien Lampiran PP No. 41 Tahun 1999. Jakarta (2011) 10. Lestiani, D.D., Santoso, M., Hidayat, A.: Karakteristik black carbon partikulat udara halus PM2.5 di Bandung dan Lembang 2004–2005. In: Prosiding Seminar Nasional Sains dan Teknologi Nullar PTNBR_BATAN, Bandung (2007) 11. Marsidi, Kamaluddin, M.T., Kurdi, F.N., Novrikasari: Correlation of climate to particulate matter in Palembang. Pollut. Res. 37(2), 279–284 (2018) 12. Maraziotis, E., Sarotis, L., Marazioti, C., Marazioti, P.: Statistical analysis of inhalable (PM10) and fine particles (PM2,5) concentrations in urban region of Patras, Greece. Glob. NEST J. 10(2), 123131 (2008) 13. Muhaimin: Permodelan dispersi polutan udara dari aktivitas PLTU Cirebon pada musim kemarau dan hujan serta penggunaan 2 cerobong asap [tesis]. Universitas Gajah Mada, Yogyakarta (2014) 14. Puspitasari, A.D.: Pola spasial pencemaran udara dari sumber pencemar PLTU dan PLTGU Muara Karang [skripsi]. Universitas Indonesia, Depok (2011) 15. Stewart, R.H.: Introduction to Physical Oceanography. Department of Oceanography Texas A & M University, Texas (2008) 16. Sya’bani, A., Chandra, I., Majid, L.I., Vaicdan, F., Barus, R.A.A., Abdurrachman. A., Salam, R.A.: Pemantauan konsentrasi PM2.5 dan CO2 berbasis low-cost sensor secara real-time di cekungan Udara Bandung Raya. Jurnal Teknologi Lingkungan 21(1), 009–015 (2020) 17. Tampubolon, A.P.C., Boedisantoso, R.: Analisis persebaran polutan karbon monoksida dan partikulat dari kebakaran hutan di Sumatera Selatan. Jurnal Teknik ITS 5(2), C160–C165 (2016) 18. Nakayama, T., Matsumi, Y., Kawahito, K., Watabe, Y.: Development and evaluation of a palm sized optical PM2.5 sensor. Aerosol Sci. Technol. 52(1), 2–12. https://doi.org/10.1080/027 86826.2017.1375078 (2018) 19. Zhang, Y., Guo, J., Yang, Y., Wang, Y., Yim, S.H.L.: Vertical wind shear modulates particulate matter pollutions: a perspective from radar wind profiler observations in Beijing, China. Rem. Sens. 12(3), 546. https://doi.org/10.3390/rs12030546 (2020)

Air Pollution Impact During Forest Fire 2019 Over Sumatra, Indonesia Prawira Yudha Kombara, Waluyo Eko Cahyono, Wiwiek Setyawati, Hana Listi Fitriana, Emmanuel Adetya, and Alvin Pratama

Abstract In 2019, there was a strong positive Indian Ocean Dipole (IOD) phenomenon and forest and land fire in several areas of Sumatra. The positive IOD phenomenon makes the Sumatra region drier, and the rainfall decreases from normal conditions so that the forest and land fires that occur are getting worse. The environmental impacts, especially the pollution impact of the forest and land fires and the 2019 positive IOD in Sumatra, have been studied in this study by analyzing the spatial pattern of the aerosol optical depth (AOD), particulate matter (PM)2.5 , and carbon monoxide (CO) parameters. Spatial analysis was carried out from August to November 2019. Then, the general results obtained that the maximum concentrations for the parameters AOD, PM2.5 , and CO occurred in September 2019. The three parameters spread to several provinces in Sumatra following the wind direction. Several areas affected by the 2019 forest and land fires are the provinces of Riau, Jambi, and South Sumatra. In October, the South Sumatra region became the area with the highest concentrations of PM2.5 and CO compared to other regions. This indicates that the forest and land fires incident was centered in South Sumatra Province in October 2019.

P. Y. Kombara (B) · W. E. Cahyono · W. Setyawati · E. Adetya Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected] H. L. Fitriana Research Center for Remote Sensing, National Research and Innovation Agency, Jakarta, Indonesia A. Pratama Department of Atmospheric and Planetary Science, Sumatera Institute of Technology, Lampung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_13

129

130

P. Y. Kombara et al.

1 Introduction In 2019, there were forest and land fires in several areas of Sumatra [1]. The impact of the 2019 forest fires caused a smoke hazard and increased the concentrations of aerosols in the form of PM2.5 and toxic gases such as carbon monoxide (CO). However, unlike in 2015, when the El Nino phenomenon also occurred, in the 2019 forest fires, the Indian Ocean Dipole (IOD) phenomenon occurred. IOD is a phenomenon that shows the variability of sea surface temperature (SST) between the western and eastern parts of the Indian Ocean. The name Dipole refers to the two sides of the Indian Ocean. When this phenomenon occurs, it will affect the climate and weather around the Indian Ocean. When the west side, of the Indian Ocean experiences an increase in SST, on the east side, there will be a decrease in the value of SST and is called a positive IOD and vice versa. While the positive IOD is active, the west side of the Indian Ocean will receive an abundant supply of water vapor from normal conditions due to the presence of a low-pressure center and make the wind blow and carry water vapor, resulting in excessive precipitation and vice versa on the east side will experience a shortage of water vapor supply and less precipitation. The opposite will happen when the negative IOD is active. The latest IOD phenomenon occurred in 2019 when it was autumn in the Northern Hemisphere. At that time, there was a strong positive IOD [2–4]. The strong positive IOD phenomenon that occurred in 2019 has resulted in the western part of Indonesia, especially some areas in Sumatra, experienced drought [5, 6]. This drought is exacerbated by the occurrence of forest and land fires. However, it should be emphasized that the forest fires incident was first caused by land acquisition using the burning method and caused forest fires occurred. Therefore, the IOD phenomenon only exacerbates the impact of forest fires. Several previous studies have examined the impacts of the 2019 IOD phenomenon on the weather in the Indonesian Maritime Continent (BMI), especially in Sumatra. Those previous studies explain the impact of drought and reduced rainfall [5, 6]. However, environmental conditions such as the variability of aerosol concentration and the distribution of smoke haze during forest fires, which coincide with the IOD phenomenon, have not been thoroughly studied. Therefore, in this paper, we try to examine the environmental impact of the atmosphere during the forest fires and the 2019 IOD phenomenon. The environmental impacts of the atmosphere that will be investigated are an increase in the concentration of aerosols, PM2.5 , and toxic gases such as CO.

2 Data and Method In this study, we used several data such as

Air Pollution Impact During Forest Fire 2019 Over Sumatra, Indonesia

131

1. Data The Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2) can be downloaded from: https://disc.gsfc.nasa.gov/dat asets/M2IMNXGAS_5.12.4/summary?keywords=aod. This data is used to visualize aerosol optical depth (AOD) and carbon monoxide (CO) parameters with a monthly and spatial–temporal resolution of 0.5° × 0.625°. MERRA-2 is a data reanalysis developed by NASA through the GMAO project [7]. In this study, we used MERRA-2 data from August–November 2019. 2. Sea surface temperature data using data from COBE-SST which can be downloaded on the page: https://psl.noaa.gov/data/gridded/data.cobe.html. COBESST data is the input data for sea surface temperature parameters used by JMA to run the climate data assimilation system (JCDAS) model. The COBE-SST data has a spatial resolution of one degree by one degree and a monthly temporal resolution. 3. ECMWF Atmospheric Composition Reanalysis 4 (EAC4) data to visualize Particulate Matter 2.5 µm (PM2.5) parameters which can be downloaded on the page: https://ads.atmosphere.copernicus.eu/cdsapp#!/dataset/cams-globalreanalysis-eac4-monthly?tab=form. This data is the fourth version of global reanalysis data released by ECMWF. The data we use has a spatial resolution of 0.75° and a monthly temporal resolution. 4. Rainfall data from CPC Global Unified Gauge-Based Analysis of Daily Precipitation can be downloaded at https://psl.noaa.gov/data/gridded/data.cpc.globalpre cip.html. The data has a monthly temporal and spatial resolution of 2.5° by 2.5°. In this study, we used two main methods, consist of the calculation of the IOD index and exploratory data analysis (EDA). The determination of the IOD index in 2019 was carried out by calculating the difference in SST anomalies in the western (50E-70E, 10S-10N) and eastern (90E-110E, 10S-0S) of Indian Ocean [8]. Furthermore, analyze the spatial pattern of SST anomalies, rainfall anomalies, AOD, PM2.5 , and CO using the EDA method when the 2019 strong positive IOD forest fires occurred.

3 Result 3.1 The Strong Positive IOD and Forest Fire 2019 By using the method of Saji et al. [8] can be obtained an index that shows the condition of IOD in 2019. Figure 1 shows the IOD index with a positive value above 1.5 for the months of October–November. The strengthening of the index value started in September and reached its peak in October with a value of 1.8. This indicates that there was a strong positive IOD phenomenon in 2019. The spatial pattern of the SST anomaly as shown in Fig. 2 confirms the results shown by the IOD index. Figure 2 shows the spatial pattern of SST anomalies from August to November 2019. In August, positive SST anomalies were still around the central and eastern

132

P. Y. Kombara et al.

Fig. 1 IOD 2019 index based on Saji et al. [8] method

areas of the Indian Ocean. In September, the positive anomaly area shifted to the west of the Indian Ocean. Then, from October to November, the area of the positive anomaly shifted in the western part of the Indian Ocean, resulting in a positive IOD phenomenon. The impact felt on the eastern part of the Indian Ocean is becoming drier due to the cooling of SST, resulting in decreased convective activity and reduced precipitation as shown in Fig. 3. Figure 3 shows precipitation anomalies in some areas of Sumatra, Indonesia. It can be seen there are some regions have negative values in September and October 2019. A negative value means some areas of Sumatra experienced a precipitation deficit as previously the study result was done by Lestari et al. [6]. This deficit triggers a drought in some regions of Sumatra. Along with the occurrence of positive IOD 2019, in some Sumatra regions, there were forest fires. Based on Fig. 4, we can see at several red points, there is smoke (orange circle) due to forest fire. The smoke moves follow the wind direction to other regions that do not experience forest fire. Figure 4 is one example that shows a forest fire event on September 22, 2019. The red points in the figure represent the hot spots when forest fires occurred. The negative anomaly of precipitation makes the forest fire last longer. In addition, the dry air due to positive IOD increases the intensity of forest fires. All of this can cause the degradation of air quality.

3.2 Environment Atmosphere Impact When forest fire occurs, the smoke will appear and emit an amount of aerosol into the atmosphere. Besides aerosol, this smoke contains particulate, and toxic gases that can harm human health, especially the respiratory system [9]. Figure 5 shows aerosol is emitted from smoke when forest fire 2019 occurs in Sumatra. Here, the aerosol is represented by aerosol optical depth (AOD) parameter. In August, the AOD in most parts of Sumatra had a magnitude of around 0.5 and moved toward the north. However, in September, the AOD magnitude increased drastically to reach the value

Fig. 2 SST anomaly during strong IOD positive 2019, a August, b September, c October, and d November

Air Pollution Impact During Forest Fire 2019 Over Sumatra, Indonesia 133

Fig. 3 Precipitation anomaly on September (left) and October (right) 2019

134 P. Y. Kombara et al.

Air Pollution Impact During Forest Fire 2019 Over Sumatra, Indonesia

135

Fig. 4 Forest fire on September 22, 2019, on Sumatra based on Suomi/VIIRS satellite capture (https://worldview.earthdata.nasa.gov/)

of one. The Sumatran area that has a high AOD value is Riau Province. We can say that there were severe forest fires during September 2019 as shown in Fig. 4. This high magnitude of AOD in September resembles the forest fire event in a few years ago when strong El Nino in 2015. At that time, the AOD magnitude reached greater than one based on Cahyono et al. [10]. As shown in Fig. 5b, the high AOD value moved to the north and northwest toward the Malacca Peninsula and West Sumatra Province due to was being carried away by the wind. In September, the background wind conditions still showed the Australian monsoon pattern which means the prevailing wind was southeasterly [11]. Therefore, neighboring countries such as Malaysia and then West Sumatra Province experienced a decrease in air quality and visibility. A high AOD value can be interpreted as smoke due to forest fires. In the next month, the AOD value decreased as in August. This decrease in the AOD value could have been caused by the rain in October. If we see the rainfall anomaly in Fig. 3 on the right, most of the Sumatra region has a positive rainfall anomaly in October. Then in November, the condition of AOD showed a decline in value than in October. We can say that in November, the forest fires started to subside. Previously, Fig. 5 shows the AOD pattern during the 2019 forest fires in Sumatra. This section shows the distribution pattern of one type of aerosol, namely PM2.5 during the 2019 forest fires in Sumatra, as shown in Fig. 6. In August (Fig. 6a), PM2.5 concentrations began to increase, and the maximum magnitude of PM2.5 occurs in the province of Jambi. PM2.5 spread to Riau Province and a little to Malaysia. In

136

P. Y. Kombara et al.

Fig. 5 Aerosol optical depth (AOD) is overlayed by vector wind at 850 hPa during a forest fire and strong positive IOD 2019, a August, b September, c October, and d November

the following month of September, it can be said that the forest fires reached its peak. In September (Fig. 6b), there are two centers of magnitude PM2.5 . The two centers occurred in Jambi and Riau Provinces. The monthly mean of PM2.5 in both provinces was exceeding by 300 ug/m3 . From these two provinces, PM2.5 spread to surrounding provinces such as Lampung, Bengkulu, North Sumatra, West Sumatra, and South Sumatra. The areas near to the forest fire location will be higher than usual [12]. Then in October (Fig. 6c), the center of the PM2.5 magnitude moved to South Sumatra Province, while in Jambi and Riau Provinces it had subsided. From South Sumatra Province, PM2.5 spread to surrounding areas such as Lampung, Jambi, and Bengkulu Provinces. In October, it can be said that more forest fires occurred in South Sumatra Province. Then in November (Fig. 6d), the concentration of PM2.5 was seen to be quite reduced but still concentrated in South Sumatra Province. As in

Air Pollution Impact During Forest Fire 2019 Over Sumatra, Indonesia

137

Fig. 5d, based on Fig. 6d can also be interpreted that in November, the forest fires had started to subside. Figure 7 shows the distribution of CO gas during the 2019 forest fires in Sumatra. In August (Fig. 7a), CO concentrations increased and were concentrated in Jambi and Riau Provinces. Then in September (Fig. 7b), CO concentrations increased drastically with the center of magnitude in Jambi Province. This increase shows that forest fires occur significantly. CO gas spreads to the provinces around Jambi because it is carried away by the wind. Other provinces affected include the provinces of Riau, South Sumatra, Bengkulu, to the provinces of North Sumatra. CO gas from these forest fires also spread to Malaysia. In the following month, October (Fig. 7c), CO gas concentrations began to decline, and the center of magnitude was in Jambi and South Sumatra Provinces. For October, there is a difference in the pattern between the parameters AOD, PM2.5 , and CO. This month, PM2.5 and CO parameters still

Fig. 6 PM2.5 dispersion during forest fire 2019, a August, b September, c October, and d November

138

P. Y. Kombara et al.

Fig. 7 Carbon monoxide (CO) dispersion during forest fire 2019, a August, b September, c October, and d November

showed high concentrations, but the AOD parameter showed the opposite. This difference may be caused by different models. The AOD parameters are derived from the MERRA-2 model, whereas the PM2.5 and CO parameters are derived from the CAMS model. Then in November (Fig. 7d), the concentration of CO gas decreased drastically and was only concentrated in South Sumatra. As in Figs. 5d, 6d and 7d can also be interpreted that forest fires were starting to subside in November.

4 Conclusion In 2019, strong positive IOD and forest and land fires occurred, and there was an increase in the concentration of AOD, PM2.5 , and CO gas in several provinces of Sumatra. The increase in concentration began to occur in August and reached its peak from September to October. The concentration centers of AOD, PM2.5 , and CO

Air Pollution Impact During Forest Fire 2019 Over Sumatra, Indonesia

139

occurred in the provinces of Riau, Jambi, and South Sumatra. At the peak of the forest fire event, aerosols in the form of PM2.5 and CO gas spread to Bengkulu, Lampung, West Sumatra, North Sumatra, and even Malaysia. In September, most forest and land fires occurred in the provinces of Riau and Jambi, while in October most fires occurred in South Sumatra. As soon as November entered, the concentrations of AOD, PM2.5 , and CO began to decline so it can be interpreted that in this month the forest and land fires began to subside. Acknowledgements We want to thank Rumah Program Kebencanaan DIPA OR KM BRIN No. SP DIPA-124.01.1.690501/2022 for funding research support. Besides, we acknowledge the use of imagery from the NASA Worldview application (https://worldview.earthdata.nasa.gov), part of the NASA Earth Observing System Data and Information System (EOSDIS).

References 1. Indonesia, C.: Membandingkan Karhutla di Indonesia Pada 2015 dan 2019. https://www.cnn indonesia.com/teknologi/20190918104533-199-431485/membandingkan-karhutla-di-indone sia-pada-2015-dan-2019%0A 2. Lu, B., Ren, H.L.: What caused the extreme Indian Ocean Dipole event in 2019? Geophys. Res. Lett. 47, 1–8 (2020). https://doi.org/10.1029/2020GL087768 3. Ratna, S.B., Cherchi, A., Osborn, T.J., Joshi, M., Uppara, U.: The extreme positive Indian Ocean Dipole of 2019 and associated Indian summer monsoon rainfall response (2021) 4. Chang, D.Y., Yoon, J., Lelieveld, J., Park, S.K., Yum, S.S., Kim, J., Jeong, S.: Direct radiative forcing of biomass burning aerosols from the extensive Australian wildfires in 2019–2020. Environ. Res. Lett. 16 (2021). https://doi.org/10.1088/1748-9326/abecfe 5. Mardiansyah, W., Setiabudidaya, D., Khakim, M.Y.N., Yustian, I., Dahlan, Z., Iskandar, I.: On the Influence of Enso and IOD on rainfall variability over the Musi Basin, South Sumatra. Sci. Technol. Indones. 3, 157 (2018). https://doi.org/10.26554/sti.2018.3.4.157-163 6. Lestari, D.O., Sutriyono, E., Sabaruddin, S., Iskandar, I.: Respective influences of Indian Ocean Dipole and El Niño-Southern Oscillation on Indonesian precipitation. J. Math. Fundam. Sci. 50, 257–272 (2018). https://doi.org/10.5614/j.math.fund.sci.2018.50.3.3 7. Services, G.E.S.D. and I., DISC), 151 Center (GES: Global Modeling and Assimilation Office (GMAO): MERRA-2 tavgM_2d_chm_Nx: 2d, 149 monthly mean, time-averaged, single-level, assimilation, carbon monoxide and ozone diagnos150 tics V5.12.4. https://disc.gsfc.nasa.gov/ 8. Saji, N.H., Vinayachandran, P.N.: A dipole mode in the tropical Indian Ocean. Lett. Nat. 401, 360–364 (1999) 9. Fitri, D.W., Afifah, N., Anggarani, S.M.D., Chamidah, N.: Prediction concentration of PM2.5 in Surabaya using ordinary Kriging method. AIP Conf. Proc. 2329 (2021). https://doi.org/10. 1063/5.0042284 10. Eko Cahyono, W., Setyawati, W., Hamdi, S., Cholianawati, N., Yudha Kombara, P., Julian Sari, W.: Observations of aerosol optical properties during tropical forest fires in Indonesia. Mater. Today Proc. 63, S445–S450 (2022). https://doi.org/10.1016/j.matpr.2022.04.113 11. Xian, P., Reid, J.S., Atwood, S.A., Johnson, R.S., Hyer, E.J., Westphal, D.L., Sessions, W.: Smoke aerosol transport patterns over the Maritime Continent. Atmos. Res. 122, 469–485 (2013). https://doi.org/10.1016/j.atmosres.2012.05.006 12. Sharma, A., Valdes, A.C.F., Lee, Y.: Impact of wildfires on meteorology and air quality (PM2.5 and O3 ) over Western United States during September 2017. Atmosphere (Basel) 13 (2022). https://doi.org/10.3390/atmos13020262

Influence Impregnation Method in the Structure of Bimetallic Ni-Zn/ZrO2 Catalyst Fildzah ‘Adany, Kiky Corneliasari Sembiring, Mustofa Amirullah, and Reva Edra Nugraha

Abstract Ni-based catalysts are an alternative to precious metal catalysts. However, monometallic Ni catalysts still have weaknesses, such as being easily deactivated and being able to agglomerate if the Ni concentration is too high. This can be overcome by adding a second metal (bimetallic catalyst) in which the structure and morphology of the bimetallic catalyst are strongly influenced by the method used. The wetimpregnation method is used to obtain a supported bimetallic catalyst where the order of metal carrying will affect the structure and morphology and is specific for each material. On the Ni-Zn/ZrO2 catalyst, which has been synthesized by coimpregnation (Cat-1) and sequentially (Cat-2) method, XRD results show that both catalysts have specific peaks of ZrO2 , NiO, and ZnO. However, the Cat-2 catalyst has lower crystallinity and smaller NiO crystallite size (14.79 nm) than Cat-1 (19.45 nm). Micrograph results show that both catalysts have irregular particle shapes. Cat-1 has a smaller particle size and a more homogeneous metal dispersion than Cat-2. The elemental analysis shows that metal concentration at the surface is close to the theoretical (10%). Further research with further characterization and application to a specific chemical reaction is needed to determine which of the two catalysts perform better. This is because each reaction has specific requirements, and each catalyst will produce different results in each reaction even though they have the same properties.

F. ‘Adany (B) · K. C. Sembiring Research Center for Chemistry, National Research and Innovation Agency (BRIN), South Tangerang, Banten, Indonesia e-mail: [email protected] M. Amirullah Research Center for Smart Mechatronics, National Research and Innovation Agency (BRIN), Bandung, West Java, Indonesia R. E. Nugraha Department of Chemical Engineering, Faculty of Engineering, Universitas Pembangunan Nasional “Veteran” Jawa Timur, Surabaya, East Java, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_14

141

142

F. ‘Adany et al.

1 Introduction Catalysts are materials that have an essential role in various chemical reactions so that reactions can run more quickly and selectively. Because of this, catalysts have an essential role in various chemical industries, and more than 80% of the chemical industry uses catalysts [1]. In addition to having a vital role in the industry, catalysts also play an essential role in efforts to mitigate global warming and climate change. Catalysts play a role in producing clean fuels such as hydrogen or biogas [2]. There are two types of catalysts based on the reaction system’s phase: homogeneous and heterogeneous catalysts. Heterogeneous catalysts have advantages compared to homogeneous catalysts, which are easy to separate from products and reactants, have high stability, and are more environmentally friendly [3]. Heterogeneous catalysts include zeolite, ion exchange resin, carbon-based catalysts, mineral–clay, and metal oxide [4]. Metal oxides are important as catalysts because they are essential in most chemical reactions, such as petroleum refining, energy conversion, hydrogenation, and biomass conversion. Due to their potential application in most industrial chemical processes, metal oxides are of great interest today to researchers. Metal oxide catalysts used in various chemical industries include TiO2 , ZnO, ZrO2 , porous and mesoporous metal oxides, multi-component mixed oxides, polyoxometallates (POMs), phosphates, and perovskites [1]. Metal oxide catalysts are usually modified to have high activity and selectivity in a reaction. Modification of the catalyst usually aims to improve its characteristics, such as hydrophobicity, specific surface area, basicity, and acidity [5]. The catalyst can be modified by depositing a metal on the surface of the support. This modification can increase selectivity and activity by increasing the catalyst’s active site [6]. One of the catalyst preparation methods is wet-impregnation, which is often used because it is easy and does not require high costs [7]. Nickel-based catalysts, such as Ni/Al2 O3 , Ni/MgAl2 O4 , Ni/ZrO2 , and Ni/SBA15, are well-established catalysts that can be used as an alternative to precious metalbased catalysts (Pd, Pt, and Ru) due to their high catalytic activity, low cost, and widespread availability [8–10]. On the other hand, the Ni-based catalyst is easily deactivated during the catalysis process, necessitating a relatively large amount of H2 in the hydrogenation reaction. Furthermore, the high Ni concentration in the support causes agglomeration that will inhibit the reaction process [11]. However, research shows that the ability of this monometallic Ni catalyst can be improved by adding other metals (bimetallic catalyst). Because of their much higher activity than monometallic catalysts, current Ni-based bimetallic catalysts are receiving special attention. Tanakabe et al. discovered that a homogeneous alloy of Co and Ni, as well as a low Ni substitution of Co, can increase the activity and stability of the NiCo/TiO2 catalyst, which is due to the catalyst’s high resistance to unwanted metal oxidation and coke [12]. Zhang et al. also found that the bimetallic NiCo/MgO catalyst outperforms other monometallic or NiMe metals (Me = Fe, Cu, Mn) in CO2 reforming gas in activity and stability [13].

Influence Impregnation Method in the Structure of Bimetallic …

143

The activity of Ni-based bimetallic catalysts has increased significantly in several reactions. However, to obtain a stable catalyst with high activity, the composition, pretreatment process, preparation method, and support characteristics must all be considered [14–16]. Ferari et al. discovered that the sequence of Co and Mo metal impregnation on carbon supports affects metal dispersion, the interaction between the oxide and carrier phases, and the catalytic activity of the hydrodeoxygenation reaction [17]. The order of metal bearing in the impregnation method also affects the structure of the catalyst, which affects its catalytic activity and is specific to each catalyst, as explained by Salerno et al. and Bacariza et al. [18, 19]. As a result, in this study, the Ni-Zn/ZrO2 catalyst will be prepared using a wetimpregnation method with an ethanolic solvent. The addition of the two metals in the support will be carried out simultaneously (co-impregnation) and sequentially impregnation. Furthermore, the impact of impregnation methods on the structure and morphology of the catalyst was investigated using X-ray diffraction (XRD) and scanning electron microscopy–energy dispersive X-ray (SEM–EDX).

2 Materials and Methods 2.1 Materials The materials used in this study include commercial zirconia (ZrO2 ) from Kanto Chemical, Ni(NO3 )2 ·6H2 O, Zn(NO3 )2 ·4H2 O, and ethanol absolute was purchased from Merck. All materials used in this work were analytical grade and were used directly without prior purification.

2.2 Catalyst Preparations Co-impregnation. The solid catalyst Ni-Zn/ZrO2 was synthesized with the following procedure with some modifications [20]. One gram of ZrO2 is added with the metal precursor (Ni 10% wt and Zn 10% wt), which has been dissolved in ethanolic solution (ethanol: aquadest = 1:1) slowly. Next, the mixture was stirred for a few hours at room temperature and heated at 65 °C to form a slurry. The slurry was dried at 100 °C overnight and calcined at 400 °C for 4 h. The solid formed is labeled Cat-1. Bimetallic catalyst impregnated in sequential steps. The bimetallic catalyst impregnated on ZrO2 was carried out in sequential steps but under the same conditions as the co-impregnation method. The first Ni solution (10% wt), dissolved in the ethanolic solution, is added to 1 g ZrO2 under stirring. The mixture was stirred for a few hours at room temperature and heated at 65 °C until a slurry was formed. Then, the slurry was dried at 100 °C overnight. The solid formed was then calcined at 400 °C for 4 h. Subsequently, a Zn solution (10% wt) was impregnated, dried, and

144

F. ‘Adany et al.

calcined under the same conditions. So, in this case, the catalyst is double calcination. Impregnation in sequential steps was used to synthesize the Ni–Zn bimetallic catalyst. The solid made in two steps was designed as Cat-2.

2.3 Catalyst Characterizations X-ray Diffraction (XRD). The crystal structure of the synthesized catalyst was characterized by XRD using PANalytical EMPYREAN. The pattern was run with copper radiation (Cu K = 1.54056 Å) with a second monochromator at 40 kV and 30 mA. The step size is 0.0260, and an angle of 2θ is between 20 and 100°. The crystallite size was calculated using the Scherrer equation shown in (1). D=

Kλ β cos θ

(1)

D is the average value of the crystal diameter, K is a constant (0.9), λ is the wavelength, θ is the diffraction angle, and β is the full width at half maximum diffraction peak (FWHM) [21]. Scanning Electron Microscopy–Energy Dispersive X-ray (SEM–EDX). SEM– EDX was used to determine the morphology and particle distribution of the catalyst. The equipment used was an SEM Hitachi SU350 scanning microscope, which was additionally equipped with a chemical composition analysis system based on energy dispersion scattering—the EDX.

3 Result and Discussion The crystal structure of the synthesized catalyst was determined using XRD. The XRD results are shown in Fig. 1, which shows that both catalysts have a diffractogram with the same pattern. Cat-1 and Cat-2 consist of ZrO2 , NiO, and ZnO phases. ZrO2 monoclinic is indicated by the appearance of a typical peak at 2θ = 17.45; 24.07; 24.46; 28.20; 31.49; 34.18° (PDF 00-024-1165). In addition, the peak of the NiO phase with the shape of a cube is also visible at 2θ = 37.24; 43.27; 74.99° ( ) and ZnO with a hexagonal shape at 2θ = 36.33; 67.96° (PDF 00-005-0664) ( ) [22]. The diffractogram in Fig. 1 shows that the typical peaks of both NiO and ZnO have low intensity, which indicates that NiO and ZnO are dispersed on the surface of ZrO2 . From the diffractogram in Fig. 1, the crystallinity of Cat-2 synthesized by the sequential impregnation method has lower crystallinity when compared to Cat-1 synthesized by the co-impregnation method. The same thing was also explained in the Romero-Sáez study (2018) that the catalyst prepared by the sequential impregnation

Influence Impregnation Method in the Structure of Bimetallic …

145

Fig. 1 XRD diffractogram of Cat-1 and Cat-2

method had lower crystallinity than the catalyst prepared by the co-impregnation method [23, 24]. The crystallite size of NiO and ZnO was measured using the Scherrer equation shown in Table 1. Crystallite size measurements were carried out at the dominant peaks of NiO at 2θ = 43.27° and ZnO at 2θ = 36.25° [25, 26]. The calculation results show that the NiO crystallite on the catalyst prepared by sequential impregnation (Cat-2) has smaller crystallite sizes. Meanwhile, the ZnO crystal size tends to be the same in Cat-1 and Cat-2, around 30.20 nm. This indicates that the smaller the crystallite size, the better the dispersion on the support surface. Figure 2 shows the surface morphology obtained from SEM of Cat-1 (Fig. 2a, b) and Cat-2 (Fig. 2c, d). From Fig. 2, the two catalysts have an irregular shape. However, the particle size of Cat-1 is more uniform when compared to Cat-2. Compared to the particle size, Cat-1 has a smaller particle size than Cat-2. The particle size of Cat-1 of 2.5–12.5 μm. Meanwhile, Cat-2 has a particle size of 5–50 μm. Figure 2b, d shows the surface morphology of the particles of the two catalysts. The elemental analysis results using EDX on Cat-1 are shown in Fig. 3, where Fig. 3a is the EDX mapping area of Cat-1. Elemental mapping results show that the particles consist of Zr (Fig. 3b), Ni (Fig. 3c), and Zn (Fig. 3d). Ni and Zn appear to Table 1 Crystallite size of the ZrO2 -based catalysts prepared by various impregnations methods

Catalyst

d NiO (nm)

d ZnO (nm)

Cat-1

19.45

30.20

Cat-2

14.79

30.21

146

F. ‘Adany et al.

Fig. 2 SEM micrograph of Cat-1 with 1000x a and 5000x b magnification and Cat-2 with 1000x c and 2000x d magnification

be homogeneously dispersed on the surface of ZrO2 . In addition, Fig. 3e shows that the percentage of Ni and Zn particles homogeneously dispersed on the ZrO2 surface is close to the theory (10%) namely 8.18 and 7.20%, respectively. The elemental analysis results on Cat-2 using EDX are shown in Fig. 4, where Fig. 4a is the Cat-2 EDX mapping area. Elemental mapping results reveal that the particles consist of Zr (Fig. 4b), Ni (Fig. 4c), and Zn (Fig. 4d). According to the elemental mapping, the dispersion of Ni is more even than Zn. This is because sequential impregnation is used, with Ni being contacted first with the support and followed by Zn [27]. The percentages of Ni and Zn on the surface of ZrO2 in Cat-2 are also close to the theoretical value of 10%, at 8.68 and 7.57%, respectively.

Influence Impregnation Method in the Structure of Bimetallic …

147

Fig. 3 SEM–EDX elemental mapping of Cat-1 where a secondary electron image and analogous elemental mapping of the element. b Zr, c Zn, d Ni, e EDX spectrum and table for the atomic and weight percentage of various elements

148

F. ‘Adany et al.

Fig. 4 SEM–EDX elemental mapping of Cat-2 where a secondary electron image and analogous elemental mapping of the element b Zr, c Zn, d Ni, e EDX spectrum and table for the atomic and weight percentage of various elements

Influence Impregnation Method in the Structure of Bimetallic …

149

4 Conclusion Both catalysts were synthesized successively. XRD diffractogram shows that both catalysts consist of ZrO2 , NiO, and ZnO phases. The NiO crystallite sizes in Cat-2 (14.79 nm) are smaller than that of Cat-1 (19.45 nm), but the ZnO crystallite size in both catalysts is almost the same, 30.20 nm. The results of SEM–EDX show that both catalysts have irregular particle shapes, with the particle size of Cat-2 being larger than Cat-1. The EDX elemental analysis showed that the amount of Ni and Zn in both catalysts was close to the theoretical amount (10%). However, further research is needed with further characterization and application to a specific reaction to determine which of the two catalysts has better performance. This is because each chemical reaction has different catalyst requirements, and each catalyst will produce different results in each reaction despite having the same properties. Acknowledgements This research is a part of the research Grant in the Research Organization of Nanotechnology and Materials-BRIN (No. 20/III.10/HK/2022). The authors also acknowledge the facilities, scientific and technical support from Laboratorium Bahan Maju Nuklir, and Advanced Characterization Laboratories Serpong National Research and Innovation Agency through E-Layanan Sains-BRIN.

References 1. Védrine, J.C.: Heterogeneous catalysis on metal oxides (2017) 2. Global climate strike. Nat Catal. 2(10), 831 (2019) 3. Rizwanul Fattah, I.M., Ong, H.C., Mahlia, T.M.I., Mofijur, M., Silitonga, A.S., Rahman, S.M.A., et al.: State of the art of catalysts for biodiesel production. Front Energ. Res. 8, 101 (2020) 4. Abdullah, S.H.Y.S., Hanapi, N.H.M., Azid, A., Umar, R., Juahir, H., Khatoon, H., et al.: A review of biomass-derived heterogeneous catalyst for a sustainable biodiesel production. Renew. Sustain. Energ. Rev. 70, 1040–1051 (2017) 5. Fatimah, I., Yanti, I., Suharto, T.E., Sagadevan, S.: ZrO2 -based catalysts for biodiesel production: a review. Inorg. Chem. Commun. 143, 109808 (2022) 6. Ghanbari, B., Kazemi, Z.F., Taheri, R.Z., Aghaei, E.: High-impact promotional effect of Mo impregnation on aluminum-rich and alkali-treated hierarchical zeolite catalysts on methanol aromatization. ACS Omega 5(21), 11971–11986 (2020) 7. Mehrabadi, B.A.T., Eskandari, S., Khan, U., White, R.D., Regalbuto, J.R.: A review of preparation methods for supported metal catalysts. Adv. Catal. 61, 1–35 (2017) 8. Erdogan, B., Arbag, H., Yasyerli, N.: SBA-15 supported mesoporous Ni and Co catalysts with high coke resistance for dry reforming of methane. Int. J. Hydrogen Energ. 43(3), 1396–1405 (2018) 9. Li, Z., Huang, Z., Ding, S., Li, F., Wang, Z., Lin, H., et al.: Catalytic conversion of waste cooking oil to fuel oil: catalyst design and effect of solvent. Energy 157, 270–277 (2018) 10. Koo, K.Y., Lee, J.H., Jung, U.H., Kim, S.H., Yoon, W.L.: Combined H2 O and CO2 reforming of coke oven gas over Ca-promoted Ni/MgAl2 O4 catalyst for direct reduced iron production. Fuel 153, 303–309 (2015) 11. Lycourghiotis, S., Kordouli, E., Sygellou, L., Bourikas, K., Kordulis, C.: Nickel catalysts supported on palygorskite for transformation of waste cooking oils into green diesel. Appl. Catal. B Environ. 259, 118059 (2019)

150

F. ‘Adany et al.

12. Takanabe, K., Nagaoka, K., Nariai, K., Aika, K.: Titania-supported cobalt and nickel bimetallic catalysts for carbon dioxide reforming of methane. J. Catal. 232(2), 268–275 (2005) 13. Zhang, J., Wang, H., Dalai, A.K.: Development of stable bimetallic catalysts for carbon dioxide reforming of methane. J. Catal. 249(2), 300–310 (2007) 14. Wu, H., Liu, J., Liu, H., He, D.: CO2 reforming of methane to syngas at high pressure over bi-component Ni-Co catalyst: the anti-carbon deposition and stability of catalyst. Fuel 235, 868–877 (2019) 15. Zhao, J., Zhou, W., Ma, J.: Pretreatment as the crucial step for biogas reforming over Ni–Co bimetallic catalyst—a mechanistic study of CO2 pretreatment. Int. J. Hydrogen Energ. 39(25), 13429–13436 (2014) 16. Wang, Z., Wang, C., Chen, S., Liu, Y.: Co–Ni bimetal catalyst supported on perovskite-type oxide for steam reforming of ethanol to produce hydrogen. Int. J. Hydrogen Energ. 39(11), 5644–5652 (2014) 17. Ferrari, M., Delmon, B., Grange, P.: Influence of the impregnation order of molybdenum and cobalt in carbon-supported catalysts for hydrodeoxygenation reactions. Carbon N Y 40(4), 497–511 (2002) 18. Bacariza, M.C., Graça, I., Lopes, J.M., Henriques, C.: Ni-Ce/zeolites for CO2 hydrogenation to CH4 : effect of the metal incorporation order. ChemCatChem 10(13), 2773–2781 (2018) 19. Salerno, P., Mendioroz, S., López, A.A.: Al-pillared montmorillonite-based NiMo catalysts for HDS and HDN of gas oil: influence of the method and order of Mo and Ni impregnation. Appl. Catal. A Gen. 259(1), 17–28 (2004) 20. Abdul Razak, N.A., Mijan, N.-A., Taufiq-Yap, Y.H., Derawi, D.: Production of green diesel via hydrogen-free and solventless deoxygenation reaction of waste cooking oil. J. Clean Prod. 366, 132971 (2022) 21. Naika, H.R., Lingaraju, K., Manjunath, K., Kumar, D., Nagaraju, G., Suresh, D., et al.: Green synthesis of CuO nanoparticles using Gloriosa superba L. extract and their antibacterial activity. J. Taibah Univ. Sci. 9(1), 7–12 (2015) 22. Hongloi, N., Prapainainar, P., Seubsai, A., Sudsakorn, K., Prapainainar, C.: Nickel catalyst with different supports for green diesel production. Energy 182, 306–320 (2019) 23. Romero-Sáez, M., Dongil, A.B., Benito, N., Espinoza-González, R., Escalona, N., Gracia, F.: CO2 methanation over nickel-ZrO2 catalyst supported on carbon nanotubes: a comparison between two impregnation strategies. Appl. Catal. B Environ. 237, 817–825 (2018) 24. Bampos, G., Bebelis, S., Kondarides, D.I., Verykios, X.: Comparison of the activity of Pd–M (M: Ag Co, Cu, Fe, Ni, Zn) bimetallic electrocatalysts for oxygen reduction reaction. Top Catal. 60(15), 1260–1273 (2017) 25. Yan, X., Zhang, Q., Zhu, M., Wang, Z.: Selective hydrogenation of benzene to cyclohexene over Ru–Zn/ZrO2 catalysts prepared by a two-step impregnation method. J. Mol. Catal. A Chem. 413, 85–93 (2016) 26. Souza, I.C.A., Manfro, R.L., Souza, M.M.V.M.: Hydrogen production from steam reforming of acetic acid over Pt–Ni bimetallic catalysts supported on ZrO2 . Biomass Bioenerg. 156, 106317 (2022) 27. Jing, J., Yang, Z., Huo, J., Bai, H., Li, W.: Metal precursor impregnation sequence effect on the structure and performance of NiCo/MgO catalyst. Int. J. Hydrogen Energ. 44(16), 8089–8098 (2019)

Assessment of Pollution and Sources of Metals in the Brantas River in East Java, Indonesia Cicik Oktasari Handayani, Hidayatuz Zu’amah, and Sukarjo

Abstract Heavy metals in river water can significantly harm the river’s environment and the surrounding community’s health. This study examines the level and spread of trace element pollution in the Brantas River. Water samples were obtained from 25 distinct sampling points along the Brantas River. The water samples contained trace metals in lead, cadmium, chrome, nickel, copper, iron, and manganese. Using a contamination factor (CoFa), contamination degree (DC), and trace element pollution index (TEPI), the element contamination level in the Brantas River was assessed. In contrast, analysis of correlation, principal component (PCA), and cluster (CA) were utilized to evaluate the origins of trace metals. In succession, the lead, cadmium, and copper content have surpassed the quality level in 20 places, nine locations, and two locations, respectively. The average values of CoFa are as follows: Pb (2.63); Cd (1.09); Cu (0.74); Ni (0.40); Fe (0.26); Cr (0.10); and Mn (0.02). The DC result indicates that 76% of river water falls into moderate-to-heavy contamination. The TEPI value is greater than 100 (129.28). Hence, the water from the Brantas River cannot be used as a source of drinking water. Statistical investigations demonstrated that most trace elements originated from different anthropogenic sources.

1 Introduction Due to the threat, it poses to aquatic ecosystems and human health, heavy metal poisoning of river water has become a global issue [1]. Heavy metals are poisonous, persistent, and accumulative [2]. Population growth, fast industrialization, and intensive agricultural practices are the primary drivers of heavy metal pollution in river water [3]. The Brantas River, at 320 km in length, is the longest in East Java Province. The town utilizes the Brantas River as a water supply for industry, domestic use, energy C. O. Handayani (B) · H. Zu’amah · Sukarjo Research Center for Horticultural and Estate Crop, National Research and Innovation Agency, Cibinong Science Center, Jl. Raya Jakarta–Bogor Cibinong, Indonesia 16915 e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_15

151

152

C. O. Handayani et al.

production, and agriculture [4]. In addition, up to 60% of the rice harvested in the province of East Java comes from fields near the Brantas River [5]. The Brantas River’s water quality has deteriorated due to several excessive anthropogenic activities throughout its length. As a result, the Brantas River is suspected of being contaminated with heavy metals [6] from industrial, household, and agricultural waste discharge. The trace metal content in river water must be evaluated based on the description of the threat heavy metal pollution in river water poses to aquatic ecosystems and human health. Determining the level of contamination and distribution of heavy metal sources in the Brantas River is the objective of this investigation.

2 Method 2.1 Research Site Brantas River is found in East Java Province at coordinates 111°31 00 –112°57 05 east longitude and 7°11 43 –8°17 35 south latitude. The Brantas River starts in Sumber Brantas Village, Batu City, and flows through Malang, Blitar, Tulungagung, Kediri, Jombang, and Mojokerto. The Brantas River is 320 km long and gets an average of 2000 mm of rain per year, of which about 85% falls during the rainy season. There are 25 points where water samples can be taken. First, choose where to take water samples by looking at river water sources, such as clean water, polluted water, and water already used. The location where water samples were collected is shown in Fig. 1 (see Fig. 1). The technique of taking water samples at each location point is very dependent on the water discharge, so the sampling technique is adjusted to the standard rules in SNI 6989.57:2008 Section 57 regarding the method of taking surface water samples. River water samples were taken using a simple water sampling device with a ballast. Water samples were prepared by filtering them using filter paper. Then, the clear water extract filtrate was measured using an (Atomic Absorption Spectrophotometer (AAS). Heavy metals analyzed include Pb, Cd, Cr, Cu, Mn, Ni, and Fe. Heavy metal analysis was carried out at the Laboratory of the IAERI, Ministry of Agriculture of the Republic of Indonesia.

2.2 Heavy Metal Contamination Assessments Contamination Factor (CoFa) The CoFa value is the ratio of metal concentration in water to metal concentration, which is the standard for river water quality that has been defined [7]. The formula

Assessment of Pollution and Sources of Metals in the Brantas River …

153

Fig. 1 Location of the Brantas River water sampling

computes the contamination factor in (1): CoFa =

Csi Cbi

(1)

C si is the metal concentration measured in the water. At the same time, C bi is the metal concentration standard for river water quality, with the classification of contamination factors referred to by Gupta et al. [8]. Contamination Degree (DC) DC is a thorough approach for measuring heavy metal contamination [9]. The formula computes the DC value in (2): DC =

n=8 1 CoFai n i=1

(2)

n is the amount of heavy metal observed. Trace Element Pollution Index (TEPI) The TEPI determines whether river water can be used to make drinking water. If the TEPI value of river water is less than 100, it is safe to drink. If the TEPI value is higher than 100, it will hurt the health of the people who drink it [10].

154

C. O. Handayani et al.

 Wi Q i TEPI =  Wi  (Mi − Ii ) Qi =  × 100 (Si − Ii )

(3) (4)

where Qi is a sub-index of the i-th parameter; W i is the unit weight of the i-th parameter; n is the number of parameters; M i is the monitored value of the i-th heavy metal parameter; I i is the outstanding value of the i-th parameter; and S i is the expected value of the i-th parameter.

2.3 Multivariate Analysis Principal component analysis (PCA) was used to extract the principal components (PC) from the sampling point and evaluate the possible sources and variations of heavy metal in river water samples. Additionally, cluster analysis, or hierarchical cluster analysis (HCA), was carried out to examine the similarities and differences between the concentration levels of all identified heavy metals in river water sampling sites.

3 Result and Discussion 3.1 Heavy Metal Concentration in the Brantas River’s Water The Brantas River water samples had the highest Pb, Cd, Cr, Cu, Mn, Ni, and Fe (mg/L) at 0.178, 0.021, 0.012, 0.025, 0.066, 0.034, and 0.289, respectively. From highest to lowest, the metal concentrations were Pb > Cd > Cu > Fe > Ni > Cr > Mn, based on the highest value. Table 1 shows the statistical description of heavy metals in river water at each place where water samples were taken. Concentrations of heavy metals Pb, Cd, and Cu in some samples have exceeded the quality standards, respectively, namely 80%, 36%, and 8% samples. In contrast, other metals such as Cr, Mn, Ni, and Fe did not exceed the quality standards.

3.2 Heavy Metal Contamination Assessments Compared to other heavy metals, the average CoFa value of Pb metal is the highest, at 2.63. Based on the average CoFa values from the highest to the lowest, respectively, they are Pb > Cd > Cu > Ni > Fe > Cr > Mn. The projected heavy metal contamination

Assessment of Pollution and Sources of Metals in the Brantas River …

155

Table 1 The statistical description of heavy metal concentrations at the Brantas River water sampling location Statistical

Pb

Cd

Cr

Cu

Mn

Ni

Fe

n

25

25

25

25

25

25

25

Average

0.079

0.011

0.006

0.015

0.007

0.020

0.078

Minimum

0

0.005

0

0.01

0

0.01

0.039

Maximum

0.178

0.021

0.012

0.025

0.066

0.034

0.289

SD

0.052

0.005

0.003

0.004

0.017

0.007

0.068

Water quality standardsa

0.030

0.010

0.050

0.020

0.400

0.050

0.300

a IGR:

22/2021[11]

factor in river water is low, with a contamination factor value of less than one. The CoFa values for Pb metal at each sampling site vary, but the overall weight is more than one. Cd and Cu metals have CoFa values greater than 1 in many places, whereas Cr, Mn, Ni, and Fe do not. The percentage of all metal contamination factors observed in several categories can be seen in Table 2. Degree of Contamination The DC is also used to determine the overall concentration of heavy metals in river water. The DC values for the moderately contaminated, moderate to extremely contaminated, and heavily contaminated types are 20%, 76%, and 4%, respectively (see Fig. 2). Most of the water test regions along the Brantas River are contaminated with heavy metals, with moderate to severe pollution being the most prevalent. Trace Element Pollution Index The TEPI value calculated based on the average trace element from all water sampling locations can be seen in Table 3. The TEPI value is 129.28 (> 100), which shows that based on the average concentration of all metals in the water in the Brantas River, it is not feasible to be used as clean water and consumed by the community. Table 2 Percentage of contamination factors in each category Heavy metal

Mean

Low degree

Moderate degree

Considerable degree

Very high degree

% Pb

2.63

20.00

44.00

36.00

0.00

Cd

1.09

64.00

36.00

0.00

0.00

Cr

0.12

100.00

0.00

0.00

0.00

Cu

0.74

92.00

8.00

0.00

0.00

Mn

0.02

100.00

0.00

0.00

0.00

Ni

0.40

100.00

0.00

0.00

0.00

Fe

0.26

100.00

0.00

0.00

0.00

156

C. O. Handayani et al.

4%

Fig. 2 DC percentage value in each category

20%

Moderately contaminated Moderately to heavily contaminated Heavily contaminated

76%

Table 3 TEPI values of the average concentration of heavy metals at all water sampling locations Para-meters

Mean value mgl/L (M i )

Standard permissible value (S i )

Highest desirable value (I i )

Unit weighting factor (W i )

Pb

0.079

0.030

0

10

Cd

0.011

0.010

0

33.33

Qi

Qi * W i

TEPI

263.49

2634.86

129.28

108.57

3618.77

Cr

0.006

0.050

0

2

11.60

23.21

Cu

0.015

0.020

0

0.05

73.68

3.68

Mn

0.007

0.400

0

0.2

1.73

0.35

Ni

0.020

0.050

0

0.5

40.26

20.13

Fe

0.078

0.300

0

3.33

26.14

87.11

Multivariate Analysis Principal component analysis was used to identify the main components associated with heavy metal sources in Brantas River water at different water sampling locations. There are two factors according to this criterion that can represent 64.71% of the total variance so that it can show information on most of the heavy metals in river water. The two main components proportions were 37.430% and 27.275%, respectively (Table 4). The first principal component (PC1) has a cumulative variance value of 37.430%. Therefore, PC1 can represent a mixed source of lithogenic and anthropogenic activity. The evaluation of heavy metals shows that Mn, Ni, and Cr do not pollute river water; this indicates that these metals may have natural sources, such as rock erosion [12]. On the other hand, river water is contaminated with Pb and Cd metals from anthropogenic sources. Metals like Cd and Pb probably come from intensive agricultural activities [13]. However, Pb and Cd metals can also come from industrial activities [14], domestic waste [15], and motor vehicles [16]. The second principal component (PC2) has a cumulative variance value of 27.275%. PC2 can be considered as representing a natural source because the value of the majority of Cu and Fe metals is still below the quality standard. The presence of Cu and Fe metals in river water is probably related to the environment around the Brantas River, which is in a volcanic environment [6].

0.67

−0.62

−0.78

−0.13

Cd

PC2

Pb

Element

PC1

Component

−0.23

−0.8

Cr 0.89

−0.03

Cu −0.37

0.63

Mn 0.04

0.62

Ni 0.74

0.34

Fe

Table 4 PCA with varimax rotation for all heavy metals found in the studied water sample

1.909

2.620

Eigen-values

27.275

37.430

% of variance

64.705

37.430

Cumulative %

Assessment of Pollution and Sources of Metals in the Brantas River … 157

158

C. O. Handayani et al.

Fig. 3 Dendrogram shows the clustering of heavy metals

The similarities and differences in the characteristics of the concentration values of all heavy metals observed between water sampling locations can be seen in HCA (see Fig. 3). The dendrogram shows the process/stages in the formation of clusters, starting from the initial stage on the far left with many clusters to the right with two clusters formed. The character of the heavy metal values of Pb, Cd, Cr, Ni, Cu, Fe, and Mn at 21 water sampling locations was almost uniform. In comparison, for the other four locations, namely, WS5, WS10, WS13, and WS22, the characters were different from the other 21 locations.

4 Conclusion There were twenty samples, nine samples, and two samples that exceeded the quality standard (Government Regulation of the Republic of Indonesia Number 22 of 2021) for the heavy metals Pb, Cd, and Cu, respectively. The average value of CoFa is as follows: Pb (2.63) > Cd (1.09) > Cu (0.74) > Ni (0.40) > Fe (0.26) > Cr (0.10) > Mn (0.02). The DC result indicates that 76% of river water falls into moderate-to-heavy contamination. The TEPI rating of 129.28 indicates that the water of the Brantas

Assessment of Pollution and Sources of Metals in the Brantas River …

159

River cannot be used as a source of drinking water for the local community. Multiple statistical tests, such as PCA, Pearson correlation matrix, and CA, showed that heavy metals originated from natural and anthropogenic sources.

References 1. Rajeshkumar, S., Liu, Y., Zhang, X., Ravikumar, B., Bai, G., Li, X.: Studies on seasonal pollution of heavy metals in water, sediment, fish and oyster from the Meiliang Bay of Taihu Lake in China. Chemosphere 191, 626–638 (2018) 2. Parween, M., Ramanathan, A.L., Raju, N.J.: Assessment of toxicity and potential health risk from persistent pesticides and heavy metals along the Delhi stretch of river Yamuna. Environ. Res. 202, 111780 (2021) 3. Luo, M., Yu, H., Liu, Q., et al.: Effect of river-lake connectivity on heavy metal diffusion and source identification of heavy metals in the middle and lower reaches of the Yangtze River. J. Hazard Mater. 416, 125818 (2021) 4. Roestamy, M., Fulazzaky, M.A.: A review of the water resources management for the Brantas River basin: challenges in the transition to an integrated water resources management. Environ. Dev. Sustain. 1–16 (2021) 5. Sujono, I.: Restorasi Air Sungai Brantas (Water Restoration of Brantas River). Osf, Surabaya (2019) 6. Mariyanto, M., Amir, M.F., Utama, W., et al.: Heavy metal contents and magnetic properties of surface sediments in volcanic and tropical environment from Brantas River, Jawa Timur Province, Indonesia. Sci. Total Environ. 675, 632–641 (2019) 7. Hoang, H.G., Lin, C., Tran, H.T., et al.: Heavy metal contamination trends in surface water and sediments of a river in a highly-industrialized region. Environ. Technol. Innov. 20, 101043 (2020) 8. Gupta, N., Yadav, K.K., Kumar, V., et al.: Appraisal of contamination of heavy metals and health risk in agricultural soil of Jhansi city, India. Environ. Toxicol. Pharmacol. 88, 103740 (2021) 9. Abrahim, G.M.S., Parker, R.J.: Assessment of heavy metal enrichment factors and the degree of contamination in marine sediments from Tamaki Estuary, Auckland, New Zealand. Environ. Monit. Assess. 136(1), 227–238 (2008) 10. Jazza, S.H., Najim, S., Adnan, M.A.: Using Heavy Metals Pollution Index (HPI) for assessment quality of drinking water in Maysan Province in Southern East in Iraq. Egypt J. Chem. 65(2), 1–2 (2022) 11. Peraturan Pemerintah No 22 Tahun 2021: Peraturan Pemerintah Nomor 22 Tahun 2021 tentang Pedoman Perlindungan dan Pengelolaan Lingkungan Hidup. Sekr Negara Republik Indones 1(078487A), 483 (2021) 12. Li, W., Qian, H., Xu, P., et al.: Distribution characteristics, source identification and risk assessment of heavy metals in surface sediments of the Yellow River, China. CATENA 216, 106376 (2022) 13. Fei, X., Lou, Z., Xiao, R., Ren, Z., Lv, X.: Source analysis and source-oriented risk assessment of heavy metal pollution in agricultural soils of different cultivated land qualities. J. Clean Prod. 341, 130942 (2022) 14. Bi, X., Zhang, M., Wu, Y., et al.: Distribution patterns and sources of heavy metals in soils from an industry undeveloped city in Southern China. Ecotoxicol. Environ. Saf. 205, 111115 (2020)

160

C. O. Handayani et al.

15. Yang, Q., Zhang, L., Wang, H., Martin, J.D.: Bioavailability and health risk of toxic heavy metals (As, Hg, Pb and Cd) in urban soils: a Monte Carlo simulation approach. Environ. Res. 214, 113772 (2022) 16. Huang, C., Zhang, L., Meng, J., et al.: Characteristics, source apportionment and health risk assessment of heavy metals in urban road dust of the Pearl River Delta, South China. Ecotoxicol. Environ. Saf. 236, 113490 (2022)

Utilization EDGAR Fifth Version as Input Emission of WRF-Chem Model for Simulating Ozone and PM2.5 Over Jakarta and Its Surroundings Area Prawira Yudha Kombara and Ninong Komala

Abstract The simulation of PM2.5 and Ozone has been conducted in this study in Jakarta and the surrounding area. PM2.5 and ozone at the surface level are one of a pollutant that can harm human health, especially the respiration system. Therefore, it is an important thing to learn the distribution pattern of the pollutants. The simulation was conducted using the WRF-Chem model utilizing Emission Database for Global Atmospheric Research (EDGAR) fifth version as anthropogenic emission input for the model. The simulation was done from 2nd until August 04, 2021. The result of the simulation was compared with observation data from The Ministry of Environment and Forestry (KLHK) and the U.S. Embassy. There are three points of observation data that be used for the comparison. Those points are Central Jakarta, Serang, and Bekasi point. Based on the comparison process, the suitability pattern between model results and observation data is still a tough thing to be reached for both pollutants. Furthermore, the daily pattern is still not similar yet too. In addition, based on Root Mean Square Error (RMSE) calculation, the PM2.5 parameter gives better result than the ozone parameter for both observation points.

1 Introduction As one of the big cities, Jakarta is still facing the problem of air pollution. The enhancement number of residents every year has resulted in massive community activities using motorized vehicles, causing the air quality in Jakarta increasingly declines. In addition, the increase in industrial and factory activities around the Jakarta area also has contributed to degradation of air quality. The problem of poor P. Y. Kombara (B) · N. Komala Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_16

161

162

P. Y. Kombara and N. Komala

air quality has resulted in degradation in the quality of life of the citizens of DKI Jakarta due to unhealthy air. Based on IQAir [1], Jakarta is ranked as the ninth most polluted city in the world. The problem of air pollution can be handled with the right policies. We can start from an academic study of how the pattern of air pollution that occurs in Jakarta. Understanding the pattern of air pollution is critical thing to arrange a right policy. Understanding the pattern of air pollution in Jakarta and its surroundings can be done by using several methods such as model simulation, spatial interpolation of ground observation data, or estimation with linear regression [2–5]. However, if we want to understand the overall pattern and mechanism of air pollution that occurs, the use of model simulation is the right choice. Simulation models for air pollution can be done using numerical models, such as the WRF-Chem model. The WRF-Chem model is a numerical model that combines the weather and chemical components of the atmosphere [6–9]. The WRF-Chem model requires meteorological and emission input data. Therefore, to produce a good simulation, the latest emission data is needed as input data for the model. Earlier this year in February 2022, the fifth version of the EDGAR emission data was released to the public. The EDGAR is the Emissions Database for Global Atmospheric Research that contains global anthropogenic emission data [10]. This latest version of EDGAR data can be used as emission input data for the WRFChem model. However, the fifth version of the EDGAR data has not been tested for air pollution simulations in Indonesia. Therefore, in this study, we tried to use the fifth version of EDGAR data for emission input of the WRF-Chem model. The pollutant parameters that will be simulated in Jakarta and its surroundings for this trial are PM2.5 and ozone gas (O3 ). These two parameters are kinds of pollutants that are often encountered and become a problem in big cities. In addition, these two pollutants are very dangerous to human health, especially in the respiratory system so they can cause death [11, 12].

2 Data and Method The data used in this study consists of 1. Data Final Reanalysis (FNL) FNL data is reanalysis data produced by NCEP-NCAR. The FNL data used in this study has a temporal resolution of 6 h and a spatial resolution of 0.25°. This FNL data is used as meteorological input data for the WRF-Chem model. The web address to download this data is as follows: https://rda.ucar.edu. 2. The fifth version of The EDGAR Data The EDGAR data is global anthropogenic emission data produced by Joint Research Centre. This data has a monthly temporal and spatial resolution of 0.1°. The version of the data used in this study is the fifth version. This data will be used as input

Utilization EDGAR Fifth Version as Input Emission of WRF-Chem …

163

data for anthropogenic emissions to run the WRF-Chem model. The fifth version of EDGAR data can be downloaded on the page: https://zenodo.org/record/6130621#. YzQOKzRBy5d. 3. KLHK air quality observation data KLHK is the ministry that handles air pollution conducts continuous observations by installing an air quality measurement instrument called AQMS throughout Indonesia. This data is used to evaluate the simulation results of the WRF-Chem model. Two stations will be used in this study, namely Serang station, Banten (−6.1208 S, 106.1731 E), and Bekasi station, West Java (−6.237452 S, 106.992943 E). 4. PM2.5 observation by U.S. Embassy In the U.S. Embassy office, central Jakarta (−6.18108 S, 106.83015 E), there is an instrument that measures PM2.5 operationally. This instrument is operated by AirNow, and they provide PM2.5 data at hourly intervals. The data can be obtained from https://www.airnow.gov/international/us-embassies-and-consulates/# Indonesia$Jakarta_Central.

2.1 Model WRF-Chem Configuration In this study, two domains were used to run the WRF-Chem model. The first domain has a spatial resolution of 5 km, and the second domain has a spatial resolution of 1 km. The domain design used in this study is shown in Fig 1. The simulation was run for three days from August 2–4, 2021.

D1

D2

Fig. 1 Domain design for WRF-Chem simulation

164

P. Y. Kombara and N. Komala

Table 1 Configuration of parameterization schemes Parameterization schemes

1st domain

2nd domain

Microphysics

WSM6

WSM6

Cumulus

BMJ

–

Boundary layer

YSU

YSU

Longwave radiation

RRTMG

RRTMG

Shortwave radiation

RRTMG

RRTMG

Chemical

MOSAIC using 8 sectional aerosol bins

MOSAIC using 8 sectional aerosol bins

In addition, to run the WRF-Chem model, a schema configuration of several parameterizations is required. Table 1 shows the list of the parameterization schemes are used.

2.2 Model Evaluation The simulation results of the model were evaluated by two methods. The two methods are correlation and root mean square error (RMSE). Correlation Correlation is a value that expresses the relationship between two variables. The range of correlation value is from −1 to 1. To calculate the correlation, you can use (1). 

  Fi − Fi Oi − Oi r=  2  2  Fi − Fi Oi − Oi

(1)

RMSE Root mean square error (RMSE) is a method used to determine the difference between two variables. The greater the RMSE value, the greater the difference between the two variables. To calculate RMSE can use (2).  RMSE =

1 N (Fi − Oi )2 i=1 N

(2)

Variables F and O represent model simulation variables and observations variables.

Utilization EDGAR Fifth Version as Input Emission of WRF-Chem …

165

3 Result 3.1 PM2.5 Parameter The results of the comparison between model simulations and observations for the PM2.5 parameter for three days at three stations are shown in Fig. 2. The simulation results of PM2.5 at the Bekasi station is more similar to the observation data (Fig. 2b) than the Central Jakarta and Serang station points (Fig. 2a, c). Even so, for the PM2.5 pattern at the Bekasi station point, there is still lag in several time windows. In general, the PM2.5 simulation results at Bekasi station are slightly overestimated; in other words, the magnitude of the model simulation is slightly larger than the observation data. Furthermore, the comparison results at the points of the Central Jakarta and Serang stations are not as good as the Bekasi station. The PM2.5 simulation results for the Serang station are quite overestimate or the simulation magnitude is greater than the observation data, so it is difficult to find a similarity between the simulation results and the observation data. The opposite happened at the Central Jakarta station, where the simulation results were very underestimated compared to the observation data. Based on the simulation results, the maximum value for three days occurs on August 3, 2021, at 03–04 AM for Bekasi and Serang stations, while for Central Jakarta stations, it occurs on August 2 at around 10 PM. The results for calculating the correlation and RMSE parameters for PM2.5 at three stations are shown in Table 2. As discussed previously, the Bekasi station points have a smaller difference between simulation results and observations than Central Jakarta and Serang stations. This can be seen from the RMSE value for Bekasi station, which is only 4.3427, while Jakarta and Serang stations are larger. Then the correlation at three stations has not shown good results yet. As previously mentioned, the pattern between the simulation results and the observation data still does not show suitability. This is confirmed by the correlation value that does not exceed 0.5 from all stations. If the simulation results and three-day observation data are composited daily, a diurnal pattern will be obtained as shown in Fig. 3. The daily pattern at Bekasi station (Fig. 3b) shows a pattern that is almost similar between simulation result and observation data. Based on the simulation result, the maximum magnitude at the Bekasi station occurred at 6 AM, while the observation data occurred three hours later. Then, the Serang station still shows a large enough difference between the simulation results and the observation data. The daily pattern from the simulation results shows two peaks that occur at 04 and 08 AM, while the observation data tends to be flatter even though there is a little fluctuation. Likewise, for the Central Jakarta station. In addition, the very large difference between the simulation results and the observation data, the suitability of the pattern has not been produced.

166

P. Y. Kombara and N. Komala

Fig. 2 Comparison of model result and ground station for PM2.5 parameter in a Central Jakarta, b Bekasi station, and c Serang station

Table 2 Correlation and RMSE result for the PM2.5 parameter in Bekasi and Serang station

Station

Correlation

RMSE

Centre Jakarta

−0.0755

314.2647

Bekasi

0.4247

4.3427

Serang

−0.0207

9.5634

The 24-h pattern for the PM2.5 parameter has been mentioned in several previous studies. The PM2.5 pattern in the Bandung Basin based on the model results can be said to have a diurnal pattern. The lowest value occurs at 10 AM and the maximum value is at 06 AM [4]. When compared to Jakarta and its surroundings area, based on the simulation results in August 2021, PM2.5 in the surrounding Jakarta area

Utilization EDGAR Fifth Version as Input Emission of WRF-Chem …

167

Fig. 3 Diurnal pattern of PM2.5 parameter in a Central Jakarta, b Bekasi station, and c Serang station

has a semi-diurnal pattern. The difference in 24-h patterns could be due to differences in meteorological conditions, topography, and land use. Furthermore, based on Hutauruk et al. [13], the PM2.5 pattern in Jakarta is similar to the diurnal pattern. This difference may be due to Hutauruk et al. [13] averaging the data for several years so that there is a possibility that the semi-diurnal pattern does not appear because of the averaging process. The reason why there is still a large discrepancy at station points other than Bekasi and the different pattern between the simulation results and the observation data has not been confirmed. Many things affect the simulation results of the WRFChem model such as initial meteorological conditions, domain design, selection of physical and chemical parameterization schemes, and emission input data. Pratama

168

P. Y. Kombara and N. Komala

and Sofyan [7] stated in their research, meteorological conditions greatly affect the dynamics of pollutants in the atmosphere, so they assimilate data for meteorological initial input to obtain more realistic initial conditions. However, in this study, we only update global anthropogenic emission data.

3.2 Ozone Parameter The comparison of simulation results and observation data for ozone parameters is shown in Fig. 4. The magnitude of the simulation results looks greater than the observation data. This shows the simulation results are overestimate. The result of overestimation occurred in two stations, either at Bekasi station or Serang station. For the suitability of the pattern, the Bekasi station almost did not show a suitable pattern between the simulation results and the observation data. This is due to the observation data pattern which tends to be more sloping than the simulation results even though there is a slight fluctuation. In the case of Serang station, there is a slight pattern match between the simulation results and the observation data.

Fig. 4 Comparison of model result and ground station for ozone parameter in a Bekasi station, and b Serang station

Utilization EDGAR Fifth Version as Input Emission of WRF-Chem … Table 3 Same as Table 2, for the ozone parameter

169

Station

Correlation

RMSE

Bekasi

−0.0310

22.0868

Serang

0.4863

21.8090

The unsuitability of the pattern and the large difference is strengthened by the results of the calculation of correlation and RMSE which are shown in Table 3. Based on the previous explanation, the unsuitability of the data pattern has not been seen at the Bekasi station. This is confirmed by the correlation value which is only −0.031, while for Serang station, there is still a pattern match even though the correlation value is still below 0.5. The difference between the magnitude of the simulation results and the observation data at the two stations is still quite large. This is shown by the RMSE value which reached 22.0868 for Bekasi station and 21.8090 for Serang station. Similar to the PM2.5 parameter, a daily composite for ozone parameter is also carried out to be able to see the daily pattern as shown in Fig. 5. Comparison of daily patterns at Bekasi station still shows a pattern discrepancy between the simulation results and the observation data. The maximum value from the observation data occurs first, while from the simulation results the maximum value occurs only a few hours later. Then for the Serang station, the maximum value based on the simulation results occurred twice, namely at 10 AM and 05 PM. In addition, pattern conformity occurs after 09 AM. Even so, the conformity of the pattern that occurs cannot be said perfect because there is a large difference in magnitude between the simulation results and the observation data. Based on Kitada et al. [2], ozone in Jakarta is produced after sunrise, after 6 AM exactly. It means the result of ozone simulation in this study matches with Kitada et al. [2] study result. We can see in Fig. 5 that ozone concentration from the simulation started to increase from 6 am. In other words, while the sun begins to rise and the photochemical reaction was begun.

4 Conclusion The simulation results for PM2.5 and ozone after replacing the fifth version of EDGAR data to the WRF-Chem model have not shown satisfactory results. There is still a discrepancy in the pattern and a large difference between the magnitude of simulation results and the observation data, especially for the ozone parameter. In general, the simulation results for the PM2.5 parameter are much better than the ozone parameter, especially at the Bekasi station. The results of the PM2.5 parameter simulation at the Bekasi station show a pattern that resembles the pattern of observation data. In addition, the value of the difference between the simulation results and the observation

170

P. Y. Kombara and N. Komala

Fig. 5 Diurnal pattern of ozone parameter in a Bekasi station, and b Serang station

data at the Bekasi station can be said to be quite small so that the simulation results can follow the magnitude of the observation data. Generally, the simulation results for both parameters can be said to overestimate the observation data except at the Central Jakarta station which results in an underestimate of PM2.5 simulation on the observation data. Acknowledgements We would like to thank to the Minister of Environment and Forestry for sharing the data of the related stations.

References 1. IQAir: World Air Quality Report. 2020 World Air Qual. Rep. 1–41 (2020) 2. Kitada, T., Sofyan, A., Kurata, G.: Numerical simulation of air pollution transport under sea/land breeze situation in Jakarta, Indonesia in dry season. NATO Sci. Peace Secur. Ser. C Environ. Secur. 243–251 (2008). https://doi.org/10.1007/978-1-4020-8453-9_27

Utilization EDGAR Fifth Version as Input Emission of WRF-Chem …

171

3. Solihah, K.I., Martono, D.N., Haryanto, B.: Identifying the best spatial interpolation method for estimating spatial distribution of PM2.5 in Jakarta. IOP Conf. Ser. Earth Environ. Sci. 893 (2021). https://doi.org/10.1088/1755-1315/893/1/012043 4. Salsabila, H., Turyanti, A., Nuryanto, D.E.: Estimation of the spatial distribution of maximum PM10 and PM2.5 concentration in Bandung City and surrounding countries using WRF-Chem Model (case study in July and October 2018). IOP Conf. Ser. Earth Environ. Sci. 893 (2021). https://doi.org/10.1088/1755-1315/893/1/012044 5. Cholianawati, N., Cahyono, W.E., Indrawati, A., Indrajad, A.: Linear regression model for predicting daily PM2.5 using VIIRS-SNPP and MODIS-Aqua AOT. IOP Conf. Ser. Earth Environ. Sci. 303 (2019). https://doi.org/10.1088/1755-1315/303/1/012039 6. Zhang, Q., Tong, P., Liu, M., Lin, H., Yun, X., Zhang, H., Tao, W., Liu, J., Wang, S., Tao, S., Wang, X.: A WRF-Chem model-based future vehicle emission control policy simulation and assessment for the Beijing-Tianjin-Hebei region, China. J. Environ. Manage. 253, 109751 (2020). https://doi.org/10.1016/j.jenvman.2019.109751 7. Pratama, A., Sofyan, A.: BANDUNG MENGGUNAKAN WRFCHEM DATA ASIMILASI PM 10 AIR POLLUTION DISPERSION ANALYSIS IN BANDUNG CITY USING WRFCHEM DATA ASSIMILATION PENDAHULUAN pengaruh penting terhadap kehidupan manusia dan ekosistem. Pencemaran udara berasal pengamatan yang dilakukan o. J. Tek. Lingkung. 26, 19–36 (2020) 8. Hong, J., Mao, F., Min, Q., Pan, Z., Wang, W., Zhang, T., Gong, W.: Improved PM2.5 predictions of WRF-Chem via the integration of Himawari-8 satellite data and ground observations. Environ. Pollut. 263, 114451 (2020). https://doi.org/10.1016/j.envpol.2020.114451 9. Grell, G.A., Peckham, S.E., Schmitz, R., McKeen, S.A., Frost, G., Skamarock, W.C., Eder, B.: Fully coupled “online” chemistry within the WRF model. Atmos. Environ. 39, 6957–6975 (2005). https://doi.org/10.1016/j.atmosenv.2005.04.027 10. Crippa, M., Solazzo, E., Huang, G., Guizzardi, D., Koffi, E., Muntean, M., Schieberle, C., Friedrich, R., Janssens-Maenhout, G.: High resolution temporal profiles in the Emissions Database for Global Atmospheric Research (2020) 11. Ismiyati, Marlita, D., Saidah, D.: Pencemaran Udara Akibat Emisi Gas Buang Kendaraan Bermotor. J. Manaj. Transp. Logistik. 1, 241–248 (2014) 12. Fitri, D.W., Afifah, N., Anggarani, S.M.D., Chamidah, N.: Prediction concentration of PM2.5 in Surabaya using ordinary kriging method. AIP Conf. Proc. 2329 (2021). https://doi.org/10. 1063/5.0042284 13. Hutauruk, R.C.H., Rahmanto, E., Pancawati, M.C.: Variasi Musiman dan Harian PM2.5 di Jakarta Periode 2016–2019. Bul. GAW Bariri. 1, 20–28 (2020)

GEMS Satellite Identification on Volcanic Ash Distribution of Mount Dukono Eruption in April 2022 Amalia Nurlatifah, Emmanuel Adetya, Asri Indrawati, Risyanto, Sumaryati, Ninong Komala, Nani Cholianawati, and Prawira Yudha Kombara

Abstract The Geostationary Environment Monitoring Spectrometer (GEMS) is geostationary satellite that is used for air quality monitoring aims. Utilization of GEMS now becomes one of the alternative solution of air quality monitoring problem in Asia. In this research, GEMS is used for monitoring the volcanic ash distribution as impact of Mount Dukono Eruption in April 2022. GEMS is used with wind hourly data from ERA-5. HYSPLIT Forward Trajectory Model is used to determine the direction and the distribution of ash from the eruption of Mount Dukono. Wind data shows that in April 2022, the prevailing wind moves from the southeast and west. This causes the volcanic ash tend to move from southeast to the northwest and east. This is in line with the results of the HYSPLIT model plot at 500, 1335, and 3000 m AGL. All plots indicate that the volcanic ash tends to move toward to the northwest and east. The results of satellite imagery from GEMS show a significant concentration of sulfur dioxide and aerosol in the North Halmahera area at the time of the incident and move to the northwest and east over time.

1 Introduction The Geostationary Environment Monitoring Spectrometer (GEMS) is the first Geostationary Earth orbit (GEO) satellite that is used for monitoring air quality [1]. GEMS was on boarding the Geostationary Korea Multi-Purpose Satellite 2 (GEOKOMPSAT-2) satellite series, which also hosts the Advanced Meteorological Imager (AMI) and Geostationary Ocean Color Imager 2 (GOCI-2). GEMS is providing some products for better understanding about air quality monitoring such as air quality A. Nurlatifah (B) · E. Adetya · Sumaryati · N. Komala · N. Cholianawati · P. Y. Kombara Research Center of Climate and Atmosphere, National Research and Innovation Agency, Central Jakarta, Indonesia e-mail: [email protected] A. Indrawati · Risyanto Laboratory Management, Research Facilities and Science, and Technology Park, National Research and Innovation Agency, Central Jakarta, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_17

173

174

A. Nurlatifah et al.

and meteorology condition monitoring, the long-range transport of air pollutants, emission source distributions, and chemical processes. Air pollution is one of the most important problems in the world. There is 4.2 million deaths per year due to stroke, lung disease, lung cancer, acute, and chronic respiratory illnesses as impact of air pollution [2]. Source of air pollution can vary, and the sources are divided into two types, namely natural sources and sources from anthropogenic activities. The source of air pollution can be from household combustion devices, motor vehicles, and industrial facilities (anthropogenic activities), forest fires, and volcano eruptions (natural source). Volcanic gases that can be the greatest potential hazards are sulfur dioxide, carbon dioxide, and hydrogen fluoride [3–7]. Volcanoes also can spew ash, a type of particulate matter air pollution, into the air for miles downwind of the eruption. Gases and aerosol that emit from volcano eruption can cause damage and hazard such as acid rain, climate change, long-range transport of ash/aerosol, decrease of visibility that can inhibit social and economic activity, and also a lot of toxic pollutants that can affect health [8–10]. So that, the volcano eruptions can cause air pollution problem that can be a source of damage in a lot of aspect such as health, social, and economy. Air quality monitoring as impact from volcano can be beneficial for managing the volcano eruption disaster. Managing the volcano eruption is very important so that the impact of that can be mitigated. Indonesia Maritime Continent as one of the areas that located in the ring of fire area is surrounded by many active volcanoes which can erupt at any time. One of this active volcano is Mount Dukono. Mount Dukono is an active and fairly remote volcano in eastern Indonesia. This mount is located in the northern end of Halmahera Island. Remote Dukono volcano on the island of Halmahera has been erupting continuously since 1933, with ash explosions, frequent ash plumes, and sulfur dioxide (SO2 ) plumes. Today, studies on monitoring volcanic activity, including eruptions, have developed a lot. Previous studies show that major volcanic eruptions emit large amounts of solid particles and volatile gases in the troposphere and the stratosphere [7, 11, 12]. Another study showed, in a statistical approach, the global distribution of volcanic SO2 degassing during the last century (1900–2000) and further indicated that each eruption (even for non-monitored ones) could affect the stratosphere, based on empirical observations [13]. This study aims to identification the distribution of ash and sulfur dioxide as impact from Mount Dukono eruption in April 2022 with GEMS.

2 Data and Method 2.1 Research Location The research location is Mount Dukono which is located in North Halmahera Province, precisely at 127.89 E and 1.693 N. The research conducted in April 24–29, 2022, 23:45 UTC/April 25–30, 2022 08:45 LT.

GEMS Satellite Identification on Volcanic Ash Distribution of Mount …

175

2.2 Data The data that is used in this study is as follows: 1. GEMS Data The GEMS data that is used in this research has daily temporal resolution and 20 km × 20 km spatial resolution. In this study, the GEMS data used include: • The slant column and the total column of SO2 were used to analyze the distribution of volcanic ash from the eruption of Mount Dukono. As we know that the volcano ash is major contained SO2 so that the SO2 concentration can reflect the activity of volcano. The SO2 algorithm fits the spectrum over the 310–326 nm (310–340 nm in volcanic regions) window for planetary boundary layer (PBL) SO2 SCD retrievals using a hybrid algorithm based on DOAS and PCA (Li et al. 2013). • Aerosol optical depth in the 550 nm spectrum which reflects the aerosol conditions in the study area during the eruption. 2. ERA-5 Wind Data ERA-5 is the fifth generation ECMWF reanalysis for the global climate and weather for the past 4–7 decades. The data assimilation system used to produce ERA-5 is the IFS Cycle 41r2 4D-Var. Wind data from ERA-5 that is used for this research has 0.25° spatial resolution and daily temporal resolution. The wind component that is used for this research is u-wind component and v-wind component.

2.3 Method 2.3.1

HYSPLIT Forward Trajectory Model

The Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) transport model is a complete system used to calculate and estimate air parcel trajectories such as transport, dispersion, chemical transformation, and deposition simulations [14]. HYSPLIT was run on the NOAA web (https://www.arl.noaa.gov/hysplit/). In this study, the HYSPLIT model is used with the output data from GDAS. This GDAS data has a spatial resolution of 0.25°. The model is run for an altitude of 500, 1335, and 3000 m AGL with a tracking time of 1 h ahead.

3 Results and Discussion On April 24–29, 2022, PVMBG (Center of Volcanology and Geological Hazard Mitigation) reported an increase in the activity of Mount Dukono, located in North

176

A. Nurlatifah et al.

Halmahera. This area is one of the areas covered by GEMS monitoring. Figure 1 shows the total value of the SO2 column at the time of the incident (April 24–29, 2022). In Fig. 1a, it can be seen that on April 24, 2022, at 23:45 UTC (April 25, 2022, at 08:45 LT), GEMS can provide an illustration that there is an increase in the concentration of the total column SO2 in the study area. This condition lasted for quite a long time, namely until the next day on April 25, 2022, at 23:45 UTC (April 26, 2022, at 08:45 LT) (Fig. 1b). Figure 1b shows the concentrated peak of the total column SO2 value in the study area. This indicates that Mount Dukono’s activity reached its maximum at the time of this incident. Mount Dukono’s activity was still recorded high the next day, namely on April 26, 2022, at 23:45 UTC (April 27, 2022, at 08:45 LT) which is also supported by the GEMS depiction of the total value of the SO2 column (Fig. 1c). Figure 1c shows that the total SO2 column value in the study area at this time still looks high. Interestingly, at this time it was seen that the total concentration of the SO2 column was spread lengthwise to the northwest of the study. This gives an indication that the SO2 resulting from the increased activity (eruption) of Mount Dukono on the third day of the study has been dispersed to the northwest. To prove it, a trajectory/dispersion model plot and wind movement analysis are needed which will be discussed in the next subchapter. The total value of the SO2 column at Mount Dukono is seen decreasing on April 27, 2022, at 23:45 UTC (April 27, 2022, at 08:45 LT) (Fig. 1d), then increasing again on April 28, 2022, at 23:45 UTC (April 29, 2022, at 08:45 LT) and tends to move

Fig. 1 GEMS image for total column amount of SO2 (molecules/cm2 × 106 ) in Mount Dukono on a April 24, 2022, b April 25, 2022, c April 26, 2022, d April 27, 2022, e April 28, 2022, and f April 29, 2022. All images are on 23:45 UTC

GEMS Satellite Identification on Volcanic Ash Distribution of Mount …

177

Fig. 2 Total column amount of SO2 at mount Dukono in April 2022 23:45 UTC

toward the northeast (Fig. 1e), then decreases again the next day, namely on April 29, 2022, at 23:45 UTC (April 30, 2022, at 08:45 LT) (Fig. 1f). To clarify the description of the total concentration value of the SO2 column due to the increased activity of Mount Dukono, Fig. 2 shows the time series of SO2 concentration against time in April 2022 including the time of the incident. As has been found in Fig. 1, it can be seen that the total column SO2 on April 25, 2022, at 23:45 UTC is the time with the highest concentration of slant column SO2 . At this time, MAGMA PVMBG ESDM reported that there had been a significant increase in activity at Mount Dukono. Figure 2 also shows that there has been a significant increase in the value of SO2 concentration at Mount Dukono during the eruption (April 24–29, 2022 23:45 UTC/April 25–30, 2022, 08:45 LT). This supports the report from PVMBG that during this period there was a significant increase in activity at Mount Dukono. From the explanation above, it can be concluded that GEMS is quite good in describing the total concentration value of the SO2 column around Mount Dukono during the period of increased activity of Mount Dukono (April 24–29, 2022, 23:45 UTC/April 25–30, 2022, 08:45 LT). In the presentation of Fig. 1c, it can be seen that the SO2 emitted from Mount Dukono is seen spreading to the northwest on April 26, 2022, 23:45 UTC/April 27, 2022, 08:45 LT. Then in Fig. 1e, it can be seen that SO2 is dispersed to the northeast on April 28, 2022, 23:45 UTC/April 27, 2022, 08:45 LT. From the two images, it can be seen that GEMS depicts the movement of SO2 resulting from the Mount Dukono eruption. However, to be able to measure its performance in describing the direction of pollutant dispersion from volcano emissions, this must also be proven by depicting the dispersion model and wind direction plots. Figure 3 shows a plot of the concentration value of the SO2 slant column along with the trend of its movement described by GEMS. Figure 4 shows the results of the forward trajectory volcano ash modeling from the HYSPLIT model which illustrates

178

A. Nurlatifah et al.

the direction in which the pollutants move, while Fig. 5 shows a plot of wind direction at an altitude of 750 mb from ERA-5 at the time of the incident. Figure 3a shows a plot of the concentration value of the SO2 slant column at Mount Dukono at the beginning of the increase in activity, namely on April 25, 2022, at 23:45 UTC/April 26, 2022, at 08:45 LT. It can be seen that there is a tendency for volcano ash to move to the northwest, although the concentration is still dominant around the study area. This is in accordance with the results of the HYSPLIT model plot presented in Fig. 4a where this plot shows that at that time the volcano ash at the three altitudes (500, 1335, and 3000 m AGL) showed a movement to the northwest. The plot of the wind direction at an altitude of 750 mb on April 26, 2022, 00:00 UTC/April 27, 2022, 09:00 LT also shows the dominant wind moving to the northwest (Fig. 5a). Figure 3b shows a plot of the concentration value of the slant column SO2 at Mount Dukono during the middle period of increased activity, namely on April 26, 2022, at 23:45 UTC/April 27, 2022, at 08:45 LT. It can be seen that SO2 or volcano ash has moved far to the northwest. This is in accordance with the results of the HYSPLIT model plot presented in Fig. 4b where this plot shows that at that time the volcano ash at all three altitudes (500, 1335, and 3000 m AGL) showed a movement to the northwest, further than the plot on the previous day, in Fig. 4a. The plot of the wind direction at an altitude of 750 mb on April 27, 2022, 00:00 UTC/April 28, 2022, 09:00 LT also shows the dominant wind moving to the northwest (Fig. 5b). Interestingly, the SO2 GEMS plot and the HYSPLIT model on this date show a further movement of SO2 compared to the previous day. This is supported and validated by the results of the plot of wind movement and speed on April 27, 2022, 00:00 UTC/April 28, 2022, 09:00 LT (Fig. 5b) where the plot results show that the wind movement is faster than before, on April 26, 2022, 00:00 UTC/27 April 2022 09:00 LT (Fig. 5a). Thus, there are indications that this faster wind speed is causing volcano ash to move faster and farther today (April 27, 2022, 00:00 UTC/April 28, 2022, 09:00 LT) compared to the previous day. Figure 3c shows a plot of the concentration value of the slant column SO2 at Mount Dukono at the end of the increase in activity, namely on April 28, 2022, at 23:45 UTC/April 29, 2022, at 08:45 LT. It can be seen that there is a tendency for volcano ash to change its movement compared to the previous two images, namely to the east. This is in accordance with the results of the HYSPLIT model plot presented in Fig. 4c where this plot shows that at that time volcano ash at two altitudes (500 and 1335 m AGL) also showed an eastward movement. The plot at these two altitudes shows the movement of volcano ash to the east although the movement is not too far away. This is supported by the plot of the wind direction at an altitude of 750 mb on April 29, 2022 00:00 UTC/April 30, 2022, 09:00 LT also shows that the wind at the study site is dominantly moving eastward at a low speed (Fig. 5c). From the explanation above, it can be concluded that GEMS describes the movement of SO2 from the volcano ash from the Mount Dukono eruption at the time of the incident which tends to move toward the northwest initially, then at the end of the volcanic ash eruption it moves to the east. This is confirmed by the results of the

Fig. 3 Slant column concentration of SO2 in mount Dukono in a April 25, 2022 23:45 UTC, b April 26, 2022 23:45 UTC, c April 28, 2022 23:45 UTC from GEMS

GEMS Satellite Identification on Volcanic Ash Distribution of Mount … 179

Fig. 4 HYSPLIT forward trajectory plot in mount Dukono on 500, 1335, and 3000 m AGL in a April 25, 2022 23:45 UTC, b April 26, 2022 23:45 UTC, and c April 28, 2022 23:45 UTC

180 A. Nurlatifah et al.

Fig. 5 Wind plot in mount Dukono on 750 mb in a April 26, 2022 00:00 UTC, b April 27, 2022 00:00 UTC, c April 29, 2022 00:00 UTC

GEMS Satellite Identification on Volcanic Ash Distribution of Mount … 181

182

A. Nurlatifah et al.

HYSPLIT plot at 500, 1335, and 3000 m AGL and the wind plot at 750 mb from the ERA-5 which shows the same movement. In addition to SO2 , aerosols are also one of the dominant components of volcanic eruptions. Aerosol optical depth (AOD) expresses the density of an area caused by aerosols. The higher the AOD value, the more concentrated the aerosol concentration in the area, and vice versa. Under normal conditions, the AOD value only ranges from 0.1 to 0.15. Figure 6 shows the AOD value during the eruption of Mount Dukono (April 25, 2022–April 29, 2022, LT). In Fig. 6a, it can be seen that on April 24, 2022, at 23:45 UTC (April 25, 2022 08:45 LT), GEMS can provide an image that there is an increase in the AOD value in the study area. The AOD value in Mount Dukono is close to 1 even though the AOD value in the surrounding area is only around 0.2–0.4 (Fig. 6a). This continued until the next day where the AOD value at Mount Dukono had value close to 1 (Fig. 6b). The peak of activity from Mount Dukono occurred on April 26, 2022 23:45 UTC/April 27, 2022, 08:45 LT where the AOD at Mount Dukono and around North Halmahera has value close to 1, and some in the surrounding area are around 0.6 (Fig. 6c). This is also in line with the depiction of the total concentration value of the SO2 column at this time (Fig. 1c) which also has a fairly high value. The depiction of the AOD value close to 1 which is quite widespread also illustrates that at this time volcano ash containing aerosols has spread in the area around the study area. AOD values close to 1 are also spread over the northwestern region of Mount Dukono. This confirms again that at this time the volcano ash from Mount Dukono tends to spread to the northwest. AOD values that are getting denser and closer to 1 becoming more widespread at Mount Dukono and its surroundings on April 27, 2022, 23:45 UTC/April 28, 2022, 08:45 LT (Fig. 6d). This indicates that there is a wider distribution of volcano ash around the study area due to the eruption of Mount Dukono. A fairly high AOD value also still occurred the next day, on April 28, 2022, 23:45 UTC/April 29, 2022, 08:45 LT (Fig. 6e). Figure 6e shows that the AOD value close to 1 occurs in the area around Mount Dukono and the surrounding eastern area. This is in line with the results of the plot in the previous discussion which stated that at this time the volcano ash from the Mount Dukono eruption was moving eastward. The AOD value decreased as Mount Dukono’s activity decreased on April 29, 2022, 23:45 UTC/April 30, 2022, 08:45 LT (Fig. 6f). Overall, GEMS is able to describe the condition of SO2 emissions from volcanoes, especially in the case of the eruption of Mount Dukono in Indonesia in April 2022. This proves that GEMS can be used for monitoring Mount Dukono’s activities. Case studies and similar results have also been reported in a study by [15]. Other study said that satellites play a significant and reliable role in monitoring volcanic activity (Ebmeier et al., 2018).

GEMS Satellite Identification on Volcanic Ash Distribution of Mount …

183

Fig. 6 GEMS image for aerosol optical depth (AOD) in mount Dukono on a April 24, 2022, b April 25, 2022, c April 26, 2022, d April 27, 2022, e April 28, 2022, f April 29, 2022. All images are on 23:45 UTC

4 Conclusion GEMS data states that there has been an increase in the concentration of SO2 and aerosols due to the increased activity of Mount Dukono on April 24–29, 2022, 23:45 UTC. On April 25–26, 2022, 23:45 UTC, volcano ash tends to move to the northwest. This is also confirmed by the results of the HYSPLIT plot which shows the movement of ash in the same direction. The wind movement at 750 mb of ERA-5 also shows consistent results where the prevailing wind in the study area tends to move to the northwest. On April 28, 2022, 23:45 UTC, volcano ash tends to move eastward. This is also confirmed by the results of the HYSPLIT plot which shows the movement of ash in the same direction. The wind movement at 750 mb of ERA-5 also shows consistent results where the prevailing wind in the study area tends to move to the east. Acknowledgements Authors would like to thank to National Institute of Environmental Research (NIER), Republic of Korea and The United Nations Economic and Social Commission for Asia and the Pacific (UNESCAP) for providing GEMS dataset and other technical/material supports for conducting air quality studies in Indonesia.

184

A. Nurlatifah et al.

References 1. GEMS Science Team: New era of air quality monitoring from space: geostationary environment monitoring spectrometer (GEMS). Bull. Am. Meteor. Soc. 101(1), E1–E22 (2020). https://doi. org/10.1175/BAMS-D-18-0013.1 2. https://www.who.int/news-room/fact-sheets/detail/ambient-(outdoor)-air-quality-and-health Accessed 19 Sept 2022 3. Bhugwant, C., Siéja, B., Bessafi, M., Staudacher, T., Ecormier, J.: Atmospheric sulfur dioxide measurements during the 2005 and 2007 eruptions of the Piton de La Fournaise volcano: implications for human health and environmental changes. J. Volcanol. Geoth. Res. 184(1), 208–224 (2009). https://doi.org/10.1016/j.jvolgeores.2009.04.012 4. Hobbs, P.V., Radke, L.F., Lyons, J.H., Ferek, R.J., Coffman, D.J., Casadevall, T.J.: Airborne measurements of particle and gas emissions from the 1990 volcanic eruptions of Mount Redoubt. J. Geophys. Res. 96(D10), 18735–18752 (1991). https://doi.org/10.1029/91JD01635 5. Gurwinder Sivia, S., Gheusi, F., Mari, C., Di Muro, A., Tulet, P.: Numerical simulations and parameterizations of volcanic plumes observed at Reunion Island. 10th EGU General Assembly, Apr 2013, Vienna, Austria. pp.EGU2013--10090. hal-00965476 6. Tabazadeh, A., Turco, R.P.: Stratospheric chlorine injection by volcanic eruptions: HCl scavenging and implications for ozone. Science 260(5111), 1082–1086 (1993). https://doi.org/10. 1126/science.260.5111.1082 7. Evans, B.M., Staudacher, T.H.: In situ measurements of gas discharges across fissures associated with lava flows at reunion island. J. Volcanol. Geoth. Res. 106(3–4), 255–263 (2001) 8. Re, M.: World Map of Natural Hazards, Munich Reinsurance Company (1998) 9. Brosnan, D.M.: Coral reefs: volcanic impacts, ecological impacts of the Montserrat volcano: a pictorial account of its effects on land and sea life. Sustainable Ecosystems Institute (2000). http://www.sei.org/impacts.html 10. Brantley, S., Myers, B.: Mount St. Helens—from the 1980 eruption to 2000 (Fact Sheet 036–00; Online Version 1.0). https://doi.org/10.3133/fs03600 11. Krueger, A.J., Walter, L.S., Bhartia, P.K., Schnetzler, C.C., Krotkov, N.A., Sprod, I., Bluth, G.J.S.: Volcanic sulfur dioxide measurements from the total ozone mapping spectrometer instruments. J. Geophys. Res. 100(7), 14057–14076 (1995). https://doi.org/10.1029/95J D01222 12. Schneider, D.J., Rose, W.I., Coke, L.R., Bluth, G.J.S., Sprod, I.E., Krueger, A.J.: Early evolution of a stratospheric volcanic eruption cloud as observed with TOMS and AVHRR. J. Geophys. Res. 104(4), 4037–4050 (1999). https://doi.org/10.1029/1998JD200073 13. Halmer, M.M., Schmincke, H.-U., Graf, H.-F.: The annual volcanic gas input into the atmosphere, in particular into the stratosphere: a global data set for the past 100 years. J. Volcanol. Geoth. Res. 115(3), 511–528 (2002). https://doi.org/10.1016/S0377-0273(01)00318-3 14. Draxler, R.R., G.D. Rolph (2003) HYSPLIT (HYbrid single-particle Lagrangian integrated trajectory) Model access via NOAA ARL READY website: http://www.arl.noaa.gov/ready/ hysplit4.html. NOAA Air Resources Laboratory, Silver Spring, MD 15. Gouhier, M., Pinel, V., Belart, J.M.C., et al.: CNES-ESA satellite contribution to the operational monitoring of volcanic activity: the 2021 Icelandic eruption of Mt Fagradalsfjall. J Appl. Volcanol. 11, 10 (2022). https://doi.org/10.1186/s13617-022-00120-3 16. Ebmeier, S.K., Andrews, B.J., Araya, M.C., et al.: Synthesis of global satellite observations of magmatic and volcanic deformation: implications for volcano monitoring and the lateral extent of magmatic domains. J Appl. Volcanol. 7, 2 (2018). https://doi.org/10.1186/s13617018-0071-3

Emission Inventory, Investigation of EKC Presence, Handling of SO2 Emission, and PBL SO2 Column Problems in Jakarta Toni Samiaji

Abstract Sulfur dioxide (SO2 ) gas produced from burning objects (including fuel) containing sulfur harms the environment. SO2 and NO2 gases can cause acid rain. Based on the results of BMG/BMKG (Meteorological and Geophysical Agency/Meteorology, Climatology, and Geophysics Council) data processing, rainwater’s power of hydrogen (pH) in Jakarta decreased from pH 5.47 in 1985 to 4.75 in 2020. In this regard, have SO2 emissions in Jakarta increased? This study aims to analyze the trend of total SO2 emission, investigate the presence of Environmental Kuznets Curve (EKC) and minimize the column and emission of SO2 in Jakarta. The data are fuel consumption, sulfur content, population, SO2 column in Planetary Boundary Layers (PBL), Gross Regional Domestic Product (GRDP), waste, and industrial production. The method used is the calculation of SO2 emissions using tier 2, the relationship between the economy and the environment based on the investigation of the existence of the EKC in Jakarta and minimizing the column and emission of SO2 with a Strength, Weakness, Opportunity, Threat (SWOT) analysis. The results found that the total SO2 emissions in Jakarta from 1970–2021 tended to increase from 41 to 119 kt. There was no EKC in Jakarta, then based on SWOT analysis, the right strategy to minimize emissions and SO2 column in Jakarta is diversification.

1 Introduction As a basis for making policies related to dealing with air pollution problems, planning for controlling air pollution in urban areas, and determining the layout of urban spaces, it is necessary to carry out an inventory of pollutant emissions and their sources, one of which is the inventory of SO2 . From previous research, SO2 inventory from fuel consumption in Jakarta in 2011 showed that SO2 emissions were from T. Samiaji (B) Research Center for Climate and Atmosphere, National Research and Innovation Agency, Jl. Sangkuriang, Cisitu, Bandung, West Java, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_18

185

186

T. Samiaji

industry and power plants at 78.22, followed by the transportation sector at 21.73 and the household sector at 0.05% [1]. Meanwhile, the trend of SO2 emissions in Jakarta carried out by Lestari (2016) shows that SO2 from 2010 to 2015 from power plants tended to decrease while those from transportation, industry, household, and commercial tended to increase [2]. This phenomenon raises the question of the trend from all these sectors when added up. International Institute for Applied Systems Analysis (IIASA), in the Asian model of Regional Air Pollution Information and Simulation (RAINS), conducts an inventory of SO2 not only from fuels but also from industrial processes [3]. The three studies did not involve sources of SO2 from burning waste, even based on the DKI (a particular area of the capital) Jakarta Provincial Sanitation Service, DKI Jakarta’s waste production per day is up to thousands of tons, especially in a year. This study will inventory SO2 emissions from fuel combustion, industrial processes, and household waste burning as a differentiator from previous research from 1970 to 2021 in Jakarta is a novelty in this study, and then, the trend is seen. So the research problem is how is the SO2 emissions trend from anthropogenic activities (fuel and household waste burning plus industrial processes) in Jakarta from 1970 to 2021. Increased fuel combustion causes economic growth but will emit SO2 gas into the air and cause adverse effects on the environment, or the environment degrades. One of economic growth describes per capita income. When per capita income is associated with environmental degradation and forms an inverted U curve, it is called the EKC [4] found. In China, Wang et al. (2016) found the EKC from 1990 to 2012 when linking SO2 emissions with Gross Domestic Product (GDP) [5], and also in 139 cities in India from 2001 to 2013 [6]. Next is in 12 European countries, when SO2 emissions per capita are related to GDP per capita in 1870–2001, the EKC is found [7]. In EKC, not only SO2 emissions but SO2 concentrations also are associated with GDP [8]. For example, Listyarini (2008) associated Jakarta GRDP with SO2 concentrations from 1993 to 2004, but the EKC was not found [9]. The concentration of SO2 from the surface to the height of the PBL forms a column of PBL SO2 . Then the research problem is if the author connects GRDP per capita with the PBL SO2 column, is the EKC found in Jakarta? If SO2 emissions and PBL SO2 columns in Jakarta increase, what efforts can reduce emissions and column SO2 ? In this regard, do not be careless in making decisions. For this reason, one way is to use a SWOT analysis [10]. People widely use SWOT analysis in strategic management [11], including in product marketing strategies [12] and dissecting business cases [13]. In addition, they use the SWOT analysis to assess the impact of a development project on the environment [14] or combine it with the cloud model for the sustainable water resources assessment [15] and others. People have not discussed the best way to reduce SO2 emissions and its column based on SWOT analysis. In this study, using SWOT analysis, the writer will look for the right strategy to reduce SO2 emissions and PBL SO2 columns in Jakarta if they increase.

Emission Inventory, Investigation of EKC Presence, Handling of SO2 …

187

2 Methods and Data The data used in this study are the sulfur content in fuel 1979–2025, fuel consumption in Jakarta 1970–2018, PBL SO2 column in 2005–2021, waste production 1990–2014, the population of Jakarta and Indonesia 1970–2019, GRDP and GDP 1970–2019, industrial production 1990–2017. The author obtained fuel consumption data from the Limited Liability Company of Pertamina and BPS (Central Statistics Agency) Jakarta Province, SO2 column National Aeronautics and Space Administration (NASA), the sulfur content the Directorate General of Oil and Gas, Ministry of Energy and Mineral Resources, and the Ministry of Environment and Forestry, population, industrial production, waste, and GRDP/GDP Jakarta and Central BPS. To calculate SO2 emissions in Jakarta, the writer use the equation E = A × EF

(1)

where E = emissions; A = activity rate; EF = emissions factor [16]. The activity rate can be in the form of the amount of fuel consumption each year. The author determined SO2 or SOx emissions from industrial processes by the number of particular goods produced or raw materials required and emission factors, as shown in 2. E pollutant = ARproduction × EFpollutant

(2)

Source [17, 18]. where E pollutant = SO2 or SOx emissions from industrial processes, A R pr oduction = the rate of activity of industrial processes, namely the number of particular goods produced or the number of raw materials needed, EF pollutant = emission factor of SO2 or SOx. The particular goods are goods that the process of their formation emits SO2 gas. According to Kuenen et al. [17], Kuenen et al. [18], Maddox [19], and Kuenen and Dellaert [20], SO2 or SOx emission factors from industrial processes depend on the type of goods produced. The researcher may calculate SO2 emissions from burning waste by the formula (2) above, only E pollutant = SO2 gas emission, A Rproduction = the rate of activity, namely the amount of garbage in the Tonnage burned annually. To make EKC, Selden and Song (1994) used the formula Y = β0 + β1 x + β2 x 2 + ε

(3)

where Y is emission per capita, β0 is a constant, β1 and β2 are the regression coefficient, x is GDP per capita, and ε is an error [21]. In another study at (3), Y is the ambient SO2 concentration, x is GRDP [9], while in this study, Y is the PBL SO2

188

T. Samiaji

column, and x is GRDP/capita. Considering that the SO2 concentration that occurs in Jakarta is not solely due to SO2 emissions from Jakarta, there is a contribution from the area around Jakarta caused by the wind that carries pollutants and clouds [22]. To overcome the problem of SO2 gas in Jakarta, the goverment of Jakarta and its surroundings need the right strategy and action. To overcome this problem, here, the author uses a SWOT analysis. Figure 1 shows a SWOT diagram; the x-axis extends from weaknesses to strengths, and the y-axis stretches from threats to opportunities. Deciding the exact strategy coordinates on the SWOT diagram, it is determined by X = TS − TW

(4)

Y = TO − TT

(5)

Ingaldi and Skurkova [24], where X are the coordinates on the x-axis, Y y-axis, TS is the total score of the internal aspects of strength, TW is weakness, TO is the external factor of opportunity, and TT is a threat. While the total score of each element is determined by TS =

ns 

R Si × W Si

(6)

RWi × WWi

(7)

i

TW =

nw  i

Fig. 1 SWOT diagram [23]

Emission Inventory, Investigation of EKC Presence, Handling of SO2 … no 

TO =

189

R Oi × W Oi

(8)

RTi × WTi

(9)

i nt 

TT =

i

Primadona and Rafiqi [25], where R Si is the rating of each factor point of strength, RWi weakness, R Oi opportunity, and RTi threat. Then W Si is the factor weight of strength, WWi weakness, W Oi opportunity, and WTi threat. In addition, the sum of each weight of each factor is equal to 1, namely [26] ns 

W Si = 1

(10)

Wwi = 1

(11)

W Oi = 1

(12)

WTi = 1

(13)

i nw  i no  i nt  i

3 Results and Discussion After analyzing the trend of total SO2 emissions in Jakarta from 1970 to 2021, the author found that emissions tend to increase by natural logarithm, as shown in (14). y = 19867 ln(x) + 40943 + ε

(14)

y is the total SO2 emission in Jakarta from 1970 to 2021, x is the year number, and ε is an error. Based on (14), the SO2 in Jakarta tends to increase from 41 to 119 kt from 1970 to 2021. Then based on the analysis of variance (ANOVA), as given in Table 1, it is found that the regression significance is 0.02 (smaller than 0.05). It means that the model shown by (14) has a 98% confidence level that is correct.

190

T. Samiaji

Table 1 The significance and regression coefficient of the independent variable

R2

Independent variable/model

0.104

Constant

40,943

0.120

Regression (ANOVA)

–

0.020

Ln(year number)

19,867

0.020

Coefficient

Sig

Fig. 2 Relationship curve of PBL SO2 column with GRDP per capita for the period 2005–2021 in Jakarta

Table 1 shows that the significance of the constant is 0.12, meaning that the value of this constant is 88% true. The value of R2 (determination factor) in Table 1 0.104 shows the importance of y in (14) 10.4% influenced by ln(x). It means the total SO2 emissions in Jakarta, 10.4%, are affected by time. In comparison, the remaining 89.6% is influenced by other factors, for example, fuel consumption, industrial processes, waste burning, and sulfur content in fuel. Figure 2 shows the relationship curve of the PBL SO2 column with GRDP per Capita for 2005–2021 in Jakarta. Based on (3), the curve has the equation. y = (5E − 06)x 2 − 0.0012x + 0.2838 + ε

(15)

However, y is the PBL SO2 column, and x is the GRDP per capita. Because the shape of the curve is U, the writer did not find EKC. In addition, when relating the ambient SO2 concentration with GRDP in Jakarta, EKC was not found [9]. Figure 2 means increasing community welfare lowers the SO2 column. Still, with increasing community welfare, the community is increasingly consumptive, meaning that fuel use and waste production also increase, thus increasing the SO2 column. Rising industrial production can cause increased waste production. Due to industrial production increases, SO2 emissions also increase, thus increasing the PBL SO2 column. The model significance shown in (15) is 0.008, so the U-shaped model is 99.2% correct. In this study, to analyze the existence of EKC in Jakarta, only GRDP/capita against the SO2 PBL column, because of the ANOVA significance of GRDP and GRDP per capita on SO2 emissions, GRDP per capita on SO2 emissions

Emission Inventory, Investigation of EKC Presence, Handling of SO2 …

191

per capita, GRDP and GRDP per capita on ambient SO2 concentrations, GRDP and GRDP per capita to the sulfate ion concentration in rainwater is higher than 0.05. The maximum significance value of 0.05 is a general requirement for a model to be accepted in statistics. For GRDP against the PBL SO2 column, although the significance of the ANOVA is less than 0.05, the regression coefficient significance for GRDP (or GRDP squared) is not defined (or cannot be determined). This model has a minimal regression coefficient, so the author considered the value of the PBL SO2 column equal to constant because the product of the regression coefficient by GRDP (or GRDP squared) is very small and thought zero. Judging from the column curve of SO2 to GRDP per capita, with the increase in people’s welfare, the environment in Jakarta, which initially experienced an improvement, became a slump. According to the equation model (15), in only six years, namely, from 2005 to 2011, the SO2 column experienced a decrease and then increased. The large SO2 column is due to the high SO2 emission. The high source of SO2 emissions is coal-fired steam power plants [27]. In addition, there are several industries not only from burning fuel but also from their industrial processes emitting SO2 gas as cement factories [28], metals [29], glass [30], oil and gas refineries [30], and petrochemicals [18]. Some of these factories are in Jakarta, and others are around Jakarta, as shown in Fig. 3. Considering what contributes to the SO2 concentration in Jakarta is not only SO2 emissions from Jakarta but also from around Jakarta [31], as well those who contribute to the PBL SO2 column Jakarta are not only SO2 emissions from Jakarta but also from outside Jakarta. These phenomena are shown in Fig. 3 of the airflow by the wind entering Jakarta [22]. So, strategies and actions are needed to reduce the value of the PBL SO2 column in Jakarta. Not only SO2 emissions from Jakarta need to be considered but also SO2 emissions outside Jakarta and the wind direction. Therefore by using SWOT analysis, the writer hopes to find the strategies to reduce emissions and the value of the PBL SO2 column in Jakarta. Table 2 shows the SWOT matrix to reduce the value of the PBL SO2 column and SO2 emissions in Jakarta and its surroundings. Based on the SWOT matrix and (4) and (5) equations, to determine the strategy for reducing the SO2 column value and SO2 emissions in Jakarta and its surroundings on the SWOT diagram, X = 3.043−1.4 = 1.643 is obtained, while Y = 2.696−3 = −0.304. The coordinates are = (1.643–0.304), so on the SWOT diagram (Fig. 1), it falls in quadrant IV or is − 0.187, so the angle the Strength Threat (ST) strategy. The tan θ = YX = −0.304 1.643 () formed by the S axis and line from the center of the coordinate to point (x, y) is −10.5° less than − 45°. This point location is in a diversification strategy [26]. This strategy means that a suitable method to reduce the value of the SO2 column and SO2 emissions in Jakarta and its surroundings is to maximize internal strength while reducing external threats. In this case, the community of Jakarta and its vicinity must implement strength factors such as regulations, use measuring equipment and information media, and utilize human and financial resources as well as possible. Meanwhile, the people must reduce threat factors such as the use of fossil fuels, motor vehicle congestion, sectoral egos, and population surges.

192

T. Samiaji

Fig. 3 Airflow (indicated by red lines) flowing into Jakarta in 2017–2020 as a result of the HYSPLIT and PROPER models during a the rainy season, b the dry season, including bringing SO2 pollutants from outside Jakarta. Source [22]

Emission Inventory, Investigation of EKC Presence, Handling of SO2 …

193

Table 2 SWOT matrix for decreasing SO2 column values and SO2 emissions Strength factors

Rating Weight Score

1. Having Air Quality Monitoring System (AQMS), measurement via satellite, High Volume air sampler, tools emission meter, anemometer

4

0.085

0.340

2. DKI Governor Regulation No. 31 of 2008 concerning emission thresholds for motor vehicle exhaust

4

0.085

0.340

3. Having human resources

3

0.064

0.191

4. Having financial resources

3

0.064

0.191

5. There is a Regional Environmental Laboratory at the Jakarta Environment Agency, BMKG

4

0.085

0.340

6. DKI Regional Regulation No. 2 of 2005 concerning Control of Air Pollution

4

0.085

0.340

7. Having information media to the public

2

0.043

0.085

8. Decree of the Governor of DKI No. 551 of 2001 concerning the determination of ambient air quality standard

3

0.064

0.191

9. DKI Governor Regulation No. 141 of 2007 concerning the use of CNG (Concentrate Natural Gas) for public transportation and local government operational vehicles

3

0.064

0.191

10. DKI Governor Regulation No. 88 of 2019 regarding traffic restrictions with an odd–even system

3

0.064

0.191

11. Decree of the Governor of Jakarta No. 545 of 2016 concerning the determination of the location and motor vehicle-free day schedule

1

0.021

0.021

12. Minister of Environment Regulation No. 12 of 2010 concerning the 1 implementation of air pollution control in the area

0.021

0.021

13. Minister of Environment and Forestry Regulation No. 1 of 2021 concerning Company Performance Rating Program in Management Environment

2

0.043

0.085

14. Law No. 32 of 2009 About the Protection and Environmental Management

2

0.043

0.085

15. The company/factory must provide Marginal External Cost, namely 1 include the cost of pollution or environmental damage in the production cost calculation

0.021

0.021

16. Governor of Jakarta, in supervising the air quality of Jakarta 1 Province, appoint a Regional Environmental Supervisory Officer at the Office of Environment Jakarta Province

0.021

0.021

17. There is SO2 transport from Jakarta to outside Jakarta

4

0.085

0.34

18. The existence of Intelligent Transport System-Area Traffic Control 1 System (ITS-ATCS)

0.021

0.021

19. The Environment Minister Decree No. 45 of 1997 concerning the Air Pollution Standard Index

1

0.021

0.021

Total

47

1

3.043

Weakness factors

Rating Weight Score (continued)

194

T. Samiaji

Table 2 (continued) 1. There has been no socialization of these policies to community/business entity

1

0.2

0.2

2. Human resources and funds are still lacking

1

0.2

0.2

3. Ineffective management and coordination

2

0.4

0.8

4. Dependence on fossil fuels is still large

1

0.2

0.2

Total

5

1

1.4

Opportunity factors

Rating Weight Score

1. Installing the SO2 emission reduction device on the chimney, exhaust 4

0.087

0.348

2. Building a tall chimney

1

0.022

0.022

3. Inventory of emissions both in Jakarta and outside Jakarta

2

0.043

0.087

4. Modeling the distribution of SO2 gas from outside Jakarta

2

0.043

0.087

5. Mixing limestone in coal burning

4

0.087

0.348

6. Controlling illegal parking

1

0.022

0.022

7. Relocating industries to areas far from Jakarta

4

0.087

0.348

8. Preparing Environmental Protection and Management Plan

3

0.065

0.196

9. Reduced fuel oil usage

3

0.065

0.196

10. Making a Strategic Environmental assessment

1

0.022

0.022

11. Analyzing environmental impacts

1

0.022

0.022

12. Switching to mass transportation

1

0.022

0.022

13. Testing SO2 emissions in motor vehicles and factory chimneys periodically

1

0.022

0.022

14. Reducing sulfur content in fossil fuels

4

0.087

0.348

15. Planting a tree that absorbs SO2 gas

1

0.022

0.022

16. Energy conservation

1

0.022

0.022

17. Supervision of compliance with emission quality standards, both mobile sources and sources are not moving

1

0.022

0.022

18. Conducting air pollution control education and training for a person 1 or entity whose activities make the air polluted

0.022

0.022

19. Energy-saving culture

1

0.022

0.022

20. Using renewable energy

4

0.087

0.348

21. Modification of fuel injectors

1

0.022

0.022

22. Eco-driving

1

0.022

0.022

23. Development of emission permit trading system

1

0.022

0.022

24. Energy diversification to environmentally friendly energy

2

0.043

0.087

Total

46

1

2.696

Threat factors

Rating Weight Score

1. Lack of public/business entity awareness of the dangers of air pollution

1

0.036

0.036

(continued)

Emission Inventory, Investigation of EKC Presence, Handling of SO2 …

195

Table 2 (continued) 2. Fossil fuels as fuel

1

0.036

0.036

3. Motor vehicle congestion

4

0.143

0.571

4. There is SO2 transport from outside Jakarta to Jakarta

1

0.036

0.036

5. Sectoral ego

2

0.071

0.143

6. Industry in Jakarta does not want to be relocated

3

0.107

0.321

7. Employees do not want to relocate

4

0.143

0.571

8. It costs money to install SO2 emission reduction devices

1

0.036

0.036

9. The existence of government projects such as toll road construction

3

0.107

0.321

10. Construction of a waste power plant

3

0.107

0.321

11. Increasing population

4

0.143

0.571

12. There is a plan to increase coal consumption in RUED P (provincial 1 region energy general plan) DKI

0.036

0.036

Total

1

3

28

4 Conclusion The study found that SO2 emissions in Jakarta from 1970 to 2021 tended to increase in a natural logarithm from 41 to 119 kt. Jakarta was no EKC where the y-axis is the PBL SO2 column while the x-axis is GRDP per capita. Based on the SWOT analysis, the right strategy to minimize SO2 emissions and SO2 column in Jakarta, viz diversification. In this case, the people of Jakarta and its surroundings must implement strength factors such as regulations, use measuring equipment and information media, and utilize human and financial resources as well as possible. Meanwhile, they also must reduce threat factors such as the use of fossil fuels, motor vehicle congestion, sectoral egos, and population surges.

References 1. Darmanto, N.S., Dan Sofyan, A.: Analysis of the distribution of air pollutants NO2 , SO2 , CO, and O2 in Jakarta using WRF-CHEM. J. Environ. Eng. 18(1), 54–64 (2012) 2. Erou, A., Fadhillah, F.: Inventarisasi emisi Provinsi DKI Jakarta. Seri Lembar Informasi Pencemaran Udara. Indonesian Center for Environmental Law. Icel.or.id (2019) 3. Amann, M., Dhoondia, J.: Rains-Asia User’s manual. International Institute for Applied Systems Analysis, A-2361 Laxenburg, Austria (1994) 4. Uchiyama, K.: Environmental Kuznets curve hypothesis. In: Uchiyama, K.(ed.). Environmental Kuznets Curve Hypothesis and Carbon Dioxide Emissions. Development Bank of Japan Research Series. https://doi.org/10.1007/978-4-431-55921-4_2 (2016) 5. Wang, Y., Han, R., Kubota, J.: Is there an environmental Kuznets curve for SO2 emissions? A semi—parametric panel data analysis for China. Renew. Sustain. Energy Rev. 54, 1182–1188 (2016) 6. Sinha, A., Bhattacharya, J.: Estimation of environmental Kuznets curve for SO2 emission: a case of Indian cities. Ecol. Ind. 72, 881–894 (2017)

196

T. Samiaji

7. Markandya, A., Golub, A., Pedroso-galinato, S.: Empirical analysis of national income and SO2 emissions in selected European Countries. Environ. Res. Econ. 35, 221–257 (2006) 8. Stern, D.I.: The rise and fall of the environmental Kuznets curve. World Dev. 32(8), 1419–1439 (2004) 9. Listyarini, S., Tarumingkeng, R.C., Fauzi, A., Hutagaol, P.: Kurva Lingkungan Kuznet: Relasi Antara Pendapatan Penduduk Terhadap Polusi Udara. Jurnal Organisasi dan Manajemen 4(1), 46–54 (2008) 10. Stacchini, A., Guizzardi, A., Mariotti, A.: Smoothing down arbitrariness in planning: From SWOT to participatory decision making. Land Use Policy 119. https://doi.org/10.1016/j.lan dusepol.2022.106213 (2022) 11. Rahim, A.R., Radjab, E.: Manajemen strategi. Cetakan 1, Lembaga Perpustakaan dan Penerbitan Universitas Muhammadiyah Makassar (2017). ISBN: 978-602-8187-60-2 12. Prastiyo, A., Fathoni, A., Malik, D.: Red ginger product marketing strategy through SWOT analysis at UD Company. Ungaran blessing. J. Manag. 4(4) (2018) ISSN: 2502–7689 13. Rangkuti, F.: Analisis SWOT: Teknik Membedah Kasus Bisnis. Gramedia, Jakarta. (1999) ISBN 979-605-718-2 14. Rathi, A.K.A.: Evaluation of project-level environmental impact assessment and SWOT analysis of the EIA process in India. Environ. Impact Assess. Rev. 67, 31–39 (2017). https://doi. org/10.1016/j.eiar.2017.08.004 15. Gao, X., Chen, L., Sun, B., Liu, Y.: Employing SWOT analysis and normal cloud model for water resource sustainable utilization assessment and strategy development. Sustainability 9, 1439 (2017). https://doi.org/10.3390/su9081439 16. Hatta, G.M., Asaad, I.: Regulation of the State Minister of the Environment No. 12 of 2010 concerning the implementation of air pollution control in the regions, Appendix II: Guidelines for the inventory of ambient air quality data and sources of air pollution (2010) 17. Kuenen, J., Berdowski, J., Most, P.V.D., Slager, J.M., Mulder, W., Hlawiczka, S., et al.: Iron and steel production. In: European Environment Agency. EMEP/EEA Air Pollutant Emission Inventory Guidebook, p. 21. Publications Office of the European Union, Luxembourg. ISBN: 978-92-9213-806-6 (2016a) 18. Kuenen, J., Appelman, W., Rentz, O., Oertel, D., Berdowski, J., Jonker, J., et al.: Chemical industry. In: European Environment Agency. EMEP/EEA air pollutant emission inventory guidebook, p. 26. Publications Office of the European Union, Luxembourg. ISBN: 978-929213-806-6 (2016b) 19. Maddox, R.R.: Gas and Liquid Sweetening, 2nd ed. Campbell Petroleum Series, Norman, OK (1974) 20. Kuenen, J., Dellaert, S.: Other metal production. In: European Environment Agency. EMEP/EEA air pollutant emission inventory guidebook, p. 5. Publications Office of the European Union, Luxembourg. ISBN: 978-92-9213-806-6 (2016) 21. Susandi, A.: The impact of international greenhouse gas emissions reduction on Indonesia. PhD thesis. Report on earth system science. Max Planck Institute for Meteorology. ISSN: 1614–1199. Hamburg (2004) 22. Myllyvirta, L., Suarez, I., Uusivuori, E., Thieriot, H., dan Anhauser, A.: Pencemaran udara lintas batas di Provinsi DKI Jakarta, Banten, dan Jawa Barat. Centre for Research on Energy and Clean Air (CREA) (2020) 23. Li, C., Negnevitsky, M., Wang, X.: Prospective assessment of methanol vehicles in China using FANP-SWOT analysis. Transp. Policy 96, 60–75 (2020). https://doi.org/10.1016/j.tra npol.2020.06.010 24. Ingaldi, M., Skurkova, K.L.: Company strategy determination in food companies using the SWOT method. Acta Technol. Agric. 3. https://doi.org/10.2478/ata-2014-0015 (2014) 25. Primadona, Y., Rafiqi, Y.: SWOT analysis on the competition strategy of the Madina Purbaratu minimarket, Tasikmalaya City. Sharia Econ. J. 2(1) (2019). P-ISSN: 2548–5032 26. Zhou, J., He, P., Qin, Y., Ren, D.: A selection model based on SWOT analysis for determining a suitable strategy for prefabrication implementation in rural areas. Sustain. Cities Soc. 50, 101715 (2019). https://doi.org/10.1016/j.scs.2019.101715

Emission Inventory, Investigation of EKC Presence, Handling of SO2 …

197

27. Song, H., Yang, M.: Analysis of Effectiveness of SO2 Emission Reduction in Shanxi, China by Satellite Remote Sensing. Atmosphere 5, 830–846. (2014)https://doi.org/10.3390/atmos5 040830 28. Zhang, T., Wu, C., Li, B., Wang, J., et al.: Linking the emission of cement plants to the sulfur characteristics of their limestones: a study of 80 NSP cement lines in China. J. Clean. Prod. 220, 200–211 (2019). https://doi.org/10.1016/j.jclepro.2019.02.133 29. Sun, X., Liu, Z., Huang, W., Ji, L., et al.: Looping upcycling SO2 into value-added H2 S by fast-induced reduction process for heavy metals treatment in the nonferrous smelting industry. Fuel 331, 125867 (2023). https://doi.org/10.1016/j.fuel.2022.125867 30. Kharol, S.K., Fioletov, V., McLinden, C.A., Shephard, M.W., et al.: The ceramic industry at Morbi is a large source of SO2 emissions in India. Atmos. Environ. 223, 117243 (2020). https:// doi.org/10.1016/j.atmosenv.2019.117243 31. Samiaji, T.: The scenario of sulfur content reduction on coal to fulfil the quality standards of SO2 concentration in the province of DKI Jakarta. Dissertation. Environmental Science Graduate School of Padjadjaran University. Bandung, Indonesia (2022)

Impact of Vortex on Rainfall as a Trigger for Tropical Cyclones (TC) in Maritime Continent of Southern Indonesia (Case Studies: Victoria, Ernie, and Seroja TC) Luthfiyah Jannatunnisa , Nurjanna Joko Trilaksono , and Muhammad Ridho Syahputra Abstract Tropical cyclones can have a destructive impact because they can directly affect weather conditions and wave height. In some cases, before becoming a tropical cyclone, it begins with a vortex that can produce weather disturbances such as rain, storms, and even tropical cyclones themselves. In this study, we investigated whether there are other possible vortex events, especially for the Maritime Continent of Southern Indonesia region. The data used are wind and relative vorticity from the reanalysis of the European Center for Medium-Range Weather Forecast (ECMWF) Reanalysis fifth Generation (ERA5) during the period of vortex events that grew into cyclones around the south of Indonesia from 2013 to 2021, then grouped for single, double, and triple vortex. A spatial rainfall intensity review was done using Global Satellite Mapping of Precipitation (GSMaP) satellite data when the vortex phenomenon occurred to assess the impact of the vortex event. Results showed that single vortex events were more common with seven events, double vortex with six events, and triple vortex with three events during 2013–2021. Single vortex formation area tends to be in the west of the Sumatra Region, double vortex formation tends to spread in the south of Indonesia, and the triple vortex occurs along the southern sea of Indonesia (west to east). The impact on rainfall spatially can be seen with the more vortex, the wider rainfall occurs in the south of Indonesia.

1 Introduction Tropical cyclones (TC)s are one of the phenomena that can cause hydrometeorological disasters in the territory of Indonesia. Although rare, these tropical cyclones L. Jannatunnisa Master Program in Earth Sciences, Faculty of Earth Science and Technology, Institut Teknologi Bandung, Bandung, Indonesia N. J. Trilaksono (B) · M. R. Syahputra Atmospheric Sciences Research Group, Faculty of Earth Science and Technology, Institut Teknologi Bandung, Bandung, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_19

199

200

L. Jannatunnisa et al.

have a destructive impact because they can directly affect weather conditions and even storm waves [1, 2]. TC is an intense vortex event in the atmosphere which is classified as a synoptic event on the atmospheric motion scale [3] because it has a life span of days to weeks and a typical size of about 200–500 km but can reach 1000 km (World Meteorological Organization—WMO). In some cases, before becoming a tropical cyclone, it begins with a vortex that can produce weather disturbances such as rain, storms, and even tropical cyclones itself [4, 5]. Previous research has shown that the vortex in southern Indonesia that has a high frequency (in the December– January–February season) is a vortex in the East Indian Ocean region, and the impact is that there is significant rainfall for the southern part of Sumatra and Java Island [6]. There is also a finding of double vortex in the northern and southern parts of the East Indian Ocean which can result in an increase in rainfall in several locations in Indonesia [7]. Therefore, from previous studies, it will be investigated whether there are other possible vortex events. The distribution of vortex events in the southern region of Indonesia has various characteristics [6]. So that a vortex characterization process will be carried out which is reviewed based on the number of vortexes, especially those that initialize the occurrence of tropical cyclones. This study aims to review vortex events in the Maritime Continent of Southern Indonesia based on differences in the number of vortex events at one time, then review the differences in their impact on rainfall.

2 Data and Method This study uses zonal and meridional wind data (u, v) as well as relative vorticity from the reanalysis of the European Center for Medium-Range Weather Forecast (ECMWF) Reanalysis fifth Generation (ERA5) during the period of vortex events that grew into tropical cyclones around the region of Southern Indonesia from 2013 to 2021. The data can be accessed at the link https://cds.climate.copernicus.eu/#!/ home. The vortex identification method will be carried out following the research procedure conducted by Chang et al. [8], namely the vortex is identified when there is a clockwise wind circulation in the southern hemisphere (southern Indonesia) at an altitude of 925 hPa and wind speed of more than 2 m/s in the southern hemisphere. Four grid points 2.5° × 2.5° from the center of circulation are located [8]. Once identified, spatial rainfall intensity is reviewed using Global Satellite Mapping of Precipitation (GSMaP; available online at https://hokusai.eorc.jaxa.jp) satellite data, by calculating rainfall anomalies, which are the differences in rainfall values (in mm/hour) during cyclone maturation with its initial vortex. It aims to evaluate the difference in the impact of these phenomena.

Impact of Vortex on Rainfall as a Trigger for Tropical Cyclones (TC) …

201

3 Results and Discussion The result will be divided into the following sections, the first section is about the vortex event in southern Indonesia, the second to fourth sections will discuss case study vortex analysis for the Victoria, Ernie, and Seroja TC, and the last section will present a discussion on the impact of all TC cases on the rainfall.

3.1 Vortex Events in Maritime Continent of Southern Indonesia The single vortex formation area tends to be in the west of the Sumatra Region (Fig. 1a) except for the case of Lili and Marcus, the double vortex tends to spread in the southern part of Java and Papua (Fig. 1b), and the triple vortex occurs along the southern waters of Indonesia (Fig. 1c). From this vortex event, one case was taken from each single, double, and triple vortex classification (marked with bold letters). Furthermore, Table 1 shows the incidence of vortexes that developed into tropical cyclones during the period 2013–2021. From the statistical results obtained, there were more single vortex events, with seven events followed by a double vortex with six events; in addition, three triple vortex events were found during the period 2013–2021.

3.2 Single Vortex on Victoria Cyclone Event (2013) Victoria TC developed in the Eastern Indian Ocean with fast and consistent cyclone movement toward the south–southeast. Figure 2 shows streamlines and vortices for initial vortex events (a and b), tropical cyclone (c and d), and cyclone dissipation (e and f). Seen on April 04 at 00 UTC, an initial vortex was formed which indicates a single vortex in the southwest of Sumatra. Then when entering the tropical cyclone stage on April 6 at 12 UTC, it was seen that there was an increase in wind speed and the relative vorticity value was more negative which tended to be circular and concentrated. To evaluate the impact of this phenomenon, an anomaly of rainfall (in mm/hour) is shown at the time of tropical cyclone and vortex initialization stage. Figure 3 shows a positive anomaly that tends to occur in southern Sumatra to several areas on the island of Java, this condition indicates that during the Victoria TC, there was an increase in rainfall in the area compared to when the initial vortex was detected. While the negative anomaly value indicates when the initial vortex is formed, there is an increase in rainfall in the waters of the eastern Indian Ocean and some areas in central Sumatra.

202

L. Jannatunnisa et al.

Fig. 1 Initial vortex in the southern Indonesian maritime continent in the period 2013–2021 for case studies of a single vortex on Victoria TC, b double vortex on Ernie TC, and c triple vortex on Seroja TC Table 1 Classification of vortex events in Maritime Continent of Southern Indonesia for case studies of vortexes that developed into tropical cyclones for the period 2013–2021

Single

Double

Triple

Manga (2020)

Nora (2018)

Seroja (2021)

Lili (2019)

Ernie (2017)

Riley (2019)

Marcus (2018)

Frances (2017)

Narelle (2013)

Flamboyant (2018)

Cempaka and Dahlia (2017)

–

Memories (2018)

Yvette (2016)

–

Victoria (2013)

Daffodils (2014)

–

Alessia (2013)

–

–

The bold letters indicate each single, double, and triple vortex classification case that was evaluated further in this study

Impact of Vortex on Rainfall as a Trigger for Tropical Cyclones (TC) …

203

Fig. 2 Figures a, c, and e are streamlines plots and b, d, and f plots of relative vorticity at an altitude of 925 hPa at the identified single vortex stage on April 4, 2013, at 00 UTC (a and b), the Victoria TC on April 6, 2013, at 12 UTC (c and d) and the dissipation stage of the cyclone or cyclone away from the study area on April 10, 2013, at 06 UTC (e and f)

Fig. 3 Anomalies of rainfall during Victoria TC and the initialization of the formation of a single vortex. Positive values (blue shaded) indicate rainfall anomalies when TC occurs, and negative values (red shaded) indicate rainfall anomalies during initialization of a single vortex

3.3 Double Vortex on Ernie Cyclone Event (2017) Cyclone Ernie grew in the southern waters of Indonesia, on April 4, southwest, and experienced a peak intensity of category five on April 7 at 12 UTC. Figure 4 shows streamlines and vorticity for initial vortex events (a and b), tropical cyclone (c and d), and cyclone dissipation (e and f) stage. On April 5, at 00 UTC, an initial vortex was formed, indicating the presence of a double vortex in the southern part of Java and around the Arafura Sea. Then when entering the tropical cyclone stage on April 7 at 00 UTC, it was seen that there was an increase in wind speed and the value of relative vorticity was more negative which tended to be circular and concentrated and the vortex in the Arafura Sea was strengthening.

204

L. Jannatunnisa et al.

Fig. 4 Figures a, c, and e are streamlined plots and b, d, and f plots of relative vorticity at an altitude of 925 hPa at the identified double vortex stage on April 5, 2017, at 00 UTC (a and b), the Ernie TC on April 7, 2017, at 00 UTC (c and d) and the dissipation stage of the cyclone or cyclone away from the study area on April 9, 2017, at 00 UTC (e and f)

Fig. 5 Same as Fig. 3, but for Ernie TC

Figure 5 shows a positive anomaly occurred in the eastern part of Java, Java Sea, North Sea of Australia, and South of Papua during the strengthening of the cyclone. While the negative anomaly value indicates when the initial vortex is formed, there is an increase in rainfall in the waters of southern Java and the Arafura Sea. Meanwhile, for the Nusa Tenggara archipelago and its surroundings, there was a reduction in rainfall due to the double vortex.

3.4 Triple Vortex on Seroja Cyclone Event (2021) Tropical Cyclone Seroja is quite well known because it has a destructive impact on East Nusa Tenggara and its surroundings. This cyclone moves slowly enough

Impact of Vortex on Rainfall as a Trigger for Tropical Cyclones (TC) …

205

over several days in the region to cause continuous heavy rains, floods, landslides, and even fatalities. Cyclone Seroja is a category three cyclone and is a cyclone that experienced the Fujiwara effect [9] with Cyclone Odette and was able to survive far to the south. Figure 6 shows streamlines and vortices for initial vortex events (a and b), tropical cyclone (c and d), and cyclone dissipation (e and f) stage. Seen on April 2 at 18 UTC, an initial vortex was formed indicating the presence of a triple vortex along the southern part of the southern Indonesian maritime continent ocean (southwest, south, and southeast). Then when entering the tropical cyclone stage on April 3 at 18 UTC, it was seen that there was an increase in wind speed and the relative vorticity value was more negative which tended to be circular and concentrated. Meanwhile, when Cyclone Seroja moved south on April 7 at 06 UTC, it was seen that the vortex that was previously in the southeast was already outside the study area, and another vortex was formed in the eastern Indian Ocean. Furthermore, Fig. 7 shows positive anomalies occurring in the southern region of Sumatra, western Java, southern Nusa Tenggara islands, and the Arafura Sea during the strengthening of the cyclone. While the negative anomaly value indicates that when the initial vortex is formed, there is an increase in rainfall in the waters of western Sumatra, the islands of Nusa Tenggara and its surroundings and the southern part of Papua. Meanwhile, for the central part of Java and Bali, there was a reduction in rainfall due to the triple vortex.

Fig. 6 Figures a, c, and e are streamlined plots and b, d, and f plots of relative vorticity at an altitude of 925 hPa at the single vortex identified stage on April 2, 2021, at 18 UTC (a and b), the Seroja TC on April 3, 2021, at 18 UTC (c and d) and the stage of dissipation of the cyclone or cyclone away from the study area on April 7, 2021, at 6 UTC (e and f)

206

L. Jannatunnisa et al.

Fig. 7 Same as Fig. 3, but for Seroja TC

3.5 Differences Impact of Rainfall on the Number of Vortex Events This section shows the impact of rainfall from vortex events quantitatively, by looking at rainfall during the initial vortex, tropical cyclone stage, and dissipation stage on five samples of weather observation stations in the study area. The locations are the Long Maritime Meteorological Station (Bandar Lampung City), Ahmad Yani Meteorological Station (Semarang, Central Java), David Constantijn Saudale Meteorological Station (Rote Ndao Regency, East Nusa Tenggara-NTT), Eltari Meteorological Station (Kupang City, NTT), and Merauke Climatology Station (Papua). In Fig. 8, the intensity of rainfall at Lampung station increased as the vortex formed until the tropical cyclone stage, then decreased after the cyclone weakened or moved away. However, in Central Java, rainfall tends to be constant and not too influenced by vortex events and tropical cyclones. Meanwhile, for other regions, observation data is not available. 50 40 30 20 10 0 04-04-2013

06-04-2013 Lampung

10-04-2013

Jawa Tengah

Fig. 8 Rainfall time series for the Victoria TC case study for Lampung (blue line) and Central Java (orange line) at the initial vortex formation (April 4, 2013), tropical cyclone stage (April 6, 2013), and cyclone dissipation stage (April 10, 2013)

Impact of Vortex on Rainfall as a Trigger for Tropical Cyclones (TC) …

207

Then for the double vortex in the case of Victoria TC is shown in Fig. 9. There is an increase in rainfall in the Central Java area until the cyclone grows, then decreases when the cyclone moves away. However, for the Lampung area, due to limited observation data, only an increase in rainfall is seen after the cyclone has moved away; this is due to the presence of strong winds of around 10–14 m/s (Fig. 4), and a new vortex system has emerged in the west of the area, Indonesia. Figure 10 shows that the availability of data is much better than the previous case study, but for Central Java, it is not available. For the cities of Kupang, Rote Ndao, and Lampung, it was seen that there was an increase in rainfall from the initial vortex to the occurrence of tropical cyclone and weakened thereafter. However, in Merauke City, there was a decrease in rainfall intensity from the initial vortex to tropical cyclone stage due to the movement of the cyclone away from that location. 10 5 0 05-04-2017

07-04-2017 Lampung

09-04-2017

Jawa Tengah

Fig. 9 Rainfall time series for the case study of Ernie TC for Lampung (blue line) and Central Java (orange line) at the initial vortex formed (April 5, 2017), tropical cyclone stage (April 7, 2017), and cyclone dissipation stage (April 9, 2017)

100 50 0 02-04-2021 Lampung

03-04-2021 Kota Kupang

Rote Ndao

07-04-2021 Kota Merauke

Fig. 10 Rainfall time series for the Seroja TC case study for Lampung (dark blue line), Central Java (orange line), Kupang City (gray), Rote Ndao (yellow), and Merauke City (light blue), at initial vortex forming (April 2, 2021), tropical cyclone stage (April 3, 2021), and cyclone weakening/removal (April 7, 2021)

208

L. Jannatunnisa et al.

4 Conclusions From the observations of vortex events in the southern part of southern Indonesian maritime continent for each vortex case that preceded tropical cyclones during the period 2013–2021, the vortex events can be divided into single, double, and even triple vortexes. This difference in the number of vortexes causes differences in the resulting impact, by reviewing the rainfall anomaly. For single vortex events in the case of Cyclone Victoria, it appears that there is one initial vortex formed and the impact of rainfall only occurs around central Sumatra. In the case of Cyclone Ernie, there were double vortexes in the waters of southern Java and the Arafura Sea, then the impact of the rainfall that occurred was an increase in the area and a decrease in rainfall around the islands of Nusa Tenggara and its surroundings. Furthermore, in the case of Cyclone Seroja, it was seen that a triple vortex was formed, with the impact of an increase in rainfall in the waters of western Sumatra, the islands of Nusa Tenggara and its surroundings, as well as the southern part of Papua. So, with the difference in the number of vortexes that are increasing (single to triple), it will provide weather disturbances, one of which is in the form of more widespread rainfall in the southern Indonesian maritime continent area, especially the southern part. Acknowledgements This work would not have been possible without the cooperation and guidance of the atmospheric sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung staff, who have supported this research and provided fruitful discussion, also Marine Technology Cooperation Research Center (MTCRC) which has funded the first author. Besides that, we thank ECMWF and JAXA GSMaP team for providing data for this research.

References 1. Suryantoro, A.: Tropical cyclones in south and southwest Indonesia from TRMM satellite monitoring and their possible links to high waves and tornadoes. Aerosp. Sci. Technol. Mag. 3(1) (2010) 2. Needham, H.F., Keim, B.D., Sathiaraj, D.: A review of tropical cyclone-generated storm surges: global data sources, observations, and impacts. Rev. Geophys. 53(2), 545–591 (2015) 3. Chan, J.C.: The physics of tropical cyclone motion. Annu. Rev. Fluid. Mech. 37, 99–128 (2005) 4. Hou, J., Wang, P., and Zhuang, S.: A new method of characterizing flow patterns of vortices and detecting the centers of vortices in a numerical wind field. J. Atmos. Oceanic Technol. (2017) 5. Emanuel, K.: Tropical cyclones. Annu. Rev. Earth Planet. Sci. 31(1), 75–104 (2003) 6. Pratama, B.E.: Study of vortex activities in the Indonesian maritime continent. Master’s Program Thesis, Institut Teknologi Bandung (2014) 7. Putra, I.K.Y.D., Trilaksono, N.J.: The impact of double vortex phenomena in eastern Indian Ocean on rainfall in western part of Indonesia. In AIP Conference Proceedings, vol. 1987, no. 1, p. 020045. AIP Publishing LLC (2018) 8. Chang, C.P., Harr, P.A., Chen, H.J.: Synoptic disturbances over the equatorial South China Sea and western maritime continent during boreal winter. Mon. Weather Rev. 133(3), 489–503 (2005) 9. Diamond, H.J., Schreck, C.J., Allgood, A., Becker, E.J., Blake, E.S., Bringas, FG., Camargo, S.J., Chen, L., Coelho, C.A., Fauchereau, N., Goldenberg, S.B., Woolley, JM.: The Tropics. Bull. Am. Meteorol. Soc. 103(8), S193–S256 (2022)

Study of the Inter-Tropical Convergence Zone (ITCZ) Movement Over the Maritime Continent Region Didi Satiadi, Ibnu Fathrio, and Anis Purwaningsih

Abstract Understanding the behavior of the ITCZ is crucial in order to understand the variability of weather and climate in the tropics. However, its behavior has not been fully understood, especially over the Maritime Continent region. This study aims to investigate the behavior of the ITCZ movement over the globe and the Maritime Continent region using a long-term monthly mean rainfall dataset from the fifth generation of ECMWF Atmospheric Reanalysis (ERA-5) from 2001 until 2020. In this investigation, the ITCZ positions were determined as the zonally averaged latitudinal position with maximum rainfall. In addition, the Australian monsoon indices (AUSMI) were used to investigate the relationship between seasonal migration of the ITCZ and the monsoon. The results of the study showed the bimodality of the ITCZ preferred locations, one in the north and another in the south of the equator. The seasonal meridional migration of the ITCZ generally followed that of the sun’s apparent position with a 1–2 month delay. We found that the movements of the ITCZ were not always gradual, but sometimes exhibited cross-equatorial rapid movement or “jump”. The seasonal migration of the sun’s apparent position caused one ITCZ position to become more dominant than the other, such that the ITCZ appeared to jump from the less to the more dominant position. Further study also showed that the ITCZ jumps generally coincided with the onset or withdrawal of the Australian monsoon. The results of this study supported and provided further proof to numerical experiments by Chao (Chao in J Atmos Sci 57:641–651, 2000), Chao and Chen (J Atmos Sci 58:2820–2831, 2001, Clim Dyn 22:447–459, 2004) using a General Circulation Model (GCM) and previous studies by Satiadi (Pergerakan Meridional dan Zonal ITCZ dalam Kaitannya dengan Curah Hujan di Wilayah Benua Maritim Indonesia, Prosiding Seminar Nasional Sains Atmosfer dan Antariksa, pp. 44–56, 2011), Satiadi and Fathrio (Jurnal Sains Dirgantara 9:1–11, 2011, Bimodalitas dan Lompatan ITCZ: Teori, Pengamatan dan Simulasi, Prosiding Seminar Nasional Sains D. Satiadi (B) · I. Fathrio · A. Purwaningsih Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, West Java, Indonesia e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_20

209

210

D. Satiadi et al.

Atmosfer dan Antariksa, pp. 68–76, 2012) using the Multi-functional Transport Satellite (MTSAT) and the Tropical Rainfall Measuring Mission (TRMM) data that showed similar results.

1 Introduction The Inter-tropical Convergence Zone (ITCZ) is a dominant phenomenon in the tropics that affects the weather and climate in this region, through modulation of wind, cloud, and rainfall. The ITCZ undergoes seasonal migration meridionally following the sun’s relative position to the earth. The ITCZ is a component of the Hadley/Walker circulation, where surface air converges and rises upward. The latent heat released in the ITCZ drives and interacts with the Hadley/Walker circulation. The ITCZ is easily visible on satellite imagery as a series of tropical synoptic waves stretching west to east. Each wave is a cloud group oscillating between connected and split states for about two weeks, during which the split conditions can create a cyclone. The ITCZ is a continuous cloud band with varying latitudes in the longitudinal direction due to zonal variations of surface conditions such as the SST, land distribution, and land characteristics [1]. Waliser and Gautier [20] provided the climatological characteristics of the ITCZ using long-term satellite’s Highly Reflective Cloud (HRC) dataset. An early theory of the ITCZ came from Charney [5] was developed based on the theory of the Convective Instability of Second Kind (CISK), to explain the mechanism of tropical cyclone formation. This theory states that the convergence of water vapor at the boundary layer acts as fuel for the growth of convection. The stronger the convection will give a feedback, further strengthening the convergence at the boundary layer. According to Charney, a large Coriolis force plays an essential role in the formation of convection, so the ITCZ is very likely to be found in polar regions, with a large Coriolis force. However, because the supply of water vapor is more abundant at the equator, ITCZ also tends to form near the equator. Subsequent research on ITCZ was carried out by Pike [13] and Sumi [19]. Pike and Sumi conducted experiments with a water planet model with uniform Sea Surface Temperature (SST) and showed the presence of ITCZ, which remained near the equator. This theory contradicts Charney’s theory which predicts that uniform SST should form ITCZ at the poles. Furthermore, an explanation of the reasons why the ITCZ tends to form near the equator which also answers the experimental results of Sumi and Pike was given by Chao [1], Chao and Chen [4]. Chao and Chen also answered the problem of the occurrence of double ITCZ in Sumi’s experiment, which was not caused by the horizontal resolution of the model used by Sumi. The water planet model experiment conducted by Chao [1] using the Global Circulation Model (GCM) has shown how important the ITCZ phenomenon is and answered the previous ITCZ problems. According to the ITCZ theory proposed by Chao, the ITCZ tends to occupy certain stable latitude positions. At this position, an equilibrium is reached between the gravitational force or the effect of the earth’s

Study of the Inter-Tropical Convergence Zone (ITCZ) Movement Over …

211

rotation and the location of the SST peak. The presence of preferred positions of the ITCZ was also found by Sikka and Gadgil [18] who found two favorable locations for a maximum cloud zone over the Indian longitudes (70° E–90° E) during April–October 1973–1977. Waliser and Somerville [21] also examined the preferred latitudinal position of the ITCZ based on observation, theory, and modeling analysis. Chao explained that the movement of the ITCZ is closely related to the change of seasons or monsoons. Chao defines a monsoon as the ITCZ located at latitudes more than 10° from the equator. Chao also showed that the meridional movement of the ITCZ undergoes a sudden jump from the equator to higher latitudes, further defined as the onset of the monsoon. Lau and Yang [11] also studied seasonal variation, abrupt transition, and intra-seasonal variability associated with the Asian Summer Monsoon using a global circulation model. They found that the onset of the Asian Monsoon is followed by the sudden jump of the equatorial ITCZ. Such an abrupt seasonal variation of the ITCZ was also observed by Hu et al [10] using the Global Precipitation Climatology Project (GPCP) daily data. Hagos and Cook [7] also studied the mechanism behind the abrupt latitudinal shift of maximum precipitation known as the West African monsoon jump using a regional climate model. They concluded that the jump is an example of multiscale interaction in the climate system, in which an intra-seasonal scale event is triggered by the smooth seasonal evolution of SSTs and the solar forcing in the presence of land–sea contrast. The monsoon theory based on the ITCZ is very different from the classical theory, which explains that the monsoon circulation is caused by the temperature difference between the land and the ocean. Chao’s experiments have proven that atmospheric circulation associated with the monsoon persists even without the land. The temperature contrast between the land and the sea only modifies the formation of the monsoon circulation and affects the location of the ITCZ. Chao argues that this new definition of monsoon applies not only to the western Pacific region, where conditions resemble the conditions of the water planet model, but also to all monsoon circulations in the tropics. The movement of the ITCZ associated with this monsoon can be explained by understanding the equilibrium between the two attractive forces acting on the ITCZ caused by the earth’s rotation (Coriolis force) and the location of the peak of the SST. The balance between these two pulling forces ultimately determines the positions favored by ITCZ. In line with the theory, [6] observed that the seasonal migration of the ITCZ was basically responsible for the variability of the Indian summer monsoon. Research has been carried out to study the movement of the ITCZ especially in the Maritime Continent region. The aim of the research is to investigate the preferred locations and the seasonal migration of the ITCZ and its relationship with the Australian Monsoon. This research is an extension of previous studies by Satiadi [14], Satiadi and Fathrio [17], Satiadi and Adikusumah [15] using the Multi-functional Transport Satellite (MTSAT), the Tropical Rainfall Measuring Mission (TRMM), and the Commonwealth Scientific and Industrial Organization (CSIRO) Global Circulation Model (GCM) data. In the current investigation, longer term (20 years) ERA-5 reanalysis data were used, and more in depth analyses were presented.

212

D. Satiadi et al.

2 Data and Methodology The gridded data of monthly average rainfall from the European Center for Medium Range Weather Forecasting Reanalysis 5 (ERA-5) was used to determine the position and seasonal migration of the ITCZ. This data was downloaded from the Climate Data Store (CDS) Catalog ([8, 9]. The data was obtained for a particular domain (180° W–180° E, 30° S–30° N) with 0.25° × 0.25° latitude–longitude resolution, for the 2001–2020 period. The position of the ITCZ was determined as the latitudinal position with the maximum rainfall. This was based on the assumption that the convergence zone of the Hadley circulation produces the tallest and thickest clouds and thus produces the most rainfall. Moreover, the mean latitudinal position of the ITCZ was calculated as the zonal mean (180° W–180° E). After the average position of ITCZ for each month was obtained, then the movement of ITCZ from month to month in a year was obtained and plotted. Furthermore, we examined the relationship between the seasonal migration of the ITCZ and the Australian Monsoon over the year by comparing the monsoon indices with the ITCZ movement over the tropical area (180° W–180° E) and the Australian Monsoon area (110°–140° E). The Australian Monsoon Index (AUSMI) data for the 2001–2015 period were obtained from the Asia Pacific Data Research Center of the International Pacific Research Center (IPRC) (http://apdrc.soest.hawaii.edu/pro jects/monsoon/daily-data.html#mon). The AUSMI was defined as an area-averaged zonal wind at 850 hPa level within an area bounded by (5° S–15° S, 110° E–140° E). The index was calculated for its monthly mean, and then averaged over 15 years (2001–2015). The seasonal change of the average indices was then compared to the ITCZ movement to investigate the relationship between the ITCZ and the monsoon.

3 Results and Discussion 3.1 Bimodality of the ITCZ The variability of the zonally averaged meridional profile of annual rainfall within 20 years (2001–2020), and its average were identified. A bimodal distribution of rainfall was indicated with two maxima, in the north of the equator (between 0° and 10° N), and in the south of the equator (between 10° S and 0°), separated by a minimum at the equator (see Fig. 1). The maxima corresponded to the average position of the ITCZ. However, the bimodal profile of the maxima was not symmetrical about the equator. The maximum in the north of the equator was generally higher (ranging from 6 to 8 mm/day) than that in the south of the equator (ranging from 4 to 6 mm/day). The asymmetry was presumably due to the difference in land/ocean composition between the northern and the southern hemispheres. The northern hemisphere (NH) has more land than ocean, while the southern hemisphere (SH) has more

Study of the Inter-Tropical Convergence Zone (ITCZ) Movement Over …

213

ocean than land. When the NH experiences summer, the air pressure will be lower in the NH than that in the SH, so that the wind will move from the SH to the NH. Conversely, when the SH experiences summer, the air pressure will be lower in the SH than that in the NH, so that the wind will move from the NH to the SH. However, because the NH is dominated by land while the SH is dominated by ocean, the summer pressure in the NH is lower than that in the SH, so that the ITCZ is pushed more strongly to the NH and spends more time in the NH. Such an asymmetry has also been found by Masunaga and L’ecuyer [12] in the East Pacific ITCZ. The zonally averaged meridional profile of annual rainfall is mainly determined by the variation of the Coriolis force and the Sea Surface Temperature (SST) with respect to latitude, so that the distribution is generally smooth as can be seen in Fig. 1. The meridional profile of the zonally averaged monthly rainfall shows a similar pattern to that of the annual rainfall, but the monthly profile depicts more variability between the months. In general, the monthly profile also indicated a bimodal and asymmetric distribution (see Fig. 2). Distinct profile for each month was presumably caused by the different apparent position of the sun during its annual cycle, which influenced the monthly distribution of rainfall. The position of the maximum rainfall, and hence the ITCZ, alternated between the NH and the SH (Fig. 2). For example, in January, February, and March, the maximum was in the SH, while in other months, it was in the NH. In January, the maxima in the SH and the NS were almost balanced. The bimodality was more pronounced from December to April, when both maxima were clearly visible. However, it was less pronounced from May to November, when only the northern maximum was clearly visible. It can be seen from Fig. 2, the shifting of the profile during its annual cycle. In January, the profile was in the southernmost position. In the following months, the profile shifted gradually to the north, the southern maximum decreased, while the northern maximum increased,

Fig. 1 Zonally averaged meridional profile of annual mean rainfall during 2001–2020

214

D. Satiadi et al.

Fig. 2 Zonally averaged meridional profile of monthly mean rainfall (2001–2020)

so that the asymmetry increased. In August, the profile reached the northernmost position. Subsequently, in the following months, the profile receded gradually to the south, the northern maximum decreased, while the southern maximum increased, so that the symmetry increased. Figure 2 also shows that the rainfall distribution in the SH was more spread out in latitude than that in the NH, which was more localized in latitude.

3.2 Seasonal Migration of the ITCZ The movement of the ITCZ during its annual cycle was not always gradual, but occasionally underwent a northward jump (between January and April) or southward jump (between December and January) (see Fig. 3). Moreover, it can be seen that the jump was actually the change of modality from the southern to the northern maximum and vice versa. The change of modality was related to the migration of the apparent position of the sun during its annual cycle. When the sun was in the SH (NH), then the southern (northern) maximum became more dominant than the other. It can be seen from Fig. 3 that the positions of the southern maximum have more variability than the northern one, which were more consistent. In general, the position of both maxima during its annual cycle followed the apparent position of the sun, which formed a sinusoidal pattern. However, the movement of the northern maximum seemed to experience a delay from that of the sun’s apparent position for about 2 months. The delay was thought to be caused by the seasonal hysteresis phenomenon, where the output (maximum rainfall) over land, which dominated the NH, lagged behind the input (the annual cycle of the sun’s apparent position). The hysteresis did not appear

Study of the Inter-Tropical Convergence Zone (ITCZ) Movement Over …

215

Fig. 3 Seasonal migration of ITCZ latitudinal position (2001–2020)

significant in the SH, which was dominated by the ocean, and hence responded very much slowly and weakly to the seasonal change of radiation. Zhang et al [22] investigated the ITCZ seasonal migration using GPCP data (1981–2010) over the Eastern Pacific, the Atlantic, the Indo-Western Pacific, and the global domain and found a similar track of the seasonal migration of the ITCZ. In February and March, the position of the northern maximum was the closest to the equator, while in August it was the farthest. In May, the southern maximum was the closest to the equator, while in January it was the farthest. In April, the positions of the maxima were the closest to each other.

3.3 ITCZ Jump and Monsoon Onset Relationships between the ITCZ and the monsoon were analyzed using a combined plot of the latitudinal position of ITCZ with the Australian Monsoon Index (AUSMI) against the month (January–December) during 2001–2015 from the ERA-5 (Fig. 4). The AUSMI was defined as an area-averaged zonal wind at 850 hPa level within an area bounded by (5° S–15° S, 110° E–140° E). The average values of AUSMI were positive in December–January–February, when the westerly wind dominated, which corresponded to the wet season in the southern part of the IMC (Fig. 4). Conversely, they were negative in other months, when the easterly wind dominated, which corresponded to the dry season in the southern part of the IMC. The AUSMI change sign from positive to negative in around March and from negative to positive between November and December. It can be seen from Fig. 4 that the northward jump of the ITCZ occurred in March, coincident with the sign change (from positive to

216

D. Satiadi et al.

Fig. 4 Position of ITCZ (180° W–180° E) vs the Australian Monsoon Index (AUSMI)

negative) of the AUSMI, which also occurred in March. However, the southward jump of the ITCZ occurred in December, a bit late (about a month) compared to the sign change (from negative to positive) of the AUSMI, which occurred in November. The northward jump was quite rapid presumably due to the fact that the NH was dominated by landmass, which attracted the ITCZ. For the same reason, the southward jump was delayed due to the fact that the southward jump was withheld by the attraction of landmass in the NH. The similar plot to Fig. 4 was employed for a different domain. This time, the zonal average was taken around the IMC, especially over the Australian Monsoon region (110° E–140° E). It can be seen from Fig. 5 that the northward jump of the ITCZ occurred in April, a bit late (about a month) compared to the sign change of the AUSMI, which occurred in March. However, the southward jump of the ITCZ in November coincided with the sign change of the AUSMI, which also occurred in November. The northward jump was delayed presumably due to the effect of the Australia Continent in this domain, which withheld the northward jump. For the same reason, the southward jump was rapid due to the attraction of the Australia Continent. Thus, the Australia continent may play an important role in determining the onset of the dry and wet season in this particular region.

4 Conclusions Research has been conducted to study the meridional movement of the ITCZ using monthly rainfall data from ERA-5 during 2001–2020. In this study, the meridional position of the ITCZ was determined as the latitudinal position with the maximum rainfall in the zonally averaged meridional profile of monthly rainfall.

Study of the Inter-Tropical Convergence Zone (ITCZ) Movement Over …

217

Fig. 5 Position of ITCZ (110°–140° E) vs the Australian Monsoon Index (AUSMI)

The zonally averaged meridional profile of annual rainfall exhibited a bimodal distribution with two maxima, one in the north of the equator and another one in the south of the equator, which corresponded to the two preferred locations of the ITCZ. The bimodal distribution was not symmetrical about the equator; the northern maximum was higher than the southern maximum. This asymmetry was presumably due to different land/sea composition between the northern and the southern hemisphere. More land in the NH and more ocean in the SH have caused the ITCZ to spend more time in the NH, which has the lowest pressure. The zonally averaged meridional profile of monthly rainfall also exhibited a bimodal distribution. However, during the annual cycle, the position of the maximum alternated between the NH and the SH depending on the apparent position of the sun. In January–February–March, the maximum tended to be in the south, while in other months in the north. The movement of the ITCZ generally followed the migration of the apparent position of the sun during its annual cycle, although it was delayed probably due to the seasonal hysteresis effect, especially in the NH. The movement of the ITCZ during its annual cycle did not always occur gradually, but occasionally experienced a northward or southward jump. The jump corresponded to the change in the modality from the southern maximum to the northern maximum and vice versa. We found that the poleward jumps occurred quite consistently every year during 2001–2020. Comparison between the ITCZ migration and the Australian Monsoon Index (AUSMI) showed that in general, the ITCZ jump corresponded to the monsoon onset. In the global domain, the northward jump coincided with the withdrawal of the Australian Monsoon, although the southward jump occurred a bit late compared to the onset of the Australian Monsoon. In the AUSMI domain, the southward jump coincided with the onset of the Australian Monsoon, although the northward jump occurred a bit late compared to the withdrawal of the Australian Monsoon. Therefore

218

D. Satiadi et al.

the ITCZ southward jump in the AUSMI domain was better in detecting the onset of the Australian Monsoon (the start of the wet season over the IMC), whereas the ITCZ northward jump in the global domain was better in detecting the withdrawal of the Australian Monsoon (the start of the dry season over the IMC). The results of this study supported and provided further proof to numerical experiments by Chao [1], Chao and Chen [2], Chao and Chen [4] using a General Circulation Model (GCM) and previous studies by Satiadi [14], Satiadi and Fathrio [17], Satiadi and Adikusumah [15] using the Multi-functional Transport Satellite (MTSAT) and the Tropical Rainfall Measuring Mission (TRMM) data that showed similar results.

References 1. Chao, W.C.: Multiple quasi equilibria of the ITCZ and the origin of monsoon onset. J. Atmos. Sci. 57, 641–651 (2000) 2. Chao, W.C., Chen, B.: Multiple quasi-equilibria of the ITCZ and the origin of monsoon onset. Part II Rotational ITCZ attractors. J. Atmos. Sci. 58, 2820–2831 (2001) 3. Chao, W.C., Chen, B.: The origin of monsoons. J. Atmos. Sci. 58, 3497–3507 (2001) 4. Chao, W.C., Chen, B.: Single and double ITCZ in an aqua-planet model with constant SST and solar angle. Clim. Dyn. 22, 447–459 (2004) 5. Charney, J.G.: Tropical Cyclogenesis and the Formation of the ITCZ. Mathematical Problems of Geophysical Fluid Dynamics. In: W. H. Reid (ed.) Lectures in Applied Mathematics American Mathematical Society, vol. 13, pp. 355–368 6. Gadgil, S.: The monsoon system: land–sea breeze or the ITCZ? J. Earth Syst. Sci. 127, 1 (2018). https://doi.org/10.1007/s12040-017-0916-x 7. Hagos, S.M., Cook, K.H.: Dynamics of the West African monsoon jump. J. Clim. 20, 5264–5284 (2007) 8. Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horanyi, A., Munoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., Simmons, A., Soci, C., Abdalla, S., Abellan, X., Balsamo, G., Bechtold, P., Biavati, G., Bidlot, J., Bonavita, M., Chiara, G., Dahlgren, P., Dee, D., Diamantakis, M., Dragani, R., Flemming, J., Forbes, R., Fuentes, M., Geer, A., Haimberger, L., Healy, S., Hogan, R.J., Holm, E., Janiskova, M., Keeley, S., Laloyaux, P., Lopez, P., Lupu, C., Radnoti, G., Rosnay, P., Rozum, I., Vamborg, F., Villaume, S., Thepaut, J.-N.: The ERA5 global reanalysis. Q. J. R. Meteorol. Soc. 2020, 1–51 (2020). https://doi.org/10.1002/qj.3803 9. Hersbach, H., Bell, B., Berrisford, P., Biavati, G., Horányi, A., Muñoz Sabater, J., Nicolas, J., Peubey, C., Radu, R., Rozum, I., Schepers, D., Simmons, A., Soci, C., Dee, D., Thépaut, J.N.: ERA5 monthly averaged data on single levels from 1959 to present. Copernicus Clim. Change Serv. (C3S) Clim. Data Store (CDS). (Accessed on 01-Sep-2022) (2019). https://doi.org/10. 24381/cds.f17050d7 10. Hu Y et al.: Abrupt seasonal variation of the ITCZ and the Hadley circulation, Geophysical Research Letters, 34, L18814. https://doi.org/10.1029/2007GL030950 11. Lau, K.M., Yang, S.: Seasonal variation, abrupt transition, and intraseasonal variability associated with the Asian summer monsoon in the GLA GCM. J. Clim. 9, 965–985 (1995) 12. Masunaga, H., L’ecuyer, T.S.: Equatorial asymmetry of the east pacific ITCZ: observational constraints on the underlying processes. J. Clim. 24, 1784–1800 (2011) 13. Pike, A.C.: Intertropical convergence zone studied with an interacting atmosphere and ocean model. Mon. Wea. Rev. 99, 469–477 (1971) 14. Satiadi, D.: Pergerakan Meridional dan Zonal ITCZ dalam Kaitannya dengan Curah Hujan di Wilayah Benua Maritim Indonesia, Prosiding Seminar Nasional Sains Atmosfer dan Antariksa, pp. 44–56 (2011). ISBN 978-979-1458-53-5

Study of the Inter-Tropical Convergence Zone (ITCZ) Movement Over …

219

15. Satiadi, D., Adikusumah, N.: Bimodalitas dan Lompatan ITCZ: Teori, Pengamatan dan Simulasi, Prosiding Seminar Nasional Sains Atmosfer dan Antariksa, pp. 68–76 (2012). ISBN 978-979-1458-64-1 16. Satiadi, D., Adikusumah, N.: ITCZ dan Monsun Indo-Australia, Prosiding Seminar Nasional Sains Atmosfer dan Antariksa, pp. 77–83 (2015). ISBN 978-979-1458-64-1 17. Satiadi, D., Fathrio, I.: Penentuan onset Monsun Di Wilayah indo-Australia Berdasarkan Lompatan ITCZ. Jurnal Sains Dirgantara 9(1), 1–11 (2011) 18. Sikka, D.R., Gadgil, S.: On the maximum cloud zone and the ITCZ over Indian longitudes during southwest monsoon. Mon. Weather Rev. 108, 1840–1853 (1980) 19. Sumi, A.: Pattern formation of convective activity over the aqua-planet with globally uniform sea surface temperature. J. Meteor. Soc. Japan 70, 855–876 (1992) 20. Waliser, D.E., Gautier, C.: A satellite-derived climatology of the ITCZ. J. Clim. 6, 2162–2173 (1993) 21. Waliser, D.E., Somerville, C.J.: Preferred latitude of the intertropical convergence zone. J. Atmos. Sci. 51(12), 1619–1639 (1994) 22. Zhang et al.: Advances in research on the ITCZ: mean position, model bias, and anthropogenic aerosol influences. J. Meteorol. Res. 35, 729–741 (2021)

Weather Condition Identification Using Edge Detection Method for Early Warning System Aisya Nafiisyanti, Farid Lasmono, Ibnu Fathrio, Risyanto, Teguh Harjana, Didi Satiadi, and Acep Catur Nugraha

Abstract This study is a preliminary examination of edge detection methods for the given data set produced by a numerical weather prediction model. This research compares three edge detection methods: Laplacian, Sobel and Canny. Each method is applied to images that represent flight weather information, specifically low-level wind shear and low level. The area coverage is the Indonesian region with -10° to 10° latitude and 95° to 145° longitude range. Moreover, hue, saturation and value (HSV) colour space is performed in the pre-processing stage to give a better performance. The characteristics of the test results are obtained and then used as the basis for selecting the most effective method to be implemented in the system. The Laplacian and Canny methods are capable of fulfilling the requirement even though the three methods generally are capable of detecting edges in the given images.

1 Introduction Image processing specifically in the atmospheric field and beyond has been used for various purposes and is implemented on different data sources. [1] applied a modified edge detection method which was developed from Canny methods on aerosol data from Lidar. Then, from this research, [2] exploited the detection results to identify Planetary Boundary Layer (PBL) height that plays a role to characterize the PBL itself. Another case was developed by [3], by extracting objects from Indian Remote Sensing Satellite (IRS) images that are derived from LISS-III, LISS-IV, and CartosatI sensors. Edge detection methods are also used by [4] to localize an Unmanned Aerial Vehicle (UAV) by detecting craters and calculating the craters’ distance using the Mahalanobis method on the Martian surface. One sector that has the potential to use image processing is aviation by informing flight weather conditions. JATAYU as an air transportation network developed by the Center of Climate and Atmospheric Research, Indonesian National Research and Innovation Agency, is a system that uses A. Nafiisyanti (B) · F. Lasmono · I. Fathrio · Risyanto · T. Harjana · D. Satiadi · A. C. Nugraha Indonesian National Research and Innovation Agency, Jakarta, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_21

221

222

A. Nafiisyanti et al.

images to convey information on its website [5] Each image describes the weather conditions of a parameter by classifying the values into several ranges. Each range is represented by a specific color. Thus, according to the classification, there might be objects in unregular shape in the image if a specific weather condition occurs. Previously, the identification process, judgment, and conclusions of determining the occurrence of specific weather conditions were drawn using human vision. Then, the warning can be delivered through the system. However, this method requires human resources and consistency and also has inefficient performance. As the system moves toward automation, we attempt to shift the early warning process from human vision to computer vision. This study uses edge detection as the approach to utilize computer vision. Detecting edges is one of the fundamental operations in image processing. It helps to reduce the amount of data (pixels) to process and maintains the necessary aspect of the image. Edges detection is a problem of fundamental importance in object extraction as it reduces image data and facilitates object detection. Edges identify object boundaries and are detected through abrupt changes in the gray level above a particular threshold. Edge detectors can exploit operators that are sensitive to the change in gray levels. The methods used in this study are traditional Sobel, Canny, and Laplacian operators. [6] stated that Canny has been the state-of-the-art method for edge detection until today and used the method for detecting abrupt changes in climate events by examining two dimensions: space and time. Also, [6] claimed that the method successfully managed to detect and quantify abrupt changes. However, other methods are widely used because they show good results, for example is Laplacian that becomes the basis for modified edge detection method in [7] by proposing the Laplacian Gradient algorithm for the detection of three-dimensional objects. Another example is [8] by proposing combination of Laplacian of Gaussian, Canny, and mathematical morphology operator for a timely and accurate ground fissures finding in coal mining area. [9] proposed modified Sobel method by generating new mask values resulting to the enhanced computation time. The purpose of this study is to find the best method for the given case.

2 Data and Method 2.1 Data The data used in this research are obtained from Weather Research and Forecast model simulation results. The model produces Air Force Weather Aviation (AFWA) data as derived data from the main simulation results. Each parameter data is stored in an image in png format with a size of 1000 × 400 pixels. An image represents the condition of a parameter in an hour. In particular, this study used Low-Level Wind Shear (LLWS) and Low-Level Turbulence (LLT) data parameters on September 8 and 16, 2022. Since each day is represented by 24 images for each parameter, then

Weather Condition Identification Using Edge Detection Method …

223

Fig. 1 a Example of wind shear image and b example of turbulence image. Both represent information at one hour, and the white is seen as transparent

there are 96 images in total used in this study. These parameters at those specific times were chosen because of the continuous data availability and have a varying range of values. LLWS value is represented in purple color gradation. See Fig. 1, there are five color gradations and RGB coded to represent different range levels. Transparent (0, 0, 0), light purple (201, 0, 255), heavy purple (93, 0, 118), trans purple (51, 0, 64), and dark purple (137, 0, 174) represent value below 1.9, 2–3.9, 4–5.9, 6–7.9, and 8 and above, respectively. Meanwhile, the LLT value is represented in orange color gradation. There are five color gradations to represent different range levels. Transparent (0, 0, 0), peach orange (255, 199, 143), rajjah (255, 182, 104), yellow-orange (255, 165, 63), and dark orange (254, 141, 1) represent value below 0.15, 0.16–0.2, 0.21–0.4, 0.41–0.8, and 0.8 and above, respectively. The idea of performing edge detection in this study is to find objects in form of blob(s) or line(s) in the color of trans purple and dark orange which represent range value 8 and above for LLWS, and 0.8 and above for LLT, respectively. The range necessarily is drawn with the lowest pixel intensity (a shape with the darkest color) within an image. The object with the lowest pixel intensity in this image represents an area whose value belongs to the highest range level. This means that the area is experiencing the most intense aviation weather events, in this case, turbulence events and wind shear events. Blob is a collection of pixels that share the same property that form irregular shapes and line is collection of interconnected pixels whose shape has two ends. This study excludes dot or single pixel because 1 pixel in the image is equivalent to 5 km2 in actual size, so when compared to the actual conditions, this area is considered very small. See Fig. 2 and the following paragraph to understand the comparison. Figure 2 is an example of severe turbulence event reported by National Weather Aviation Service Center (NWASC) on October 26, 2022. The incident occurred over southwestern Ohio near Indianapolis when the small aircraft fly at 3000 feet. The map shown has a scale of 2 cm:100 km. The black circle marked the estimated location of the turbulence and has a diameter of 1.5 cm or equals to 75 km. Thus, the circle has an area of 4,417.86 km2 or 883 pixels in the image used in this study.

224

A. Nafiisyanti et al.

Fig. 2 Example of reported severe turbulence in aviation. source: https://twitter.com/NWSAWC/ status/1585292028178534403?ref_src=twsrc%5Etfw%7Ctwcamp%5Etweetembed%7Ctwterm% 5E1585292028178534403%7Ctwgr%5E2321ce5efa2d9c12b67f788d96649c1db78d747b%7Ct wcon%5Es1_&ref_url=https%3A%2F%2Fwww.newsweek.com%2Fsevere-turbulence-plane-alm ost-flips-over-1755237

Weather Condition Identification Using Edge Detection Method …

225

2.2 Method Sobel operator. Sobel is one of the popular methods among other operators for edge detection. It is based on the gradient/first-order derivative to perform convolution to find an edge by using a horizontal mask or vertical mask [10]. One of the advantages of this method is that it reduces noises in the image and gives brighter and clearer edges [11]. The vertical mask is simply the transposition of the horizontal mask as described in Fig. 3 The convolution process of an image starts from the top to bottom for the vertical mask and left to right for the horizontal mask. It is a multiplication of a pixel intensity value with its neighbors then weighted by the mask. As an illustration, Fig. 4 is the subset of pixels of an image. The gradient of P4 is determined by using Eqs. 1–7. Comparison as the weighing process is then performed to define whether Px4 is an edge pixel or not by using Eq. 3. Px4 is set as an edge pixel if the gradient is greater than the threshold, on the contrary, it is set as a non-edge pixel [12].

Fig. 3 a Horizontal mask and b vertical mask

Fig. 4 3*3 mask of image subset

G x = f1 − f2

(1)

f 1 = (P x6 + 2 × P x7 + P x8 )

(2)

f 2 = (P x0 + 2 × P x1 + P x2 )

(3)

G y = f3 − f4

(4)

226

A. Nafiisyanti et al.

f 1 = (P x2 + 2 × P x5 + P x8 )

(5)

f 1 = (P x0 + 2 × P x3 + P x6 )

(6)

  G = |G x | + G y 

(7)

Canny Operator. The canny operator is another well-known method for edge detection. As well as Sobel, this method’s advantage is that it can reduce noise. Traditionally, there are four stages of the Canny method [13]. The first is image smoothing using the Gaussian function, this step reduces the effect of the noises by using 3*3, 5*5, 7*7 mask, etc. The second step is calculating the gradient to obtain gradient magnitude and gradient direction. The calculation is tended to identify the drastic change in the gray area [14]. The third step is performing Non-Maxima Suppression (NMS). NMS uses 3*3 neighbors to perform interpolation of gradient magnitude along the gradient direction. If the gradient magnitude is greater than the interpolation of the two-gradient direction, then the pixel is selected as an edge point, otherwise, it is marked as a non-edge point. This selection supports the ideal final image where the edge’s line is preferably thin to get a precise and unilateral edge. The last step is double thresholding to detect and connect edges. High and low thresholds are set and compared to the gradient magnitude of the pixel. If the gradient magnitude is greater than the high threshold, then the pixel is marked as an edge point, and if the gradient magnitude is less than the low threshold, the pixel is marked as the non-edge point. As the result of this step is discontinuous, the rest pixel which are connected to the edge point is marked as the edge point. Laplacian operator. The Laplacian operator is based on zero-crossing/second-order derivative. As well as Canny, this method performs better in edge localization so that it gives an accurate location to the edge (Anas et al., 2019). However, it is sensitive to noise due to double derivation. The calculation is denoted in Eq. 8 [15]. ∇ 2 f = f (x + 1, y) + f (x − 1, y) + f (x, y + 1) + f (x, y − 1) − 4 f (x, y) (8) where x and y are the coordinates of the pixel. Laplacian method uses a 3*3 mask as a filter as the following HSV Color Space. Hue, Saturation, and Value (HSV) color space selection is implemented to perform edge detection better. Hue reflects the color that is dominantly observed by humans, saturation is the shade of the color, and value is the intensity or the tone of the color [16]. HSV is obtained from RGB using the following equation: H = arccos 

1 2 (2R 2

− G − B)

(R − G) − (R − B)(G − B)

(9)

Weather Condition Identification Using Edge Detection Method …

S=

max(R, G, B) − min(R, G, B) max(R, G, B) V = max(R, G, B)

227

(10) (11)

This color space also has been used in various image recognition systems [17]. Thus, HSV color space is chosen to fit the idea of choosing the closest channel in which color space portrays shapes with the darkest color.

3 Result and Analysis 3.1 Preprocessing Intuitively, channel V will be chosen since the study worked on pixel intensity, but that was not applicable for both parameters. LLT images are formed by a collection of pixels, the majority of which are orange gradations, while the rest is a few transparent areas (which are then converted to white). As seen in Fig. 1(b), there is less contrast compared to Fig. 1(a). In the channel V calculation result, the edges that define an object are not contrasting. In addition, the V channel captures all the edges in the image, including the edges of objects with non-low-intensity pixels. Meanwhile, the saturation channel provides better contrast for objects with low pixel intensity because it can differentiate the tone of each orange, even though it still captures another edge. Thus, channel S is selected for LLT, whereas channel V is selected for LLWS because the original image has stronger contrast since it has more transparent areas. The saturation channel is incapable to provide better differentiation since it is compared to white (maximum RGB).

3.2 Processing According to Fig. 5(a–f), it is shown that the three methods are capable of detecting the edges of the objects in the images. They are also able to detect the smallest object in dots and lines form, which is supposedly identified as noise. Further explanation is methods performance paragraphs. However, there are some differences described in the following paragraphs. Canny and Laplacian construct interconnected dots to form an edge for a blob. Meanwhile, Sobel constructs a blob with dotted lines as seen in Fig. 5(e) and (f). Besides, the interconnected dots in Canny and Laplacian form thinner lines compared to lines in Sobel. Although the results of Canny and Laplacian look similar, there is a difference that can be seen in Table 1.

228

A. Nafiisyanti et al.

Fig. 5 Implementation results of Sobel, Canny, and Laplacian operator on LLWS and LLT images

Table 1 Number of objects counted on generated edge detection image results

Methods

LLWS

LLT

Sobel

74

40

Canny

54

8

Laplacian

43

7

To confirm the performance of the method, we implement object counting on each image result. The counting process results in a different number of objects in each method. As stated in Table 1, Laplacian collects the fewest number of objects, Canny identifies more objects than Laplacian, whilst Sobel collects the biggest number of objects on both datasets. Although it is stated in [18] that Laplacian is sensitive to noise, Canny appears to detect more objects than Laplacian. Since this study uses the traditional Canny method, the operator detects too detailed edges that causes the method is prone to noise [19], which similarly occurred in [20]. Different from the other two methods, Sobel does not connect the points marked as edges. Moreover, Sobel does not apply edge thinning so that the points are considered as one object. These reasons affect the calculation. Thus, the object counting process failed to recognize the original object.

Weather Condition Identification Using Edge Detection Method …

229

The difference in the number of objects in the image of each method affects the determination of the object’s central coordinate. Each method returns different locations of the objects. Take an example on Canny and Laplacian that give similar results, some coordinates shift slightly from one to another. This may affect the information conveyed on aviation.

4 Conclusion Canny, Laplacian, and Sobel methods generally are capable of detecting edges in the given images. However, Sobel does not apply object thinning on each detected edge point/edge pixel, and each point next to each other is not interconnected, so that each point is considered as an object. Different from Sobel, Laplacian, and Canny form thinner point and create full blobs or lines which perfectly recognized as an object. The three methods are susceptible to noise since they detect detailed edges. In conclusion, traditional Laplacian and Canny methods are capable of fulfilling the requirement. Meanwhile, traditional Sobel needs improvement to meet the requirement.

References 1. Xiang, Y., Ye, Q., Liu, J., Zhang, T., Fan, G., Zhou, P., Lü, L., Liu, Y.: Retrieve of planetary boundary layer height based on image edge detection. Chin. J. Lasers 43 (2016) 2. Dang, R., Yang, Y., Hu, X.M., Wang, Z., Zhang, S.: A review of techniques for diagnosing the atmospheric boundary layer height (ABLH) using aerosol Lidar data, Remote Sens. 11 (2019) 3. Katiyar, S.K., Arun, P.V.: Comparative analysis of common edge detection techniques in context of object extraction, 50, 68 (2014) 4. Higashino, S.I., Teruya, T., Yamada, K.: Position identification using image processing for uav flights in martian atmosphere. J. Robot. Mechatronics 33, 254 (2021) 5. I. N. R. and I. Agency, JATAYU, http://jatayu.brin.go.id/new-jatayu/hello.php 6. Bathiany, S., Hidding, J., Scheffer, M.: Edge detection reveals abrupt and extreme climate events. Am. Meteorol. Soc. 33 (2020) 7. Al-Anssari, J., Naser, I., Ralescu, A.: Three-dimensional laplacian spatial filter of a field of vectors for geometrical edges magnitude and direction detection in point cloud surfaces. In: Proceeding 2019 IEEE international conference on humanized computing and communication HCC 2019, p. 83 (2019) 8. Xu, D., Zhao, Y., Jiang, Y., Zhang, C., Sun, B., X. He, Using improved edge detection method to detect mining-induced ground fissures identified by unmanned aerial vehicle remote sensing. Remote Sens. 13 (2021) 9. Vinista, P., Joe, M.M.: A novel modified sobel algorithm for better edge detection of various images. Int. J. Emerg. Technol. Eng. Res. 7, 25 (2019) 10. Ahmed, A.S.: Comparative study among Sobel, Prewitt and canny edge detection operators used in image processing. J. Theor. Appl. Inf. Technol. 15 19 (2018) 11. Setiadi, D.R.I.M., Jumanto, J.: An enhanced LSB-image steganography using the hybrid cannySobel edge detection. Cybern. Inf. Technol. 18, 74 (2018) 12. Nausheen, N., Seal, A., Khanna, P., Halder, S.: A FPGA based implementation of Sobel edge detection. Microprocess. Microsyst. 56, 84 (2018)

230

A. Nafiisyanti et al.

13. Song, R., Zhang, Z., Liu, H.: Edge connection based canny edge detection algorithm. Pattern Recogn. Image Anal. 27, 740 (2017) 14. Sekehravani, E.A., Babulak, E., Masoodi, M.: Implementing canny edge detection algorithm for noisy image. Bull. Electr. Eng. Inf. 9, 1404 (2020) 15. Bouganssa, I., Sbihi, M., Zaim, M.: Laplacian edge detection algorithm for road signal images and FPGA implementation. Int. J. Mach. Learn. Comput. 9, 57 (2019) 16. Hema, D., Kannan, S.: Interactive color image segmentation using HSV color space. Sci. Technol. J. 7, 37 (2019) 17. Qazanfari, H., Hassanpour, H., Qazanfari, K.: Content-based image retrieval using HSV color space features. Int. J. Comput. Inf. Eng. 13, 537 (2019) 18. Anas, R., Elhadi, H.A., Ali, E.S.: Impact of edge detection algorithms in medical image processing. World Sci. News 118, 130 (2019) 19. Liu, L., Liang, F., Zheng, J., He, D., Huang, J.: Ship infrared image edge detection based on an improved adaptive canny algorithm. Int. J. Distrib. Sens. Netw. 14 (2018) 20. Feng, Y., Zhang, J., Wang, S.: A new edge detection algorithm based on canny idea. In: AIP Conference Proceeding, vol. 1890 (2017)

Numerical Simulation of MCC Evolution Over Borneo Island Using a High-Resolution Model, Case Study: April 14–15, 2012 Ibnu Fathrio and Trismidianto

Abstract The evolution of the mesoscale convective complex (MCC), over Borneo on April 14–15, 2012, was successfully simulated using the Weather Research and Forecasting (WRF) model. This experiment uses ERA5 dataset as the initial and boundary condition on a 3 km spatial resolution. Meanwhile, the Betts–Miller–Janjic (BMJ), The WRF single moment 6 (WSM6), and the Yonsei University (YSU) were then selected as schemes for cumulus, microphysics, and planetary boundary layer, respectively. The model successfully captures the phase, duration, and lowest cloud top temperature of MCC. Despite the model’s inability to simulate comparable spatial patterns of cloud distribution, this simulation shows promising results in simulating MCC evolution. A brief assessment of topography’s role in MCC evolution is also discussed by contrasting normal topography with flat topography simulation.

1 Introduction Since the 1980s, researchers have been studying mesoscale convective complexes (MCC), a long-lived and quasi-circular type of mesoscale convective systems (MCS) [1, 2]. The majority of MCC occurs in the northern hemisphere during the night, which lasts 10 h [3, 3]. The MCC is responsible for up to 18% of total precipitation in the Great Plains in the USA. According to a previous study [5], MCC contributes up to 60% of total rainfall in Subtropical South America. Given the importance of MCC in predicting hazardous weather, improving numerical model skills to predict MCS for weather forecasting and climate prediction is critical. As the climate warms, MCSs may become more frequent and intense [6].

I. Fathrio (B) · Trismidianto National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_22

231

232

I. Fathrio and Trismidianto

[7] has documented MCC observation on the maritime continent. The MCC accounts for up to 20% of total rainfall in Indonesia. The MCC matures more on the continent than on the sea, particularly near mountainous areas. It has characteristics similar to the global population of MCCs. MCC populations are higher in the boreal spring, yet lower in the boreal summer. A previous study by Yulihastin et al. [8] has documented an effort to study the evolution of MCC using regional models on the maritime continent. They emphasized the importance of using up-todate SST when simulating the long-lived MCC with persistent precipitation. Yulihastin et al. [9] successfully simulates the offshore propagation of MCC in southern Sumatra. They emphasized the well-simulated cold pool, which maintains favorable low-temperature conditions for the MCC. Borneo, one of the largest islands in the maritime continents, has the highest contribution of the MCC to rainfall, up to 20% [10]. Borneo Island is home to five major cities as well as Indonesia’s future capital city. The study of MCC that contributes to cause extreme rainfall is critical. Assessment of the regional model’s abilities in simulating the evolution of the MCC over Borneo is required to improve the accuracy of the model prediction of the MCC. Unfortunately, only a few studies have discussed the evaluation of model performance to simulate MCC over Borneo. As a result, the goal of this study is to assess the performance of the regional model in simulating the evolution of MCC over Borneo Island. To evaluate the model, a study case of MCC on April 14–15, 2012, as discussed in [11], was chosen as a reference. In addition, this study emphasizes the role of mountainous regions in the northern part of Borneo in the initiation of the MCC.

2 Data and Method This study focuses on a case of MCC event that occurred on the island of Borneo on April 14–15, 2012. Based on [1, 7] recognized this case as one of the convective systems that meets the MCC characteristics. Readers are encouraged to see [11] for more information on MCC characteristics in the maritime continent. To simulate the evolution of MCC, the Weather Research and Forecast (WRF) model [12] was utilized. The long-lived MCC is characterized using cloud top temperature from the MTSAT satellite’s IR-1 channel [13]. This simulation consists of two nested domains with spatial resolutions of 9 km and 3 km, respectively (Fig. 1). The WRF Single Moment 6 (WSM6) microphysics scheme, the Yonsei University planetary boundary layer scheme, the Dudhia and RRTM schemes for shortwave and longwave radiation, respectively, the MM5 similarity for surface layer models, and the NOAH land surface model are all used in this simulation. The Betts–Miller–Janjic (BMJ) cumulus scheme is used for the first domain (9 km), but no cumulus scheme is used for the second domain (3 km) because convection is assumed to be resolved

Numerical Simulation of MCC Evolution Over Borneo Island Using …

233

Fig. 1 Domain of simulation consists of the first domain (D01) and the second domain (D02) with a spatial resolution of 9 km and 3 km, respectively. The gray-shaded area denotes the topography greater than 500 m in height

explicitly. The European Center for Medium-Range Weather Forecasting (ECMWF) Re-Analysis 5 (ERA5) hourly reanalysis dataset [14] was used for the boundary and initial conditions. The time integration is two days, beginning at 1800 UTC on April 13, 2012, with a 12-h spin-up time. To investigate the influence of topography on the initiation of MCC, the second simulation has been carried out by setting all topography over the continent in the first and second domains to one-meter height. Herein, the flat topography simulation is referred to as FLAT simulation, while the normal topography simulation is referred to as CTRL simulation.

3 Results and Discussion The initiation stage of the MCC occurs on 15–19 LT (GMT + 7), April 14, 2012, when the convective cells are found in the northern, middle, and southern parts of the continent (Fig. 2a-d). A combination of mountain-valley breezes [11, 15] in the northern and middle parts of the continent, which are dominated by mountainous regions, and sea breezes contribute to the initiation of convective activity over the

234

I. Fathrio and Trismidianto

Fig. 2 Contour map of cloud top temperature derived from the observation a–d adapted from [10]. The initiation stage of MCC was on April 14, 2012 for a 1300 LT, b 1500 LT, c 1600 LT, and d 1700 LT. WRF output for cloud top temperature is shown at a similar time for e 1300 LT, f 1500 LT, g 1600 LT, and h 1700LT April 15, 2022, respectively. Cloud top temperature units are in Kelvin

middle and the coast of Borneo, respectively. The model simulates the initiation stage with a less convective cell found in the continent’s middle and southern regions. The MCC reaches a mature stage around midnight, when a broad convective area spreads across the middle, southern, and western parts of the continent (Fig. 3a-d). Convective cells tended to propagate landward to the middle of the continent. At the same time, expanding convective cells with cloud top temperatures of less than 200 K can be found in the southern part. The WRF model simulates the mature stages of MCC, showing the intensification and merging of convective cells in the continent’s western and middle regions (Fig. 3f-h). In comparison with the observation, the model also reasonably simulates the minimum cloud top temperature. However, there was not a single major convective cell simulated by the model in the south. The merging and weakening of convective cells in the western and middle parts of the continent are shown in the decaying stage (Fig. 4a-d), along with increased divergence of surface wind. Simultaneously, the land breeze promotes a squall line in the continent’s south that propagates southward. This squall line takes six hours to reach the north Coast of Java Island. The model successfully simulates the decaying stages of MCC in the western part of the continent with surface wind divergence (Fig. 4e-h). The southern squall line has not yet been identified, but the land breeze is already present.

Numerical Simulation of MCC Evolution Over Borneo Island Using …

235

Fig. 3 Contour map of cloud top temperature derived from the observation a–d adapted from [10]. The growth stage of MCC is shown at a 2100 LT, while the mature stage is shown at b 2200 LT on April 14, 2022; c 0000 LT, and d 0100 LT on April 15, 2022. WRF output for cloud top temperature is shown at a similar time for e 2100 LT, and f 2200 LT on April 14, 2022; g 0000 LT, and h 0100 LT on April 15, 2022, respectively. Cloud top temperature units are in Kelvin

A few hours later, as the sea breeze returns, new convective cells form along the north and south coasts and in the middle of the continent (Fig. 5a-d). Meanwhile, the squall line in the south of the continent hits Java Island in the afternoon and lasts for 5–6 h before decaying. The model simulates the post-MCC stage and the squall-line propagation reasonably well (Fig. 5e-h). However, models simulate the presence of the squall line 4–5 h later than the actual squall line.

3.1 Influence of Topography in the Initial Stage of MCC Figure 6 depicts the differences between CTRL and FLAT simulations, where topography enhances convective activity during the initiation and growth stages. Figure 6a– c shows developing convective cells to the west of the mountainous region in CTRL simulation (highlighted by the orange box in Fig. 6b), whereas less convective cells are visible in FLAT simulation. Figure 6c–e depicts how convective cells grow to the south of the mountainous region and then propagate south-westward (highlighted

236

I. Fathrio and Trismidianto

Fig. 4 Contour map of cloud top temperature derived from the observation a–d adapted from [10]. The decaying stage of MCC was shown at a 0300 LT, b 0500 LT, c 0700 LT, while the dissipation stage of MCC was shown at d 0900 LT on April 15, 2022. WRF output for cloud top temperature is shown at a similar time for e 0300 LT, f 0500 LT, g 0700 LT, and h 0900LT April 15, 2022, respectively. Cloud top temperature units are in Kelvin

by green boxes in Fig. 6). Following that, the convective cells merge with other convective cells growing in the continent’s west to form larger convective cells. FLAT simulation, on the other hand, shows fewer convective cells growing in mountainous regions to the south. On the west coast of the continent, the FLAT simulation generates convective cells earlier than the CTRL simulation, as demonstrated by Tan et al. [16] in a no-topography experiment. This experiment supports previous findings that most continental MCC prefers high-elevation areas, which contribute to the generation of MCC in the middle of the continent [15].

4 Conclusion Overall, the WRF model could simulate the evolution of MCC occurred on April 14–15, 2021, across the Borneo continent, from initiation to dissipation and postMCC stages. Although there are differences in spatial distribution and the WRF model simulates smaller cell size than the observed MCC, the evolution phase and

Numerical Simulation of MCC Evolution Over Borneo Island Using …

237

Fig. 5 Contour map of cloud top temperature derived from the observation a–d adapted from [10]. The post-MCC stages are shown at a1300 LT, b 1500 LT, c 1700 LT, and d 1900 LT on April 15, 2022. WRF output for cloud top temperature is shown at a similar time for e 1300 LT, f 1500 LT, g 1700 LT, and h 1900 LT on April 15, 2022, respectively. Cloud top temperature units are in Kelvin

minimum cloud top temperature are successfully simulated. Some characteristics, such as the role of land-sea breeze, the influence of topography in the initiation stage, convective cell merging, and the decaying stage denoted by surface wind divergence, are simulated correctly. The model also simulates the post-MCC stage, such as an induced squall line that moves toward Java Island. Furthermore, flat topography simulation was performed to confirm the role of the mountainous region, which gives rise to more convective cells, thereby increasing the strength and coverage of MCC over Borneo. This study allows researchers to delve deeper into the underlying mechanism of MCC life, which can last for several hours across Borneo. In the future, it will be interesting to see if this result is affected by the parameterization and initial-boundary condition used.

238

I. Fathrio and Trismidianto

Fig. 6 Contour of simulated cloud top temperature of 210 K for CTRL simulation (red) and FLAT simulation (blue) at a 1900 LT, b 2000 LT, c 2100 LT, d 2200 LT on April 14, 2002, e 0000 and f 0200 LT on April 15, 2022. Orange circles highlight the presence of convective cells in CTRL. The gray-shaded area denotes the topography greater than 500 m in height. Orange and green boxes highlight the differences between CTRL and FLAT experiments

References 1. Maddox, R.A.: Mesoscale convective complexes Bull. Amer. Meteor. Soc. 61, 1374–1387 (1980) 2. Houze, R.A.: Structure and dynamics of a tropical squall line system. Mon. Wea. Rev. 105, 1540–1567 (1977) 3. Fritsch, J.M., Kane, R.J., Chelius, C.R.: The contribution of mesoscale convective weather system to the warm season precipitation in the United States. J. Clim. Appl. Meteorol. 25, 1333–1345 (1986) 4. Laing, A.G.: Contribution of mesoscale convective complexes to rainfall in Sahelian Africa: estimates from Geostationary Infrared and Passive Microwave Data. J. Appl. Meteorol. 38, 957–964 (1999) 5. Durkee J.D., Mote T.L., Shepherd, M.J.: The contribution of mesoscale convective complexes to rainfall across, subtropical South America. Int. J. Clim. 22 (2009) 6. Schumacher, R.S., Rasmussen, K.L.: The formation, character and changing nature of mesoscale convective systems. Nat. Rev. Earth Environ. 1, 300–314 (2020) 7. Trismidianto.:The global population of mesoscale convective complexes (MCCs) over Indonesian maritime continent during 15 Years. IOP Conf. Ser.: Earth Environ. Sci. 166, 012040 (2018)

Numerical Simulation of MCC Evolution Over Borneo Island Using …

239

8. Yulihastin, E., Nuryanto, D.E., Muharsyah, R.: Improvement of heavy rainfall simulated with SST adjustment associated with mesoscale convective complexes related to severe flash flood in Luwu, Sulawesi, Indonesia. Atmosphere 12(11), 1445 (2021) 9. Yulihastin, E., Fathrio, I., Nauval, F., Saufina, E., Harjupa, W., Satiadi, D., Nuryanto, D.E.: Convective cold pool associated with offshore propagation of convection system over the East Coast of Southern Sumatra, Indonesia. Adv. Meteorol. (2021) 10. Trismidianto, Yulihastin, E., Satyawardhana, H., Nugroho, J.T., Ishida, S.: The contribution of the mesoscale convective complexes (MCCs) to total rainfall over Indonesian maritime continent. IOP Conf. Ser. Earth Environ. Sci. 54, 012027, (2017) 11. Trismidianto.: Characteristics of the oceanic MCC, continental MCC, and coastal MCC over the Indonesian maritime continent. IOP Conf. Ser. Earth Environ. Sci. 149, 012024 (2018) 12. Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Liu, Z., Berner, J., Wang, W., et al.: A description of the advanced research WRF model version 4. National Center for Atmospheric Research: Boulder, CO, USA, 145, (2019) 13. Trismidianto, H.T.W., Ishida, S., Moteki, Q., Manda, A., Iizuka, S.: Development processes of oceanic convective systems inducing the heavy rainfall over the western coast of Sumatra on 28 October 2007. Sola 12, 6–11 (2016) 14. Hersbach, H., Bell, B., Berrisford, P., Biavati, G., Horányi, A., Muñoz Sabater, J., Nicolas, J., Peubey, C., Radu, R., Rozum, I., Schepers, D., Simmons, A., Soci, C., Dee, D., Thépaut, J.-N.: ERA5 hourly data on single levels from 1959 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS) (2018). Accessed on 01–07–2022 15. Trismidianto, Yulihastin, E., Satyawardhana, H., Ishida, S.: A composite analysis of the mesoscale convective complexes (MCCs) development over the Central Borneo and its relation with the propagation of the rainfall systems. IOP Conf. Ser. Earth Environ. Sci. 54(1), 012036 (2017). IOP Publishing 16. Tan, H., Ray, P., Barrett, B., Dudhia, J., Moncrieff, M., Zhang, L., Zermeno-Diaz, D.: Understanding the role of topography on the diurnal cycle of precipitation in the maritime continent during MJO propagation. Clim. Dyn. 58, 3003–3019 (2022)

Tropopause Height Variation Toward Different Land-Sea Convection Activities in Java Using GNSS-RO Data Khanifah Afifi

and Nurjanna Joko Trilaksono

Abstract The exchange of substances like ozone and water vapor is facilitated by the tropopause layer, which is the layer between the stratosphere and troposphere. Convective activity is one of the factors that affects the tropopause’s height. The variation in tropopause height may be influenced by the distinct differences in diurnal convective activity between land and sea regions. The purpose of this research is to investigate diurnal variations in the height of the tropopause layer are affected by differences in convective activity between land and sea. The study area is Java Island and the sea that surrounds it (6°–9° S and 105°–115° E), and the study period runs from December 1, 2019, to November 30, 2020. The Constellation Observing System for Meteorology Ionosphere and Climate (COSMIC-2), Global Navigation Satellite System Radio Occultation (GNSS-RO) data, and Lapse Rate Tropopause methods are utilized to determine the height of the tropopause layer. In the meantime, the temperature black body gradient was used to estimate the convective activity. The findings demonstrated that, following the rise in tropopause height, the increase in convective activity over the ocean was more dominant in the late afternoon and early morning, while it was more dominant over land from noon until the afternoon in local time. The pattern of cloud formation in the ocean area is most closely matched by the tropopause height during March to May (MAM) season, whereas in the land area during December to February (DJF) season.

K. Afifi Master Program in Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia N. J. Trilaksono (B) Atmospheric Sciences Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_23

241

242

K. Afifi and N. J. Trilaksono

1 Introduction The Indonesian Maritime Continent (MC) has the main characteristic of diurnal variation of convective activity [1]. The existence of contrasting differences between diurnal convective activity in the land and sea areas is thought to influence the variation in tropopause height. Previous research showed that the diurnal variation in tropopause height was not significant enough in the Surabaya area [2]. However, that study used radiosonde data that had poor temporal resolution because radiosonde measurements were only performed at 00.00 and 12.00 UTC, so it was considered insufficient to study the diurnal variability of the tropopause. In addition, that study did not separate land and sea convection. The separation of land and sea convection is very important because there are contrasting differences in their diurnal variations [3]. Global Navigation Satellite System Radio Occultation (GNSS-RO) is a remote sensing technique that uses GNSS signals received by Low Earth Orbit (LEO) satellites by utilizing atmospheric refractivity. The GNSS signal will be deflected by the atmosphere before being captured by the LEO satellite and forming a deflection angle. The bending angle will produce an atmospheric refractivity profile. GNSS-RO provides data in global coverage and has a vertical resolution of 100 m [4]. RO data is of the highest quality at altitudes between 8 and 35 km and would be very useful for tropopause studies [5]. In addition, GNSS-RO data is also capable of providing data over the ocean area, which radiosondes and monitoring stations cannot provide. Although many studies have investigated tropopause height variations in the tropics over long timescales such as annual [6], seasonal [7], and monthly [8], relatively few studies have identified tropopause height variations on diurnal timescales. Several studies have also been conducted to determine the effect of strong convection on tropical tropopause variations. However, the effect of differences in convective activity between land and sea on variations in tropopause height still needs to be studied further. In this study, the identification of variations in the height of the tropopause layer and its relation to differences in convective activity between land and sea will be developed, especially on the island of Java using diurnal GNSS-RO data.

2 Data and Method In this study, GNSS-RO data provided by the Constellation Observing System for Meteorology Ionosphere and Climate (COSMIC-2) has a vertical resolution of 100 m and a spatial resolution of about 150 km [3]. COSMIC-2 is the result of cooperation between the USA and Taiwan which was launched on June 25, 2019. COSMIC-2 data can produce 5000 bending angle profiles and high-resolution refractivity per day in the tropics and subtropical [9]. The data taken is temperature and altitude from the wetPf2 atmospheric profile data. In addition, black body temperature (TBB)

Tropopause Height Variation Toward Different Land-Sea Convection …

243

data from the IR1 channel of the Himawari-8 satellite is used to obtain the cloud top temperature which will then be used as a proxy in identifying convective activity. The Himawari-8 satellite is a geostationary satellite operated by the Japan Meteorological Agency (JMA). Himawari-8 was launched in October 2014 and is located at 140° E to observe eastern Asia and western Pacific regions. Himawari-8 satellite data was used in this research from Kochi University (http://weather.is.kochiu.ac.jp/archive-e. html). This data has a horizontal resolution of 0.05° × 0.05° and an hourly temporal with Portable Gray Map (PGM) format. The data period used is December 1, 2019– November 30, 2020, with the study area located at 6°–9° S and 105°–115° E. The method used in this study includes the separation of data between land and sea using the Python global-land-mask module which has a resolution of 1 km. The latitude/longitude point of the data is checked whether it is on land or at sea and then grouped into land and ocean data before further processing. The height of the tropopause was identified using the minimum Lapse Rate Tropopause (LRT) method based on the definition from the World Meteorological Organization (WMO). WMO defined tropopause as the height at which the lapse rate decreases to 2° C/km or less, provided that within 2 km above this point, the average lapse rate does not exceed 2° C/km [10]. The convective activity was obtained from the average seasonal TBB gradient. The TBB gradient is the difference between the TBB value of a certain hour and the TBB value of the previous hour [11]. Both tropopause height variations and convective activity patterns were grouped into eight local time bins with 3-h intervals (7–10, 10–13, 13–16, 16–19, 19–22, 22–1, 1–4, and 4–7 local time (UTC + 7)).

3 Result 3.1 Seasonal Tropopause Altitude Diurnal Pattern The diurnal pattern of tropopause height identified by the minimum LRT method for each season is presented in Fig. 1 (sea) and Fig. 2 (land). There is a diurnal variation of tropopause height in both sea and land areas. In general, the height of the tropopause in the sea area has a peak in the afternoon until the evening and early in the morning. Meanwhile, in the land, the peak is in the afternoon until late in the evening and late at night. The maximum tropopause height above the ocean can be observed in the morning of each season. In December to February (DJF), March to May (MAM), and June to August (JJA) seasons, the maximum tropopause height occurs at 07.00–10.00 LT, while in September to November (SON) season the maximum height occurs at 04.00– 07.00 LT. Toward noon the height of the tropopause above the ocean in all seasons decreases and begins to increase in the afternoon and into the evening. Meanwhile, the maximum tropopause height above the land can be observed at 10.00–13.00 LT during the DJF season, at 07.00–10.00 LT during the MAM and JJA seasons, and at 19.00–22.00 LT during the SON season. During the day in the land, the height of the

244

K. Afifi and N. J. Trilaksono

Fig. 1 Boxplot of the minimum LRT height above the ocean during the a DJF, b MAM, c JJA, and d SON seasons for the period December 1, 2019–November 30, 2020. The orange line is the median value, and the green triangle is the average value

tropopause tends to increase in almost all seasons except the JJA season. This is in accordance with research conducted by Suneeth [12]. Tables 1 and 2 show the variation in tropopause height in the ocean and land areas calculated using the Interquartile Range (IQR). The tropopause height in the ocean area ranges from 15 to 18 km with the highest variation of tropopause height occurring in the DJF season at 10.00–13.00 LT at 1.24 km and the lowest in the JJA season at 19.00–22.00 at 0.59 km. Meanwhile, in the land area, the tropopause altitude ranges from 14.7 to 18.1 km with the highest variation of tropopause height in the DJF season at 19.00–22.00 LT at 1.29 km and the lowest in the JJA season at 13.00–16.00 LT at 0.84 km. In general, the tropopause height variation is greater over land than over sea. This is in accordance with research conducted by Gettelman et al. [13].

Tropopause Height Variation Toward Different Land-Sea Convection …

245

Fig. 2 Same as Fig. 1 but for minimum LRT height above the land Table 1 IQR values of the average minimum LRT height in the sea area Season DJF MAM JJA SON

07-10 0.97 0.84 0.43 0.45

10-13 1.24 0.89 0.55 0.42

13-16 1.05 0.42 0.39 0.63

Local me 16-19 19-22 0.78 1.17 0.76 0.81 0.51 0.59 0.57 0.66

22-01 0.32 0.69 0.52 0.50

01-04 0.74 0.48 0.35 0.38

04-07 0.61 0.60 0.42 0.50

Table 2 IQR values of the average minimum LRT height in the land areas Season DJF MAM JJA SON

07-10 0.87 1.03 0.48 0.27

10-13 0.94 0.46 0.35 0.45

13-16 1.00 0.46 0.84 0.37

Local me 16-19 19-22 0.79 1.29 0.35 0.51 0.47 0.30 0.52 0.50

22-01 0.60 0.74 0.32 0.35

01-04 0.94 0.64 0.40 0.67

04-07 0.50 1.09 0.20 0.89

246

K. Afifi and N. J. Trilaksono

(a)

(b)

Fig. 3 Spatial plot of the TBB variable of the Himawari-8 satellite IR1 channel (cloud top temperature) at a January 27, 2020, at 09.00 LT and b March 21, 2020, at 09.30 LT. The blue dot is the location of the occultation profile

3.2 Effect of Convective Activity on Tropopause Altitude The minimum tropopause height (14.7 km) was identified on March 21, 2020, at 09.30 LT, and the maximum height (18.1 km) was identified on January 27, 2020, at 09.00 LT in the land areas. Figure 3 shows the spatial pattern of cloud top temperatures at both times. In the land area on January 27 at 09.00 LT, there were convective clouds (deep convection events) characterized by TBB < 240 K [14]. While on March 21, 2020, at 09.30 LT, there were no convective clouds (clear sky conditions) characterized by TBB > 240 K. This indicates that at the time the tropopause was at maximum altitude, there are clouds with strong convection. In this study, the average seasonal gradient of TBB for 24 h was used to identify the convective activity that occurred. The convective activity identified as a cloud growth process is characterized by a negative gradient. Cloud dissipation is indicated by a positive gradient value. The value of the TBB seasonal mean gradient for 24 h for land and sea areas has a clear difference between afternoon and evening. Based on Fig. 4, during the DJF season in the land area from the afternoon to evening the TBB gradient shows a negative value which indicates the occurrence of the cloud formation process. The peak of the negative TBB gradient occurred at 13.00–14.00 LT. At night, the TBB gradient shows a positive value associated with cloud dissipation which is thought to produce convective rain. Meanwhile, in the DJF season in the ocean area at 13.00–19.00, 20.00–21.00, and 23.00–00.00 LT, the process of cloud formation occurs, and in the early morning, cloud thinning occurs. In the MAM transition season in the land, it is 11.00–18. 00 LT occurring in the process of cloud formation with a peak at 13.00–14.00 LT, and at night until sunrise, there is cloud thinning. While in the ocean area at 14.00–18.00, 19.00–21.00, and 23.00–00.00 LT, the process of cloud formation occurs, and in the morning, there is cloud thinning. In the JJA season, the TBB gradient value is not very volatile and tends to be around the value 0, indicating the least convective activity, both growth and cloud dissipation. This is probably because the JJA season in Java is experiencing a dry season, so the atmospheric conditions tend to be cloudless, sunny, and rarely rainy. Meanwhile,

Tropopause Height Variation Toward Different Land-Sea Convection …

247

Fig. 4 Seasonal mean TBB gradient values between land and sea areas a DJF season, b MAM, c JJA, d SON. The orange line represents the ocean area, and the blue line represents the land area

during the SON transition season, both on land and in the ocean, there is a very high fluctuation of data that indicates the cloud growth process followed by cloud dissipation.

4 Discussion In general, the diurnal variation of convective activity over the ocean in the form of cloud formation is more dominant in the late afternoon and early morning, while on land, cloud formation is more dominant in the noon until the afternoon local time, in sync with the diurnal variation in tropopause height. The diurnal pattern of TBB gradients in both DJF and MAM seasons shows that above the land, cloud formation is more dominant in the afternoon to evening local time. This can happen because land and sea have different responses to the reception of solar radiation. On the land, surface heating due to the reception of solar radiation takes place faster than in the ocean so that on land convective activity reaches its peak during the day, which in turn causes convective rain above the land to occur in the afternoon until evening [11]. Nitta and Sekine [1] stated that over the continents and large islands, convection reaches its maximum intensity in the late afternoon to evening, probably due to strong

248

K. Afifi and N. J. Trilaksono

surface heating throughout the day. Local convective activity can cause an increase in the intensity of turbulence in the upper troposphere [15]; therefore, variations in the height of the tropopause may also be influenced by turbulence.

5 Conclusion The effect of differences in convective activity between land and sea on variations in the height of the tropopause in Java during the four seasons: DJF, MAM, JJA, and SON has been identified using GNSS-RO data. There is a synchronous pattern of convective activity to variations in tropopause height in both sea and land areas. Convective activity over the seas around the island of Java is more dominant in the late afternoon and early in the morning, while over the land of Java, cloud formation is more dominant in the noon until the afternoon local time following the increase in tropopause height. In addition, it is known that seasonally, the height of the tropopause in the seas of Java Island that most closely matches the pattern of cloud formation is the MAM season, while on the land it is the DJF season. Acknowledgements The authors wish to thank University for Corporation of Atmospheric Research (UCAR) for providing the GNSS-RO data and Kochi University for providing the Himawari-8 data. The first author was supported by the Marine Technology Cooperation Research Center (MTCRC).

References 1. Nitta, T., Sekine, S.: Diurnal variation of convective activity over the tropical western pacific. J. Meteorol. Soc. Jpn. 72(5) (1994) 2. Prayuda, S.: Analysis of the Lapse Rate Tropopause (LRT) Characteristics and its relation to MJO activities in Surabaya. In: Proceedings of the STMKG Earth and Atmospheric (2018) 3. Johnston, B.R, Xie, F., Liu, C.: The effects of deep convection on regional temperature structure in the tropical upper troposphere and lower stratosphere. J. Geophys. Res. Atmos. 123(3) (2018) 4. Kursinski, E.R, Hajj, G.A, Schofield, J.T, Linfield, R.P, Hardy, K.R.: Observing Earth’s atmosphere with radio occultation measurements using the global positioning system. J. Geophys. Res. Atmos. 102(19) (1997) 5. Foelsche, U., Borsche, M., Steiner, A.K, Gobiet, A., Pirscher, B., Kirchengast, G., Wickert, J., Schmidt, T.: Observing upper troposphere-lower stratosphere climate with radio occultation data from the CHAMP satellite. Clim. Dyn. 31(1) (2008) 6. Seidel, D.J., Ross, R.J., Angell, J.K., Reid, G.C.: Climatological characteristics of the tropical tropopause as revealed by radiosondes. J. Geophys. Res. Atmos. 106(D8) (2001) 7. Liu, Y., Xu, T., Liu, J.: Characteristics of the seasonal variation of the global tropopause revealed by COSMIC/GPS data. Adv. Space Res. 54(11). (2014) 8. Rieckh, T.: Tropopause characteristics from GPS radio ocultation data. Master Thesis. Wegener Center for Climate and Global Change. University of Graz (2013) 9. Schreiner, W. S., Weiss, J. P., Anthes, R. A., Braun, J., Chu, V., Fong, J., Hunt, D., Kuo, Y. H., Meehan, T., Serafino, W., Sjoberg, J., Sokolovskiy, S., Talaat, E., Wee, T. K., & Zeng, Z. COSMIC-2 Radio occultation constellation: first results. Geophys. Res. Lett. 47(4) (2020)

Tropopause Height Variation Toward Different Land-Sea Convection …

249

10. WMO, M.: A three-dimensional science. WMO Bull. 6, 134–138 (1957) 11. Kusumawati, Y., Effendy, S., Aldrian, E. Spatial and temporal variations of convective rain on Java Island based on satellite imagery. Agromet J. 22(1) (2008) 12. Suneeth, K.V., Das, S.S, Das, S.K.: Diurnal variability of the global tropical tropopause: results inferred from COSMIC observations. Clim. Dyn. 49(9–10) (2017) 13. Gettelman, A., Salby, M.L, Sassi, F.: Distribution and influence of convection in the tropical tropopause region. J. Geophys. Res. Atmos. 107(9–10) (2002) 14. Mehta, S. K., Ratnam, M. V., Murthy, B. V. K. Variability of the tropical tropopause over Indian monsoon region. J. Geophys. Res. Atmos. 115(14) (2010) 15. Fujiwara, M., Yamamoto, M.K., Hashiguchi, H., Horinouchi, T., Fukao, S.: Turbulence at the tropopause due to breaking kelvin waves observed by the equatorial atmosphere radar. Geophys. Res. Lett. 30(4) (2003)

Sensitivity Analysis of Constructed Analogue Statistical Downscaling Method for Extreme Rainfall Prediction Trinah Wati , Tri Wahyu Hadi , Faiz Rohman Fajary , Ardhasena Sopaheluwakan , and Lambok M. Hutasoit

Abstract Sensitivity analysis of a statistical downscaling (SD) method is to assess a method’s shortcomings, evaluate the reliability, compare the uncertainty, and finetune to see which parameter of the model can be changed. This study aims to analyse the sensitivity of constructed analogue (CA) SD method by evaluating the reliability of gaining extreme information and comparing the uncertainties of CA experiment schemes. There are four schemes, including multilinear regression (MLR), weight mean (WMEAN), MLR with principal component analysis/PCA (PC-MLR), and WMEAN with PCA (PC-WMEAN). The comparison of the schemes was based on predictors’ uncertainty analysis, distribution, and extreme value using several metrics such as Verification Rank Histogram (VRH), Probability Density FunctionSkill Score (PDF-SS), and Brier Score (BS). The predictor is the reanalysis dataset ERA-20C, while the predictand is rainfall from 11 observation rain gauge stations in Java Island, with leave one-year out cross-validation from 1970 to 2010. The results of VRH show the WMEAN and PC-WMEAN schemes have under-dispersed ensemble members. Meanwhile, the MLR shows positive bias, and PC-MLR shows negative bias. The MLR scheme has the highest score on PDF-SS. Although underestimated and overestimated, the PDF plot shapes of MLR and PC-MLR schemes are closer to observation than WMEAN and PC-WMEAN. The MLR and PC-MLR schemes are more capable of estimating daily and pentad rainfall of 50 and 100 mm than the WMEAN and PC-WMEAN schemes. This study implicates the development of the CA SD method for predicting/estimating extreme rain.

T. Wati (B) · T. W. Hadi · F. R. Fajary · L. M. Hutasoit Graduate Program of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung 40132, Indonesia e-mail: [email protected] T. Wati · A. Sopaheluwakan Indonesia Agency for Meteorology, Climatology, and Geophysics, Jakarta 10610, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_24

251

252

T. Wati et al.

1 Introduction Statistical downscaling (SD) is one of the techniques to generate high-resolution regional climate models based on large-scale information from either reanalysis data or global climate models (GCM). The SD method with an analogue approach was previously developed for weather prediction [1, 2, 3]. Other studies [4, 5] used the method to gain local climate statistics consistent with synoptic-scale atmospheric conditions of GCM. The application of the analogue approach requires a long-term database to cover weather patterns or atmospheric conditions that may appear in future. The CA method has an advantage in the analogy of a local scale climate pattern with a synoptic-scale climate pattern, so that the spatial structure of the local climate is well presented in the simulation results. The method minimizes biases and uses the absolute value of the predictand in the process of prediction or estimation. The capability of the CA SD method depends on the similarity between the patterns defined in the analogue selection [4]. Previous studies [6, 7] used the minimum Euclidean distance metric to calculate the similarity degree of weather patterns from predictors; another is cosine similarity [8, 9]. However, the shortcomings of the CA SD method were found, i.e. increasing spatial coherence in the result field, decreasing the final result’s temporal variation caused by averaging process, and too much drizzle (light rain) production [10]. This study aims to analyse the sensitivity of the CA SD method by evaluating the reliability of gaining extreme information and comparing the uncertainties of CA experiment schemes.

2 Data and Method The predictor data employ ERA-20C’s 3-hourly averaged to daily zonal (U) and meridional (V ) wind parameters at 850 hPa from 1970 to 2010. We derive U and V wind parameters as vectors to scalar values of stream function (ψ) or Psi and velocity potential (χ ) or Phi using a decomposition of Helmholtz. The study uses a multi-window approach [8, 9], getting 14 ensemble members from seven monsoon windows to obtain the probabilistic rainfall prediction or estimation. Meanwhile, the predictand data employ daily rainfall of 11 stations in Java Island with the same periods. The location of monsoon windows and the stations are visualized in Fig. 1a and b. There are two stages in CA: diagnosis and prognosis. Firstly, the diagnosis stage selects a subset of historical weather patterns of GCM predictors in training periods that are similar to the target periods. The selection of analogue uses the cosine → similarity/S(u) method [15] to calculate the degree of similarity between vectors − a − → (u) in training periods and a (t) in target periods using the equation: − → → a (u).− a (t)  − S(u) = − →  a (u)→ a (t)

(1)

Sensitivity Analysis of Constructed Analogue Statistical Downscaling …

253

Fig. 1 Location of monsoon windows [8] (a) and observation stations, (b) The monsoon windows are (1) Maritime Continent [11], (2) Australian Monsoon [12], (3) Webster and Young Monsoon [13], (4, 5) Indian Summer Monsoon [14], and (6, 7) Western North Pacific Summer Monsoon [14]

The subset number assigned is 30 subsets producing a constructed analogue and determining the most suitable multilinear combination [5, 8]. The application of EOF or PCA reduces the atmospheric circulation field’s degree of freedom from predictors [4, 9]. The formulas for PCA are as follows: F(u) =

M 

c(u)k a(u)k

(2)

c(t)k a(t)k

(3)

k=1

F(t) =

M  k=1

where F(u) and F(t) are predictors and c and a are spatial and temporal components in training and target periods. M is the number of significant components (PCs), usually between 5 and 10. The PC scores are normalized to give equal weight to the variance described by all PCs. The combination of 30 analogue subsets is applied in the prognosis stage using the predictand data to predict or estimate the predictand variable in the target time. The formula of analogue construction can be MLR [9], WMEAN [16], or other regression methods. The formulas of analogue construction in this study are as follows: RCA (t) =

30 

Bn Rn (t) + C for MLR

(4)

n=1

RCA (t) =

30 

Wn Rn for WMEAN

(5)

n=1

where RCA (t) is the CA simulation rainfall at the target periods (t) and n is the sum of the best analogue subsets at the diagnosis stage. B is the regression coefficient obtained between the predictors from the diagnosis stage for MLR, and W is the

254

T. Wati et al.

weighting between RMSE and the correlation coefficient from the diagnosis stage for WMEAN. There are four CA schemes of the sensitivity analysis, i.e. MLR, WMEAN, MLR with principal component analysis (PC-MLR), and WMEAN with PCA (PCWMEAN). We use cross-validation leave one-year out for the CA SD running. The comparison of the four schemes is based on predictor uncertainty, distribution, and extreme values analysis. Verification Rank Histogram (VRH) is a dispersion error measure of ensemble predictions representing observational uncertainty at a daily timescale. The shape of VRH that is uniform or flatter indicates the ensemble predictions close to the observations [17]. Distribution analysis uses the PDF Skill Score (SS) metric [18–20]; the formula is as follows: SS =

nb 

min( f mk , f ok )

(6)

k=1

where f mk and f ok are relative frequencies of rain days (days with precipitation > 0.5 mm). A value in the kth bin belongs to the histograms of the CA SD result and observation, whereas Nb is the bin number of the empirical PDF calculation. The result of CA SD perfectly simulates the observed PDF with a maximum value of 1. The verification for extreme values uses the Brier Score (BS) of 50 and 100 mm rainfall events daily and 100 mm on pentad data (five days accumulation). The formula is as follows: BS =

N 1  ( f n − O n )2 N n=1

(7)

where N is the number of sample data and f n is the probability of CA SD results of daily and pentad rainfall events (50 and 100 mm). The formulation uses binary possibilities, and On is an observation (value 1 for events that occur and 0 for nonoccurrence). BS has the best skill at zero and the worst at one.

3 Result and Discussion The resulting period of the leave one-year out cross-validation approach is 1970 to 2010 at a daily timescale. Figure 2 shows VRH diagrams of four CA SD schemes at Banyuwangi Station in East Java. The VRH of the MLR schematic in Fig. 2a shows a positive bias of ensemble members [17]. Meanwhile, the VRH of WMEAN and PCWMEAN schemes look similar and show under-dispersive or less spread of ensemble members. In contrast to the MLR scheme, the PC-MLR scheme exhibits a negative bias. All stations have a similar pattern of VRH, with the sample at Banyuwangi Station as a representation.

Sensitivity Analysis of Constructed Analogue Statistical Downscaling …

255

Fig. 2 Diagrams of VRH at Banyuwangi Station for CA SD scheme of MLR (a), WMEAN (b), PC-MLR (c), and PC-WMEAN (d)

The distribution analysis of the four schemes, as seen in Fig. 3a–d, compares the empirical PDF of rain days at Semarang Climate Station. It can be seen that the PDF graph of the WMEAN (b) and PC-WMEAN (d) schemes are different from the observations, while the PDF MLR (a) and PC-MLR (c) schemes are closer, but the PC-MLR scheme is very overestimated. The SS quantification measure for all stations presented in Table 1 shows the MLR scheme has the highest average of 0.80, ranging between 0.68 and 0.87. The lowest SS is the PC-MLR scheme with an average of 0.51, ranging from 0.31 to 0.83. Figure 3 is the representation for all stations that have similar PDF pattern. The extreme value analysis uses the BS metric for 50 mm and 100 mm daily and 100 mm on pentad data presented in Figs. 4, 5 and 6. We consider analysing the pentad data due to the accumulation of rainfall for several days associated with flood events. The BS for 50 mm daily rainfall in Fig. 4 shows almost the same skills for all schemes, except for the MLR scheme, which has more representative stations with lower skills (blue). The BS for 100 mm daily rainfall in Fig. 5a–d shows a more significant number of stations in the MLR and PC-MLR schemes with better skills than in the WMEAN and PC-WMEAN schemes, while the BS for 100 mm pentad rain in Fig. 6 shows that the PC-WMEAN scheme has the lowest skill compared to other schemes.

256

T. Wati et al.

Fig. 3 Graphs of rain day empirical PDF at Semarang Climate Station for MLR (a), WMEAN (b), PC-MLR (c), and PC-WMEAN (d) Table 1 SS of four CA SD schemes

CA SD Schemes

PDF skill score (SS) Mean

Min

Max

MLR

0.80

0.68

0.87

WMEAN

0.65

0.55

0.73

PC-MLR

0.51

0.31

0.83

PC-WMEAN

0.65

0.56

0.74

Fig. 4 BS of 50 mm daily rain for MLR (a), WMEAN (b), PC-MLR (c), and PC-WMEAN (d)

Sensitivity Analysis of Constructed Analogue Statistical Downscaling …

257

Fig. 5 BS of 100 mm daily rain for MLR (a), WMEAN (b), PC-MLR (c), and PC-WMEAN (d)

Fig. 6 BS of 100 mm pentad rain for MLR (a), WMEAN (b), PC-MLR (c), and PC-WMEAN (d)

4 Summary Sensitivity analysis of four CA SD schemes in this study assesses uncertainty (VRH), distribution (SS), and extreme values (BS). The MLR and PC-MLR schemes have better distribution representation and a more significant number of stations with better extreme representatives than the WMEAN and PC-WMEAN schemes. The WMEAN and PC-WMEAN schemes based on the VRH show low dispersion, indicating less represented extreme values. PCA treatment did not affect the WMEAN scheme as seen in the same VRH pattern on WMEAN and PC-WMEAN schemes. On the contrary, in the MLR scheme, which has a positive bias VRH pattern, the PCA treatment gives more effect, resulting in the VRH pattern becoming negatively biased. It should be considered a calibration treatment to the result of CA SD for obtaining a

258

T. Wati et al.

more extreme value representation and closer to observed data. This study implicates the development of the CA SD method in estimating or predicting extreme rainfall.

References 1. Lorenz, E.N.: Atmospheric predictability as revealed by naturally occurring analogues. J. Atmos. Sci. 26, 636–646 (1969) 2. Van den Dool, H.M.: Searching for analogues, how long must we wait? Tellus A 46(3), 314–324 (1994) 3. Zorita, E., Hughes, J.P., Lettemaier, D.P., Von Storch, H.: Stochastic characterization of regional circulation patterns for climate model diagnosis and estimation of local precipitation. J. Clim. 8(5), 1023–1042 (1995) 4. Zorita, E., Von Storch, H.: The analog method as a simple statistical downscaling technique: comparison with more complicated methods. J. Clim. 12(8), 2474–2489 (1999) 5. Hidalgo, H.G., Dettinger, M.D., Cayan, D.R.: Downscaling with constructed analogues: daily precipitation and temperature fields over The United States (2008) 6. Timbal, B., McAvaney, B.J.: An analogue-based method to downscale surface air temperature: application for Australia. Clim. Dyn. 17(12), 947–963 (2001) 7. Anuchaivong, P., Sukawat, D., Luadsong, A.: Statistical downscaling for rainfall forecasts using modified constructed analog method in Thailand. Sci. World J. (2017) 8. Syahputra, M.R.: Implementation on constructed analogue on seasonal rainfall forecast from CFS (Climate Forecast System) output; case study Island of Java and Sumatra. Master Thesis. Institut Teknologi Bandung. (2013). In Bahasa 9. Surmaini, E., Hadi, T.W., Subagyono, K., Puspito, N.T.: Prediction of drought impact on rice paddies in west java using analogue downscaling method. Indonesia J. Agric. Sci. 16(1), 21–30 (2015) 10. Pierce, D.W., Cayan, D.R., Thrasher, B.L.: Statistical downscaling using localized constructed analogues (LOCA). J. Hydrometeorol. 15(6), 2558–2585 (2014) 11. Robertson, A.W., Moron, V., Qian, J., Chang, C., Tangang, F., Aldrian, E., Koh, T.Y., Liew, J.: The maritime continent monsoon. The global monsoon system: research and forecast, 85–98 (2011) 12. Kajikawa, Y., Wang, B., Yang, J.: A multi-time scale Australian monsoon index. Int. J. Climatol. 30(8), 1114–1120 (2010) 13. Webster, P.J., Yang, S.: Monsoon and ENSO: selectively interactive systems. Q. J. R. Meteorol. Soc. 118, 877–926 (1992) 14. Wang, B., Wu, R., Lau, K.M.: Interannual variability of the Asian summer monsoon: contrasts between the Indian and the western North Pacific-East Asian monsoons. J. Clim. 14(20), 4073–4090 (2001) 15. Garcia, E.: Cosine Similarity Tutorial. https://www.researchgate.net/publication/327248753 (2015) 16. Fernández, J., Sáenz, J.: Improved field reconstruction with the analog method: searching the CCA space. Climate Res. 24, 199–213 (2003) 17. Hamill, T.M.: Interpretation of rank histograms for verifying ensemble forecasts. Mon. Weather Rev. 129(3), 550–560 (2001) 18. Perkins, S.E., Pitman, A.J., Holbrook, N.J., McAneney, J.: Evaluation of the AR4 climate models’simulated daily maximum temperature, minimum temperature, and precipitation over Australia using probability density functions. J. Clim. 20(17), 4356–4376 (2007) 19. Chhin, R., Yoden, S.: Ranking CMIP5 GCMs for model ensemble selection on regional scale: case study of the Indochina region. J. Geophys Res. Atmos. 123(17), 8949–8974 (2018) 20. Wati, T., Hadi, T.W., Sopaheluwakan, A., Hutasoit, L.M.: Statistics of the performance of gridded precipitation datasets in Indonesia. Adv. Meteorol. (2022)

Meteorological Factors Influencing Coastal Flooding in Semarang, Central Java, Indonesia, on 23 May 2022 Teguh Harjana, Eddy Hermawan, Risyanto, Anis Purwaningsih, Dita Fatria Andarini, Ainur Ridho, Dian Nur Ratri, and Akas Pinaringan Sujalu Abstract It has been well established that coastal flooding is caused by heavy rains, storm surges, and high tidal waves that potentially lead to tremendous damage. Using brightness temperature (TBB) data from the Himawari satellite and ocean surface wind data from ERA5, we have investigated meteorological factors influencing strong wind and increased seawater level over the northern coast of Semarang which caused the coastal flood on 23 May 2022. Our results indicated that there are two potential meteorological factors that contribute to the coastal flooding during that period. Firstly, the formation of MCS over the ocean on the northern coast of Semarang led to heavy rain over the coast and strong-surface wind speed that potentially enhances ocean tidal waves toward Semarang. Secondly, the unusually strong-surface easterly winds cause the increase in seawater level through the Ekman pumping mechanism. As for the latter, the strong easterly wind stress led to strong Ekman transport to the south of the flow (i.e., to the northern coast of Java), causing a large net transport of seawater toward this region as a result of a balance between Coriolis and turbulent (wind) drag forces. Our results provided a new perspective on the factors influencing the increased seawater level intruding into Semarang during the coastal flooding on 23 May 2022.

T. Harjana (B) · E. Hermawan · Risyanto · A. Purwaningsih · D. F. Andarini Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung 40173, Indonesia e-mail: [email protected] A. Ridho Cerdas Antisipasi Risiko Bencana Indonesia (CARI), Bandung 40293, Indonesia D. N. Ratri Meteorological, Climatological, and Geophysical Agency (BMKG), Jakarta 10720, Indonesia Droevendaalsesteeg, Wageningen University and Research, 6708 Wageningen, The Netherlands A. P. Sujalu Universitas, 17 Agustus 1945 Samarinda, Samarinda 75123, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_25

259

260

T. Harjana et al.

1 Introduction Semarang is one of the big cities in Indonesia and is located on the northern coast of Java Island. Semarang city is an urban area and has a specific problem related to the coastal environment, and it has been reported that the coastal area is a subside relative to the mean sea level [1]. The condition that the topography is under the sea level rise, due to the effects of land subsidence, makes Semarang vulnerable to coastal floods [2–8]. On 23 May 2022, Semarang experienced a tidal flood disaster that caused the breaking of the coastal embankment in the Tanjung Emas Area. The Semarang City Disaster Mitigation Agency (BPBD) on Tuesday announced that the breaking of the coastal embankment causes seawater to flood the land with inundations of 40 cm to 1.5 m. Moreover, Tanjung Emas Maritime Meteorological Station Semarang BMKG recorded the height of the tide before the embankment broke at 15.00 WIB (08Z) reaching 2 m [9]. It is also recorded by BPBD that the coastal flooding had affected 1255 households [10]. The Semarang tidal flood during this day occurred in the phase of the moon perbani (neap tides), so the influence on the tides was not as high as in the phase of the full moon (spring tides) [11]. The existence of a storm that occurred at 02.00 WIB early in the morning in the Java Sea and the presence of persistent winds since a few days earlier are suspected to be one of the triggers for the Semarang tidal flood [11]. In this paper, the factors that affect strong winds and sea level increases that caused coastal floods on 23 May 2022, on the northern coast of Semarang, will be investigated using satellite and reanalysis data. The detail of the data and method that are employed for this study is written in Sect. 2. Then, the results and the discussion are described in Sect. 3. Furthermore, this is followed by the conclusions and future works in Sect. 4.

2 Data and Methods The fifth-generation reanalysis data from The European Center for Medium-Range Weather Forecasting (ECMWF ERA5) [12] was used in this study. Data used from ERA5 is Mean Sea Level Pressure (MSLP), u and v components of wind at 10 m, 925 hPa wind, and Sea Surface Temperature (SST). The Global Satellite Mapping of Precipitation (GSMaP) gauge-corrected version-7 standard precipitation product was used to investigate the presence of rain that occurs in the Java Sea [13]. The spatial resolution of the GSMaP data is 0.01 × 0.01 degrees, and the temporal resolution is 1 h (https://sharaku.eorc.jaxa.jp). In addition, the presence of the Mesoscale Convective System (MCS) was identified using Temperature Black Body (TBB) from Band-13 (IR) of Himawari-8 satellite, using the “Grab ‘em, Tag ‘em, Graph ‘em” (GTG) algorithm [14]. Criteria for the MCS phase are described in Table 1.

Meteorological Factors Influencing Coastal Flooding in Semarang … Table 1 Criteria for MCS phase

261

Physical characteristics

Criteria

BT(10.4)

243 K

Size

10,000 km2

Duration

Size and temperature definition must be met for a period of 3 h

Initiation

Size and temperature definition are first satisfied

Termination

Size and temperature definition are no longer satisfied

Mature

Minimum mean of cloud temperature definition must be met

The GTG method has been widely used in analyzing the role of MCS in Indonesia, for example, in the case of heavy rains in Jakarta in January 2013 [15], Semarang [16], and New Capital City Nusantara of Indonesia [17]. In addition, Ekman transport and pumping velocity were also calculated which were derived from wind stress at an altitude of 10 m. Ekman transport (EMT, m3 s−1 m−1 ) is calculated using the formula used by Dieng AL [18]. EMTx =

τy τx , EMT y = − ρw f ρw f

(1)

where ρ w is the water density (1025 kg m−3 ), f is the Coriolis factor (1/s), and the wind stress τ (Pa) is calculated using τx = ρa · C D ·



u 2 + v 2 · u, τ y = ρa · C D ·



u 2 + v2 · v

(2)

where ρ a is the density of air (1.25 kg m−3 ), C D is the coefficient of drag, and u, v are components of wind at 10 m above sea level. Ekman pumping velocity (EPV, ms−1 ) is calculated using the curl of wind stress with the equation: EPV =

1 ∇ ×τ ρw f

(3)

3 Result and Discussion A breaking seawater embankment of Tanjung Emas Semarang caused a coastal flood event of up to 1.5 m on 23 May 2022 at 08Z (15 WIB). Figure 1b denotes the evolution of sea level height at Tanjung Emas Semarang from 22 May 2022 at 00Z to 24 May

262

T. Harjana et al.

Fig. 1 Figures a and b show the geographical location of the northern coast of Java, and c shows a time series of sea wave heights at the time of the tidal wave on 23 May 2022

2022 at 00Z. It is worth noting that the high seawater level occurred at around 05Z– 11Z on 22 and 23 May 2022 with a height between 180 and 210 cm. Specifically, the seawater level significantly increased since 23 May 2022 at 05Z and reached a peak of approximately 210 cm at 08Z and 09Z. This maximum sea level height affected the collapse of the seawater embankment and hence tidal flooding in many areas over Semarang and its surroundings. In addition, the hourly GSMaP precipitation data also shows heavy rainfall occurred in the Java Sea, next to the north of Central Java, from 22 May 2022 at 20Z till 23 May 2022 at 05Z (the GSMaP rain map is not shown here).

3.1 The MCS Evolution Ahead of Coastal Flooding To understand the meteorological factors that modulate the increased seawater level during the coastal flooding in Semarang, we analyzed the evolution of MCS before and during the tidal flood event (see Fig. 2). A convective cloud began to develop on 22 May 2022 at 19Z in the coastal area of Central Java and was identified as the pre-MCS phase (Fig. 2a). This cloud grew rapidly (around 1 h), which showed the onset of the MCS initiation phase on 22 May 2022 at 20Z. At the same time, a rainfall of 5 mm/h occurred in the MCS areas. Then, the MCS developed to the growth stage, where the size of clouds increased further and became two cloud cells

Meteorological Factors Influencing Coastal Flooding in Semarang …

263

Fig. 2 Evolution of MCS (shaded) overlaid with GSMaP rain (yellow contour) in phases: a preMCS, b initiation, c growth, d mature, e dissipation, and f decay

to the north (Java Sea) and south area (the coastal and mainland) on 23 May 2022 at 00Z. Those MCSs propagated to the north and merged after five hours, which created a single deep MCS in the ocean. As a result, this mature stage of MCS produced heavy rainfall of 10 mm/hour over the coastal area of Central Java and the Java Sea. Thereafter, the MCS began to dissipate gradually until it reached the decay stage, which was followed by a decrease in rainfall intensity. Overall, the evolution of MCS in the north of Semarang and surrounding areas beginning with the pre-MCS and ending with the decay stage during 12 h induced the increased storm surge over the northern coast of Java Island and hence enhanced the seawater mass due to the heavy precipitation.

3.2 Evolution of Convection and Near-Surface Easterly Wind Figure 3 shows the evolution of convection (represented by cloud top temperature with TBB colder than 255 K), vertical velocity, and near-surface wind in each stage of MCS development as described in the previous section. A negative value of vertical velocity was identified in the coastal areas of Central Java since the pre-MCS stage. Interestingly, the highest magnitude of negative vertical velocity occurred over the ocean in the north of Semarang during the growth stage of MCS (23 May 2022 at 00Z), which is associated with the intense downward motion. It also can be seen that persistent easterly wind was observed along the Java Sea before and during the tidal flooding event, with the magnitude varying between 7 ms−1 and 15 ms−1 . The near-surface easterly wind modulated the easterly wind

264

T. Harjana et al.

Fig. 3 Evolution of TBB (shaded), 925 hPa wind (vector), and vertical velocity (contour) associated for each phase of MCS

stress, and it induced the Ekman transport to the south areas of the northern coast of Java Island. The details of Ekman transport will be furtherly discussed in the next section.

3.3 Ekman Mass Transport (EMT) and Ekman Pumping Velocity (EPV) Figure 4 shows the evolution of Ekman transport (EMT, vector) and Ekman pumping velocity (EPV, shaded) at 06Z (13 WIB) from 12 May 2022 to 23 May 2022. As seen in Fig. 4, on 12 May 2022, EMT is heading north, and on 14 May 2022, EMT is heading south. Several days before the tidal wave, this EMT consistently headed south toward the northern coast of Java Island which caused seawater to accumulate in the north of the island of Java. Ekman transport EMT on 23 May 2022 is much larger than the previous days. Likewise, the EPV on 23 May 2022 had the lowest negative value compared to the previous days. A negative EPV indicates downwelling, and a positive EPV indicates upwelling. The EPV time series every 3 h from 10 May 2022 to 24 May 2022 is shown in Fig. 5.

Meteorological Factors Influencing Coastal Flooding in Semarang …

265

Fig. 4 Evolution of daily EPV (shaded) and EMT (vector) from 12 May 2022 at 06Z to 23 May 2022 at 06Z

Fig. 5 Time series of EPV from 10 May 2022 to 24 May 2022 for the part of the northern coast area (108.5–110.5 E; 6.5–7 S)

266

T. Harjana et al.

4 Conclusion Using TBB data from Himawari satellite and ocean surface wind data from ERA5, we have investigated meteorological factors influencing strong wind and increased seawater level over the northern coast of Semarang that caused the coastal flood on 23 May 2022. Our results suggested two potential meteorological factors that play an important role: (1) the formation of MCS over the ocean on the northern coast of Semarang that led to heavy rain over the coast and strong-surface wind speed that potentially enhances tidal waves toward Semarang Island and (2) unusually strongsurface easterly winds leading to increase seawater level through the Ekman pumping mechanism as a result of a balance between Coriolis and turbulent (wind) drag forces. Although our results did not yet quantify how much such factors contributed to the total increase in seawater level as a whole, our results really provided a new perspective on the factors influencing the increased seawater level intruding into Semarang during the coastal flooding on 23 May 2022, in addition to the previously known perspective of astronomical (tidal) forcing. Acknowledgements We would like to thank Dr. Sandro W. Lubis from the Pacific Northwest National Laboratory (PNNL), who has given a lot of time for discussion, input, and improvement so that this paper can be completed. This paper was written with financial support from the DIPA of the Aeronautics and Space Organization, National Research and Innovation Agency (BRIN), in 2022. Author Contribution All authors are the main contributors, have read the manuscript, and declare no conflict of interest.

References 1. Sutanta, H.: Spatial modeling of the impact of land subsidence and sea level rise in a coastal urban setting, case study: Semarang, Central Java, Indonesia. M.Sc. thesis, International Institute for Geo-Information and Earth Observation, ITC, Enschede, The Netherlands (2002) 2. Irawan, A.M., Marfai, M.A., Munawar, I.R., Nugraheni, S., Gustono T., Hasti., Rejeki, A., Widodo, A., Rikha, R., Mahmudiah., Faridatunnisa, M.: Comparison between averaged and localised subsidence measurements for coastal floods projection in 2050 Semarang, Indonesia. Urban Clim. 35, 100760 (2021). https://doi.org/10.1016/j.uclim.2020.100760 3. Marfai, M.A., dan King, L.: Potential vulnerability implications of coastal inundation due to sea level rise for the coastal zone of Semarang City, Indonesia. Environ. Geol. (2007). https:// doi.org/10.1007/s00254-007-0906-4 4. Al Dianty, M., Arbaningrum, R., Putuhena, F.J.: The linkage of effect climate change for determining design flood of Tenggang River. Geogr. Tech. (2020). https://doi.org/10.21163/ gt_2020.151.17 5. Koch, M.: Intergrading earth & disaster science to enable sustainable adaptation & mitigation. In: Engineering, Information and Agricultural Technology in the Global Digital Revolution (2020). https://doi.org/10.1201/9780429322235-2 6. Kurniawati, W., Mussadun, Nugraha, M.F.: Spatial expression of Malay Kampung Semarang in facing flood disaster. In: IOP Conference Series: Earth and Environmental Science (2020). https://doi.org/10.1088/1755-1315/409/1/012049

Meteorological Factors Influencing Coastal Flooding in Semarang …

267

7. van de Haterd, J., Budiyono, B., Darundiati, Y.H., Spaan, E.: Environmental change and health risks in coastal Semarang, Indonesia: importance of local indigenous knowledge for strengthening adaptation policies. Cities Heal. (2020). https://doi.org/10.1080/23748834.2020. 1729451 8. Verrest, H., Groennebaek, L., Ghiselli, A., Berganton, M.: Keeping the business going: SMEs and urban floods in Asian megacities. Int. Dev. Plan. Rev. (2020). https://doi.org/10.3828/idpr. 2020.3 9. BMKG Homepage, https://pasut.maritimsemarang.com. Accessed 2022/08/13 10. Tempo Homepage, https://en.tempo.co/read/1594634/coastal-flooding-hits-semarang-thousa ndsof-families-affected. Accessed 2022/05/24 11. CNN Homepage, https://www.cnnindonesia.com/teknologi/20220602182348-199-804098/4penyebab-banjir-rob-semarang-siklus-bulan-hingga-badai. Accessed 2022/07/20 12. Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., et al.: The ERA5 global reanalysis. Q. J. R. Meteorol. Soc. 146, 1999–2049 (2020). https://doi.org/10.1002/qj.3803 13. Kubota, T., Shige, S., Hashizume, H., et al.: Global precipitation map using satellite-borne microwave radiometers by the GSMaP project: production and validation. IEEE Trans. Geosci. Remote Sens. 45(7), 2259–2275 (2007). https://doi.org/10.1109/TGRS.2007.895337 14. Whitehall, K., Mattmann, C.A., Jenkins, G., Rwebangira, M., Demoz, B., Waliser, D., Kim, J., Goodale, C., Hart, A., Ramirez, P., et al.: Exploring a graph theory based algorithm for automated identification and characterization of large mesoscale convective systems in satellite datasets. Earth Sci. Inform. 8, 663–675 (2015). https://doi.org/10.1007/s12145-014-0181-3 15. Nuryanto, D.E., Pawitan, H., Hidayat, R., Aldrian, E.: The occurrence of the typical mesoscale convective system with a flood-producing storm in the wet season over the Greater Jakarta area. Dyn. Atmos. Ocean. 96, 101246 (2021). https://doi.org/10.1016/j.dynatmoce.2021.101246 16. Hermawan, E., Lubis, S.W., Harjana, T., Purwaningsih, A., Risyanto., Ridho, A., Andarini, D.F., Ratri, D.N., Widyaningsih.R.: Large-scale meteorological drivers of the extreme precipitation event and devastating floods of early-February 2021 in Semarang, Central Java, Indonesia. Atmosphere (2022). https://doi.org/10.3390/atmos13071092 17. Purwaningsih, A., Lubis, S.W., Hermawan, E., Andarini, D.F., Harjana, T., Ratri, D.N., Ridho, A., Risyanto, Sujalu, A.P.: Moisture origin and transport for extreme precipitation over Indonesia’s New Capital City, Nusantara in August 2021. Atmosphere 13, 1391 (2022). https://doi.org/10.3390/atmos13091391 18. Dieng, A.L., Ndoye, S., Jenkin, G.S., Sail, S.M., Gaye, A.T.: Estimating zonal Ekman transport along coastal Senegal during passage of Hurricane Fred, 30–31 August 2015. Springer Nature Appl. Sci. 3, 588 (2021). https://doi.org/10.1007/s42452-021-04578-5

Assessment of Predictability of Convective-Induced Turbulence Event Using High-Resolution Model, Case Study: Hong Kong Airlines Incident on 6 May 2016 Ibnu Fathrio, Aisya Nafiisyanti, Ina Juaeni, Muhammad Arif Munandar, Dita Fatria, Anis Purwaningsih, Fadli Nauval, Alfan Sukmana Praja, Elfira Saufina, Didi Satiadi, Teguh Harjana, Wendi Harjupa, and Risyanto Abstract This study investigates the predictability of the mesoscale convective system (MCS) that caused severe turbulence on Hong Kong Airlines flight in the south of Banjarmasin on 6 May 2016 at 1830 UTC. This study is carried out by using The Weather Research and Forecasting (WRF) model by using high spatial resolution and the Global Forecast System (GFS) dataset as the primary forcing for boundary and initial conditions. The result shows that the turbulence could be well predicted by the model that is initiated on 6 May 2016 at 0600 UTC or 18 h before the incident happened. Despite differences in spatial distribution and timing of precipitation compared to the observation, this study also highlights the importance of using a high spatial resolution of 1 km to capture the convective-induced turbulence successfully.

1 Introduction Hong Kong commercial aircraft was reported to experience severe turbulences on 6 May 2016 [1]. The plane departs from Denpasar Airport with an intended destination of Hong Kong Airport. KNKT reports [1] documented that severe turbulence hit the aircraft at 1830 UTC above the Java Sea, near Southern Kalimantan coastline. The weather radar displayed a magenta color in about 9 km distance indicating the severe intense area of rainfall and turbulence associated with the Cumulonimbus cloud. This incident injured 3 flight attendants and 11 passengers, which made the aircraft return to Bali Airport. According to KNKT reports, the aircraft used weather I. Fathrio (B) · A. Nafiisyanti · I. Juaeni · D. Fatria · A. Purwaningsih · F. Nauval · A. S. Praja · E. Saufina · D. Satiadi · T. Harjana · W. Harjupa · Risyanto National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected] M. A. Munandar Indonesian Agency for Meteorological, Climatological, and Geophysics, Jakarta, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_26

269

270

I. Fathrio et al.

Fig. 1 Prognosis chart of flight level (FL) 250–630 valid for 6 May 2016 at 1800 UTC issued by WAFC London. This figure is obtained from Fig. 5 on KNKT report [1]

information issued by World Area Forecast Center (WAFC) London. The prognosis chart displayed a potential cumulonimbus cloud development area that unfortunately excludes the location where the aircraft experienced severe turbulence (Fig. 1). This issue could be related to the model skill in predicting convective systems in the maritime continent. The spatial resolution of the model is an important factor in simulating convective systems in the tropics. A previous study [2] and [3] highlighted that 1 km horizontal resolution is required to properly affect the mesoscale convective system (MCS) in the maritime continent. Meanwhile, based on WAFC Management Report on March 2017–February 2018 obtained from [4], the WAFC still used global models with a spatial resolution of 17 km before 2017, which is too coarse to resolve the convective system. Therefore, higher spatial resolution models shown could complement the WAFC models [5]. Therefore, this study aims to evaluate the model’s skill in predicting the Hong Kong turbulence incident by utilizing higher spatial resolution regional models up to 1 km. In addition, we also consider which initiation time of model predictions better simulates the incident in terms of timing and amplitude.

Assessment of Predictability of Convective-Induced Turbulence Event …

271

2 Data and Method In this study, Weather Research Forecasting (WRF) [6] regional model was used to investigate the turbulence incident at 1830 UTC on 6 May 2016 in high spatial resolution. This study refers to [2] for selecting parameterization schemes, which has successfully simulates the evolution of MCS in the Southern Sumatra. In all domains, the model uses WRF Single Moment 6 (WSM6) as the microphysics scheme and Yonsei University as the planetary boundary layer scheme, and in all domains, Dudhia and RRTM schemes as shortwave and longwave radiation scheme, respectively, MM5 similarity as surface layer models, and NOAH land surface model as land surface scheme. The use of Betts-Miller-Janjic (BMJ) for the cumulus scheme in coarse resolution displays a promising result, as previously shown in [7]. Therefore, all domains utilize BMJ as cumulus scheme, except the innermost domain. The cumulus scheme on the innermost domains was set to the explicit (no cumulus scheme used). The GFS products with a 0.25-degree spatial resolution are used as the initial and boundary conditions of simulation obtained from [8]. The product is available four times a day at 0000 UTC, 0600 UTC, 1200 UTC, and 18,000 UTC. All GFS input has a time lag of about 6 h to be available for download. Assuming the first 12 h is used as a spin-up time of simulation, and considering the incident happened at 1830 UTC, the latest GFS input that could be used is GFS initiated on 6 May 2016, at 0600 UTC. This data is available online around 1200 UTC, which is about six and a half hours before the turbulence happened. This study carried out a sensitivity test study by utilizing three different initial times of GFS input: 5 May 2016 at 1800 UTC, 6 May 2016 at 0000 UTC, and 6 May 2016 at 0600 UTC. Meanwhile, the sensitivity test on spatial resolution experiment was investigated for 1 km and 3 km (Fig. 2). To evaluate the 1 km domain (D03), the nesting domain consists of two parent domains with a spatial resolution of 9 km (D01) and 3 km (D02). In the case of 3 km, it only consists of one parent domain with a 9 km spatial resolution (D01). Table 1 displays the name andthe details of the six experiments. In addition, GSMaP precipitation data [9] and IR1 channel #13 (10.41 µm) of the Himawari-8 satellite were used as precipitation reference and proxy of cloud top temperature [1], respectively.

3 Results and Discussion The evolution of the two mesoscale convective systems (MCS) is shown in Fig. 3. The first MCS, located to the east of the turbulence location, starts to grow at 1000 UTC (Fig. 3c), and the second MCS develops later at 1600 UTC to the west of turbulence location. The second MCS was primarily responsible for inducing turbulence hitting the aircraft at 1800 UTC (Fig. 3g) when all MCS reached mature stages. Eventually, the MCS dissipates after 1800 UTC (Figs. 3h and 3i). Figure 4 shows hourly precipitation derived from the GSMaP precipitation dataset, implying the turbulence

272

I. Fathrio et al.

Fig. 2 WRF simulation domain. Details of the simulation scenario are explained in Table 1. For sensitivity test on spatial resolution, RUN 1, RUN 2, and RUN 3 simulation were run on three domains: D01, D02, and D03, while RUN 4, 5, and 6 simulations were run on two domains D01 and D02. The black square indicates the estimated location where the aircraft experiences turbulence based on KNKT reports

Table 1 WRF simulation scenario Simulation name

GFS initial time

Spatial resolution of innermost domain and examined domain number

RUN I

1800 UTC, 5 May 2016

1km (D03)

RUN 2

0000 UTC, 6 May 2016

1km (D03)

RUN 3

0600 UTC, 6 May 2016

1km (D03)

RUN 4

1800 UTC, 5 May 2016

3km (D02)

RUN 5

0000 UTC, 6 May 2016

3km (D02)

RUN 6

0600 UTC, 6 May 2016

3km (D02)

incidence was related to the MCS that propagates southward. Furthermore, the incident happened near the area where the highest precipitation occurs (Fig. 4g, j) as the MCS reached the mature phase. It is expected that model that could properly predict the phase and spatial distribution of precipitation could also predict the turbulence accurately. The skill of models in predicting the precipitation was examined, as shown in Fig. 5. Generally, all experiments show southward propagation of the rains. Yet, they simulate the early timing of precipitation peak, and the precipitation ends earlier than that of the observation. They also artificially generate the rainfall north of 3S, as

Assessment of Predictability of Convective-Induced Turbulence Event …

273

Fig. 3 Contour of black body temperature obtained from TBB-13 channel of Himawari-8 satellite on. The black box indicates the location of turbulence

Fig. 4 Precipitation contour derived from GSMaP displayed in a–i 2D distribution and j latitudetime contour of precipitation averaged over 114° E to 115° E. The black square and X marker indicate the turbulence event location. Precipitation units are in mm hour−1

274

I. Fathrio et al.

shown in RUN 1, RUN 4, and RUN 5. The cause of the early timing of precipitation peak on RUN 1 and RUN 4 is unknown. However, they start the simulation earlier with less data assimilated than in RUN 2 which may degrade the accuracy. On the other hand, RUN 3 and RUN 6 underestimate the precipitation amplitude all the time in the north of 3.8S. The simulated precipitation also presents at 1500 UTC which is later than the other experiment. This could be related to shorter pre-conditioning periods of the model to simulate the MCS. Of all experiments, RUN 2 shows better results with the closest timing of precipitation peak to the observations. Following previous analysis by Kim and Chun [10], this study examines the turbulence by estimating the gradient Richardson number and Brunt–Väisälä frequency. RUN 2 also successfully simulates the location of turbulence at a flight level of 410 (~12.5 km height) that is closest to the actual event both in timing, altitude, and spatial location, as can be seen in Fig. 6 (f-j). Greater turbulence kinetic energy is collocated with a strong gradient Richardson number less than 0.25 s−1 indicating severe turbulence and associated with the coolest top cloud temperature implying active convective activity. These features are not shown in RUN 1 and RUN 3. This suggests that the selection of GFS initiation time is crucial.

Fig. 5 Latitude-time contour of precipitation averaged over 114° E to 115° E for a RUN 1, b RUN 2, c RUN 3, d RUN 4, e RUN 5, and f RUN 6 experiments. The X marker indicates the turbulence incident

Assessment of Predictability of Convective-Induced Turbulence Event …

275

Fig. 6 Contour at ~ 12 km height or 410 flight level for cloud top temperature of 210 K (blue), Richardson number of 0.25 (purple), and turbulent kinetic energy (shaded) with yellow and orange colors that denote the value of 5 m2 s−2 and 10 m2 s−2 . The top panels a–e, middle panels f–j, and bottom panels k–o are for RUN 1, RUN 2, and RUN 3, respectively

Figure 7 shows that 3 km-simulation: RUN 4, RUN 5, and RUN 6 also fail to simulate the timing of turbulence. This implies that higher spatial resolution of 1 km could increase prediction skill in the precipitation phase. Although RUN 5 uses a similar GFS initiation time (on 6 May at 0000 UTC) as RUN 2, it predicts the turbulence event early on 1400 UTC. The turbulence happened at the top of the growing convective cloud, as shown in Fig. 8a–b, around 12 km height. The negative value of Brunt–Väisälä frequency indicates an unstable condition that contributes to a small gradient Richardson number. The cloud dissipates at 1830 UTC, yet the turbulence remains with comparable strength.

4 Conclusion This study has shown that the WRF model could properly predict the incident of turbulence on Hong Kong Airlines at 1830 UTC on 6 May 2016 with an appropriate selection of GFS initiation time and running in higher spatial resolution of 1 km with three nested domain strategies. The simulated precipitation shows better timing and spatial distribution of rainfall related to the MCS activity. Moreover, this study recommends a pre-conditioning period of the model about 18 h before the peak of

276

I. Fathrio et al.

Fig. 7 Same as Fig. 6 but for RUN 4 a–e, RUN 5 f–j, and RUN 6 k–o

Fig. 8 Same as Fig. 6 but for RUN 2 on a 1800 UTC and c 1830 UTC. Latitude-height cross-section along the solid red line shown in a and c for cloud water mixing ratio of 0.05 g kg−1 (gray), cloud ice mixing ratio of 0.2 g kg−1 (blue), gradient Richardson number (black) of 0.25, and negative value of Brunt–Väisälä (s−1 ; magenta-filled) at b 1800 UTC and d 1830 UTC

Assessment of Predictability of Convective-Induced Turbulence Event …

277

the MCS. The reason for the better prediction of GFS on 6 May 2016 at 0000 UTC (RUN 2) than GFS on 5 May 2016 at 1800 UTC (RUN 1 and RUN 4) could be related to more assimilated data being ingested by the former one.

References 1. KNKT: Aircraft Serious Accident Investigation Report. KNKT.14.05.14.04 Preliminary Report. 1–4 (2016) 2. Yulihastin, E., Fathrio, I., Nauval, F., Saufina, E., Harjupa, W., Satiadi, D., Nuryanto, D.E.: Convective cold pool associated with offshore propagation of convection system over the East Coast of Southern Sumatra, Indonesia. Adv. Meteorol. (2021) 3. Yulihastin, E., Nuryanto, D.E., Muharsyah, R.: Improvement of heavy rainfall simulated with SST adjustment associated with mesoscale convective complexes related to severe flash flood in Luwu, Sulawesi Indonesia. Atmosphere 12(11), 1445 (2021) 4. www.icao.int, last accessed 2022/06/22 5. Rais, A.F., Putra, R.M., Fitrianto, M.A., Hermansyah, T.: A preliminary comparative study on the feasibility of a multipurpose numerical weather model for prediction of cumulonimbus Clouds in Indonesia. In: 2022 International Conference on Science and Technology, pp. 1–6. IEEE (2022) 6. Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Liu, Z., Berner, J., Wang, W., et al.: A description of the advanced research WRF model version 4, National Center for Atmospheric Research: Boulder, CO, USA, 145, (2019) 7. Fonseca, R.M., Zhang, T., Yong, K.-T.: Improved simulation of precipitation in the tropics using a modified BMJ in the WRF model. Geosci. Model Develop. 8(9), 2915–2928 (2015) 8. National Centers for Environmental Prediction,: NCEP GFS 0.25 degree global forecast grids historical archive. NCAR Research Data Archive, accessed 11 June 202, (2015) 9. Kubota, T., Shige, S., Hashizume, H., et al.: Global precipitation map using satellite-borne microwave radiometers by the GSMaP project: production and validation. IEEE Trans. Geosci. Remote Sens. 45(7), 2259–2275 (2007) 10. Kim, J.H., Chun, H.Y.: A numerical simulation of convectively induced turbulence above deep convection. J. Appl. Meteorol. Climatol. 51(6), 1180–1200 (2012)

The Effects of the Cross Equatorial Northerly Surge (CENS) on Atmospheric Convection and Convergence Over Jakarta and the Surrounding Area Didi Satiadi, Anis Purwaningsih, Wendi Harjupa, Trismidianto, Dita Fatria Andarini, Fadli Nauval, Elfira Saufina, Fahmi Rahmatia, Ridho Pratama, Teguh Harjana, Risyanto, Ibnu Fathrio, Eddy Hermawan, Mutia Yollanda, and Dodi Devianto Abstract The effect of the Cross Equatorial Northerly Surge (CENS) on the diurnal cycle of convection and moisture convergence over Jakarta and the surrounding area has been investigated. The data used in this study was the CENS indices, the Convective Available Potential Energy (CAPE), the Convective Inhibition Energy (CINH), the Vertically Integrated Moisture Fraction Convergence (VIMFC), and dew point temperature and rainfall from the fifth generation of the European Center for Medium-Range Weather Forecasting (ECMWF) Atmospheric Reanalysis (ERA5) during 2014. The data was depicted in a graph against the diurnal as well as the seasonal cycle, thus illustrating the multi-scale variability. The results showed a very strong influence of the CENS phenomena on the diurnal cycle of the CAPE, the CINH, the VIMFC, and rainfall in Jakarta and the surrounding area, over both the land and the sea areas. We found that in general, the CENS tended to suppress convection, but enhances moisture convergence. The important results of this study were that the increase of rainfall that occurred during active CENS was not actually caused by the increase of convection, but rather due to the increase in moisture convergence.

1 Introduction Various atmospheric phenomena from synoptic to mesoscale significantly influence rainfall variability over the IMC. The Northerly Cold Surge (NCS) is one of the prominent synoptic disturbances that was responsible for the occurrence of heavy D. Satiadi (B) · A. Purwaningsih · W. Harjupa · Trismidianto · D. F. Andarini · F. Nauval · E. Saufina · F. Rahmatia · R. Pratama · T. Harjana · Risyanto · I. Fathrio · E. Hermawan Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, West Java, Indonesia e-mail: [email protected] M. Yollanda · D. Devianto Mathematics Department, Andalas University, Padang, West Sumatera, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_27

279

280

D. Satiadi et al.

rainfall, especially in the western part of Java Island. It often occurs during the peak of Boreal Winter (December–January–February; DJF) with the duration from days to a week [1]. Determination of the NCS has been defined and developed over time using different meteorological variables. Riehl classified the NCS as differences in pressure between 115° E and Hong Kong along 30° N that exceed 10 hPa [2]. Moreover, the NCS consists of two stages, a significant increase in pressure (Ps ) and a steep drop in dew point (T d ) [3]. Lau and Chang concluded that based on the surface condition, the NCS can cause a sharp drop in surface temperature (4°–5°), and a decrease of humidity showed as a drastic drop of dew point (1°–2°), the enhancement of surface pressure (±3 hPa), and the strengthening of the northerly wind (5.5 m/s) through 12– 24 LT [4]. Another study by Wu and Chan suggested that the surge is tracked when the pressure over Siberia reaches 1045 hPa. Then, the onset of the NCS in Hong Kong is identified by a rapid change in its synoptic condition around 12–14 LT [5]. A more recent study from Chang et al. defined the NCS as a transient gravitational wave-like motion caused by a wind-pressure imbalance, which allows the rapid propagation of energy from the mid-latitudes to the tropics [1]. Chang introduced a quantitative cold surge index by averaging the 925 hPa meridional wind over 110° E–117.5° E and 15° N and determining the average of more than 8 m/s as threshold for active cold surge [1]. Lim et al. provided an updated definition of the South China Sea cold surge based on low-level wind speed at 850 hPa over domain 5°–10° N and 107°–115° E and mean sea level pressure (MSLP) over domain 18°–22° N and 105°–122° E [6]. Furthermore, Abdillah et al. showed quantitative evidence linking the low-latitude CS to polar cold air outbreak in Siberia and defined multiple pathways of CS in the tropics [7]. The propagation of the NCS triggers the increasing rainfall intensity over some areas. The NCS is initially categorized as dry air masses. However, along its trajectory, from the mid-latitudes and sometimes extends to the South China Sea, air parcels are moistened by the ocean. This air mass propagation is strengthening the East Asia Hadley Cell associated with the enhanced outflow in the upper troposphere [8]. More specifically, the anomaly of cold air flow is weakened by an intensified surface heat flux. Thus, the cold characteristics of initial air mass are eliminated during the cold air masses propagation to the low latitude. Sometimes, this strong cold air outbreak can penetrate far to the south region and cross the equator and is called the Cross Equatorial Northerly Surge (CENS). The CENS as a dominant low-level circulation triggers deep convection and eventually leads to torrential rainfall over the equatorial region including the northern part of Java Island [9–14]. Over Jakarta, a strong NCS crossing the equator line played a critical role for several devastating flooding events, for example, in February 2013 [14], February 2020 [15], and February 2021 [16]. Furthermore, the complex local circulations (such as land–sea breeze and mountain valley wind) and their interaction with synoptic disturbances (such as the NCS and the CENS) play a vital role for the mechanisms of precipitation over the Java Island. There is a difference between rainfall patterns over the land and the ocean in the western part of Java Island. The increase of rainfall over the land (ocean) occurred in the afternoon (morning) [17], whereas in the coastal area of West Java, the maximum rainfall occurred frequently in the early morning (around 01 LT) during

The Effects of the Cross Equatorial Northerly Surge (CENS) …

281

the wet season. Moreover, the interaction of local topography with the CENS might affect the variability of the amplitude and intensity of diurnal precipitation. The CENS modulated the early morning precipitation and extreme rainfall throughout the northern coast of West Java [9–20]. Nonetheless, the mechanism of the multiscale interactions in modulating diurnal rainfall variability remains unclear [21]. Therefore, a thorough study addressing the mechanism is discussed in this paper. This study aims to investigate the dynamic process of diurnal rainfall over the Jakarta area under the influence of CENS. We focus on analyzing the influence of the CENS on the convective and moisture convergence activities over the land and the sea areas utilizing several parameters, namely the Convective Available Potential Energy (CAPE), the Convective Inhibition (CINH), and the Vertically Integrated Moisture Flux Convergence (VIMFC). Convective processes were represented by the CAPE and the CINH. The CAPE is an indication of the instability or stability of the atmosphere that is used to assess the potential for convection development. Potential energy is represented by the total excess buoyancy, whereas the CINH measures the amount of energy that will inhibit an air parcel from rising from the surface to the level of free convection [22]. The convergence activities were represented by the VIMFC, which indicates the potential of moisture to converge over a certain location. Overall, this study gives understanding about how convective and convergence activities undergo over the land and the sea and thus influence the diurnal rainfall variation and how the CENS influences these activities.

2 Data and Methods 2.1 Data This study used the hourly gridded dataset (0.25° × 0.25°) from the fifth-generation European Center for Medium-Range Weather Forecasts (ECMWF) Atmospheric Reanalysis of the global climate (ERA5) [23]. We obtained several parameters on a single level such as the CAPE, the CINH, the VIMFC, and precipitation from http:// www.ecmwf.int/. We retrieved all the mentioned data for two locations (Onshore Jakarta: 6.4° S, 106.8° E; and Offshore Jakarta: 5.5° S, 106.8° E) representing the land and the sea areas (see Fig. 1). These parameters were employed to investigate the convective and moisture convergence activities associated with the CENS in 2014. The year 2014 was selected for this study, since the CENS phenomena were relatively strong and quite persistent, occurring intermittently for more than two weeks. To identify the CENS, the daily mean wind and wind stress obtained from the Advanced Scatterometer (ASCAT) and Daily ASCAT (DASCAT) with field gridded data (0.25° × 0.25°) were used. The data was accessed for the same period as the ERA5 data, from ftp://ftp.ifremer.fr/ifremer/cersat/products/gridded/. The ASCAT is a real aperture radar with vertically polarized antennas operating at 5.255 GHz frequency (C-band) [24]. It transmits a long pulse with linear frequency modulation (“chirp”).

282

D. Satiadi et al.

Fig. 1 Map of the study area in the western part of Java Island. The red marks denote the locations under study (Onshore Jakarta: 6.4° S, 106.8° E and Offshore Jakarta: 5.5° S, 106.8° E)

2.2 Methods The Identification of the CENS. To determine the occurrences of the CENS, we applied a method from Hattori et al. (2011) by calculating its daily indices. The CENS index is identified by an area average of 925 hPa meridional wind over the area of 0°–5° S and 105°–115° E. Then, a threshold of 5 m/s was used to define an active phase of CENS [19]. Analysis of Convective and Convergence Processes. Convective processes were analyzed by investigating the diurnal and seasonal cycles of the CAPE and the CINH. The CAPE is the term used for energy available in the atmosphere for convection to occur. The CAPE is basically the type of energy available to lift an air parcel from the level of free convection to the level of neutral buoyancy [25]. Positive CAPE values point to conditions that are not stable, and there are growing chances of hail and thunderstorms. It is one of the most frequently used indicators of meteorological conditions that are convenient for the incidence of energetic precipitation events along with lightning, wind shear, and hailstorms [26, 27] . According to a study conducted by Chou and Neelin (2004), in convective zones, positive CAPE is maintained by increasing moisture content [28]. Conversely, the CINH measures the amount of energy that will prevent an air parcel from rising from the surface to the level of free convection [22]. Convergence processes were analyzed by investigating the diurnal and seasonal cycles of the VIMFC. The VIMFC is the horizontal rate of flow of moisture (water vapor, cloud liquid, and cloud ice), per square meter across the flow, for a column of air extending from the surface of the earth to the top of the atmosphere. Its horizontal divergence is the rate of moisture spreading outward from a point, per square meter.

The Effects of the Cross Equatorial Northerly Surge (CENS) …

283

This parameter is positive for moisture that is spreading out, or diverging, and negative for the opposite, for moisture that is concentrating, or converging. The VIMFC has a strong correlation with the rain occurrence [29, 30].

3 Results and Discussion 3.1 The Effects of CENS Over the Land Area The hourly data from ERA5 and the CENS indices during 2014 was plotted to examine diurnal rainfall in relation to the conditions of the CAPE, the CINH, and the VIMFC under the influence of the CENS phenomena. Figure 2 shows rainfall (Fig. 2a), the CAPE (Fig. 2b), the CINH (Fig. 2c), the VIMFC (Fig. 2d), dew point temperature (Fig. 2e), and the CENS indices (Fig. 2f) against hour (y-axis) and day (x-axis) during 2014 over the land area (Onshore Jakarta) (6.4° S, 106.8° E). Although the CENS usually occurs only in DJF, one-year data was plotted to provide comparison with the normal conditions. Results indicate that there was a change in temporal distribution of diurnal rainfall during the CENS phenomenon compared to that without the phenomenon (see Fig. 2a). Rainfall over the land area generally increased in the late afternoon around 12–18 local time. This result was expected presumably due to the intensification of atmospheric convection caused by the heating of the land surface by solar radiation during the day. However, during active CENS (in January and also February), rainfall occurrences seemed to be extended from the early morning to the evening. The change of the magnitude of the CAPE during the CENS phenomenon was also identified (see Fig. 2b). Over the land area, the CAPE generally increased in the afternoon at around 10–16 local time, at an earlier time than that of rainfall. This result was expected presumably due to the atmospheric destabilization caused by the heating of the land surface due to solar radiation during the day. However, the growth of the CAPE tended to be suppressed during active CENS. The behavior of CINH under the influence of the CENS was analyzed (see Fig. 2c). Over the land area, the CINH generally increased in the morning and in the evening and relatively reduced in the afternoon. The increase of the CINH in the morning was important for allowing the accumulation of the CAPE in the afternoon. Subsequently, the decrease of CINH in the afternoon was important for allowing the release of the CAPE in the form of convection and rainfall in the late afternoon. After the process, the CINH increased again in the evening until early morning, which suppressed convection. Similar to that of the CAPE, the CINH development seemed to be suppressed during active CENS. The CINH decreased possibly due to the wind convergence between the CENS and the land breeze, which provided a mechanical lifting through the inhibition layer [31]. Furthermore, to investigate the moisture convergence process, the VIMFC was analyzed during and without the CENS phenomenon (see Fig. 2d). Over the land

284

D. Satiadi et al.

Fig. 2 a Rainfall, b CAPE, c CINH, d VIMFC, e dew point temperature, and f CENS indices over the land area (Onshore Jakarta) in 2014

area, the VIMFC generally increased in the late afternoon. This result was expected due to the intensification of atmospheric convection and also convergence caused by the heating of the land surface due to solar radiation during the day. Note that the onset of the VIMFC was more in phase with the rainfall than that of the CAPE, whereas during the active CENS (in January and also February), the VIMFC seemed to be extended from the early morning to the evening. The temporal distribution of the VIMFC was also more spread out during negative CENS indices. Figure 2f shows the daily CENS indices during 2014. Positive (negative) indices mean that the wind propagated northward (southward). The magnitude of the indices indicated the wind strength. The CENS was active when the southerly wind speed was greater than 5.5 m/s, indicated by the blue lines. It can be seen from Fig. 2e that a relatively strong and persistent CENS occurred in January and also a brief one in February 2014.

The Effects of the Cross Equatorial Northerly Surge (CENS) …

285

Overall, the interaction between the CENS and convection or convergence processes over the land areas can explain which factor, whether convective or convergence, was responsible for the change in rainfall in terms of its timing and intensity. When the CENS was active, the CAPE and the CINH tended to be significantly suppressed, but the rainfall tended to be increased (See Fig. 2). The decrease of the CAPE development during active CENS was presumably due to the advection of the cold and dry wind gusts from the north, which reduced the dew point temperature (See Fig. 2f) and hence reduced the CAPE. Koseki (2013) has also shown that stronger monsoon does suppress convection due to lower the equivalent potential temperature. In contrast, the VIMFC tended to be enhanced and more extended in time from the early morning until the evening during active CENS. The results indicated that the increase of rainfall over the land during active CENS was not related to the increase of convection, which was actually suppressed, but was more related to the increase of the moisture convergence. The decrease of CINH was also an important factor in this mechanism, since it allowed the forced convection to take place easily, despite the lower CAPE.

3.2 The Effects of CENS Over the Sea Area The same analysis as in Sect. 3.1 was conducted for the sea area to examine the diurnal rainfall in relation to the conditions of the CAPE, the CINH, and the VIMFC under the influence of the CENS and to compare the results between the land and the sea areas. Figure 3 shows rainfall (Fig. 3a), the CAPE (Fig. 3b), the CINH (Fig. 3c), the VIMFC (Fig. 3d), dew point temperature (Fig. 3e), and the CENS indices (Fig. 3f) against hour (y-axis) and days (x-axis) during 2014 over the sea area (Offshore Jakarta) (5.5° S, 106.8° E). In general, there was a contrast between the phase of diurnal rainfall profile over the land and that over the sea areas. It can be seen from Fig. 3a that rainfall over the sea area generally increased during the night. This result was expected presumably due to the intensification of atmospheric convection caused by the heat and water vapor released by the sea surface during the night. Over the sea area, the CENS also influenced the intensity of rainfall. During the active CENS (in January and also February), there was an increase of rainfall during the night. Moreover, the effects of CENS on the CAPE and the CINH profiles over the sea area were identified (Fig. 3b, c). Over the sea area, the CAPE was relatively high during the day and the night, but there was enhancement of CAPE particularly during night (see Fig. 3b). This result was also expected presumably due to the enhancement of atmospheric destabilization caused by the heating of the sea surface during the night. It can be seen from Fig. 3b that the CAPE development tended to be significantly suppressed during active CENS. Moreover, the CINH over the sea generally increased in the late afternoon and relatively reduced during the night (see Fig. 3c). The increase of the CINH in the late afternoon was important for allowing the accumulation of the CAPE in the evening. Subsequently, the decrease of CINH

286

D. Satiadi et al.

Fig. 3 a Rainfall, b CAPE, c CINH, d VIMFC, e dew point temperature, and f CENS indices over the sea area (Offshore Jakarta) in 2014

was important for allowing the release of the CAPE in the form of convection and rainfall during the night. The CINH seemed to be suppressed during active CENS, possibly due to the wind convergence, which provided a mechanical lifting through the inhibition layer. It can be seen from Fig. 2d that over the sea area, the VIMFC was generally increased during the night. This result was expected presumably due to the intensification of atmospheric convection and also convergence caused by the heat released by the sea surface during the night. Figure 2d also shows that during negative CENS indices, the VIMFC was more intensified, especially when the CENS phenomena occurred (in January and also February). Overall, the results over the sea were similar to that over the land area regarding the effects of CENS on convection, moisture convergence, and rainfall. When CENS was active, the CAPE as well as the CINH tended to be significantly suppressed, but the

The Effects of the Cross Equatorial Northerly Surge (CENS) …

287

VIMFC and rainfall tended to be increased during the night and early morning (see Fig. 3). The decrease of the CAPE development during active CENS was presumably due to the advection of cold and dry wind gusts from the north, which reduced the dew point temperature (See Fig. 2f) and hence reduced the CAPE. Therefore, the increase of rainfall over the sea area during active CENS over the sea area was not related to the increase of convection, which was actually suppressed, but was more related to the increase of moisture convergence. The elimination of CINH was also an important factor in this mechanism, since it allowed the forced convection to take place easily, despite the lower CAPE.

4 Conclusions Based on the analysis described in the previous sections, we found a very strong influence of the CENS on the diurnal cycle of convection, moisture convergence, and rainfall over Jakarta that was found both over the land and the sea areas. In general, the CENS was found to suppress the growth of the CAPE during the day and the night, over both the land and the sea. The decrease of the CAPE development during active CENS was presumably due to the advection of cold and dry wind gusts from the north, which reduced the dew point temperature and stabilized the atmosphere. The effect of the CENS on the CINH was similar to that on the CAPE. The CENS generally reduced the formation of the CINH significantly during the day and the night, over both the land and the sea, possibly due to the wind convergence between the CENS and the land breeze, which provided a mechanical lifting through the inhibition layer. In contrast, the CENS was found to enhance the VIMFC over both the land and the sea. Therefore, one of the important results from this study was that the increase of rainfall that occurred during active CENS was not related to the increase of convection, which was actually suppressed, but was more related to the increase of moisture convergence, which was more dominant when the CENS was active. The elimination of CINH was also an important factor in this mechanism, since it allowed the forced convection to take place easily, despite the lower CAPE. Acknowledgements This research is supported by the National Research and Innovation Agency (BRIN) of the Republic of Indonesia. We are grateful for the availability of the fifth-generation European Center for Medium-Range Weather Forecasts (ECMWF) Atmospheric Reanalysis of the global climate (ERA5) data downloaded from http://www.ecmwf.int and also the ASCAT and DASCAT data downloaded from ftp://ftp.ifremer.fr/ifremer/cersat/products/gridded/.

288

D. Satiadi et al.

References 1. Chang, C.P., Harr, P.A., Chen, H.J.: Synoptic disturbances over the equatorial South China Sea and western Maritime Continent during boreal winter. Mon. Weather Rev. 133, 489–503 (2005) 2. Riehl, H.: Surface wind over the South China Sea during the northeast monsoon season. Navy Weather Res. Facility Tech. 24, 22–68 (1968) 3. Chang, C.P., Millard, J.E., Chen, G.T.J.: Gravitational character of cold surges during winter MONEX. Mon. Wea. Rev. 111, 293–307 (1983) 4. Lau, K.M., Chang, C.P.: Planetary scale aspects of the winter monsoon and atmospheric teleconnection. In: Chang, C.P., Krishnamurti, T.N., (eds.) Monsoon Meteorology, pp. 161–202. Oxford University Press (1987) 5. Wu, M.C., Chan, J.C.L.: Surface features of winter monsoon surges over South China. Mon. Weather Rev. 123, 662–680 (1995) 6. Lim et al.: Impacts of Boreal winter monsoon cold surges and the interaction with MJO on Southeast Asia Rainfall. J. Clim. 30(11), 4267–4281 (2017). https://doi.org/10.1175/JCLI-D16-0546.1 7. Abdillah et al.: Cold surge pathways in East Asia and their tropical impacts. J. Clim. 34(1), 157–170 (2021). https://doi.org/10.1175/JCLI-D-20-0552.1 8. Huang, P. et al.: Interannual variability of regional Hadley circulation intensity over western pacific during Boreal winter and its climatic impact over Asia-Australia Region. JGR Atmosphere 123(1), 344–366. https://doi.org/10.1002/2017JD027919 9. Wu, P., Hara, M., Fudeyasu, H., et al.: The impact of trans-equatorial monsoon flow on the formation of repeated torrential rains over Java Island. SOLA 3, 093–096 (2007) 10. Tangang, F.T., Juneng, L., Salimun, E., et al.: On the roles of the northeast cold surge, the Borneo vortex, the Madden-Julian Oscillation, and the Indian Ocean Dipole during the extreme 2006/2007 flood in southern Peninsular Malaysia. Geophys. Res. Lett. 35(14), 1–6 (2008) 11. Yokoi, S., Matsumoto, J.: Collaborative effects of cold surge and tropical depression-type disturbance on heavy rainfall in Central Vietnam. Mon. Weather Rev. 136, 3275–3287 (2008) 12. Chen, T.C., Tsay, J.D., Matsumoto, J., Alpert, J.: Development and formation mechanism of the Southeast Asian Winter Heavy Rainfall Events around the South China Sea. Part I: Formation and Propagation of Cold Surge Vortex. J. Clim. 28, 1417–1443 (2015) 13. Hermawan, E., Lubis, S.W., Harjana, T., Purwaningsih, A., Risyanto, Ridho, A., Andarini, D.F., Ratri, D.N., Widyaningsih, R.: Large-scale meteorological drivers of the extreme precipitation event and devastating floods of early-February 2021 in Semarang, Central Java, Indonesia. Atmosphere 13(7), 1092 (2022) 14. Yulihastin, E., Trilaksono, N.J.: Evolution of heavy rainfall in Jakarta flood case 2013 based on COSMO Model. Asia Pacific Center for Ecohydrology-UNESCO Research Center for Limnology-Indonesian Institute of Sciences, 167 (2014) 15. Saufina, E., Trismidianto, Risyanto, Fathrio, I., Harjupa, W.: Impact of cross equatorial northerly surge (CENS) on Jakarta heavy rainfall and its interaction with tropical cyclones (Case study: 18–25 February 2020). AIP Conf. Proc. 2366, 050002 (2021) 16. Purwaningsih, A., Tank, A.K., Vila, J.: Atmospheric conditions associated with northerly surge, Borneo vortex and madden Julian oscillation during the extreme rainfall cases in early 2021 over the western part of the maritime continent. In: Yulihastin, E., Abadi, P., Sitompul, P., Harjupa, W. (eds.) Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, 2021. Springer Proceedings in Physics, vol. 275. Springer, Singapore (2022) 17. Yulihastin, E., Fathrio, I., Trismidianto, Nauval, F., Saufina, E., Harjupa, W., Satiadi, D., Nuryanto, D.E.: Convective cold pool associated with offshore propagation of convection system over the East Coast of Southern Sumatra, Indonesia. Adv. Meteorol. (2021) 18. Yulihastin, E., Hadi, T.W., Ningsih, N.S., Syahputra, M.R.: Early morning peaks in the diurnal cycle of precipitation over the northern coast of West Java and possible influencing factors. Ann. Geophys 38, 231–242 (2020)

The Effects of the Cross Equatorial Northerly Surge (CENS) …

289

19. Hattori, M., Mori, S., Matsumoto, J.: The cross-equatorial northerly surge over the maritime continent and its relationship to precipitation patterns. J. Meteorol. Soc. Japan Ser. II(89), 27–47 (2011) 20. Mori, S., Hamada, J.I., Hattori, M., et al.: Meridional March of Diurnal Rainfall over Jakarta, Indonesia, Observed with a C-band Doppler Radar: an overview of the HARIMAU 2010 Campaign. Progress Earth Planetary Sci. 5(47) (2018) 21. Yulihastin, E.: Propagation of convective systems associated with early morning precipitation and different northerly background winds over Western Java. J. Meteorol. Soc. Japan. Ser. II 100(1), 99–113 (2022). https://doi.org/10.2151/jmsj.2022-005 22. DeAngelis, A.M., Qu, X., Zelinka, M.D., Hall, A.: An observational radiative constraint on hydrologic cycle intensification. Nature 528, 249–253 (2015) 23. Hersbach, H., Bell, B., Berrisford, P., Biavati, G., Horányi, A., Muñoz Sabater, J., Nicolas, J., Peubey, C., Radu, R., Rozum, I., Schepers, D., Simmons, A., Soci, C., Dee, D., Thépaut, J.-N.: ERA5 hourly data on single levels from 1959 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS) (2018). Accessed 2022/05/23 24. Klaes, K.D.: The EUMETSAT Polar System. Comprehensive Remote Sensing, 192–219 (2018) 25. Moncrieff, M.W., Miller, M.J.: The dynamics and simulation of tropical cumulonimbus and squall lines. Q. J. R. Meteorol. Soc. 102, 373–394 (1976) 26. Zawadzki, I., Torlaschi, E., Sauvageau, R.: The relationship between mesoscale thermodynamic variables and convective precipitation. J. Atmos. Sci 38, 1535–1540 (1981) 27. Singh, C., Ganguly, D., Dash, S.K.: Dust load and rainfall characteristics and their relationship over the South Asian monsoon region under various warming scenarios. J. Geophys. Res. Atmos. 122, 7896–7921 (2017) 28. Chou, C., Neelin, J.D.: Mechanisms of global warming impacts on regional tropical precipitation. J. Clim. 17, 2688–2701 (2004) 29. Kosum, C., Luadsong, A., Aschariyaphotha, N.: Vertically integrated moisture flux convergence over Southeast Asia and its relation to rainfall over Thailand. Pertanika J. Sci. Technol. 26(1), 235–246 (2018) 30. Darand, M., Pazhoh, F.: Vertically integrated moisture flux convergence over Iran. Clim. Dyn. 53, 3561–3582 (2019). Author, F.: Article title. Journal 2(5), 99–110 (2016) 31. Colby, F.P., Jr.: Convective inhibition as a predictor of convection during AVE-SESAME II. Mon. Weather Rev. 112(11), 2239–2252 (1984)

Comparison Influencing of El-Nino Southern Oscillation and Indian Ocean Dipole on Rainfall Variability During the Asian Winter and Summer Monsoon Over Indonesian Maritime Continent Trisni Hadiningrum, Deni Okta Lestari, and Trismidianto Abstract The Indian Ocean Dipole (IOD) and El-Nino Southern Oscillation (ENSO) are part of the atmospheric phenomena that considerably influence rainfall variability in Indonesia. This study has reported the impact of ENSO and IOD on Indonesian rainfall variability when they interact with each other. This analysis uses daily rainfall data from Aphrodite, sea surface temperature (SST), horizontal wind, Ocean Nino Index (ONI) index, and IOD index from 1961 to 2010. The results of this study indicate that the effect of LN and NIOD that co-occurs on rainfall variability in Indonesia is more robust in the AMJJASO than in the NDJFM. The impact of LN and PIOD simultaneously on rainfall is more visible in the NDJFM than in the AMJJASO. The effect of NIOD on rainfall is more dominant than EN when it co-occurs in the NDJFM and the AMJJASO. The impact of NIOD + EN on rain in Java is more significant during AMJJASO than during NDJFM. Events of PIOD and EN that co-occur will cause drought in almost all of Indonesia, and a high decrease in rainfall will occur on the island of Java during AMJJASO. The influence of LN on rainfall is more dominant than NetIOD during NDJFM and AMJJASO. The effect of LN + NetIOD on increasing rainfall in Indonesia is higher during AMJJASO than during NDJFM. The impact of EN, when it co-occurs with NetIOD, on the increase in rainfall in Sumatra is higher during NDJFM than during AMJJASO. The influence of PIOD and NetENSO on drought in Indonesia is quite strong if the two phenomena co-occur during AMJJASO.

T. Hadiningrum (B) · D. O. Lestari Department of Atmospheric and Planetary Science, Sumatera Institute of Technology, Lampung, Indonesia e-mail: [email protected] Trismidianto Research Center for Climate and Atmospheric, National Research and Innovation Agency (BRIN), Bandung 40173, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_28

291

292

T. Hadiningrum et al.

1 Introduction Indonesian Maritime Continent (IMC) lies at the heart of the tropical region’s warm pool. It is one of the world’s most convectively active areas, affecting the global climate system. It is located between two continents (Asia and Australia) and between two oceans (the Indian and the Pacific Oceans) [1]. The IMC is a unique geographical region composed of a complex system of mountainous islands and consists of five large (Papua, Kalimantan, Sumatera, Sulawesi, and Java) and more than 13,000 smaller islands. It has about 5000 km or 1/8 of the equatorial circumference in longitude [2]. This individual IMC situation recognized the central importance of this region of heating as one of the primary energy sources for the entire global circulation system [1]. As it is, the air is primarily humid, and the enhanced cloudiness indicates massive exchanges of energy that are fundamentally important in the general circulation of the global atmosphere. Of this region, none is more important in global climate dynamics than the maritime continent region because of its role in providing energy for the operation of the north–south tropical Hadley cell and the east–west Walker circulation, both essential components of the global circulation [3]. McBride [4] and recently, Slingo et al. [5] have shown that the IMC heat source is a significant driver of global circulation. The unique IMC conditions also make IMC weather and climate variability influenced by various weather systems and atmospheric phenomena occurring at different temporal and spatial scales, with time range ranging from intra-diurnal (e.g., sea breezes) to inter-annual (e.g., El-Nino Southern Oscillation) and spatial scales ranging from a few kilometers (e.g., individual cloud cells), to thousands of kilometers (e.g., the Madden–Julian Oscillation), to planetary-scale ENSO oscillations. The effect of these atmospheric phenomena occurs when these phenomena occur singly or interact with each other. One of the interactions that is part of the atmospheric phenomenon that has a considerable influence on the rainfall variability in Indonesia is the interaction between the El-Nino Southern Oscillation (ENSO) and the Indian Ocean Dipole (IOD). ENSO is an irregularly periodical variation in winds and sea surface temperatures over the tropical Eastern Pacific Ocean, affecting much of the tropics and subtropics. The warming phase is known as El Niño, and the cooling phase is known as La Niña. IOD, also known as the Indian Niño, is an irregular oscillation of sea surface temperatures in which the western Indian Ocean becomes alternately warmer and colder than the eastern part of the ocean. The previous research has shown that ENSO and IOD impact rainfall in Indonesia, where most of Indonesia have a positive rainfall anomaly ranging from 100 to 300 mm/day, except for Sumatra Island, with a rainfall anomaly of −200 mm/day. The same condition occurred in September–October–November and December– January–February, where only a few areas in Sumatra, Java, Kalimantan, and Sulawesi had negative rainfall anomalies [6]. Seasonal analysis shows that rainfall is low during June, July, and August, while high precipitation occurs in almost all parts of Indonesia during the DJF season [7]. The effect of ENSO and IOD has

Comparison Influencing of El-Nino Southern Oscillation and Indian …

293

been reported by the previous studies using data analysis for 3-month periods (i.e., DJF (December, January, February), MAM (March, April, May), JJA (June, July, August), SON (September, October, November)). From several previous studies, there has never been a detailed explanation of the effect of ENSO and IOD interactions that co-occur in two different seasonal conditions. Therefore, this study aims to compare how much influence the interaction between ENSO and IOD has on the two seasons in Indonesia. However, this study uses a season division based on the Asian monsoon division, namely Winter Asian Monsoon as the rainy season, which is often referred to as November to March (NDJFM), and Boreal Summer Monsoon as the dry season occurs from April to October (AMJJASO) [8].

2 Data and Method We use the ENSO and IOD index to analyze the condition of ENSO and IOD. The ENSO index is seen from the SST anomaly, Nino 3.4, and IOS Tahiti-Darwin, defined as the difference in the Sea Level Pressure (SLP) anomaly in the East Pacific, Tahiti (17.6° S, 149.6° W) with the SLP in the West Pacific, Darwin (12.4° S, 130° E). The Ocean Nino Index is obtained by representing the 3-month average of the ERSST.v5 SST anomaly in the Niño 3.4 region (5° N–5° S, 120° –170° W). The index data can be downloaded via http://www.cpc.ncep.noaa.gov/data/indices/oni.ascii.txt. The IOD index is seen from the difference in the western equatorial Indian Ocean SST anomaly (50° E–70° E/10° S–10° N) with the Indian Ocean SST anomaly off the coast of Sumatra (90° E–110° E/10o S—Equator). This IOD index can be downloaded free from the following link: https://psl.noaa/gov/data/timeseries/DMI. We classified the incidence of ENSO and IOD during the study period from January 1961 to December 2010. In this study, rainfall forecast data use data from APHRODITE. Asian Precipitation-Highly-Resolved Observational Data Integration Toward Evaluation’s (APHRODITE) gridded precipitation is a set of long-term (1951 onward) continentalscale daily products based on a dense network of rain-gage data for Asia, including the Himalayas, South, and Southeast Asia and mountainous areas in the Middle East. The gridded products are available for four sub-domains—Monsoon Asia, Middle East, Russia, and Japan—and a combined domain. The time-varying data have a 0.25 × 0.25 or 0.5 × 0.5 degree horizontal resolution in each part, except for Japan, which has a 0.05 × 0.05 degree horizontal resolution. Climatological daily mean precipitation and temperature data are available for Monsoon Asia at 0.05 × 0.05 degree resolution. This study used this data from January 1961 to December 2010. At this link, the data can be downloaded http://aphrodite.st.hirosaki-u.ac.jp/download/. To analyze wind and sea surface temperature, we used product data from the National Centers for Environmental Prediction (NCEP) and the National Center for Atmospheric Research (NCAR) with a resolution of 2.5° × 2.5° in the period January 1961 to December 2010. NCEP/NCAR Reanalysis 1 is an analysis/estimation of a complex project system to perform data assimilation using past data from 1948 to

294

T. Hadiningrum et al.

the present [9]. This data can be downloaded for free from the following link: https:// psl.noaa.gov/data/gridded/data.ncep.reanalysis.html. In this study, a composite spatial analysis was carried out for each predetermined event condition. By grouping the data into two seasons, such as AMJJASO and NDJFM, then it composite data when the ENSO and IOD phenomena occur. After that, we composite data on rainfall, SST, and wind. Then, we look for anomalies from the two data by subtracting the composite data at the time of the incident from the overall composite data for each season. To make it easier to do the analysis, we do some abbreviations, namely LaNina = LN, ElNino = EN, Negative IOD = NIOD, Positive IOD = PIOD, Neutral ENSO = NetENSO, Neutral IOD = NetIOD.

3 Result and Discussion 3.1 Statistical Data of ENSO and IOD Figure 1 shows the frequent interactions between ENSO and IOD in the AMJJASO season, indicated by the dashed and standard red lines, and in the NDJFM season,

Fig. 1 Time series of monthly ENSO and IOD during AMJJASO and NDJFM over events in 1961–2010

Comparison Influencing of El-Nino Southern Oscillation and Indian …

295

indicated by the black lines. Figure 1 shows several LN occurrences simultaneously as NIOD, even with a high index, and the event of EN along with NIOD or PIOD in the AMJJASO and NDJFM seasons.

3.2 Interaction ENSO with IOD When the LN coincides with NIOD during the NDJFM season, there is an increase in rainfall around eastern Kalimantan, East Java, northern Sumatra, and almost all of Sulawesi, with a slight increase in intensity (see Fig. 2a). About 20 mm/month, other areas experienced a decrease in rainfall. The increase in precipitation is very significant in almost all of Indonesia when LN and NIOD co-occur during the AMJJASO season, as shown in Fig. 2b. In fact, an increase in rainfall with a reasonably high intensity reaching 60–80 mm/month has occurred in almost all areas in Java and Kalimantan. This event shows that LN and NIOD in rainfall variability in Indonesia are more robust in the AMJJASO season than in the NDJFM season. The effect of LN and PIOD co-occurs on rainfall during the NDJFM season, it can be seen that the influence of LN is quite dominant compared to PIOD, where the increase in rainfall only occurs in Java and eastern Indonesia, while in Sumatra, almost all experience a decrease in rainfall as shown in Fig. 2c [10]. Compared to the AMJJASO season, during the NDJFM season, rainfall increases only occur in

Fig. 2 Rainfall anomaly during NDJFM on a LN + NIOD and c LN + PIOD and during AMJJASO on b LN + NIOD and d LN + PIOD. Unit in mm/month

296

T. Hadiningrum et al.

Papua (see Fig. 2d). Meanwhile, almost all other areas experienced a decrease in precipitation. Even the island of Java, which previously shared a significant increase in rainfall during NDJFM, was not seen during the AMJJASO season. This event shows that the simultaneous influence of LN and PIOD on rainfall is more visible in the NDJFM season than in the AMJJASO season. The impact of LN also seems to be more dominant than PIOD because, during PIOD, Indonesian territory should experience dryness [11]. When EN and NIOD co-occur, the dominant influence of NIOD compared to EN on rainfall in Indonesia both in the NDJFM and AMJJASO seasons, with a significant increase in rainfall in several areas in western Indonesia, especially in Sumatra and Java (See Fig. 3a, b). Regions that experience increased rainfall are more comprehensive during the AMJJASO season than during the NDJFM season. The increase in precipitation with an intensity of about 40–60 mm/month is seen evenly in Java during the AMJJASO season (see Fig. 3a) compared to NDJFM, only seen around East Java with a small intensity. About 20 mm/month (see Fig. 3b), therefore, EN + NIOD on rainfall in Indonesia is more significant in the AMJJASO season than in the NDJFM season. EN and PIOD made Indonesia dry when this phenomenon occurred [12]. In Fig. 3c, d, almost all of Indonesia experience drought or decreased rainfall when EN and PIOD co-occur in the NDJFM or AMJJASO season. There was increased precipitation around northern Sumatra in both seasons, and other phenomena may need to be studied further. The decrease in rainfall during NDJFM ranges from 60– 80 mm/month in almost all of Indonesia (Fig. 3c). The decline in rainfall even reaches

Fig. 3 Rainfall anomaly during NDJFM on a EN + NIOD and c EN + PIOD and during AMJJASO on b EN + NIOD and d EN + PIOD. Unit in mm/month

Comparison Influencing of El-Nino Southern Oscillation and Indian …

297

80–100 mm/month around the island of Java when the two phenomena co-occur in the AMJJASO season (Fig. 3d).

3.3 Interaction of ENSO and IOD If One is Neutral Figure 4a, b shows the effect of LN on NetIOD during the NDJFM and AMJJASO seasons. The influence of LN is more dominant than NetIOD during NDJFM and AMJJASO, with an increase in rainfall around central and eastern Indonesia. Sumatra looks very dry; there is no increase in rainfall. The increase in rainfall broadly more during the AMJJASO season than NDJFM, where there was a fairly even increase in the islands of Kalimantan, Sulawesi, and Papua with small intensities of around 0– 20 mm/month (see Fig. 4b). However, the increase in rainfall in northern Kalimantan is more significant during NDJFM than during AMJJASO (see Fig. 4a). This shows the LN effect at the same time as NetIOD on rainfall in northern Kalimantan is more significant during the NDJFM season than during AMJJASO. The effect of EN during NetIOD is also clearly seen in Fig. 4c, d, which causes drought in almost all of central and eastern Indonesia during the NDJFM and AMJJASO Seasons. NetIOD conditions indicate an increase in rainfall in Sumatra and parts of Kalimantan. The increase in rainfall in Sumatra is higher during NDJFM than during AMJJASO. However, the distribution of areas experiencing increased

Fig. 4 Rainfall anomaly during NDJFM on a LN + NetIOD and c EN + NetIOD and during AMJJASO on b LN + NetIOD and d EN + NetIOD. Unit in mm/month

298

T. Hadiningrum et al.

rainfall in Kalimantan is wider during AMJJASO than during NDJFM, with almost the same intensity. Then, the effect of EN during Net IOD on rainfall in Sumatra is higher during NDJFM than during AMJJASO. However, in Kalimantan, the impact is more significant during AMJJASO than during NDJFM [13]. NetENSO conditions will make it dry in central and eastern Indonesia when they co-occur with NIOD during the NDJFM season (see Fig. 5a). However, this did not happen during the AMJJASO season because there was an increase in rainfall in some areas of eastern Kalimantan and almost all of Papua although with a low intensity of around 0–20 mm/month (see Fig. 5b). The effect of NIOD that co-occurs with NetENSO on rainfall is more visible during AMJJASO than NDJFM; this is indicated by an increase in precipitation with low intensity in almost all of Java and southern Sumatra. However, it is seen in parts of Sumatra during NDJFM. This shows that the effect of NIOD on increasing rainfall in Sumatra and Java is more dominant than NetENSO, which acts more as a decrease in rainfall intensity when the two phenomena co-occur. PIOD conditions should make western Indonesia dry [14]. This can only be seen during the AMJJASO season (see Fig. 5c) but not during NDJFM (See Fig. 5d). The significant increase in rainfall on the islands of Sumatra, Kalimantan, Papua, and Java when PIOD co-occurs with NetENSO during the NDJFM season is thought to be due to the influence of NetENSO followed by other phenomena that need further follow-up. During AMJJASO, the increase in rainfall was only visible over the island of Papua, while other areas experienced drought. This shows that the influence of

Fig. 5 Rainfall anomaly during NDJFM on a NetENSO + NIOD and c NetENSO + PIOD and during AMJJASO on b NetENSO + NIOD and d NetENSO + PIOD. Unit in mm/month

Comparison Influencing of El-Nino Southern Oscillation and Indian …

299

PIOD and NetENSO on drought in Indonesia is quite strong if the two phenomena co-occur during AMJJASO.

3.4 SST and Wind Anomaly Analysis Figure 6a, b shows LN conditions that co-occur with NIOD, where there is a decrease in sea surface temperature around the Pacific Ocean and also in the Indian Ocean, while in Indonesia, there is an increase in sea surface temperature. This resulted in the Indonesian region experiencing wet conditions that received a supply of water vapor from the west and east. Figure 6a, b strengthens the analysis of Fig. 2a, b, where the influence of LN and NIOD that co-occurs is more substantial during AMJJASO than during NDJFM due to the decrease in sea surface temperature in the Pacific and Indian Oceans that is greater in intensity during the AMJJASO season than during NDJFM. The decrease in sea surface temperature is very high in the Indonesian region when the PIOD co-occurs with the LN in the NDJFM season (see Fig. 6c) [15]. This strengthens the analysis of Fig. 2c, namely the decrease in rainfall in almost all Indonesian regions. This decrease in sea surface temperature also occurred in Indonesia during AMJJASO when these two phenomena occurred (See Fig. 6d). This makes parts of Indonesia dry, as shown in Fig. 2d.

Fig. 6 Composite of SSTs anomaly (Shaded; °C) and 850 hPa wind anomaly (vector; m/s) during a LN + NIOD (NDJFM), b LN + NIOD (AMJJASO), c LN + PIOD (NDJFM), d LN + PIOD (AMJJASO), e EN + NIOD (NDJFM), f EN + NIOD (AMJJASO), g EN + PIOD (NDJFM) and h EN + PIOD (AMJJASO)

300

T. Hadiningrum et al.

Figure 6c, d shows the occurrence of EN that coincided with NIOD in the NDJFM and AMJJASO seasons. It is clear that there is a decrease in sea surface temperature in eastern Indonesia caused by EN, while in western Indonesia, there is an increase in sea surface temperature due to NIOD (See Fig. 6c). This is what causes an increase in rainfall in west Indonesia, while the eastern part is experiencing drought as shown in Fig. 3a, b. Lack in Indonesia extends to almost all of Indonesia, when EN cooccurs with PIOD due to an increase in sea surface temperature in the Pacific and Indian oceans, causing the Indonesian region to experience a decrease in sea surface temperature, which causes drought (see Fig. 6e, f). This strengthens the analysis of Fig. 3c, d. Figure 7a, b strengthens the analysis of Fig. 4a, b, where when the LN and NetIOD co-occur, the decrease in sea surface temperature in the Pacific Ocean causes an increase in sea surface temperature in Indonesia so that it becomes a hot pool which causes an increase in rainfall in Indonesia. However, the rise in the rain only occurs in parts of central and eastern Indonesia with low intensity; this is due to the influence of uncertain NetIOD conditions, which can make Indonesian areas wet or dry. Figure 7c, d clearly shows a decrease in sea surface temperature in the Pacific Ocean, indicating the presence of EN. This makes Indonesia dry, and even if there is an increase in rainfall, as shown in Fig. 4c, 4d, it is due to the NetIOD condition, which causes an increase in sea surface temperature in the western part of Indonesia. When NetENSO conditions, the sea surface temperature in the Pacific Ocean is also uncertain, as is what happened in Indonesia. However, the concurrent occurrence of NIOD has increased sea surface temperature, which is visible in Indonesia, as shown in Fig. 7c, d. This causes an increase in rainfall in the western part of

Fig. 7 Composite of SSTs anomaly (Shaded; °C) and 850 hPa wind anomaly (vector; m/s) during a LN + NetIOD (NDJFM), b LN + NetIOD (AMJJASO), c EN + NetIOD (NDJFM), d EN + NetIOD (AMJJASO), e NetENSO + NIOD (NDJFM), f NetENSO + NIOD (AMJJASO), g NetENSO + PIOD (NDJFM), and h NetENSO + PIOD (AMJJASO)

Comparison Influencing of El-Nino Southern Oscillation and Indian …

301

Indonesia, as shown in Fig. 5a, 5d. Meanwhile, when NetENSO co-occurs with PIOD, it should also cause drought in Indonesia, as shown in Fig. 7e, 7f, where there is a decrease in sea surface temperature in Indonesia. However, Fig. 7e is not in sync with Fig. 5c, where many areas experience an increase in rainfall; this is possible because of the influence of other phenomena that co-occur as the PIOD or NetENSO. Figure 7h supports the statement of Fig. 5d, which shows the drought in several parts of Indonesia due to a decrease in sea surface temperature in Indonesia.

4 Conclusion From this study, it can be concluded that the effect of LN and NIOD that co-occurs on rainfall variability in Indonesia is more robust in the AMJJASO season than in the NDJFM season, with higher and more evenly distributed intensity in Kalimantan and Java. The effect of LN and PIOD simultaneously on rainfall is more visible in the NDJFM season than in the AMJJASO season. The influence of the LN also seems to be more dominant than the PIOD because, during the PIOD, parts of Indonesia must experience drought. The effect of NIOD on rainfall is more prevalent than EN when it co-occurs both in the NDJFM season and the AMJJASO season, with a significant increase in precipitation in some areas in western Indonesia, especially in Sumatra and Java. The effect of NIOD + EN on rainfall in Java is more critical during AMJJASO than during NDJFM. Events of PIOD and EN that co-occur will cause drought in almost all of Indonesia, and a high decrease in rainfall will occur on the island of Java during AMJJASO. The influence of LN on rainfall is more dominant than NetIOD during NDJFM and AMJJASO, with an increase in precipitation around central and eastern Indonesia. The effect of LN + NetIOD on increasing rainfall in Indonesia is higher during AMJJASO than during NDJFM. The impact of EN, when it co-occurs with NetIOD, on the increase in rainfall in Sumatra is higher during NDJFM than during AMJJASO. However, the effect was more significant in Kalimantan during AMJJASO than during NDJFM. NetENSO conditions will dry in central and eastern Indonesia simultaneously with NIOD during the NDJFM season. However, this did not happen during the AMJJASO season because there was an increase in rainfall in some areas of eastern Kalimantan and almost all of Papua, although with low intensity. The influence of PIOD and NetENSO on drought in Indonesia is quite strong if the two phenomena co-occur during AMJJASO. The patterns of increasing and decreasing sea surface temperature anomalies and wind movements described in Figs. 6 and 7 are in sync and strengthen the analysis of the previous images regarding the influence of ENSO and IOD on rainfall in Indonesia.

302

T. Hadiningrum et al.

Acknowledgements We are grateful for the availability of APHRODITE data which can be accessed at http://aphrodite.st.hirosaki-u.ac.jp/download/ and the ENSO index and IOD index sourced from this link http://www.cpc.ncep.noaa.gov/data/indices/oni.ascii.txt and https://psl.noaa/ gov/data/timeseries/DMI. The availability of NCEP/NCAR Reanalysis data can be downloaded for free from the following link: https://psl.noaa.gov/data/gridded/data.ncep.reanalysis.html. Declaration of Competing Interest The authors report that they have no conflict of interest. All the authors of this paper are the main contributors. They have assisted in processing, analyzing, and interpreting the data and preparing the report as a whole.

References 1. Ramage, C.S.: Role of a tropical “maritime continent” in the atmospheric circulation. Mon. Weather Rev. 96, 365–369 (1968) 2. Yamanaka, M.D.: Physical climatology of Indonesian maritime continent: an outline to comprehend observational studies. Atmos. Res. Elsevier. https://doi.org/10.1016/j.atmosres.2016. 03.017 3. Sturman, A.P., Tapper, N.J.: The weather and climate of Australia and New Zealand Oxford University Press, Melbourne 476 (1996) 4. McBride, J.L.: Indonesia, Papua New Guinea, and Tropical Australia: The Southern Hemisphere summer monsoon. Meteorology of the Southern Hemisphere, Meteor. Monogr. (49), Amer. Meteor. Soc., 89–99 (1999) 5. Slingo, J., Inness, P., Neale, R., Woolnough, S., Yang, G.Y.: Scale interactions on diurnal to seasonal timescales and their relevance to model systematic errors. Ann. Geophys. 46, 139–155 (2003) 6. Yulihastin, E., Febrianti, N.: Impacts of El Nino and IOD on the Indonesian Climate mechanism of air-sea interaction to change of diurnal rainfall over java View project Ais-sea interaction an it’s impact on anomalously-wet dry season view project impacts of El Nino and IOD on the Indonesian Climate. Available: https://www.researchgate.net/publication/323783989 (2009) 7. Lestari, D.O., Sutriyono, E., Sabaruddin, S., Iskandar, I.: Respective influences of Indian Ocean Dipole and El Niño-Southern Oscillation on Indonesian Precipitation. J. Math. Fundam. Sci. 50(3), 257–272 (2018). https://doi.org/10.5614/j.math.fund.sci.2018.50.3.3 8. Luh Putu Trisnawati, N.: Analisis Dampak El-Nino Terhadap Curah Hujan Pada Bulan Basah Dan Kering Di Kintamani (2018) 9. NCEP-NCAR Reanalysis 1, Physical Sciences Laboratory 10. Aditya, F.: Analisis Variabilitas Curah Hujan Di Kalimantan Barat Tahun 1991–2020, Jurnal Meteorologi Klimatologi dan Geofisika 2(3) (2021) 11. As-syakur R.A.: Pola Spasial Hubungan Curah Hujan Dengan ENSO Dan IOD Di Indonesia - Observasi Menggunakan Data TRMM 3B43 (2012) 12. Nur’utami, M., Hidayat, R.: Influences of IOD and ENSO to Indonesian rainfall variability: role of Atmosphere-ocean Interaction in the Indo-pacific Sector Procedia Environ. Sci. (2016) 13. Safril, A.: Rainfall variability study in Kalimantan as an impact of climate change and El Nino. AIP Publishing. 978-0-7354-4064-7 (2020) 14. Kurniadi, A., Wllwr, E., Min, K.S., Seong, M.G.: Independent ENSO and IOD impacts on rainfall extremes over Indonesia. Int. J. Climatol. 41, 3640–3656 (2021) 15. Kusuma D.W., Murdimanto, A., Aden, L.Y., Sukresno, B., Jatisworo, D., Hanintyo, R.: Sea surface temperature dynamics in Indonesia. IOP Conf. Ser. Earth Environ. Sci. 98(1). https:// doi.org/10.1088/1755-1315/98/1/012038

The Distribution and Characteristics of Mesoscale Convective Complex (MCC) and Its Relation with Rainfall During Madden-Julian Oscillation (MJO) Conditions Over Indonesia Trismidianto Abstract One of the atmospheric phenomena that significantly affects Indonesian rainfall is MJO, and the other phenomenon with a different time scale with MJO is MCC which rainfall until causing storms. This study was conducted to the analysis of MCC during MJO. MCCs were identified and tracked for 15 years (2001–2015) over IMC by infrared satellite imagery using an algorithm that combines criteria of cloud coverage, eccentricity, and cloud lifetime. The identification of the MJO phase is carried out using the MJO index from RMM1 and RMM2. The results showed that the existence of MCC can strengthen the influence of the MJO on increasing rainfall in Indonesia. MCC is more common in areas that have a high convective activity, which indicates the presence of MJO. The existence of MCC which occurs concurrently with MJO will increase its influence on the distribution of rainfall and does not even rule out the possibility of causing extreme rainfall in Indonesia. The effect of active MJO followed by the presence of MCC on rainfall in Indonesia is higher when compared to when MJO is not active. Monthly, in phases 2, 3, and 4, MCC mostly appeared in November, while in phases 1 and 5, MCC mainly occurred in March. MCC often appears in January when the MJO is in phases 6 and 7, and MCC is more commonly found in February when the MJO is in phase 8. The largest MCC cloud core size during the mature phase is found in phase 5. MCC with large average size is found in Region A in almost all phases. The average MCC size found has a cloud core size ranging from 100,000 to 200,000 km2 , but the largest sizes above 800,000 km2 are found during phases 3 and 4 of the MJO. The most extended average duration of about 12.1 was found in MCC, which occurred during phase 2 of the MJO. Most of the MCCs that occurred were found on the continents for each MJO phase.

Trismidianto (B) Research Center for Climate and Atmospheric, National Research and Innovation Agency (BRIN), BRIN Kawasan Sains dan Teknologi Samadikun, Cisitu Sangkuriang, Bandung 40135, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_29

303

304

Trismidianto

1 Introduction Indonesia is one of the most active convective regions in the world that affects the global climate system [1]. One of the convective systems that dominate weather and climate in Indonesia is the mesoscale convective system (MCSs). In general, there are three classifications of MCS, namely squall line, which is a type of MCS in the form of liner cluster clouds that usually form along cold fronts. Bow echoes are a type of MCS that is shaped like a bow. Bow echo size is relatively smaller than other types. A special type of MCS is mesoscale convective complexes (MCCs), the most widely observed MCS with the largest size and long life span in the world [2]. The phenomenon of MCC has been studied by many researchers for more than a few decades in the world [3–9]. MCC does not only occur in middle latitudes but also occur in tropical areas, including Indonesia [3]. Trismidianto [10] found that 1028 MCCs occurred in Indonesia during the 15 years from 2001 until 2015, the average duration of the MCCs over IMC was approximate ~ 9.5 h, and the maximum duration was 10 h. Those observed in this study were usually nocturnal and reached a maximum at midnight. Trismidianto et al. [11] stated that MCCs contributed to total rainfall for 15 years over the Indonesian Maritime Continent by up to 20%. The most significant contribution was concentrated over Central Kalimantan, the South China Sea, the Indian Ocean, and Papua Island. MCC can result in bad weather and continuous rain [3, 11, 12]. They also concluded that MCC does not only affect rainfall in the location where the MCC occurs but also affects rainfall in the area around the MCC [11, 12]. The relationship between MCC and other phenomena that have a significant influence on the distribution of rainfall in Indonesia has been explained by several previous researchers. The link between ENSO and MCC has been reported by some earlier researchers, for example, Velasco and Fritsch [13] suggested that El Niño might play a role in MCC activities. Laing and Fritsch [14] also told the possibility of a relationship between El Niño and MCCs. Trismidianto and Satyawardhana [15] stated that the existence of the MCC was also accompanied by increased rainfall intensity at the locations of the MCC occurrences for all ENSO events. Rustiana et al. [16] stated that rainfall distribution during MCC was higher when the events of negative IOD compared to other events, especially in the western maritime of Sumatra until Kalimantan. In addition to the relationship between MCC and ENSO and IOD, several studies have explained the relationship between MCC and MJO, one of which is Aiqiu et al. [17], which states that the influence of MCC on significant rainfall is also supported by the existence of MJO phase 4. The substantial increase in rainfall when phase 4 active MJO is followed by the MCC. However, Aiqiu et al. [17] only took a case study of one MCC that occurred during MJO phase 4. As how the characteristics of MCC during various MJO conditions had not been explained in detail by the previous research; therefore, using MCC data that had been found by Trismidianto [10], this study will explain how the impact of MCC during several MJO conditions against rainfall distribution in Indonesia, due to both of atmospheric phenomena usually cause of floods in Indonesia.

The Distribution and Characteristics of Mesoscale Convective Complex …

305

2 Data and Method MCC, in this study, was identified by inputting the temperature, latitude, and longitude values for each cloud shield pixel obtained from IR to a modified version of a computerized MCC program using MATLAB based on characteristics of MCCs [3]. Maddox [1] stated that MCC has a cloud shield with continuous brightness temperature (TBB)—32° C (in this study converted to 241 K) and must have an area of 100,000 km2 . The interior cold cloud has TBB −52° C (in this study converted to 221 K) and must have an area of 50,000 km2 for over six hours, although the cloud shield does not have to maintain an eccentricity of 0.7 for its entire life cycle. TBB data obtained from Himawari generation infrared satellite imagery consisting of Geostationary Meteorological Satellite (GMS-5), Multi-functional Satellite Imagery Transport Satellite (MTSAT 1R and MTSAT 2), Himawari 8 and Geostationary Operational Environmental Satellites (GOES-9) with a spatial resolution of 0.05° × 0.05°. This data can be downloaded at http://weather.is.kochiu.ac.jp/sat/GAME. The analysis was conducted classifying MCC events in the Asian Cold Monsoon, November to March (NDJFM), in several MJO phases. The interpolated outgoing longwave radiation (OLR) as a proxy for convection obtained from NOAA/NESDIS. The identification of the MJO phase is carried out using the MJO index, which is the daily time series effectively without the need for conventional time filtering and is called real-time multivariate MJO series 1 (RMM1) and 2 (RMM2). RMM1 and RMM2 provide information about the state of the MJO throughout the tropics and have been shown by Wheller and Hendon [18] to be true in all seasons. As long as MJO activity was identified, RMM1 led RMM2 around the fourth cycle, indicating eastward propagation of the MJO along the equator. MJO index data were obtained from the Climate Prediction Center (CPC) Website, NOAA (https://www.cpc.ncep. noaa.gov/products/precip/CWlink/MJO/). The MJO is considered strong (weak) if the amplitude is higher (less) than 1. Furthermore, the MJO event is divided into eight phases, where each phase of the MJO indicates the location of the MJO convective center. The convective center of the MJO propagates eastward from west Africa (phase 1) to the east, passing over the Indian Ocean (phases 2 and 3), MC (phases 4 and 5), migrating to the western Pacific (phases 6 and 7), and decaying at phase 8. However, in this study, we divided active MJO and inactive MJO by defining active and inactive MJO based on Lim et al. [19] and Xavier et al. [20], but there was a slight difference in active MJO by adding phase 5 during strong conditions. Active MJO is defined during phases 2, 3, 4, and 5, which have an amplitude value greater than or equal to 1, while inactive MJO is all days with MJO phases 2, 3, 4, and 5 with an amplitude value less than 1 added all the days when MJO in phases 1 and 5–8. The MCC information is used to explain the MCC characteristics over the three longitude regions (90–105E as Region A, 106–125E as Region B, and 126–140E as Region C) consisting of the area interior size, duration of the life cycle, and interior eccentricity of MCC. In addition to analyzing the characteristics and distribution of MCC during these MJO phases, an analysis of the relationship between MCC and

306

Trismidianto

rainfall was also carried out using estimated rainfall data from GSMaP data [21] from JAXA (http://sharaku.eorc.jaxa.jp/GSMaP/). The spatial distribution of rainfall from GSMaP has spatial and time resolutions of 0.1° × 0.1° and has time observation near real time. 850 hPa wind data obtained from ERA5, which provides hourly estimates of a large number of atmospheric, land, and oceanic climate variables. The data cover the Earth on a 30 km grid and resolve the atmosphere using 137 levels from the surface up to 80 km. ERA5 includes uncertainties for all variables at reduced spatial and temporal resolutions.

3 Result and Discussion 3.1 MJO Climatologies The movement of the MJO that propagates from the Indian Ocean to the Pacific Ocean via Indonesia can be seen in Figs. 1a, b. MJO propagation is identified by the presence of convective activity as indicated by the negative OLR anomaly as shown in Fig. 1a. In phase 1, the MJO begins to occur in the Indian Ocean near India. In phases 2 and 3, MJO is already in the Indian Ocean near Sumatra Island. MJO was already above Indonesia when MJO entered phases 4 and 5. MJO was still in parts of eastern Indonesia, especially on Papua Island when MJO was in phases 6 and 7. MJO had already left Indonesia for the Pacific Ocean during phase 8 of the MJO. This convective activity is followed by an increase in rainfall as shown in Fig. 1b. During phase 1, increased rainfall was seen in the Indian Ocean around India where the MJO started. However, there is an interesting thing that shows an increase in rainfall over the island of Sumatra, even though there is no convective activity in the area. An increase in high rainfall occurred in the Indian Ocean, Sumatra Island, and some parts of Java seen when MJO was in phases 2 and 3. During phases 4 and 5, MJO was above Indonesia which caused an increase in rainfall over Indonesia. Increased rainfall on parts of the island of Papua was seen when the MJO was in phase 6. During phases 7 and 8, there was no visible increase in rainfall over Indonesia because the MJO had moved into the Pacific Ocean leaving Indonesia.

3.2 MCC Distribution During MJO Condition 3.2.1

Distribution of MCC in Each Phase of MCC

There are 508 MCCs that occur in the Asian winter in the month of NDJFM. Figure 2a shows the distribution of MCC events during the phases of the MJO. Increased rainfall occurred around the MCC event. In phase 1 to phase 4, increased rainfall is concentrated in the Indian Ocean and the island of Sumatra where there are many

The Distribution and Characteristics of Mesoscale Convective Complex …

307

Fig. 1 Composites of a OLR anomaly (Wm−1 ) and b rainfall anomaly (mmday−1 ) for eight phases of the MJO cycle for November through March in the period of 2001–2015

MCCs. Increasing rainfall is moving to central and eastern Indonesia as the MJO system moves which are followed by many MCC events around the region. In phases 6 to 8, MJO has left Indonesian territory but there is still an increase in rainfall in Indonesian areas around the location of the MCC. It is estimated that the MCC that occurs during phases 6 to 8 is not caused by the presence of MJO but is possible due to local factors because the MCC that occurs is more dispersed and not concentrated in one place. In MJO phase 1, there were 46 MCCs that occurred in Indonesia, which were concentrated in Region A, around the Indian Ocean near Sumatra, and this was also followed by an increase in rainfall. MCC events in MJO phase 1 were also seen around Regions B and C but on a small scale so the effect on rainfall was not that significant. There were 64 MCCs that occurred in phase 2 of the MJO which were concentrated in the Indian Ocean and Kalimantan Island, but the average MCC size was medium. Similar to phase 1, the MCC in phase 2 of the MJO also affects the rainfall increase in Region A. In phase 2, the distribution of MCC is more on the Sumatra Island and also on the Borneo Island. The impact of MCC on increasing rainfall in Region A is very significant, as seen in MJO phase 3. This is also in accordance with the results of Yulihastin et al. [22] that there is MJO modulation related to the formation of the MCS which is suspected of being MCC in Jakarta. There were 89 MCC events that occurred during phase 3 of the MJO focused on Region A and Region B, around the Indian Ocean near Sumatra Island and above Borneo Island with a large-scale average size. The presence of MCC that occurs in phase 3 of the MJO is believed to be able to increase the influence of MJO on increasing rainfall in Regions A and B, including in Java. Even though there are no MCC events on the island of Java, there are several MCC events that occur around the Java Island. Almost the same as in phase 3, a total of 70 MCC that occurred in phase 4 MJO also affected an increase in rainfall in Regions A and B which had the same pattern in phase 3, but the intensity of the increase in rainfall was smaller than in phase 3. MJO phase 4 is also seen to have an effect on

308

Trismidianto

Fig. 2 Composites of rainfall anomaly in mmday-1 (shaded) during eight phases of the MJO cycle for November through March in 2001–2015 in condition a MCC occur, b no MCC, and c composites of OLR anomaly in Wm-1 (shaded) and 850 hPa wind in ms-1 (vector). The circle refers to the distribution of MCC occur, where the size of the circle indicates the size of the MCC when it occurs in the mature phase

increasing rainfall in parts of Region C, and this is also seen when MCC occurs in phase 4 of the MJO. The effect of MCC and MJO on increasing rainfall in Regions B and C can be seen in phase 5, where there were 52 MCC events which were more widely spread in Regions B and C although there were not many large ones. The growth of several MCCs in the Indian Ocean region can also be affected by MJO, or local scale influences from atmospheric phenomena but this is not discussed in the results of this study. The number of MCCs in MJO phases 6, 7, and 8, respectively, were 76, 56, and 55 events, including a fairly large number, but the average was on a small scale. In phases 6, 7, and 8, MCC is more widely spread in Regions B and C

The Distribution and Characteristics of Mesoscale Convective Complex …

309

but on a small scale. The effect of MCC on increasing rainfall in Indonesia in phases 6, 7, and 8 is not too significant compared to other phases. When compared with Fig. 2b, it can be seen that the increase in rainfall during the MJO followed by the presence of an MCC event is higher than when there is no MCC. This shows that the presence of MCC increases the influence of MJO on the distribution of rainfall in Indonesia. A large number of MCC distributions at the MJO location also indicates that there is MJO modulation of the development of the MCC. In phases 6, 7, and 8, when the MJO is without MCC, there is no significant increase in rainfall in the Indonesian. However, when compared to the MJO event followed by MCC, it can be seen that there was an increase in rainfall in the area around the MCC occurrence in phases 6, 7, and 8. This shows that the MCC that occurs in phases 6, 7, and 8 is not modulated by MJO but is possible due to local factors or diurnal convection, Fig. 2c reinforces the statement in Fig. 2a that the pattern of convective activity around the MCC event in each phase of the MJO is the same as the pattern of increasing rainfall, especially in phases 1–5, this reinforces the statement that the presence of MJO modulates the development of the MCC. The percentage of extreme rainfall events in each MJO phase followed by the presence of MCC and without MCC using the 95th percentile threshold is shown in Fig. 3. From Fig. 3a, it can be seen that extreme rainfall events with a high percentage reaching more than 20% almost all occur around the location of the occurrence of MCC in each phase of MJO. When compared with Fig. 3b, it can be seen that the percentage of extreme rainfall events is not very significant when MJO events are not followed by MCC events that occur in almost all MJO phases. This shows that the presence of MCC that occurs simultaneously with MJO will increase its influence on the distribution of rainfall, and it is even possible that it can cause extreme rainfall in Indonesia.

Fig. 3 Composites of the percentage of extreme rainfall based on percentile 95 during eight phases of the MJO cycle for November through March in 2001–2015 in condition a MCC occur and b no MCC

310

3.2.2

Trismidianto

Distribution of MCC During MJO Active and Inactive

Figure 4 shows the distribution of MCC events when the MJO is active (Fig. 4a), and the MJO is inactive (Fig. 4b). It can be seen that the distribution of MCC events when the MJO is active is indeed less than when the MJO is inactive, but the effect on increasing rainfall intensity is higher when the MJO is active. During the active MJO, the area where the MCC occurs is always followed by a distribution of high rainfall increases, and this is different from when the MJO is inactive. The distribution of the increase in high rainfall is not seen in all areas of the MCC occurrence. This strengthens the analysis result that the presence of an active MJO initiates the occurrence of MCC, but this needs to be continued in a more detailed study related to the MJO mechanism in initiating the occurrence of this MCC. The existence of MCC when the MJO is active strengthens the influence of the MJO in increasing rainfall in the Indonesian Territory. However, MCC is not only initiated by the MJO, and it can also be influenced by other atmospheric phenomena, including local-scale phenomena, which are not discussed in the results of this study. The strengthening of the analysis in Fig. 4c, d is shown in Fig. 4c, d, which shows the same pattern of convective activity as the pattern of increasing rainfall when the MJO is active, and the MJO is inactive. The formation of convective activity when the MJO is active is also shown in Fig. 4c due to the strengthening of the wind. The percentage of extreme rainfall events is very significant, especially around MCC locations when the MJO is active (see Fig. 4e). This high percentage of extreme rainfall is not seen when the MJO is not active as shown in Fig. 4f even though there are many MCC occurrences. This shows that there is an interaction between MJO and MCC in influencing rainfall variability in Indonesia, and it is even possible to cause extreme rainfall.

3.3 Characteristics of MCC During MJO Condition Statistically, as shown in Fig. 5a, it can be seen that most MCC events were found in phase 3, which is one of the active phases of the MJO. Phases 2, 4, and 5 also occur quite a lot of MCC but are still below the number of MCC events during phase 6. There is a large number of MCC events in phase 6, due to the influence of other atmospheric phenomena or local scale phenomena because the MCC formed is primarily small scale. Monthly, in phases 2, 3, and 4, MCC mostly appeared in November, while in phases 1 and 5, MCC mainly occurred in March. MCC often appears in January when the MJO is in phases 6 and 7. MCC also is more commonly found in February when the MJO is in phase 8. Figure 5b shows that MCC is most widely found in Region B in each phase of the MJO, followed by Region A, whereas in region C, the least MCC was found in each phase. Figure 5c shows that, on average, the largest MCC cloud core size during the mature phase is found in phase 5, followed by phases 3 and 4, where the average value is above 150 thousand km2 . MCC with a large average size is found in Region A in almost all phases of the MJO except for phase 6, where the average large size

The Distribution and Characteristics of Mesoscale Convective Complex …

311

Fig. 4 Composites of rainfall anomaly in mmday−1 (shaded) during MCC occur in a MJO active and b MJO inactive. Composites of OLR anomaly in Wm-1 (shaded) and 850 hPa wind in ms-1 (vector) during c MJO active and d MJO inactive. The percentage of extreme rainfall based on percentile 95 during e MJO active and f MJO inactive, for November through March in the period 2001–2015. The circle refers to the distribution of MCC occur, where the size of the circle indicates the size of the MCC magnitude when it occurs mature phase

MCC is found in Region C. Figure 5d shows the distribution of the number of MCC based on size, which is divided into three categories, namely small scale (50,000– 149,999 km2 ), medium scale (150,000–299,999 km2 ), and large scale (> = 300,000 km2 ). The average MCC that occurs in each MJO phase is small scale. Medium-scale MCCs were mostly found in phases 2 and 3, while large MCCs were found during phases 3 and 4 of the MJO. Figure 6a shows that the average MCC size found has a cloud core size ranging from 100,000–200,000 km2 , but the largest sizes above 800,000 km2 are found during phases 3 and 4 of the MJO, and some large sizes above 500,000 km2 are also found during phases 1, 6, and 8 MJO. We define the MCC duration as the duration between the time of the initial phase and the time of the dissipation phase based on the previous research [10]. The frequency and mean of the distribution of the duration of all MCC during every phase of MJO are shown in Fig. 6b. The average duration of MCC events in each phase was about 11 h, but the most extended average duration of about 12.1 was found in MCC during phase 2 of the MJO. Short-lived MCC often occurs in phase 6 of the MJO, and medium-lived MCC is usually found during phase 3 of the

312

Trismidianto

Fig. 5 Number of MCC events in a every month and b every region during MJO in its eight phases for November through March in 2001–2015. Area size of interior cloud of MCC events during the mature phase in c every region and d based on size category during MJO in its eight phases for November through March in 2001–2015

MJO, while long-lived MCC often occurs when the MJO is in phase 2. The average duration of MCC events is longer than MCC overall. Trismidianto [10] found that the mean duration was approximate ~ 9 0.5 h, and the maximum duration is 10 h. The distribution and average duration are significantly larger than the distribution and average for MCCs in southern Africa (~9.5 h; [23]). However, it is slightly shorter than the global average (~10 h; [14]) and much less than those found in subtropical South America (14 h; [24]). In most of the MCCs, around 686 events occurred during 8–12 h. Only 20 events are the long-lived MCCs that occurred over 20 h. In this study, we defined the oceanic (continental) MCCs as a system that reaches the maximum extent while positioned over the ocean (land) as the following method by Morel and Senesi [25] and Blamey and Reason [23]. This research follows the method by Ogino [26] and Mori [27] to define the coastal region. The coastal MCC is defined as the MCCs that occur near the coast between the ocean and the continent during MCC reaching maximum extent or mature stage. The frequency of MCC occurrences on the coast, continent, and ocean in this study is shown in Fig. 6c. Most of the MCCs that occurred were found on the continents for each MJO phase, with the highest number in phase 3 and phase 6. The number of coastal MCC and ocean MCC events mostly occurred during phase 3 MJO.

4 Conclusion The increase in rainfall during the MJO followed by the presence of an MCC event is higher than when there is no MCC. The increase in rainfall is also higher in the

The Distribution and Characteristics of Mesoscale Convective Complex …

313

Fig. 6 a Box and plot whisker for area size of interior cloud of MCC events during the mature phase in in every phase of MJO for November through March in 2001–2015. Frequency distribution of duration time of MCCs, b based on duration category and c over the coastal, continent, and ocean during MJO in its eight phases for November through March in 2001–2015

presence of MCC during active MJO than when MJO is not active. The percentage of extreme rainfall events is not very significant when MJO events are not followed by MCC events that occur in almost all MJO phases. However, extreme rainfall events often occur around the MCC event location in each phase of the MJO. This shows that the existence of MCC which occurs concurrently with MJO will increase its influence on the distribution of rainfall and does not even rule out the possibility of causing extreme rainfall in Indonesia. The effect of active MJO followed by the presence of MCC on rainfall in Indonesia is higher when compared to when MJO is not active. This shows that the presence of MCC increases the influence of MJO on the distribution of rainfall in Indonesia. A large number of MCC distributions at the MJO location also indicates that there is MJO modulation of the development of the MCC. In phases 6, 7, and 8, when MJO is without MCC, there is no significant increase in rainfall in Indonesia. However, when compared to the MJO event followed by MCC, it can be seen that there is an increase in rainfall in the area around the MCC event in phases 6, 7, and 8. This indicates that the MCC that occurs in phases 6, 7, and 8 is not modulated by MJO but is possible due to local factors or diurnal convection. Monthly, in phases 2, 3, and 4, MCC mostly appeared in November, while in phases 1 and 5, MCC mostly occurred in March. MCC often appears in January when the MJO is in phases 6 and 7, and MCC is more commonly found in February when the MJO is in phase 8. The largest MCC cloud core size during the mature phase is found in phase 5, followed by phases 3 and 4, where the average value is above 150 thousand km2 . MCC with a large average size is found in Region A in almost all phases of the MJO except for phase 6 where the average large size MCC is found in Region C. the average MCC size found has a cloud core size ranging from 100,000–200,000 km2 , but the largest sizes above 800,000 km2 are found during

314

Trismidianto

phases 3 and 4 of the MJO. The average duration of MCC events in each phase was about 11 h, but the longest average duration of about 12.1 was found in MCC that occurred during phase 2 of the MJO. The frequency of MCC occurrences on the coast, continent, and ocean in this study is shown in Fig. 6c. Most of the MCCs that occurred were found on the continents for each MJO phase, with the highest number in phase 3 and phase 6. The number of coastal MCC and ocean MCC events mostly occurred during phase 3 MJO. Acknowledgements Thank you so much for the three sources of data, which we can get for free, and the data are beneficial to this research. This research uses data TBB from Himawari generation data (http://weather.is.kochiu.ac.jp/sat/GAME). MJO index data were obtained from the Climate Prediction Center (CPC) Website, NOAA (https://www.cpc.ncep.noaa.gov/products/precip/CWl ink/MJO/), and GSMaP satellite (http://sharaku.eorc.jaxa.jp/GSMaP/).

References 1. Ramage, C.S.: Role of a tropical “maritime continent” in the atmospheric circulation. Mon. Weather Rev. 96, 365–369 (1968) 2. Houze, R.A.Jr. Mesoscale convective system. Rev. Geophys, American Geophisical Union, 43 (2004) 3. Maddox, R.A.: Mesoscale convective complexes. Bull. Amer. Meteor. Soc. 61, 1374–1387 (1980) 4. Miller, D., dan Fritsch, J.M.: Mesoscale convective complexes in western pacific region. Amer. Meteor. Soc. 119, 2978–2992 (1991) 5. Laing, A.G, dan Fritsch, J.M.: Mesoscale convective complexes over the Indian monsoon region. Am. Meteorol. Soc. 121, 2254–2263 (1993) 6. Velasco, I., dan Fritsch, J.M.: Mesoscale convective complexes in the America. J. Geophys.. Res. 92, 9591–9613 (1987) 7. Durkee, J.D., dan Mote, T.L.: A climatology of warm season mesoscale convective complexes in subtropical South America. Int. J. Clim., 12 (2009) 8. Yuan, J., Houze, R.A.: Global variability of mesoscale convective system anvil structure from A-Train satellite data. J. Clim. 23, 5864–5888 (2010). https://doi.org/10.1175/2010JCLI3671.1 9. Trismidianto.: A study of mesoscale convective complexes (MCCs) activities in the western Indian Ocean and their effects on convection over Sumatra Island. Tesis Magister, FITB, ITB (2012) 10. Trismidianto.: The global population of mesoscale convective complexes (MCCs) over Indonesian Maritime Continent during 15 Years IOP Conf. Ser.: Earth Environ. Sci. 166, 012040, 1–17 (2018). https://doi.org/10.1088/1755-1315/166/1/012040 11. Trismidianto, Yulihastin, E., Satyawardhana, H., Nugroho, J.T., Ishida, S.: The contribution of the mesoscale convective complexes (MCCs) to total rainfall over Indonesian Maritime Continent. IOP Conf. Ser. Earth Environ. Sci. 54 (2017). https://doi.org/10.1088/1755-1315/ 54/1/012027 12. Trismidianto, Hadi, T.W., Ishida, S., Moteki, Q., Manda, A., Iizuka, S.: Development processes of oceanic convective systems inducing the heavy rainfall over the western coast of Sumatra on 28 October 2007. Sci. Online Lett. Atmos. (SOLA). 12, 6−11 (2016). https://doi.org/10. 2151/sola.2016-002 13. Velasco, I., Fritsch, J.M.: Mesoscale convective complexes in the America. J. Geophys. Res. 92, 9591–9613 (1987)

The Distribution and Characteristics of Mesoscale Convective Complex …

315

14. Laing, A.G., Fritsch, J.M.: The global population of mesoscale convective complexes. Q. J. R. Meteorol. Soc. 123, 389–405 (1997) 15. Trismidianto, Satyawardhana, H.: Mesoscale convective complexes (MCCs) over the Indonesian maritime continent during the ENSO events IOP Conf. Ser. Earth Environ. Sci. 149 012025, 1–10. https://doi.org/10.1088/1755-1315/149/1/01202 16. Rustiana, S., Trismidianto, Satyawardhana, H.: The influence of ENSO and IOD during mesoscale convective complex (MCC) to rainfall in Indonesia. IOP Conf. Ser. Earth Environ. Sci. 303 012006 (2019). https://doi.org/10.1088/1755-1315/303/1/012006 17. Aiqiu, L.F., Zakir, A., Trismidianto.: Analysis of mesoscale convective complex during madden Julian oscillation phase 4 (case study: heavy rain in Cilacap on Sept 16–17, 2016) IOP Conf. Ser. Earth Environ. Sci. 303, 012063 (2019). https://doi.org/10.1088/1755-1315/303/1/012063 18. Wheeler, M.C., Hendon, H.H.: An all-season real-time multivariate MJO index: development of an index for monitoring and prediction. Mon. Weather Rev. 132(8), 1917–1932 (2004) 19. Lim, S.Y., Marzin, C., Xavier, P., et al.: Impacts of boreal winter monsoon cold surges and the interaction with MJO on Southeast Asia rainfall. J. Clim. 30, 4267–4281 (2017) 20. Xavier, P., Rahmat, R., Cheong, W.K., Wallace, E.: Influence of Madden-Julian Oscillation on South East Asia rainfall extremes—observations and predictability. Geophys. Res. Lett. 41, 4406–4412 (2014) 21. Setiawati, M.D., Miura, F.: Evaluation of GSMaP daily rainfall satellite data for flood monitoring: case study—Kyushu Japan. J. Geosci. Environ. Prot. 4, 101–117. ISSN Online: 2327-4344 (2016) 22. Yulihastin1, E., Trismidianto, Satyawardhana, H., Nugroho, G.A.: MJO modulation on diurnal rainfall over West Java during pre-monsoon and strong El Niño periods. IOP Conf. Ser. Earth Environ. Sci. 54 (2017). https://doi.org/10.1088/1755-1315/54/1/012029 23. Blamey, R.C., Reason, C.J.C.: The role of mesoscale convective complexes in southern Africa summer rainfall. J. Clim. 26 (2012) 24. Durkee, J. D., Mote, T.L., Shepherd, M.J.: The contribution of mesoscale convective complexes to rainfall across, subtropical South America. Int. J. Clim. 22 (2009) 25. Morel, C., Senesi, S.: A climatology of mesoscale convective systems over Europe using satellite infrared imagery II: characteristics of European mesoscale convective systems Q. J. R. Meteorol. Soc. 128, 1973–1995 (2002) 26. Ogino, S.Y., Yamanaka, M.D., Mori, S., Matsumoto, J.: How much is the precipitation amount over the tropical coastal region? J. Clim. 29, 1231–1236 (2016) 27. Mori, S., Hamada, J.I., Tauhid, Y.I., Yamanaka, M.D., Okamoto, N., Murata, F., Sakurai, N., Hashiguchi, H., Sribimawati, T.: Diurnal land-sea rainfall peak migration over Sumatra Island, Indonesian Maritime Continent, observed by TRMM satellite and intensive rawinsonde soundings Mon. Wea. Rev 132, 2021–2039 (2004)

Diurnal Rainfall Pattern in Riau Islands as Observed by Rain Gauge and IMERG Data Ravidho Ramadhan , Helmi Yusnaini, Marzuki Marzuki , Zahwa Vieny Adha, Mutya Vonnisa, and Robi Muharsyah

Abstract Diurnal rainfall is a dominant local phenomenon in the maritime continent due to the land–sea interaction. Factors controlling rain on small tropical islands may differ from large islands, resulting in the difference in the diurnal rainfall pattern. In this work, we investigated the diurnal rainfall in small islands of Riau Islands using automatic rain gauge (ARG) and Integrated Multi-satellite Retrieval for GPM (IMERG) data from December 2015 to October 2019. Diurnal rainfall in Riau Islands was observed in terms of precipitation amount (PA), precipitation frequency (PF), and precipitation intensity (PI). Mean PA, PF, and PI values from gauge observations show variation values, that is, 0.16–0.35 mm/h (PA), 1.67–3.20% (PF), and 7.04– 19.11 mm/h (PI). Furthermore, the values of PA, PF, and PI also show different peaks time. The peak time of PA by rain gauge and IMERG observation occurs at 1100– 1300 LST (Batam Island), 0300–0500 LST (Singkep Island), and 1100–1400 LST (Bintan Island), the PF peak occurs at 1100–1300 LST (Batam and Bintan Island) and 0500–0700 LST (Singkep Island), and the peak of PI tends to be fluctuated. This precipitation system’s peak time is also dependent on rain duration. IMERG data also observed this pattern of PA, PF, and PI. IMERG data captured rainfall peaks migration from the ocean to the prominent islands in the Riau Islands, such as Batam, Bintan, and Lingga Islands.

1 Introduction The Riau Islands is a province in Indonesia that is located in the northeast of Sumatra Island. It consists of more than 2400 islands with a total area of around 8200 km2 [1]. R. Ramadhan · H. Yusnaini · M. Marzuki (B) · Z. V. Adha · M. Vonnisa Department of Physics, Universitas Andalas, Padang 25163, Indonesia e-mail: [email protected] R. Muharsyah Agency for Meteorology, Climatology and Geophysics of Republic Indonesia, Jakarta, Indonesia

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_30

317

318

R. Ramadhan et al.

The number of islands is the largest among all provinces in Indonesia. Eventually, most of the population is focused on a few main islands, such as Batam and Bintan Islands. The population density in the Riau Islands is ranked 10th in Indonesia, with a population density of 258 people/km2 [2]. The geographical condition of the Riau Islands, which is surrounded by oceans and located in the tropical region, causes hydrometeorological disasters that are very potential to occur over there. For example, hydrometeorological disasters are the only cause of natural disasters in the Riau Islands in 2021 [3]. It is composed of floods and cyclones that impacted almost 10 thousand residents. Thus, adequate government planning is needed to minimize the impact of the hydrometeorological disaster. Adequate planning can only be realized if detailed information about rainfall patterns is discovered. One of the most local dominant phenomena of rainfall on tropical islands is diurnal cycles. Diurnal cycles are the result of strong land–sea interactions related to the land–sea breeze cycle [4]. A previous study was carried out to investigate the diurnal pattern of rainfall over the Indonesian Maritime Continent (IMC) [5–10]. However, most of these studies focus on the diurnal pattern on the big islands (>2000 km2 ) in Indonesia, such as Sumatra, Java, Borneo, and Papua. Study on diurnal rainfall patterns on small islands such as the Riau Islands is still minimal. Thus, this study aims to investigate the pattern of diurnal rainfall in the Riau Islands region to analyze the diurnal characteristics of rainfall on small islands in the IMC.

2 Data and Methods This study focuses on the most densely populated area of the Riau Islands Province. The regencies included in this study are Karimun, Bintan, and Lingga regencies. These regencies also included two municipalities that are Batam and Tanjung Pinang. In the observation area, there are five prominent islands with the largest size, namely Batam, Bintan, Lingga, Singkep, and Kundur Islands. Although these islands are relatively more extensive than others, these islands are not more than 1000 km2 in size. Details of the arrangement of the islands can be seen in Fig. 1. Diurnal rainfall pattern observations in Riau Islands were analyzed using four automatic rain gauges (ARG) from the Meteorology, Climatology and Geophysics Agency (BMKG) station from December 2015 to October 2019. The distribution of the stations used can be seen in Fig. 1. One of the stations is located on Singkep Island (ARG 01), one station is on Batam Island (ARG03), and two stations are located on Bintan Island (ARG02 and ARG04). ARG records rain observations every 10 min with an observation threshold of 0.2 mm, so that rainfall below 0.2 mm is not counted. Furthermore, the spatial pattern of diurnal rainfall is analyzed using grid rainfall data from satellite precipitation products (SPPs). SPPs data used in this study is Integrated Multi-Satellite Retrieval for GPM (IMERG) V06 data with a resolution of 0.1−30 min. IMERG data was used from 2015 until 2020. The accuracy and capability of the IMERG V06 data have been evaluated for the IMC region [11–13].

Diurnal Rainfall Pattern in Riau Islands as Observed by Rain Gauge …

319

Fig. 1 Position and topography of Riau Islands (a) and location of ARG (b)

The diurnal rainfall pattern in Riau Islands is characterized by hourly precipitation amount (PA), precipitation frequency (PF), and precipitation intensity (PI) [5, 6]. First, PA is defined as the rate of accumulation of rainfall during the entire observation period to the total observation (mm/h). Second, PF is defined as the ratio of the number of counted rainfall events to total epoch data (%). Third, PI is defined as the ratio of accumulated rainfall to number of hourly data that counted as rainfall events (mm/h). A more detailed definition of the method of calculating PA, PF, and PI can be seen in Marzuki et al. [6].

320

R. Ramadhan et al.

The PA, PF, and PI values will be considered for each ARG point and the IMERG grid for every hour (0–23) in the Local Time Standard/LST unit (UTC + 7). Moreover, this hourly value of PA, PF, and PI can be used to determine the peak time of the diurnal rainfall pattern in the Riau Islands. Furthermore, this study also observed the effect of rain duration on the diurnal rainfall pattern in the Riau Islands, Indonesia. The rainfall duration is divided into short-duration rainfall (6 h).

3 Result and Discussions The amount of annual rainfall in the Riau Islands is influenced by the duration rainfall category. The percentage of each duration rainfall category for the total rainfall events from ARG and IMERG data can be seen in Fig. 2. All ARGs show a dominant percentage of short-duration rainfall (>80%). On the other hand, the percentage of long-duration rainfall is very rare ( 20% to total annual rainfall in Melbourne, Sydney, and Brisbane and global areas [16] when taking into account the median. Consistent with the previous study [16], Jakarta and Kalimantan have a typical peak of rainy season occurring in Dec–Feb and the highest wday and REs also take place in this season. Figure 4 shows the variations of frequency in wday, R95p, R99p seasonally, and it is found that for the Jakarta region, there is a clearly distinguishable frequency between rainy (Dec–Feb) and dry (Jun–Aug) seasons of about 30% (wday), 2% (R95p), and 0.5% (R99p). Meanwhile, for the Kalimantan, the difference frequency is nearly 15% (wday), 0.7% (R95p), and 0.2% (R99p). This result illustrates that frequency for both wdays and REs over Jakarta is more strongly affected

330

S. Lestari et al.

Fig. 3 Whisker and box plots representing distributions of a, b wday, c, d R95p, and e, f R99p over Jakarta and Kalimantan during 1980–2008. Box displays range of interquartile with the minimum and maximum are shown in whiskers. Horizontal black line shows the median, and green triangle shows the mean. a, c, e shows frequency of wday, R95p, R99p (frequency, %), while b, d, f presents relative contribution of wday, R95p, R99p to total rainfall (contribution, %)

by the monsoonal system compared to Kalimantan. The monsoonal system has also regulated the pattern in REs with the highest peak which occurs in Jan–Feb (Jakarta) and Nov–Dec (Kalimantan), while the lowest peak exists in Jun–Aug (Jakarta) and Sep (Kalimantan). The most striking result is that during the dry seasons (Jun-Sep), the REs over Kalimantan are almost doubled in frequency compared to REs over Jakarta in the same season.

3.2 Large-Scale Conditions Vertical velocity is analyzed to understand synoptic environment that might be favorable to initiate REs development. Figure 5 illustrates the 500-hPa vertical velocity (hereinafter referred to as 500-omega) selected from the nearest grids of locations of rainfall stations at the same time as the occurrence of rainfall event (median, R95,

Comparison of Statistical Properties of Rainfall Extremes Between …

331

Fig. 4 Seasonal variations of frequency (%) in wday, R95p, and R99p over a Jakarta and b Kalimantan regions during 1980–2008

R99p). The negative values of 500-omega indicate the upward motion of the air, while positive 500-omega shows the downward motion of the air. Median rainfall, R95p, and R99p generally occur corresponding to local condition of raising air at around 0 and −0.4 Pa S−1 at Jakarta and Kalimantan (Fig. 5). Spatial distributions of 500-omega show that the strong ascent associated with R95p also exists on Jakarta and Kalimantan, with a stronger ascent found over the mountainous region (Figures not shown). In addition to 500-omega, speed and direction of 850-hPa winds indicate that strong westerly wind (2 ms−1 ) is predominant over Jakarta (Fig. 6a). But, over Kalimantan, wind direction varies with the westerly wind occurs in the west, southeasterly wind in the north and southeast, while southwesterly in the northeast with the highest wind speed exists in the northeast (2 ms−1 ) (Fig. 6b). Strong westerly and southwesterly winds might induce convergence over the central and eastern part of Kalimantan responsible for the development REs. A different wind direction between Jakarta and

Fig. 5 Scatter plots between median rainfall, R95p, and R99p and 500-omega from station datasets and ERA-5 analysis, respectively, over a Jakarta b and Kalimantan during 1980–2008. The negative values of 500-omega indicate ascent, while positive values show descent

332

S. Lestari et al.

Fig. 6 Spatial maps of speed and direction of the 850-hPa wind from ERA-5 reanalysis, over (a) Jakarta (b) and Kalimantan during 1980–2008. The magnitude of wind speed is shown in the filled contour (ms−1 ) while the wind direction is indicated by the vector

Kalimantan is likely related to “dipole pattern” induced by complex local–regional physical processes as reported by the previous studies [18, 19]. However, incorporating a comprehensive analysis of physical background responsible for the different features of REs over both regions shall be done in future studies.

4 Conclusion This study has shown characteristics of distribution in rainy days (wday) and rainfall extremes of the 95th percentile (R95p) and the 99th percentile (R99p) analyzed from observational datasets during 1980–2008. Results indicate that in the Megacity Jakarta, maximum intensity of 95th and 99th percentile of rainfall is 65 and 120 mm, respectively, and is larger compared to Kalimantan (62 and 100 mm, respectively). This indicates more intense rainfall happens in Jakarta and its surrounding regions (Fig. 2). However, compared to Kalimantan, over Jakarta, the frequency of wday, R95p, and R99p shows a lower percentage implying that Jakarta has less frequent but higher intensity of rainfall and REs rather than Kalimantan (Fig. 3). Yet, although Jakarta has a lower frequency of REs, relative contribution of R95p to total rainfall is high (25%) (Fig. 3d). Over Jakarta, there is a distinct seasonal feature of wday shown by the different frequency between wet and dry seasons of about 30% (Fig. 4a). The most notable result is that during the dry seasons (Jun-Sep), the REs over Kalimantan are almost doubled in frequency compared to REs over Jakarta. Rainfall extremes are apparent to be initiated by a large-scale pattern of upward motion of air in the vicinity area in Jakarta and Kalimantan (Fig. 5). However, features of wind speed and directions are different between those two regions, with the strong westerly wind prevailing over Jakarta and its surrounding regions. Meanwhile, the strong westerly wind and southwesterly winds exist in the western and northeastern parts of Kalimantan, respectively. The convergence area of wind seems to play a role in generating REs over the Kalimantan region. However, this is a preliminary

Comparison of Statistical Properties of Rainfall Extremes Between …

333

study. Further research shall explore a more detailed explanation of the underlying mechanism which causes the different features of REs over the Megacity Jakarta and Kalimantan. Acknowledgements This research has been supported by the National Research and Innovation Agency under the Cruise Day Facilitation Program with grant number 376/II/FR/3/2022 and RIIM program Batch 1 with grant number 65/II.7/HK/2022. We thank the Indonesian Agency for Meteorology, Climatology, and Geophysics (BMKG) for providing rainfall datasets.

References 1. Trenberth, K.: Changes in precipitation with climate change. Climate Res. 47(1), 123–138 (2011) 2. Merz, B., Basso, S., Fischer, S., Lun, D., Blöschl, G., Merz, R., ... Schumann, A.: Understanding heavy tails of flood peak distributions. Water Resour. Res. 58(6), (2022) 3. Allan, R.P., Soden, B.J.: Atmospheric warming and the amplification of precipitation extremes. Science 321(5895), 1481–1484 (2008) 4. Groisman, P.Y., Knight, R.W., Easterling, D.R., Karl, T.R., Hegerl, G.C., Razuvaev, V.N.: Trends in intense precipitation in the climate record. J. Clim. 18(9), 1326–1350 (2005) 5. O’Gorman, P.A., Schneider, T.: The physical basis for increases in precipitation extremes in simulations of 21st-century climate change. Proc. Natl. Acad. Sci. 106(35), 14773–14777 (2009) 6. White, B.A., Jakob, C., Reeder, M.J.: Fundamental ingredients of Australian rainfall extremes. J. Geophys. Res.: Atmosph. 127(17) (2002) 7. Ramage, C.S.: Role of a tropical “maritime continent” in the atmospheric circulation. Mon. Weather Rev. 96(6), 365–370 (1968) 8. O’Gorman, P.A., Schneider, T.: The physical basis for increases in precipitation extremes in simulations of 21st-century climate change. Proc. Natl. Acad. Sci. 106, 14773–14777 (2009) 9. Lestari, S., King, A., Vincent, C., Karoly, D., Protat, A.: Seasonal dependence of rainfall extremes in and around Jakarta, Indonesia. Weather and Climate Extremes 24 (2019) 10. Lestari, S., Protat, A., Louf, V., King, A., Vincent, C., Mori, S.: Subdaily rain-rate properties in western Java analyzed using C-band Doppler radar. J. Appl. Meteorol. Climatol. 61(9), 1199–1219 (2022) 11. Lockhoff, M., Zolina, O., Simmer, C., Schulz, J.: Representation of precipitation characteristics and extremes in regional reanalyses and satellite-and gauge-based estimates over western and central Europe. J. Hydrometeorol. 20(6), 1123–1145 (2019) 12. Alexander, L.V., Fowler, H.J., Bador, M., Behrangi, A., Donat, M.G., Dunn, R., Funk, C., Goldie, J., Lewis, E., Rogé, M.: On the use of indices to study extreme precipitation on sub-daily and daily timescales. Environ. Res. Lett. 14, 125008 (2019) 13. Abidin, H.Z., Andreas, H., Gumilar, I., Fukuda, Y., Pohan, Y.E., Deguchi, T.: Land subsidence of Jakarta (Indonesia) and its relation with urban development. Nat. Hazards 59(3), 1753–1771 (2019) 14. Zhang, C.: Borneo–a future gem of Asia-Pacific. In: The 33rd Annual Pan-Pacific Conference PPBA (2016) 15. WMO.: Guidelines on analysis of extremes in a changing climate in support of informed decisions for adaptation. World Meteorological Organization (2009). Accessed 19 Sept 2020. https://library.wmo.int/doc_num.php?explnum_id=9419 16. Warren, R.A., Jakob, C., Hitchcock, S.M., White, B.A.: Heavy versus extreme rainfall events in southeast Australia. Q. J. R. Meteorol. Soc. 147(739), 3201–3226 (2021)

334

S. Lestari et al.

17. Aldrian, E., Dwi Susanto, R.: Identification of three dominant rainfall regions within Indonesia and their relationship to sea surface temperature. Int. J. Climatol. 23, 1435–1452 (2003) 18. Qian, J.H., Robertson, A.W., Moron, V.: Diurnal cycle indifferent weather regimes and rainfall variability over Borneoassociated with ENSO. J. Clim. 26(5), 1772–1790 (2013) 19. Kurniadi, A., Weller, E., Min, S.K.: Seong, and Min-Gyu: independent ENSO and IOD impacts on rainfall extremes over Indonesia. Int. J. Climatol. 41(6), 3640–3656 (2021)

Analysis of Mesoscale Convective Complex (MCC) that Often Appears on the Northern Coast of the Borneo Island Trismidianto

Abstract Mesoscale convective complex (MCC) was portrayed as an organized ensemble of convective elements, whose life cycle is longer than that of the individual convective parts and the largest of the convective storms. MCC was identified by using the program in MATLAB by inputting the temperature, latitude, and longitude values for each cloud shield pixel obtained from brightness temperature (TBB) data based on characteristics of MCC. The data used are a combination of satellite data and reanalysis data. This study analyzes 29 MCC events frequently found in the same location on the Northern Coast of Borneo Island. The study results show that MCCs that occur on the Northern Coast of Borneo Island on average have a long-life span and tend to be nocturnal MCC. The development of the MCC on the Northern Coast of Borneo Island was supported by the presence of surface convergent winds and land breezes. The new convective system began to appear during which the MCC decayed and migrated westward after the MCC dissipated. The MCC on the Northern Coast of Borneo Island has an impact on rainfall, and the MCCs in this region are possibly one of the triggers to the development of the other convective system over the South China Sea. From this study result, if MCC occurs on the Northern Coast of Borneo Island, we can predict its movement and influence.

1 Introduction One of the convective systems that dominates weather and climate over the Indonesian Maritime Continent (IMC) is the mesoscale convective systems (MCS). MCS is a complex of organized storm clouds on a larger scale than ordinary/individual storms but smaller thanextratropical cyclones and usually lasts for a few hours or Trismidianto (B) Research Center for Climate and Atmospheric, National Research and Innovation Agency (BRIN), Bandung 40173, Indonesia e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_32

335

336

Trismidianto

more [1]. A particular type of MCS is mesoscale convective complexes (MCCs), and the most widely observed MCS with the largest size and long-life span globally [2]. The phenomenon of MCC has been studied by the previous researchers for more than a few decades in the world [2−7]. MCC over IMC has been also studied by the previous studies [8−13]. MCC plays a very active role in the distribution of energy, momentum, and mass in the atmosphere and occurs over a broad spectrum of time and space scales [12, 14]. MCCs have their organization; they are not passive components in atmospheric flow but interact with large-scale dynamics [15−17]. MCC triggers were strongly correlated with orography, and maximums of MCC triggers were all observed near ridges [8−10, 11]. Trismidianto [8] found that there were 1028 MCCs that occurred during the period 2001–2015 at IMC. Of the1028 MCCs found, there were nine areas as favorite locations occurring of MCCs. Some are above the oceans and coastal areas, and others are above the inland areas and highlands [8, 10, 11]. The nine areas are the Indian Ocean near Sumatra Island, along the West Coast of Sumatra Island, South China Sea on the Northern Coast of Borneo Island, Central Kalimantan around 3°S– 1°N; 111°–115°E, East Kalimantan, around 1°–4°N; 116°–120°E, Makassar Strait near the coast of Kalimantan Island, Central Sulawesi about 4°–2°LS; 120°–123°E, Merauke above Papua Island at locations around 8°–5°S and 135.5°–140°E, and Cendrawasih Bay near the coast of Papua Island at around 132.5°–136.5°E and 4°–1°S [8, 10−11]. Of the nine favorite areas for MCC occurrence, several have been discussed in the previous paper by Trismidianto [8, 11−13]. For example, Trismidianto et al. [12] have discussed the development of MCC in Central Kalimantan, but the discussion has focused more on analyzing the atmospheric conditions when MCC occurred in the region. Trismidianto [13] has described the characteristics and mechanisms of MCC events in the Indian Ocean around Sumatra. While the characteristics of MCC that occur in the South China Sea on the North Coast of Borneo Island have never been discussed, the characteristics and movement mechanisms of MCC in each location are often different. So, it is interesting to know the characteristics and development and movement of MCC in the South China Sea on the North Coast of Borneo Island using a composite analysis. This is expected to be able to find out how the influence of the MCC on the rainfall distribution around it so that it can be an analysis of extreme rainfall predictions when the MCC occurs at the same location.

2 Data and Method MCC was identified by inputting the temperature, latitude, and longitude values for each cloud shield pixel obtained from the equivalent black body temperature (TBB), derived from the hourly infrared data to a modified version of a computerized MCC program using MATLAB based on characteristics of MCCs [2]. The MCC information is used to explain the characteristic of MCC by analyzing the number of MCC occurrences, the duration time of the MCC life cycle, the area size of the

Analysis of Mesoscale Convective Complex (MCC) that Often Appears …

337

cloud, and eccentricity and analyzing the evolution characteristics of MCC. TBB data obtained from Himawari generation infrared satellite imagery consisting of Geostationary Meteorological Satellite (GMS-5), Multi-functional Satellite Imagery Transport Satellite (MTSAT 1R and MTSAT 2), and Himawari 8 and Geostationary Operational Environmental Satellites (GOES-9) with a spatial resolution of 0.05° × 0.05°. This data can be downloaded at https://www.data.jma.go.jp/mscweb/en/him awari89/cloud_service/cloud_service.htm. The estimated rainfall data, corresponding to the MCCs in this research, were obtained from the real-time TRMM TMPA (TMPA-RT) 3B41RT v7 dataset, which has hourly temporal resolution and 0.25° × 0.25° spatial resolution. (Huffman, J.G. 2013). These data could be downloaded from ftp://disc2.nascom.nasa.gov/data/ TRMM/Gridded/3B41RT/. Using the spatial analysis method, we analyze surface wind and atmospheric conditions using ECMWF (European Center for MediumRange Weather Forecasts) ERA5 and CCMP data. ERA5 provides hourly estimates of a large number of atmospheric, land, and oceanic climate variables. The data cover the Earth on a 30 km grid and resolve the atmosphere using 137 levels from the surface up to 80 km. ERA5 includes uncertainties for all variables at reduced spatial and temporal resolutions [18]. The surface wind data were obtained from the CCMP, which covers the global ocean for 20 years with 6-hourly temporal resolution and 25-km spatial resolution. The dataset is produced using a variational analysis method to combine extensive cross-calibrated multiple satellite datasets with in situ data analyzes [19]. To explain the similarities and consistency of the evolution and propagation of the MCCs in the other several regions, this research is utilized composite analysis adapted from McAnelly and Cotton [20−21]. We create a composite region by taking the region that is most frequently concentrated of MCCs that occurred in the South China Sea on the Northern Coast of Borneo Island, as shown in Fig. 1a in symbol SCS. The region which was chosen also has the grid locations of the MCC center, and the duration of the time life cycle is almost the same by taking the point of the life cycle of the nearest hour. The example composite analysis approach to the evolution of the MCC development is illustrated in Fig. 1b, where the timing of the TBB-defined life cycle characteristics is indicated along an hourly time axis for each MCC. In this study, analysis was also carried out on four important stages, which are also referred to as critical stages in the growth of MCC, i.e., initial, mature, decay, and dissipation or post-MCC stage, following the previous research, Cotton et al. [22], which have defined eight stages in the life cycle. Of an MCC: MCC-12 h, preMCC, initial, growth, mature, decay, dissipation, and post-MCC; however, the most important period for an MCC is four of the critical stages.

338

Trismidianto

Fig. 1 a Geographical distribution of the MCCs in the IMC during 15 years (2001–2015). Locations are for the MCC at the time of the maximum extent of the interior cloud size area when the mature stage, source of figure from [8, 10, 11]. b Illustrated of the stratification the MCC composite. The time shows in LT

3 Result and Discussion 3.1 MCC Characteristics A total of 29 MCCs were found to occur frequently in the same area in the South China Sea on the Northern Coast of Borneo Island. Variations in actual MCC totals per month, as shown in Fig. 2a, are reported with a prominent peak of MCC activity in October and followed by January, April, and June. From the results of Trismidianto [8], it was found that many MCCs occurred in April and October. In this study, we define the MCC duration as the duration between the time of the initial stage and the time of the dissipation stage. The frequency average distribution of the duration of all MCC systems in each month is shown in Fig. 2a in the line graph. The average duration was approximately ~11.53 h. The average duration of MCC life is greater than the overall MCC results globally at IMC, which averages around 9.5 h [8]. However, it is slightly longer than the global average (~10 h; [4]) and much less than those found in subtropical South America (14 h; [6]). The long-lived MCCs occurred in June, October, and November. Most of the short-lived MCCs occurred in May, February, and December. Figure 2b shows that the MCC that frequently occurs on the Northern Coast of Borneo Island has an average area size of the interior cloud of MCC during the mature stage of around 101,000–150,000 km2 . The MCC eccentricity that often occurs on the Northern Coast of Borneo Island generally ranges from 0.90–0.95 (see Fig. 2c). In general, MCC in IMC usually develops from late afternoon to nearly midnight (see Fig. 3). The most significant frequency of initiation events occurred between the time of 1700–2300 LT. LT is local time (UTC + 7). The MCC system reaches maximum size levels after midnight or early morning at mostly between 0100– 0500 LT, and then, MCC begins to decay in the early morning hours after sunrise, especially around 0700–1100 LT. The MCC system dissipated from noon to late afternoon, around 1200–1400 LT. This indicates that the life cycle characteristics of

Analysis of Mesoscale Convective Complex (MCC) that Often Appears …

339

Fig. 2 a Number of MCCs that occur and the duration of the MCCs each month, b area size of the interior cloud of MCC during the mature stage of MCC, and c number of MCC eccentricity at the time of maximum extent

Fig. 3 Frequency distribution of all MCCs during the critical stage (initial, mature, decay, and dissipation or post-MCC stage)

MCC tend to be nocturnal, which is following the characteristics of MCC in globally that occur in IMC [8]. This result is similar to the previous studies that documented regional populations of MCCs were found to be nocturnal, among others; MCC in the Americas and the Western Pacific region, Southern Africa [2, 3, 5, 6, 23]. It follows that the global population is also predominantly nocturnal [24].

3.2 MCC Evolution Using Composite Method Figure 4 illustrates the average time of TBB in the hours of initiation, maturation, decay, and dissipation stages for the 29 MCC cases using the composite method. The MCC that occurs on the Northern Coast of Borneo Island begins to develop from small-scale clouds on the Northern Coast of Borneo Island (see Fig. 4a) and begins to develop after one day from the initial phase even the interior cloud or

340

Trismidianto

cloud core has begun to be seen (see Fig. 4b). The interior cloud and cloud shield of the MCC are clearly visible from the day before the mature phase to the day after (see Fig. 4c-e). During the mature phase, the MCC looks very large, even covering the mainland and coastal areas. The interior cloud begin to shrink two days after the mature phase which indicates that MCC will begin to enter the decay phase (Fig. 4f). When it enters the decay phase (Fig. 4g), the interior cloud of MCC starts to disappear followed by the cloud shield from MCC starts to decrease in scale. The position of the MCC during the decay phase is also different compared to the location when the MCC is newly formed and mature, this shows that there has been a movement of the MCC toward the ocean. Two days after the decay phase, the MCC begins to move toward the ocean, leaving the North Coast of Borneo Island as the initial location of MCC formation (see Fig. 4h). Figure 4i and j shows that MCC is starting to dissipate spreading to the South China Sea, and it is also seen that there is distribution around the mainland from the Borneo Island. The development of MCC in this region is likely generated by the diurnal cycle of cloud systems from the South China Sea [25]. They describe that the diurnal cycle over the South China Sea disappears during the day and migrates to Borneo Island. The system reaches its maximum size around morning. The system begins to decay and dissipate in the morning to noon when the cloud system migrates into the South China Sea after the MCC dissipates. However, this system is likely related to the development of the diurnal cycle over the South China Sea as a cloud trigger. According to Houze [25], the diurnal cycle over the South China Sea begins to develop at midnight. Figure 5 illustrates the effect of MCC on rainfall distribution which shows the same pattern as the evolution of MCC development from TBB data. It can be seen that the peak of maximum rainfall does occur when the MCC reaches its maximum

Fig. 4 Horizontal distribution of composite of the TBB for 29 MCC composite in this study area. The time describes the average timing of the hours of initiation, mature, decay, and dissipation stage of MCC in terms of the normalized life cycle using composite method. The unit for TBB is Kelvin

Analysis of Mesoscale Convective Complex (MCC) that Often Appears …

341

Fig. 5 Same with Fig. 4 but for the horizontal distribution of rainfall composite for 29 MCC composite in this study area

peak. When the MCC decayed, it was also seen that there was a distribution of rainfall toward the South China Sea and also some around Kalimantan. This shows that the MCC that occurred on the Northern Coast of Borneo Island affected the distribution of rainfall in several areas in Kalimantan.

3.3 Cold Pool and Convergent Wind Support the MCC Development Figure 6a shows that the development of MCC on the North Coast of Borneo Island is triggered by the presence of surface convergent winds that interact with land breezes to form a small-scale convective system. The existence of surface convergent wind support from all directions, especially from the South China Sea, which is getting stronger, makes the convective system that was small in scale become an MCC system, as shown in Fig. 6b. During the mature MCC phase, the cold pools also began to be indicated by the surface potential temperature (theta) which is relatively smaller than the surrounding area, even though the land breeze was not seen clearly. According to Engerer et al. [26], a cold pool is an area of downdraft air cooled by evaporation that spreads out horizontally beneath a precipitating cloud. The difference in surface theta could have acted as a trigger for the development of new convective systems to form along the leading edge of the cold pool that began to be seen in the decay stage of the MCC, as shown in Fig. 6c. The divergent outflow due to the cold pool and in conjunction with the evening sea breeze makes the new convective system migrate widely throughout the surrounding area of the MCC system area, predominantly along the South China Sea and some of over Kalimantan Region, as shown in Fig. 6c and d.

342

Trismidianto

Fig. 6 Horizontal distribution of the composite analysis of the average rainfall for 29 MCC cases in this study (shaded), wind surface vector anomaly (vector), and surface theta (contour) for a initial stage, b mature stage, c decay stage, d dissipation stage

Figure 7a and b shows that during the initial and mature phases, there is a surface convergent wind in the vicinity of the MCC occurrence. This reinforces the statement in Fig. 6a and b that there is support from the surface convergent wind in the development of MCC. During the decay and dissipation phases (Fig. 7c and d), there is a surface divergent wind which also strengthens the analysis from Fig. 6c and d which shows that there is a divergent outflow when the MCC decays and dissipates. At the upper level, there is a divergent area around the MCC location during the initial and mature phases (see Fig. 7e and f). The area of upper layer divergence is also visible during the decay and dissipation phases but appears to be starting to shift from the location where the MCC system was formed (see Fig. 7g and h).

Analysis of Mesoscale Convective Complex (MCC) that Often Appears …

343

Fig. 7 Composite analysis of the geopotential height (m, black contour) and divergence (10–5 s-1, shaded) in 1000 hPa (a–d) and 200 hPa (e–h) during initial stage (a and e), mature stage (b and f), decay stage (c and g), and dissipation stage (d and h) using ECMWF ERA5 data

Figure 8a shows that during the initial and mature phases, there is a very strong convergence in the surface layer. The presence of this convergence wind is also seen in Fig. 8b, with the easterly wind in the surface layer toward the MCC system. Meridional winds on the surface layer also show movement from north to south toward the MCC formation system (see Fig. 8c). Surface convergence winds during the initial and mature phases shown in Fig. 8 support the statements in Fig. 6a and b. While the surface divergence winds shown in Fig. 6c and d cannot be shown in Fig. 8a where there is no wind divergence in the surface layer during the decay and dissipation phases, weak divergence winds are seen in the intermediate layer. However, surface divergence winds can be demonstrated by the presence of westerly winds from the MCC system (see Fig. 8b) and northerly winds from the MCC system (see Fig. 8c).

Fig. 8 Composite analysis of the a divergence (10–6 s-1), b zonal wind (m/s), and c meridional wind (m/s) for 29 MCC case with average latitude 2–5 N and longitude 112–115 E during MCC phase of initial, mature, decay, and dissipation using ECMWF ERA5 data

344

Trismidianto

Figure 9 also strengthens the analysis of Fig. 6 regarding the presence of a cold pool when the MCC reaches its maximum or mature stage and sometime after the mature stage, which causes a cold pool outflow that triggers the occurrence of new convective cells. This is indicated by an increase in pressure during the mature phase of MCC, followed by a decrease in the value of equivalent potential temperature (theta-E) and theta as well as a decrease in humidity and an increase in surface temperature. The cold pools were associated with theta and theta-E decreases [27] and generated by these individual convective cells in an MCS typically spread out over the surface and combine to form a large mesoscale cold pool covering a contiguous area on the scale of the entire MCS [1]. The formation of a cold pool when MCC occurs will trigger the formation of new convective cells which will cause continuous rain because the new convective cells that are formed induce convective surrounding them to form other new convective systems [9]. This is what makes MCC can cause heavy and even extreme rainfall. Figure 10 shows the percentage of extreme rainfall events during the MCC in this case study. It can be seen that extreme rainfall occurred during the development of MCC and reached its peak. However, during MCC decay and dissipation, several areas also experienced extreme rainfall due to the MCC.

4 Conclusion The results of the study show that as many as 29 MCCs that frequently occur on the North Coast of Borneo Island have a long-life span with an average duration of around ~11.53 h, which is longer than the average duration of all MCCs that occur in Indonesia, which is around 9.5 h [8]. The characteristics of the MCC life cycle tend to be nocturnal which are the same as the global MCC characteristics that occur in Indonesia [8]. The MCC that occurs on the North Coast of Borneo Island develops triggered by surface convergent winds which are indicated as westerly and northerly winds from the South China Sea which are supported by the presence of land breeze, as illustrated in Fig. 11. The system reaches its maximum limit due to interaction with the mainland winds the night that follows the rainfall is also getting the maximum. Cold pools begin to form when MCC reaches the mature phase. This cold pool triggers the formation of new convective cells which will cause continuous rain because the new convective cells that are formed induce convective surrounding them to form other new convective systems. New convective systems begin to emerge during the MCC in the decay and dissipation phases. New convective cells migrate westward after the MCC disappears. The MCC on the Northern Coast of Borneo Island has an impact on rainfall, and the MCCs in this region are possible as one of the triggers to the development of the other convective system over the South China Sea by Houze [25]. Houze [25] showed that the morning precipitation over the sea near the Northwest of Borneo Island is due to the convergence of the onshore and monsoon winds.

Analysis of Mesoscale Convective Complex (MCC) that Often Appears …

345

Fig. 9 Composite of surface pressure (hPa) and mean sea level pressure (hPa) in the above figure, equivalent potential temperature (Theta-E), and potential temperature (Theta) in unit Kelvin (middle figure), and the below figure shows the relative humidity (%) and surface temperature (K) during initial-1, initial, initial + 1, mature-1, mature, mature + 1, decay-1, decay, decay + 1, dissipation-1, dissipation, and dissipation + 1. All of this data is from ERA5. The area between the two red lines is an indication of the time of the cold pool

346

Trismidianto

Fig. 10 Percentage of extreme rainfall events during MCC in the stage of initial, mature, decay, and dissipation using the composite method

Fig. 11 Schematic representations of the evolution and propagation of the MCC over this study related to the development of the convective system over the South China Sea. Light and dark gray areas illustrated the MCC cloud shield and convective clouds, respectively

Analysis of Mesoscale Convective Complex (MCC) that Often Appears …

347

References 1. Houze, R.A. Jr.: Mesoscale convective system. In: Review of Geophisics, American Geophisical Union, vol. 43 (2004) 2. Maddox, R.A.: Mesoscale convective complexes. Bull. Amer. Meteor. Soc. 61, 1374–1387 (1980) 3. Miller, D., dan Fritsch, D., J.M.: Mesoscale convective complex in Western Pasific Region. Amer. Meteor. Soc. 119, 2978–2992 (1991) 4. Laing, A.G., dan Fritsch, J.M.: Mesoscale convective complexes over the Indian Monsoon Region. American Meteorol. Soc. 121, 2254–2263 (1993) 5. Velasco, I., dan Fritsch, J.M.: Mesoscale convective complexes in the America. J. Geophysic. Res. 92, 9591–9613 (1987) 6. Durkee, J.D., dan Mote, T.L.: A climatology of warm season Mesoscale convective complexes in subtropical South America. Int J. Clim, 12 (2009) 7. Yuan, J., Houze Jr, R.A.: Global variability of mesoscale convective system anvil structure from A-Train satellite data. J. Climate 23, 5864–5888 (2010). https://doi.org/10.1175/2010JC LI3671.1 8. Trismidianto.: The global population of Mesoscale convective complexes (MCCs) over Indonesian maritime continent during 15 Years. In: IOP Conference Series: Earth and Environmental Science, vol.166 (2018). https://doi.org/10.1088/1755-1315/166/1/012040 9. Trismidianto, Hadi, T.W., Ishida, S., Moteki, Q., Manda, A., Iizuka, S.: Development processes of Oceanic convective systems inducing the heavy rainfall over the Western Coast of Sumatra on 28 October 2007. Scient. Online Lett. on the Atmos. (SOLA). 12, 6−11 (2016). https://doi. org/10.2151/sola.2016-002 10. Trismidianto, Yulihastin, E., Satyawardhana, H., Nugroho, J.T., Ishida, S.: The contribution of the Mesoscale convective complexes (MCCs) to total rainfall over Indonesian Maritime Continent. In: IOP Conference Series: Earth and Environmental Science, vol. 54, (2017a) https://doi.org/10.1088/1755-1315/54/1/012027 11. Trismidianto.: Characteristics of the oceanic MCC, continental MCC, and coastal MCC over the Indonesian maritime continent. In: IOP Conference Series: Earth and Environmental Science, vol. 149, (2018). https://doi.org/10.1088/1755-1315/149/1/012024 12. Trismidianto, Yulihastin, E., Satyawardhana, H., Ishida, S.: A composite analysis of the Mesoscale convective complexes (MCCs) development over the Central Kalimantan and its relation with the propagation of the rainfall systems. In: IOP Conference Series: Earth and Environmental Science, vol. 54, (2017). https://doi.org/10.1088/1755-1315/54/1/012036 13. Trismidianto.: The large-scale meteorological condition during the critical stage of Mesoscale convective complexes (MCCs) over Indian Ocean. In: IOP Conference Series: Earth and Environmental Science, vol. 374, (2019). https://doi.org/10.1088/1755-1315/374/1/012039 14. Ashley, W.S.: A distribution of mesoscale convective complexe rainfall in the United State. American Meteorol. Soc. 131, 3003–3301 (2003) 15. Maddox, R.A.: Large scale meteorological conditions associated with midlatitude mesoscale convective complexes. Mon. Wea. Rev. 111, 1475–1493 (1983) 16. Hartmann, D.L., Hendon, H.H., Houze Jr, R.A.: Some implications of the mesoscale circulations in tropical cloud clusters for large-scale dynamics and climate. J. Atmos. Sci. 41, 113–121 (1984). https://doi.org/10.1175/1520-0469(1984)041,0113: SIOTMC.2.0.CO;2 17. Chen, S.S., Houze, R.A., Jr.: Diurnal variation and life-cycle of deep convective systems over the tropical Pacific warm pool. Q. J. R. Met. Soc. 123, 357–388 (1997) 18. Dee, D.P., Uppala, S.M., Simmons, A.J., Berrisford, P., Poli, S., Kobayashi, S., Andrae, U., Balmaseda, M.A., Balsamo, G., Bauer, D., Bechtold, P., Beljaars, A.C.M., Van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A.J., Haimberger, L., Healy, S.B., Hersbach, H., Holm, E.V., Isaksen, L., Kallberg, K., Kohler, M., Matricardi, M., McNally, A.P., Monge-Sanz, B.M., Morcrette, J.J., Park, B.K., Peubey, C., De Rosnay, P., Tavolato, C., Thepaut, J.N., Vitart, F.: The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Quart. J. Roy. Meteor. Soc. 137, 553–597 (2011)

348

Trismidianto

19. Atlas, R., Hoffman, R.N., Ardizzone, J., Leidner, S.M., Jusem, J.C., Smith, D.K., Gombos, D.: A cross-calibrated, multiplatform ocean surface wind velocity product for meteorological and oceanographic applications. Bull. Amer. Meteor. Soc. 92, 157–174 (2011) 20. McAnelly, R.L., Cotton, W.R.: Meso-β-scale characteristics of an episode of meso-α-scale convective complexes. Mon Wea Rev. 114, 1740–1770 (1986) 21. McAnelly, R.L., Cotton, W.R.: The precipitation life cycle of mesoscale convective complexes over the central United States. Mon. Wea. Rev. 117, 784–808 (1989) 22. Cotton, W.R., George, R.L., Wetzel, P.J., McAnelly, R.L.: A long-lived mesoscale convective complex. Part I: The mountain-generated component. Mon. Wea. Rev. 111, 1983−1918 (1983) 23. Blamey, R.C., Reason, C.J.C.: The role of mesoscale convective complexes in Southern Africa summer rainfall. J. Clim 26 (2012) 24. Laing, A.G. dan Fritsch, J.M, The global population of mesoscale convective complexes. Q. J. R. Meteorol. Soc. 123, 389–405 (1997) 25. Houze, R.A., Jr.: Cloud clusters and large-scale vertical motions in the tropics. J. Meteorol. Soc. Jpn. 60, 396–410 (1982) 26. Engerer, N.A., Stensrud, D.J., dan Coniglio, M.C.: surface characteristics of observed cold pools. Monthly Weather Rev. 136, 4839–4849 (2008) 27. Tompkins, A.M.: Organization of tropical convection in low vertical wind shears: the role of cold pools. J. Atmos. Sci. 58, 1650–1672 2001

Impact of Cold Surge Based on Its Strength on Rainfall Distribution in Western Indonesia Alfan Sukmana Praja

and Trismidianto

Abstract Cold surge (CS) is a mass of cold air carried by the north–south (meridional) wind circulation due to high pressure disturbances in the Siberian region that flows to the equator and south through the north coast of Java and is one of the synoptic scale weather phenomena that has a significant influence during Asian winter monsoon (NDJFM). This study was conducted to analyze the impact of CS based on the category of strength to the distribution of rainfall in Indonesia. Analysis of CS events was carried out by calculating the CS index which is the average meridional wind speed at the level of 850 hPa in the region of 110−115 °E and 7.5−12.5 °N. Based on its strength, CS is divided into three, namely weak CS (8–10 ms-2), moderate CS (10–12 ms-2), and strong CS (>= 12 ms-2). Rainfall data are obtained from 20 years GSMaP data and wind data and other parameters from ERA 5 data. The results showed the area that experiences an increase in rainfall decreases when the strength of the CS increases. However, rainfall intensity increased in the South China Sea, Karimata Strait, and the Java Sea. When CS is weak, there is generally an increase in rainfall from day-1 to day+ 1. While for moderate CS, the increase in rainfall starts when a cold surge occurs until day+ 2. Lastly, when CS is strong, an increase in rainfall and potential for cold surges has occurred from day-2 to day+ 2. Strong CS can increase rainfall more in the Java Sea area than Java Island, whereas weak and moderate CS increase rainfall more in Java Island than in the Java Sea from day-1 to day+1.

1 Introduction Indonesia’s strategic geographical location has attracted world researchers to observe the dynamics of climate and weather in Indonesia. As a tropical area, Indonesia has high rainfall and changes yearly. Changes in rainfall on a seasonal scale in A. S. Praja (B) · Trismidianto Research Center for Climate and Atmosphere, National Research and Innovation Agency, BRIN KST Samadikun, Jl. Cisitu Sangkuriang, Bandung, West Java, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_33

349

350

A. S. Praja and Trismidianto

Indonesia are heavily influenced by Monsoon, ENSO, MJO, and so on. Changes in the daily scale can occur by disturbances from the north called a cold surge. The value of fluctuations in air pressure and wind on a large scale can trigger a cold surge phenomenon, namely a large mass flow of air from mainland Siberia in the meridional east season accompanied by a decrease in temperature and air pressure in Hong Kong at a surface layer of 1000–850 hPa. The cold surge will flow through the South China Sea, Karimata strait, the Bangka Islands to the Java Sea. Increased rainfall in the Java Sea, which can reach the mainland, can trigger flooding in areas near the coast, such as Jakarta. Therefore, cold surge analysis is crucial to increase awareness in dealing with floods due to high rainfall, especially from November to March. In these months, the rainfall on the island of Java tends to be high, along with frequent cold surges. The relationship between MJO and cold surge has been studied in some parts of Indonesia [1–3]. Generally, when MJO is active, cold surges can increase rainfall even up to 150% [1]. However, the type of cold surge that often causes extreme rain in western Indonesia has not been further elaborated. Thus, this information is very important to anticipate the occurrence of high rainfall. The impact of cold surges on the tropics is that it can trigger cumulus convection so that cold surges can increase extreme rainfall and precipitation in several parts of western Indonesia [1, 2]. Cold surges can occur once or twice a month for up to 2 weeks [4]. The high potential intensity of rainfall in an area can be seen from water vapor transport [5, 6]. The previous research on cold surges has not specifically divided the influence of cold surge strength on rainfall in western Indonesia. So in this study, the authors divide the cold surge into three categories, namely weak, moderate, and strong, and explain its influence on rainfall distribution in western Indonesia. A brief explanation of the MJO and extreme rain is provided along with the case study.

2 Data and Method The data used in this study include daily rainfall data in the month of NDJFM (November, December, January, February, and March) as months that have high rainfall rates [7] and correspond to the time of frequent occurrence of cold surges [8]. Rainfall data from 2000 to 2019 were obtained from GSMap, downloaded from sharaku.eorc.jaxa.jp/GSMaP, with a spatial resolution of 0.1 × 0.1 degrees. This study also used zonal wind data and meridional winds with the same period used in rainfall data. Wind data are obtained from ECMWF which can be downloaded at apps.ecmwf.int/datasets/data/interimfull-daily/ with a spatial resolution of 0.25 × 0.25 degrees. We also use outgoing long-wave radiation(OLR) [9] data from NCEP/NCAR. Analysis of CS events was carried out by calculating the CS index, which is the average meridional wind speed at the level of 850 hPa in the region of 110–115 °E and 7.5–12.5 °N [10]. Based on its strength, CS is divided into three, namely, weak

Impact of Cold Surge Based on Its Strength on Rainfall Distribution …

351

CS (8–10 ms-2), moderate CS (10–12 ms-2), and strong CS (>= 12 ms-2) [4]. All of 193 CS events consisting of 140 weak CS (73%), 37 moderate CS (19%), and 16 strong CS events (8%). In this study, we defined active MJO as when MJO is active in the Indian Ocean (phases 2 and 3) and active in the maritime continent (phases 4 and 5) with an amplitude greater than or equal to 1, otherwise called inactive MJO. The calculation of extreme rainfall uses the 95th percentile of the cumulative daily precipitation distribution [11]. The measure of moisture transport is obtained from the total water vapor carried according to the movement of the wind [1, 5].

3 Result and Discussion 3.1 Rainfall Distribution During CS Event In western Indonesia, when CS occurs, rainfall in the month of NJDFM increases to 21 mm/day, especially along the South China Sea route, the Karimata Strait to the Java Sea, and parts of the Indian Ocean to southern Java (Fig. 1). Meanwhile, when there is no CS, there is relatively no rainfall anomaly in western Indonesia. Based on Fig. 2, the wind direction in the weak, moderate, and strong CS events is relatively the same, mostly blowing from the South China Sea, the Karimata Strait to turning toward the Java Sea. However, in strong CS, a vortex around Kalimantan triggers the Borneo vortex [12]. The highest increase in rainfall intensity in the South China Sea area along the CS route, the Karimata Strait to the Java Sea, occurs when the CS is strong. However, the highest increase in rainfall intensity in the western part of Java and the Indian Ocean near Sumatra was caused by moderate CS. Besides the increase in rainfall intensity, there is a decrease in rainfall intensity in parts of the Indian Ocean near Sumatra when the CS is strong.

Fig. 1 Rainfall Distribution during CS dan No CS

352

A. S. Praja and Trismidianto

Fig. 2 Rainfall anomaly when CS is weak, moderate, and strong

A comparison of extreme rainfall patterns when CS is weak, moderate, and strong is presented in Fig. 3 following. When CS is weak, the highest extreme rainfall occurs around the South China Sea and parts of the Java Sea, Batam, and northern Bangka waters. At moderate CS, extreme rainfall events with an average of up to 30% occur in the South China Sea, North Brunei, and North Java to the Java Sea. Meanwhile, when CS is strong, extreme rainfall occurs up to 35% in the South China Sea, north of the island of Kalimantan, to the Java Sea. Generally, CS strong produces higher extreme rainfall in western Indonesia than when CS is weak, and CS is moderate. The previous studies have shown that 1–2 days of lag time are required for cold surges to increase rainfall [1]. This lag time analysis aims to see the movement of each type of CS and rainfall before and after the occurrence of CS. In Fig. 4 lag time analysis, the increase in rainfall occurs when CS is weak and moderate in the Java Sea area. As for CS strong, from day-1 to day+2, there was a fairly consistent increase in rainfall in the Java Sea area. From day-2 until day+2 in weak CS, there is no significant change on the island of Sumatra and Kalimantan island with anomaly changes ranging from 0 to 4 mm. Around the island of Java, the most significant increase in rainfall anomaly occurs

Fig. 3 Percentage of extreme rainfall events

Impact of Cold Surge Based on Its Strength on Rainfall Distribution …

353

Fig. 4 Lag time analysis day-2 to day+2 for weak, moderate, and strong CS

when day-1 in weak CS. When CS is moderate, there is a decrease in rainfall of up to 12 mm from day-2 to day +2 on the islands of Sumatra and Kalimantan. However, in the western part of Java Island, the highest rainfall is from day-2 to day+2. When CS strong occurs, the series of days of the occurrence of CS have different intensities between West Kalimantan and East Kalimantan. The western part of Kalimantan generally has a decrease in rainfall, while the eastern part experiences an increase in rainfall from day-2 to day +2. High rainfall occurs on day-2 and day-1, while on CS day up to day+ 2, very high rainfall occurs in the South China Sea when CS is strong. When CS is weak, there is generally potential for cold surges from day-1 to day+1, when CS is moderate, CS day to day+2 still occurs CS. Meanwhile, when CS is strong, from day-2 to day +2 cold surge, an increase in rainfall still occurs in the South China Sea.

3.2 Impact of CS During MJO Figure 5 shows the difference between events when MJO is active and inactive for weak, moderate, and strong CS. The influence of MJO on rainfall is relatively strong in the Indian Ocean, West Sumatra, especially when CS is weak and CS is moderate. When CS is weak with MJO takes place, rainfall increases up to 15 mm, while when inactive MJO only reaches 6 mm. Even when the CS is moderate, the active MJO generally increases in rainfall by up to 18 mm compared to the inactive MJO, which typically decreases in Sumatra’s West Indian Ocean region.

354

A. S. Praja and Trismidianto

Fig. 5 CS based on its strength when MJO is active and inactive

When MJO is active, the distribution of the increase in rainfall gets narrower in contrast to the increasing strength of the CS. When CS is weak, almost all areas experience increased rainfall from the Indian Ocean to Kalimantan. However, when the CS is moderate, parts of northern Sumatra and the Riau Islands decrease in rainfall intensity. When the CS was strong, the areas that experienced a decrease in rainfall were more expansive: the Indian Ocean, West Sumatra, Riau Islands, North Sumatra, and parts of southern Java. In contrast to when the MJO is inactive, the distribution of the increase in rainfall when the CS is weak occurs in the South China Sea, most of the Java Sea, and parts of the Indian Ocean near Sumatra. The increase distribution becomes narrower when CS is moderate, which only occurs in the South China Sea, the Karimata Strait, and a small part of the Java Sea. However, when CS is strong, the distribution of the increase in rainfall extends to cover the South China Sea, Java Sea, southern Java Island, and parts of Sumatra. The OLR value reaches −70 when MJO is inactive for strong CS in the South China Sea, even lower than when MJO is active, which only reaches −50. However, this is the opposite in the Java Sea area, which has a lower OLR value when MJO is active, namely −50, and when MJO is inactive, −30. In general, the OLR value when MJO is active has a short range of −40 to −50 in the South China Sea area, both when CS is weak, moderate, or strong. But when MJO is inactive, there is a significant difference when CS is weak and CS is moderate, which only reaches − 20 to −30 against strong CS, which can reach −70.

Impact of Cold Surge Based on Its Strength on Rainfall Distribution …

355

Fig. 6 Moisture transport when weak, moderate, and strong CS

3.3 Moisture Transport and Case Study Analysis Moisture transport for weak and moderate CS tends to be influenced by flows from the Indian Ocean. Meanwhile, moisture transport in pure strong CS is obtained only from the north, not from the Indian Ocean (Fig. 6). Observation data in the Kemayoran area represent the occurrence of high-intensity rainfall and flood events in the Jakarta area during weak, moderate, and strong cold surges. Rainfall in Fig. 7 represents weak, moderate, and strong CS on 17 January 2014, 3 February 2018, and 13 January 2009, with 147.9, 50.7 and 102 mm/day, respectively [13]. Based on the Hovmoller diagram in Fig. 7, day-1 to day+ 1, strong CS can increase rainfall more in the Java Sea area than Java Island, whereas weak and moderate CS increase rainfall more in Java Island than the Java Sea. This increase is also triggered by the moisture transport in Fig. 6, which explains that the flow from the Indian Ocean to the island of Java is not significant when the CS is strong.

Fig. 7 Time-latitude cross sections of meridional wind and precipitation average of 100°–120°E when a weak, b moderate, and c strong cold surge from day-1 to day+1

356

A. S. Praja and Trismidianto

4 Conclusion The area that experiences an increase in rainfall decreases when the strength of the CS increases. However, rainfall intensity increased in the South China Sea, Karimata Strait, and the Java Sea. The increase in rainfall intensity and the narrowing of the distribution area also occur when CS and MJO are active. In contrast to inactive MJO, although the intensity of rainfall increases, the distribution area narrows during moderate CS and expands again during strong CS. Therefore, the percentage of extreme rainfall events in weak and moderate CS is relatively high for the South China Sea and North Java Sea areas. As for the percentage of extreme rainfall in Strong CS, the coverage is wider distributed in the South China Sea, the Karimata Strait to the Java Sea, and the southwest of the island of Java. The lag time analysis of each CS has differences related to the increase in rainfall and the potential for CS two days before to 2 days after. When CS is weak, there is generally an increase in rainfall from day-1 to day+1. While for moderate CS, the increase in rainfall starts when a cold surge occurs until day+2. Lastly, when CS is strong, an increase in rainfall and potential for cold surges has occurred from day-2 to day+2. Moisture transport for weak and moderate CS is influenced by flow from the Indian Ocean, while it is not for strong CS. Based on the case study, strong CS can increase rainfall more in the Java Sea area than Java Island, whereas weak and moderate CS increase rainfall more in Java Island than in the Java Sea from day-1 to day+1.

References 1. Fauzi, R.R., Hidayat, R.: Role of cold surge and MJO on rainfall enhancement over Indonesia during East Asian winter monsoon. In: IOP Conference Series: Earth and Environmental Science. vol. 149(1), IOP Publishing (2018) 2. Lim, S.Y., Marzin, C., Xavier, P., Chang, C.P., Timbal, B.: Impacts of boreal winter monsoon cold surges and the interaction with MJO on Southeast Asia rainfall. J. Clim. 30(11), 4267–4281 (2017) 3. Pang, B., Lu, R., Ling, J.: Impact of cold surges on the Madden-Julian oscillation propagation over the Maritime Continent. Atmosp. Sci. Lett. 19(10), e854 (2018) 4. Chang, C.P., Harr, P.A., Chen, H.-J.: Synoptic disturbances over the equatorial South China Sea and western maritime continent during boreal winter. Monthly Weather Rev. 133(3), 489–503 (2005) 5. Zhou, T.J., Yu, R.C.: Atmospheric water vapor transport associated with typical anomalous summer rainfall patterns in China. J. Geophys. Res. D: Atmos. 110(8), 1–10 (2005) 6. Wicaksono, G.B., Hidayat, R.: Extreme rainfall in Katulampa associated with the atmospheric circulation. Proc. Environ. Sci. 33, 155–166 (2016) 7. Xavier, P. et al.: Influence of Madden-Julian oscillation on Southeast Asia rainfall extremes: observations and predictability. Geophys. Res. Lett. 41(12), 4406–4412 (2014) 8. Yulihastin, E., Hadi, T.W., Ningsih, N.S.: Diurnal rainfall propagation relate to cold surge-cold tongue interaction over the northern coast of West Java. In: IOP Conference Series: Earth and Environmental Science. vol. 303(1), IOP Publishing (2019) 9. Liebmann, B., Smith, C.A.: Description of a complete (interpolated) outgoing longwave radiation dataset. Bull. Am. Meteor. Soc. 77(6), 1275–2127 (1996)

Impact of Cold Surge Based on Its Strength on Rainfall Distribution …

357

10. Koseki, S., Koh, T.Y. ,Teo, C.K.: Borneo vortex and mesoscale convective rainfall. Atmosp. Chem. Phys. 14(9), 4539–4562 (2014) 11. Reed, A.T., Stansfield, A.M., Reed, K.A.: Characterizing long Island’s extreme precipitation and its relationship to tropical cyclones. Atmosphere 13(7), (2022) 12. Saragih, R.M., Fajarianti, R., Winarso, P.A.: Atmospheric study of the impact of Borneo vortex and Madden-Julian oscillation over Western Indonesian maritime area. J. Phys.: Conf. Ser. 997(1), IOP Publishing (2018) 13. BMKG Homepage. https://dataonline.bmkg.go.id. Last Accessed 08 Oct 2022

Numerical Simulation of Low-Level Wind Shear at Soekarno-Hatta Airport Associated with Landward Propagation of Mesoscale Convective System Dita Fatria Andarini, Muhammad Arif Munandar, and Ibnu Fathrio

Abstract Low-level wind shear (LLWS) is one of meteorological phenomena that can produce major risks for aircraft. This study aims to investigate the occurrence of low-level wind shear in Soekarno-Hatta Airport, Jakarta (12 December 2019), using a high-resolution numerical weather prediction of the weather research forecasting (WRF) model. Moreover, the hourly rainfall data from Global Satellite Mapping of Precipitation (GSMaP) was utilized to analyze the convective system. According to pilot reports (PIREPS), strong LLWS was recorded at around 10:00–11:00LT. The rainfall distribution of GSMaP confirmed that the LLWS event occurred with the presence of a convective system that propagated from the North of Java Island. The maximum rainfall was also observed around the airport, with an intensity of about 30 mm. The WRF model successfully simulated the phase and amplitude of precipitation to explain this convective activity. In addition, there were two LLWS events observed from the model results; 09:50–10:00LT and 10:30LT in the east and west airport, respectively. The surface wind speed of more than 10 m/s associated with strong downdraft, divergence, and high magnitude of radar reflectivity (>40 dB) induced those LLWS processes.

1 Introduction Vertical wind shear, defined as the variation of wind direction and speed by height, is a crucial parameter in the atmosphere that could be associated with small to largescale atmospheric phenomena. At low-level altitudes, the occurrence of significant D. F. Andarini (B) · I. Fathrio National Research and Innovation Agency, Jakarta, Indonesia e-mail: [email protected] I. Fathrio e-mail: [email protected] M. A. Munandar Indonesian Agency for Meteorological, Climatological and Geophysics, Jakarta, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_34

359

360

D. F. Andarini et al.

vertical wind shear could be dangerous for aircraft. It could happen when the airplane experiences a microburst produced below the base of a thunderstorm. The microburst causes the plane to encounter strong headwinds and tailwinds and displaces the aircraft distance from the ideal take-off and landing path [1]. The previous studies have shown how improvement of numerical weather prediction (NWP) skill results in better low-level wind shear (LLWS) prediction. Storm and Basu (2010) employed weather research forecasting (WRF) simulation to forecast low-level wind shear in the United States Great Plains as an exchange for costly observation and better spatial–temporal coverage [2, 3]. Their results revealed the dependency of the model on utilizing a planetary boundary layer scheme in simulating the wind shear. Moreover, [4] and [5] also used a high-resolution model up to 200 m to predict airflow variation at the Hong Kong International airport as a possible cause of low-level wind shear occurrence. Similarly, [6] evaluated the output of high spatio-temporal resolution of NWP to predict LLWS in the runway corridor of Hon Kong International Airport. It depicted positive skills for short-term prediction. In terms of the implementation of data assimilation to improve the prediction results, [7] successfully applied LIDAR to improve WRF’s ability to detect LLWS events. They highlighted that LIDAR data assimilation could improve the forecast for a short duration after the events, while increased vertical level had less contribution to improving the forecast result. A more recent study by Choi et al. [8] also implemented statistical analysis of ensemble model output to simulate the low-level wind shear at several airports in South Korea, where the results showed a better prediction than the raw ensemble method. Indonesia is an archipelago country, consisting of five major islands and thousands of small islands. Thus, aircraft is a vital transportation means to connect all areas of the nation. More than 250 airports are available in Indonesia for the domestic and international flights. In this case, information on flight weather disturbances, especially LLWS, is crucial in supporting the safety of Indonesian aviation. However, there are only a few studies that utilize the use of NWP and observations in predicting LLWS in Indonesia. One of them is the research conducted by Sasmito et al. [9], which analyzed the Lion Air plane crash in 2013 and showed the potential impact of LLWS. Therefore, this study re-simulated one of the LLWS events related to mesoscale convection activity on 12 December 2019, around 10:00–11:00 local time (LT) at Soekarno-Hatta International Airport, Jakarta. This study aims to evaluate the skill of WRF models run in high spatial resolution for convective-induced LLWS cases in Jakarta airport that has not been investigated in the previous studies.

2 Data and Method WRF experiments were applied to simulate LLWS events on 12 December 2019, around 10:00–11:00LT, recorded by a pilot report. Unfortunately, no maximum wind data was obtained when an LLWS disturbance occurred from the LLWS warning system at the airport. In detail, the WRF model configuration consists of 3 nested

Numerical Simulation of Low-Level Wind Shear at Soekarno-Hatta …

361

domains with a spatial resolution of 9 km, 3 km, and 1 km, respectively (see Fig. 1). The WRF model configuration applied in this study was the WRF Single Moment 6 (WSM6) microphysics scheme and Yonsei University planetary boundary layer scheme. Regarding the cumulus scheme, the Betts-Miller Janjic scheme was implemented in Domain 1 (D01) and Domain 2 (D01). Meanwhile, Domain 3 (D03) used no cumulus scheme to simulate convection explicitly. These configurations had successfully implemented to simulate the propagation of convective systems that induced heavy rainfall over southern Sumatera [10] and diurnal precipitation over the Maritime Continent [11]. Furthermore, the evaluation was conducted only on the D03, which is assumed to have better skill in elucidating the convective system as the cause of the LLWS. In addition, the Air Force Weather Agency (AFWA) diagnostic tools in WRF were also applied for further analysis. It allows us to analyze the weather disturbances on the aircraft from specific variables such as low-level wind shear, wind gust, and turbulence [12]. In order to evaluate the convective system during the LLWS event, gridded rainfall data from the Global Satellite Mapping of Precipitation gage-corrected (GSMaP_gage) standard version 7 was employed for the period of 12 December 2019, 04:00LT to 14:00LT [13]. This data has 0.1 × 0.1 degrees of horizontal resolution and a 1-h interval. The GSMaP precipitation was used as a reference to evaluate the convective system that occurred during the incidence. In addition, this study also utilized the horizontal wind component data in surface layer from the European

Fig. 1 Simulation domain of WRF model. The red boxes represent domain 1 (D01), domain 2 (D02), and domain 3 (D03) for 9, 3 and 1 km resolution, respectively

362

D. F. Andarini et al.

Center for Medium-Range Weather Forecast reanalysis 5 (ERA-5) [14]. The spatial resolution of this data is 0.25 × 0.25 degree, while the temporal resolution is 1 h.

3 Results and Discussion 3.1 Mesoscale Convective System A westward propagation of the mesoscale convective system (MCS) was detected originating from the Java Sea, north of central Java, from the night to the morning, as shown from the spatial distribution of hourly rainfall in Fig. 2. The convective system initiated at around 04:00LT. It then reached its mature stage next to the north of western Java at 07:00LT. Three hours later, the southern part of this mesoscale convective system propagated southwestward and hit the Northern part of Java Island, including Soekarno-Hatta airport. The maximum precipitation was observed at about 30 mm per hour when it reached the airport. Eventually, the MCS decayed at 14:00 LT. Generally, the MCS activity induced heavy rainfall over the airport at around 10:00LT. In 09:50–10:30 LT, pilot reports informed that the strong low-level wind shear was identified. Unfortunately, there were no actual wind speed data at the airport during that period. According to the horizontal wind components from ERA-5 data, a strong northerly surface wind occurred over the study area since 04:00LT. This condition was persistent and it increased at about 10:00LT and 12:00LT with a magnitude of more than 5 m/s. Moreover, the peak of precipitation was also found in the same period. As a result, it confirmed that low-level wind shear was generated by the convective systems. This is also in agreement with the previous study that the unstable

Fig. 2 Spatial distribution of precipitation from GSMaP superimposed with ERA-5 surface wind on 12 December 2019 at a 04:00 LT and b 05:00, and 12 December 2019 at c 07:00 LT, d 10:00 LT, e 12:00 LT, and f 14:00 LT. X marker denotes the location of the airport

Numerical Simulation of Low-Level Wind Shear at Soekarno-Hatta …

363

atmospheric conditions are conducive to convective cloud development, and hence, are favorable for the LLWS process [15].

3.2 WRF Simulation The propagation of the convective system was clearly seen from the 1 km rainfall simulated by the WRF model (Fig. 3). Firstly, the convective system concentrated around the north of Central Java at around 05:00LT. Then, it propagated to the southwest (land area) and reached a peak in the airport at about 10:00LT. In this stage, the rainfall reached its maximum value of around 30 mm, which is consistent with the observation data, as shown in Fig. 2. Eventually, the convection system began to decay in the next two hours. From this result, it is worth noting that the WRF model successfully simulated the evolution of the convective system both in amplitude and phase. To investigate a more detailed analysis of the LLWS event induced by the convection system, this study also explored other parameters simulated by a high-resolution WRF model, such as surface wind and derived radar reflectivity. Figure 4 shows the spatial distribution of derived radar reflectivity superimposed with the divergence and wind vector. The surface wind speed reached a value of greater than 10 m/s twice associated with strong downdraft, divergence pattern of surface wind, and high magnitude (> 40 dBz) of simulated radar reflectivity as the index of heavy precipitation. This pattern could be associated with a microburst that spreads out at the surface, generating a strong surface wind speed or LLWS. However, the recorded downdraft speed seemed weaker than the previous study [16], and models simulated the location of LLWS events 5–10 km from the airport.

Fig. 3 Evolution of rainfall (filled contour) and surface wind (vector) simulated by the WRF model on 12 December 2019 at a 04:00 LT and b 05:00 LR, c 07:00 LT, d 10:00 LT, e 12:00 LT, and f 14:00 LT. X marker denotes the location of the airport

364

D. F. Andarini et al.

Fig. 4 Spatial distribution of wind speed (filled contour) and surface wind divergence (solid purple line) on 12 December 2021 at a 09:45 LT, b 09:54 LT, c 10:00 LT, d 10:20 LT, e 10:24 LT, and f 10:27 LT. Only values (multiplied by 1000 for clarity) of 1.5, 3 and 6 (units in m−1 ) are shown for surface wind divergence. Location of vertical velocity less than 2 m/s in 0–2 km height is denoted by the orange cross marker. Derived radar reflectivity greater than 40 dBz at 2 km height is denoted by the red line contour. The first LLWS event and second LLWS are represented in top panels (a, b, c) and lower panels (d, e, f), respectively. All variables are estimated by using AFWA diagnostic tool. X marker denotes the location of the airport

There were two LLWS events that occurred in the period of 09:45LT–10:30LT. The first LLWS was identified around the east of the airport at about 09:50–10:00 LT, indicated by the strong easterlies with the magnitude greater than 15 m/s (Fig. 4a– c). The high amplitude of radar reflectivity also dominated the western area of the airport. Moreover, the second LLWS was observed next to the west of the airport at around 10:24LT, as shown in Fig. 4d–f. This consistent high magnitude of radar reflectivity and divergence favored the environment to produce the second LLWS event. Furthermore, Fig. 5 displays the distribution of LLWS and wind gusts during the first and second LLWS events. A strong LLWS and maximum wind gust near the airport with values greater than 10 m/s (~20knots) occurred during the first event (Fig. 5a-c). Similarly, the high value of LLWS was also found nearby the airport (next to the west) in the second event, as depicted in Fig. 5d-f. Those results are consistent with the high intensity of reflectivity and strong easterly surface wind during the two LLWS events, as previously explained in Fig. 4.

4 Conclusion WRF simulation on high spatial resolution (1 km) has successfully simulated the evolution of mesoscale convective systems that triggered strong low-level wind

Numerical Simulation of Low-Level Wind Shear at Soekarno-Hatta …

365

Fig. 5 Contour of wind gust (red) and low-level wind shear (purple) that represent wind speed of 10 m/s for the first event at a 09:45 LT, b 09:54 LT, c 10:00 LT, and the second event at d 10:20 LT, e 10:24 LT, and f 10:27 LT on 12 December 2019. All variables are estimated by using AFWA diagnostic tool. X marker denotes the location of the airport

shear in the vicinity of Jakarta International Airport around 09:50–10:30LT on 12 December 2019. The phase and maximum precipitation were simulated comparably with the observation from the GSMaP data. Further analysis of LLWS and wind gust was carried out by utilizing AFW diagnostic tool to reveal the presence of significant LLWS and wind gust. The model results identified two LLWS events; 09:50–10:00LT at the east and around 10:24LT at the west of the airport. Those events were modulated by more than 10 m/s of the surface wind associated with strong downdraft, divergence, and high magnitude of radar reflectivity (>40 dB). The location of the LLWS event cannot be determined since there was no available data. Nevertheless, WRF models could predict the possible location of LLWS events 5–10 km away from the airport, which should be considered as an alert for the takeoff and landing of the aircraft. This could also be partly caused by the selection of parameterization schemes in the parent domains, especially the cumulus scheme as a crucial factor in determining the strength, timing, and location of convection. Therefore, sensitivities studies of the model should be further investigated in future studies.

References 1. HKO, I.: Windshear and Turbulence in Hong Kong–information for pilots. Hong Kong Observatory and International Federation of Air Line Pilots’ Associations (2005) 2. Storm, W., Basu, S.: The WRF model forecast-derived low-level wind shear climatology over the United States Great Plains. Energies 2010(3), 258–276 (2010). https://doi.org/10.3390/en3 020258

366

D. F. Andarini et al.

3. Skamarock, W.C., Klemp, J.B., Dudhia, J., Gill, D.O., Liu, Z., Berner, J., Wang, W. et al.: In: A description of the Advanced Research WRF Model Version 4, National Center for Atmospheric Research, Boulder, CO, USA, pp. 145. (2019) 4. Chan, P.W., Hon, K.K.: Performance of super high resolution numerical weather prediction model in forecasting terrain-disrupted airflow at the Hong Kong international airport: case studies. Meteorol. Appl. 23(1), 101–114 (2016) 5. Chan, P.W., Hon, K.K.: Observation and numerical simulation of terrain-Induced windshear at the Hong Kong International Airport in a planetary boundary layer without temperature inversions. Adv Meteorol Artic ID (2016) 6. Hon, K.: Predicting low-level wind shear using 200-m-resolution NWP at the Hong Kong international airport. J. Appl. Meteorol. Climatol. 59(2), 193–206 (2020) 7. Li, L., Xie, N., Fu, L., Zhang, K., Shao, A., Yang, Y., Ren, X.: Impact of lidar data assimilation on low-level wind shear simulation at Lanzhou Zhongchuan international airport, China: a case study. Atmosphere 11, 1342 (2020). https://doi.org/10.3390/atmos11121342 8. Choi, H.-W., Kim, Y.-H., Han, K., Kim, C.: Probabilistic forecast of low level wind shear of Gimpo, Gimhae, Incheon and Jeju international airports using ensemble model output statistics. Atmosphere 12, 1643 (2021). https://doi.org/10.3390/atmos12121643 9. Sasmito, A., Permana, D.S., Praja, A.S., Haryoko, U.: Pengaruh Microburst dan Low-Level Wind Shear (LLWS) Pada Kasus Kecelakaan Pendaratan Pesawat Lion Air Tanggal 13 April 2013 Di Bali. Jurnal Meteorologi dan Geofisika 21(1), 1–8 (2020) 10. Yulihastin, E., Fathrio, I., Nauval, F., Saufina, E., Harjupa, W., Satiadi, D., Nuryanto, D.E.: Convective cold pool associated with offshore propagation of convection system over the East Coast of Southern Sumatra, Indonesia. Adv. Meteorol. (2021) 11. Fonseca, R.M., Zhang, T., Yong, K.-T.: Improved simulation of precipitation in the tropics using a modified BMJ in the WRF model. Geoscientif. Model Developm. 8(9), 2915–2928 (2015) 12. Creighton, G., Kuchera, E., Adams-Selin, R., McCormick, J., Rentschler, S., Wickard, B.: AFWA diagnostics in WRF (2014) 13. Kubota, T., Shige, S., Hashizume, H., et al.: Global precipitation map using satellite-borne microwave radiometers by the GSMaP project: production and validation. IEEE Trans. Geosci. Remote Sens. 45(7), 2259–2275 (2007) 14. Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., Nicolas, J., Peubey, C., Radu, R., Schepers, D., et al.: The ERA5 global reanalysis. Q. J. R. Meteorol. Soc. 146, 1999–2049 (2020). https://doi.org/10.1002/qj.3803 15. Lin, C., Zhang, K., Chen, X., Liang, S., Wu, J., Zhang, W.: Overview of low-level wind shear characteristics over Chinese mainland. Atmosphere 12(5), 628 (2021) 16. Bolgiani, P., Fernández-González, S., Valero, F., Merino, A., García-Ortega, E., Sánchez, J.L., Martín, M.L.: Simulation of atmospheric microbursts using a numerical mesoscale model at high spatiotemporal resolution. J. Geophys. Res.: Atmosp. 125(4), e2019JD031791 (2020)

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia Using the VECM Model Mutia Yollanda, Wendi Harjupa, Dodi Devianto, Dita Fatria Andarini, Fadli Nauval, Elfira Saufina, Anis Purwaningsih, Wendi Harjupa, Trismidianto, Teguh Harjana, Risyanto, Fahmi Rahmatia, Ridho Pratama, and Didi Satiadi Abstract A study of two global circulations crossing Indonesia, the CrossEquatorial Northerly Surge (CENS) and the Madden–Julian Oscillation (MJO), can trigger extreme rainfall occurrences over Indonesia, has been carried out. The purpose of this study is to predict the presence of these circulations and the impact of these circulations on the distribution of rain throughout Indonesia. The territory of Indonesia was divided into 14 clusters to determine the effect of CENS and MJO and observe the relation of which with extreme rainfall patterns in each area. The method used to predict these three phenomena is a statistical approach called vector autoregressive (VAR), which indicates the occurrence of global circulation and extreme rain in Indonesia. In the studied area, there is a cointegration or long-term relationship between variables, which can be anticipated using the vector error correction model (VECM). It is also found that the presence of CENS and MJO circulation can be predicted with the model error rate measured based on the MAPE values of 1.8518 and 0.0619. When CENS is active (inactive), rainfall occurrences occur more (less) in the southern part of Indonesia, respectively. Moreover, the VECM model provides a pretty good prediction for the variables CENS, MJO, and rainfall intensity in 14 clusters of Indonesia.

1 Introduction Hydrometeorological disaster data recorded by the National Disaster Management Agency (BNPB) from 2000 to 2022 show that hydrometeorological disasters such as floods and landslides are the most frequent disasters that have a significant effect on M. Yollanda (B) · D. Devianto Department of Mathematics and Data Science, Andalas University, Padang, Indonesia e-mail: [email protected] W. Harjupa · D. F. Andarini · F. Nauval · E. Saufina · A. Purwaningsih · W. Harjupa · Trismidianto · T. Harjana · Risyanto · F. Rahmatia · R. Pratama · D. Satiadi Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_35

367

368

M. Yollanda et al.

life (https://gis.bnpb.go.id/ [1]). For example, based on the BNPB statement, from January to September 2020 alone, 99% of disasters were recorded as hydrometeorological disasters. The highest number of incidents was the flood as many as 791 times and followed by other disasters, including 573 tornadoes, 387 landslides, 314 forests and land fires, 26 tidal waves or abrasion, 22 droughts, 13 earthquakes, and volcanic eruptions. 5. The total number of natural disasters is 2131 incidents. A number of these incidents have a consequence of lost life and property. The BNPB recorded 322 fatalities, 454 were injured, and 4,481,641 were displaced and affected by the disaster. While infrastructure damage, the disaster impacted the housing sector, the 31,749 housing units, 627 educational facilities, 653 worship facilities, and 128 health facilities were damaged (cnbcindonesia.com 01/10/22) [2]. One of the causes of hydrometeorological disasters is the high intensity of rainfall that occurs. Rain is the result of a long process. The rainfall process results in the transformation and retransformation of mechanical energies. The influence of weightlessness keeps the clouds hanging over the sky, resulting in freefall until their growing weight is counterbalanced by gravity. Consequently, huge clouds are broken into water molecules that hit the earth as rain after traveling under gravity through the troposphere, acquiring heat via air friction, and rising in tropospheric temperature with height [3]. The previous research found that the high rainfall in the Indonesian cluster was caused by the large number of atmospheric circulations that occurred in both the Cross-Equatorial Northerly Surge and the Madden–Julian Oscillation [4–10]. The Madden–Julian Oscillation (MJO) is associated with enhanced low-level convergence and ascent in the synoptic scale, while winds in the upper troposphere diverge from the updraft. This circulation is accompanied by enhanced deep convection. Its cycle is split up into eight phases, each of which corresponds to 1/8 of the phases. An individual MJO event can last anywhere from 30 and 60 days. The increase and decrease in rainfall in the Indonesia Maritime Continent (IMC) also depend on whether the MJO convection center stops over the eastern Indian Ocean, passes through the IMC, or moves to the Western Pacific [11]. Teleconnection associated with the MJO is strongest when the MJO convection passes through Indonesia [12]. Indonesia is a known barrier to the eastward propagation of the MJO [13]. The CENS is a dominant low-level circulation that triggers deep convection and eventually leads to heavy rainfall and flood event over the equator cluster, including the northern part of Java Island [14–17]. For example, the study investigated that the strongest northerly wind played a critical role in the devastating flood event in Jakarta [14]. Similarly, a strong CENS crossing the equator line induced extreme rainfall, and a major flood occurred over Jakarta in February 2013 [18]. Many studies have been carried out to see the effect of oscillations that pass through Indonesia on the occurrence of rain in Indonesia, including analyzing the combined impact of several of these oscillations, which aims to better understand the distribution of rain throughout Indonesia and of course can predict the rainfall that occurs. Rain prediction has been developed using many methods, including a numerical weather prediction (NWP) [19]. Prediction with NWP requires a highly capable computer to carry out the prediction process for a large area and considerable

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia …

369

funding. For this reason, other methods are needed to predict extreme rain events with resources that are not too large. Besides predicting rain events, the predictions of global events such as MJO and CENS are also needed to provide early warning that the circulation can produce rain with high intensity. The VAR model is a development of the autoregressive (AR) time series model with more than one variable. All variables in the VAR model are assumed to be endogenous variables (variables whose value is determined by other variables in the model) and are interrelated. The application of VAR models to climate and weather changes attracts researchers to study past behavior, present, and future of the variables used and then used the VAR model to predict the value of the climate and weather variables for the future period [20–23]. With the case of cointegration between variables, the VAR model has not been able to model these variables, so a correction model is needed, namely the vector error correction model (VECM) for error correction in the previous VAR model. The VECM model is starting to attract attention from various fields because it can determine the relationship between one variable and another. In the case of COVID-19, the VECM technique is approached to see the relationship between air pollution and COVID-19 hospitalizations in the state of Kuwait [24]. This paper will use the rainfall data, CENS, and MJO as endogenous and interrelated variables. Based on the crucial things above, this paper aims to use the VAR method to predict rain events, CENS, and MJO to provide alerts to related parties. This paper is divided into several parts, namely Sect. 2, describing the data and methods. Section 3 contains the results and discussion, and Sect. 4 includes the conclusions of this study.

2 Data and Methodology 2.1 CENS, MJO, and Rainfall Data The rainfall data used in this study were obtained from the Global Satellite Mapping and Precipitation (GSMaP) gage-corrected (GSMaP_gage) version-7 from 2009 to 2019. GSMaP gage is a standard product of GSMaP that has been corrected by daily rain gages from NOAA CPC Gage-Based Analysis [25]. This data have a horizontal resolution of 0.1 × 0.1 degrees and can be accessed from https://sharaku.eorc.jaxa. jp/GSMaP was employed to deploy the vector autoregressive (VAR) model. To identify CENS indices, we used the daily wind component data (mean wind and wind stress field) from advanced scatterometer (ASCAT) and daily ASCAT (DASCAT). ASCAT is an actual aperture radar, operating at 5.255 GHz frequency (C-band) and using vertically polarized antennas [26]. It transmits a long pulse with linear frequency modulation (“chirp”). These data have a spatial resolution of 0.25° ×

370

M. Yollanda et al.

0.25° degrees and can be accessed from ftp://ftp.ifremer.fr/ifremer/cersat/products/ gridded/. Moreover, MJO indices data were obtained from the Climate Prediction Center (CPC) of the National Oceanic and Atmospheric Administration (NOAA) at https://www.cpc.ncep.noaa.gov/products/precip/CWlink/MJO/. The MJO Index-2 was selected over 100°E over 18 years from 2001–2018.

2.2 Building of Vector Error Correction Model 2.2.1

Stationarity Data

Before forming the vector autoregressive (VAR) model, the time series data first perform some tests on the initial data, whether the data are stationary or not. Therefore, the stationarity of time series data concerning variance or mean value can be carried out successively using the Box-Cox transformation [27] and the augmented Dickey–Fuller (ADF) test [28].

2.2.2

Model Vector Autoregressive (VAR)

The VAR model is an extension of the autoregressive (AR) time series model with more than one variable. All variables in the VAR model are assumed to be endogenous variables (variables whose value is determined by other variables in the model) and are interrelated. The general form of the VAR model with a lag p, VAR(p), can be calculated as follows [28]: X p,t = φ0 +

p 

φi X i,t−i + et , i = 1, 2, . . . , p, t = 1, 2, . . .

(1)

i=1

where X p,t is the dependent variable (X 1,t , X 2,t , . . . , X p,t ) sized p × 1, X i,t−1 is the independent variable (X 1,t−i , X 2,t−i , . . . , X p,t−i ), sized p × 1, φ0 is vector intercept size p × 1, φi is a coefficient matrix of size p × p for each i = 1, 2, . . . , p, et is the remainder vector (e1,t , e2,t , . . . , e p,t ) sized p × 1.

2.2.3

Cointegration Test

If data are cointegrated, then the modeling for that data cannot use the VAR model. The basis of cointegration is that many time series data can deviate from their average in the short term but move together. The hypothesis in Johansen’s test is as follows:

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia …

371

H0 : There are as many as r , where r = 0, 1, . . . , k − 1 cointegration equation (no cointegration or long-term relationship between variables). H1 : There are cointegration equations (cointegration occurs between variables).

2.2.4

Vector Error Correction Model (VECM)

VECM is a restricted form of VAR due to the existence of non-stationary but cointegrated data forms. The VECM specification specifies the long-term relationship of endogenous variables to converge to their cointegration relationship but still allows the presence of short-term dynamics. The VECM has the following equation: X p,t = φ p,0 + X p,t−1 +

p−1 

i∗ X p,t−i + et

(2)

i=1

where X p,t = X p,t−1 + X p,t , ∗j = −    = − 1 − 1 − 2 − . . . −  p .

2.2.5

p i= j+i

i , j = 1, 2, . . . , p − 1, and

Optimum Lag Determination

VAR estimation is very sensitive to the lag length used. Therefore, it is necessary to know the optimal lag before estimating the VAR. Determination of the optimal amount of lag can be determined using Akaike information criterion (AIC), Schwarz information criterion (SC), final prediction error (FPE), or Hannan Quinnon (HQ). Determination of the optimal lag using the information criteria by selecting the measures with the smallest value [29]. This optimal lag is applied to eliminate the autocorrelation problems in the VAR model so that they will no longer arise.

2.2.6

Impulse Response Functions (IRFs)

After modeling using the VAR model, it is necessary to have a method that can characterize the dynamic structure generated by VAR. IRF shows the response of each endogenous variable over time to the shock of the variable itself and other endogenous variables. Thus, IRF is used to see the contemporary effect of a dependent variable if it gets a shock or innovation from the independent variable by one standard deviation.

2.2.7

Mean Absolute Percentage Error (MAPE)

MAPE is a useful measure to measure the error of the estimated value of the model. The error in the estimated value of the model expresses an average absolute

372

M. Yollanda et al.

percentage of residual, which is written as follows [30]:   N 1  Yt − Yˆ  M AP E =   × 100% N t=1  Yt 

(3)

where N is the number of predictions made, Yt is the actual data, and Yˆt is the data forecast using the VAR model. The MAPE value as the modeling accuracy is indicated as “excellent” if the MAPE value < 10%, “good” if the MAPE value is 10–20%, “sufficient” if the MAPE value is 20–50%, and “inaccurate” if the MAPE value > 50%.

2.3 Rain Cluster Division In finding out the rainfall distribution in Indonesia, the Indonesian cluster is divided into 14 clusters. Because we use global-scale atmospheric circulation, the area of the cluster division is quite large. Some parts of Indonesia are not included in the territorial division due to many ocean areas, such as Maluku and East Nusa Tenggara (Fig. 1 and Table 1). The research area is divided to be 14 clusters. The division is just to simply analysis for the method that we used. It is very important to make the cluster size smaller, which can impact our prediction result. The prediction will be more precise. However, it will generate many clusters, and of course, it will cause the analysis to become more difficult since we will discuss the impact of circulations on the many clusters.

Fig. 1 Division of research area

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia …

373

Table 1 Study area coordinate Locations

Longitude 1

Longitude 2

Latitude 1

Latitude 2

1

96.3

97.3

3.8

4.8

2

100

101

0

−1

3

104

105

−3.8

−4.8

4

107

108

− 6.3

−7.3

5

109

110

−6.7

−7.7

6

111

112

−7

−8

7

117

118

−8

−9

8

110.1

111.1

1.2

0.2

9

110.1

111.1

−1.2

−2.2

10

114

115

1.2

0.2

11

114

115

−1.2

−2.2

12

121

122

1.2

0.2

13

119.6

120.6

−4

−5

14

132

133

−0.7

−1.7

3 Result and Discussion In this chapter, we will discuss vector autoregressive modeling on MJO amplitude, CENS, and rainfall time series data in Indonesia which is partitioned into 14 clusters.

3.1 Time Series Data Plot The data used are daily amplitude, CENS, and rainfall data from 14 clusters in Indonesia starting from January 2001 to December 2019 with 6939 data. The initial step is to plot daily data on amplitude, CENS, and rainfall from 14 clusters in Indonesia which can be shown in Fig. 2. Figure 2 shows Indonesia’s amplitude, CENS, and rainfall plots have an up-anddown trend over time. It means that the time series data for amplitude, CENS, and rainfall in Indonesia are not stationary or do not fluctuate at a certain number. It shows a transformation step so that the time series data to be analyzed becomes stationary in terms of the mean and variance. The pattern of peak rainfall events in Fig. 2 shows a clear relationship between the CENS index and peak rainfall that clustered rainfall can be seen in 4, 5, 6 and 7.

374

M. Yollanda et al.

Fig. 2 Time series data plot

3.2 Data Transformation The next step is to perform the Box-Cox transformation and augmented Dickey– Fuller test that will be displayed as follows: Based on Table 2, because the statistical value of the ADF test is less than the critical point value, namely −1.96, the reject H0 . It means that all variables do not have a unit root, or in other words, the data are stationary with respect to the middle value. While the Box-Cox parameters contained in Table 2 can be used to transform the variance values. Table 2 Box-Cox parameters and Dickey–Fuller augmented test statistical values Variable

Parameter Box-Cox ADF test

Variable

Parameter Box-Cox ADF test

CENS

1.902137

−4.644505 Cluster 7

−0.8781

−13.28698

Amplitude

0.907747

−2.031191 Cluster 8

0.016774

−16.15948

Cluster 1

−0.40141

−2.645304 Cluster 9

−0.22655

−17.04338

Cluster 2

−0.23157

−14.82022 Cluster 10

−0.03125

−14.56749

Cluster 3

−0.30515

−4.958705 Cluster 11

−0.16825

−15.77734

Cluster 4

− 0.11661

−15.20612 Cluster 12

−0.58326

−13.1753

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia …

375

Table 3 Results of the AIC, HQ, SC, and FPE criteria Lag

Criteria AIC

HQ

SC

FPE

1

−5.5221E + 01

2

−5.5365E + 01

−5.5069E + 01

−5.4793E + 01*

1.0422E-24

−5.5070E + 01*

−5.4535E + 01

3

9.0249E-25

−5.5385E + 01

−5.4948E + 01

−5.4152E + 01

8.8474E-25*

4

−5.5373E + 01

−5.4794E + 01

− 5.3739E + 01

8.9460E-25

5

−5.5336E + 01

−5.4614E + 01

−5.3299E + 01

9.2870E-25

6

−5.5314E + 01

−5.4450E + 01

−5.2875E + 01

9.4905E-25

7

−5.5277E + 01

−5.4270E + 01

−5.2435E + 01

9.8509E-25

3.3 Determining Optimum Lag The following are the scores for the Akaike information criterion (AIC), Schwarz information criterion (SC), Hannan-Quinn information criterion (HQ), and Akaike’s final prediction error (FPE): Table 3 shows that the lowest AIC values occur in lag 3, HQ in lag 2, SC in lag 1, and FPE in lag 3. With the diversity of the results of the optimal criteria, it is necessary to rank between criteria. After the average ranking is obtained for the four criteria, the best ranking result is lag 3. It means that the VAR model will be optimal at lag 3.

3.4 Cointegration Test The cointegration test indicates the possibility of a long-term relationship between variables. A cointegrated pair of variables indicates a long-term relationship. The result of the cointegration test shows the trace statistic of 107.9833 which is greater than the critical values of 9.24. It means that the null hypothesis H0 is rejected, or in other words, there are cointegration or long-term relationships between variables.

3.5 VECM Due to the indication of cointegration, the VECM model will be formed as a model built after correcting the previous VAR model error. Let X 1 as CENS, X 2 as amplitude, and X i , i = 3, 4, . . . , 16 as the rainfall for each cluster in Indonesia, respectively. The following are the parameters of the VECM model.

376

M. Yollanda et al.

Table 4 shows the parameter estimation of CENS, MJO, and rainfall of 14 clusters in Indonesia. The obtained parameter values are implemented to create the formula of VECM.

3.6 Test of Impulse Response Functions (IRFs) Due to the indication of cointegration, the VECM model will be formed as a model that was built after correcting the previous VAR model error. The following are the parameters of the VECM model. Figure 3 shows how the movement occurred due to the shock from the CENS phenomenon on rainfall in 14 clusters of Indonesia. It can be seen that before the movement of CENS, all clusters in Indonesia were at their equilibrium point. Then in the next period, several clusters began to show their response to the CENS phenomenon, both positive responses that occurred in clusters 1, 2, 3 and 4. In contrast, the rest gave a negative response to the CENS phenomenon. Then in the next period, it was seen that the movement of the amount of rainfall in 14 clusters of Indonesia began to reach the equilibrium point, which means that the response given to all 14 clusters began to reach stability by showing that the standard deviation was convergent toward a value of 0. Figure 4 shows how the movement occurred due to the shock of the amplitude phenomenon on rainfall in 14 clusters of Indonesia. It can be seen that before the movement of amplitude, all clusters in Indonesia were at their equilibrium point. Then in the next period, several clusters began to show their response to the CENS phenomenon, both negative responses that occurred in the 6, 7, 8, 9, 10 and 13 clusters to the amplitude phenomenon. In the second period, it can be seen that there are all 14 cluster’s rainfall responses to the amplitude phenomenon started to converge to point 0, except for clusters 10 and 11. In the movement given the amplitude phenomenon to rainfall in 10 clusters, rainfall in the cluster begins to show a negative response to rainfall. In contrast to the rainfall in cluster 11, the amplitude phenomenon began to respond positively. Then in the next period, it was seen that the movement of the amount of rainfall in 14 clusters of Indonesia began to reach the equilibrium point, which means that the response given to all 14 clusters began to reach stability by showing that the standard deviation was convergent toward a value of 0.

3.7 Comparison of the Original Data and the Estimated Data Using the VAR Model Figure 5 shows how the data on the estimated CENS variables, amplitude, and intensity of rainfall in 14 clusters of Indonesia compare with the observed data. Based on Fig. 5, it can be seen that the data from the VECM modeling (blue) are getting close

3.6e-04 0.002 −

−

−

−

−

−

−

−

−

−

−0.615 −

−0.349 −

−0.544 −

−

0.737

X2

X3

X4

X5

X6

X7

X8

X9

X10

X11

−

0.002

X4

−0.031 −

−0.025 −

−

−

−

−

−

−0.447 −

−

−

X12

X13

X14

X15

X16

−

−

−

−

−

0.716

−

−

−

−

0.033

0.049

0.059

−

−

−

−

0.038

−

−

−

−

0.054

−

−

−

−

−

−

0.137

−

−

−

−

−

−

0.023

−

−

−

−

X12

−

−

−

−

−

−

−

X13

−

−

−0.104 0.037

−0.036 0.045

−0.042 −

−

−

−

0.04

−

0.04

−

0.047

−

−

−

0.037

−

−0.068 -0.483

−0.785 0.050

X15

−

−

−

−

−

−

−

−

−

0.039

0.035

−

−

−0.188

−0.592

−0.184

−0.086

−0.052

−0.150

−0.061

−

−0.094

0.043

−0.498 −

−0.441 −0.029 −

−

−0.057 0.053 −0.480 −

−

0.164 −0.092 0.765

−0.049 0.033

0.047

−

−

−

−

−0.006

ECT 9.8e-04 −4.5e-04

−

X16

−0.046 −0.057 0.131

−

−

−

−

−

−

−

−

−

−

−0.003 −

X14

−0.118 −

-0.048

−

−0.466 −0.109 −

−0.045 −

0.037

−

−0.071 −

−

0.008

−

−0.026 -0.044

−0.459 0.031

−0.152 −0.059 0.079 −

−

−0.093 −

−0.048 −0.058 − -0.069

X11

−0.061 −

−

−0.453 −

−0.049 −

−

X10

−0.005 −0.005 −0.004 −0.007 -

X9

−0.513 −

−0.040 0.049

−

−

0.071

0.075

−

−

−

−

−0.499 −

−0.460 − −

−

−

X8

−0.070 -

−

−

X7

−0.099 −

−

−

0.004

X6

-0.465 −

−0.054 −

0.086

0.184

0.070

0.057

−

0.023

-0.042

0.031

0.022

0.036

−0.657 0.108

−

0.027

0.043

−

0.172

−

−

X5

−0.474 −0.031 −

−

−

−2.52

X1

X3

X2

X1

Model Parameter

Table 4 Parameter estimate of VECM for CENS, amplitude, and rainfall in Indonesia

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia … 377

378

M. Yollanda et al.

Fig. 3 Effect of CENS on rainfall intensity in 14 clusters

Fig. 4 Effect of amplitude on rainfall intensity in 14 clusters

to the observed value (red). It indicates that the VECM modeling results obtained are pretty good for predicting the values of CENS, amplitude, and intensity of rainfall for 14 clusters in Indonesia for the next period.

3.8 VECM Modeling Accuracy Values To determine the accuracy of the VECM model, the following is the MAPE value for each variable in Table 5. In explaining the model’s accuracy for each variable, the MAPE value is required. The MAPE value that is lower than 10 indicates that the model for each variable has better performance to be applied in forecasting for the next period. Based on Table 5, all MAPE values are approaching zero or lower than 10% so that all the VECM

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia …

Fig. 5 Grafik Perbandingan data Aktual dan Pemodelan VECM

379

380

M. Yollanda et al.

Table 5 MAPE values Variable

MAPE(%)

Variable

MAPE(%)

Variable

MAPE(%)

CENS

1.852

Cluster 5

0.489

Cluster 10

0.967

Amplitude

0.062

Cluster 6

0.578

Cluster 11

0.821

Cluster 1

0.945

Cluster 7

0.438

Cluster 12

0.611

Cluster 2

1.327

Cluster 8

0.730

Cluster 13

0.548

Cluster 3

0.626

Cluster 9

0.741

Cluster 14

0.683

Cluster 4

0.478

−

−

−

−

models obtained are very good at predicting the value of each variable for the future period.

3.9 Forecasting Results for the Following Period The forecasting value will be shown along with the forecasting results’ lower and upper limits on January 1st 2019 in this following Table 6. Table 6 shows the forecasting value for Indonesia’s CENS, amplitude, and rainfall. Besides, there are also the lower and upper values as the forecasting ranges for one period ahead. Based on Table 6, most of the actual values are already within the Table 6 Forecasting value in one day ahead Variable

Forecast

Lower

Upper

The actual value

Conclusion

CENS

−5.034

−6.854

−3.214

−4.068

Within range

Amplitude

0.333

−0.292

0.959

0.370

Within range

Cluster 1

−1.290

−29.969

27.388

0.000

Within range

Cluster 2

15.679

−19.224

50.582

0.004

Within range

Cluster 3

24.637

4.943

44.331

0.731

Out of range

Cluster 4

37.144

21.088

53.201

25.171

Within range

Cluster 5

47.288

29.676

64.901

12.357

Out of range

Cluster 6

10.425

−6.285

27.135

12.059

Within range

Cluster 7

6.451

−9.339

22.241

11.508

Within range

Cluster 8

29.091

5.167

53.016

18.055

Within range

Cluster 9

29.368

4.077

54.660

9.339

Within range

Cluster 10

32.813

10.066

55.560

61.631

Out of range

Cluster 11

28.009

3.267

52.751

29.537

Within range

Cluster 12

4.695

−16.260

25.650

21.674

Within range

Cluster 13

20.549

3.035

38.063

20.369

Within range

Cluster 14

−0.454

−21.555

20.646

0.153

Within range

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia …

381

forecasting range except for clusters 3, 5 and 10. They are still out of the forecasting range because daily forecasts are still relatively unstable and fluctuate freely.

4 Conclusions Rainfall in the 14 clusters over Indonesia is forecasted to be stable and within the prediction range in the upcoming periods. Similar trends are also observed for the CENS and amplitude of the MJO where the predicted data are not beyond the lower and upper forecasted value. In other words, this condition is the same as in the previous periods fluctuating at the same point, which neither overestimates nor underestimates. Furthermore, the comparison of rainfall prediction results in those clusters with VECM, and actual data can be seen by the mean absolute percentage error (MAPE) test values which were barely under 1 (good performance) for the whole study areas. In addition, the amplitude of the MJO data also shows a fairly enough performance in which the MAPE values are close to zero. Unfortunately, the MAPE value of forecasting CENS is clearly observed at more than 1%, precisely at 1.852. Somehow, this value remains considerably categorized as a good performance. To sum up, the statements above clearly indicate that the VECM performs well in forecasting the CENS, amplitude of the MJO, and rainfall over Indonesia. Therefore, this VECM is recommended to be implemented in developing the early warning system to mitigate and manage the upcoming hazards in Indonesia. For getting better results and more supporting evidence of this performance, however, it is suggested to carry out further this study in predicting other global phenomena such as Cold Surge (CS), Southerly Surge (SS), and Indian Oscillation Dipole (IOD) and including their relationships with extreme rainfall in Indonesia.

References 1. BNPB.: Geoportal Data Bencana Indonesia. https://gis.bnpb.go.id/. Last Accessed 04 Oct 2022 2. Dob.: RI Supermarket Bencana, Sampai September Ada 2000 Bencana—CNBC Indonesia. https://www.cnbcindonesia.com/news/20201001104931-4-190809/ri-supermarket-bencanasampai-september-ada-2000-bencana. Last Accesed 01 Oct 2022 3. Amajama, J.: Physics of rainfall. J. Scientif. Eng. Res. 3(1), 51–54 (2016) 4. As-syakur, A.R., Osawa, T., Miura, F., Nuarsa, I.W., Ekayanti, N.W., Dharma, I.G.B.S., Adnyana, I.W.S., Arthana, I.W., Tanaka, T.: Maritime continent rainfall variability during the TRMMera: the role of monsoon, topography and El Niño Modoki. Dyn. Atmos. Oceans 75, 58–77 (2016) 5. Peatman, S.C., Matthews, A.J., Stevens, D.P.: Propagation Of the Madden–Julian oscillation through the Maritime Continent and scale interaction with the diurnal cycle of precipitation. Q. J. R. Meteorol. Soc. 140, 814–825 (2014) 6. Saufina, E., Trismidianto, Risyanto, Fathrio, I., Harjupa, W.: Impact of cross-equatorial northerly surge (CENS) on Jakarta heavy rainfall and its interaction with a tropical cyclone

382

7.

8.

9.

10.

11. 12.

13.

14.

15.

16. 17.

18.

19.

20.

21.

M. Yollanda et al. (Case study: 18–25 February 2020). In: AIP Conference Proceedings, vol. 2366, pp. 050002. (2021). https://doi.org/10.1063/5.0059995 Purwaningsih, A., Tank, A.K., Vila, J.: Atmospheric conditions associated with northerly Surge, Borneo Vortex and Madden Julian Oscillation during the extreme rainfall cases in early 2021 over the Western Part of the Maritime Continent. In: Yulihastin, E., Abadi, P., Sitompul, P., Harjupa, W. (eds.) Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, 2021. Springer Proceedings in Physics, vol. 275, Springer, Singapore (2022). https://doi.org/10.1007/978-981-190308-3_57 Lestari, D.O., Sutriyono, E., Sabaruddin, Iskhaq Iskandar.: Respective influences of Indian Ocean Dipole and El NiñoSouthern Oscillation on Indonesian precipitation. J. Mathem. Fundamental Sci. 50(3), 257–272 (2018) Kurniadi, A., Weller E., Ki Ming, S., Gyu Seong, M.: Independent ENSO and IOD impacts on rainfall extremes over Indonesia. Int. J. Climatol. 41(2) (2021). https://doi.org/10.1002/joc. 7040 Hadi, T., Horinouchi, T., Tsuda, T., Hashiguchi, H., Fukao, S.: Sea-breeze circulation over Jakarta, Indonesia: a climatology based on boundary layer radar observations. Mon. Weather Rev. 130(9), 2153–2166 (2001) Zhang, C.: Madden–Julian oscillation: bridging weather and climate. Bulletin of the American Meteorol. Soc. 94, 1849–1870 (2013). https://doi.org/10.1175/BAMS-D-12-00026.1 Adames, Á.F., Wallace, J.M.: Three-dimensional structure and evolution of the MJO and its relation to the mean flow. J. Atmos. Sci. 71(6), 2007–2026 (2014). https://doi.org/10.1175/ JAS-D-13-0254.1 Nitta, T., Mizuno, T., Takahashi, K.: Multi-scale convective systems during the initial phase of the 1986/87 El Niño. J. Meteorol. Soc. Japan Ser. II 70(1B) 447–466 (1992). https://doi.org/ 10.2151/jmsj1965.70.1B_447 Wu, P., Hara, M., Fudeyasu, H., Yamanaka, M.D., Matsumoto, J., Syamsudin, F., Sulistyowati, R., Djajadihardja, Y.S.: The impact of trans-equatorial monsoon flow on the formation of repeated torrential rains over Java Island. Sola 3, 93–96 (2007) Tangang, F.T., Juneng, L., Salimun, E., Vinayachandran, P.N., Seng, Y.K., Reason, C.J.C., Behera, S.K., Yasunari, T.: On the roles of the northeast cold surge, the Borneo vortex, the Madden-Julian oscillation, and the Indian Ocean Dipole during the extreme 2006/2007 flood in southern Peninsular Malaysia. Geophys. Res. Lett. 35, L14S07 (2008). https://doi.org/10. 1029/2008GL033429 Yokoi, S., Matsumoto, J.: Collaborative effects of cold surge and tropical depression-type disturbance on heavy rainfall in central Vietnam. Mon. Weather Rev. 136(9), 3275–3287 (2008) Chen, T.-C., Tsay, J.-D., Matsumoto, J., Alpert, J.: Development and formation mechanism of the Southeast Asian winter heavy rainfall events around the South China Sea. Part I: formation and propagation of cold surge vortex. J. Climate 28(4), 1417–1443 (2015). http://www.jstor. org/stable/26194409 Yulihastin, E, Trilaksono, N.J.: Evolution of heavy rainfall in Jakarta flood case 2013 based on COSMO model. In: Asia Pacific Centre for Ecohydrology-UNESCO Research Center for Limnology-Indonesian Institute of Sciences, 167 (2014). https://doi.org/10.13140/2.1.3913. 9525 Pu Z., Kalnay E.: Numerical weather prediction basics: models, numerical methods, and data assimilation. In: Duan, Q., Pappenberger, F., Thielen, J., Wood, A., Cloke, H., Schaake J. (eds.) Handbook of Hydrometeorological Ensemble Forecasting. Springer, Berlin, Heidelberg (2018) Shahin, M.A., Ali, M.A., Ali, A.B.M.S.: Vector autoregression (VAR) modeling and forecasting of temperature, humidity, and cloud coverage. In: Computational Intelligence Techniques in Earth and Environmental Sciences. Springer, Dordrecht (2014) Adenomon, M.O., Oyejola, B.A.: Forecasting meteorological time series data with a reduced form vector autoregressive (Var) model and three univariate time series techniques: a comparative study. Soc. Basic Sci. Res. Rev. 2, 139–152 (2014)

Prediction of CENS, MJO, and Extreme Rainfall Events in Indonesia …

383

22. Ramli, I., Rusdiana, S., Basri, H., Munawar, A.A., Azelia, V.: Predicted rainfall and discharge using vector autoregressive models in water resources management in the high hill Takengon. In: IOP Conference Series: Earth and Environmental Science, vol. 271 (2019) 23. Abdallah, W., Abdallah, N., Marion, J., Oueidat, M., Chauvet, P.: A vector autoregressive methodology for short-term weather forecasting: tests for Lebanon. SN Appl. Sci. 2, 1–9 (2020) 24. Alsaber, A.R., Setiya, P., Al-Sultan, A.T., Pan, J.: Exploring the impact of air pollution on COVID-19 admitted cases. Japanese J. Statist. and Data Sci. 5, 379–406 (2022) 25. Chen, M., Shi, W., Xie, P., Silvia, V.B.S., Kousky, V.E., Higgins, R.W., Janowiak, J.E.: Assessing objective techniques for gauge-based analyses of global daily precipitation. J. Geophys. Res. 113, D04110 (2008). https://doi.org/10.1029/2007JD009132 26. Klaes, K.D.: The EUMETSAT polar system. Comprehensive Remote Sensing 192–219 (2018) 27. Brockwell, P.J.: In: Introduction to Time Series and Forecasting, 3rd edn. Springer Nature, Switzerland (2016) 28. Wei, W.W.S.: Time Series Analysis. Addison Wesley, New York (2006) 29. Lutkepohl, H.: New Introduction to Multiple Time Series Analysis. Springer, Berlin (2005) 30. Gentle, J.: Statistical Analysis of Financial Data: With Examples In R. CRC Press, Boca Raton (2020)

Implementation of Zero Runoff to Reduce Runoff Discharges in Timbang Langsa Village, Langsa City Eka Mutia, Ellida Novita Lydia, Wan Alamsyah, and Danil Rahmad Priatna

Abstract Factors causing flooding include high rainfall, reduced water catchment areas, blockages due to indiscriminate garbage disposal, tree cutting and reduced forest area. The reduction in the catchment area can lead to an increase in the amount of rainwater runoff that occurs on the surface. Flooding can also occur due to the dimensions of the channel that are not able to accommodate the flood discharge or the collector channel that is not able to accommodate the water discharge from the Interceptor Channel. Langsa City is currently carrying out relocation activities for residents living on the banks of the Krueng Langsa river. The relocation site is located in Timbang Langsa village. This community relocation area was originally an oil palm plantation owned by PT. Timbang Langsa who is no longer productive. Changes in land use are seen as one of the causes of flooding in the area. Based on this, it is necessary to develop a way to reduce Runoff discharge, by making an alternative land cover. The purpose of this study is to reduce the flood discharge that occurs with a change from the value of runoff coefficient (C). The research method used is a survey method. The results obtained that the runoff discharge (Q) value was significantly reduced on the land cover as a mixed garden with a runoff discharge (Q) value ranging from 0.0077 to 0.0718 m3 /sec from the previous runoff discharge (Q) value ranging from 0.0187 to 0.0979 m3 /sec.

E. Mutia · E. N. Lydia (B) · W. Alamsyah · D. R. Priatna Universitas Samudra, Jalan Prof. Dr. Syarif Thayeb, Meurandeh, Langsa-Aceh, Indonesia e-mail: [email protected] E. Mutia e-mail: [email protected] W. Alamsyah e-mail: [email protected] D. R. Priatna e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_36

385

386

E. Mutia et al.

1 Introduction Rapid population growth can cause serious environmental problems in the world [1–3]. The increase in population will certainly be followed by development developments, causing reduced infiltration functions and the construction of artificial drainage networks can also change the natural path of water/runoff [4–6]. Surface runoff is an important component in the hydrological cycle that occurs in urban areas. This is related to changes in land use which result in changes in land cover in urban areas [7]. The consequences that can occur from population growth include increased land cover, soil compaction, vegetative cover, development of rainwater drainage systems, infiltration, groundwater recharge and reduced evapotranspiration. Surface runoff and peak flow will increase in volume and magnitude because the runoff will be quickly discharged into water bodies [8, 9]. Langsa City is currently making improvements to the riverbank area which was originally a densely populated area to be used as a waterfront tourist destination. The community previously lived on the banks of the river and then relocated to a new place. The relocation site was originally an oil palm plantation. Changes in land use at the two locations are an effort to reduce the flood discharge that occurs, but currently in the relocation area, flooding is still happening and it is difficultly to have enough clean water. This phenomenon is interesting to study to obtain an appropriate method to overcome the problem of flooding in coastal areas. The purpose of this study is to obtain an appropriate runoff coefficient based on the planned land cover to reduce runoff discharge. The reduction of runoff discharge aims to make the infiltration that occurs greater than runoff; this aims to overcome floods in Timbang Langsa village.

2 Methods 2.1 Study Area The research will be conducted in the relocation area of Timbang Langsa village, Langsa Baro sub-district, Langsa City, Aceh, Indonesia (see Fig. 1). Geographically, the research location is located between latitude 4°32 4.35 N−4°32 7.24"N and longitude 97°55 28.22 E−97°55 59.29"E in Langsa City, Aceh, with an area of 0.097 km2 [10]. There is a lot of rainfall throughout the year in Langsa. The most rainfall in Langsa city occurs in Mei, with an average rainfall of 170 mm. January has the least rainfall, with an average rainfall of 1010 mm [11]. The main types of vegetation found at the research site were weeds, ornamental plants, fruit plants and tubers. The livelihoods of the majority of the population are entrepreneurship, employees and traders.

Implementation of Zero Runoff to Reduce Runoff Discharges …

387

Fig. 1 Research site map

2.2 Data Collection and Processing The data needed in this study are measurement data of the research area, elevation data of research area, rainwater catchment delineation data and Qanun Kota Langsa number 12 of 2013 concerning the Langsa City Spatial Plan for 2012−2032. This type of research is quantitative research with a descriptive survey method. The research stage consists of three stages, namely the preparation stage, the survey stage and the data processing and analysis stage. • Stages of preparation At this stage, the activities carried out are secondary data collection consisting of maximum daily rainfall data for the city of Langsa in 2009 – 2020. • Survey stage in the field At this stage, the activities carried out are primary data collection consisting of data on the area of the flow area through declination and elevation measurements. The primary data were obtained by surveying the field. • The data processing and analysis stage consist of: – The stages of research implementation begin with the delineation of rainwater catchment areas – Survey data calculation analysis

388

E. Mutia et al.

– Hydrological analysis to determine runoff discharge using a modified rational formula [12] Q = 0.278C.Cs.I.A

(1)

Description: Q: C: Cs: I: A:

Runoff discharge Runoff coefficient Retention coefficient Rain intensity Area of stream.

• The final stage is to recommend the zero runoff concept to be used

3 Results and Discussion 3.1 Runoff Coefficient The studied conducted in a residential area which is relocation location for people who previously lived on the banks of the Krueng Langsa river. The allotment of land at this location is a coupling housing, a yard that functions as a green open space and an asphalt road. By Fig. 1, it can be seen that the study area is divided into 33 watersheds. The existing watersheds have almost the same shape and number of houses as one. So what will be shown in Tables 1, 2, 3 and 4 are several watersheds, namely A, C, N, O, Q, Z6 and Z7 watersheds. The runoff coefficient (C) will be calculated for the C value for housing and yard, while the C value for asphalt is not taken into account. This is because the runoff discharge that will be accommodated is runoff that is in the yard of each house. Table 1 Runoff Area Runoff area

Number of houses (couple)

The yard area (km2 )

Residential area (km2 )

Area of runoff (km2 )

Value of C 0.3638

A

6

0.0063

0.0004

0.0067

C

12

0.0010

0.0011

0.0021

0.4807

N

22

0.0033

0.0018

0.0051

0.4374

O

32

0.0032

0.0026

0.0058

0.4612

Q

12

0.0026

0.0009

0.0035

0.4159

Z6

2

0.0013

0.0005

0.0018

0.4226

Z7

2

0.0015

0.0002

0.0017

0.3787

Implementation of Zero Runoff to Reduce Runoff Discharges …

389

Table 2 Runoff coefficient value Runoff area

Area of runoff (km2 )

C value (land)

C value (grass)

C value (mixed garden)

C value (paving)

A

0.0067

0.3638

0.3260

0.1276

0.6945

C

0.0021

0.4807

0.3136

0.3136

0.3136

N

0.0051

0.4374

0.1480

0.0477

0.3342

O

0.0058

0.4612

0.4616

0.3613

0.6477

Q

0.0035

0.4159

0.2099

0.2099

0.2099

Z6

0.0018

0.4226

0,2016

0.0650

0.4552

Z7

0.0017

0.3787

0.4114

0.2749

0.6650

Table 3 Runoff discharge (m3 /det) Runoff area

Area of runoff (km2 )

Q value (land)

Q value (grass)

Q value (mixed garden)

Q value (paving)

A

0.0067

0.0337

0.0297

0.0100

0.0711

C

0.0021

0.0979

0.0937

0.0718

0.1353

N

0.0051

0.0746

0.0700

0.0459

0.1153

O

0.0058

0.0317

0.0299

0.0208

0.0482

Q

0.0035

0.0260

0.0238

0.0130

0.0460

Z6

0.0018

0.0785

0.0729

0.0437

0.1289

Z7

0.0017

0.0187

0.0169

0.0077

0.0341

Table 4 Clean water needs Stream area

Number of Q houses value—m3/s (couple) (soil)

Q value—m3/s (mixed garden)

Water needs (m3/day)

Reservoir (3 days)—m3

Home

Flow area

Home

Flow area 11,736

A

6

0.0337

0.0100

0.652

3912

1956

C

12

0.0979

0.0718

0.652

7824

1956

23,472

N

22

0.0746

0.0459

0.652

14,344

1956

43,032

O

32

0.0317

0.0208

0.652

20.864

1956

62.592

Q

12

0.0260

0.0130

0.652

7824

1956

23,472

Z6

2

0.0785

0.0437

0.652

1304

1956

3912

Z7

2

0.0187

0.0077

0.652

1.304

1956

3912

E. Mutia et al.

Time (Hours)

390 0.6000 0.5000 0.4000 0.3000 0.2000 0.1000 0.0000

Tc jam

0.5023

0.4563

0.2781

0.2306

0.1719 0.0300

A

C

N O Q Runoff Area

0.0293

Z6

Z7

Fig. 2 Time concentration

The area of the study area is 0.097 km2 , with the value of C in each flow area in Table 1 following (the value of C is calculated for only a few different typical flow areas). The value of C is taken with the assumption that all open land is clay. The value of C in each flow area ranges from 0.3 to 0.5. The lowest value of C found in runoff area A. In that area, the house area is only 5.97% of the area of runoff A. The highest value of C found in flow area C with a Residential area of 52.38%. This shows that the more green open space, the greater the infiltration of flood discharge.

3.2 Time Concentration (Tc) The time concentration (Tc) will determine the length of time the water flows. This calculation is to estimate the time needed to accommodate the flow in a container for further storage as a source of clean water. The calculation of the concentration time can be in Fig. 2. In Fig. 2, it can be seen that the lowest concentration time is in the Z6 flow area with a Tc value of 0.0293 and the highest Tc is in the O flow area with a value of 0.5023. The value of Tc depends on the shape of the flow area. The shape of the short flow area will result in a shorter time compared to the shape of the elongated flow area [13].

3.3 Runoff Discharge (Qt) Runoff discharges the research site to calculate how much discharge is generated at the site. Runoff discharge will then be reduced by harvesting rainwater so the flood problem can be resolved (Fig. 3). Runoff discharge ranges from 0.02 to 0.10 m3 , depending on the overall flow area outside.

Runoff Discharge (m3/det)

Implementation of Zero Runoff to Reduce Runoff Discharges …

0.1200 0.1000 0.0800 0.0600 0.0400 0.0200 -

391

0.0979 0.0785

0.0746 0.0337

A

C

Qt m3/det

0.0317 0.0260

N O Q Z6 Runoff Area

0.0187

Z7

Fig. 3 Runoff discharge

3.4 Clean Water Needs Everyone’s need for clean water is different. In Indonesia, the need is based on SNI 6728. 1: 2015 concerning the preparation of the spatial balance of natural resources, the water demand for medium-sized cities is 100–125 L/per person/day [14]. Clean water at the research site is sourced from the Regional Drinking Water Company (PDAM) Tirta Kemuning, Langsa City. The community is still experiencing difficulties with clean water because its distribution is not yet smooth. At the research location, it is assumed that if in 1 house there are 5 people, then the need for clean water is 625 L/person/day. We will try to fulfil this need for clean water by harvesting rainwater from runoff.

3.5 Reduction of Flood Discharge Reduction of runoff that occurs at the research site will be carried out by changing the C value. Several alternative land covers will be tried to make recommendations on which one has a lower C value so that it will reduce the resulting Q (Fig. 4). Based on Tables 2 and 3, the smallest C value and Q value occur in land cover used as mixed gardens. So if the open space in each housing is mostly used as mixed gardens, it can reduce runoff discharge by 25−70%. The next step is to reduce runoff by harvesting rainwater. The amount of clean water needed in each watershed will be different according to the number of houses built. The largest water demand is in the O runoff area, because it has more houses than other runoff area, but has a low Q value. Further research is needed to calculate the adequacy of clean water when compared to the existing plan.

392

E. Mutia et al. Z7

Q Value Comparison

Runoff Area

Z6 Q

Q value (paving)

O

Q value (mixed garden)

N

Q value (grass) Q value (land)

C A 0.0000

0.0500

0.1000

0.1500

Runoff Discharge (m3/det)

Fig. 4 Comparison of runoff discharge value

4 Conclusion The appropriate runoff coefficient value is a C value of 0.1276−0.2749 with mixed garden land cover types. Land cover with mixed gardens reduces runoff discharge by 25−70%, so that the community using mixed garden land cover types will reduce the flood discharge that occurs. Acknowledgements The author would like to thank LPPM and PM of Samudra University who have provided research funds and all related parties who have helped so that this research can run smoothly. This research was funded by the University of Samudra’s DIPA in 2022.

References 1. Liu, W. et al.: Identifying city-scale potential and priority areas for retrofitting green roofs and assessing their runoff reduction effectiveness in urban functional zones. J. Clean Prod. 332 (2022). https://doi.org/10.1016/j.jclepro.2021.130064 2. Chan, F.K.S., et al.: ‘Sponge City’ in China—a breakthrough of planning and flood risk management in the urban context. Land Use Policy 76, 772–778 (2018). https://doi.org/10.1016/j.lan dusepol.2018.03.005 3. Fang, C., et al.: Modeling regional sustainable development scenarios using the urbanization and eco-environment coupler: case study of Beijing-Tianjin-Hebei urban agglomeration, China. Sci. Total Environ. 689, 820–830 (2019). https://doi.org/10.1016/j.scitotenv.2019.06.430 4. Petroselli, A., Wał˛ega, A., Mły´nski, D., Radecki-Pawlik, A., Cupak, A., Hathaway, J.: Rainfallrunoff modeling: a modification of the EBA4SUB framework for ungauged and highly impervious urban catchments. J. Hydrol. (Amst) 606 (2022). https://doi.org/10.1016/j.jhydrol.2021. 127371 5. Lizárraga-Mendiola, L. et al.: Hydrological design of two low-impact development techniques in a semi-arid climate zone of central Mexico. Water (Switzerland) 9(8) (2017). https://doi.org/ 10.3390/w9080561

Implementation of Zero Runoff to Reduce Runoff Discharges …

393

6. Walega, A., Salata, T.: Influence of land cover data sources on estimation of direct runoff according to SCS-CN and modified SME methods. Catena (Amst) 172, 232–242 (2019). https:// doi.org/10.1016/j.catena.2018.08.032 7. Chang, B., Wherley, B., Aitkenhead-Peterson, J.A., McInnes, K.J.: Effects of urban residential landscape composition on surface runoff generation. Sci. Total Environ. 783 (2021). https:// doi.org/10.1016/j.scitotenv.2021.146977 8. Hong, W., Guo, R., Tang, H.: Potential assessment and implementation strategy for roof greening in highly urbanized areas: a case study in Shenzhen, China. Cities 95 (2019). https:// doi.org/10.1016/j.cities.2019.102468 9. Walsh, C.J., et al.: Principles for urban stormwater management to protect stream ecosystems. Freshwater Sci. 35(1), 398–411 (2016). https://doi.org/10.1086/685284 10. Badan Pusat Statisik Kota Langsa,: Kecamatan Langsa Bar˝o Dalam Angka 2022. Langsa, Badan Pusat Statistik Kota Langsa (2022) 11. Badan Pusat Statistik Kota Langsa.: Kota Langsa dalam Angka (Langsa Municipality in Figures) 2022. Langsa, Badan Pusat Statistik Kota Langsa (2022). 12. Mutia, E., Alamsyah, W.: Penataan Jaringan Drainase Berdasarkan Tata Ruang Kota Langsa. Jurutera 2(1), 38–47 (2015) 13. Suripin.: Sistem Drainase Perkotaan yang BerkelanjutanSistem Drainase Perkotaan yang Berkelanjutan. Jakarta, Penerbit Andi (2004) 14. Afriyanda, R., et al.: Analisis Kebutuhan Air Bersih Domestik Di Desa Penjajap Kecamatan Pemangkat Kabupaten Sambas. PWK, Laut, Sipil, Tambang 6(2), 1–11 (2018)

Monthly Rainfall Prediction Using Vector Autoregressive Models Based on ENSO and IOD Phenomena in Cilacap Fadli Nauval, Mutia Yollanda, Dodi Devianto, Wendi Harjupa, Dita Fatria Andarini, Elfira Saufina, Anis Purwaningsih, Fahmi Rahmatia, Ridho Pratama, Trismidianto, Teguh Harjana, Risyanto, and Didi Satiadi Abstract A study of El Nino-Southern Oscillation (ENSO) and Indian Oscillation Dipole (IOD) as climate elements to predict extreme rainfall is critical to be carried out. This study aims to forecast monthly rainfall by using Vector Autoregressive (VAR) model in the Cilacap. Rainfall data were obtained by the Global Satellite Mapping of Precipitation (GSMaP) and ENSO and IOD by the National Oceanic and Atmospheric Administration (NOAA) from March 2000 to December 2018. The data was used to predict monthly rainfall for the next twelve months in 2019. The VAR model with a minimum lag of length 2 was considered to be the best lag and selected to model the rainfall in the study area. Response function reveals that changes in rainfall significantly affect changes in rainfall after some time lags. The results of the actual data accuracy test with predictive data from the Vector Autoregressive (VAR) model were conducted with the root mean square error (RMSE) and the mean absolute percentage error (MAPE). The RMSE value was 0.3291, 0.047, 0.061 and the MAPE value was 87.0671, 0.1845, 0.217. Rainfall in the Cilacap is predicted to be stable in the future. This condition is the same as the rainfall in the previous periods, which fluctuated at the same point.

1 Introduction Indonesia, which is well-known as Indonesia Maritime Continent (IMC) is a county with a tropical climate that is vulnerable to extreme rainfall [1]. Moreover, the IMC experiences the impact of climate change in which the air temperature in Indonesia has increased by 0.30 °C since 1900. This increase in temperature occurred F. Nauval (B) · W. Harjupa · D. F. Andarini · E. Saufina · A. Purwaningsih · F. Rahmatia · R. Pratama · Trismidianto · T. Harjana · Risyanto · D. Satiadi Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, Indonesia e-mail: [email protected] M. Yollanda · D. Devianto Department of Mathematics, UNAND, West Sumatera, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_37

395

396

F. Nauval et al.

throughout the season. Changes in weather and seasons are marked by an increase in rainfall (La Nina) that occurs in an area, while in other regions, there is a decrease in rainfall (El Nino) of 2–3% [1]. Variability in the IM is known to be modulated by two ocean–atmosphere phenomena, El Nino–Southern Oscillation (ENSO) and the more recently described Indian Ocean Dipole (IOD; [2, 3]). The possible linkages between ENSO and IOD are also an active area of investigation [4]. Furthermore, although climate teleconnections are of a global scale, local modulations—forced, for example, by fine-scale topography of terrain or land-sea contrasts—may play an important role. Therefore, understanding the role of multiscale interaction in regional climate is critical for climate prediction and change studies and applications in weather- and climaterelated risk management. Java Island is Indonesia’s most populated island and most important industrial and agricultural region. It is located in the deep tropics of the Southern Hemisphere, at the center of the Asian–Australian monsoon region [5]. It is zonally elongated and slightly tilted in the northwest-southeast direction, with a central mountain range that runs the length of the island. Based on data on disaster events in Indonesia from 2001 to 2019 released by the National Disaster Management Agency (BNPB), it is known that Java Island, especially Central Java Province, is the area with the highest hydrometeorological disasters, both floods, landslides, and strong winds (BNPB disaster plot). The spatial distribution of disaster events, especially floods in Central Java, shows that the most dominant area experiencing flooding is Cilacap (spatial plot of the distribution of flood events in Central Java), so it is necessary to predict rain in the Cilacap area as an effort to mitigate disasters in the area. Figure 1 depicts Indonesia’s number of hydrometeorological disasters from 2000 to 2019. Overall, the hazard taking place the most in the given period was flood, and the area often experiencing the catastrophic was Jawa Tengah. Figure 2 illustrates how many events of flood disasters across the Central Java province were observed in each district from 2000 to 2019. In general trend, Cilacap was an area that dealt with flood disasters the most. Moreover, the number of flood hazards in Central Java dominantly occurred in the southern and northern parts in the given period. As for southern Central Java, the rate of flood events was higher than the middle areas such as Cilacap, Banyumas, Kebumen, and Wonogiri at 155, 66, 56, and 62, respectively, while the areas of Purbalingga, Banjarnegara, and Wonosobo only experienced floods at less than 20 events in the last two decades. Furthermore, a similar trend in the southern part also took place in the northern areas. Changes in rain patterns every day and every month are very important to take action. More accurate predictions are needed to avoid large losses of property and lives. Rain prediction using a numerical model has been carried out. One of the models used is Numerical Weather Prediction (Pu and Kalnay 2018). There are still shortcomings where the accuracy of using numerical models is still small [6]. For this reason, another approach is needed so that it can produce more accurate rain predictions. One of the common prediction methods is the Vector Autoregressive (VAR) method. VAR was initially developed by an Econometric expert, Christopher A.

Monthly Rainfall Prediction Using Vector Autoregressive Models …

397

Fig. 1 The number of occurrences in Indonesia

Fig. 2 Distribution of flood disasters in Central Java

Sims, as an alternative model approach to the multiple equation models with consideration of a theoretical approach that aims to be able to capture economic phenomena well [7]. VAR is a dynamic equation system. The Vector Autoregressive (VAR) model can be used to predict time series data for more than one variable, whereas in this model, we do not need to distinguish between independent variables and variables [7]. In the VAR model, all variables are endogenous (bound). With a high level of

398

F. Nauval et al.

accuracy, the VAR model can be used for operational purposes so that it can be a reference for those who need it. In the meteorology field, the VAR model has been used to improve weather forecast system. In Lebanon, the parameters of weather such as temperature, humidity, and precipitation are implemented to evaluate the built model [8]. The results were very promising (average of about 96.67% of precision between real values and three predicted parameters: temperature minimum, maximum humidity and precipitation) in the field of short-term weather parameter forecasting. Moreover, the similar parameters were also trained to carry out the establishment of the VAR model in Bangladesh to enhance Climate Change observation [9]. The VAR model was found to be the best. Structural analyses were performed using forecast error variance decomposition and impulse response function. These structural analyses divulged that the temperature, humidity, and cloud coverage would be interrelated and endogenous in future. Finally, temperature, humidity, and cloud coverage were forecasted from January 2011 to December 2016 using the best selected model VAR. Therefore, the utilization of precipitation data is necessary to deepen further in order to take a preventive action toward upcoming hazards so that the high rate of floods occurrences in Cilacap is critical to conduct the research of the use of the VAR model by using the ENSO and IOD data.

2 Methodology 2.1 Data 1. Flood Data BNPB Data on disaster events in Indonesia obtained from the National Disaster Management Agency (BNPB) 2. Global Satellite Measurement of Precipitation (GSMAP) Rainfall data were obtained from Global Satellite Mapping of Precipitation (GSMaP_MVK). GSMaP_MVK is a product that integrates passive microwave radiometer data with infrared radiometer data in order to have high temporal (1 hour) and spatial (0.1 degrees) resolution global precipitation estimates. The product (GSMaP_MVK) is produced based on a Kalman filter model that refines the precipitation rate propagated based on the atmospheric moving vector derived from two successive IR images. The detail of the algorithm can be found in [10]. Data dapat diunduh di https://sharaku.eorc.jaxa.jp/GSMaP. 3. Sea Surface Temperature (SST) Data ENSO and IOD Index data from SST data downloaded from NOAA.

Monthly Rainfall Prediction Using Vector Autoregressive Models …

399

2.2 Method 1. Stationarity Data Before forming the Vector Autoregressive (VAR) model, the time series data first perform some tests on the initial data, whether the data is stationary or not. Therefore, the stationarity of time series data concerning variance or mean value can be carried out successively using the Box-Cox transformation (Brockwell, 2016) and the Augmented Dickey-Fuller (ADF) test [11]. 2. Model Vector Autoregressive (VAR) The VAR model is an extension of the Autoregressive (AR) time series model with more than one variable. All variables in the VAR model are assumed to be endogenous variables (variables whose value is determined by other variables in the model) and are interrelated. The general form of the VAR model with a lag p, VAR(p), can be calculated as follows [12]: X p,t = φ0 +

p 

φi X i,t−i + et , i = 1, 2, . . . , p, t = 1, 2, . . .

(1)

i=1

where; X p,t is the dependent variable (X 1,t , X 2,t , . . . , X p,t ) sized p × 1, X i,t−1 is the independent variable (X 1,t−i , X 2,t−i , . . . , X p,t−i ), sized p × 1, φ0 is vector intercept size p × 1, φi is a coefficient matrix of size p × p for each i = 1, 2, . . . , p, et is the remainder vector (e1,t , e2,t , . . . , e p,t ) sized p × 1. 3. Cointegration Test If data is cointegrated, then the modeling for that data cannot use the VAR model. The basis of cointegration is that many time series data can deviate from their average in the short-term but move together. The hypothesis in Johansen’s test is as follows: H0 : there are as many as r , where r = 0, 1, . . . , k − 1 cointegration equation (no cointegration or long-term relationship between variables). H1 : there are cointegration equations (cointegration occurs between variables). 4. Vector Error Correction Model (VECM) VECM is a restricted form of VAR due to the existence of non-stationary but cointegrated data forms. The VECM specification specifies the long-term relationship of endogenous variables to converge to their cointegration relationship but still allows the presence of short-term dynamics. The VECM has the following equation: X p,t = φ p,0 + X p,t−1 +

p−1  i=1

i∗ X p,t−i + et

(2)

400

F. Nauval et al.

where X p,t = X p,t−1 +

X p,t , ∗j

=−

p 

i , j = 1, 2, . . . , p − 1, and

i= j+i

 = −(1 − 1 − 2 − . . . −  p

(3)

5. Optimum Lag Determination VAR estimation is very sensitive to the lag length used. Therefore, it is necessary to know the optimal lag before estimating the VAR. Determination of the optimal amount of lag can be determined using Akaike Information Criterion (AIC), Schwarz Information Criterion (SC), Final Prediction Error (FPE), or Hannan Quinnon (HQ). Determination of the optimal lag using the information criteria by selecting the measures with the smallest value [12]. This optimal lag is applied to eliminate the autocorrelation problems in the VAR model so that they will no longer arise. 6. Impulse response functions (IRF) After modeling using the VAR model, it is necessary to have a method that can characterize the dynamic structure generated by VAR. IRF shows the response of each endogenous variable over time to the shock of the variable itself and other endogenous variables. Thus, IRF is used to see the contemporary effect of a dependent variable if it gets a shock or innovation from the independent variable by one standard deviation. 7. Mean Absolute Percentage Error (MAPE) MAPE is a useful measure to measure the error of the estimated value of the model. The error in the estimated value of the model expresses an average absolute percentage of residual, which is written as follows:   N 1  Yt − Yˆ  M AP E =   × 100% N t=1  Yt 

(4)

where N is the number of predictions made, Yt is the actual data, and Yˆt is the data forecast using the VAR model. Figure 3 indicates the topography of the Cilacap area. It is clearly seen that Cilacap is located on Java Island and close to the Indian Ocean. Interaction between land and sea affects the uniqueness of rainfall patterns as studied in this research.

Monthly Rainfall Prediction Using Vector Autoregressive Models …

401

Fig. 3 Topography of Cilacap

3 Result and Discussion 3.1 Time Series of Rainfall Data and the ENSO Index, and the IOD Figure 4a depicts a graph of rainfall time series data, IOD anomaly, and ENSO anomaly for the monthly period starting from March 2000 to December 2018. Moreover, it can be seen that the three time series data still fluctuate randomly and do not move around a fixed value. That is, these three data must be stationary to the variance and the mean so that the fluctuations in the data are more directed at a certain interval and the movement is centered on a fixed value. Therefore, a Box-Cox transformation is needed to overcome the problem of stationary variance and differencing methods to make the time series data move at a middle value. Figure 4b shows that there is a clear relationship between the ENSO, IOD, and peak rainfall. Those three data indicate stationarity one another. Thus, the time series data can be easily analyzed in terms of the mean and variance.

3.2 Determining Optimum Lag Determination of the optimal amount of lag can be determined using Akaike Information Criterion (AIC), Schwarz Information Criterion (SC), Final Prediction Error

402

F. Nauval et al.

Fig. 4 Time series of anomaly data of mohly rainfall a before differencing, b after differencing

(FPE) or Hannan Quinnon (HQ). Determination of the optimal lag using the information criteria is obtained by selecting the criteria that have the smallest value (Table 1). VAR Estimation Results: Estimated coefficients for equation Rain: R AI N (t) = 134.5828899 + 0.2835074R AI N (t − 1) + 26.6413343I O D(t − 1) + 26.4434951E N S O(t − 1) + 0.1647146R AI N (t − 2) − 27.4772601I O D(t − 2) − 24.5771315E N S O(t − 2) (5) Table 1 Criteria of optimum lag Lag

Criteria AIC

HQ

SC

FPE

1

−17.98138

−17.9051

−17.79263

1.551650e-08

2

−18.33049

−18.19702

−18.00018*

1.094459 e-08*

3

−18.30355

−18.11287

−17.83168

1.124497e-08

4

−18.30727

−18.05939

−17.69384

1.121851e-08

5

−18.30654*

−18.00146*

−17.55155

1.111766e-08

6

−18.31614

−17.95385

−17.41959

1.15E-08

* represents the criteria of optimum lag given by the data

Monthly Rainfall Prediction Using Vector Autoregressive Models …

403

Estimated coefficients for equation IOD: I O D(t) = 0.0553247229 − 0.0001777911R AI N (t − 1) + 0.7539071521I O D(t − 1) + 0.1444971616E N S O(t − 1) + 0.0000895992R AI N (t − 2) − 0.1165418043I O D(t − 2) − 0.1303052468E N S O(t − 2)

(6)

Estimated coefficients for equation ENSO: E N S O(t) = −0.0622970607 − 0.0002176436R AI N (t − 1) + 0.2096981972I O D(t − 1) + 1.4310148716E N S O(t − 1) + 0.0005191951R AI N (t − 2) − 0.2639312138I O D(t − 2) − 0.5001126226E N S O(t − 2)

(7)

It can be seen in Fig. 5 that there are fluctuations in rainfall intensity in the first month of stability, and then fluctuations occur until the 5th month, which then stabilizes in the 5th month and so on. While the IRF from ENSO for Rainfall, it can be seen that there is a fluctuation on the 2nd day, and in the following month, there is stability at the equilibrium point. Table 2 shows that the performance of the VAR model predicts the rain, IOD, and ENSO data in the following month. It can be clearly seen that those three data are

Fig. 5 Orthogonal impulse response

404

F. Nauval et al.

Table 2 Forecasting interval of VAR model Rain Jan 2019

Forecast

Lower

Upper

CI

Actual

201.6998

27.27569

376.1238

174.4241

326.038

Within predicted range

0.194283

−0.18436

0.572929

0.378646

0.452

Within predicted range

0.192421

−0.24527

0.630111

0.43769

0.520

Within predicted range

IOD Jan 2019 ENSO Jan 2019

in the predicted range. It also indicates that the VAR model successfully predicts the upcoming index or value in the forecasted data. In order to evaluate the performance, statistical evaluation was carried out. Figure 6 indicates that VAR model estimation is able to evaluate the performance of those three data. Moreover, the relationship between VAR model estimation and observation is significantly strong for ENSO and IOD data in particular. Somehow, VAR model only predicts a narrow range in terms of middle value only. In other words, the VAR model is not able to predict both the highest and the lowest peak of rainfall compared to observation data (Table 3).

Fig. 6 Model VAR estimation

Monthly Rainfall Prediction Using Vector Autoregressive Models … Table 3 VAR model accuration

405

Root mean squared error Mean absolute percentage (RMSE) error (MAPE) Rainfall 87.06771

0.3291528

IOD

0.1845223

0.04796626

ENSO

0.2178605

0.06163001

4 Conclusion Rainfall in Cilacap is predicted to be stable and within the prediction range in the upcoming periods. Similar trends are also observed for the IOD and ENSO, where the predicted data is not beyond the lower and upper forecasted values. In other words, this condition is the same as in the previous periods fluctuating at the same point, which neither overestimates nor underestimates. Furthermore, the comparison of rainfall prediction results in such a study area with the VAR model and actual data can be seen by the mean absolute percentage error (MAPE) test values which were barely under 1 (good performance) in the study area. In addition, the IOD and ENSO data also indicate a better performance, in which the MAPE values are completely close to zero. Unfortunately, the RMSE value of forecasting rainfall is clearly observed more than one, which is precisely at 87.06771. Somehow, the value of RMSE for the IOD and the ENSO remains considerably categorized as a good performance where the value of which is less than one. To sum up, the aforementioned statements indicate crystal clear that the VAR model performs well in forecasting the IOD and ENSO and is fairly moderate in predicting rainfall in Cilacap. Therefore, this VAR model is recommended to be implemented in developing the early warning system to mitigate and manage the upcoming hazards in Indonesia. For getting better results and more supporting evidence of this performance, however, it is suggested to further carry out this study in predicting rainfall, the IOD, and the ENSO in other areas in Indonesia.

References 1. Chang, C.P., Wang, Z., McBride, J., Liu, C.-H.: Annual cycle of Southeast Asia– Maritime Continent rainfall and asymmetric monsoon transition. J. Climate 18, 287–301 (2005) 2. Saji, N.H., Goswami, B.N., Vinayachandran, P.N., Yamagata, T.: A dipole mode in the tropical Indian Ocean. Nature 401, 360–363 (1999) 3. Ashok, K., Guan, Z., Saji, N.H., Yamagata, T.: Individual and combined influences of the ENSO and Indian Ocean Dipole on the Indian summer monsoon. J. Climate 17, 3141–3155 (2004b) 4. Wang, X., Wan, C.: Different impacts of various El Nin˜o events on the Indian Ocean Dipole. Clim. Dyn. 2014(42), 991–1005 (2014)

406

F. Nauval et al.

5. Ramage, C.S.: Role of a tropical “maritime continent” in the atmospheric circulation. Mon. Weather Rev. 96, 365–369 (1968). https://doi.org/10.1175/1520-0493(1968)096%3c0365: ROATMC%3e2.0.CO;2 6. Li, J.: Assessing the accuracy of predictive models for numerical data: Not r nor r2 , why not? Then what? (2007) 7. Widarjono, A.: Ekonometrika Teori dan Aplikasi untuk Ekonomi dan Bisnis. Edisi kedua. Yogyakarta, Ekonisia (2007) 8. Abdallah, W., Abdallah, N., Marion, J., Oueidat, M., Chauvet, P.: A vector autoregressive methodology for short-term weather forecasting: tests for Lebanon. SN Appl. Sci. 2, 1–9 (2020) 9. Shahin, M.A., Ali, M.A., Ali, A.B.M.S.: Vector Autoregression (VAR) Modeling and Forecasting of Temperature, Humidity, and Cloud Coverage. Computational Intelligence Techniques in Earth and Environmental Sciences. Springer, Dordrecht (2014) 10. Ushio, T., Sasashige, K., Kubota, T., Shige, S., Okamoto, K., Aonashi, K.: A kalman filter approach to the global satellite mapping of rainfall (GSMaP) from combined passive microwave and infrared radiometric data. J. Meteorol. Soc. Japan 87A, 137–151 (2009) 11. Lutkepohl, H.: New Introduction to Multiple Time Series Analysis. Springer, Berlin (2005) 12. Wei, W.W.S.: Time Series Analysis: Univariate and Multivariate Methods. Addison-Wesley Publishing Co., USA, United State of America (2006)

Extreme Rainfall Clusters in Borneo and Their Synoptic Climate Causes Narizka Nanda Purwadani , Mohamad Rahman Djuwansah , Muhammad Rais Abdillah , Faiz Rohman Fajary , and Ida Narulita

Abstract A detailed study of the relationship between synoptic climate phenomena and high rainfall is essential to predict the occurrence of hazardous rainfall. We use Global Satellite Mapping of Precipitation (GSMaP) data and ERA-5 Reanalysis data to identify extreme rainfall events and the triggering anomalies. Extreme rainfall was identified as a 97th percentile maximum threshold. Then, k-means clustering applied to cluster spatial patterns of extreme rainfall. The influencing climatic factors anomaly reveals the synoptic climate phenomena in the Western Maritime Continent that provoke extreme rainfall on Borneo Island. Strong northerly moisture transport in December-January–February and March–April-May seasons may be related to cold surge events and lead to heavy rainfall in northern or western Borneo. Westerly wind anomalies toward the Philippines that can occur in all seasons may be associated with Madden Julian Oscillation and cause heavy rainfall around North Kalimantan. Also, positive sea surface temperature anomalies in the Java Sea may trigger wet anomalies around Central Kalimantan.

N. N. Purwadani Master Program in Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia M. R. Djuwansah Research Centre for Environment and Clean Technology, National Agency for Research and Innovation, Bandung, Indonesia M. R. Abdillah (B) · F. R. Fajary Atmospheric Sciences Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia e-mail: [email protected] I. Narulita Research Centre for Climate and Atmospheric Science, National Agency for Research and Innovation, Bandung, Indonesia Faculty of Engineering, University of Indonesia, Depok, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_38

407

408

N. N. Purwadani et al.

1 Introduction Flood is one of Indonesia’s most frequent hydrometeorological disasters, including on Borneo Island. Social media-based data in 2015–2021 from a global database of flood events record floods in Borneo can occur in all parts of the island in all seasons [1]. Still, the highest flood events occur in the December-January–February (DJF) season. One of the phenomena that trigger flooding is extreme rainfall. Therefore, a study on extreme rainfall in Borneo is needed so flood disaster mitigation can run more effectively. Previous climatological studies have examined the phenomena affecting rainfall variability around the Borneo region and noted the importance of synoptic disturbances [2, 3]. However, the spatial pattern of extreme rainfall that commonly occurs on the island of Borneo due to those synoptic disturbances is still not well understood. Extreme rainfall events seem to recur from time to time according to repeated atmospheric circulation conditions [4]. Thus, it is possible to classify the spatial patterns of extreme rainfall [5]. Our study focuses on finding the spatial pattern of extreme rainfall that commonly occur on Borneo Island and its relationship with synoptic phenomena to increase the predictability of extreme rainfall, as one of the flood triggers.

2 Methodology 2.1 Data Daily rainfall data is taken from near-real-time gauge-adjusted rainfall of the Global Satellite Mapping of Precipitation (GSMaP) at 0.1° × 0.1° grids [6], which could be effective for extreme rainfall monitoring in Asia–Pacific [7]. This data can be downloaded using the FTP address hokusai.eorc.jaxa.jp. Specific humidity, wind fields, and sea surface temperature (SST) are obtained from the European Center for Medium-Range Weather Forecasts Reanalysis v5 (ERA5) and are used for a composite analysis of atmospheric conditions during extreme rainfall events. ERA5 has a 0.25° horizontal resolution, provided at hourly intervals [8], and is free for download from the climate data store website https://cds.climate.copernicus.eu/. All the datasets are retrieved for the study period of 2001–2021.

2.2 Methods GSMaP rainfall data are masked on Borneo Island to exclude extreme rainfall in the ocean. An extreme rainfall day is identified if one or more grid points of daily rainfall in Borneo are greater or equal to its 97th percentile of the cumulative distribution

Extreme Rainfall Clusters in Borneo and Their Synoptic Climate Causes

409

function of daily rainfall maxima for the study period. The identified date is classified as the date of the extreme rainfall event. To reveal dominant spatial patterns of extreme rainfall events in Borneo, k-means clustering is applied to the correlation matrix of extreme rainfall events in each season (DJF, March–April-May (MAM), June-July–August (JJA), and September– October-November (SON)). The number of clusters in each season is chosen using the trial-and-error method. For each season, k-means clustering is applied for 2, 3, 4, or 5 clusters. Then, the number of clusters with the most representative spatial patterns (no similar pattern between clusters) is chosen. A composite analysis of moisture flux, SST, and wind anomalies is used to determine synoptic conditions that trigger extreme rainfall events in each cluster. Moisture flux (MF) is obtained from integrating specific humidity (q) and wind field (V) from the surface to 700 hPa: 1 MF = g

700  hPa

qV dp

(1)

surface

where g is the gravitational constant and p is pressure. Then the divergence of moisture flux (MFD) can be calculated using MFD = ∇ · MF

(2)

Positive (negative) values of MFD denote divergence (convergence) areas.

3 Results 3.1 Temporal Distribution of Extreme Rainfall The value of extreme rainfall refers to the above-mentioned threshold from 2001– 2021 (191 mm/day). During the study period, there were 231 total extreme rainfall events, which vary per year (Fig. 1a). The most frequent occurrences were in 2019, with twenty-six events. In the monthly distribution (Fig. 1b), extreme rainfall mostly occurs in December, January, February, and May, or DJF for the quarterly seasons.

3.2 Spatial Distribution Borneo extends around the equator for which heavy rainfall can occur throughout the year. However, each season has a different circulation pattern so it may trigger different spatial patterns of extreme rainfall. The clustering result shows four clusters

410

N. N. Purwadani et al.

Fig. 1 Yearly (a) and monthly (b) number of extreme rainfall events during the period 2001–2021 in Borneo

(C1, C2, C3, and C4) of extreme rainfall in DJF and MAM and three clusters (C1, C2, and C3) in JJA and SON. Note that C1 in DJF is not related to C1 in other seasons. C1 in DJF means the first spatial pattern of extreme rainfall in the DJF season. Clusters of extreme rainfall in each season show a certain spatial pattern that is different from other clusters. Each cluster may be caused by similar triggers but may also be by different triggers.

3.3 The Synoptic Patterns of Extreme Rainfall Events DJF (Fig. 2). Extreme rainfall events in C1 occur when negative SST anomalies cover the entire South China Sea (SCS) with northerly moisture transport anomalies from the Philippines. This moisture transport converges in northern Borneo and causes extreme rainfall in the Sabah region. In C2, there are positive SST and easterly moisture transport anomalies in the Java Sea. These positive SST anomalies moisten southern Borneo and trigger extreme rainfall in Central Kalimantan and South Kalimantan. SST and moisture transport anomalies at C3 are slightly like C1. The difference lies in the moisture transport at C3 originating from SCS and converging in western Borneo, thus causing extreme rainfall in West Kalimantan and Sarawak region. In C4, there are positive SST anomalies around northern Borneo and westerly moisture transport anomalies to the Philippines. These conditions cause wet anomalies in North Kalimantan and SCS. MAM (Fig. 3). In C1, extreme rainfall events occur when there are westerly moisture transports anomalies in Borneo Island that converge around western Sulawesi. This moisture convergence causes heavy rainfall in East Kalimantan and South Kalimantan. C2 in MAM is similar to C1 in DJF with negative SST anomalies in SCS and northerly moisture transport from the Philippines. This moisture transport causes moisture convergence in northern Borneo and extreme rainfall around the Sabah region. C3 in MAM is slightly similar to C2 in DJF with positive SST anomalies and easterly moisture transport in the Java Sea, thus causing rainfall anomalies around

Extreme Rainfall Clusters in Borneo and Their Synoptic Climate Causes

411

Fig. 2 Clusters composite anomalies of [top] SST (shaded) and 10 m wind (arrows); [middle] vertically integrated 1000–700 hPa moisture divergence (shaded) and moisture transport (arrow); and [bottom] rainfall (shaded) in DJF season. The number in brackets indicates the number of members in each cluster

Central Kalimantan. C4 in MAM is quite similar to C4 in DJF with westerly moisture transport anomalies to the Philippines and heavy rainfall around North Kalimantan and SCS. JJA (Fig. 4). Extreme rainfall events in JJA occur when the SST anomaly is positive in almost all of the Western Maritime Continent (WMC). The difference lies in the direction of moisture transport. In C1, there are northeasterly moisture transport anomalies from the Philippines and SCS. This moisture transport converges on the SCS, the Karimata Strait, and the west of the Java Sea, causing rainfall anomalies on the coast of this part of Borneo. C2 in JJA is slightly similar to C4 in DJF and C4 in MAM with westerly moisture transport anomalies to the Philippines and heavy rainfall around North Kalimantan and SCS. C3 shows a similar case to C1 but with weaker northeasterly moisture transport anomalies in SCS and stronger anomalies in the Philippines and the Celebes Sea. This moisture transport converges in Makassar Strait and causes extreme rainfall in East Kalimantan and South Kalimantan. SON (Fig. 5). Extreme rainfall events in SON occur when there is northeasterly moisture transport from SCS. This moisture transport turns west at C1 and converges around the southern part of SCS. This moisture convergence causes extreme rainfall around the Sarawak region. In C2, the moisture transport turns south to the Karimata Strait. Moreover, there are positive SST anomalies around the Java Sea in C2. This

412

N. N. Purwadani et al.

Fig. 3 Same with Fig. 2 but in MAM season

warm SST moistens the southern part of Borneo, thus causing heavy rainfall in Central Kalimantan. In C3, the moisture transport turns east resulting in westerly moisture transport anomalies to the Philippines. This pattern is analogous with C4 in DJF, C4 in MAM, and C2 in JJA with wet anomalies in North Kalimantan.

4 Discussion The Asian monsoon has the most significant effect on rainfall patterns in WMC [9, 10]. The complex terrain of the archipelago strongly influences the local wind circulation and rain distribution in WMC [10]. Interannual climate phenomena, such as ENSO and IOD, affect the amount of annual rainfall [11, 12]. In the intraseasonal time range, synoptic phenomena such as strong northerly wind known as Cold Surge (CS) [13], vortices around the waters and mainland of Borneo known as Borneo Vortex (BV) [2], and Madden Julian Oscillation (MJO) [14] interfere with the variation of monsoonal rainfall in the region. Referring to the average climate regime pattern in WMC [15], we hypothesized several synoptic climate phenomena causing rainfall anomalies in Borneo. Based on previous analysis, this study can identify the intensified northerly wind known as CS as a possible specific synoptic climate affecting the heavy rainfall in northern Borneo at DJF C1 and MAM C2 and in western Borneo at DJF C3. The difference in

Extreme Rainfall Clusters in Borneo and Their Synoptic Climate Causes

413

Fig. 4 Same with Fig. 2 but in JJA season

the location of the wet anomalies may occur due to the difference in the direction of moisture transport in the CS, from the Philippines in DJF C1 and MAM C2, and from SCS in DJF C3. The wet anomaly around North Kalimantan at DJF C4, MAM C4, JJA C2, and SON C3 may be related to the MJO phenomenon with convective center around the Philippines because the wind patterns in the four clusters show westerly wind anomalies to the Philippines. This is in line with the statement that westerly winds exist in and to the west of the center of MJO large-scale convection [14]. Positive SST anomalies in the Java Sea may also have an important role in extreme rainfall events around Central Kalimantan such as in DJF C2, MAM C3, and SON C2. This preliminary conclusion still needs confirmation by observing more in detail the relationship between rainfall intensity and the indexes of CS, MJO, and vorticity particularly that related to the BV. Meanwhile, knowledge about the influence of other climate variations, both synoptic and interannual, on extreme rainfall in Borneo still requires further investigation and will be the focus of further studies.

414

N. N. Purwadani et al.

Fig. 5 Same with Fig. 2 but in SON season

5 Conclusion This study has clustered the spatial patterns of extreme rainfall in Borneo during 2001–2021 to investigate the synoptic conditions associated with each cluster. There are several synoptic patterns associated with extreme rainfall in Borneo. Strong northerly moisture transport with negative SST anomalies occurring in the DJF and MAM seasons may be associated with CS events and cause heavy rainfall around northern or western Borneo. Westerly wind anomalies with positive SST around northern Borneo that can occur in all seasons may be related to the MJO with convective centers around the Philippines and causing heavy rainfall around North Kalimantan. In addition, positive SST anomalies around the Java Sea may also have an important role in triggering extreme rainfall events around Central Kalimantan. Acknowledgements We would like to thank the Earth Sciences Research Organization BRIN for its financial support through the Disaster Home Financing Program (WBS2-15) for the fiscal year 2022. This study was also partially supported by the FITB-ITB Collaborative Research PPMI.

Extreme Rainfall Clusters in Borneo and Their Synoptic Climate Causes

415

NNP was also partly supported by the project titled “Marine Science & Technology Cooperation (MTCRC) between Korea and Indonesia” funded by the Ministry of Oceans and Fisheries, Korea.

References 1. de Bruijn, J.A., de Moel, H., Jongman, B., et al.: A global database of historic and real-time flood events based on social media. Sci Data 6, 311 (2019). https://doi.org/10.1038/s41597019-0326-9 2. Chang, C.-P., Harr, P.A., Chen, H.-J.: Synoptic disturbances over the equatorial South China Sea and Western Maritime Continent during boreal winter. Mon. Weather Rev. 133, 489–503 (2005). https://doi.org/10.1175/MWR-2868.1 3. Lim, S.Y., Marzin, C., Xavier, P., et al.: Impacts of boreal winter monsoon cold surges and the interaction with MJO on Southeast Asia rainfall. J. Clim. 30, 4267–4281 (2017). https://doi. org/10.1175/JCLI-D-16-0546.1 4. Jacobeit, J., Homann, M., Philipp, A., Beck, C.: Atmospheric circulation types and extreme areal precipitation in Southern Central Europe. Adv. Sci. Res. 14, 71–75 (2017) 5. Gvoždíková, B., Müller, M., Kašpar, M.: Spatial patterns and time distribution of central European extreme precipitation events between 1961 and 2013. Int. J. Climatol. 39, 3282–3297 (2019). https://doi.org/10.1002/joc.6019 6. Kubota, T., Aonashi, K., Ushio, T., et al.: Global satellite mapping of precipitation (GSMaP) products in the GPM era. In: Satellite Precipitation Measurement. pp. 355–373. Springer (2020) 7. Tashima, T., Kubota, T., Mega, T., et al.: Precipitation extremes monitoring using the nearreal-time GSMaP product. IEEE J. Selected Topics in Appl. Earth Observ. Remote Sens. 13, 5640–5651 (2020) 8. Hersbach, H., Bell, B., Berrisford, P., et al.: The ERA5 global reanalysis. Q. J. R. Meteorol. Soc. 146, 1999–2049 (2020). https://doi.org/10.1002/qj.3803 9. Robertson, A.W., Moron, V., Qian, J.-H., et al.: The maritime continent monsoon. In: The Global Monsoon System: Research and Forecast, pp. 85–98. (2011) 10. Chang, C.-P., Lu, M.-M., Lim, H.: Monsoon convection in the Maritime Continent: interaction of large-scale motion and complex terrain. Meteorol. Monographs 56, 6.1–6.29 (2016). https:// doi.org/10.1175/AMSMONOGRAPHS-D-15-0011.1 11. Aldrian, E., Dümenil Gates, L., Widodo, F.H.: Seasonal variability of Indonesian rainfall in ECHAM4 simulations and in the reanalyses: the role of ENSO. Theoret. Appl. Climatol. 87, 41–59 (2007) 12. Narulita, I., Fajary, F.R., Syahputra, M.R., et al.: Spatio-temporal rainfall variability of equatorial small island: case study Bintan Island, Indonesia. Theor. Appl. Climatol. 144, 625–641 (2021). https://doi.org/10.1007/s00704-021-03527-x 13. Abdillah, M.R., Kanno, Y., Iwasaki, T., Matsumoto, J.: Cold surge pathways in East Asia and their tropical impacts. J. Clim. 34, 157–170 (2021). https://doi.org/10.1175/JCLI-D-20-0552.1 14. Zhang, C.: Madden-Julian oscillation. Rev. Geophys. 43 (2005). https://doi.org/10.1029/200 4RG000158 15. Chang, C.-P., Wang, Z., McBride, J., Liu, C.-H.: Annual cycle of Southeast Asia—Maritime Continent rainfall and the asymmetric monsoon transition. J. Clim. 18, 287–301 (2005). https:// doi.org/10.1175/JCLI-3257.1

Analysis of Wind Variations and Differences on the Airport Runways: a Case Study at I Gusti Ngurah Rai Airport, Bali Kadek Sumaja and Amanda Pasa Kencana

Abstract Surface wind conditions are strongly influenced by several factors, such as geographical location, seasons, and temperature differences. I Gusti Ngurah Rai Airport is located on the narrow land on Bali Island and surrounded by the ocean. Thus, its runway is prone to significant wind variation due to the changing season as well as the influence of land and sea breeze. Moreover, there is an essential issue in determining the runway in use for the takeoff and landing process when there are significant differences in wind direction and speed between two runways. Therefore, an analysis was carried out to determine the variation of wind direction and speed on runways 09 and 27 in I Gusti Ngurah Rai, Bali. The 10 years of observation data were retrieved from AWOS, then processed to define the wind data in both runways. The wind rose method was utilized to analyze the wind pattern on both runways in each season. Then, the delta calculation was utilized to determine the significant difference in wind direction and speed on the two runways. The result showed that the wind variation on both runways were generally similar and the significant wind direction differences mainly occurred during the transition season. Moreover, the highest number of significant wind direction differences generally emerged before noon and at night while the lowest number was in the evening. Lastly, wind speeds above 10 KT mostly occurred on runway 09 when the wind direction in both runways was significantly different.

1 Introduction Wind information on the airport runway is crucial to support efficiency and safety during the takeoff and landing process of aircraft [1–3]. The takeoff and landing K. Sumaja (B) I Gusti Ngurah Rai Meteorological Office, Indonesian Agency for Meteorology Climatology and Geophysics, Bali, Indonesia e-mail: [email protected] A. P. Kencana Bandung Institute of Technology, Bandung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_39

417

418

K. Sumaja and A. P. Kencana

activities are strongly affected by the wind pattern on the runway. The aircraft started the process from one end point of the runway which is known as the runway in use. I Gusti Ngurah Rai Airport has a single runway, which stretches from east to west, where the eastern touch point is called runway 27 and the touch point at the western end of the runway is called runway 09. In performing safe and efficient takeoff and landing process, the aircraft must go downwind [2]. Thus, reliable wind information is crucial at each end of the runway. To provide this information, along the runway at the I Gusti Ngurah Rai Airport were installed three wind sensors as part of the Automatic Weather Observation System (AWOS) at the eastern (runway 27), the middle, and the western point (runway 09). The wind direction in Bali Island, where I Gusti Ngurah Rai Airport is located, changes every season due to the influence of the monsoon [4]. Moreover, the runway at I Gusti Ngurah rai airport is flanked by the ocean in the west and east. Therefore, wind variations are most likely to arise due to local influences such as land and sea breezes as well as heat capacity and surface roughness between sea and water [5]. This geographical condition not only leads to the occurrence of crosswind, wind gust, and wind shear but also significant differences in wind direction between both runways [6]. Therefore, this study aims to analyze the wind direction variation and its differences on both I Gusti Ngurah Rai Airport’s runways to determine the time occurrence and frequency of these significant differences.

2 Data and Methods This study utilized wind observation data on runway 09 and runway 27 from wind sensors of the Automatic Weather Observation System (AWOS) for 10 years period (2012–2021), as shown in Fig. 1. The time used is in Coordinated Universal Time (UTC) as recorded in AWOS. The wind direction and speed were retrieved from 30 min’ weather data known as MET REPORT, and weather data when significant weather conditions occurred known as SPECIAL REPORT. Wind direction is the direction from which the wind blows and wind speed defines how strong the wind blows. The wind data was shortened based on the completeness of the wind data on the two runways, so around 97% of total data were passed pre-processing and utilized in this research. In addition, this study also employed the 10 years general wind data at I Gusti Ngurah Rai Airport to classify the runway wind data analysis into a group with similar patterns. Then, the data was divided based on the direction and wind speed on runway 09 and 27 using Microsoft Excel. Furthermore, it was categorized into certain periods based on the normal wind [7, 8]. Thus, four wind periods were obtained, namely December-January–February (DJF) was a westerly wind, May–June-July–AugustSeptember (MJJAS) was an easterly wind, March–April (MA) was a transition period from westerly to easterly wind, and October–November (ON) was a transition period from easterly to westerly wind, as shown in Fig. 2.

Analysis of Wind Variations and Differences on the Airport Runways …

419

Fig. 1 Runways 09 and 27 at I Gusti Ngurah Rai Airport

Fig. 2 Wind and rainfall in Bali period a DJF, b MA, c MJJAS, and d ON

The analysis is carried out by looking for the difference in wind direction between runways 09 and 27. The wind direction is measured between 1 and 360° (in form of a cycle from the north, east, south, west, and back to the north) and the highest difference would be 180°. Therefore, when the difference value is more than 180° , the value that would be used in the analysis is the result of subtraction between 360° and the initial value. For instance, if the difference in wind direction between both runways is 200° , then the value that will be used is 360° –200° = 160° . Then, referring to the BMKG regulations regarding significant changes or differences in wind direction, the results of the difference in wind direction to be analyzed are those whose values are above 60° [9, 10]. Then from the difference in the significant wind direction, it will be selected for direction with a speed above 10 Knots (KT), and time of occurrence for each season. The wind difference was considered to interfere with

420

K. Sumaja and A. P. Kencana

the takeoff and landing process because it complicates the selection of the runway in use [1].

3 Result and Analysis 3.1 Wind Pattern in Runways 09 and 27 The wind patterns on runway 09 were generally similar to those of runway 27, as shown in Figs. 3 and 4. For the DJF, the wind direction on runway 09 varied from east to southwest, in contrast on Runway 27 it only differed from north to west. Furthermore, in the MA and ON periods, it had more variation. The wind on runway 09 was blown almost from every direction, slightly different from the wind direction conditions on runway 27 with no wind blowing from the southwest. A unique condition occurs in the MJJAS period, where the wind was blown only from the east and southeast. This indicates that the transition season, MA and ON, have a high potential for significant wind direction differences between runways 09 and 27. It is due to the uncertain wind direction during the transition season caused a high fluctuation in wind movement on the runway and lead to an increase in the wind direction difference on runway 09 and 27 [11]. Moreover, some wind variation occurred during the DJF period, especially in runway 09. This period is the peak of the rainy season in Bali, so it has high air instability and a lot of convective clouds [8, 12]. This convective cloud during its mature state produced numerous wind gusts which most likely affect the wind on runway 09 where located beside the sea that has fewer obstacles [13]. Although the possibility of its occurrence is uncertain, wind gusts usually occur due to the influence of storms, convective clouds, and air instability [14, 15]. The MJJAS period had the smallest wind variation. It is because this period is the eastern monsoon season. This easterly wind carried less water vapor and has a more stable atmospheric condition that led to a lack of cloud growth and stable air movement [4].

Fig. 3 Wind pattern in runway 09 period a DJF, b MA, c MJJAS, and d ON

Analysis of Wind Variations and Differences on the Airport Runways …

421

Fig. 4 Wind pattern runway 27 period a DJF, b MA, c MJJAS, and d ON

3.2 Significant Wind Difference Between Runway 09 and 27 Significant wind direction difference The direction discrepancy in wind data between runways 09 and 27 is then analyzed for significant events or when the difference is above 600 . The number of occurrences of significant differences was categorized according to periods of normal wind data (DJF, MA, MJJAS, and ON) and compared with the total incidence in each season to determine the percentage of the occurrence. The results were shown in Fig. 5a, where the ON period has the highest percentage of occurrence (18%) of significant wind direction difference between runway 09 and 27. Then followed by months MA, MJJAS, and DJF with 14%, 13%, and 10%, respectively. Thus, it is necessary to put more attention out for the high incidence of significant wind differences in October– November. Since this period is the beginning of the rainy season in Bali, it would cause the potential for disturbance to increase [8, 12]. Furthermore, wind speed data, when the difference in wind direction on runway 09 and 27 was significant, was selected as the maximum wind speed. This aimed to determine the potential occurrence of maximum wind disturbance on this significant 35

(a)

15%

Wind speed (Knot)

percentage (%)

20%

10% 5%

MA

MJJAS

ON

22

25 17

15

21 15

15

15

5

0% DJF

(b)

RWY 09 RWY 27 28 27

-5

DJF

MA

MJJAS

ON

Fig. 5 a Occurrence of significant wind direction difference between runway 09 and 27, b Maximum wind speed during significant difference in wind direction between runway 09 and 27

422

K. Sumaja and A. P. Kencana

wind variation. The maximum wind speed on run-way 09 is always higher than that of runway 27, as shown in Fig. 5b. Although the MA Period has the highest maximum wind speed (28 KT) that occurs on runway 09, the DJF period has a high maximum wind speed on both runways (above 17 KT). The MJJAS and ON periods have almost the same maximum wind speed, namely 21–22 KT on runway 09 and 15 KT on runway 27. These results indicate that airlines and navigation parties should pay more attention to the DJF month because of the significant difference in wind direction and speed. Moreover, the maximum winds on both runways were considerably high and this period is the peak of the rainy season in Bali [8, 12, 13]. In addition, the significant wind direction difference between the two runways was displayed in the hourly range to find out at what time the significant wind direction difference on the two runways occurred. The results showed that the number of significant differences in wind direction generally occurred in the morning, then increased during the day, decreased at night, rosed in the middle of the night, and slowly dropped to the morning. This was caused by the location of the runway which is located between two oceans. Thus, the wind variation most likely occurred due to the differences in heat capacity and surface roughness between sea and land. The boundary area between land and ocean has a characteristic of local wind that changes periodically in form of the sea and the land breeze that happens diurnally [11]. These breezes tend to influence the monsoonal wind on this runway to depend on its speed [5]. The most significant difference in direction during DJF and ON periods mostly occurred at 14 UTC and 1 UTC, respectively. The ON month also shows the number of significant wind direction differences in the afternoon (18 to 19 UTC) and night (13−14 UTC). Furthermore, significant differences in MA and MJJAS months occurred in the morning at 2 UTC and in the evening at 12 and 13 UTC. Generally, the lowest difference in wind direction across all seasons was in the evening from 9 to 10 UTC (Fig. 6). Wind Speed when the difference in wind direction is significant Wind speed above 10 knots on runway 09 and 27 during significant wind direction difference was displayed as the frequency of occurrence, as shown in Fig. 7a−d. 8 percentage (%)

DJF

MA

MJJAS

ON

6 4 2 0 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Time (UTC)

Fig. 6 The occurrence (%) of significant wind direction difference between runway 09 and 27

Analysis of Wind Variations and Differences on the Airport Runways …

423

Generally, when the wind direction in both runways was significantly different, the frequency of wind speeds above 10 KT on runway 09 was always higher than on runway 27. This is due to runway 09 is directly adjacent to the open ocean compared to runway 27 in the east, which has bays and narrow straits as its borders. This would generate a factor where the direct heating of the sun during the day is greater than the heating by the earth’s surface [16]. Thus, the influence of sea breezes on runway 09 was mostly higher since the sea breezes have a greater speed than land breezes [17]. A wind speed of 11 KT was also the most frequent in each period. Wind speeds above 10 KT on runway 27 mainly occurred in the DJF and MA periods, followed by ON and the least occurred in the MJJAS periods. Therefore, the DJF and MA periods could potentially disrupt takeoff and landing activities. In addition, the significant wind data which has wind speed above 10 KT was classified into the normal wind periods. It can be seen in Fig. 8 that the DJF period has the most events, followed by MA, then the MJJAS and ON periods. Although there were many incidents of significant wind direction differences, only a few have velocities above 10 KT that occur simultaneously on both runways. As a result, there would be a rare circumstance for the Air Traffic Controller (ATC) to face the challenging situation of determining the runway in use. Because when the wind direction on both runways is significantly different, but the speed on one runway is low, and the other is high, the runway chosen for takeoff and landing would be the runway with the highest wind speed. However, if both runways have high wind 90

RWY-09

Number of events

70 Number of events

40

(a) RWY-27

50 30

(b)

30 RWY-09

RWY-27

20 10

10 0 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

-10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 Wind Speed (Knot)

Wind Speed (Knot)

RWY-09

RWY-27

0 11 12 13 14 15 16 17 18 19 20 21 22 Wind Speed (Knot)

(d) Number of events

Number of events

(c) 500

50 RWY-09

RWY-27

0 11

12

13

14 15 16 17 18 Wind Speed (Knot)

19

20

21

Fig. 7 Wind speed when the difference of wind direction was above 60° for period a DJF, b MA, c MJJAS, and d ON

424 Fig. 8 Frequency of wind speed above 10 knots on both runways 09 and 27 for wind direction difference of more than 60°

K. Sumaja and A. P. Kencana

15

11

10

5

5 0

DJF

MA

MJJAS

ON

speeds, the ATC will find it difficult to determine the runway in use for the takeoff and landing process. Such conditions rarely occurred at I Gusti Ngurah Rai airport, because from 186,631 observation data, there were only 22 incidents. Therefore, it can be said that the wind conditions at I Gusti Ngurah Rai airport are very supportive for flight operations services.

4 Conclusion The wind patterns on runway 09 were generally similar to those of runway 27 and the significant differences of wind direction mostly occur during the transition season MA and ON. Moreover, the highest number of significant wind direction differences mostly occurred before noon (2−3 UTC) and at night (13−14 UTC) while the lowest number mostly occur in the evening (10−11 UTC). Lastly, the maximum wind speed as well as the frequency of wind speeds above 10 KT on run-way 09 was always higher than that of runway 27 when the wind direction in each runway was significantly different.

References 1. Federal Aviation Administration.: Takeoffs and departure climbs. Airpl. Fly. Handb.1–14. (2013) 2. van Es, G.W.H., Karwal, A.K.: Safety aspects of tailwind operations. Flight Saf. Found. (2001). [Online]. Available: https://www.skybrary.aero/bookshelf/books/1148.pdf 3. Connor, A.O., Kearney, D.: Evaluate the effect of turbulence on aircraft during landing and take-off phases. Int. J. Aviat. Aeronaut. Aerosp. 5(4) (2018). https://doi.org/10.15394/ijaaa. 2018.1284 4. Tjasyono, H.B.: The character of rainfall in the Indonesian monsoon. In: International Symposium Equatorial Monsoon Systems, pp. 1–11. (2008) 5. Tjasyono, B., Harijono, S.W.B.: Meteorologi Indonesia : Awan dan Hujan Monsun.Volume 2, 1st edn, vol. 2, Jakarta, Badan Meteorologi dan Geofisika (2007) 6. Domenico Cimini, F.S.M., Visconti, G.: In: Integrated Ground-Based Observing Systems. Berlin, Heidelberg: Springer (2011). https://doi.org/10.1007/978-3-642-12968-1 7. Pusparini, N., Winardi, Irmawan, D.: Analisa angin zonal dalam menentukan awal musim hujan di Bali bagian selatan. Bul. Fis. 16(2), 1–10 (2015) 8. BMKG.: Informasi Hujan Bulanan Wilayah Bali. (2022). http://balai3.denpasar.bmkg.go.id/ info-hujan-bulanan Accessed 28 Sep 2022

Analysis of Wind Variations and Differences on the Airport Runways …

425

9. ICAO.: ICAO Annex 3—Meteorological Service for International Air Navigation. 17th edn, no. July 2010. International Civil Aviation Organization (2001) 10. BMKG.: Pembuatan dan Penyampaian METAR dan SPECI dalam Pelayanan Informasi Cuaca untuk Penerbangan. Indonesia (2017) 11. Rifai, A., Rochaddi, B., Fadika, U., Marwoto, J., Setiyono, H.: Kajian Pengaruh Angin Musim Terhadap Sebaran Suhu Permukaan Laut (Studi Kasus : Perairan Pangandaran Jawa Barat). Indones. J. Oceanogr. 2(1), 98–104 (2020). https://doi.org/10.14710/ijoce.v2i1.7499 12. BMKG-Staklim Jembrana.: Buletin Prakiraan Musim Hujan 2020/2021, Jembrana (2021) 13. Risuana, I.G.S., Hendrawan, I.G., Suteja, Y.: Distribusi Spasial Total Padatan Tersuspensi Puncak Musim Hujan Di Permukaan Perairan Teluk Benoa, Bali. J. Mar. Aquat. Sci. 3(2), 223 (2017). https://doi.org/10.24843/jmas.2017.v3.i02.223-232 14. Seregina, L.S., Haas, R., Born, K., Pinto, J.G.: Development of a wind gust model to estimate gust speeds and their return periods. Tellus A Dyn. Meteorol. Oceanogr. 66(1), 22905 (2014). https://doi.org/10.3402/tellusa.v66.22905 15. Jungo, P., Goyette, S., Beniston, M.: Daily wind gust speed probabilities over Switzerland according to three types of synoptic circulation. Int. J. Climatol. 22(4), 485–499 (2002). https:// doi.org/10.1002/joc.741 16. Kurnia Anzhar, Y.S.B.S.: Pola Angin Laut dan Angin Darat di Daerah Ujung Lemahabang, Semenanjung Muria. J. Pengemb. Energi Nukl. 2(4), 199–206 (2000). https://doi.org/10.17146/ jpen.2000.2.4.2024 17. Pahlevi, A.R.: Simulasi Interaksi Angin Laut dan Bukit Barisan dalam Pembentukan Pola Cuaca di Wilayah Sumatera Barat Menggunakan Model WRF-ARW. Pros. Semin. Nas. Metod. Kuantitatif 1(1), 7–17 (2017)

Diurnal Variation of Rainfall Over Bangka Belitung Islands Determined from Rain Gauge and IMERG Observations Helmi Yusnaini, Zahwa Vieny Adha, Ravidho Ramadhan , Marzuki Marzuki , and Robi Muharsyah Abstract The diurnal cycle of rainfall is the primary circulation of the atmosphere, which is highly dependent on an area’s topographic conditions, land surface, and land-sea contrast. Understanding diurnal characteristics are beneficial in providing an overview of regional weather and the influence of topography on the diurnal cycle of rainfall. From previous research, it can be seen that there are differences in the diurnal cycle between the western and eastern parts of Sumatra. However, research in the eastern part of Sumatra is very limited, especially on small islands. Therefore, in this study, the diurnal rainfall characteristic on small islands of east Sumatra, i.e., Bangka Belitung Islands, from December 2015 to October 2019, was investigated using rain gauge and Integrated Multi-satellite Retrieval for GPM (IMERG) data. The diurnal characteristics of rainfall are defined as precipitation amount (PA), precipitation frequency (PF), and precipitation intensity (PI). The average PA, PF, and PI on the island by rain gauge (IMERG) observation, respectively, are 0.15–0.27 mm/h (0.22– 0.35 mm/h), 2.69–4.88% (6.15–8.59%), and 5.14–7.87 mm/h (3.35–3.84 mm/h). Rain peaks from PA, PF, and PI dominated the early afternoon (1200–1500 LST). Meanwhile, the PI peaks by IMERG show different peaks between the southern and northern parts of the island, i.e., early morning (0200 LST).

1 Introduction The diurnal cycle of rainfall is a local pattern of rainfall distribution caused by an intense convection process (interaction between the surface and the atmosphere). These cycles tend to be driven by local circulations such as land, sea, mountain, and valley winds [1]. As an area that receives high solar radiation yearly, this diurnal cycle is significant in the tropics, especially the Indonesian maritime continent (IMC) [2]. H. Yusnaini · Z. V. Adha · R. Ramadhan · M. Marzuki (B) Department of Physics, Universitas Andalas, Padang 25163, Indonesia e-mail: [email protected] R. Muharsyah Agency for Meteorology, Climatology and Geophysics of Republic Indonesia, Jakarta, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_40

427

428

H. Yusnaini et al.

IMC condition with complex topography, variation in shape and size of the island, and unique shoreline shape causes inhomogeneous diurnal cycles [3]. In addition, the diurnal circulation is also strongly influenced by global atmospheric interactions such as monsoons, Madden–Julian Oscillation (MJO), ENSO, and other possible factor. The characteristics and mechanisms of the diurnal cycle of rainfall for the Sumatra region have been carried out by many researchers [4, 5, 6, 7]. As-Syakur et al. [5] found that the peak of diurnal rainfall was observed in the early afternoon in the Barisan Mountains and the small islands on the east coast, such as Bangka and Belitung. Its study used data from Tropical Rainfall Measuring Mission multisatellite analysis (TRMM 3B42) version 07 for 17 years. The latest research on the diurnal characteristics of rainfall in the Sumatra region was carried out by Marzuki et al. [6 & 7] by utilizing the Automatic Rain Gauge (ARG) and Integrated Multisatellite Retrievals for the Global Precipitation Measurement (IMERG) final-run product version-6 for a period of 5 years (2015–2019). They found that the peak accumulation (PA), frequency (PF), and intensity (PI) of rainfall on Bangka Island were dominant during the day (1400 Local Standard Time/ LST (UTC + 7)). The average PA and PF are slightly low in the area compared to areas in the Barisan Mountains, both ARG and IMERG observations. Observations of the diurnal characteristics of rainfall in the Bangka and Belitung areas are still very limited [8, 9]. Prasetya et al. [9] found that topography greatly influences and controls local circulation, such as land and sea breeze, in forming convective clouds around the hills on the Bangka and Belitung Islands. They also found that the most robust sea breeze activity was observed during the March to May (MAM) period. Supari and Setiawan [8] found that rainfall patterns in the Bangka and Belitung areas are strongly influenced by the movement of the Inter-Tropical Convergence Zone (ITCZ) and monsoons. Due to limited research on the diurnal characteristics of rainfall for the Bangka and Belitung areas, here we will describe the climatology of diurnal characteristics by utilizing data with good quality, namely ARG and IMERG, following research that has been done previously by Marzuki et al. [6, 7, 10].

2 Data and Methods This study uses surface rainfall data from 8 stations of an automatic rain gauge (ARG) which spread across Bangka Island (7 stations) and Belitung Island (1 station). The data has an observation period from August 2015 to October 2019 and is provided by the Meteorology, Climatology, and Geophysics Agency (BMKG). ARG data has a temporal resolution of 10 min with a minimum sensitivity of 0.2 mm/10 min. Furthermore, the satellite data used is IMERG V06 data with an observation period of 2015–2020. The IMERG data is grid data with a spatial resolution of 0.1 × 0.1 and a temporal 30 min. More detailed information about these data can be read in the report of Tan et al. [11]. This data is provided free to the public by NASA which

Diurnal Variation of Rainfall Over Bangka Belitung Islands Determined …

429

can be accessed on the web https://disc.gsfc.nasa.gov/datasets/. Several studies have carried out IMERG data validation with surface rainfall data. Validation of hourly IMERG data for the Sumatra area has been carried out by Ramadhan et al. [12, 13], and Yusnaini et al. [14]. Bangka Belitung Islands is divided into two main islands, namely Bangka Island and Belitung Island, which are small islands to the east of Sumatra Island with an area of 81.725 km2 [8]. The topographical conditions of the Bangka and Belitung Islands can be seen in Fig. 1. These two islands generally have a distinctive topography consisting of hilly areas with the highest peak of 705 m above sea level and directly adjacent to the sea [9]. The distribution of ARG station coordinates can be seen in Fig. 1b. The diurnal characteristics of rainfall were analyzed according to Marzuki et al. [6, 7]. Precipitation amount (PA) is rainfall accumulation from all available observation data (mm/h). Precipitation frequency (PF) is the ratio of the total data that detects the occurrence of rain to the total of all data under observation (%). Precipitation Intensity (PI) was calculated from the total rainfall at the time of the rainfall (mm/h). In addition, this study also groups rain events into three different durations, i.e., 6 h. Rain events are defined as events within a specific period without data gaps for more than one hour [6, 7, 15, 16]. Rain that occurs after a significant data gap of one hour will be classified as a new rain event.

3 Result and Discussion Figure 2 shows the annual mean of annual rainfall from IMERG observations for the 2015–2020 observation period. The average annual rainfall is found to be relatively high on the Bangka and Belitung Island, namely 2500 to 3400 mm/year. Uniquely, the average rainfall on Belitung Island is higher (3400 mm/year) than on Bangka Island (2500 mm/year). Although lower than Belitung Island, rainfall on Bangka Island shows a different distribution between West Bangka (2653 mm/year) and South Bangka (2525 mm/year). Differences in the spatial distribution of annual rainfall on the Bangka and Belitung Islands were also observed from the TRMM Multi-satellite Precipitation Analysis 3B42 Real-Time product at daily scale (TMPA 3B42RT) for ten years by Supari and Setiawan [8]. They found that rainfall was smaller in the northern part of Bangka Belitung and higher in the central part of West Bangka and North of South Bangka (1800 and 3000 mm/year, respectively). The spatial distribution of observations from TMPA 3B42RT differs from IMERG observation. This condition is due to the difference in the observation data used of the two satellite data and the difference in spatial resolution between the two satellites, i.e., 0.25° (TMPA 3B42RT) and 0.1° (IMERG). Based on the grouping of rain types carried out by Aldrian et al. [17], Bangka and Belitung Islands are included in type B. The peak of rain is strongly influenced by the phenomenon of the ITCZ and monsoon, with two peaks of rain (April and December) [8, 17].

430

H. Yusnaini et al.

Fig. 1 Position and topography of Bangka and Belitung Islands a and location of ARG b

In this work, rainfall will be defined if the rainfall is 0.2/10 min from ARG and 0.6 /30 min from IMERG. Figure 3 shows the percentage of rain events with three different durations from the ARG and IMERG observations. ARG observations show that more than 80% of rain events on Bangka and Belitung Islands originate from short-duration rain events (< 3 h), and rain events which more than 6 h are rare (smaller than 5%) (Fig. 2a). Similar to ARG observations, IMERG also shows the Bangka and Belitung are dominated by rain with a small duration of 3 h (61.67 and 51.61%) (Fig. 2b).

Diurnal Variation of Rainfall Over Bangka Belitung Islands Determined …

431

Fig. 2 Spatial distribution of annual rainfall by IMERG observation during 2015–2020

Fig. 3 Distribution of percentage of rain events for duration of < 3 h, 3–6 h, > 6 h under observation of ARG a dan IMERG b, c, d

Rainfall events with a 3–6 h duration from IMERG observations showed a higher percentage on Bangka Island (38.56%) and lowered on Belitung Island (35.31%). Meanwhile, the incidence of rain with a duration of >6 h was found to be very low over the two islands (7.50–15.00%) compared to the surrounding seas, especially around the Karimata Strait. The dominance of rain with a duration of 0.5 × 1010 Jm−2 ) mostly contained in the shallow basin region of the Karimata Strait, the Natuna Sea, the Java Sea, and the Arafura Sea. The OHC standard deviations are larger (> 0.2 × 1010 Jm−2 ) in the western Pacific, along the main ITF route including the Sulawesi Sea and the Banda Sea, offshore of the west (south) Sumatra (Java) region, and the Indian Ocean along 5° S−13° N and 80−90° E. The ITF is likely responsible for these significant standard deviation values. In contrast, the OHC standard deviations are lower (< 0.05 × 1010 Jm−2 ) in the Karimata strait and the Java Sea (Fig. 1c). In order to understand the seasonal variability of upper OHC in the IMC, we first assessed the seasonal cycle spatial patterns in the IMC region. The magnitude of the seasonal cycle of OHC for the 300 m from 1990 to 2021 is shown in Fig. 1d. In general, the seasonal cycle of OHC displays a similar pattern to the OHC standard deviations. It is observed that the magnitude of the seasonal cycle within the marginal sea of the IMC, viz. the Sulawesi Sea, the Java Sea, and the Karimata Strait, is lower (< 1 × 1010 Jm−2 ) than the Banda Sea, which is reaching about 2 × 1010 Jm−2 . A comparison of the temporal means of the seasonal cycle of the marginal seas in the IMC over the period 1990 to 2021 is presented in Fig. 2. The mean seasonal cycle in the Banda Sea and Sulawesi is higher than in the Java Sea and Karimata Strait. The mean seasonal cycle of OHC in the Banda Sea exhibits a maximum OHC value of about 25 × 109 Jm−2 during the northwest monsoon season in February and March (Fig. 2a). Minimum OHC value of about 23 × 109 Jm−2 is observed during the southeast monsoon season from July to September. The seasonal cycles of OHC in the Banda Sea suggest that maximum and minimum are likely influenced by the seasonally reversing monsoon (Fig. 2a). Further comparison of mean seasonal cycles of the OHC among the Sulawesi Sea, the Java Sea, and the Karimata Strait identifies clearer bimodal variations (Fig. 2b–d). Predominantly higher OHC in those basins was observed during the southeast monsoon season (May to August) and pre-northwest monsoon (October to November). It appears that the pre-northwest monsoon wind has a clear influence on the seasonal cycle pattern in the Sulawesi Sea, the Java Sea, and the Karimata Strait. In the Sulawesi Sea, the peak and bottom values of OHC reach about 24.20 × 109 Jm−2 and 23.50 × 109 Jm−2 in June and February, respectively. The highest seasonal cycle of OHC in the Java Sea and the Karimata Strait attains a maximum of about 54.50 × 108 Jm−2 and 40.50 × 108 Jm−2 , respectively, in the month of May/June, and secondary high values in November (Fig. 2c and d). To explore the increasing rate of upper OHC at diverse spatial locations, the linear tendency in the OHC for the 0–300 m layers is quantified for the entire observation

Warming of the Upper Ocean in the Indonesian Maritime Continent

493

Fig. 2 Mean seasonal cycle of OHC over the period 1990 to 2021 in the domain-averaged over: a the Banda Sea; b the Sulawesi Sea; c the Java Sea; d the Karimata Strait

period. In Fig. 3, we show the linear decadal trend of OHC in the IMC computed for the 0–300 m layers from 1990 to 2021. The OHC linear trend is positively significant for most regions, ranging from 0.2 × 108 Jm−2 decade−1 to 7.0 × 108 Jm−2 decade−1 . The trend with larger values is found in the vicinity of the Halmahera eddy reaching up to 7.0 × 108 Jm−2 decade−1 . The Halmahera eddy is associated with the convergence of warm sea surface temperatures. Another maximum of OHC trends (~6.8 × 108 Jm−2 decade−1 ) is observed along the western Pacific. This is a region where the sea surface temperature resides very warm (> 28 °C) the whole year (Fig. 3). In general, the eastern basins in the IMC such as the Banda Sea and Sulawesi Sea have higher positive trends than the western basin as the Java Sea and Karimata Strait. The Banda Sea and Sulawesi Sea trends denote positive increases ranging from 3 × 108 Jm−2 decade−1 to 6 × 108 Jm−2 decade−1 . We evaluate the relationship of OHC variability in the IMC to the ENSO. Figure 4 shows the time series of deseasonalized OHC anomaly averaged over the Banda Sea, the Sulawesi Sea, the Java Sea, and the Karimata Strait in the upper 300 m in reference to the 1990 to 2020 baseline period, along with their linear decadal trend. The interannual and interdecadal variability of upper OHC in the Banda Sea and the Sulawesi Sea is quite similar in terms of their magnitude and their phase as well

494

M. F. A. Ismail et al.

Fig. 3 Geo-distribution of OHC decadal trend (× 108 J m−2 decade−1 ) for the 0–300 m in the IMC over the period of 1990 to 2021. Black lines exhibit 500 m isobath. The black stippled points indicate a significant trend where p-values less than 0.05

(Fig. 4a and b). In contrast, the OHC variability in the Java Sea and the Karimata Straits has similar patterns (Fig. 4c and d), albeit its magnitude is one order lower than the Banda Sea and Sulawesi Sea. The OHC variability in the Banda Sea and the Sulawesi Sea negatively correlates with the ENSO. The years with negative and positive anomalies of OHC in the Banda Sea and the Sulawesi Sea coincide with the strong El Niño and La Niña events, respectively (Fig. 4a and b). In contrast, the variability of OHC in the Java Sea and the Karimata Strait is such that the anomaly is positive (negative) during the El Niño (La Niña) events with a lag of about 3−6 months (Fig. 4c and d). It is clear from Fig. 4 that there is an increasing trend of OHC anomaly in our study region. Using the Mann–Kendall rank statistic test, these trends are evaluated for significance and determined at a 99.9% significant level. The Sulawesi Sea has the highest decadal warming trend of OHC of about 4.2 * 108 J m−1 decade1 , followed by the Banda Sea reaching up to 4.2 * 108 J m−1 decade1 . The OHC trend in the Java Sea and the Karimata Strait is about 2.0 * 107 J m−1 decade1 and 1.7 * 107 J m−1 decade1 , respectively.

4 Conclusions In this study, the variability of upper OHC over the period of 1990 to 2021 in the IMC is investigated based on the OCEAN5 reanalysis outputs. We show that the upper OHC in the internal seas of the IMC viz. the Banda Sea, the Sulawesi Sea, the Java Sea, and the Karimata Strait tend to increase, though it is spatially distinct. In addition, we have documented the seasonal, interannual, and interdecadal evolution

Warming of the Upper Ocean in the Indonesian Maritime Continent

495

Fig. 4 Time series of deseasonalized OHC anomaly (J m−2 ) of 0–300 m layer relative to 1990 to 2020 baseline climatology spatially averaged over: a the Banda Sea, b the Sulawesi Sea, c the Java Sea, and d the Karimata Strait with the climatological seasonal cycle removed. e Temporal distribution of the multivariate ENSO index over the same period in a to d. Blue (red) shadings denote the two strong El Niño (La Niña) periods identified by the multivariate ENSO index exceeding ± 1. Straight green line indicates the linear trend of OHC year−1

of upper OHC in the Banda Sea, the Sulawesi Sea, the Java Sea, and the Karimata Strait and quantified their long-term trend. For most regions, the OHC linear trend is positive, ranging from 0.2 × 108 Jm-2 decade-1 to 7.0 × 108 Jm-2 decade-1 . Compared to the western basins, including the Java Sea and Karimata Strait, the eastern basins of the IMC, viz. the Banda Sea and Sulawesi Sea, exhibit stronger positive tendencies. The ENSO and the OHC fluctuation in the Banda and Sulawesi Seas are inversely correlated. Strong El Niño and La Niña episodes both occur during the years with negative and positive OHC anomalies, particularly in the Banda Sea and the Sulawesi Sea, respectively. Acknowledgements The OHC data was downloaded from the Copernicus Climate Change Service (C3S) Climate Data Store (https://doi.org/10.24381/cds.67e8eeb7). Mochamad Furqon Azis Ismail was supported by the Alexander von Humboldt-Stiftung and contributed as the main contributor. Asep Sandra Budiman, Abdul Basit, Erma Yulihastin, Herlina Ika Ratnawati, Dewi Surinati, Adi Purwandana, Widodo Setiyo Pranowo, Subekti Mujiasih, Rahaden Bagas Hatmaja, and Praditya Avianto contributed as associate contributors to this paper.

496

M. F. A. Ismail et al.

References 1. IPCC.: The Physical Science Basis: Working Group I Contribution to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (2014) 2. Trenberth, K.E., Fasullo, J.T., Balmaseda, M.A.: Earth’s energy imbalance. J. Climate 27, 3129–44 (2014) 3. Levitus, S., Antonov, J., Boyer, T.: Warming of the world ocean, 1955–2003. Geophys. Res. Lett. 32 (2005) 4. Levitus, S., Antonov, J.I., Boyer, T.P., Baranova, O.K., Garcia, H.E., Locarnini, R.A., Mishonov, A.V., Reagan, J.R., Seidov, D., Yarosh, E.S.: World ocean heat content and thermosteric sea level change (0–2000 m), 1955–2010. Geophys. Res. Lett. 39 (2012) 5. Levitus, S., Antonov, J.I., Boyer, T.P., Stephens, C.: Warming of the World Ocean. Science 80(287), 2225–2229 (2000) 6. Trenberth, K.E., Fasullo, J.T.: Tracking earth’s energy: from El Niño to global warming. Surv. Geophys. 33, 413–426 (2012) 7. Xue, Y., Balmaseda, M.A., Boyer, T., Ferry, N., Good, S., Ishikawa, I., Kumar, A., Rienecker, M., Rosati, A.J., Yin, Y.: A comparative analysis of upper-ocean heat content variability from an ensemble of operational ocean reanalyses. J. Clim. 25, 6905–6929 (2012) 8. Cheng, L., Trenberth, K.E., Palmer, M.D., Zhu, J., Abraham, J.P.: Observed and simulated full-depth ocean heat-content changes for 1970–2005. Ocean Sci. 12, 925–935 (2016) 9. Palmer, M.D., Good, S.A., Haines, K., Rayner, N.A., Stott, P.A.: A new perspective on warming of the global oceans. Geophys. Res. Lett. 36 (2009) 10. Yoneyama, K., Zhang, C.: Years of the Maritime Continent. Geophys. Res. Lett. 47, e2020GL087182 (2020) 11. Iskandar, M.R., Ismail, M.F.A., Arifin, T., Chandra, H.: Marine heatwaves of sea surface temperature off south Java. Heliyon 7, e08618 (2021) 12. Ismail, M.F.A.: Characteristics of marine heatwaves off West Sumatra derived from highresolution satellite data. Hunan Univer. Natural Sci. 48, 130–136 (2021) 13. Ismail, M.F.A., Gerhaneu, N.Y., Yulihastin, E., Ratnawati, H.I., Purwandana, A.: Assessment of marine warming in Indonesia: a case study off the coast of West Sumatra. IOP Conf. Ser. Earth Environ. Sci. 718, 012006 (2021) 14. Ismail, M.F.A., Ribbe, J., Arifin, T., Taofiqurohman, A., Anggoro, D.: A census of eddies in the tropical Eastern boundary of the Indian Ocean. J. Geophys. Res. Oceans 126, 1–9 (2021) 15. Budiman, A.S., Bengen, D.G., Nurjaya, I.W., Arifin, Z., Ismail, M.F.A.: The spatio-temporal variability of chlorophyll-A and its physical variables in the South Java Sea shelf. Hunan University Natural Sci. 48 (2021) 16. Hersbach, H., Bell, B., Berrisford, P., Biavati, G., Horányi, A., Muñoz Sabater, J., Nicolas, J., Peubey, C., Radu, R., Rozum, I.: ERA5 hourly data on single levels from 1979 to present. Copernicus Climate Change Service (C3S) Climate Data Store (CDS) (2018) 17. Zuo, H., Alonso-Balmaseda, M., Mogensen, K., Tietsche, S.: OCEAN5: the ECMWF ocean reanalysis system and its real-time analysis component. ECMWF Tech. Memo (2018) 18. Zhang, T., Hoell, A., Perlwitz, J., Eischeid, J., Murray, D., Hoerling, M., Hamill, T.M.: Towards probabilistic multivariate ENSO monitoring. Geophys. Res. Lett. 46, 10532–10540 (2019)

Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) During Storm Intensification Episodes in the Early Dry Season Period Namira Nasywa Perdani, Ankiq Taofiqorahman, Erma Yulihastin, Rahaden Bagas Hatmaja, Gammamerdianti, Eka Putri Wulandari, Noersomadi, and Haries Satyawardhana Abstract Previous studies have mentioned that ocean–atmosphere interaction over the Maritime Continent played a prominent role in determining climate and weather variability on the intraseasonal variation scale in Indonesia. One of them could be proven by the anomalously wet-dry season over Java Island, which was generated by the warming of sea surface temperatures (SST) in the Indonesian Sea. However, the mechanism of the ocean–atmosphere interaction in these local seas in generating weather disturbances such as thunderstorms was still poorly understood. This study investigates the ocean–atmosphere interaction in the southern seas of Indonesia in inducing storms during the anomalously wet-dry season episodes. Here, we explored satellite and reanalysis datasets, including ocean and atmosphere variables in subdaily time resolution and various spatial resolutions, i.e., 0.25°, 0.083° and 0.1° derived from The European Centre for Medium-Range Weather Forecast (ECMWF) reanalysis 5th Generation (ERA5), Copernicus Marine Environment Monitoring Service (CMEMS), and Global Satellite Mapping of Precipitation (GSMaP), respectively. Those variables were investigated over the Southeast Tropical Indian Ocean (SETIO) during the early dry season of May–June 2021. Our finding described that during these episodes, the increase of rainfall over southwestern Indonesia did not correspond to the position of the Intertropical Convergence Zone (ITCZ) and might have been influenced by air-sea interaction over a regional scale. We also confirmed that storm activities over southwestern Indonesia were associated with strengthening sea surface current (SSC) along the west coast of Sumatra and forming eddies in the southern Java Sea. We also noticed that variability in both SSC and SST in the southern Java Sea, which occurred on a sub-daily scale, may influence a dynamic weather pattern over southwestern Indonesia. N. N. Perdani (B) · A. Taofiqorahman Marine Sciences Department, Faculty of Fisheries and Marine Sciences, Padjajaran University, Bandung, Indonesia e-mail: [email protected] E. Yulihastin · R. B. Hatmaja · Gammamerdianti · E. P. Wulandari · Noersomadi · H. Satyawardhana National Research and Innovation Agency, Jakarta, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_46

497

498

N. N. Perdani et al.

1 Introduction The ocean–atmosphere interaction is a coupling system that can influence weather and climate cycles in the world. Both are exchanged heat energy and momentum to create a dynamic between the ocean and the atmosphere. The ocean–atmosphere dynamics play an important role in triggering convective activity in Indonesia Maritime Continent (IMC) from local to large scale [1]. On a local scale, land-sea breezes interaction can increase convective activity that causes local rainfall due to the diurnal SST variability [2–4]. In global scales, phenomena such as ENSO, IOD, and MJO influenced convective activity in various time and spatial scales. The complex dynamics of the ocean–atmosphere interaction in the IMC may produce seasonal anomalies in Indonesia, particularly on Java Island. A previous study stated that the anomalously wet in dry season over Java was most influenced by local ocean 37% compared to remote forcings [5]. Therefore, the warming SST in southern Java affects convective activities on the island of Java during the dry season. Furthermore, previous studies have suggested that the slightest change in SST can cause large-scale changes in atmospheric circulation [6–9]. However, the dynamical processes over surface level between air-sea components over southern waters of IMC in increasing convective activity associated with storms during the anomalously wet in dry season were still less of knowledge. Therefore, this study aims to investigate the ocean–atmosphere interaction in the southern waters of Indonesia in triggering storms during the anomalously wet in dry season by exploring three storm episodes during May to June 2021.

2 Methods We use satellite and reanalysis data of ocean and atmospheric parameters to investigate the air-sea interaction in developing convective activity associated with the convective storm in southern IMC. Hourly oceanic parameter data such as sea surface temperature (SST) and sea surface currents (SSC) are taken from the Copernicus Marine Environment Monitoring Service (CMEMS) (https://marine.copernicus.eu/), with a 0.083° × 0.083° spatial resolution. The hourly atmospheric parameter data such as Outgoing Longwave Radiation (OLR), surface temperature (2 m), latent heat, and divergence wind (850 mb) were obtained from the European Center for MediumRange Weather Forecast (ECMWF) reanalysis 5th Generation (ERA5) with a spatial resolution of 0.25° × 0.25° [10]. Furthermore, hourly rainfall data are obtained from the Global Satellite Mapping of Precipitation (GSMaP) with rain gauge data validation [11]. The GSMaP_gauge data have been developed to fill the data gaps in the rainfall forecast between satellite and rain gauge data due to differences in spatial and temporal resolution [12]. Therefore, there is an increasing spatial resolution of GSMaP from 0.25° × 0.25° to 0.1° × 0.1° for the GSMaP_gauge used in this study.

Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) …

499

The area study is located between 20–3°S; 95–120°E, so-called Southeast Tropical Indian Ocean (SETIO). To describe the mean state of the atmosphere–ocean condition, we used a synoptical analysis of the average May–June 2021 and defined the results as a background condition during the early period of the dry season. Furthermore, we perform a composite analysis to investigate the diurnal variation in every six-hourly data. For SST anomaly, we calculate the anomaly by subtracting diurnal data from the data average for two months (May–June).

3 Results 3.1 Background Conditions Globally, convective activity in the IMC during the early dry season 2021 does not correspond to the Intertropical Convergence Zone (ITCZ) position, as seen in Fig. 1a and b. However, the condition of convective activity in May showed normal conditions, but for June, it did not show a strong signal of the dry season. In this condition, the sun’s position should move closer to the north latitude, and the Australian monsoon winds begin to strengthen, causing cooling of SST in southern Indonesia. Still, in this case, it will not occur in June 2021. This month shows the SST warming south of Indonesia, and SETIO is shown in Fig. 1c. The increasing of SST in the SETIO area causes a vortex to form, which is indicated by a wind divergence of 850 mb (Fig. 1d). This condition describes globally that the beginning of the dry season in Indonesia in 2021 shows a wet-dry season signal based on OLR, SST Climatological Anomaly, and 850 mb wind divergence data.

3.2 Ocean–Atmosphere Interaction in Sub-Daily Cases We did an average of May−June 2021 every six hours, which shows that convective activity is still concentrated in the SETIO region during the night-to-early morning time, early morning-to-morning time, morning-to-middle day, afternoon-to-nighttime (Fig. 2a), and its OLR (Fig. 2b). This group of convective clouds in the SETIO area is centered above the eddies as shown in Fig. 2d. On the other hand, we can detect diurnal SST variability between night-to-early morning time, early morningto-morning time, morning-to-middle day time, and afternoon-to-night-time. In the early morning-to-morning time, the SST anomaly (Fig. 2c) decreased in the coldest conditions. In contrast, in the afternoon-to-night-time, the SST anomaly increased in the warmest conditions among other periods. We can conclude that the influence of the ocean dominates to maintain convective activity in the SETIO over the atmosphere, such as temperature at 2 m (Fig. 2e) and latent heat (Fig. 2f) during the early dry season.

500

N. N. Perdani et al.

Fig. 1 Averages for May (left) and June (right) of 2021 for: a rainfall derived from GSMaP satellite data, b OLR carried from NCEP/NCAR reanalysis, c SST anomaly, and d 850 mb wind divergence from CMEMS and ERA5 reanalysis data, respectively

Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) …

501

Fig. 2 Average May–June 2021 of six-hourly data from left to right: night-to-early morning time (20:00−01:00 LST); early morning-to-morning time (02:00−07:00 LST); morning-to-middle day time (08:00−13:00 LST); and afternoon-to-night-time (14:00−19:00 LST) for: a rainfall, b OLR, c SST anomaly, d SSC, e temperature at 2 m, and f latent heat

502

N. N. Perdani et al.

We did breakdown scales on convective activity daily, divided into 3 cases with the high precipitation affecting Java Island (Fig. 3). The first case was on May 27−28, 2021. On May 27 at night, around the southwest of the island of Sumatra, rain clouds began to form (Fig. 4a). The storm got bigger spreading to the coast of Sumatra on May 28, 2021, in the early morning-to-morning time; this can be seen from its OLR cover (Fig. 4b). The storm is getting bigger caused of the increasing of SST (Fig. 4c) in the sea near Sumatra due to the Kelvin waves signed by the increasing of SSC near Sumatera Island (Fig. 4d). Rain causes a drop in temperature at 2 m (Fig. 4e) in the early morning-to-middle day time periods. Then, there is another convective activity at 20S formed due to the increase of latent heat near Australian waters (Fig. 4f). We can conclude that the forming of convective clouds in May 2021 was influenced by the presence of Kelvin waves, thus increasing SST in the western waters of Sumatra, while further south the convective activity was caused by an increase in latent heat. The next case is concentrated convective clouds over the island of Java (Fig. 5b) and the detection line-shaped storm (Fig. 5a). This convective activity is formed in the south of Java due to an increase in SST in the area as seen from its SST anomaly (Fig. 5c). The increasing of SST occurs due to the mass of water carried by SSC (Fig. 5d) called Indonesian Through Flow (ITF) from the Makassar Strait to the Indian Ocean through the Lombok strait. Then, the rain clouds propagated to the north of Java in the early morning-to-middle day time, but the movement of these rain clouds was caused by temperature at 2 m in the Java Sea. Same as the first case, there is convective activity at the southern Latitude. This rain cloud is not caused by SST variability but rather by high latent heat values near the Australian waters (Fig. 5f).

Fig. 3 Time series of averages precipitation, latent heat flux, temperature at 2 m, and SST for: a 27−28 May 2021 (case 1), b 16−17 June 2021 (case 2), and c 23−24 June 2021 in SETIO area

Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) …

503

Fig. 4 Same as Fig. 2, but for case 1

The following week there was another line-shaped storm. Initially, a storm had begun to form during the night-to-early morning on June 23 (Fig. 6a). Then the next day, on June 24, the storm pattern became clearer to form a line, and there were multi-cells of rain clouds in the north of the island of Java. It can be seen from the OLR (Fig. 6b) that convective activity in the north is moving toward the south, which makes convective storms even more amplified in the morning-to-middle daytime on June 24. The factor that triggered the concentration of convective clouds was the

504

N. N. Perdani et al.

Fig. 5 Same as Fig. 2, but for case 2

increase in SST (Fig. 6c) due to ITF (Fig. 6d) and latent heat (Fig. 6f). Rain clouds make the affected area experience a temperature at 2 m drops with a value reaching 26 °C shown in Fig. 6e.

Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) …

505

Fig. 6 Same as Fig. 2, but for case 3

4 Conclusion We investigated Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) in The Early Dry Season Period causing a lot of convective activity to occur above the Maritime Continent; it is proven that in June 2021, convective activity is not strongly associated with the position of the ITCZ. The factor that maintains convective activity

506

N. N. Perdani et al.

is the influence of ocean–atmosphere interactions in local waters. We found that on a six-hourly average for two months, convective activity formed just above the eddies in the SETIO area. In addition, the ocean factor, such as SST carried by the ITF from the Makassar strait through the Lombok strait, supplies the warm surface water mass, causing increased SST variability in southern Indonesia. This can be proven by several cases of storms that occur sub-daily in southern Indonesia, dominated by ocean factors compared to atmospheric factors if the storm occurs near the coast of the island of Java. Previous studies have investigated that the influence of local waters dominates the occurrence of the anomaly wet-dry season on the island of Java compared to the impact of global disturbances [5]. It differs from the convective activity that occurs close to Australian waters; an increase in latent heat value causes this. This study still has limitations on the other parameters to observe the effect of the air-sea interaction in the SETIO area, causing the dry season anomaly in Indonesia. Therefore, for further studies, it is necessary to consider other ocean– atmosphere parameters to improve the weather model in Indonesia. In addition, this study reviewed that updates to the data inputted in the weather model need to be carried out on a sub-daily to capture convective activity on the IMC with higher resolution. Acknowledgements This research has been supported by the National Research and Innovation Agency under the Cruise Day Facilitation Program with grant number 376/II/FR/3/2022.

References 1. Xue, P., Malanotte-Rizzoli, P., Wei, J., Eltahir, E.A.B.: Coupled Ocean-atmosphere modeling over the Maritime Continent: a review. J. Geophys. Res. Ocean. 125(6), 1–18 (2020) 2. Birch, C.E., et al.: Scale interactions between the MJO and the western Maritime Continent. J. Clim. 29(7), 2471–2492 (2016) 3. Im, E.S., Eltahir, E.A.B.: Simulation of the diurnal variation of rainfall over the western Maritime Continent using a regional climate model. Clim. Dyn. 51(1–2), 73–88 (2018) 4. Thompson, B. et al.: A high-resolution atmosphere–ocean coupled model for the western Maritime Continent: development and preliminary assessment, vol. 52, pp. 7–8. Springer, Berlin, Heidelberg (2019) 5. Suaydhi, Y.E.P.M.F., Sofiati, I.: Oceanic effect on precipitation development in teh Maritime Continent during anomalously-wet dry seasons in Java Indones. J. Geogr. 53(3), 328–339 (2021) 6. Bjerknes, J.: Monthly weather reyiew atmospheric teleconnections from the equatorial Pacific Mon. Weather Rev. 97(3), 163–172 (1969) 7. Hendon, H.H.: Indonesian rainfall variability: impacts of ENSO and local air-sea interaction. J. Clim. 16(11), 1775–1790 (2003) 8. Zhang, C., Anderson, S.P.: Sensitivity of intraseasonal perturbations in SST to the structure of the MJO. J. Atmos. Sci. 60(17), 2196–2207 (2003) 9. Li, Y., et al.: Assessing the role of the ocean–atmosphere coupling frequency in the western Maritime Continent rainfall Clim. Dyn. 54(11–12), 4935–4952 (2020) 10. Hersbach, H., et al.: The ERA5 global reanalysis Q. J. R. Meteorol. Soc. 146(730), 1999–2049 (2020)

Air-Sea Interaction Over Southeast Tropical Indian Ocean (SETIO) …

507

11. Kubota, T., et al.: Global satellite mapping of precipitation (GSMaP) products in the GPM Era Adv. Glob. Chang. Res. 67, 355–373 (2020) 12. Bui, H.T., Ishidaira, H., Shaowei, N.: Evaluation of the use of global satellite–gauge and satellite-only precipitation products in stream flow simulations. Appl. Water Sci. 9(3), 1–15 (2019)

Anomalous Sea Surface Temperature and Chlorophyll-a Induced by Mesoscale Cyclonic Eddies in the Southeastern Tropical Indian Ocean During the 2019 Extreme Positive Indian Ocean Dipole Rahaden Bagas Hatmaja, M. Rizqi Ramadhan, Sigit Kurniawan Jati Wicaksana, and Suaydhi Abstract The positive Indian Ocean Dipole (pIOD) event in 2019 was notable as the strongest event in the last two decades. This event is characterized by cold sea surface temperature anomaly (SSTA), high chlorophyll-a concentrations (CHL), and intense mesoscale eddy activity off the coast of Java in the southeast tropical Indian Ocean (SETIO), coincident with upwelling season in June to November. In this study, the modulation effects of mesoscale eddies on the local SSTA and CHL are investigated by analyzing daily reanalysis and remote-sensed data from January to December 2019. The variations of SSTA and CHL are dominated by upwelling-induced SSTA and CHL up to ~ 80%. Nevertheless, the eddy-induced SSTA and CHL have remarkable contribution about ~ 25% to their monthly variation, especially during upwelling season that coincident with the peak of pIOD from June to September. In this period, higher number and stronger cyclonic eddies (CEs) are generated, which significantly contributes to the colder SSTA and higher CHL. Moreover, stronger CEs induce colder SSTA and higher CHL by both horizontal advection, which transports cold, high-CHL water from the coast to 400 km offshore, and vertical upwelling, which raises nutrient-rich cold water from the subsurface to the surface.

R. B. Hatmaja (B) · Suaydhi Research Centre for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected] M. R. Ramadhan Oceanography Department, Faculty of Earth Sciences and Technology, Bandung Institute of Technology, Bandung, Indonesia S. K. J. Wicaksana Research Centre for Oceanography, National Research and Innovation Agency, Jakarta, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_47

509

510

R. B. Hatmaja et al.

1 Introduction In 2019, the Indian Ocean experienced the strongest positive Indian Ocean Dipole (IOD) event in the last two decades [1]. This event was characterized by low sea surface temperature (SST) anomalies and high chlorophyll-a (CHL) concentrations off the coast of Java in the southeastern tropical Indian Ocean (SETIO), while reversed characteristics in the western pole near Somalian coast [1, 2]. In Fig. 1a, the daily Dipole Mode Index (DMI) from 1 January to 31 December 2019, which is calculated as the difference of the SST anomaly (SSTA) in the eastern equatorial Indian Ocean (90–110 °E, 10 °S–0 °N) and the SSTA in the western equatorial Indian Ocean (50– 70 °E, 10 °S–10 °N), shows that the pIOD event initiated in May and reached its peak in October up to ~ 2.5 °C, then weakens in the following months. The average SSTA and CHL concentrations in October reached over ~ –3 °C and ~ 0.5 mg/m3 on the east IOD zone, respectively. Moreover, these anomalous ocean activities were attributed to the enhanced upwelling during positive IOD (pIOD) event that advected higher-nutrient cold deep water to the surface [1, 3]. However, this significant effect of wind-driven coastal upwelling on SETIO was merely limited to 100 km from the coast [3].

Fig. 1 a Time series of Dipole Mode Index (DMI) from 1 January to 31 December 2019 and distribution of mean b sea surface temperature anomaly (SSTA) as well as c chlorophyll-a concentration (CHL) on the eastern pole of the Indian Ocean Dipole (IOD) during positive IOD’s peak in October 2019

Anomalous Sea Surface Temperature and Chlorophyll-a Induced …

511

In addition, the widespread of SSTA and CHL for more than 200 km offshore during pIOD event are also controlled by high cyclonic eddies (CE) activity [4], as the result of baroclinic instabilities that are associated with the South Java Current (SJC), South Equatorial Current (SEC), and Indonesian Throughflow (ITF) on this region [5–8]. Additionally, the wind forcing of pIOD event can enhance the vertical shear of the background currents and favor energetic CEs off Java coast [4]. Previous study has investigated the effects of mesoscale eddies on the CHL bloom off the Java coast during the 2006 pIOD event and showed that the peak CHL occurred around the peak phase of the IOD, and the active CEs contributed to eddy upwelling that increased the offshore CHL [9]. On the other hand, another study demonstrated that the anti-cyclonic eddy (AE) activities in the SETIO were stronger during the 1994 IOD event, which countered the anomalous cooling [10]. These results imply a probable linkage between mesoscale eddies and SSTA, as well as CHL, on the interannual timescale. Therefore, the investigation of relative effects of mesoscale eddies on the SSTA and CHL distribution off the coast of Java during 2019 IOD, as the recent event, is needed.

2 Data and Methods 2.1 Data The assessment of SSTA and CHL of this study is conducted by using daily SSTA and CHL data from 1 January to 31 December 2019, which were retrieved from NOAA Optimum Interpolation SST V2 High-Resolution Dataset on a 0.25° global grid (https://psl.noaa.gov/) and Copernicus Global Ocean Colour (CopernicusGlobColour) Satellite Observations product on a 4 km global grid (https://marine. copernicus.eu/), respectively. The daily SSTA data were calculated relative to a 1971–2000 climatology. Moreover, the mesoscale eddies were identified based on daily merged satellite altimetry multimission data, which includes sea level anomaly (SLA) and surface geostrophic currents data with a spatial resolution of 0.25° for the same period. SLA is defined as the height of water over the mean sea surface, that is computed with respect to a twenty-year mean reference period (1993–2012). These data were collated during the following satellite altimetry missions: Jason1/2/3, Sentinel-3A, HY-2A, SARAL/AltiKa, Cryosat-2, ERS1/2, TOPEX/Poseidon, ENVISAT, GFO, and ERS1/2, which are provided by the E.U. Copernicus Marine Environment Monitoring Service (https://marine.copernicus.eu/).

512

R. B. Hatmaja et al.

2.2 Eddy Detection and Tracking Methods The analysis of this study is conducted for eddies with a lifetime more than ten days [11]. In addition, only eddies generated between 105–117 °E and 6–12 °S are tracked. The eddies were identified and tracked using the hybrid eddy census algorithm. The hybrid algorithms combine a geometric criterion with an Okubo-Weiss parameter resulting in a more reliable detection from SLA compared to only using an OkuboWeiss parameter or a geometric criterion [12]. In this hybrid method, the presence of a geostrophic eddy is defined as a coherent region detected within a closed SLA contour with minimum 2 cm and an Okubo-Weiss parameter below −2 × 10−12 s−2 . Moreover, a Hanning filter is applied twice on the Okubo-Weiss parameter [12, 15]. The Okubo-Weiss parameter (W ) is computed as follows [13, 14]: W = Sn2 + Ss2 + ξ 2 Sn =

∂v ∂u ∂v ∂u ∂u ∂v − , Ss = + ,ξ = + ∂x ∂y ∂x ∂y ∂x ∂y

(1) (2)

where u and v are the horizontal velocity components in the x and y directions. The eddy tracking method basically assumes that an eddy detected in one SLA map is the same eddy identified in the following SLA map if the distance in a nondimensional property space is minimal. This methods follows the following equation [16]:

X e1,e2

     X 2  R 2  ξ 2  = + + X0 R0 ξ0

(3)

where X is the spatial range between the eddy cores X e1 and X e2 , R is the variation of diameter, ξ is the variation of vorticity, X 0 is a characteristic length scale (25 km), R0 is a typical eddy radius scale, and ξ0 is a characteristic vorticity (10 − 5 s − 1).

3 Results and Discussion 3.1 The Characteristics of Mesoscale Eddies in the Southeastern Tropical Indian Ocean In 2019, a total of 96 CEs and 80 AEs were detected, which distributed more than 100 km offshore (Fig. 2a). Areas of a higher total number of CEs and AEs are dominated in the vicinity of the SEC, across the eastern edge of the SJC near 115 °E

Anomalous Sea Surface Temperature and Chlorophyll-a Induced …

513

Fig. 2 a The spatial distribution and b monthly number of detected mesoscales eddies in the Southeastern Tropical Indian Ocean (SETIO) in 2019. Black contour denotes offshore distance with interval of 100 km. Black box denotes the area of interest

and 12 °S, within the Savu Sea, and the ITF region [6, 8, 17]. Therefore, the analysis of mesoscale eddies’ influence on SSTA and CHL in SETIO will be focused on 6° to 12 °S and 105° to 117 °E (area of interest). In addition, temporal variations of mesoscale eddies in Fig. 2b show fluctuation of CEs’ and AEs’ formation. Number of AEs’ formation dropped dramatically in May and escalated in June, as the number of CEs showed reversed dynamics. However, the number of CEs multiplied in the following months until its peak in September. This rapid growth, as well as more significant number of CEs than AEs, is suggested as the influence of strengthened ITF outflow transport in Lombok Strait that modulates the formation of CE during the peak of pIOD event in this region [4, 18]. As the CE is more dominant than AE in 2019, the mesoscale eddies’ characteristics focus on CE properties (Table 1). Along with significant number of CEs’ formation from July to October, the size of CEs was also increasing. From this period, the mean amplitude (radii) of the CEs ranges from 3.97 cm (89.1 km) to 6.52 cm (91.6 km). Besides its most significant number, the CEs in September also live longest, with mean lifetime extended than 50 days or almost two months later. Table 1 Monthly cyclonic eddy properties derived from tracked eddies, for mean lifetime (τ ), mean amplitude (η), and mean radii (R) Month η (cm) R (km)

Jan 4.86

Feb 4.29

Mar 6.16

Apr 3.88

May 4.22

Jun 5.03

Jul 3.97

Aug 4.76

Sep 6.52

Oct 5.66

Nov 4.71

Dec 4.43

91.0

93.1

88.6

87.7

92.5

94.7

89.1

77.7

89.3

91.6

88.3

88.0

τ (days) 26.0

26.9

35.1

31.3

22.3

22.0

20.5

25.6

51.5

25.3

20.3

15.3

514

R. B. Hatmaja et al.

3.2 The Variations of Eddy-Induced Sea Surface Temperature Anomalies and Chlorophyll-a Concentration Based on spatial distribution SSTA, CHL, and CE during pIOD event, the analysis of eddy-induced SSTA and CHL is focused on the area of interest that discussed on the previous section (see Fig. 2a). To examine the impact of CE on SSTA and CHL variation, the regional-averaged SSTA and CHL are divided into two categories based on offshore distance: upwelling region (0–100 km offshore) and mesoscale eddies region (100–400 km offshore). Based on monthly variation of mean SSTA and CHL in the SETIO (Fig. 3), both parameters show similar characteristics that are dominated by upwelling-induced SSTA and CHL up to ~ 80%, especially during the upwelling season coincident with pIOD’s generation from June to October [19]. Nevertheless, the eddy-induced SSTA and CHL have remarkable contribution (~25%) to its monthly variation from June to September, which is associated to rapid growth of CEs’ number (Fig. 2b). It can be concluded that CE has significant impact in modulating SSTA and CHL in SETIO during the 2019 pIOD event, especially from June to October, which peaked up to –3.3 °C and 1.7 mg/m3 . This impact can be explained through vertical upwelling that raises nutrient-rich cold water from the subsurface to the surface, as well as horizontal advection that transports cold water with high CHL near the coast to the offshore area. In addition to inspect the influence of CE on the SSTA and CHL’s variations, the relationship between the eddy’s amplitude that characterized by the SLA of the eddy’s center and the mean SSTA, as well as CHL inside the eddies, is discussed furthermore. As shown in Fig. 4, stronger CEs (with larger SLA in the center) induce colder SSTA and higher CHL. Nonetheless, the strong correlation between CE’s strength, SSTA, and CHL is significant to the eddies with amplitude larger than 6 cm. It is suggested because several newborn eddies, with an amplitude of less than 6 cm, were formed near to the upwelling region, thus resulting in a bias in the eddy-induced SSTA and CHL variations on these ranges.

Fig. 3 Monthly mean of area-averaged a SSTA and b CHL on upwelling’s region (0−100 km offshore) and mesoscale eddies’ region (100−400 km offshore) over the area of interest

Anomalous Sea Surface Temperature and Chlorophyll-a Induced …

515

Fig. 4 Eddy-induced a sea surface temperature anomaly and b chlorophyll-a concentration as a function of cyclonic eddies’ amplitude. The curves are second-order polynomial fits

4 Conclusion The pIOD event in 2019 is characterized by cold SSTA, high CHL, and intense CE in SETIO, which peak in October. The modulation effects of CE on the offshore SSTA and CHL distribution within 400 km in the SETIO were investigated using satellite data from January to December 2019. Based on eddies detection and tracking methods, 96 CEs and 80 AEs are detected and tracked. These eddies dominated in the vicinity of the SEC, across the eastern edge of the SJC near 115 °E and 12 °S, within the Savu Sea, and the ITF region. In addition, temporal variations of mesoscale eddies show that the number of CEs grows rapidly from June until its peak in September. It is suggested as the influence of strengthened ITF outflow transport in Lombok Strait modulates the formation of long-lived CE. Moreover, the size of CEs was growing along with its swift growth, with the mean amplitude (radii) of the CEs ranging from 3.97 cm (89.1 km) to 6.52 cm (91.6 km). Furthermore, monthly variation of mean SSTA and CHL in the SETIO is dominated by upwelling-induced SSTA and CHL up to ~80%, especially during the upwelling season coincident with pIOD. However, the eddy-induced SSTA and CHL have remarkable contribution by ~25%, which is associated to rapid growth of CEs’ number. It can be concluded that CE has significant impact on modulating SSTA and CHL in SETIO during the 2019 pIOD event from June to October. This impact can be explained through both vertical upwelling process of nutrient-rich cold water that rises from the subsurface to the surface, as well as horizontal advection that transport cold water with high CHL near the coast to the offshore area. Moreover, the eddy-induced SSTA and CHL are strongly correlated with the eddy amplitude, as characterized by SLA at the eddy center, with strong CEs inducing colder SSTA and higher CHL anomalies. It is interesting to note that this close relation is merely significant to the eddies with amplitude larger than 6 cm, due to its bias as the result of the formation of newborn eddies, with an amplitude of less than 6 cm, that firstly generated close to the upwelling region. Nevertheless, this study results in a new insight for understanding the physical-biogeochemical oceanic processes in SETIO as the eastern pole of IOD, which led for fisheries management in this region.

516

R. B. Hatmaja et al.

Acknowledgements The authors gratefully acknowledge the E.U. Copernicus Marine Service Information and Giovanni for providing satellite data. This research was supported by the Indonesian National Research and Innovation Agency under Aeronautics and Space Research Fund Program [No. 6/III/HK/2022].

References 1. Shi, W., Wang, M.: A biological Indian Ocean Dipole event in 2019. Sci Rep 11, 2452 (2021). https://doi.org/10.1038/s41598-021-81410-5 2. Saji, N.H., Goswami, B.N., Vinayachandran, P.N., Yamagata, T.: A dipole mode in the tropical Indian Ocean. Nature 401, 360–363 (1999). https://doi.org/10.1038/43854 3. Hatmaja, R.B., Munthe, C.C., Yulihastin, E., Pramudia, K.E.: Atmospheric response to the Southern Java upwelling variability associated with positive Indian Ocean Dipole Event. pp. 25– 37. (2022) 4. Yang, G., Zhao, X., Li, Y., et al.: Chlorophyll variability induced by mesoscale eddies in the Southeastern tropical Indian Ocean. J. Mar. Syst. 199, 103209 (2019). https://doi.org/10.1016/ j.jmarsys.2019.103209 5. Azis Ismail, M.F., Ribbe, J., Arifin, T., et al.: A census of Eddies in the tropical Eastern boundary of the Indian Ocean. J Geophys Res Ocean 126, 1–9 (2021). https://doi.org/10.1029/2021JC 017204 6. Feng, M., Wijffels, S.: Intraseasonal Variability in the south equatorial current of the East Indian Ocean. J. Phys. Oceanogr. 32, 265–277 (2002). https://doi.org/10.1175/1520-0485(200 2)032%3c0265:IVITSE%3e2.0.CO;2 7. Jia, F., Wu, L., Qiu, B.: Seasonal modulation of eddy kinetic energy and its formation mechanism in the Southeast Indian Ocean. J. Phys. Oceanogr. 41, 657–665 (2011). https://doi.org/ 10.1175/2010JPO4436.1 8. Yu, Z., Potemra, J.: Generation mechanism for the intraseasonal variability in the IndoAustralian basin. J. Geophys. Res. 111, C01013 (2006). https://doi.org/10.1029/2005JC 003023 9. Iskandar, I., Sasaki, H., Sasai, Y., et al.: A numerical investigation of eddy-induced chlorophyll bloom in the southeastern tropical Indian Ocean during Indian Ocean Dipole—2006. Ocean Dyn. 60, 731–742 (2010). https://doi.org/10.1007/s10236-010-0290-6 10. Ogata, T., Masumoto, Y.: Interactions between mesoscale eddy variability and Indian Ocean dipole events in the Southeastern tropical Indian Ocean—case studies for 1994 and 1997/1998. Ocean Dyn. 60, 717–730 (2010). https://doi.org/10.1007/s10236-010-0304-4 11. Faghmous, J.H., Frenger, I., Yao, Y., et al.: A daily global mesoscale ocean eddy dataset from satellite altimetry. Sci. Data 2, 150028 (2015). https://doi.org/10.1038/sdata.2015.28 12. Halo, I., Backeberg, B., Penven, P., et al.: Eddy properties in the Mozambique Channel: A comparison between observations and two numerical ocean circulation models. Deep Res. Part II Top Stud. Oceanogr. 100, 38–53 (2014). https://doi.org/10.1016/j.dsr2.2013.10.015 13. Okubo, A.: Horizontal dispersion of floatable particles in the vicinity of velocity singularities such as convergences. Deep Sea Res. Oceanogr. Abstr. 17, 445–454 (1970). https://doi.org/10. 1016/0011-7471(70)90059-8 14. Weiss, J.: The dynamics of enstrophy transfer in two-dimensional hydrodynamics. Phys. D Nonlinear Phenom. 48, 273–294 (1991). https://doi.org/10.1016/0167-2789(91)90088-Q 15. Chelton, D.B., Schlax, M.G., Samelson, R.M.: Global observations of nonlinear mesoscale eddies. Prog. Oceanogr. 91, 167–216 (2011). https://doi.org/10.1016/j.pocean.2011.01.002 16. Penven, P.: Average circulation, seasonal cycle, and mesoscale dynamics of the Peru current system: a modeling approach. J. Geophys. Res. 110, C10021 (2005). https://doi.org/10.1029/ 2005JC002945

Anomalous Sea Surface Temperature and Chlorophyll-a Induced …

517

17. Qiu, B., Mao, M., Kashino, Y.: Intraseasonal variability in the Indo-Pacific throughflow and the regions surrounding the Indonesian Seas. J. Phys. Oceanogr. 29, 1599–1618 (1999). https:// doi.org/10.1175/1520-0485(1999)029%3c1599:IVITIP%3e2.0.CO;2 18. Sprintall, J., Révelard, A.: The Indonesian throughflow response to Indo-Pacific climate variability. J. Geophys. Res. Ocean 119, 1161–1175 (2014). https://doi.org/10.1002/2013JC 009533 19. Saji, N.H., Yamagata, T.: Structure of SST and surface wind variability during Indian Ocean Dipole mode events: COADS observations. J. Clim. 16, 2735–2751 (2003). https://doi.org/10. 1175/1520-0442(2003)016%3c2735:SOSASW%3e2.0.CO;2

Modeling of Surface Current-Driven Water Pollution and ASEAN Countries Index Water Quality Assessment in the Rupat Strait, Riau Province, Indonesia Koko Ondara

and Ulung Jantama Wisha

Abstract Seawater degradation has become a concern in the strategic location of Rupat Strait, where various industries exist. This study aims to determine the currentdriven water pollution in the Rupat Strait. We directly surveyed water quality parameters (pH, salinity, temperature, and DO). The other environmental parameters, such as BOD5, ammonia, phosphate, nitrate, TSS, and heavy metals, were analyzed in the laboratory. Two-dimensional hydrodynamic of MIKE21 flow model FM was simulated to understand the sea current surface patterns whereby the modeled tidal current data was superimposed with water quality to predict the current-driven pollutant. Water quality and environmental parameters showed a different response to the tidal current pattern. The assessed parameters were above (below) the quality standard, indicating water pollution and degradation, particularly for nutrients and heavy metals, with a deviation of more than 0.03 mg/L compared to the quality standards for marine waters. The highest pollutant accumulation was detected in the middle of the strait, where the current magnitude ranged from 0.02 to 0.21 m/s. In contrast, in the north and south gates of the waterway, the pollutant concentration was not too significant, even though the current profiles tend to get more robust. Aside from hydro-oceanographic features, the sources and intake of industrial waste-induced pollution play a substantial role in determining the pollution level within the Rupat Strait.

Koko Ondara and Ulung Jantama Wisha are the main contributor of this manuscript. K. Ondara (B) · U. J. Wisha Research Center for Oceanography, National Research and Innovation Agency, 14430 Jakarta, Indonesia e-mail: [email protected] U. J. Wisha Department of Physics and Earth Sciences, University of the Ryukyus, Nishihara 903-0213, Japan © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_48

519

520

K. Ondara and U. J. Wisha

1 Introduction Rupat Strait is a small channel positioned between the Malacca Strait and Rupat Island, Riau Province, with a channel length of approximately 72.4 km and a width ranging from 3.8 to 8 km. Rupat Strait is a significant area, becoming the center of oil and gas industries with a dense shipping activity. Rupat Island is composed of peatland, resulting in a vulnerable coastal area when an extreme oceanographical state occurs at certain times [1]. This condition has become a significant concern for a long time because the Rupat Strait supports tourism, fishery, and industry [2]. On the other hand, the rapid development in the surrounding Rupat Island leads to a downgrade in the water quality. It is in line with the increase of industrial wastes and oil spills that jeopardize the marine biota and even the local people. The primary pollutant in the Rupat Strait is derived from offshore oil spilling, ballast water wastes from tanker ships and other water transportation modes [3–5], and heavy metal from any coastal activities [6–9]. The primary source of heavy metals is from industry, mining, household, and agricultural wastes [10–12]. The increased heavy metal and other pollutant concentrations will directly affect the biota’s survival ability [13]. It can accumulate within the body of marine biota and in the sediment [6, 13, 14, 16–20]. An integrated environmental study is crucial because water pollution potency is very alarming in the Rupat Strait and may be controlled by tidal currentsinduced transport mechanisms. Furthermore, the influence of oceanographic factors on triggering pollutant distribution within the strait is not apparent yet, and this aspect should be investigated. Therefore, this study aims to determine the water mass transport mechanisms and their influence on pollutant distribution within the Rupat Strait based on hydrodynamic modeling, spatial analysis approaches and the water quality results will be compared to the ASEAN countries index water quality assessment.

2 Materials and Methods 2.1 Study Site Rupat Strait is a semi-enclosed channel formed by Rupat Island and the eastern coast of Sumatra (Dumai City). Administratively, it belongs to Riau Province, Indonesia (Fig. 1a). Geographically, it is positioned between 101.2−101.8° East and 1.6−2.15° North. Malacca Strait borders this water area in the north and south, Dumai City in the west, and Rupat Island in the East. The most significant economic activity in the surrounding Dumai City is industrial sectors, such as the crude oil industry and oil and gas product distribution. Moreover, the port, fishery, and anthropogenic activities increase water pollution as the wastes

Modeling of Surface Current-Driven Water Pollution and ASEAN …

521

Fig. 1 Research location map, water quality sampling site (a); model mesh (b)

flow into the strait in the form of hydrocarbon (sulfide, ammonia, phenol, mercaptan, and other organic and acidic compounds) [13]. Loading and unloading activities in the port also frequently leave the liquid waste into the ocean, where oil spilling is undoubtedly avoided [21].

2.2 Hydrodynamic Model Simulation MIKE21 flow model FM with unstructured triangle mesh [22] was employed to determine the current surface patterns within the Rupat Strait. It was simulated from June 1, 2022 to December 28, 2022, representing two significant states of moon phase-induced tidal elevation (spring and neap tidal conditions) [23]. In the meshing stage (Fig. 1b), bathymetry and coastline data were provided by the bathymetry map of the Indonesian Navy, calibrated using in situ bathymetry survey result using single beam echo sounder Teledyne Odom collected in September 2020. A tidal model predicting the surface elevation fluctuation was used as boundary conditions of the model developed. The model setup is given in Table 1.

2.3 Water Sampling and Analysis The field sampling was conducted from August 29, 2022 to September 5, 2020. Nine sampling sites were chosen to represent the water condition of Rupat Strait (Fig. 1a). All areas were surveyed during the same tidal conditions (ebb tides), considering the land-sourced waste intake to the water. It commenced from the Northern gate of Rupat Strait and ended at the Southern waterway (Fig. 1). We considered the

522

K. Ondara and U. J. Wisha

Table 1 Flow model flexible mesh setup Parameter

Applied in the simulation

Time

Number time of steps 210 Time steps interval 3600 Start date 1/6/2020 01:00 to 28/12/2020 23:00

Mesh boundary

Indonesian navy bathymetry map calibrated by field measurement in 2020 with six open boundaries

Solution technique

Time integration: low order, fast algorithm Space discretization: low order, fast algorithm Minimum time step: 0.01 s Maximum time step: 3600 s Critical CFL number: 0.8

Flood and dry

Drying depth: 0.005 m Flooding depth: 0.05 m Wetting depth: 0.1 m

Initial water level

-0.37 m

Density

Density type: Barotropic Reference temperature: 30 °C Reference salinity: 28 PSU

Bed resistance

Manning number: 0.22 m1/3 /s

nature of Rupat Strait and the location of industry and settlement areas to choose the sampling site (Fig. 2). Water bodies were sampled at a depth of 1 m underneath the surface level using a Nansen bottle. The water sampled was then stored in a plastic bottle. It was then analyzed in the marine laboratory of Riau University. The analyzed parameters consisted of BOD5, ammonia, phosphate, nitrate, and heavy metals (nickel, zinc, lead, copper, and cadmium). The result was then compared to the standard quality established by ASEAN Marine Water Quality criteria (AMWQC), Indonesia, Thailand, the Philippines, and Malaysia [24, 25]. An in situ survey (Fig. 1a) using the TOA-DKK water quality checker was also done at the same site. All environmental data were then spatially analyzed using an inverse distance weighted spatial analysis tool to generate the distribution map of the pollutant within the study area. The distribution map was then overlaid with the hydrodynamic model result sampled with the same tidal conditions. In this case, overlaying the low spring tidal current pattern is feasible since the water quality survey was done during ebb tides.

Modeling of Surface Current-Driven Water Pollution and ASEAN …

523

Fig. 2 Tidal current patterns in the Rupat Strait during spring high tidal conditions (a); spring low tidal conditions (b); neap high tidal conditions (c); neap low tidal conditions (d)

3 Results and Discussion 3.1 Tidal Current Patterns in the Rupat Strait Based on tidal harmonic analysis, the semi-diurnal co-tidal showed a predomination in the study area where the M2 and S2 amplitudes were the highest among the significant tidal constituents (Table 2). The Formzahl was 0.34, showing that the mixed tide predominates the study area with prevailing semi-diurnal. Moreover, according to the least-square-based analysis, within 18.6 years of prediction, the highest elevation reaches 1.53 m above the mean sea level, and the lowest level is 1.46 m beneath the mean sea level. The average tidal range was 2.99 m. Table 2 Tidal harmonic analysis results Constituents M2 Amplitude (m)

0.77

Phase lag (°) 258.8

S2 0.36 82.8

N2 0.11 245.7

K2 0.11 112.4

K1

O1 0.14

−12.3

P1 0.25

−37.3

0.05 155.1

M4 0.00 15.1

MS4 0.00 118.6

524

K. Ondara and U. J. Wisha

Within a channel with semi-diurnal tide predomination, the phase lags of M2 and M4 determine the flood or ebb current dominance [26]. The value of 2g M2 − g M4 was between 90° and 270°, indicating that the study area has more extreme and shorter flood tides than the ebb currents (flood dominant) [27]. During the spring high tidal conditions (Fig. 2a), the current velocity in the Northern gate of Rupat Strait reached 0.82 m/s, where the northerly current flowed Southward, passing the Rupat channel. In contrast, the more robust southerly flow ranging from 0.5 to 1.05 m/s entered the strait via the Southern gate. The water mass was confluence in Mampu Island in the middle of the waterway, decreasing velocity and directional alteration [28]. The current speed in this area ranged from 0.15 to 0.49 m/s during spring flood tides. Unlike the spring high tidal phase, the ebb currents flowed in the opposite direction, getting out of the strait with a magnitude ranging from 0.07 to 0.15 m/s (Fig. 2b). During the neap phases, the current direction showed the same pattern as the spring one. However, the velocity decreased significantly, ranging from 0.04 to 0.32 m/s and 0.02 to 0.18 m/s during flood and ebb tides. Overall, the current motions are more robust during the spring phase of tides, influencing the transport mechanism throughout the strait [30].

3.2 The Spatial Distribution of Environmental Parameters The influence of hydrodynamics within the Rupat Strait on distributing the pollutant intake is shown in Fig. 3. Overall, the current tidal regimes did not solely control the distribution and accumulation of pollutants. Except for the total suspended solid (TSS) parameter, the highest concentration was identified in the middle of the strait (around Mampu Island). Water temperature ranged from 30.6 to 31.4 °C with the highest value observed in the surrounding Mampu Island as the southerly and northerly tidal currents flowed toward the middle of the strait, bringing the higher (lower) temperature derived from the Northern (Southern) Malacca Strait (Fig. 4a). Higher TSS concentration was found in the Southern Rupat Strait, where the current motions were more robust in this area (>1 m/s). The TSS ranged from 90.3 to 145.9 mg/L (Fig. 3b). On the other hand, the biological oxygen demand (BOD5) ranged from 8.4 to 9.4 mg/L. Even though the deviation was not too significant, a higher concentration was observed in the surrounding Mampu Island as the current regimes transported the water mass toward the middle of the strait. The nutrient parameter distribution showed the same patterns for ammonia, nitrate, and phosphate (Fig. 3d−f), where the higher concentration was identified in the middle of the strait. The average N:P ratio was too high (197.6:1), showing that the nutrient pollution was alarming, resulting in imbalanced water conditions. The super high ammonia concentration could cause this state to be detected throughout the study area. According to [29], ammonia can be derived from the solution of organic nitrogen generated by microbes and fungi (ammonification) and from minerals entering water

Modeling of Surface Current-Driven Water Pollution and ASEAN …

525

Fig. 3 Spatial distribution of water quality parameters superimposed with tidal current pattern during sampling time; temperature a; TSS b; BOD5 c; ammonia d; nitrate e; and phosphate f

Fig. 4 Heavy metal concentration in the Rupat Strait (Cd and Ni × 10−1 ). The black lines denote the standard deviation of every data

bodies through soil erosion. However, industrial and anthropogenic wastes also increase the intake of nutrients within the Rupat Strait. Concerning marine pollutants, the most dangerous pollutant is heavy metals. Since the concentration of heavy metals is difficult to determine in the water bodies, we only identified several dominant heavy metals detected in the study area, such as cadmium, copper, zinc, and Nickel (Fig. 4). All sampling stations identified the other heavy metal parameters (copper and zinc) except nickel and cadmium. Zinc

526

K. Ondara and U. J. Wisha

was detected at stations 1 and 3, while cadmium was only detected at stations 1, 3 and 4, situated in the Northern gate of Rupat Strait, ranging from 0.008 to 0.2 mg/L. Generally, a significant concentration of nickel and copper was observed in the Northern strait, with a magnitude from 0.04 to 0.39 mg/L. The remnant station located in the Southern gate of Rupat Strait, the lower concentration of nickel and cadmium tended to be uniform, ranging from 0.032 to 0.2 mg/L. Due to the nature of heavy metal tending to be settled in the sediment, its quantification in the water bodies could not be used for determining the influence of water motions in transporting the heavy metal compounds throughout the Rupat Strait [31]. However, the industrial wastes and shipping activities are more contributed to elevating heavy metal content in the semi-enclosed strait of Rupat. Table 3 gives the assessment of the environmental parameters toward several quality standards. Overall, almost all parameter concentrations exceeded the quality standards in the ASEAN region, primarily the TSS and nutrients. These results indicate that the Rupat Strait may be polluted since the elevated industrial, shipping, and anthropogenic wastes are undoubtedly avoided. However, long-term monitoring is necessary to control marine pollutions in the Rupat Strait. Table 3 Comparison of the environmental parameters to several quality standards Parameter

Average value

Quality standard

Pollution status

TSS (mg/L)

126.11

I = 80, M = 50, T = nd, P = 30, A = nd

Highly polluted

BOD5 (mg/L)

8.84

I = 20, M = nd, T = nd, P = nd, A = nd

Unpolluted

Ammonia (mg/L)

5.56

I = 0.3, M = 0.05, T = 0.07, P = nd, A Highly polluted = 0.07

Phosphate (mg/L)

5.64

I = 0.015, M = 0.075, T = 0.045, P = nd, A = 0.045

Highly polluted

Nitrate (mg/L)

0.86

I = 0.08, M = 0.06, T = 0.02, P = nd, A = 0.06

Highly polluted

Nickel (mg/L)

0.01

I = 0.05, M = 0.06, T = 0.02, P = nd, A = 0.06

Unpolluted

Zinc (mg/L)

0.29

I = 0.05, M = 0.05, T = 0.05, P = nd, A = 0.05

Highly polluted

Copper (mg/L)

0.22

I = 0.008, M = 0.003, T = 0.008, P = 0.02, A = 0.008

Highly polluted

Cadmium (mg/L)

0.004

I = 0.001, M = 0.002, T = 0.005, P = 0.01, A = 0.01

Slightly polluted

(Water quality standard for marine biota: I = in Indonesia, M = in Malaysia, T = in Thailand, P = in The Philippines, A = ASEAN Marine Water Quality Criteria (AMWQC), nd = no data)

Modeling of Surface Current-Driven Water Pollution and ASEAN …

527

4 Conclusion To conclude, the water motions within the Rupat channel are more dynamic during spring tidal phases, primarily at the Southern gate of the strait. It is categorized as a dominant flood channel with more intense flood currents. The hydrodynamic variability within the Rupat Strait does not solely control the distribution of the environmental parameters assessed in this study. The most impactful parameters were temperature, TSS, BOD5, and nutrients. Copper is the most found heavy metal in the Rupat Strait, where the concentration exceeds the standard quality. Based on the average concentrations of environmental parameters, Rupat Strait is categorized as a polluted water area caused by intensive wastewater from industrial and anthropogenic processes. It is recommended that regular monitoring is necessary to prevent the long-term impact of marine pollution. Acknowledgements We would like to thank The Research Institute for Coastal Resources (RICRV) for research funding in 2020 and those who have helped in completing the field survey and this article.

References 1. Putra, S: Analisis Laju Perubahan Garis Pantai Mengunakan Metode Linier Regression Rate (LRR) dan Metode End-Point Rate (EPR) (Studi Kasus di Pulau Rupat Kabupaten Bengkalis). Riau University (2014) 2. Sandhyavitri, A., Fatnanta, F., Ramadhan, H.R., Suprayogi, I.: Combination of a Coastal Vulnerability Index (CVI) and social economic approaches in prioritizing the development of Riau Coastlines. Indonesia. MATEC Web Conf. 276, 02006 (2019). https://doi.org/10.1051/matecc onf/201927602006 3. Thamrin, K., Anwar, A., Mubarak, M.: Pola Penyebaran Minyak di Perairan Selat Rupat Dumai, Riau. J. Ilmu Lingkung. 1(1) (2007) 4. Arabova, Z.M., Arabov, M.S., Prochorov, E.M., Saadati, P.A.: Problems and ways to reduce anthropogenic impact on waters of The World Ocean. In: Vestnik of Astrakhan State Technical University. Series: Marine Engineering and Technologies. pp. 41–47. (2019). https://doi.org/ 10.24143/2073-1574-2019-3-41-47 5. Tian, Y., Chen, J., Huang, J., Lu, S.: Research of ship antifouling paint pollutants analysis and treating technique. Chinese J. Environ. Eng. 11(2), 1164–1170 (2017). https://doi.org/10. 12030/j.cjee.201601127 6. Riani, E., Sudarso, Y., Cordova, M.R.: Heavy metals effect on unviable larvae of dicrotendipes simpsoni (diptera: Chironomidae), a case study from Saguling Dam Indonesia. AACL Bioflux 7(2), 76–84 (2014) 7. Riani, E.: Perubahan Iklim dan Kehidupan Akuatik (Dampak pada Bioakumulasi Bahan Berbahaya dan Beracun & Reproduksi). IPB Press, Bogor (2012) 8. Gavhane, S.K., Sapkale, J.B., Susware, N.K., Sapkale, S.J.: Impact of heavy metals in riverine and estuarine environment: a review. Res. J. Chem. Environ 25(5), 226–233 (2021) 9. Gao, L.: Microplastics aged in various environmental media exhibited strong sorption to heavy metals in seawater. Mar. Pollut. Bull. 169 (2021). https://doi.org/10.1016/j.marpolbul.2021. 112480

528

K. Ondara and U. J. Wisha

10. Safitri, N.A., Hamidy, R.: Konsentrasi Logam Berat (Cd dan Pb) pada Sedimen Permukaan Perairan Teluk Bayur Provinsi Sumatera Barat Indonesia. J. Environ. Sci. 2(3), 85–94 (2009) 11. Bartzas, G., Tsakiridis, P.E., Komnitsas, K.: Nickel industry: Heavy metal(loid)s contamination—sources, environmental impacts, and recent advances on waste valorization. Current Opinion in Environ. Sci. Health 21 (2021). https://doi.org/10.1016/j.coesh.2021.100253 12. Luo, P.: Heavy metals in water and surface sediments of the Fenghe River Basin, China: assessment and source analysis. Water Sci. Technol. 84(10–11), 3072–3090 (2021). https:// doi.org/10.2166/wst.2021.335 13. Syahminan, E., Riani, S., Rifardi, A.: Telaahan Logam Berat Pb dan Cd pada Sedimen di Perairan Barat Laut Dumai—Riau. J. Pengelolaan Sumberd. Alam dan Lingkung 5(2), 133–140 (2015) 14. Mulyani, S. Lani T.I.G., Sujana, A.E.: Identifikasi Cemaran Logam Pb dan Cd Pada Kangkung Yang Ditanam di Daerah Kota Denpasar. Bumi Lestari J. Environ. 12(2) (2012) 15. Çelik, A.: Heavy metal monitoring around the nesting environment of green sea turtles in Turkey. Water. Air. Soil Pollut. 169(1–4), 67–79 (2006). https://doi.org/10.1007/s11270-0061562-0 16. Stankovi´c, S., Jovi´c, M., Mihajlovi´c, M.L.D., Joksimovi´c, J., Tanaskovski, B.: Metal pollution determined by pollution indices for sea grass P. oceanica and surface sediments. Arch. Biol. Sci. 67(1), 91–101 (2015). https://doi.org/10.2298/ABS140410010S 17. Ali, R., Bream A.: The effects of sewage discharge on the marine Gastropod Gibbula sp., collected from the coast of Al-Hanyaa, Libya. Egypt. Acad. J. Biol. Sci. B. Zool. 2(2), 47–52, (2010). https://doi.org/10.21608/eajbsz.2010.15889 18. Marcovecchio, J.E., Moreno, V.J., Perez, A.: Determination of heavy metal concentrations in biota of Bahia Blanca. Argentina. Sci. Total Environ. 75(2–3), 181–190 (1988). https://doi.org/ 10.1016/0048-9697(88)90031-9 19. Gómez-Parra, A., Forja, J., DelValls, M.T., Sáenz, A.I., Riba, I.: Early contamination by heavy metals of the Guadalquivir estuary after the Aznalcollar mining spill (SW Spain). Mar. Pollut. Bull. 40(12), 1115–1123 (2000). https://doi.org/10.1016/S0025-326X(00)00065-5 20. Tack, F., Verloo, M.: Chemical speciation and fractionation in soil and sediment heavy metal analysis: a review. Int. J. Environ. Anal. Chem. 59(2–4), 225–238 (1995). https://doi.org/10. 1080/03067319508041330 21. Syahminan, S., Rifardi, R., Ruli, E.: Vertical content an analysis crude oil at core in Dumai coastal waters (2011) 22. Sutherland, G., Foreman, M., Garrett, C.: Tidal current energy assessment for Johnstone Strait, Vancouver Island. Proc. Inst. Mech. Eng. Part A J. Power Energy 221(2), 147–157 (2007). https://doi.org/10.1243/09576509JPE338 23. Brooks, D., Townsend, A.D.W.: Variability of the coastal current and nutrient pathways in the eastern Gulf of Maine. J. Mar. Res. 47(2), 303–321 (1989). https://doi.org/10.1357/002224089 785076299 24. ASEAN Secretariat: ASEAN Marine Water Quality Management Guidelines and Monitoring Manual. 1st ed. Jakarta, New Millenium Pty Ltd (2008) 25. Ministry of Natural Resources and Environment.: Malaysia Marine Water Quality Criteria and Standard. pp. 1. (2016), [Online]. Available: http://www.doe.gov.my/ 26. Wisha, U.J., Wijaya, Y.J., Hisaki, Y.: Real-time properties of hydraulic jump off a tidal bore, its generation and transport mechanisms: a case study of the Kampar River Estuary, Indonesia. Water 14 (2561), (2022). https://doi.org/10.3390/w14162561 27. Wisha, U.J., Wijaya, Y.J., Hisaki, Y.: Tidal bore generation and transport mechanism in the Rokan River Estuary, Indonesia: Hydro-oceanographic perspectives. Reg. Study. Mar. Sci. 52 (102309), (2022). https://doi.org/10.1016/j/rsma.2022.102309 28. Sutherland, G., Foreman, M., Garrett, C.: Tidal current energy assessment for Johnstone Strait, Vancouver Island. Proc. Inst. Mech. Eng. Part A J. Power Energy 221(2), 147–157, (2007). https://doi.org/10.1243/09576509JPE338 29. Peirong, Z., Wei, L.: Use of fluidized bed biofilter and immobilized Rhodopseudomonas palustris for ammonia removal and fish health maintenance in a recirculation aquaculture system. Aquac. Res. 44(3), 327–334 (2013). https://doi.org/10.1111/j.1365-2109.2011.03038.x

Modeling of Surface Current-Driven Water Pollution and ASEAN …

529

30. Ondara, K., Husrin, S.: Characteristic of breaking waves and analysis of sediment transport in Teluk Kendari. Jurnal Ilmu dan Teknologi Kelautan Tropis, (2018). https://doi.org/10.2944/ jitkt.v9i2.19293 31. Chakraborty, S., Chakraborty, P., Nath, B.N.: Lead distribution in coastal and estuarine sediments around India. Mar. Pollut. Bull. 97(1–2), 36–46 (2015). https://doi.org/10.1016/j.marpol bul.2015.05.056

Identification of Seasonal Water Mass Characteristics in West Sumatra Waters Gandhi Napitupulu , Ivonne M. Radjawane , Nabila Afifah Azuga, Khafid Rizki Pratama, Naffisa Adyan Fekranie, and Hansan Park

Abstract The eastern Indian Ocean is western tropical waters with very complex dynamic and physical processes affecting the area along the line of the West Sumatra coast. This study aims to analyze the water masses in the West Sumatra waters adjacent to the equatorial latitude based on the seasonality of December–February (DJF), March–May (MAM), June–August (JJA), and September–November (SON) from 2003 to 2018 which uses World Ocean Database (WOD) data. The tendency of the observed water mass characteristics to be dominant in the surface layer is more concentrated by the influence of water masses from the area around West Sumatra and intrusion from the surrounding waters. Meanwhile, in the middle to deep layers, more water mass patterns are observed from the western Indian Ocean region with a temperature range of −10 ˚C and salinity of 34.6−36.7 PSU. Characteristics based on seasonality are more indicative of a water mass mixing process. The DJF season tends to have a more uniform sea surface temperature than the JJA season. In addition, transitional seasons I and II (MAM and SON) tend to approach the main seasonal conditions. This is evidenced by the different Mixed Layer Depth (MLD) contours in the DJF season, which are thinner than in the JJA season. The temperature uniformity in the horizontal profile seems to vary in the JJA season compared to the DJF, so the thermocline observed in the JJA season is deeper. The identified water column density conditions show the same pattern seasonally. G. Napitupulu · N. A. Azuga · K. R. Pratama Department of Earth Sciences, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia I. M. Radjawane (B) Oceanography Research Group, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia e-mail: [email protected] N. A. Fekranie Department of Oceanography, Faculty of Earth Sciences and Technology, Institut Teknologi Bandung, Bandung, Indonesia I. M. Radjawane · H. Park Korea-Indonesia Marine Technology Cooperation Research Center, Bandung Institute of Technology, Bandung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_49

531

532

G. Napitupulu et al.

1 Introduction The western waters of Sumatra are a unique and complex area because their geographical location is crossed by the equator and flanked by the Asian continents–the Australian continent and the Indian Ocean–the Pacific Ocean [1], making these waters dynamic [2]. It is due to the significant influence of the monsoon system, Sumatra’s west coast, and Indonesian Crossflow (ARLINDO/Arus Lintas Indonesia), the Indian Ocean Dipole (IOD), and El-Nino Southern Oscillation (ENSO) phenomena on the dynamics of the water mass characteristics of Sumatra’s western waters [3]. The mass of water is formed from the composition of temperature, salinity, and density at a certain water depth [4]. Temperature, salinity, density, and pressure will develop stratification in the waters vertically, affecting the stability of the water mass [5]. Water mass identification is accomplished by analyzing temperature and salinity changes [6]. The water temperature is divided vertically into three layers: the mixed layer, the thermocline layer, and the inner layer. The mixed layer is formed by the stirred water mass caused by the movement of winds, currents, and tides on the sea surface, resulting in a homogeneous temperature value in the layer [7]. Seasonally, the east monsoon temperature is lower than the west monsoon temperature. It is inversely proportional to the salinity value, higher during the east monsoon and lowers during the west monsoon [6]. Seasonal winds, also known as monsoons, are seasonal winds that regularly blow due to differences in the intensity of land heating by the solar radiation system, causing differences in air mass pressure [8]. Winds carry masses of cold and moist air during the west monsoon, which occurs from October to March, causing rain in various affected locations. During the east monsoon, which arises from April to September, the winds carry dry air masses, resulting in a dry season in the affected areas [9]. Wind plays a vital role in moving and stirring the surface water mass [5]. However, wind movements do not affect the mixing and movement of water masses in deep waters (500–2000 m depth) because the speed of the current generated by the wind becomes zero at this depth, which is known as the depth without movement or the Ekman depth [10]. The eastern Indian Ocean water mass is dominated by two major streams that originate in the Subtropical Indian Ocean (SIC) and the Australasian Mediterranean Sea [4]. Another study identified five types of water masses in the Indian Ocean’s northeastern waters, including Bengal Bay Water (BBW), South Indian Central Water (SICW), Indian Equatorial Water (IEW), Subtropical Lower Water (SLW), and Northern Salinity Minimum Water (NSMW) [11]. The water mass in the Indian Ocean around western Australia consists of Antarctic Bottom Water (ABW), Lower Circumpolar Deep Water (LCDW), Upper Circumpolar Deep Water (UCDW), Antarctic Intermediate Water (AIW), Australasian Mediterranean Water (AMW), Subantarctic Mode Water (SMW), and South Indian Central Water (SCW) [12]. Another study suggested that the Equatorial Undercurrent (EUC) moves high salinity water masses to Indonesian waters [13]. The Arabian Sea High Salinity

Identification of Seasonal Water Mass Characteristics in West Sumatra …

533

Water (ASHSW) moves from the western Indian Ocean to the upper thermocline in the eastern Indian Ocean. Many studies have been conducted on water masse’s characteristics in the eastern Indian Ocean waters, but few have focused solely on the waters of western Sumatra. Therefore, this study aims to investigate the seasonal characteristics of water mass in the western waters of Sumatra and to identify the factors that influence mixing in the western waters of Sumatra when viewed from vertical and horizontal cross-sections of these waters.

2 Study Area and Methods The study area includes the western waters of Sumatra at coordinates of 0−10° south latitude and 90−100° east longitude (Fig. 1). The data used in this study was obtained from the World Ocean Database (WOD) and processed using Ocean Data View (ODV) software from 2003 to 2018. The data is available for download at https://www.ncei.noaa.gov/products/world-ocean-database. The salinity, temperature, and pressure (depth) data from 2003 to 2018 were utilized to determine the density and mass properties of the water. The water mass in four seasons: the western season (DJF), the transitional season I (MAM), the east season (JJA), and the transitional season II (SON), were analyzed separately by displaying a T-S diagram, vertical distribution, and horizontal distribution at 0−100 m and 0−1000 m (Table 1). Fig. 1 Map of the study area

534

G. Napitupulu et al.

Table 1 Characteristics of global water mass temperature and salinity (Source Emery, 2001) Layer

Indian Ocean

Pacific Ocean

Upper waters (0–500 m)

Bengal Bay Water (BBW) 25.0–29 °C, 28.0–35.0‰) Arabian Sea Water (ASW) (24.0–30.0 °C, 35.5–36.8‰) Indian Equatorial Water (IEW) (8.0–23.0 °C, 34.6–35.0‰) Indonesian Upper Water (IUW) (8.0–23.0 °C, 34.4–35.0‰) South Indian Central Water (SICW) (8.0–25.0 °C, 34.6–35.8‰)

Pacific Subarctic Upper Water (PSUW) (3.0–15.0 °C, 32.6–33.6‰) Western North Pacific Central Water (WNPCW) (10.0–22.0 °C, 34.2–35.2‰) Eastern North Pacific Central Water (ENPCW) (12.0–20.0 °C, 34.2–35.0‰) Eastern North Pacific Transition Water (ENPTW) (11.0–20.0 °C, 33.8–34.3‰) Pacific Equatorial Water (PEW) (7.0–23.0 °C, 34.5–36.0‰) Western South Pacific Central Water (WSPCW) (6.0–22.0 °C, 34.5–35.8‰) Eastern South Pacific Central Water (ESPCW) (8.0–24.0 °C, 34.4–36.4‰) Eastern South Pacific Transition Water (ESPTW) (14.0–20.0 °C, 34.6–35.2‰)

Intermediate waters (500-1500 m)

Antarctic Intermediate Water (AAIW) (2-10°C, 33.8–34.8‰) Indonesian Intermediate Water (IIW) (3.5–5.5°C, 34.6–34.7‰) Red Sea-Persian Gulf Intermediate Water (RSPGIW) (5-140 C, 34.8–35.4‰)

Pacific Subarctic Intermediate Water (PSIW) (5.0–12.0 °C, 33.8–34.3‰) California Intermediate Water (CIW) (10.0–12.0 °C, 33.9–34.4‰) Eastern South Pacific Intermediate Water (ESPIW) (10.0–12.0 °C, 34.0–34.4‰) Antarctic Intermediate Water (AAIW) (2–10 °C, 33.8–34.5‰)

Deep and abyssal waters (1500 m-bottom)

Circumpolar Deep Water (CDW) Circumpolar Deep Water (1.0–2.0°C, 34.62–34.73‰) (CDW) (0.1–2.0 °C, 34.62–34.73‰) Subantarctic Surface Water (SASW) (3.2–15.0 °C, 34.0–35.5‰)

Identification of Seasonal Water Mass Characteristics in West Sumatra …

535

3 Results and Discussion 3.1 Water Mass Characteristics The temperature-salinity (T-S) diagram shows a decrease in temperature from the surface (0 m) to a depth of about 5000 m. The characteristics of the water mass are arranged by the composition of temperature, salinity, and density [5]. The surface layer temperature in Sumatra’s western waters is 32 °C, with salinity values ranging from 32.5 to 32.8 PSU, and the density value is 22.5 kg/m3 . The temperature value decreases significantly as depth increases, with a minimum value of 3 °C at 5000 m. It is due to differences in the penetration of sunlight into the water, which causes the surface layer to be warmer than the deep layer [15]. Temperatures are higher at shallow depths and lower at deeper water layers [16]. Meanwhile, the salinity value increases as the depth increases, with a maximum salinity value of 34.8 PSU at a depth of 5000 m. At the surface layer, salinity tends to be lower due to the inclusion of freshwater sources in the area [17]. The density of water also increases along with depth. The maximum density value is 27.5 kg/m3 . The density of seawater varies depending on salinity and temperature. The density of water is lighter in the surface layer due to the freshwater influx. It gets heavier in deeper waters and becomes constant at 2000 m [10]. According to the T-S diagram analysis (Fig. 2), there are six types of water masses found in Sumatra’s western waters at coordinate positions 0−10° south latitude and 90°−100° east longitude. The water masses are Bengal Bay Water (BBW), Indonesian Upper Water (IUW), South Indian Central Water (SICW), Antarctic Intermediate Water (AAIW), Red SeaPersian Gulf Intermediate Water (RSPGIW), and Circumpolar Deep Water (CDW). BBW water masses dominate the surface layer (depth 0−120 m) with temperatures ranging from 25 to 29 °C and salinities ranging from 28 to 35 PSU. At a depth of 150−650 m, the mass of water is dominated by SICW, which has a temperature range of 8−25 °C and a salinity range of 34.6−35.8 PSU, followed by IUW, which has a temperature range of 8−23 °C and a salinity range of 34.4−35 PSU. At a depth of 650−1000 m, a mass of water is identified as the RSPGIW, with temperature values ranging from 5 to 14 C and salinity values ranging from 34.8 to 35.4 PSU. Another water mass that dominates at this layer is AAIW, which has a temperature range of 1−10 °C and a salinity range of 33.8−34.8 PSU. While at depths of 3000−5000 m, the dominating water is CDW, which has a temperature of 1−2 °C and a salinity of 34.62−34.73 PSU. The characteristics of the water mass in the western waters of Sumatra indicate intense mixing. According to the T-S diagram, the mixing is speculated to occur due to the influence of internal waves from tides and strong global currents. The Indian Ocean plays a significant role in the global current system, causing continuous mixing of water masses in the western area of Sumatra. One of the areas for the generation and propagation of internal solitary waves (ISW) is the southern part of the Andaman and Nicobar Islands, which border the northwest waters of Sumatra. It is caused by

536

G. Napitupulu et al.

Fig. 2 Characteristics of water masses

a strong barotropic semidiurnal tidal current from the Indian Ocean hitting the ridge off the coast of Breueh Island to the northwest. The internal wave formation along the ridge is primarily caused by barotropic tidal currents of the semidiurnal component (M2 ). The major axis of the M2 ellipse is oriented northeast (flood) and southwest (ebb) with a strong sea level elevation amplitude and magnitude that reach about 0.5 m and 5.0 m/s, respectively. During high tide, the magnitude of the barotropic tidal current from the Indian Ocean that cuts along the ridge dramatically increases, and the current direction is divided into several directions. Some of the current flows to the southeast as it enters the passage between Pulau Weh, Pulau Breueh, and Sumatra and then flows partly across the north side of Pulau Weh to the east before turning east or southeast. The strait between Nasi Island and mainland Sumatra also contributes to the tidal currents (Prasetya 2020).

3.2 Seasonal Variability of Water Mass Temperature Variations Figure 3 shows the variation of water mass for the four seasons. The temperature values at the surface and before the thermocline layer at low-latitude stations are higher than at high-latitude stations during the western season (Fig. 3a), as shown

Identification of Seasonal Water Mass Characteristics in West Sumatra …

537

by an isothermal line rising by 28 °C. The SST distribution is clear in the minimum range at higher latitudes during the western season. The high value of SST at lowlatitude stations is caused by the intensity of solar radiation, which tends to be below the equator during this period. The westerly monsoon winds bring cooler air from the continents of Asia and Europe that moves toward mainland Australia through the Northeast Indian Ocean. Strong mixing has also been observed to bring SICW water mass from India [11]. Similar to the pattern observed during the west monsoon period, intense mixing occurred, shown by mixing occurrence at a shallower depth and wider range of water mass column. It happened because the surface winds that blew during this period

Fig. 3 Variation of Water Mass Characteristics during a December–January–February (DJF) b March–April–May (MAM) c June–July–August (JJA) d September–October–November (SON)

538

G. Napitupulu et al.

were not as strong, and the direction was more random. As a result, the mixing strength of the surface layer is not as homogeneous. In contrast to previous periods, the SST value is generally higher toward stations with higher latitudes during the east monsoon period of June–August (JJA). Furthermore, during the transitional season II (SON), wind direction and currents are unstable and weak. The temperature difference between low and high-latitude stations is not particularly noticeable. The water mass is generally unstable beneath the thermocline layer (200−250 m). The influence of currents and topography causes water mass parcels to be distorted. Compared to other seasons, the east monsoon has the most instability due to strong mixing caused by tides in the western part of Sumatra [18]. Tides potentially generate mixing when an internal wave occurs [23]. In western Sumatra waters, turbulent flow is common and results in turbulent mixing that distorts the water mass (lower density water masses can be above higher density water masses). Salinity Variations The diagrams show that seasonal variation of the salinity profile is insignificant (the difference between seasons is only about 0.2−0.3 PSU). Salinity values in the surface layer are typically lower, ranging from 33 to 34.5 PSU. The highest surface salinity is identified during the east monsoon, when the western region of Sumatra receives the least rainfall, resulting in a reduced supply of fresh water entering the waters and, presumably, high salinity values. The salinity increases from 35 to 35.5 PSU at 200−500 m. Salinity values range from 34.5 to 34.8 PSU at 1000−3000 m.

3.3 Seasonal Variability of Horizontal Temperature and Salinity The temperature distribution in the longitude horizontal section (Fig. 4a) shows that the thermocline layer is deeper in the east than in the west. It indicates that the water mass was transported from the West Indian Ocean to the East Indian Ocean. According to Yoga and Karmen (2013), the surface temperature distribution in the Indonesian region ranges from 27.8 to 32 °C. Water temperatures, particularly in Indonesian waters, are influenced by the cycle of changing seasons. Aside from the season, the temperature of water mass is influenced by the sun’s intensity, depth, and the land around it [3]. The temperature profile in a seawater column reveals that the seawater column is divided into three main layers: the mixed surface layer, the thermocline layer, and the deep layer. To more precisely determine the three layers, the upper and lower thermocline limits are calculated using the assumption that a layer is considered a thermocline layer if the temperature gradient (100 m) is equal to or larger than 2 °C. However, it is not the case for West Sumatran waters. As shown in Fig. 4a, the decrease of thermocline depth in the eastern area indicates an upwelling process. The thermocline in longitude

Identification of Seasonal Water Mass Characteristics in West Sumatra …

539

Fig. 4 Horizontal cross-section of a temperature and b salinity in longitude for the depth of 500 m in West Sumatra waters

92 E occurred at 0−120 m, while it occurred at shallower than 100 m in longitude 98 E. Besides the upwelling caused by underwater currents, the bathymetric conditions also play a role in determining the thermocline depth in the region. Figure 4b shows salinity contours that vary with longitude. It shows that the salinity values are higher at longitudes 92−94 E than at longitudes 98 E, and therefore, the salinity pattern seems to move from west to east. Interestingly, salinity is almost uniformly in the 35 PSU range at depths of more than 100 m. It happened due to a mass transfer of water from the West Indian Ocean, an intrusion from the more saline western water mass (saline). The dominant water mass flow toward the east is at a depth of 80−150 m from latitude 1.2° south latitude to 1.5° north latitude. Meanwhile, the current direction is westward in the surface layer and at depths of 250−500 m. There is a current reversal between the Equatorial Under Current (EUC) and the current in the surface layer [13] (Fig 6). The wind is stronger during the east and transition seasons and lasts longer. As a result, the thermocline layer will be deeper. Strong winds will cause stronger currents and waves, stirring the water mass in the marine water column, causing the mixed layer to become deeper and the thermocline layer to descend [19]. Generally, as the wind speed above the water increases, the thermocline column becomes thinner, and the thermocline gets deeper. The results show that the temperature in the west is higher than in the east, indicating an upwelling process that occurs in the waters of West Sumatra. The characteristics of the water mass carried from the West Indian Ocean allow it to remain in West Sumatra’s waters. The West Indian Ocean has a broader thermal process than the

540

G. Napitupulu et al.

eastern region. This concept reveals that the temperature of the waters, particularly those in Indonesia, is strongly influenced by the cycle of changing seasons. Aside from the season, the temperature is also influenced by the intensity of the sun, depth, and land. The uniform temperature tends to exist at the 0−100 m layer for the western region and the 80 m for the eastern region. Meanwhile, water mass underneath the thermocline layer tends to vary toward depth. During the DJF season, the water mass distributed across the eastern Indian Ocean has low salinity. Meanwhile, the mass of water distributed in this area has higher salinity during the JJA season. It happens due to precipitation processes formed during the DJF period, such as the Madden Julian Oscillation (MJO) that causes rain to fall across Sumatra’s western region. As a result, the depth of the mixed layer may be shallower. Meanwhile, the wind transports water masses with high salt due to low precipitation along Sumatra’s western coast during the JJA season. The movement of water masses from the west to the east of the Indian Ocean becomes slow as the easterly wind strengthens and heads toward China’s mainland. It can be compared to Fig. 5, where varying temperatures tend to decrease at lower latitudes (8° south latitude) compared to higher latitudes (2° south latitude). Furthermore, the visible depth tends to decrease at higher latitudes with a thermocline depth range of 0−50 m. Figure 7a shows that in the west monsoon, the vertical contour of density varies from 22 to 24 kg/m3 and tends to occur at the 0−150 m thermocline. Density values are higher at depths of 200−500 m, ranging from 26 to 30 kg/m3 . Meanwhile, during the transitional season I (MAM), the density contours of 22−24 kg/m3 occurred at a depth of shallower than 100 m, and the density contours of 26−27 kg/m3 occurred at a depth of 250−500 m. During the east monsoon (JJA), the density ranging from 22 to 23 kg/m3 appeared at 0−120 m and the density of 26−27 kg/m3 at a depth of more than 200 m.

Fig. 5 Vertical cross-section of a temperature and b salinity in latitude for the depth of 500 m in West Sumatra waters

Identification of Seasonal Water Mass Characteristics in West Sumatra …

541

Fig. 6 Spatial plot for Mixed Layer Depth in a December–January–February (DJF) 2016 and b June–July–August (JJA) 2016 in West Sumatra waters

Fig. 7 Vertical cross-section of density (kg/m3) a December–January–February (DJF) b March– April–May (MAM) c June–July–August (JJA) d September–October–November (SON)

4 Conclusion The temperature declined from the surface layer to a depth of 5000 m, whereas the salinity increased with depth. The western waters of Sumatra have a maximum temperature of 32 °C and a minimum temperature of 3 °C, while the highest salinity value is 34.8 PSU, and the lowest value is 32.5 PSU. In general, the distribution pattern of SST during the west monsoon and transitional season I tend to be higher at low latitudes. Meanwhile, the SST value at high latitudes rises during the east monsoon. During transitional season II, however, the temperature difference between low and high-latitude locations is not very significant. The T-S diagram analysis reveals six different types of water masses in Sumatra’s western waters, including Bengal Bay Water (BBW), Indonesian Upper Water (IUW), South Indian Central Water (SICW), Antarctic Intermediate Water (AAIW), Red Sea-Persian Gulf Intermediate Water

542

G. Napitupulu et al.

(RSPGIW), and Circumpolar Deep Water (CDW). Horizontally, the temperature distribution in Sumatra’s western and the horizontal temperature profile in Sumatra’s western waters indicate the presence of an upwelling phenomenon, defined by a temperature differential of more than 2 °C/100 m of water depth. In the western half of the waterways, the salinity profile is higher. During the east monsoon and the transition season, the thermocline layer becomes deeper due to strong winds that trigger the mixing of water masses until the mixing layer. Meanwhile, the thermocline layer tends to be shallower during the west monsoon. The salinity value of the waters in the west monsoon tends to be lower than in the east monsoon. The density profile at a depth of 0−150 m ranges from 22 to 24 kg/m3 in the west season and transitional season I, 22 to 23 kg/m3 in the east season, and 21−22 kg/m3 in the transitional season II. Meanwhile, at depths more than 200 m, density levels during the west season range from 26 to 30 kg/m3 , whereas density values during transitional seasons I, east season, and transitional season II range from 26 to 27 kg/m3 . Acknowledgements The author would like to thank the Korea-Indonesia MTCRC (Marine Technology Cooperation Research Center) Scholarship for funding during the research. Thanks also go to the World Ocean Database (WOD) for providing and accessing the data used in this research.

References 1. Supriyadi, E., Hidayat, R.: Identification of upwelling area of the Western territorial waters of Indonesia from 2000 to 2017. Indonesian J. Geograp. 52(1), 105–111 (2020) 2. Napitu, R., Surbakti, H., Diansyah, G.: Identifikasi Karakteristik Massa Air Perairan Selat Bangka Bagian Selatan. Maspari J.: Marine Sci. Res. 8(2), 91–100 (2016) 3. Purba, M.: Dinamika perairan selatan Pulau Jawa-Pulau Sumbawa saat Muson Tenggara. Torani 17(2), 140–150 (2007) 4. Purba, N.P., Pranowo, W.S., Faizal, I., Adiwira, H.: Temperature-Salinity stratification in the Eastern Indian Ocean using argo float. In: IOP Conference Series: Earth and Environmental Science, June, vol. 162(1), pp. 012010. IOP Publishing (2018) 5. Al Tanto, T., Hartanto, T., Purba, M., Pranowo, W.S.: Karakteristik Massa Air di Perairan Barat Daya Pulau Sumba, Provinsi Nusa Tenggara Timur. Jurnal Kelautan Nasional 15(1), 23–36 (2020) 6. Siregar, S.N., Sari, L.P., Purba, N.P., Pranowo, W.S., Syamsuddin, M.L.: Pertukaran massa air di Laut Jawa terhadap periodisitas monsun dan Arlindo pada tahun 2015. Depik 6(1), 44–59 (2017) 7. Nofiyanti, K., Kunarso, K., DK.: A. R. D. T. Kajian Kedalaman Mixed Layer Dan Termoklin Kaitannya Dengan Monsun Di Perairan Selatan Pulau Jawa (Doctoral dissertation, Diponegoro University) (2020) 8. Pentury, R., Pietersz, J.H., Tuapattinaja, M.A., Pello, F.S., Huliselan, N.V., Hulopi, M., Tupan, C.I.: Potensi Komunitas Mangrove Pantai Tial Kabupaten Maluku Tengah. TRITON: Jurnal Manajemen Sumberdaya Perairan 16(2), 68–76 (2020) 9. Simanjuntak, P.P., Safril, A.: Analisa angin zonal dan meridional dalam menentukan awal musim hujan di Kota Jambi. Jurnal Teori dan Aplikasi Fisika 8(1), 43–50 (2020) 10. Rofiqa, Z., Muliadi, M., Risko, R.: Estimasi Potensi Tenaga Arus Laut Permukaan sebagai Pembangkit Listrik di Perairan Selatan Selat Makassar. Prisma Fisika 6(3) (2005) 11. Al Ayubi, M.A., Surbakti, H., Mbay, L.O.N.: Identifikasi Massa Air Di Perairan Timur Laut Samudera Hindia (Doctoral dissertation, Sriwijaya University) (2012)

Identification of Seasonal Water Mass Characteristics in West Sumatra …

543

12. Woo, M., Pattiaratchi, C., Feng, M.: Hydrography and water masses in the south-eastern Indian Ocean. SFRME Collaborative Research Project. School of Environmental Systems Engineering, Australia (2006) 13. Kusmanto, E., Siswanto, S.: Analisis Masa Air Dan Estimasi Transport Arus Bawah Ekuator Pada Bujur 90° E Selama Indonesia Prima 2017. Jurnal Meteorologi dan Geofisika 19(2), 59–69 (2019) 14. Emery, W.J.: Water types and water masses. Ocean circulation. Science 1556–1567 (2003) 15. Santoso, A.D.: Pemantauan hidrografi dan kualitas air di Teluk Hurun Lampung dan Teluk Jakarta. Jurnal Teknologi Lingkungan 6(3) (2011) 16. Sidabutar, E.A., Sartimbul, A., Handayani, M.: Distribusi suhu, salinitas dan oksigen terlarut terhadap kedalaman di Perairan Teluk Prigi Kabupaten Trenggalek. JFMR (Journal of Fisheries and Marine Research) 3(1), 46–52 (2019) 17. Kalangi, P.N., Mandagi, A., Masengi, K.W., Luasunaung, A., Pangalila, F.P., Iwata, M.: Sebaran suhu dan salinitas di Teluk Manado. Jurnal Perikanan dan Kelautan Tropis 9(2), 70–75 (2013) 18. Purwandana, A., Purba, M., Atmadipoera, A.S.: Distribusi Percampuran Turbulen di Perairan Selat Alor (Distribution of Turbulence Mixing in Alor Strait). ILMU KELAUTAN: Indonesian J. Marine Sci. 19(1), 43–54 (2014) 19. Hutabarat, M.F., Purba, N.P., Astuty, S., Syamsudin, M.L., Kuswardani, A.R.: Variabilitas lapisan termoklin terhadap kenaikan mixed layer depth (MLD) di Selat Makassar. Jurnal Perikanan Kelautan 9(1) (2018) 20. Prasetya, I.A.: Studi Gelombang Internal di Kawasan Perairan Pulau Weh, Aceh (2020) 21. Wyrtki, K.: Physical oceanography of the Southeast Asian waters (Vol. 2). University of California, Scripps Institution of Oceanography (1961) 22. Sulaiman, A.: Turbulensi Laut Banda. Jakarta, Badan Pengkajian dan Penerapan Teknologi (2000)

The Climate Comfort and Risk Assessment for Tourism in Bali, Indonesia Kadek Sumaja, I Kadek Mas Satriyabawa, Sindy Maharani, and Weny Anggi Mustika

Abstract Tourism is one of the main economic resources in Bali that is mostly affected by climate condition which tends to change. However, it is still unclear how climate change may affect tourism in tropical archipelagos like Bali Island. Moreover, climate comfort is one of the most important factors that must be considered to increase the number of tourists visiting because it may alter the pattern of tourist visits and preferences. Therefore, this study will analyze the comfort level of the tourism climate using the Holiday Climate Index (HCI) and humidity index (humidex) methods. Moreover, the climate comfort level in observation points is mapped to analyze the comfortable level for each month in Bali. In addition, HCI and humidex values will be combined with the number of foreign visitors to Bali to analyze the relationship between climate comfort indicators and tourist visits. The results of this study indicated that there were six quite comfortable months in the majority area of Bali, while the most comfortable month for traveling in Bali was in August and the least comfortable month was in December and January. The number of tourist arrivals to Bali had a relationship with the HCI and humidex values. The results of this study are expected to assist tourists, the tourism sector, and local governments in making pre-travel decisions based on climatic comfort information and tourism policy to enhance the Bali tourism business.

1 Introduction Tourism is one of the main economic resources in Bali. However, unfortunately, it is one of the economic sectors that is mostly affected by climate conditions which tends K. Sumaja (B) · I. K. M. Satriyabawa · S. Maharani I Gusti Ngurah Rai Meteorological Office, Indonesian Agency for Meteorology Climatology and Geophysics, Bali, Indonesia e-mail: [email protected] W. A. Mustika Regional III Office, Indonesian Agency for Meteorology Climatology and Geophysics, Bali, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_50

545

546

K. Sumaja et al.

to change. The most of tourist attractions in Bali are related to outdoor and social activities, such as sightseeing, adventure (trekking, rafting, surfing), and cultural tourism [1, 2]. Therefore, the condition of the climate parameters will affect most of the tourist preferences in Bali. However, the potential impact of climate change on tourism in tropical archipelagos like Bali Island is largely unknown. Climate conditions could affect the number of tourists visits as well as tourists alternatives and it is one of the key elements that must be considered to increase the number of tourists visiting [3, 4]. Moreover, climate comfort is one of the most important factors that must be considered into increasing the number of tourists visiting because it may alter the pattern of tourist visits and tourist preferences [5]. Thus, information about the climate conditions in a tourist destination is considered substantial. However, the abundance of information related to local climate will only confuse travelers. Consequently, simpler information is imperative to represent the climate comfort of a region. Some of the methods are the Holiday Climate Index (HCI) and the humidity index (humidex). The Holiday Climate Index (HCI) highlights the significance of climate as a factor in sustaining the competitiveness of coastal and urban destinations. HCI is affected by various climate parameters such as maximum and average temperature, relative humidity, cloud cover, precipitation, and wind speed. Also, it measures how hot we feel by combining the effects of perceived warm temperatures and humidity. HCI has a strong influence on tourist visits especially international tourists, because they need to adapt to climate conditions in destination area [6]. Several studies related to climate comfort that had been carried out in Bali and other regions of Indonesia consist of the Tourism Climate Index (TCI), Temperature Humidity Index (THI), and Holiday Climate Index (HCI). However, this study only used a few observation points and did not display monthly information neither numerically nor spatially, so it was less representative. The information on these two convenience indexes would aid tourists and related stakeholders to design their trips and to develop tourist areas. Thus, this study aims to analyze and map the comfort level of the tourism climate using the Holiday Climate Index (HCI) and humidity index (humidex). Moreover, finding the most comfortable month and the relationship between climate comfort indicators and tourist visits.

2 Data and Methods Data of this research consists of BMKG’s land observation data and ERA5 reanalysis data (https://cds.climate.copernicus.eu/) from 2001 to 2020 in 24 observation points (Fig. 1). This data comprises maximum and average temperature (°C), relative humidity (%), cloud cover (%), precipitation (mm/day), and wind speed (km/h). Furthermore, the number of tourists arriving in Bali is obtained from Bali Statistical Agency (https://bali.bps.go.id) from 2008 to 2019. The latest data was from 2019 because this research would like to analyze tourist visits in normal occurrences

The Climate Comfort and Risk Assessment for Tourism in Bali, Indonesia

547

before the COVID-19 pandemic [8]. Climate parameters will be calculated and rated according to the following formula and table [7, 9]. Firstly, thermal comfort (TC) was calculated using (1). Then TC as well as other climate parameters were rated according to Table 1. Furthermore, HCI urban and beach were calculated using the rated value. HCI beach comprised of observation points categorized as coastal with an elevation below 10 m, and the other observation points were calculated using HCI urban as shown in (2) and (3). Humidex was calculated using (4). Then, the HCI and humidex values were converted into HCI criteria and humidex criteria according to Table 2. Finally, those criteria were plotted into monthly climate comfort maps using Q-GIS software.

Table 1 Parameter rating Rating Thermal comfort (°C) Precipitation (mm/day) Wind speed (km/hr) Cloud cover (%) 10

23–25

0,00

1–9

11–20

9

20–22 or 26

< 3,00

10–19

1–10 or 21–30

8

27–28

3,00–5,99

0 or 20–29

0 or 31–40

7

18–19 or 29–30

−

−

41–50

6

15–17 or 31–32

−

30–39

51–60

5

11–14 or 33–34

6,00–8,99

−

61–70

4

7–10 or 35–36

−

−

71–80

3

0–6

−

40–49

81–90

2

37–39 or (−1)–(−5)

9,00–12,00

−

91–99

1

≤ -6

−

−

100

0

≥ 39

> 12,00

50–57

−

-1

−

> 25,00

−

−

-10

−

−

> 70

−

Table 2 HCI and humidex

Index

HCI

Index

Humidex

0–19

Dangerous

< 20

No discomfort

20–39

Unacceptable

20–29

Little discomfort

40–49

Marginal

30–39

Some discomfort

50–59

Acceptable

40–45

Great discomfort

60–69

Good

> 45

Dangerous

70–79

Very good

−

−

80–89

Excellent

−

−

90–100

Ideal

−

−

548

K. Sumaja et al.

Fig. 1 Research location

T C = 0.8T +

R H + Tx 500

(1)

HCI Urban = 2(T C) + 4( A) + 3(P) + (W )

(2)

HCI Beach = 4(T C) + 2(A) + 3(P) + (W )

(3)

5 Humidex = T + ∗ 9

   RH 7.5∗T ( ) 6.112∗10 237.7+T ∗ − 10 100

(4)

TC is thermal comfort (°C), Tx is the maximum temperature (°C), T is the average temperature (°C), RH is relative humidity (%), A is cloud cover (%), P is precipitation (mm/day), and W is wind speed (km/h). The association between the HCI or humidex and the number of visitors would be measured using the correlation coefficient (r) as shown in (5), where S i is the HCI or humidex, Gi is the tourist visit, σ S and σ G are their standard deviations, and n is the number of data pairs.

The Climate Comfort and Risk Assessment for Tourism in Bali, Indonesia

n r=

549

  − S) G i − G (n = 1)σs σG

i=1 (Si

(5)

3 Result and Analysis 3.1 Climate Parameters in Bali The results showed the variation of each climate parameter in Bali (Fig. 2). The average annual temperature ranged from 24.2 to 26.6 °C and the average maximum temperature in Bali ranged from 27.6 to 30.2 °C. The annual average temperature and the highest maximum temperature occurred in November and hit the lowest in August. Furthermore, the lowest and highest wind speed was blown in March and August, respectively, from 4.0 to 10.8 km/hr. The average daily rainfall in Bali was distributed from 1.2 to 11.3 mm/day, where January and February were the wettest months and August was the driest month in Bali. Humidity was classified from 80 to 86%, the highest was in January and the lowest was in September (Fig. 2). The average cloud cover in Bali varied between 54 and 88%, where the most of clouds covered the sky in December, January, and February. Therefore, December, January, and February were the most humid months, with the highest rainfall, the hottest, and the most cloud cover, whereas in July and August, the opposite condition occurred, with fewer clouds, the lowest temperature, lower humidity, less rainfall, and stronger wind speed. This result was consistent with earlier studies that showed temperature, rain, and cloud cover were the main climate factors that affect how comfortable the climate was at the destination [6, 7, 10–12]. 30 90% 20

70%

10

50%

0

30% JAN

FEB

MAR

T-ave (C) Rain (mm/day)

APR

MAY

JUN

JUL

AUG

SEP

OCT

T-max (C) RH (%)

Fig. 2 Climate parameters in Bali during 20 years periods (2001–2020)

NOV

DEC

Wind (km/h) Cloud (%)

550

K. Sumaja et al.

3.2 Holiday Climate Index (HCI) and Humidity Index (Humidex) The Holiday Climate Index (HCI) in Bali differed from an ‘unacceptable’ to an ‘excellent’ level as shown in Fig. 3. May to September was the month with the highest HCI value in Bali, where the range of HCI values in all regions was ‘very good’ and ‘excellent’. Furthermore, the comfort level increased in July and August, when the most areas of Bali were in the ‘excellent’ category. Then, in June and September, the Bali area was dominated by the ‘very good’ HCI category, with a small percentage of ‘excellent’ and ‘good’ ratings appearing in May. The period with the next highest HCI category was October, November, and April. In October, most areas of Bali were in the very ‘good’ category, and there were still ‘excellent’ ones. Still, in some areas, they began to become ‘marginal’ and ‘acceptable’ categories in central Bali and the southwestern coast of Bali. Then in November, most of region in Bali was categorized as HCI ‘good’, with some ‘marginal’ and ‘acceptable’ and only slightly ‘excellent’ in the highlands area. In April, the Bali region was divided into HCI ‘good’ and ‘very good’ ratings with almost equal areas. December to March was the months with the lowest HCI values. In March, Bali province was dominated by ‘good’ and ‘acceptable’ criteria, followed by a few ‘marginal’ areas, and began to appear ‘unacceptable’ on the north coast of Bali. Then, December, January, and February were the most uncomfortable months in Bali because most areas of Bali were ranked as ‘marginal’. In February, almost a third of the Bali area was categorized as marginal. Although in December, there were several areas at the tip of West Bali and Nusa Penida Island which was measured as ‘good’ category, the ‘marginal’ category areas also began to expand. Even, the ‘unacceptable’ class started to appear in the southern Bali region. In January, the marginal areas were almost the same as in December. Still, the ‘unacceptable’ areas had expanded and covered the main tourist areas on the island of Bali, such as Denpasar, Kuta, Jimbaran, and Nusa Dua. This outcome was following earlier studies, which showed that these monthly climate comfort levels changed and peaked throughout the dry season [6, 10, 13]. The humidex category ranges from great ‘discomfort’ to little ‘discomfort’. The lower the humidex value means the more comfortable the environment. July to October was the months when Bali Island had a ‘little discomfort’ category that covered the highlands in northeast Bali. During July, August, and September periods, this category had been spreading to the northeast coast of Bali. However, from January to June, the humidex value in Bali was less comfortable, where most of this island was in the category of ‘some discomfort’, with only a few areas of a ‘little discomfort’. November and December were the most uncomfortable months, because apart from most areas of Bali being ranked as ‘some discomfort’ and a ‘little discomfort’, several ‘discomfort’ areas had emerged. ‘Great discomfort’ areas were widely spread on the southern and northern coasts of the western part of Bali.

The Climate Comfort and Risk Assessment for Tourism in Bali, Indonesia

551

Jan

Feb

Mar

Apr

May

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Fig. 3 Holiday climate index (HCI) and humidity index (humidex)

K. Sumaja et al.

380

36

380

34

340

32

300

30

260

420

DEC

NOV

SEP

OCT

JUL

AUG

Visitor

JUN

DEC

NOV

SEP

OCT

JUL

(a)

AUG

JUN

APR

MAY

260

FEB

40

MAR

300 JAN

50

APR

340

60

HUMIDEX

MAY

70

38

FEB

80

420

MAR

Visitor

JAN

HCI

thousands

90

thousands

552

(b)

Fig. 4 Relationship between HCI and humidex with international visitors

3.3 Relation Between HCI, Humidex, and International Visitor The influence of the value of climate comfort on international tourist visits can be seen from the correlation value between the two elements (Fig. 4). The correlation between HCI and Humidex with international visitors is 0.76 and −0.86. It shows that HCI had a very strong positive relationship and humidex had a very strong negative relationship with international tourist arrivals. Thus, when the HCI value is low, tourist visits were also low and contrarily. As regards humidex, the number of tourist visits increased as the value of the humidex decreased. Since the higher the humidex value, the more uncomfortable the conditions in the area. Thus, foreign tourists who are dominated by visitors from high latitudes country would feel uncomfortable [14]. This result was relevant to the previous research that states climate comfort information such as HCI and humidex was one of the crucial factors that should take into account to boost the number of visitors [7, 9, 15]. For future consideration, this result will help the tourism sector and local governments in making pre-travel decisions based on climatic comfort information and tourism policy to enhance the Bali tourism business. Moreover, this can be the basis for further research on the mitigation plan to overcome extreme weather or climate impact on tourism.

4 Conclusion Based on the results and analysis, the HCI rating in Bali was ‘good’ to ‘excellent’ from June to September, it was ‘unacceptable’ to ‘good’ from December to March and the other months indicate a transitional HCI condition. Moreover, the humidex category of ‘little’ to ‘some discomfort’ occurred for most months, except November and December which have ‘great discomfort’ classes. The HCI rating and humidex showed that the most comfortable months for traveling in Bali were June, July, and August. The number of tourist arrivals to Bali was directly proportional to HCI and

The Climate Comfort and Risk Assessment for Tourism in Bali, Indonesia

553

inversely proportional to humidex, where the correlation value of these two indices was very strong.

References 1. Bali, D.P.P.: Daya Tarik Wisata Candi. Dinas Pariwisata Provinsi Bali (2018). https://disparda. baliprov.go.id/ Accessed 27 Sep 2022 2. Bali, D.P.P.: Analisa Pasar dan Indeks Kepuasan Wisatawan Mancanegara. Dinas Pariwisata Provinsi Bali, Denpasar (2019) 3. Susanto, J., Zheng, X., Liu, Y., Wang, C.: The impacts of climate variables and climate-related extreme events on island country’s tourism: evidence from Indonesia. J. Clean. Prod. 276, 124204 (2020). https://doi.org/10.1016/j.jclepro.2020.124204 4. Craig, C.A., Feng, S.: A temporal and spatial analysis of climate change, weather events, and tourism businesses. Tour. Manag. 67, 351–361 (2018). https://doi.org/10.1016/j.tourman.2018. 02.013 5. Scott, D., Rutty, M., Amelung, B., Tang, M.: An inter-comparison of the holiday climate index (HCI) and the tourism climate index (TCI) in Europe. Atmosphere (Basel) 7(6), 23–28 (2016). https://doi.org/10.3390/atmos7060080 6. Dini, L.R.Z., Sobirin.:Tingkat Kenyamanan Iklim Di Pulau Bali Berdasarkan tourism climate index. In: 8th Industrial Research Work National Seminar, pp. 678–684. (2017) 7. Hasanah, N.A.I., Maryetnowati, D., Edelweis, F.N., Indriyani, F., Nugrahayu, Q.: The climate comfort assessment for tourism purposes in Borobudur Temple Indonesia. Heliyon 6(12), e05828 (2020). https://doi.org/10.1016/j.heliyon.2020.e05828 8. Provinsi Bali, B.P.S.: Banyaknya Wisatawan Mancanegara Bulanan ke Bali Menurut Pintu Masuk (Orang) (2022). https://bali.bps.go.id Accessed 25 Sep 2022 9. Demiroglu, O.C., Saygili-Araci, F.S., Pacal, A., Hall, C.M., Kurnaz, M.L.: Future holiday climate index (HCI) performance of urban and beach destinations in the Mediterranean. Atmosphere (Basel) 11(9), 1–30 (2020). https://doi.org/10.3390/ATMOS11090911 10. Yazdanpanah, H., Barghi, H., Esmaili, A.: Effect of climate change impact on tourism: a study on climate comfort of Zayandehroud River route from 2014 to 2039. Tour. Manag. Perspect. 17, 82–89 (2016). https://doi.org/10.1016/j.tmp.2015.12.002 11. Scott, D., Gössling, S., De Freitas, C.R.: Preferred climates for tourism: case studies from Canada, New Zealand and Sweden. Clim. Res. 38(1), 61–73 (2008). https://doi.org/10.3354/ cr00774 12. Amelung, B., Nicholls, S.: Implications of climate change for tourism in Australia. Tour. Manag. 41, 228–244 (2014). https://doi.org/10.1016/j.tourman.2013.10.002 13. Scott, D., Mcboyle, G.: Using a ‘tourism climate index’to examine the implications of climate change for climate as a tourism resource. Tourism and Recreation. International Society of Biometeorolog, Porto Carras, pp. 69–88. (2002) 14. Provinsi Bali, B.P.S.: Banyaknya Wisatawan Mancanegara yang Datang ke Bali Menurut Kebangsaan, 2014–2020 (2022). https://bali.bps.go.id Accessed 25 Sept 2022 15. He, P., Qiu, Y., Wang, Y.D., Cobanoglu, C., Ciftci, O., Liu, Z.: Loss of profit in the hotel industry of the United States due to climate change. Environ. Res. Lett. 14(8) (2019). https:// doi.org/10.1088/1748-9326/ab2dce

Wind Effect on Spectral Ratio Analyses of Acoustic Waves Excited by Volcanic Explosion: Preliminary Result of Application at Sakurajima Volcano, Japan Mohammad Hasib, Takeshi Nishimura, Albertus Sulaiman, Titi Anggono, Syuhada, Febty Febriani, Cinantya Nirmala Dewi, Aditya Dwi Prasetio, and Trinugroho Abstract We investigate the wind effect on spectral ratio analyses of acoustic waves (infrasound) at Sakurajima volcano, Japan. Our preliminary result shows that spectral amplitude ratios of acoustic waves are different for different stations. This result implies that the propagation and site effects are not effectively removed. To clarify why the spectral ratio method does not work, we suppose that temporal variations of the wind in the atmosphere around the crater and volcano change the propagation paths of acoustic waves. We use assimilation wind data from European Centre for Medium-Range Weather Forecasts (ECMWF) to quantify the wind speed. Time ◦ ◦ resolution is 6 h, and spatial resolution is 0.125 × 0.125 . The acoustic signals are recorded by Japan Meteorological Agency (JMA). In the present study, we examine the spectral ratio of acoustic waves at Sakurajima volcano from January 2012 to December 2013 without and within data selection during low wind speed (< 3.3 m/s). Our results show that the spectral ratios of acoustic waves at different stations are similar when we select acoustic data during low wind speeds (< 3.3 m/s). This result indicates that the propagation paths of acoustic waves change due to wind conditions around the volcano.

M. Hasib (B) · T. Anggono · Syuhada · F. Febriani · C. N. Dewi · A. D. Prasetio · Trinugroho Research Center for Geological Disasters, National Agency of Research and Innovation (BRIN), Bandung, Indonesia e-mail: [email protected] T. Nishimura Department of Geophysics, Tohoku University, Sendai, Japan A. Sulaiman Research Center for Climate and Atmosphere, BRIN, Bandung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_51

555

556

M. Hasib et al.

1 Introduction Sakurajima volcano is one of the most active volcanoes in Japan and it is located on the island of Kyushu in southern Japan at the position of 31.58 N, 130.65 E. Sakurajima volcano generates explosion earthquakes every year with Vulcanian-type eruption [1]. Since the volcanic explosion occurs, the volcano effectively evokes not only seismic waves, but also the acoustic waves. In this study, we are only focused on the investigation of acoustic waves excited by volcanic explosion. Sakurajima volcano frequently exploded from January 2012 to December 2013 (see Fig. 1). The availability of the acoustic signal is satisfied. Application of the spectral ratio method on the volcanoes was first applied by Hasib et al. in [2]. They used the spectral ratio method on the seismic waves of explosion earthquakes and explained that the process can retrieve the source information of explosion earthquakes. The spectral ratio method retrieves the source information by eliminating the geometrical spreading, instrument response, and site effect [3, 4]. In this study, we apply the spectral ratio to acoustic waves. We examine the spectral ratio of the acoustic waves to clarify the source information of acoustic waves excited by the volcanic explosions. When a volcanic explosion occurs, the seismic and acoustic signals are evoked efficiently. Hence, the concept of spectral ratio on acoustic waves may be similar to the seismic waves. The difference between seismic and acoustic signals of the volcanic explosion is in the radiation medium. 450 Volcanic earthquake Explosion

400

350

Earthquake Account

300

250

200

150

100

50

r

be r

be

em

em ec

ov N

D 13 20

be

r

ob er

em

ct 13

20 13

O 20

ly

st 3

Se pt

Ju 13

Au gu

20

3 20 1

20 1

ay M

Ju ne

13

13

20

20

h

Ap ril 3

M 13 20

20 1

ar c

ry

ar y

ra br u

Fe

13

be r

Ja nu

13 20

20

be r

em

em

ec

ov

D

N

12 20

be r

ob er ct 12 20

em

O

pt Se 12

20 12

ly

st

Ju 12

Au gu

20

2 20 1 20

ay M

Ju ne

12

12

20

20

h ar c

Ap ril 2

M 12

20 1

y ra r 20

br u

Fe 12

20

20

12

Ja

nu ar

y

0

Fig. 1 Seismic activity at Sakurajima volcano from January 2012 to December 2013 recorded by Japan Meteorological Agency (JMA). The dark green bar indicates the volcanic earthquake. The red bar indicates the volcanic explosion

Wind Effect on Spectral Ratio Analyses of Acoustic Waves Excited …

557

Acoustic wave radiates via the atmosphere around the volcano. Since the investigation of the spectral ratio of acoustic waves in our analysis is about two years. We should carefully consider the wind condition around the Sakurajima volcano. The contamination of atmospheric condition may be affected the acoustic waves [5–7]. Therefore, we examine the results of the spectral ratio of acoustic waves within and without wind condition selection to clarify the wind effect on acoustic wave propagation. We marked that we only clarified the wind effect on the spectral ratio of acoustic waves based on observation data in this study.

2 Method We use the acoustic station is operated by Japan Meteorological Agency (JMA). The acoustic waves are measured by infrasonic microphones (ACO Ltd, Type 7144, 0.1–100 Hz). All the acoustic station is distributed at about 2.91–4.97 km away from the active crater (see Fig. 2). We apply the spectral ratio method in acoustic signal to understand the source information of volcanic explosion in more detail. However, the important remark is the applicability of the spectral ratio method on acoustic waves evoked by volcanic explosion since this condition is the first application on acoustic waves. First, we apply the spectral ratio without any data selection. We analyse 1164 acoustic signals excited by the volcanic explosion at Sakurajima volcano that occurred from January 2012 to December 2013. Second, we classify the acoustic event based on their amplitude into four classes (Class I, II, III, and IV) to examine the spectral ratio (see Fig. 3a). The smallest class (Class I) is the reference in our analysis. Third, we calculate the average of spectra in each class. Fourth, we take the ratio Class IV, III, and II to Class I. At last, we calculate the spectral ratio for different stations. The propagates of acoustic signal evoked by volcanic explosion may be disturbed by wind conditions around the crater (see Fig. 4). Hence, we calculate the spectral ratio into two states which are the calculation of acoustic signal without and within wind selection to understand the effect of wind on spectral ratio. For the analysis of acoustic signal during low wind speed (within wind selection), we select the acoustic event by assimilation wind data from European Centre for Medium-Range Weather Forecasts (ECMWF) to understand the temporal variation of wind conditions around the Sakurajima volcano during January 2012 to December 2013 [8, 9]. The assimilated wind data is obtained from reanalysis of the global atmosphere since ◦ ◦ 1979 with the time resolution of 6 h, and spatial resolution of 0.125 × 0.125 . The procedure of assimilated wind data extraction from the database can be explained as follows: first, we select and extract the wind data at different elevations corresponding to atmospheric pressure level (see Fig. 5). Second, we select the area calculation (around Sakurajima volcano) at a position longitude of 130, 62 E-130, 74 E, and latitude of 31, 54 N-31, 62 N, then extract the wind data. Third, we select the wind data starting from 31 m above

558

M. Hasib et al.

Showa crater

Fig. 2 Distribution of acoustic stations at Sakurajima volcano. The red triangle represents the active crater (Showa crater). The blue squares represent the acoustic station. The arrow indicates the Sakurajima Volcano location

Without wind selection

Within wind selection

Fig. 3 Acoustic event classification in our analysis from January 2012 to December 2013, a without wind selection, and b within wind selection (low wind speed). Double-sided arrow indicates the range of class in our analysis

sea level (minimum elevation) up to 800 m above sea level (Showa summit elevation). One example of the result of simulated wind extraction can be seen in Fig. 6. We characterize the wind condition based on the Beaufort wind scale [11]. At last, we select the acoustic events that occur during the low wind speed condition is characterized as a light breeze, in which wind speed is less than 3.3 m/s ( 50% will be found in the eastern part of Aceh, the central part of North Sumatra, most of West Sumatra except the northwest coast, almost all of Jambi except the eastern part, the northern part, and a few areas in the western part of South Sumatra, while for other areas are in the range of 20–50%. For the RCP8.5, the area that will experience an increase in productivity of > 50% on irrigated rice fields extends to almost most of Aceh (except the western part), North Sumatra, and West Sumatra. The area that will experience an increase of between 10 and 20% is wider on rainfed rice fields, especially in the province of Riau, Lampung, and small islands in the western part of Sumatra Island, especially for the RCP4.5 scenario. Rice productivity will increase by > 50% almost dominating, especially at RCP8.5. Different projection results will be found in the Riau Islands Province and a small area in the Bangka Belitung Province. At this location, most areas will be projected a decline in productivity of up to 50%, mainly because of a change toward a minimum temperature increase of 1 °C and a decrease of rainfall up to 200 mm month−1 . Projection of Rice Productivity in 2026–2035 on Irrigated and Rainfed Rice Fields in RS at Planting Time on October 20. The projection of rice productivity for the planting time on October 20, 2026–2035, will increase for the RCP4.5 and RCP8.5 scenarios. The increase in productivity with the RCP8.5 scenario will be lower than RCP4.5. The decrease in productivity on irrigated rice fields only will occur on Simeulue Island with a decrease of 0–10%. Rice productivity on rainfed rice fields will be projected to increase for RCP4.5 and RCP8.5. The increase in productivity for the RCP8.5 will be smaller than the change in productivity for the RCP4.5. In general, the increase in productivity for the RCP4.5 and RCP8.5 will reach more than 50% for most of Sumatra (Fig. 4).

Rice Projection in Sumatra by 2045 Regarding Climate Projection …

RS (30th Sept), irrigated, RCP4.5 in 2026-2035

RS (30th Sept), irrigated, RCP8.5 in 2026-2035

RS (30th Sept), rainfed, RCP4.5 in 2026-2035

RS (30th Sept), rainfed, RCP8.5 in 2026-2035

675

Fig. 3 Projection of changes in rice productivity on irrigated and rainfed rice fields in RS for the RCP4.5 and RCP8.5 scenarios at planting time on September 30 in the period of 2026–2035

Projection of Rice Productivity in 2026–2035 on Irrigated and Rainfed Rice Fields in DS at Planting Time on April 20. The projection of rice productivity in the period of 2026–2035 in DS at planting time on April 20 with the RCP4.5 scenario shows that most areas of Sumatra will experience an increase in productivity of 0– 30%, while for the RCP8.5 scenario, increase in productivity ranges from 0 to 20% (Fig. 5). The areas that experience a decline in yields of 0–20% will be wider, namely the Aceh region, the west coast (including Simeulue, Nias, Mentawai Islands), and the east coast (including Riau Islands, Bangka Belitung Islands) Sumatra Island. The highest productivity increase for the planting date of April 20 about 30–40% will be found in the scenario of RCP8.5 on rainfed rice fields located in the western part of South Sumatra and Lampung. Overall, the decline in production in rainfed rice fields will be more extensive than in irrigated rice fields, because of less rainfall and water availability.

676

E. R. Dewi et al.

RS (20th Oct), irrigated, RCP4.5 in 2026-2035

RS (20th Oct), irrigated, RCP8.5 in 2026-2035

RS (20th Oct), rainfed, RCP4.5 in 2026-2035

RS (20th Oct), rainfed, RCP8.5 in 2026-2035

Fig. 4 Projection of changes in rice productivity on irrigated and rainfed rice fields in RS for the RCP4.5 and RCP8.5 scenarios at planting time on October 20 in the period of 2026–2035

Projection of Rice Productivity in 2026–2035 on Irrigated and Rainfed Rice Fields in DS at Planting Time on May 10. Figure 6 shows that most of the west coast of Sumatra to the middle part is projected to increase in productivity on irrigated and rainfed rice fields in DS at planting time on May 10 for both RCP scenarios. Concurrently, rice productivity will be projected to decrease in the North of Aceh Province and the east coast. Areas experiencing decreased productivity will be wider at RCP8.5 compared to RCP4.5. Projection of Rice Productivity in 2036–2045 on Irrigated and Rainfed Rice Fields in RS at Planting Time on September 30. The increase in productivity will be found throughout Sumatra Island for both RCP4.5 and 8.5, on irrigated and rainfed rice fields, except in some areas in the Riau Islands Province which are projected to decline (Fig. 7). Overall changes in productivity will increase in 2036–2045 which are the same as in 2026–2035, starting from 10–20% to > 50%. Most changes from 30–40% to > 50%. For RCP4.5 irrigated rice fields, productivity will increase to > 50%. It will be found in the eastern part of Aceh, the central part of North Sumatra,

Rice Projection in Sumatra by 2045 Regarding Climate Projection … DS (20th Apr), irrigated, RCP4.5 in 2026-2035

DS (20th Apr), irrigated, RCP8.5 in 2026-2035

DS (20th Apr), rainfed, RCP4.5 in 2026-2035

DS (20th Apr), rainfed, RCP8.5 in 2026-2035

677

Fig. 5 Projection of changes in rice productivity on irrigated and rainfed rice fields in DS for the RCP4.5 and RCP8.5 scenarios at planting time on April 20 in the period of 2026–2035

most of West Sumatra except the northwest coast, almost all of Jambi except the eastern part, the northern part, and a few areas in the western part of South Sumatra. For RCP8.5, the increase in productivity will be broader in scope, spanning most of Aceh, North Sumatra, almost all of West Sumatra, the western part of Riau, almost all of Jambi and northern parts, and a few areas in the western part of South Sumatra. On rainfed rice, productivity increase will be almost the same as in irrigated rice fields. Productivity will increase by > 50% almost dominating Sumatra, especially for RCP8.5. A decline in productivity by up to 50% is projected to occur in the Riau Islands Province and a small part of the Bangka Belitung Province. Projection of Rice Productivity in 2036–2045 on Irrigated and Rainfed Rice Fields in RS at Planting Time on October 20. Changes in rice productivity with the planting date of October 20 for the period 2036–2045 will experience a significant increase in productivity reaching more than 50% for most areas of Sumatra both on irrigated and rainfed rice fields for RCP4.5 and RCP8.5. The decline in production

678

E. R. Dewi et al. DS (10th May), irrigated, RCP4.5 in 2026-2035

DS (10th May), irrigated, RCP8.5 in 2026-2035

DS (10th May), rainfed, RCP4.5 in 2026-2035

DS (10th May), rainfed, RCP8.5 in 2026-2035

Fig. 6 Projection of changes in rice productivity on irrigated and rainfed rice fields in DS for the RCP4.5 and RCP8.5 scenarios at planting time on May 10 in the period of 2026–2035

will be only seen on Simeulue Island for RCP4.5 on both irrigated and rainfed rice fields (Fig. 8). Projection of Rice Productivity in 2036–2045 on Irrigated and Rainfed Rice Fields in DS at Planting Time on April 20. Rice productivity in 2036–2045 at planting time on April 20 will experience an increase for both RCP scenarios on irrigated and rainfed rice fields (Fig. 9). A productivity increase of 0–20% is projected to occur in the central part of Sumatra Island from North Sumatra to South Sumatra. Meanwhile, a decrease in productivity between 0 and 20% will occur on the west coast of Sumatra, Aceh, and the east coast of Sumatra (Riau, Riau Islands, Bangka Belitung). On irrigated rice fields, the pattern formed will be not so different between scenarios RCP4.5 and RCP8.5, but on rainfed rice fields, RCP4.5 shows that the area experiencing a yield decrease of 0−10% will be wider than RCP8.5. The highest increase in productivity will occur in the RCP8.5 scenario of rainfed rice fields which is projected to increase by 30–40% located in the western part of South Sumatra and Lampung, but this increase only will occur in a small part of Sumatra. Overall, the

Rice Projection in Sumatra by 2045 Regarding Climate Projection … RS (30th Sept), irrigated, RCP4.5 in 2036-2045

RS (30th Sept), irrigated, RCP8.5 in 2036-2045

RS (30th Sept), rainfed, RCP4.5 in 2036-2045

RS (30th Sept), rainfed, RCP8.5 in 2036-2045

679

Fig. 7 Projection of changes in rice productivity on irrigated and rainfed rice fields in RS for the RCP4.5 and RCP8.5 scenarios at planting time on September 30 in the period of 2036–2045

area that will experience a decrease in yield will be more rainfed than irrigated rice fields. Projection of Rice Productivity in 2036–2045 on Irrigated and Rainfed Rice Fields in DS at Planting Time on May 10. Figure 10 shows the projection of rice productivity on irrigated and rainfed rice fields in DS with a planting time of May 10. For RCP4.5 in 2036–2045, rice productivity will be almost the same as RCP4.5 in 2026–2035. The increase in productivity will occur in most of the west coast of Sumatra to the central part, while in the North of Aceh Province and the east coast will show a decrease. Areas experiencing decreased productivity will increase under the RCP8.5 scenario. In DS, the results of the CSIROMK3.6 projection on both planting times show an increase in productivity, except for the planting date of May 10, both on irrigated and rainfed rice fields for the projections of 2036–2045, especially for RCP8.5. Generally, the increase in productivity in the 2026–2035 projection will be greater for the RCP8.5 scenario (Fig. 11). This condition is expected to occur because the planting time conditions are quite suitable for each phase of plant growth and the availability

680

E. R. Dewi et al.

RS (20th Oct), irrigated, RCP4.5 in 2036-2045

RS (20th Oct), irrigated, RCP8.5 in 2036-2045

RS (20th Oct), rainfed, RCP4.5 in 2036-2045

RS (20th Oct), rainfed, RCP8.5 in 2036-2045

Fig. 8 Projection of changes in rice productivity on irrigated and rainfed rice fields in RS for the RCP4.5 and RCP8.5 scenarios at planting time on October 20 in the period of 2036–2045

of water and other climatic elements, such as radiation intensity, temperature, and air humidity support optimal plant growth. Plant metabolism is directed by temperature, and optimum temperature can increase growth and prevent crop damage. Solar radiation drives crop production, while relative humidity regulates crop transpiration and water balance [13, 14]. On the other hand, in the years 2036–2045, the decline in productivity occurred due to the dynamics of climate elements that were not by plant growth and development. In RS, the results of the analysis using the CSIROMK3.6 model show that average productivity will increase, with the RCP8.5 scenario having an impact on increasing yields that will be greater than RCP4.5 (Fig. 12). Different scenarios and planting times have different impacts on average productivity depending on the climate projections of the model; this shows that the adaptation options made must be adapted to the climatic conditions in each region of Sumatra considering the spatially different impacts of climate change such as previously described.

Rice Projection in Sumatra by 2045 Regarding Climate Projection …

DS (20th Apr), irrigated, RCP4.5 in 2036-2045

DS (20th Apr), irrigated, RCP8.5 in 2036-2045

DS (20th Apr), rainfed, RCP4.5 in 2036-2045

DS (20th Apr), rainfed, RCP8.5 in 2036-2045

681

Fig. 9 Projection of changes in rice productivity on irrigated and rainfed rice fields in DS for the RCP4.5 and RCP8.5 scenarios at planting time on April 20 in the period of 2036–2045

682

E. R. Dewi et al.

DS (10th May), irrigated, RCP4.5 in 2036-2045

DS (10th May), irrigated, RCP8.5 in 2036-2045

DS (10th May), rainfed, RCP4.5 in 2036-2045

DS (10th May), rainfed, RCP8.5 in 2036-2045

Fig. 10 Projection of changes in rice productivity on irrigated and rainfed rice fields in DS for the RCP4.5 and RCP8.5 scenarios at planting time on May 10 in the period of 2036–2045

Rice Projection in Sumatra by 2045 Regarding Climate Projection …

683

Fig. 11 Changes in the average rice productivity in Sumatra Island on irrigated and rainfed rice fields in DS

Fig. 12 Changes in the average rice productivity in Sumatra Island on irrigated and rainfed rice fields in RS

684

E. R. Dewi et al.

4 Conclusions Different scenarios and planting times have different impacts on the average rice productivity depending on the climate projections of the model, and this shows that adaptation options must be adapted to the climatic conditions in each region. The projection of rice productivity with the CSIROMK3.6 model in RS at planting time on September 30 and October 20 shows that in RCP4.5 and RCP8.5 in 2026– 2035 and 2036–2045, both on rainfed and irrigated rice fields have increased by more than 0% to more than 50%. Meanwhile, the projection of rice productivity in DS with initial planting on April 20 and May 10 in 2026–2035 and 2036–2045 on rainfed and irrigated rice fields shows that most of Sumatra Island has increased by more than 0–30% at RCP4.5 and increased 0–20% on RCP8.5. The CSIROMK3.6 model has a high increase in productivity, and this is indicated by a large difference in the radiation intensity data, where the average value of the radiation intensity of the model is sufficiently high. The average value of the radiation intensity of the CSIROMK3.6 for RCP4.5 and RCP8.5 scenarios also looks higher than the historical average value. Adaptation options to maintain rice productivity by setting the planting time in RS and DS show differences in productivity. So, it is important to determine the right planting start for optimum productivity. To get the right adaptation options, it is necessary to run the model by changing the factors that affect productivity including the amount and irrigation system, fertilization, planting time, and rice variety.

References 1. Word Food Program.: How Climate Change Affects Hunger (2016). http://www.wfp.org/cli mate-change. Last Accessed 21 July 2016 2. Ministry of Agriculture.: Grand design pengembangan lumbung pangan beroreintasi eskpor di wilayah perbatasan, Indonesia Ministry of Agriculture, Jakarta (2017) 3. Surmaini, E., Faqih, A.: Kejadian Iklim Ekstrem dan Dampaknya Terhadap Pertanian Tanaman Pangan di Indonesia. Junal Sumberdaya Lahan 10(2), 115–128 (2016) 4. Sopaheluwakan, A., Satyaningsih, R., Linarka, A., Rizal, J., Mulsandi, A., dan Kadarsah.: Analisa Proyeksi Perubahan Iklim di Indonesia dengan Menggunakan Data Coordinated Regional Downscaling Experiment South East Asia (CORDEX-SEA). Cordex Meeting, short paper (2016) 5. World Meteorological Organization.: The Role of Climatological normal in a changing climate Technical Report WMO/TD-No. 1377; WCDMP-No.61 (2007). https://library.wmo.int/doc_ num.php?explnum_id=4546 6. World Meteorological Organization.: https://public.wmo.int/en/media/pressrelase/no-997-sci entist-urge-more-fequent-updates-of-30-year-climate-baselines-keep (2014) 7. World Meteorological Organization.: WMO Guidelines on the Calculation of Climate Normals Technical Report WMO-No.1203 (2017). https://library.wmo.int/doc_num.php?explnum_id= 4166 8. IPCC.: Climate Change 2014; Synthesis Report. Contribution of Working Groups I, II, and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change Technical

Rice Projection in Sumatra by 2045 Regarding Climate Projection …

9.

10.

11. 12.

13. 14.

685

Report. Intergovernmental Panel on Climate Change (2014). https://www.ipcc.ch.report.ar5. syr/ Keating, B.A., Carberry, P.S., Hammer, G.L., Probert, M.E., Robertson, M.J., Holzworth, D., Huth, N.I., Hargreaves, J.N.G., Meinke, H., Hochman, Z., McLean, G., Verburg, K., Snow, V., Dimes, J.P., Silburn, M., Wang, E., Brown, S., Bristow, K.L., Asseng, S., Chapman, S., McCown, R.L., Freebairn, D.M., Smith, C.J.: An overview of APSIM, a model designed for farming system simulation. Europ. J. Agronomy 18, 267–288 (2003) Susanti, E., Dewi, E.R., Surmaini, E., Sopaheluwakan, A., Linarko, A., Syahputra, M.R.: The projection of rice production in Java Island to support Indonesia as the world food granary. In: E3S Web of Conferences, vol. 306, pp. 01011. (2021) Probert, M.E., Carberry, P.S., McCown, R.L., Turpin, J.E.: Simulation of legume-cereal systems using APSIM Australian. J. Agricul. Res. 49, 317–327 (1998) Luo, Z., Wang, E., Fillery, I.R.P., Macdonald, L.M., Huth, N., Baldock, J.: Modelling soil carbon and nitrogen dynamics using measurable and conceptual soil organic matter pools in APSIM. Agric. Ecosyst. Environ. 186, 94–104 (2014) Hatfield, J.L., Prueger, J.H.: Temperature extremes: effect on plant growth and development. Weather and Climate Extremes 10, 4–10 (2015) Ferrante, A., Manani, L.: Agronomic management for enhancing plant tolerance to abiotic stresses: high and low values of temperature, light intensity, and relative humidity. Horticulturae 4, 21 (2018)

Multiyear La Niña Events and Poor Harvest of Sea Salt in Madura Island Rikha Bramawanto

and Suaydhi

Abstract Previous research stated that simultaneous La Niña and negative IOD caused sea salt harvest failure in Indonesia. In the last two decades, these happened in 2010 and 2016. It caused a scarcity of sea salt and a significant increase in salt prices. Currently, Indonesian salt farmers are upset because La Niña, which began in 2020, continues until now (mid-year 2022), becoming a multiyear La Niña. Multiyear La Niña events also occurred in 1973–1975 and 1998–2000. This article compares sea salt production and microclimate conditions in the current multiyear La Niña with the past multiyear La Niña. The results show that the three multiyear La Niña events affect the poor harvest of sea salt production in the Madura District. Understanding the character of multiyear La Niña can be an anticipatory effort for salt stakeholders to face multiyear La Niña in the future.

1 Introduction Indonesia’s position is between two oceans, the Indian and the Pacific Oceans. It makes Indonesia get global phenomena effect in the form of irregular oscillations of sea surface temperature in the Indian and Pacific Oceans, known as Indian Ocean Dipole (IOD), and El Niño Southern Oscillation (ENSO) [1, 2]. These two phenomena are closely related to the formation of rainfall in Indonesia [3–5]. The anomaly of negative sea surface temperatures in the Indian Ocean (negative IOD) and the Pacific Ocean (La Niña) makes the surface temperature of the waters in Indonesia warmer, triggering the formation of more rain than its normal conditions [6, 7]. Otherwise, the positive sea surface temperature anomaly in the Indian Ocean (positive IOD) and the Pacific Ocean (El Niño) triggers high rainfall in both oceans. This condition makes the weather in Indonesia drier than usual and can result wider drought [8, 9]. In addition, Indonesia occupies the equatorial region with a tropical R. Bramawanto (B) · Suaydhi Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_63

687

688

R. Bramawanto and Suaydhi

climate and only has two seasons, the rainy and the dry season. Indonesia is also affected by the Asian–Australian monsoon system [10, 11]. Climate variability in Indonesia associated with El Niño and La Niña events [12]. The climate variability affects the community’s livelihoods and many sectors in Indonesia, including health, agriculture, marine and fisheries, forestry, etc. IOD and ENSO affect inter-annual variability in the incidence of dengue hemorrhagic fever and malaria in all provinces in Indonesia [13]. The short rainy season during the 2015 El Niño reduced the main crops (rice, corn, and soybean) in Langkat, North Sumatera, Indonesia [14]. Analysis of historical data on El Niño events and rice harvests in Java and Bali shows the need for water storage, preparation of drought-resistant varieties, crop diversification, and early warning systems as adaptation strategies in rice farming in Indonesia [15]. There are differences in the characteristics of seasonal upwelling in the waters south of Java during the 2010 strong La Niña and super El Niño 2015 [16]. Coral geochemical records in Indonesia show the effects of ENSO and IOD on salinity and SST [17]. The intensity of El Niño has an impact on the severity of land and forest fires in fire-prone areas of provinces in Indonesia [18]. The Australian Bureau of Meteorology (BoM) and the World Meteorological Organization (WMO) have confirmed that the century’s first “triple dip” La Niña is ongoing [19, 20]. The closest “triple dip” La Niña events in the past occurred in 1973–1975 and 1998–2000 [21] (Fig. 1). “Triple dip” La Niña is a multiyear La Niña phenomenon that occurs in three consecutive years [22]. Experts claim that the “triple dip” La Niña has affected temperatures and rainfall patterns, triggering floods and droughts in various parts of the world [23]. The ongoing drought in southern South America and the Horn of Africa suggests a “triple dip” La Niña impact [24]. Several meteorological agencies in Asia and Australia warned of the flood risk caused by this phenomenon [25]. Currently, salt farmers in Indonesia are upset because the last La Niña sequentially continues from late 2020 to the present (mid-2022), which inhibits salt production [26]. The rain that persists throughout the dry season disrupts sea salt production. This abnormal condition is one of climate variability impacts on the availability of salt stocks in Indonesia. Previous studies have stated that strong La Niña coupled with

La Nina years -2 -1.5 -1 -0.5 1970 1972 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022

0

Fig. 1 La Niña years (−0.5 = weak, −1 = moderate, −1.5 = strong) and the appearance of the “triple” dip La Nina from 1970 to present. Source https://ggweather.com/enso/oni.htm (processed data)

Multiyear La Niña Events and Poor Harvest of Sea Salt in Madura Island

689

negative IOD can cause salt harvest failure in Indonesia [27]. These extreme conditions occurred in 2010 and 2016. The salt harvest failure in 2016 triggered a scarcity of salt [28]. It caused the price of salt increases abnormally [29]. Unfortunately, the salt policy in Indonesia did not consider the forecast of climatic variabilities in the following years to anticipate that kind of condition. This article compares sea salt production and microclimate conditions in the current multiyear La Niña with the past multiyear La Niña in Madura. This information can enrich the data to predict the arrival of multiyear La Niña in the next period, especially in Madura and its surroundings.

2 Data and Methods In this study, the monthly average total precipitation data is from the fifth-generation atmosphere reanalysis data of the European Centre for Medium-Range Weather Forecasts (ERA5 ECMWF) [30] with a horizontal resolution of 0.25° × 0.25°. The study area is the island of Madura and the surrounding waters at latitudes −6 to −8 0 S and longitudes 112 to 115 0 E. The climatology is taken as the period 1970–2021. The Nino 3.4 index and dipole mode index are from NOAA ESRL Physical Sciences Laboratory. The Nino 3.4 index and dipole mode index are based on SST dataset from the Hadley Centre Sea Ice and Sea Surface Temperature dataset (HadISST) spanning from 1990 to 2021 with a horizontal resolution of 1° latitude × 1° longitude [31]. Old salt production data (1990–2018) were obtained from PT Garam and Ministry of Marine Affairs and Fisheries (KKP). The latest salt production data (2019–2021) were taken from the annual report of PT Garam. The salt referred to in this study is salt made from sea water which is processed in ponds by the community and PT Garam. The data are processed using MS excel and ocean data view ODV 5.6.2. The monthly precipitation patterns for triple dip La Niña events in 1973– 1975, 1998–2000, and 2020–2022, at three salt farms in Madura Island (Sumenep, Pamekasan, and Sampang districts), were compared to each other. In this study, we illustrate monthly average total precipitation spatial distribution to determine the meteorological conditions during the triple dip La Niña events, the La Niña events combined with negative IOD and normal ENSO. Then an analysis of the characteristics of salt production was carried out in several La Niña events, whether they only occurred for one year, two years, or three years successively. June and October are critical points in the salt-making process. The salt harvest period will come late when the June rainfall is still high. The salt harvest period will be short when it starts to rain in October. To compare the critical point when occurs of multiyear La Niña, La Niña + negative IOD simultaneously, and ENSO normal, then June and October precipitation data are taken in 1993, 1999, 2010, 2014, 2016 and 2021 as a representation of the three conditions above. The rainfall in 1993 and 2014 associated with normal ENSO led to normal salt production. The 2010 and 2016 rainfall related to simultaneous La Niña and negative IOD events caused salt harvest failure. However, the rain in 1999 and 2021 associated with the La Niña Multiyear

690

R. Bramawanto and Suaydhi

event resulted in low salt production. To simplify the analysis, the monthly mean data of total precipitation is plotted in a spatial distribution.

3 Result and Discussion

Raw salt producon (tonnes)

The high rainfall over the last three years, especially during the dry season, has been causing a decline in salt production in Indonesia. The salt production in 2020 and 2021 is 1.36 and 1.09 million tons, while the salt harvest is estimated to be around 0.6 million tons or even lower in 2022. Indonesia’s salt production in the last three years was below the average from 2010–2022 (about 1.6 million tons). If the calculation does not include salt harvest failures in 2010 and 2016, then the salt production for the last three years will have a wider deviation (Fig. 2). Low salt production also occurred in 2011, 2013, and 2017. The low salt harvest in 2011 and 2017 occurred after crop failures in previous years. The salt production ratio between below average to above average was 8:5 in the last 12 years. To find out condition of salt production in the past during the triple dip La Niña, we need salt production data from 30 to 50 years back. It was hard to find valid data on salt production in Indonesia before 2010. Therefore, this study focuses on salt production in Madura Island, especially in salt ponds owned by PT Garam, in Sumenep, Pamekasan, and Sampang districts. We managed to obtain the best available data on salt production since 1990. The results of combining data on the occurrence of ENSO events, represented by a Nino3.4 index, with salt production in Madura mostly show concomitant patterns between La Nina events with decreased production and El Nino with increased salt production in the last 32 years. (Fig. 3). Of the fourteen La Niña (weak, moderate, Indonesia's raw salt producon 3,000,000 2,000,000 1,000,000 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 202 2022 Year Naonal raw salt producon Raw salt producon average from 2010 to 2021 Raw salt producon average (minus failure harvest years)

Fig. 2 Indonesia’s raw salt production from 2010 to 2021. Salt production in 2022 is the prediction by the Ministry of Marine Affairs and Fisheries [32]

691 500000 450000

2

400000

1.5

350000

1

300000

0.5

250000

0

200000

-0.5

150000

-1

100000

Nino index

3 2.5

-1.5

50000 0 DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ DJF ASO AMJ

-2

ton

Multiyear La Niña Events and Poor Harvest of Sea Salt in Madura Island

90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23

Nino 3.4 index

Year

PT Garam's raw salt producon

Fig. 3 PT Garam’s raw salt production from 1990 to 2022 in various ENSO conditions. A fullfledged El Niño or La Niña episode was classified after exceeding the threshold for at least five consecutive overlapping three-month seasons

strong) events since 1990, there were six times (42.9%) moderate salt harvests in 1995, 2005, 2007, 2008, 2017, and 2020, and four times (28.6%) little harvests in 1999, 2011, 2013 and 2021, two times (14.25%) very little harvest in 1998 and 2000, and two times (14.25%) crop failure in 2010 and 2016. The strong La Niña in 1999 and 2007 did not mean that salt production had dropped drastically. Crop failure and very little salt harvest can occur when the previous year there was a strong El Niño as happened in 1997–1998, 2009–2010, and 2015–2016. Strong La Niña both occurred at the start of the triple dip La Niña in 2020 and 1998, and there was a high salt production in the previous year (2019 and 1997). However, the salt production character of the triple dip La Niña 2020–2022 is different from that of the triple dip La Niña 1998–2000. The difference is strongly suspected to be caused by a phenomenon that occurred in the previous year. A weak El Niño in 2019 followed by a strong La Niña did not drastically reduce salt production in 2020. Meanwhile, a strong El Niño in 1997 followed by a strong La Niña made salt production decrease drastically in 1998. In Madura, the triple dip La Niña rainfall in 2020–2022 generally has a similar pattern to the “triple dip” La Niña in 1998–2000 and 1973–1975 (Fig. 4). The drought in the “triple dip” La Niña event lasts only about 2–3 months, so it can be said to be a wet dry season. Generally, the dry season in Madura is about 5–6 months. The driest conditions only occur in the months of June–July–August. The shortest drought in the “triple dip” La Niña event occurred in June 1998. The dry season that was not completely dry occurred in 1973. The highest rainfall occurred in 2020– 2022. The pattern of rainfall fluctuations in the Sumenep, Pamekasan, and Sampang districts is relatively the same, although Sampang rainfall is slightly higher than that in Pamekasan and Sumenep districts. June monthly average rainfall in the Madura waters, especially the northern part of the island, was still high when La Niña and negative IOD occurred in 2010 and 2016,

692

R. Bramawanto and Suaydhi

Fig. 4 Comparison of monthly average total precipitation in Sumenep, Pamekasan, and Sampang on three “triple dip” La Niña events in 1973–1975, 1998–2000, and 2020–2021

around 5–7.5 mm. During the multiyear La Niña (1999 and 2021) and normal ENSO (1993 and 2014), the monthly average rainfall in June is relatively the same, around 0–2.5 mm. The weather conditions in June 2021 and 1993 were wetter than in June 2021 and 2014 (Fig. 5a). The monthly average rainfall in October shows a significant difference. The island of Madura, especially the Sampang District, entered the rainy season with monthly average rainfall of 10–12.5 mm in October 2010 and 2016. The monthly average total precipitation on Madura Island was relatively small, 0– 2.5 mm, in October 1999 and 2021. Meanwhile, the general condition of the Madura Island was still dry in October 1993 and 2014 (Fig. 5b). This study focused on several La Niña events as factors that affect the decline of salt production. We get some interesting information after combining PT Garam’s salt production data with past La Niña events. Although not all La Niña events caused a decrease in production or failure to harvest salt in PT Garam’s ponds, salt stakeholders should be wary of any La Nina events, both single and multiyear La Nina events. There was 57.1% poor salt harvest during La Nina events in the last 32 years. We need to do further analysis involving other climatic variables such as IOD, Asian–Australian monsoon, and so on to identify La Nina species that can significantly reduce salt production, including whether it is within the Modoki criteria or not. Comprehensive knowledge of the influence of climate on raw salt production is very strategic, considering the high demand for salt, especially for industrial purposes, which reaches 3.7 million tons per year. In the future, the government must be able to provide the latest valid weather-climate information and accurate forecasts for salt farmers and import policymakers. On the other hand, the opportunity to utilize the supporting resources for salt production is still very potential, both extensively and intensively.

Multiyear La Niña Events and Poor Harvest of Sea Salt in Madura Island

693

Fig. 5 Comparison of monthly average rainfall spatial distribution in June and October at the time of La Niña + negative IOD, multiyear La Niña, and ENSO neutral events

4 Conclusion The multiyear La Niña has had an impact on the decline in salt production in Indonesia. Multiyear La Niña did not result in salt harvest failure as in 2010 and 2016, but the harvest was below the normal average. Experience and understanding of the character of multiyear La Niña can be an anticipatory step for salt stakeholders to predict the arrival of multiyear La Niña in the next period and face multiyear La Niña in the future.

694

R. Bramawanto and Suaydhi

References 1. Wieners, C.E., Dijkstra, H.A., de Ruijter, W.P.M.: The interaction between the Western Indian Ocean and ENSO in CESM. Clim. Dyn. 52, 5153–5172 (2019) 2. Dewi, Y.W., Wirasatriya, A., Sugianto, D.N., Helmi, M., Marwoto, J., Maslukah, L.: Effect of ENSO and IOD on the variability of Sea surface temperature (SST) in Java Sea. In: IOP Conference Series: Earth Environ 530. The 5th International Conference on Tropical and Coastal Region Eco Development 17–18 September 2019, Semarang, Indonesia (2019) 3. Nur’utami, M.N., Hidayat, R.: Influences of IOD and ENSO to Indonesian rainfall variability: role of atmosphere-ocean interaction in the Indo-Pacific sector. Proc. Environ. Sci. 33, 196–203 (2016) 4. Saputra, A., Hermawan, E., Darmawan, D.: Model development of Niño 3.4 and Indian Ocean Dipole (IOD) Anomalies teleconnection. J. Sains Dasar 7(2), 111–116 (2018) 5. Narulita, I.: ENSO and IOD impact to rainfall variability on Cerucuk watershed, Belitung Island (In Indonesian). Jurnal Tanah dan Iklim 41(1), 45–60 (2017) 6. Lim, E.P., Hendon, H.H.: Causes and predictability of the negative Indian Ocean Dipole and its impact on La Niña during 2016. Sci. Rep. 7, 12619 (2017) 7. Lestari, D., Sutriyono, E., Kadir, S., Iskandar, I.: Impact of 2016 Weak La Niña Modoki event over the Indonesian Region. Int. J. GEOMATE. 17, 156–162 (2019) 8. Avia, L., Sofiati, I.: Analysis of El Niño and IOD phenomenon 2015/2016 and their impact on rainfall variability in Indonesia. In: IOP Conference Series: Earth and Environmental Science, vol. 166, pp. 012034. (2018) 9. Lestari, D.O., Sutriyono, E., Sabaruddin, Iskandar, I. : Respective influences of Indian Ocean Dipole and El Niño-Southern oscillation on Indonesian precipitation. J. Math. Fund. Sci. 50(3), 257–272 (2018) 10. Wheeler, M.C., McBride, J.L.: Australian-Indonesian monsoon. In: Intraseasonal Variability in the Atmosphere-Ocean Climate System. Springer Praxis Books. Springer, Berlin, Heidelberg (2005) 11. Duan, Y., Liu, H., Yu, W., Liu, L., Yang, G., Liu, B.: The onset of the Indonesian-Australian summer monsoon triggered by the first-branch Eastward-propagating Madden–Julian Oscillation. J. Clim. 32(17), 5453–5470 (2019) 12. Xie, S., Annamalai, H., Schott, F.A., McCreary, J.P., Jr.: Structure and mechanisms of South Indian Ocean climate variability. J. Clim. 15(8), 864–878 (2002) 13. Kurnianingsih, Wirasatriya, A., Lazuardi, L., Kubota N., Ng, N.: IOD and ENSO-related time series variability and forecasting of Dengue and Malaria incidence in Indonesia. In: 2020 International Symposium on Community-centric Systems (CcS), pp. 1–8. IEEE (2020) 14. Ramija, K.E., Sudrajat, A., Irwandi, H., Ariantono, J.Y.: Influence of El Niño 2015/2016 on climate variability and production of main crops in Langkat regency. Agromet 35(2), 98–107 (2021) 15. Naylor, R.L., Battisti, D.S., Vimont, D.J., Falcon, W.P., Burke, M.B.: Assessing risks of climate variability and climate change for Indonesian rice agriculture. PNAS 14(19), 7752–7757 (2007) 16. Atmadipoera, A.S., Jasmine, A.S., Mulia, P., Kuswardani, A.R.T.D.: Upwelling characteristics in the Southern Java waters during strong La Niña 2010 and super El Niño 2015. Jurnal Ilmu Dan Teknologi Kelautan Tropis 12(1), 257–276 (2020) 17. Cahyarini, S.Y., Henrizan, M.: Coral based-ENSO/IOD related climate variability in Indonesia: a review. In: IOP Conference Series: Earth and Environmental Science, Global Colloquium on GeoSciences and Engineering, 18–19 October 2017, Bandung, Indonesia, vol. 118 (2017) 18. Nurdiati, S., Sopaheluwakan, A., Septiawan, P.: Spatial and temporal analysis of El Niño impact on land and forest fire in Kalimantan and Sumatra. Agromet 35(1), 1–10 (2021) 19. BBC Homepage: https://www.bbc.com/weather/features/62890361. Last Accessed 21 Sep 2022 20. WMO Homepage: https://public.wmo.int/en/media/press-release/wmo-predicts-first-%E2% 80%9Ctriple-dip%E2%80%9D-la-ni%C3%B1a-of-century. Last Accessed 21 Sep 2022

Multiyear La Niña Events and Poor Harvest of Sea Salt in Madura Island

695

21. Climateextremes Homepage: https://climateextremes.org.au/wp-content/uploads/Multi-yearLa-Nina-Events-ARC-Centre-of-Excellence-for-Climate-Extremes.pdf. Last Accessed 21 Sep 2022 22. WMO Homepage: https://public.wmo.int/en/our-mandate/climate/el-ni%C3%B1ola-ni%C3% B1a-update. Last Accessed 21 Sep 2022 23. Nature Homepage: https://www.nature.com/articles/d41586-022-01668-1. Last Accessed 21 Sep 2022 24. VOA Homepage: https://www.voanews.com/a/prolonged-la-ni%C3%B1a-likely-to-worsendrought-in-horn-of-africa/6614144.html. Last Accessed 21 Sep 2022 25. ABC Homepage: https://www.abc.net.au/news/2022-09-13/bom-declares-third-la-nina-sum mer-weather-flood-risk/101424100. Last Accessed 21 Sep 2022 26. RRI Homepage: https://rri.co.id/sumenep/daerah/1470732/cuaca-tidak-menentu-produksigaram-di-pamekasan-terhambat?utm_source=news_main&utm_medium=internal_link& utm_campaign=General%20Campaign. Last Accessed 21 Sep 2022 27. Bramawanto, R., Abida, R.F.: The climatology aspect review ( ENSO and IOD ) against the production of salt in Indonesia (In Indonesian). Jurnal Kelautan Nasional 12(2), 91–99 (2017) 28. Helmi, A., Sasaoka, M.: Dealing with socioeconomic and climate-related uncertainty in smallscale salt producers in rural Sampang Indonesia. J. Rural Stud. 59, 88–97 (2018) 29. Harini, S., Mulyanto, A.D., Iriany, A., Astuti, C.C.: Interpolation methods of spatial distribution salt price during scarcity period in Malang Raya using geographic information system. Int. J. Adv. Trends in Comput. Sci. Eng. 8(6), 3120–3129 (2019) 30. Hersbach, H., Bell, B., Berrisford, P., Hirahara, S., Horányi, A., Muñoz-Sabater, J., ……. Thépaut, J.: The ERA5 global reanalysis. Quart. J. Royal Meteorol. Soc. Part A 146(730), 1999–2049 (2020) 31. Rayner, N.A., Parker, D.E., Horton, E.B., Folland, C.K., Alexander, L.V., Rowell, D.P., Kent, E.C., Kaplan, A.: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res. Atmos.108(D14) (2003) 32. Pusat Riset Kelautan Homepage: http://pusriskel.litbang.kkp.go.id/index.php/en/publikasi/pre diksi-produksi-garam. Last Accessed 24 Sep 2022

Climate Indicators Triggering Attacks of Rice Stem Borer as Early Detection Information Suciantini, Erni Susanti, Elza Surmaini, Misnawati, and Yudi Riadi Fanggidae

Abstract In rice cultivation, one of the climate-related disasters that often occurs is the attack of plant pests and diseases. In Indonesia, especially on Java Island, the primary pest of rice plants with the most widespread attack is the rice stem borer (RSB). The purpose of this paper is to compile the relationship between RSB attacks and climate parameters. In the future, the relationship obtained will be used as input for preparing early detection model, as one of the adaptations technology efforts. The relationship between RSB attack and climate parameters was analyzed using multiple regressions equation to get the best model. The climate parameters used include; minimum air temperature, maximum air temperature, average air temperature, humidity, and rainfall. The attack area data is used from seven districts in Central Java. The analysis results show that the distribution of additional monthly rice stem borer attacks in seven districts in Central Java which mainly occurs in the rainy season (October–March) and the highest in March. Climatic parameters that have a high and significant correlation to the added area of RSB attack are maximum air temperature and humidity.

1 Introduction Agricultural production, especially food crops, is greatly affected by variability and climate change [1–4]. Climate change can have a negative impact by increasing crop damage due to the disasters it causes. Climate-related disasters can trigger a decline in food production, especially rice crops. One of the climate-related disasters to that can disrupt production is the attack of pests and plant diseases [5]. Increased exposure, increased status, decreased respiration, and the rate of pests and diseases in Indonesia are related to climate change [6]. Suciantini (B) · E. Susanti · E. Surmaini · Misnawati · Y. R. Fanggidae The Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_64

697

698

Suciantini et al.

Rice stem borer (RSB) is one of the most significant pests in Indonesia, apart from rats and brown planthoppers [7, 8]. The growth of rice plants can be attacked by RSB at all stages [9], from the nursery to the mature stage [10]. In Indonesian terms, they are known as sundep (dead hearts) and beluk (white ear heads), which represent RSB in the vegetative and generative phases [11–14]. At the beginning of planting, the intensity of the RSB attack is high, while in the productive phase, it decreases [10]. In addition, [15] stated that the dynamic nature of the pest population is influenced by the environmental constraints. Climate is very influences on the abundance of RSB [16]. Climate parameters are thought to be the trigger for pest and disease explosions, including the RSB. Population distribution, development level, and pest phenology are primarily determined by air temperature [12]. However, the study of climate elements for early detection of pest and disease outbreaks is still minimal. The purpose of this paper is to compile the relationship between RSB attacks and climate parameters. In the future, the relationship obtained will be used as input for the preparation of early detection models, as one of the adaptation technology efforts on agricultural production to avoid the adverse effects of plant attacks of pests and diseases. Prevention of development pests and plant diseases can reduce loss of crop yields, resulting in economic losses can be reduced.

2 Material and Methods The study was distributed from January to December 2021. Developing an early detection model of RSB attack data in Central Java Province (Regencies of Batang, Brebes, Tegal, Pekalongan, Pemalang, City of Tegal, and City of Pekalongan) with a data period of 2008–2019. The data was obtained from The Center for Protection of Food Crops, Horticulture, and Plantations (BPTPHP) in Central Java Province. The materials needed are long series of pest attack data and climate data in food crop production centers. Based on the availability of climate data, the observed climate data include; minimum air temperature, maximum air temperature, average air temperature, humidity, and rainfall. The climate data used in this study is observation data from the Cacaban station, Tegal Regency, Central Java Province, and Climate data derived from NASA POWER. The climate data used for analysis is data from the BMKG’s Tegal climate station, which was obtained from BMKG’s data online. The period of the data obtained was 2005–2020 year. The BMKG’s Tegal station daily climate data consists of rainfall data, T maximum, T minimum, T mean, and air humidity (RH). NASA POWER data was used to fill in blank data from observational data. NASA POWER data was accessed via the link https://power.larc.nasa.gov/data-access-vie wer/ by entering the coordinates of the Cacaban Station, Tegal (Latitude − 6.8672, Longitude 109.1372). The filling in blank data was preceded by analyzing the bias correction of NASA POWER data on observational climate data using regression equations.

Climate Indicators Triggering Attacks of Rice Stem Borer as Early …

699

(1) = observational data, α: regression constant, β: regression coefficient, and λ: NASA POWER data. After the regression equation was obtained, it was used to calculate the corrected NASA POWER data. This corrected NASA POWER data was used to fill in the unavailable observation data [17, 18]. The data analysis method used is correlation analysis followed by multiple linear regression analysis with a stepwise procedure. Analysis of the relationship between the extent of pest attack and climate parameters in several studies [19] used multiple linear regression equations. Climatic factor data was used as an independent variable (x), and data on the extent of rice stem borer attacks as a response variable (Y ). The regression equation is as follows: Y = a + b1 x1 + b2 x2 + b3 x3 + b4 x4 + b5 x5

(2)

where: Y = area of attack of RSB, x = climatic elements, a = intercept, and b= regression coefficient.

3 Result and Discussion 3.1 Result Complement Data with NASA POWER NASA POWER data was chosen because the data is available, easy to use, and userfriendly [20]. Figures 1 and 2 show a comparison of observational data with data from NASA POWER; the two types of data show almost similar patterns in RH and temperature, but there are variations in rainfall. The difference in the highest value was obtained on October I. The result of NASA POWER is more significant than that recorded from the observation station.

3.2 The Distribution of Additional Monthly Rice Stem Borer (RSB) Attack The rice stem borer (RSB) is the primary pest of rice plants with the most widespread attack in Indonesia, especially on the island of Java [21]. In this paper, the data used is the added area of RSB attacks in Batang, Brebes, Tegal, Pekalongan, Pemalang regencies, and the City of Tegal and Pekalongan. The highest attack area of the seven regencies/cities was Pemalang Regency, followed by Pekalongan and Batang regencies. Stem borer attacks fluctuate from year to year. An extreme event of additional RSB attacks occurred in 2012 in Pemalang Regency. Figure 3 shows that

700

Suciantini et al.

Fig. 1 Plotting of 2-week average rainfall and RH observations with rainfall and RH data from POWER data for the 2008–2019 data period

Fig. 2 Plotting of 2-week average data T max , T min , and T mean (Tegal station) with POWER data

the average added area of RSB attack is 300 ha per 2 weeks, and in the last two years, it is seen that the added area of RSB attack has decreased below the average value. The decrease of the attack can occur due to several things, including caused of simultaneous planting in a large area (1000 ha) with a span of 15 days [22]. In general, many factors influence the dynamics of pest development, including: crop

Climate Indicators Triggering Attacks of Rice Stem Borer as Early …

701

Fig. 3 Dynamics of RSB added area for the period 2008–2019 total of seven districts/cities

cultivation, cropping patterns, the presence of natural enemies, control methods, and climate change [23]. The distribution of additional monthly RSB attacks in seven regencies/cities of Central Java (Fig. 4) shows that RSB attacks occur mainly in the rainy season (October–March) and the highest in March. This statement is in line with the opinion of [16], which stated that the explosion of RSB attacks occurs when the intensity of rainfall is high. The area added attacks are increase in January, February, March, May, and June. However, in July, it was found that data stating that the area added to the attack had reached almost 2500 ha. This data is outlier or extreme observation. Knowing the magnitude of the spike in the area added to the attack, it is necessary to study at what phase the pest is present and the factors that may interfere with its ecological balance because this magnitude is very useful for controlling the pest [24]. From May to September, widespread incidents of RSB attacks continue. This result is shown by the relatively high minimum observation value, around 300 ha on May and June. The median monthly attack area added from January to July was in the range between 500 and 1000 ha, except for February, where it was > 1000 ha, the highest average monthly area added attack area. The median area added by RSB from August to December was lower at < 500 ha. In December, the median additional attack area increased from the previous months. This result indicates that in the rainy season, the size of attack is higher than in the dry season [16, 25]. Based on the illustration in the boxplot, the monthly maximum attacks occurred on March (> 1500 ha). Also in March, the variation in the area of addition of RSB was most varied, with the lowest quartile of ± 100 ha, a median of 600–700 ha, the highest quartile of 1200 ha, and a maximum value of around 1600 ha (Fig. 4).

702

Suciantini et al.

Fig. 4 Distribution of additional monthly RSB attacks in seven regencies/cities of Central Java

3.3 The Relationship of Area Affected by Rice Stem Borer (RSB) Attack with Climatic Parameters To see the relationship between climate parameters and the attack of the RSB, a regression analysis was performed between the climate parameters lag1 to lag3 with the attack area of RSB. Table 1 shows that the five climate parameters with a relatively high and significant correlation are T max and RH. The maximum T correlation was highest at lag1 (2 weeks before the attack), which was −0.474, and humidity at lag3 (1.5 months (6 weeks) before the attack), which was 0.418. From the results of the correlation analysis, it can be seen that there is an inverse relationship between temperature and the area of attack of the RSB. The lower the air temperature, the greater the attack area. This result is in line with the statement of [26], which stated that the increase in RSB (yellow stem borer) attacks was caused by low temperatures and high humidity. The difference in RH is directly proportional to the extent of the attack. RSB is very important to watch out for, especially for areas that have experienced two rice plantings. Table 1 Correlation of area affected by RSB attack with climatic parameters The added area of attack (ha)

T max (°C)

T min (°C)

Tmean (°C)

Rainfall (mm)

RH (%)

Lag0

− 0.398

− 0.032

− 0.220

0.250

0.313

Lag1

− 0.474

0.024

− 0.227

0.287

0.394

Lag2

− 0.405

0.105

− 0.126

0.326

0.397

Lag3

− 0.406

0.185

− 0.075

0.317

0.418

The bold letters represents the two highest correlations between RSB attacks and climate parameters

Climate Indicators Triggering Attacks of Rice Stem Borer as Early …

703

Table 2 Correlation of log area affected by stem borer attack with climatic parameters The added area of attack (ha)

T max (°C)

T min (°C)

T mean (°C)

Rainfall (mm)

RH (%)

Lag0

− 0.345

− 0.142

− 0.250

0.189

0.193

Lag1

− 0.350

− 0.099

− 0.236

0.160

0.240

Lag2

− 0.338

− 0.015

− 0.167

0.214

0.290

Lag3

− 0.331

0.093

− 0.085

0.207

0.347

The bold letters represents the two highest correlations between RSB attacks and climate parameters

The additional area of the RSB attack was also tried to be logarithmic, then seeing the correlation, and it turned out that the correlation value was lower than without logging, as shown in Table 2. However, the climatic elements that had an effect are the same, namely RH at lag3 (1.5 months before the attack) which was 0.347 and maximum temperature at lag1 (2 weeks before the attack) which was -0.350. The dominant climatic factors that are the key to the spread of RSB attacks were analyzed using stepwise regression. The results of the stepwise regression between RSB attack area and climate parameters obtained five dominant climate parameters in step 7, namely the maximum temperature at Lag1 and Lag3, average temperature, and humidity at Lag1 and Lag3. The basis for regression decision-making stepwise used is the lowest Mallows Cp, the lowest S, and R-sq (coefficient of determination) is 26,66%, and R-sq Adjusted (coefficient of determination corrected) is 25,35% (the highest). A low R-sq shows that only 26,66% make up the model, while the rest is determined by unknown factors. Based on stepwise regression, the equation obtained is: RSBattack area = 3697−119Tmax (Lag1)−125Tmax (Lag3) + 170Tmean −12.7RH(Lag1) + 7.7RH(Lag3) To know the triggering factors for the high prevalence of RSB attacks, it is necessary to look at the data starting from 6 weeks before the attack. The maximum temperature data to watch out for ranges from 28 to 33 °C, especially between 30.5 and 33 °C, the average temperature is 26–29 °C, and the RH is between 70 and 90%. According to [27], for the growth of yellow stem borer eggs, the optimum temperature is around 24–29 °C. If the temperature and humidity conditions are simultaneously met or are in the temperature and humidity range, then it can be initial information to anticipate attacks. For this reason, the data provider needs to pay attention to the data obtained from the field periodically. By studying the correlation and climatic parameters that trigger RSB attacks, it is expected to predict crop loss due to RSB attacks. Furthermore, the relationship obtained can be used as an early warning, part of the adaptation technology to reduce the risk of agricultural losses.

704

Suciantini et al.

4 Conclusions The rice stem borer (RSB) attack was significantly triggered by maximum temperature and humidity. In seven districts/cities in Central Java, RSB attacks mostly occur in the rainy season and are highest in March. Based on stepwise regression analysis, the determinant predictors were obtained, namely maximum temperature at Lag1 and Lag3, average temperature, and RH at Lag1 and Lag3. The R-sq received is relatively low at 26.66%. The maximum temperature data to watch out for ranges from 28 to 33 °C, especially between 30.5 and 33 °C, the average temperature is 26–29 °C, and the RH is between 70 and 90%. There are still many processes to go through in developing a tool for an early warning system. The correlation between RSB attacks and climate parameters is only the initial stage for the preparation of an early warning system, so this activity should be continued so that it can be used as a reference for climate information for agriculture.

References 1. Apriyana, Y., Susanti, E., Suciantini, Ramadhani, F., Surmaini, E.: Analysis of climate change impacts on food crops production in dry land and design of information system. Informatika Pertanian 25(1), 69–80 (2016) 2. Surmaini, E., Faqih, A.: Climate events and their impacts on food crop in Indonesia. Jurnal Sumberdaya Lahan 10(2), 115–128 (2016) 3. Ruminta.: Analysis of decreasing production of paddy due to climate change in Bandung district West Java. Jurnal Kultivasi 15(1), 37–45 (2016) 4. Kotir, J.H.: Climate change and variability in Sub-Saharan Africa: a review of current and future trends and impacts on agriculture and food security. Environ. Dev. Sustain 13, 587–605 (2011) 5. Maulana, W., Suharto, W.: Response of some varieties of rice (Oryza Sativa L.) to Pest Borer and “Walang Sangit” (Leptocorisa acuta Thubn.) attack. Agrovigor 10(1), 21–27 (2017) 6. Wiyono, S.: Perubahan iklim dan ledakan hama penyakit tanaman. Keanekaragaman Hayati Di Tengah Perubahan Iklim : Tantangan Masa Depan Indonesia. KEHATI, Jakarta (2007) 7. Susanti, E., Surmaini, E., Estiningtyas, W.: Climate parameters as indicators of early warning attack on pest and diseases of plant. Jurnal Sumberdaya Lahan 12(1), 59–70 (2018) 8. Nurhayati, E., Koesmaryono, Y., Impron.: Predictive modeling of rice yellow stem borer population dynamics under climate change scenarios in Indramayu. In: The 3rd International Seminar on Sciences “Sciences On Precision And Sustainable Agriculture” (ISS-2016). IOP Conf. Series: Earth and Environmental Science vol. 58. Bogor, Indonesia (2017). 9. Cahyono, G.R., Nurmahaludin.: Rancang bangun alat perangkap hama tanaman padi menggunakan arduino mega 2560. Jurnal Poros Teknik 7(2), 54–105 (2015) 10. Pertiwi, E.N., Mudjiono, G., Rachmawati, R.: Hubungan populasi ngengat penggerek batang padi yang tertangkap perangkap lampu dengan intensitas serangan penggerek batang padi di sekitarnya. Jurnal HPT 1(2), 88–95 (2013) 11. Aryantini, L.T., Supartha., Wijaya, I.N.: Population abundance and rice stem borer attack on rice in Tabanan regency. Jurnal Agroekoteknologi Tropika 4(3), 203–212 (2015) 12. Baehaki.: Hama penggerek batang padi dan teknologi pengendalian. Iptek Tanaman Pangan 8(1), 1–14 (2013)

Climate Indicators Triggering Attacks of Rice Stem Borer as Early …

705

13. Rahaman, M.M., Stout, M.J.: Comparative efficacies of next-generation insecticides against yellow stem borer and their effects on natural enemies in rice ecosystem. Rice Sci. 26(3), 157–166 (2019) 14. Jasrotia, P., Khippal, A., Yadav, J., Kashyap, P.L., Kumar, S., Singh, GP.: Effect of weather variables on the incidence of yellow stem borer (Scirpophaga incertulas W.) and leaf folder (Cnaphalocrocis medinalis G.) in rice. J. Cereal Res. 11(3), 247–251 (2019) 15. Junaedi, E., Yunus, M., Hasriyanty.: Parasitoids and it is parasitisme on white rice stem borer (Scirpophaga innotata WALKER) in two different altitudes of rice fields (Oryza sativaL.) in District of Sigi. e-J.Agrotekbis 4(3), 280–287 (2016) 16. Koem, S., Koesmaryono, Y., Impron.: Population phenology modeling of rice yellow stem borer Scirpophaga incertulas (Walker) based on climate effect. Jurnal Entomologi Indonesia 11(1), 1–10 (2014) 17. Rodrigues, G.C., Braga, R.P.: Evaluation of NASA POWER reanalysis products to estimate daily weather variables in a hot summer mediterranean climate. Agronomy 11, 1207 (2021) 18. Marzouk, O.A.: Assessment of global warming in Al Buraimi, sultanate of Oman based on statistical analysis of NASA POWER data over 39 years, and testing the reliability of NASA POWER against meteorological measurements. Heliyon 7, e06625 (2021) 19. Susanti, E., Ramadhani, F., June, T., Amien, I.: The use of climate information for developing early warning system to brown plant hopper attack on paddies. Jurnal Tanah dan Iklim 30, 47–58 (2009) 20. Rodrigues, G.C., Braga, R.P.: Estimation of daily reference evapotranspiration from NASA POWER reanalysis products in a hot summer Mediterranean climate. Agronomy 11, 2077 (2021). https://doi.org/10.3390/agronomy11102077 21. Misnawati, M., Boer, R., June, T., Faqih, A.: Perbandingan metodologi koreksi bias data curah hujan Chirps. Limnotek 25(1), 18–29 (2018) 22. Sinaga, A., Rohaeni, W.R., Marbun, O.: Studi perkembangan populasi hama penggerek batang padi hasil tangkapan light trap di Kecamatan Cilamaya Wetan – Karawang MK 2013. Seminar Nasional Agro Inovasi berbasis Sumberdaya Lokal dalam Meningkatkan Kesejahteraan Masyarakat dan Petani 23. Ilyas, A., Djufry, F.: Correlation and regression analysis pest population and natural enemy dynamic of certain varieties rice after implementation of superior IPM in Bone South Sulawesi. Informatika Pertanian 22(1), 29–36 (2013) 24. Makarim, A.K., Widiarta, I.N., Hendarsih, S., Abdulrachman, S.: Petunjuk Teknis Pengelolaan Hara dan Pengendalian Hama Penyakit Tanaman Padi Secara Terpadu. Departemen Pertanian, Jakarta (2003) 25. Misnaheti, B., Aisyah, D.: Tren perkembangan penggerek batang pada tanaman di Sulawesi Selatan. http://www.peipfi-komdasulsel.org/wpcontent/uploads/2011/06/410-415TREN-PERKEMBANGANPENGGEREK-Misnaheti.pdf. Last accessed 10 Mar 2021 26. Patel, S., Singh, C.P.: Seasonal incidence of rice stem borer, Scirpophaga incertulas (Walker) on different varieties of rice in relation to weather parameters. J. Entomol. Zool. Stud. 5(3), 80–83 (2017) 27. Saleh, T.W., Buri, N., Saragih, A.A.: Keragaan hama, penyakit dan musuh alami pada budidaya beberapa varietas padi gogo di lahan sawah. In Prosiding Temu Aplikasi Teknologi & Seminar Nasional Pertanian dan Peternakan: Akselerasi Inovasi Pertanian Era Industri 4.0 Mendukung Sapira. Hal, pp. 163–170 (2020)

Improvement of the Cropping Index and Farmers’ Resilience in Rainfed Fields Through the Application of Climate Smart Agriculture Aris Pramudia , Abriani Fensionita, Yunita Fauziah Rahim, Asis Purwoko, Andriarti Kusumawardani, and Muhammad Takdir Mulyadi Abstract One of the causes of low agricultural production on rainfed land is cultivation techniques that are highly dependent on rainfall conditions, with a high risk of crop failure in the second planting season. This condition is further exacerbated by farmers’ limited understanding and skills in adapting to rainfall variability. This paper presents an experience applying climate smart agriculture to rainfed agriculture in Lawang Subdistrict, Malang Regency, East Java Province, Indonesia. Several technics were implemented, including setting a planting schedule according to rainfall conditions, cropping pattern technique with the optimum composition of commodity types on water availability, conducting simple observations of rainfall and soil water content levels, making biopores, and application of biopesticides and organic fertilizers. The learning has some positive impacts, including early planting can be accelerated by 20–30 days from October III–November I to September III–October I, the cropping index increases from 2 times to 4–5 times a year, improving the cropping patterns from rice-paddy or rice-fallow into maize-rice-maize-peanuts-long beans, seedling age is shorter than 20–30 days after sowing to 12–15 days after sowing. Before the implementation of CSA, farmers experienced harvests 1–2 times a year, with a 60% risk of crop failure in the second planting season. After implementing CSA, farmers have harvested 4–5 times with several various commodities. Rice production also increased from 4.0 to 4.2 t/ha to 6.5–7.0 t/ha. However, CSA is a specific location technology; therefore, the application of CSA in rainfed fields still needs to be developed in other areas or regions.

A. Pramudia (B) Research Center for Climate and Atmospheric, National Research and Innovation Agency, Kota, Bandung, West Java 40173, Indonesia e-mail: [email protected] A. Fensionita · Y. F. Rahim · A. Purwoko · A. Kusumawardani · M. Takdir Mulyadi Directorate of Food Crops Protection, Directorate General of Food Crops, Indonesia Ministry of Agriculture, Jakarta 12520, Indonesia A. Pramudia Innovation Center for Tropical Science, Bogor 16113, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_65

707

708

A. Pramudia et al.

1 Introduction Rainfed paddy fields are fields for rice cultivation that utilize water sources only from rainfall [1, 2]. This field is usually in a flat area or a relatively wide basin but has no water source other than rainfall. Plants cultivated in rainfed rice fields are generally seasonal food crops, such as rice and palawija. Food crop production on rainfed land is generally lower than on irrigated fields because it has limited planting time, which only depends on rainfall conditions [3, 4] and has low soil fertility [5]. In agricultural areas of Indonesia, rainfed areas generally can only be planted two times a year [6]. The second planting usually coincides with the end of the rainy season or the beginning of the dry season, so it has a high risk of crop failure. This condition is further exacerbated by the limited understanding and skills of the farmer in anticipating and adapting to rainfall fluctuation and climate variability. To overcome the problem of food crop agriculture on rainfed fields, one solution is to applicate some climate smart agriculture concepts [7]. Climate smart agriculture (CSA) is an approach that transforms and reorients agricultural production systems and food value chains so that they both support sustainable agriculture and ensure food security in conditions of climate change. Based on the three main pillars, the CSA concept aims to: (1) Increase agricultural productivity and income sustainably, (2) implement adaptation and build resilience to climate change, and (3) reduce and, or eliminate greenhouse gas emissions, where possible [8]. This paper presents an experience of applying climate smart agriculture to food agriculture on rainfed land in Lawang Subdistrict, Malang Regency, East Java Province, Indonesia. Several things were implemented, including setting a planting schedule according to rainfall conditions, designing intercropping cropping patterns with the composition of adaptive types of commodities to water availability conditions, conducting simple rainfall observations, observing simple soil water content levels, making biopores, using biological pesticides, as well as the provision of organic fertilizers on the farmers’ fields.

2 Methodology 2.1 Study Location The study was conducted in rainfed rice fields managed by the Farmers’ Group Aman Makmur in Srigading Village, Lawang Subdistrict, Malang Regency. The rainfall data used to represent the study location is rainfall data from the Karangploso Climatology Station, obtained from the BMKG online, accessed via the link http://dataon line.bmkg.go.id/data_iklim [9]. Other cultivation technical information is obtained through the Srigading Village Farmers Group, Lawang Subdistrict, Malang Regency. This study was carried out in 2020–2021.

Improvement of the Cropping Index and Farmers’ Resilience in Rainfed …

709

2.2 Rainfall Analysis for Improving the Cropping Calendar Recommendations and the Cropping Patterns In determining the planting schedule on rainfed land, farmers referred to the rules, namely (1) planted rice as the first commodity, (2) thus, the first planting schedule waited for the start of the rainy season, and (3) the next planting was applied after the previous plant has harvested. For the first planting, farmers estimated the start of the rainy season by the frequency of heavy rain days. Referring to the criteria for determining the initial rainy season from the BMKG, the first planting approach was when the rainfall intensity during three successive decades is > 50 mm/decade [10]. Rainfall prediction information was obtained from BMKG. The innovation of the study introduced in mentoring farmers in this study was to modify the rules in determining the planting schedule and crops management, as follows: (1) the first commodity to be cultivated does not have to be rice, but plants that have fewer water requirements, (2) the planting time can be faster than the beginning of the rainy season, (3) the next planting is applied without having to wait for the harvest of the previous crop, (4) land space is prepared for planting the next crop, and (5) other modifications adapted in the field. Modifications to the planting schedule and plant composition settings applied in the field were as follows: (1) The first crop was maize. (2) The second crop was rice with the jajar legowo 2:1 technique. (3) The third crop was maize, planted when the rice was two months old, in an empty row in the jajar legowo technique. (4) The fourth crop was peanuts, planted on former rice fields after the rice was harvested. (5) The fifth crop was long beans planted in rows of maize that had been harvested. The approach taken in planting maize is when the intensity of rainfall at the beginning of planting and during the corn growing season is > 25 mm/decade (Table 1).

2.3 Capacity Building of Farmers and Improvement of Land Quality To improve land quality and farmer capacity, and to increase land productivity [11], the following steps were taken: (1) use biological pesticides to control the pest, (2) applicate biological fertilizers on the land, (3) design biopores on the land, (4) teach farmers how to measure rainfall, and (5) teach farmers how to monitor groundwater level conditions. A simple rainfall gauge was made from a milk canister, and the

710

A. Pramudia et al.

Table 1 Application of CSA in the field and its purpose CSA application

Goals

1. Rainfall analysis

Determining the optimum planting time and adjusting the water availability

2. Management of cropping patterns and crop rotation

Implement the effective combination of crop commodities

3. Application of an effective cultivation system Combining the jajar legowo 2:1 technique, inserting planting time, intercropping 4. Improvement of land quality 4.1. Use of biological pesticides

Application of healthy and environmentally friendly agricultural cultivation

4.2. Use of biofertilizer

Utilization of agricultural waste for green manure

4.3. Designing biopores

Provision of pore space to increase water availability

5. Increasing the farmer’s capacity 5.1. Rainfall measurement

Farmers can measure rainfall simply using a milk canister and a ruler

5.2. Groundwater monitoring

Farmers can monitor the soil water level simply using a paralon and a ruler

rainfall that was collected in the canister was measured using a ruler. Monitoring the soil water level was carried out by making a control hole from a paralon as deep as 40 cm and measuring the water level using a ruler (Table 1).

3 Results 3.1 Rainfall Pattern The research location has a monsoon rainfall pattern, which is a monomodal pattern that is closely related to Indonesia or Southeast Asia monsoon wind circulation [12]. With this pattern, the research location alternates between one wet period and one dry period. In the agricultural calendar, the wet period is usually called the rainy season, while the dry period is usually called the dry season. The annual rainfall in the study area is 2041 mm/year, spreading 1732 mm in the rainy season and 309 mm in the dry season. With this annual rainfall, the study area is categorized as a rainfed agricultural area [1] with a moderate climate [13]. There are six wet months with an intensity of more than or equal to 200 mm/month, and six dry months with a power of less than or equal to 100 mm/month (Fig. 1). The Agroclimate Zone is C3, meaning it has a potential planting period of six months, especially for rice plants [14].

Improvement of the Cropping Index and Farmers’ Resilience in Rainfed …

711

400

Fig. 1 Figure of monthly precipitation fluctuations and annual rainfall in Malang

5$,1)$// PPPR

350

$QQXDO 5DLQIDOO

 PP\U

300 250 200 150 100 50 0 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

3.2 Improvement of the Crop Planting and Cropping Pattern Management The learning process on farmers’ cultivation illustrates that the first planting follows the start of the rainy season, at October II–III or November I–II [15, 16]. The commodity applied in the first growing season is rice, which is grown conventionally using the tiled spacing technique. In the second planting season, several commodities applied by farmers include rice, maize, and other crops. The selection of commodities is adjusted to the conditions of rainfall experienced by farmers (Fig. 2). To introduce the implementation of CSA for improving the cropping schedules and the cropping patterns, farmers were suggested to do not to apply the continuously flooded rice. Farmers could apply maize which requires less water than rice for the first crop. By applying maize as the first crop, the planting time can be carried out 20–30 days earlier, from September III–October I (Fig. 2). After harvesting during the first planting season, the season is at the peak of the rainy season. Maize harvest is done by cutting the base of the stem. The second planting season is carried out around December II–III or January I–II. In this second planting season, rice plants were applied with the jajar legowo 2:1 planting technique. The planting jajar legowo 2:1 is a rice planting technique by emptying one row of rice plants and tucking the plants in that row into two rows on either side. Thus, there are two consecutive rows of rice plants with closer spacing between plants and interspersed with the third row of empty rows [17]. The technique of irrigating water for rice crops is adjusted to the rainfall conditions during the growing season. If water through rainfall is considered sufficient, then irrigation water can be applied through continuous inundation. However, if the water through rainfall is estimated to be limited, the provision of water can be carried out continuously for 3–5 days, following drying conditions also for 3–5 days, alternately (intermittent technique).

712

A. Pramudia et al.

Fig. 2 Average rainfall conditions, the scheme of existing cropping patterns, and the recommendations for alternative cropping patterns that apply the CSA concept to rainfed land in Lawang Subdistrict, Malang Regency

When the rice plants are two months old or enter the generative phase and do not experience flooding again, the third planting is carried out around March I–II, applying maize. Planting is carried out on empty rows between 2-row rice plants that enter the generative phase. So simultaneously on the land, there are generative phase rice plants and vegetative phase maize plants. When the maize plant enters the generative phase, namely in March III–April I, the rice harvest is carried out by slashing, but the remaining forage plants are left on the land. After the rice was harvested in April II–III, the fourth planting was carried out on the former rice fields. The fourth crop applied is peanuts. Thus, there is an intercropping system between corn plants in the generative phase and peanut plants. In May II–III, maize was harvested by leaving the maize stalks on the land. After the maize harvest, if it is predicted that there will be sufficient rainfall, the fifth planting will be carried out in May III–June I by applying long beans. Long bean seeds are planted in rows between maize stalks still left in the field. When the long bean plant grows taller, the remaining maize stalks serve as vines for the long bean plants. Thus at this stage, there is an intercropping system between the peanut and the long bean. In June II–III, peanuts were harvested. The last is the harvest of long beans can be done in August II–III. Using that cropping pattern, farmers experience 4–5 planting times in a year using various food crop commodities.

Improvement of the Cropping Index and Farmers’ Resilience in Rainfed …

713

Fig. 3 Installation of biopores, simple rainfall measurement using milk canister, and monitoring of soil water levels in the fields

3.3 Improvement of Land Quality and Farmer Capacity To prepare better fields, biopores were made in the field and the application of biofertilizers. The biopores were built with a size of about 50 cm × 50 cm with a depth of 100–150 cm. Biopore holes are filled with organic litter to facilitate absorption and the soil water-holding capacity. Improvement of farmer capacity is carried out through technology transfer and assistance in implementing technology carried out by farmers on the land. After going through the process of assisting with the application of CSA on rainfed rice fields, farmers can measure rainfall using canister tubes and can monitor the depth of the soil water table on the field using pipes and rulers (Fig. 3).

3.4 The Impact of CSA Implementation The learning and mentoring process of CSA implementation for farmers has several positive impacts, including: • The planting time of the first cropping season can be accelerated up to 20–30 days. • The cropping index increased from 2 times a year to 4–5 times a year. • The seedling duration can be shortened from 20 to 30 days after sowing to 12– 15 days after sowing. • The harvest index with 1–2 times a year and a 60% risk of crop failure in the second planting season, can be increased to 4–5 times a year with minimum risk of crop failure. • Rice production increased from 4.0 to 4.2 t/ha to 6.5–7.0 t/ha.

714

A. Pramudia et al.

Table 2 Added value was obtained after the application of CSA to the field Treatment/parameter

Before CSA implementation

After mentoring and CSA implementation

1. Cropping calendar

October II–III/November I–II

September III–October I

2. Cropping pattern

Rice-palawija (maize/peanut/beans) Rice-rice

Maize-rice-maize-peanut-long beans

3. Cropping seasons

2 times a year

4–5 times a year

4. Crop rotations

Sequential, conventional

Combination of sequential, rotational, intercropping, inserting planting time

5. Rice seed

Not certified, 75–100 kg/ha

Certified, 30 kg/ha

6. Seedling duration

20–30 days after sowing

12–15 days after sowing

7. Organic fertilizer

–

1–2 t/ha

– Urea

400 kg/ha

150 kg/ha

– Phonska

–

150 kg/ha

– SP36

100 kg/ha

–

– ZA

100 kg/ha

–

9. Pest management

Chemical pesticide

POC, biological pestiside, MOL

10. Harvest

1–2 times In the second harvest, 60% failed due to a lack of water

4–5 times Supported by the management of cropping patterns, biopores, and intermittent irrigation

11. Rice productivity

4.0–4.2 t/ha

6.5–7.0 t/ha

8. Chemical fertilizer

• Farmer’s knowledge and skills in weather observation, weather data utilization, and other CSA applications are increased. • Soil fertility has been increasing, indicated by the application of fertilizer which is getting lower from year to year (Table 2). As with other research, they are engineering the implementation of various technologies in rainfed agriculture that can increase productivity, soil fertility, and farmers’ income and capacity [18, 19].

4 Conclusion Farming systems, by applying the principles of climate smart agriculture, such as engineering cropping schedules and crop patterns, increasing groundwater infiltration, and monitoring rainfall and groundwater levels, can increase the cropping index, harvest index, and agricultural production.

Improvement of the Cropping Index and Farmers’ Resilience in Rainfed …

715

Applying climate smart agriculture through engineering schedules and cropping patterns could change farmers’ habits. However, because this is not easy, this process will be an ‘exciting’ challenge in the development of CSA-based agriculture. However, it is believed that the application of CSA is very site-specific, where one land or area can be different from another land or area. For this reason, the application of CSA on rainfed or dry land still needs to be developed or replicated in other regions or areas. Acknowledgements Acknowledgments are conveyed to the Directorate of Food Crop Protection, Directorate General of Food Crops, Ministry of Agriculture and Rainfed Farmers’ Group Aman Makmur in Srigading Village, Lawang Subdistrict, Malang Regency, East Java Province, Indonesia, for sharing information and field experiences so that this paper is compiled.

References 1. Singh, M., Tiwari, N.K., Kumar, N., Dabur, K.R., Dehinwal, A.K.: Dry and rainfed agriculturecharacteristics and issues to enhance the prosperity of Indian farming community. Pharmacol. Life Sci. Bull. Env. Pharmacol. Life Sci. 6(10), 32–38 (2017) [Online]. Available: http://bepls. com/OCT_2017/6.pdf 2. Mevada, K.D., Poonia, T.C., Saras, P., Deshmukh, S.P.: Rainfed agriculture and watershed management 6th semester B.Sc. (Tech) Agri. Course No. : Agron 6. In: 10 Title of Course: Rainfed Agriculture and Watershed Management. S.K. Kataria & Sons (2021) 3. Sulaiman, A.A., Candradijaya, A., Syakir, M.: Technological advancement and the economic benefit of Indonesian rain-fed farming development. Adv. Agric. (2019). https://doi.org/10. 1155/2019/9689037 4. Baig, M.B., Shahid, S.A., Straquadine, G.S.: Making rainfed agriculture sustainable through environmentally friendly technologies in Pakistan: a review. Int. Soil Water Conserv. Res. 1(2), 36–52 (2013). https://doi.org/10.1016/S2095-6339(15)30038-1 5. Benauli, A.: Kajian Status Hara N, P, K Tanah Pada Sawah Tadah Hujan (Studi Kasus Tiga Desa di Kecamatan Beringin). Agrosains J. Penelit. Agron. 23(1), 55 (2021). https://doi.org/ 10.20961/agsjpa.v23i1.49239 6. Arifin, S.A.A., Sadat, M.A.: J. Agribisnis 9(2), 130–138 (2021). [Online]. Available: https:// ejournals.umma.ac.id/index.php/agribis/article/view/1080 7. Pramudia, A., Apriyana, Y., Susanti, E., Suciantini, Harmanto.: Pertanian cerdas iklim indonesia: konsep dan teknologi. In: Sulaiman Y, Listianingsih W (eds) Pertanian Cerdas Iklim Indonesia: Konsep dan Teknologi. Bekasi, pp. 175–192 (2022) 8. FAO: Climate Smart Agriculture, Sourcebook. Rome (2013) 9. BMKG: Data-Online BMKG (2022). http://dataonline.bmkg.go.id/data_iklim 10. BMKG: Prakiraan Musim Kemarau 2020 di Indonesia. BMKG (2020) 11. Erythrina, E., et al.: Assessing opportunities to increase yield and profit in rainfed lowland rice systems in Indonesia. Agronomy 11(4), 1–15 (2021). https://doi.org/10.3390/agronomy1104 0777 12. Tjasyono, H.B.: The character of rainfall in the Indonesian monsoon. In: International Symposium on Equatorial Monsoon System, pp. 1–11 (2008). Available: http://file.upi.edu/Direkt ori/SPS/PRODI.PENDIDIKAN_IPA/BAYONG_TJASYONO/Kumpulan_Makalah/The_Cha racter_of_Rainfall.pdf 13. Susanti, E., et al.: Pemutakhiran Peta Sumberdaya Agroklimat Indonesia untuk Mendukung Perencanaan Pertanian. In: Updating of the Agro-climate Resources Map of Indonesia to

716

14. 15. 16.

17.

18.

19.

A. Pramudia et al. Support Agricultural Planning, vol. 45, no. 1, pp. 47–58 (2021) [Online]. Available: http:// ejurnal.litbang.pertanian.go.id/index.php/jti/article/view/12370 Oldeman, L., Las, I., Darwis, S.N.: An Agroclimatic Map of Sumatra, vol. 52. Central Research Institute for Agriculture, Bogor (1979) Pramudia, A., et al.: Cropping calendar analysis for dry season 2020 in Indonesia. IOP Conf. Ser. Earth Environ. Sci. 648(1) (2021). https://doi.org/10.1088/1755-1315/648/1/012117 Apriyana, Y. et al.: The integrated cropping calendar information system: a coping mechanism to climate variability for sustainable agriculture in Indonesia. Sustainability 13(11):1–22 (2021) [Online]. Available: https://www.mdpi.com/2071-1050/13/11/6495 Nengah Surata Adnyana, I.: Efektifitas Sistem Tanam Jajar Legowo 2:1 Dengan Sistem Tegel Terhadap Produktivitas Padi Sawah Di Subak Babakan Cangi, Desa Batuan Kaler, Kecamatan Sukawati, Kabupaten Gianyar. dwijenAGRO 10(2):127–133 (2020). https://doi.org/10.46650/ dwijenagro.10.2.1037.127-133 Rahmadiyah, R., Tanjung, F., Hariance, R.: Analisis Perbandingan Usahatani Padi Sawah Irigasi Dengan Padi Sawah Tadah Hujan Di Kecamatan Koto Tangah Kota Padang. JOSETA J. Socio Econ. Tripical Agric. 1, 9–23 (2019) [Online]. Available: http://joseta.faperta.unand. ac.id/index.php/joseta/article/view/177 Widyantoro, Toha, H.M.: Optimalisasi Pengelolaan Padi Sawah Tadah Hujan Melalui Pendekatan Pengelolaan Tanaman Terpadu. Pros. Semin. Nas. Serealia Meningkat. Peran Penelit. Serealia. Menuju Swasembada Pangan Berkelanjutan, pp. 648–657 (2010)

Applications of Soil Conditioner Polyacrylamide to Suppress Runoff and P (Phosphorus) Nutrients Loss at the Sweet Corn Cultivation Under Climate Change Issue Niken Ida Lovita, Ali Rahmat, Yuliana Eva Agasi, Sendi Purnama Hidayat, Yogina Lestari Ayu Situmorang, and Dwi Rustam Kendarto Abstract Climate change is a phenomenon that treats the environment. Climate change will trigger extreme weather conditions such as high rainfall, which promotes runoff, flooding, and landslide. High rainfall will disintegrate soil particles that can cause surface runoff. The surface runoff will cause the transport of top soil, which contains a lot of nutrients for the plants, and this will make the land infertile and decrease its productivity. To control surface runoff and soil productivity can be done by applying a soil amendment, such as polyacrylamide (PAM). Polyacrylamide is an organic-anionic polymer that could stabilize the soil structure by maintaining a water-permeable pore structure while forming the soil surface. This study aims to determine the effect of polyacrylamide on the reduction of surface runoff and P nutrient loss in sweet corn cultivation. The results showed that polyacrylamide had an effect in reducing surface runoff and P nutrient loss. Polyacrylamide with a dose of 60 kg/ha is the optimal dose that can reduce surface runoff by 14% and reduce the P nutrients loss by 12.16% during eight rainstorms. Meanwhile, for yield and sweet corn’s weights, the dose of polyacrylamide 15 kg/ha was the most optimal.

1 Introduction Global warming and climate change are two of the most extensively researched and debated environmental issues [1]. Climate change is defined as a change in the state of the climate that can be identified (e.g., using statistical tests) by changes in the mean and variability of its properties over an extended period of time, typically decades N. I. Lovita · Y. E. Agasi · S. P. Hidayat · Y. L. A. Situmorang · D. R. Kendarto Department Agriculture Engineering, Faculty of Agricultural Industrial Technology, Padjadjaran University, Sumedang, Indonesia A. Rahmat (B) Research Center for Limnology and Water Resources, National Research and Innovation Agency, Jakarta Pusat, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_66

717

718

N. I. Lovita et al.

or longer [2, 3]. Changes in global climate and the Earth’s hydrological cycle [4–6] have increased heavy rainfall, increasing the risk of surface runoff and flooding [7, 8]. Surface runoff is one of the causes of high critical land in Bogor Regency. High rainfall is one of the main causes of high surface runoff, especially in Indonesia [9]. The average rainfall in Bogor Regency is in the range of 100–150 mm/day which is included in the very high category [10]. The high surface runoff will cause the loss of topsoil which contains a lot of nutrients transported by surface runoff. Transport of nutrients by surface runoff will cause eutrophication of surface water bodies (lakes, reservoirs, and rivers). Due to the high rainfall and steep slopes, the soil is prone to erosion and surface runoff. The efforts can be made to mitigate the high surface runoff by improving the soil’s physical properties. The physical properties of the soil can be improved by using soil amendments, one of which is polyacrylamide (PAM). PAM is watersoluble and high molecular weight. PAM is frequently used as a soil amendment to improve infiltration, reduce surface runoff, prevent soil erosion, and prevent nutrient loss [11]. Previous research in China by Jiang et al. [12] found that PAM at 20 kg/ha is the optimal dose for reducing surface runoff and can reduce phosphorus transport in surface runoff by 27%. According to Kabede et al. [13], PAM at up to 40 kg/ha is the optimal dose for reducing surface runoff. According to the above explanation, using PAM to control runoff and nutrient loss is one strategy for dealing with climate change. However, the effectiveness of PAM must be investigated because different locations and times will produce different results. The purpose of this study was to determine the effectiveness of PAM at five different dosage levels in suppressing surface runoff and P nutrient loss in tropical soils in Indonesia under sweet corn cultivation.

2 Methods The research was carried out at the Research Center for Limnology and Water Resources, National Research and Innovation Agency, Cibinong, Bogor, West Java, in January – May 2022. The materials used in this study were sweet corn seeds, manure, KCl fertilizer, TSP fertilizer, urea fertilizer, and polyacrylamide (PAM). This study was arranged using a completely randomized design method (CRD) consisting of five treatment doses and repeated as many as rain events that occurred. The doses of polyacrylamide applied in this study were: P0 (0 kg/ha), P1 (15 kg/ha), P2 (30 kg/ha), P3 (45 kg/ha), P4 (60 kg/ha) using plants sweet corn trial with a plot size of 2 × 6 m totaling 5 plots. The data analysis used in this study was the ANOVA statistical test. If there was a significant difference, then the LSD test was carried out at an error level of 5%. There are three components analyzed in this study, namely surface runoff (runoff), nutrients (P), and components of sweet corn production (cob length, cob diameter, weight of cob with cob, and weight of cob without cob). Surface runoff (runoff) is calculated using the following formula [14]:

Applications of Soil Conditioner Polyacrylamide to Suppress Runoff …

719

Fig. 1 Plot illustration

Ro = V /A

(1)

Information: Ro = Runoff (mm) V = Volume of runoff (ml) A = Cross-sectional area of the plot (m2 ). The P nutrients were analyzed using the 25% HCL extraction method at the ICBB Laboratory - PT. Indonesia’s Biotechnology Biodiversity. Sweet corn production components were measured using a ruler, caliper, and scales. Tillage is processed to speed up the planting process. The tools used for processing the soil are hoes, machetes, and shovels. Tillage is carried out by clearing the land of weeds. The land is divided into five plots, with each plot measuring 2 m × 6 m. Each plot was separated using a zinc plate (Fig. 1). The fertilizer used is 10 tons/ha for manure, 65.49 kg/ha for TSP fertilizer, 200 kg/ha for urea, and 124.46 kg/ha for KCl. Before sowing on the surface of the land, the fertilizers are mixed until homogeneous with a predetermined dose for each plot. After loosening the soil, basic fertilizer is sprinkled over the entire surface of the soil and then distributed the fertilizer equally. After a few hours, PAM will be applied with a predetermined dose for each plot and the plot will be left for 7–10 days before planting. Sweet corn seeds were planted with a spacing of 75 cm × 40 cm and 2 seeds per planting spot. The irrigation is carried out by depending on the condition of the land and on the rainfall that occurs. It will be harvested when the sweet corns are 60–80 days old. Sampling will be carried out on the day after the rain, taken every morning at 9.00 WIB. Surface runoff samples were taken from the conductor, which was used as a reservoir for runoff and erosion on the land. Surface runoff and erosion water samples were separated using a sieve. Surface runoff samples were calculated using

720

N. I. Lovita et al.

Fig. 2 Surface runoff sample

a measuring cup. In addition, surface runoff is also calculated from the evaporation of water during the erosion sample oven process (Fig. 2).

3 Methods 3.1 General Condition of Research Land The research land has a latosol soil type with characteristics of red–yellow soil, slow permeability, and clay texture. In addition, the slope is 44%, which is in the steep category. The condition research land has a steep slope and also has very slow soil permeability, which can increase the possibility of runoff (Table 1). Table 1 Physical and chemical properties of the soil in the research area

Soil properties

Value

Category

Texture: % Silt

8

Clay

% Sand

9

Clay

% Clay

83

Permeability (cm/h)

0.93

Very slow

P nutrients (mg/100 g)

69

–

Applications of Soil Conditioner Polyacrylamide to Suppress Runoff …

721

3.2 Effect of Polyacrylamide on Surface Runoff The runoff will occur when rain falls at a high enough intensity with a relatively long duration. The kinetic energy of rain will destroy soil aggregates, so the crushed soil particles will clog the soil pores. Blockage of soil pores by finer particles causes surface runoff because the soil cannot absorb water. The results of the ANOVA analysis showed that the PAM (polyacrylamide) treatment was not significantly different in all treatments (P > 0.05) (Table 2). PAMtreated plots generally produced lower surface runoff than those without PAM treatment (P0). For six rainfall events, namely the first rain event to the sixth rain event, P1 always produces the lowest runoff while P4 produces the largest runoff. In addition to P4, P0 also generated large surface runoff for six rainfall events. The cause of the large surface runoff at P0 is the absence of PAM granules scattered on the soil surface. Successive rain events caused the destruction of soil aggregates and the formation of seals on the soil which caused an increase in surface runoff at P0. The high runoff at P4 is caused by the higher the PAM dose level, the smaller the soil pores will make it difficult for rainwater to seep into the soil. The results of this analysis are in accordance with research conducted by Kabede et al. [13] that higher PAM (polyacrylamide) dose levels resulted in greater surface runoff at the beginning of the rain event. P1 resulted in the lowest runoff for the six rainfall events. The low runoff at P1 was due to the 15 kg/ha PAM dose that could produce optimal aggregate stability and low soil viscosity compared to higher PAM doses to suppress runoff that occurred. In the last two rain events, the runoff at P4 decreased significantly and became the lowest of the other treatments. The drastic decrease in surface runoff at P4 was caused by the soil drying up in the before two last rains, which caused the soil viscosity to decrease. When it rained, the PAM ware still in the soil absorbed some of the water, reducing runoff at higher PAM doses. For eight rainfall events, P3 and P0 produced the most surface runoff when compared to other plots. Many non-growing or dead crops contributed to the significant increase in surface runoff volume at P3. In this case, vegetation serves to intercept and control the amount of rainwater that falls to the ground surface; due to many plants was dead, the function of plant to intercept and control the water reach the ground surface was decreasing or does not work, with the consequence will promote high runoff. Table 2 Results of ANOVA analysis between treatments and surface runoff Component

Treatments P0

P1

P2

P3

P4

Runoff (mm)

1252.3a

1259.264a

1459.753a

1691.305a

1085.361a

Description: similar letter notation means there is no significant difference at the Duncan test level which has a value of 5%

722

N. I. Lovita et al.

Higher PAM doses can increase surface runoff at the beginning of a rain event but can reduce erosion. This follows the research conducted by Agasi [15], which stated that the 60 kg/ha PAM dose was the optimal dose for reducing erosion due to surface runoff. PAM with a dose of 15 kg/ha can suppress surface runoff at the beginning of the rain event, but its effectiveness decreases in the last rain event. With the highest dose of 60 kg/ha, PAM increased the surface runoff at the beginning of the rain event but could reduce the surface runoff in the last few rain events. The use of PAM at a dose of 30 kg/ha did not produce a large surface runoff during eight rain events. The use of PAM doses that are too high is not good for the environment, and also, from an economic point of view, the costs incurred are quite large, so their use is not effective as well as the use of PAM doses that are too low in effectiveness cannot last until the last rain event. The dose of PAM used in suppressing surface runoff was not significantly different between treatments, so it can be concluded that the PAM dose of 30 kg/ha was the optimal dose in reducing surface runoff during the eight rain events in this study, both for environmental safety and also for the economy.

3.3 Effect of Polyacrylamide on P Nutrients Loss One of the important pathways for the P nutrients loss from agricultural land is surface runoff. Surface runoff carries important soil particles, one of which is P nutrients loss. The results of ANOVA analysis showed that PAM treatment was significantly different in the P nutrients loss (P > 0.05). Table 3 shows that the largest P nutrient loss due to runoff and erosion is in the P0. P4 resulted in the smallest P nutrient loss during the eight rainfall events. Figure 3 shows that during the six rains, from the first rain to the sixth rain, the amount of P nutrients loss decreased according to the increasing dose of PAM in the soil. P4 always resulted in the smallest P nutrient loss, while the P0 always resulted in the largest P nutrient loss for six rainfall events. The plot with PAM treatment had lower P nutrient loss compared to P0 without PAM treatment. The high P nutrients loss in P0 was due to the absence of PAM in the soil, which caused the soil to be easily carried away by surface runoff which contained nutrients. P4, with a higher PAM dose level compared to other treatments, produced soil aggregates that were very binding to one another. During rain, the soil and nutrients were not easily carried away by surface runoff. Table 3 Results of ANOVA analysis between treatments and P nutrients loss Component

Treatments P0

P1

P2

P3

P4

P nutrients loss (mg/100 g)

82.2300c

78.7250b

74.2813a

74.4475a

71.8550a

Description: similar letter notation means there is no significant difference at the Duncan test level which has a value of 5%

Applications of Soil Conditioner Polyacrylamide to Suppress Runoff …

723

Fig. 3 P nutrients loss during eight rainstorm

Table 4 Comparison of the effect of PAM on the loss of nutrient N with control plot Treatment

P Nutrients loss (%)

P1

78.725

P2

74.28125

P3

74.4475

P4

71.855

Treatment plots

Control plot (P0) 82.23

Difference

Percentage drop (%)

3.505

4.26

7.94875

9.66

7.7825 10.375

9.46 12.16

There was a drastic increase in P nutrient loss in P1 and P3 during the sixth and seventh rain events. The increase in P nutrient loss in P1 could be due to the insufficient dose level of PAM in P1 to retain the P nutrients loss by surface runoff. This increase in P nutrients loss could be caused because after seven rain events the effectiveness of PAM decreased with increasing time and intensity of rain. The increase in P nutrients loss in P3 could be due to the death of the vegetation in P3 toward the end of the rain event (Table 4). The PAM dose rate of 60 kg/ha reduced P nutrients loss by 12.16% during eight rainy events. This result follows the research conducted by Agasi [15], which stated that the PAM dose of 60 kg/ha was the optimal dose for reducing erosion so that the nutrient loss would be less in this plot. P0 always yielded the greatest P nutrients loss for the eight rainfall events. The P nutrients loss in the P0 without PAM application was 3 times greater than that in P4.

3.4 Effect of Polyacrylamide on Sweet Corn Production Sweet corn was planted on February 18, 2022, but after ten days after planting, some sweet corn plants did not grow so replanting was carried out. Replanting is done at the planting point that did not grow on March 1, 2022 (Table 5). The results of the analysis showed that polyacrylamide (PAM) had an effect on the diameter of the cobs (P < 0.05), but it was not significantly different for the weight of

724

N. I. Lovita et al.

Table 5 Effect of polyacrylamide production of sweet corns Treatment

With corn husks leaves

Without corn husks leaves

Corn weight (g)

Corn length (cm)

Corn diameter (cm)

Corn weight (g)

Corn length (cm)

Corn diame ter (cm)

P0

231.36a

5.4a

25.62a

162.29a

4.71a

17.18a

P1

296.13a

5.91a

29.00ab

203.21a

4.82a

19.76b

P2

266.08a

5.65a

28.31ab

18,826a

4.8a

1812ab

P3

246.91a

5.53a

27.37ab

17,677a

4.73a

1893ab

P4

297.35a

5.86a

31.06b

18,207a

4.62a

1956ab

Description: similar letter notation means there is no significant difference at the Duncan test level which has a value of 5%

the cobs and the length of the fruit, whether with husk leaves or not. The yields in plot 4 produced the largest fruit weight with husk leaves and diameter of fruit with husk leaves compared to other plots, while the highest fruit length with husk leaves was in plot 1. The results of the largest fruit weight without husk leaves were found in plot 1 as well as fruit length and fruit diameter without husk leaves. Yields increased with the PAM application when compared to control plots without PAM application. Plots with PAM application can increase by 25% of fruit weight without husk leaves, 2% of fruit length without husk leaves, and 15% of fruit diameter without husk leaves than the control plot. This is because PAM can increase the availability of nutrients for plants. Increased crop yields also can be caused by PAM where expected can increase water infiltration in the soil so that the availability of water for plants is met. PAM increases the movement of water on the soil surface so that more water flows into the planting row compared to the control plot. The dose level of 15 kg/ha contained in plot 1 is the most optimal dose in increasing crop yields because in plot 1, infiltration is always high from the beginning of the rain event to the last rain event during the study. This can also be seen in a surface runoff where plot 1 always produces low surface runoff when compared to plots treated with other PAMs.

4 Conclusion The conclusion of the research is that the 30 kg/ha dose is optimal in reducing runoff by14%. The PAM dose of 60 kg/ha is the optimal dose in suppressing the P nutrients loss by 12.16%. A dose of 15 kg/ha is the optimal dose to increase sweet corn yields.

Applications of Soil Conditioner Polyacrylamide to Suppress Runoff …

725

References 1. Rahmat, A., Zaki, M.K., Effendi, I., Mutolib, A., Yanfika, H., Listiana, I.: Effect of global climate change on air temperature and precipitation in six cities in Gifu Prefecture, Japan. J. Phys.: Conf. Ser. 1155, 012070 (2019) 2. Sturrock, R.N., Frankel, S.J., Brown, A.V., Hennon, P.E., Kliejunas, J.T., Lewis, K.J., Worrall, J.J., Woods, A.J.: Climate change and forest diseases. Plant. Pathol. 60(1), 133–149 (2011) 3. Rahmat, A., Mutolib, A.: Comparison air temperature under global climate change issue in Gifu city and Ogaki City, Japan. Indones. J. Sci. Technol. 1, 37–46 (2016) 4. Allen, M.R., Ingram, W.J.: Constraints on future changes in climate and the hydrologic cycle. Nature 419, 224–232 (2002) 5. Held, I.M., Soden, B.J.: Robust responses of the hydrological cycle to global warming. J. Climate. 19, 5686–5699 (2006) 6. Wentz, F.J., Ricciardulli, L., Hilburn, K., Mears, C.: How much more rain will global warming bring? Science 317, 233–235 (2007) 7. Trenberth, K.E.: Conceptual framework for changes of extremes of the hydrological cycle with climate change. Clim. Change 42, 327–339 (1999) 8. Trenberth, K.E.: Changes in precipitation with climate change. Clim. Res. 47, 123–138 (2011) 9. Kartasapoetra, A.G.: Teknologi Konservasi Tanah dan Air. Rineka Cipta. Jakarta (2005) 10. Badan Pusat Statistik Kabupaten Bogor. Jumlah Curah Hujan Menurut Bulan dan Pos Hujan (mm/hari). Badan Meteorologi Klimatologi dan Geofisika Stasiun Klimatologi Bogor. https:// bogorkab.bps.go.id/indicator/151/80/1/jumlah-curah-hujan-menurut-bulan-dan-stasiun-poshujan.html (2022). 11. Entry, J.A., Sojka, R.E.: The efficacy of polyacrylamide to reduce nutrient movement from an irrigated field. Trans. Am. Soc. Agric. Eng. 46(1), 75–83 (2003) 12. Jiang, T., Teng, L., Wei, S., Deng, L., Luo, Z., Chen, Y., Flanagan, D.C.: Application of polyacrylamide to reduce phosphorus losses from a Chinese purple soil: a laboratory and field investigation. J. Environ. Manage. 91(7), 1437–1445 (2010) 13. Kebede, B., Tsunekawa, A., Haregeweyn, N., Mamedov, A.I., Tsubo, M., Fenta, A.A. Meshesha, D.T., Masunaga, T., Adgo, E., Abebe, G., Berihun, M.L.: Effectiveness of polyacrylamide in reducing runoff and soil loss under consecutive rainfall storms. Sustainability 12(4), 1–18 (2020) 14. Ziliwu, Y.: Pengaruh Beberapa Macam Tanaman Terhadap Aliran Permukaan dan Erosi. Universitas Diponegoro. http://eprints.undip.ac.id/11329/1/2002MTS1772.pdf (2002) 15. Agasi, Y.E.: Kajian Aplikasi Soil Conditioner Polyacrylamide Untuk Menekan Erosi dan Kehilangan Unsur Hara N (Nitrogen) Di Lahan Budidaya Jagung Manis. Under Graduate Thesis. Padjajaran University (2022)

Elemental Analysis of Breadnut Seed Biochar and Its Potential Application as a Soil Amendment Sukamto and Ali Rahmat

Abstract Soil acidification can be promoted by long-term cultivation and the excess use of fertilizer. Moreover, increased precipitation due to climate change, can exacerbate leaching and contribute to soil acidification which correlates to the decrease in soil fertility. It has significant adverse impacts not only on agricultural production but also on economic aspects. To overcome this problem, various ways have been applied to improve soil fertility by reducing soil acidity. Recently, waste-based biochar products were used as a promising soil amendment. In this study, breadnut seed-based biochar has been successfully produced by using thermal treatment at 250 and 350 °C. The elemental composition has been evaluated by using X-ray fluorescence spectrometer (XRF). According to the elemental analysis, there is no significant difference in the amount of elemental composition for the different thermal treatments. Generally, breadnut seed-based biochar obtained at 250 °C shows higher nutrient content such as Mg, P, S, K, Ca, and Fe which are 0.716%, 5.476%, 1.757%, 71.063%, 19.685%, and 0.583%, respectively. The presence of these nutrients on the breadnut seed-based biochar increases its potential application for soil amendment to reduce soil acidity and enhance soil fertility. Increasing soil fertility and increasing crop productivity was one strategy for climate change mitigation indirectly.

1 Introduction The 2022 Global Report on Food Crises (GRFC) reported the highest food insecurity which is nearly 25%. Due to the COVID-19 pandemic, conflict, and extreme climate change, more than 193 million people in 53 countries are suffering from food crisis [1]. Moreover, the increasing of population around the world leads this problem turn to worse. On the other hand, various efforts have been conducted to Sukamto Hokkaido University, Sapporo, Japan A. Rahmat (B) National Research and Innovation Agency, Jakarta Pusat, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_67

727

728

Sukamto and A. Rahmat

face this issue in the last decade such as food diversification and agriculture intensification. However, aiming to increase agriculture production for fulfilling food needs, agriculture intensification promotes the excessive use of fertilizers that can reduce soil quality such as soil acidification. Moreover, because of rising temperatures and changes in precipitation (amount and frequency), climate change is expected to have a significant impact on soils and ecosystems. Increased precipitation, can exacerbate leaching and contribute to soil acidification [2]. Soil acidification changes the soil quality by decreasing soil pH < 5.5 which can be caused by the deposition of acid [3] and excessive use of ammonium-based fertilizers [4]. The decrease in soil pH causes an increase in aluminum and manganese toxicities toward crops as well as a decrease in the content of phosphorus and nutrient ions such as calcium, potassium, magnesium, and molybdenum [5]. This condition has a negative impact on soil quality which results in a decrease in crop yields. Therefore, appropriate, and efficient steps must be taken to maintain or improve soil acidity so that global food availability can be achieved, and the world food crisis can be overcome. Recently, various materials have been applied for recovering soil acidification such as lime, gypsum, red mud, organic waste, and residual crops. These materials show good abilities to increase soil pH and show cation exchange. However, the high heavy metals contents on these materials can cause soil contamination and promotes dangerous side effect for environmental [6]. Owing by this limitation, previously, some researchers have focused to use biochar as the soil amendment. Bicohar was obtained from pyrolyzed feedstocks at high temperatures ranging from 200 to 700 °C, biochar has shown excellent ability to improve soil quality compared to other organic materials [7]. Generally, biochar is alkaline in nature due to the presence of Mg, K, Ca, and Fe ions. It promotes liming effects when biochar is applied into the acid soil which contribute to the ameliorate soil acidity and increase of crop production [8]. With the respect of biochar application for agronomic purposes, the content of nutrient ions on the biochar is one of the most important aspects that determine its ability to improve soil quality. The concentration of nutrient ions depends on the used feedstock and pyrolysis temperature. Moreover, due to different pyrolysis temperature, toxic elements may be presented on the biochar that also can promote dangerous impact for plant growth, environment, and human health [9]. Therefore, the elemental analysis is a crucial way to ensure that biochar contains sufficient and safe ions for plant growth and production. In this study, breadnut seed biochar was created by using pyrolysis method, where the temperature of pyrolysis at 250 and 350 °C. To evaluate the elemental content of biochar, non-destructive and environmentally friendly analysis method was applied, where X-ray fluorescence spectrometry (XRF) has been used as the main technique. This study not only offers environmentally friendly method to analysis total elemental composition of biochar, but also promote the use of organic waste by product such as breadnut seed to be functional materials as a soil amendment for agronomic application.

Elemental Analysis of Breadnut Seed Biochar and Its Potential …

729

Fig. 1 Dried raw breadnut seed (a) and biochar product (b)

2 Methods 2.1 Material Breadnut seed was obtained from market. Breadnut seed waste was collected and then washed by using clean water. After wash, breadnut seed waste was air-dried for 5 h after that coarsely blended. The blended breadnut seed then was dried in an oven at 105 °C for 24 h.

2.2 Biochar Preparation and Characterization Dried breadnut seed was placed and burned in a furnace at 250 and 350 °C for 4 h [10, 11]. The obtained biochar then was labeled with BBc250 and BBc350 for biochar that was obtained at 250 °C and 350 °C, respectively. The raw material and biochar product is shown in Fig. 1. Both BBc250 and BBc350 were milled and sieved to 355 µm in size. The biochar powder was analyzed by using X-ray fluorescence spectrometer (XRF) (Omnian ED-XRF Panalytical Epsilon 3 XLE) at Lampung Advanced Characterization-BRIN. The data was analyzed descriptively by comparing the total elemental composition of biochar obtained at 250 and 350 °C.

3 Results and Discussions 3.1 Elemental Analysis of Biochar Product For several thousand years, biochar which is a solid product obtained from pyrolyzed biomass has been well known and produced for various application such as heat production, metallurgical application, building material, and agriculture. Generally,

730

Sukamto and A. Rahmat

Fig. 2 Element (a) and oxide (b) composition of BBc250 and BBc350

biochar can be obtained from decomposition of biomass during pyrolysis process. The conversion process is depending on the temperature and characteristic of feedstock that use as the starting material. In this study, pyrolysis process has been conducted in the temperature 250 and 350 °C, and it is referred to as mild pyrolysis (torrefaction). Torrefaction is used to retain and concentrate most of the energy content in the solid product and increase mechanical properties of the products significantly [12]. During the torrefaction process, there are three temperature zones: (1) non-reactive zone (from 50 to 150 °C), (2) reactive zone (from 150 to 200 °C), and (3) destructive zone (from 200 to 300 °C) which includes some biomass reactions such as dehydration, devolatilization and carbonization of hemicellulose, depolymerization and devolatilization/softening of lignin, and depolymerization devolatilization of cellulose [13]. In this study, after pyrolyzed for 4 h under 250 and 350 °C, the obtained biochar was analyzed by using X-ray fluorescence spectrometer (XRF), and the results are depicted in Fig. 2. According to the graph in Fig. 2, it can be shown that biochar produces from mild pyrolysis at 250 and 350 °C contain various mineral compositions. The highest element composition and oxide composition is achieved by potassium. Where in elements form the percentage are 71.06% and 70.229% for BBc250 and BBc350, respectively. Moreover, the percentage potassium oxide form on the biochar which reached up to 60% for both BBc250 and BBc350. Beside potassium, calcium in form element and oxides lied on the second place after potassium. Another mineral that also presented on the BBc250 and BBc350 are Mg, P, S, Mn, Fe, Cu, Zn, Rb, Sr, and Sn. However, for BBc350 there is no Sn content detected on the element as well as its oxide. Naturally, wet breadnut seed contains these minerals including K, Ca, P, Fe, Cu, Mn, Na, and Mg [14]. Generally, during the biochar production, the carbonaceous solid is expected to form. However, in this study, carbon element is not formed and detected. It might be caused by the low temperature that used in the pyrolysis process. In terms of very high carbon content (up to 95%) on biochar, high temperature close to 1000 °C is required.

Elemental Analysis of Breadnut Seed Biochar and Its Potential …

731

3.2 Potential Application of Biochar as Soil Amendment The increase of world population across the globe promotes significant effect not only on the social aspect but also economics. Recently, this condition turned to worse due to the conflict, COVID-19 pandemic, and extreme climate change. Apart from all the most crucial needs, food availability is the vital thing that should be fulfilled. According to the 2022 Global Report on Food Crises (GRFC), more than 193 million people in 53 countries are suffering from food crisis. Owing to that problem, agriculture sector plays important role to face food crisis. Alternative programs have been enrolled to overcome the problem including agriculture intensification which then promotes the excessive use of fertilizers. Instead of reducing food crisis, the use of excessive fertilizer gives dangerous impact due to soil acidification and quality degradation. Generally, soil acidification is a natural process which can be occurred almost in tropical and subtropical regions. It is caused by several processes such as H+ ions precipitation, presence of acidic gasses including SO2 and NO3 , and the application of ammonium-based fertilizers (NH4 + ) [15, 16]. Soil acidification is indicated by the decrease of soil pH ( 1 km and λz < 1 km) ranges. We discuss the daily temperature perturbations, the deviation from the mean temperature profile, to evaluate the atmospheric planetary-scale waves. The monthly mean temperature perturbations were performed to explore the radiative variations in the stratosphere, at 2030 km altitude, associated with the regular wind variations about 28 months high above the equatorial region known as the Quasi-Biennial Oscillation [23]. We analyze the temperature perturbations using CHAMP UCAR, the static stability using Brunt Vaisala frequency variations from CHAMP RISH, and potential energy variations with CHAMP WEGC. The static stability and potential energy are determined by the following formula: N 2 = g/T [dT /dz + g/cp]

(1)

 E p = 1 2(g/N )2 (T  /T )2

(2)

with N 2 is brunt vaisala frequency, g is gravitational acceleration, z is altitude, cp is specific heat constants for constant pressure, Ep is potential energy, T is temperature, and T is mean temperature.

768

Noersomadi and N. Cholianawati

To support our analysis, we utilize the zonal wind data from the NCEP reanalysis data provided by the NOAA Physical Science Laboratory, Boulder, Colorado, USA, from their website at https://psl.noaa.gov.

3 Results 3.1 Individual Comparison Among Three Products Figure 1 shows the comparison of individual temperature product retrievals of CHAMP UCAR (dry and wet), CHAMP RISH (dry), and CHAMP WEGC (dry and wet). CHAMP UCAR provides the product reach to 60 km, while others were limited up to 40 km. The dry profiles indicate unrealistic values below ~8 km. We are able to see the RISH profile depicts very high vertical resolution in the stratosphere. Meanwhile, UCAR and WEGC products show smoother temperature profiles at 1530 km altitude intervals. All CHAMP temperature products describe an oscillation in stratosphere.

Fig. 1 Temperature profiles from the three product retrievals on January 15 2003, at 33.77 E and 11.39 N. Dry and wet profiles of UCAR and of WEGC are shifted by 5 K, 10 K, 15 K, and 20 K from the dry profile of RISH, respectively. The magnitudes of all dry profiles below about 8 km are unrealistic temperature data

Comparison of CHAMP GPS Radio Occultation Dry Temperature …

769

Next, we re-plot profiles in Fig. 1 into Fig. 2 by focusing on the altitude of 15– 22 km, which is called the tropopause layer and lower stratosphere regions. We observed detailed fluctuations in the CHAMP RISH profile, which indicates double minima temperature, 189.5 K at 16.8 km and 189 K at 17.6 km. On the other hand, UCAR and WEGC show a single minimum temperature at 17.6 km with slightly different values. This individual example resulting an issue regarding tropopause definition in the tropics. RISH wave optic profile suggested multiple minima, which may affect the vertical derivative of temperature and cold point temperature [24]. Both dry and wet profiles retrieved by WEGC almost coincided with each other in this altitude range. However, UCAR dry and wet describe as rather dissimilar though less than 1 K, particularly above 20 km. The wave optics applied by RISH could detect small amplitudes of short vertical scale fluctuations within 17–22 km. For further discussion, we examined analyzing dry profiles only. We investigated typical temperature profiles in the region of (a) PS, (b) PN, (c) MS, (d) MN, and (e) TR, as shown in Fig. 3 to Fig. 6. We considered 15 January, 15 April, 15 July, and 15 October as the representative dates of mid-season of northern hemisphere winter, spring, summer, and autumn, respectively. Table 1 describes the details location and time of each occultation. In the northern winter season (summer in the southern), the minimum temperature of PS is observed at 215 K at ~9 km, while in PN the coldest temperature reaches 195 K at 22 km (Fig. 3a–b). The altitude of the minimum temperature in MS is between 15 and17 km, which seems close to the altitude of

Fig. 2 Same profiles as in Fig. 1 concerning near the tropopause layer. No profile shifting is applied in this figure

770

Noersomadi and N. Cholianawati

Fig. 3 Typical temperature profiles in the a polar south (PS); b polar north (PN); c mid-latitude south (MS); d mid-latitude north (MN); and e tropics (TR) of the three product retrievals on January 15, 2003 (northern hemisphere winter). Both UCAR and WEGC profiles are shifted by 10 K and 20 K, respectively. Details of occultation information of each profile is described in Table 1

minimum temperature in TR. Meanwhile, it is seen the minimum peak temperature 210 K at ~10 km in MN (Fig. 3c–e). In the northern spring, we observed a local minimum temperature of 210 K at ~9 km in PS and 217 K at ~8 km in PN (Fig. 4a–b). On the other hand, the typical temperature profiles show similar patterns with slightly different minimum temperatures in the three regions of MS, MN, and TR (Fig. 4c–e). The minimum temperature in TR is the coldest temperature among other regions. Figure 5 shows the opposite situation, as indicated in Fig. 3, particularly between the PS and PN, also among MS and MN. In the northern autumn, when the southern region experiences spring, the temperature pattern of PS (Fig. 6a) is still similar to the pattern in southern winter (Fig. 5a) with the altitude of minimum temperature reach ~20 km. The local minimum temperature profiles in PN are found in 9–11 km (Fig. 5b and 6b). Figures 5d and 6d describe an identical temperature pattern in the MN region during northern summer and autumn. The results show consistency with the global height and latitude temperature pattern by [25, 26]. Overall, we noted the RISH wave optics profile depicts small-scale fluctuations in the 17–30 km altitude interval in all samples. We also found the altitude of local minimum temperature profiles in the polar regions in the summer hemisphere is around 9–11 km. For the mid-latitude regions, the altitude of minimum temperature is observed at ~15 km during the spring and autumn seasons. Meanwhile, we noticed

Comparison of CHAMP GPS Radio Occultation Dry Temperature …

771

Table 1 Time (HH:MM:SS UTC) and location of occultation event shown in Figs. 3, 4, 5 and 6 Date (2003) Polar South (PS)

Mid-latitude South (MS)

Tropics (TR)

Mid-latitude North (MN)

Polar North (PN)

15 Jan

15 Apr

15 Jul

15 Oct

Time

00:07:51

02:20:48

18:12:32

01:59:38

Lon

29.27E

121.57 W

146.76E

18.52E

Lat

72.69S

89.76S

76.42S

66.07S

Time

17:17:30

03:37:47

09:10:36

06:18:39

Lon

136.3E

64.90E

7.42E

135.77E

Lat

30.78S

32.11S

40.83S

43.77S

Time

05:47:36

00:20:26

14:45:03

17:40:02

Lon

129.60E

84.99E

108.32E

114.52E

Lat

6.70 N

4.32 N

1.67S

6.62 N

Time

02:11:07

12:08:56

11:29:29

16:17:45

Lon

41.43E

114.77E

147.80E

114.10E

Lat

49.51 N

39.06 N

32.68 N

37.56 N

Time

11:44:16

10:47:18

22:03:50

01:01:48

Lon

45.68E

142.57E

134.64E

31.87E

Lat

70.08 N

80.57 N

83.67 N

67.53 N

Fig. 4 Same as in Fig. 3 on April 15, 2003 (northern spring)

772

Fig. 5 Same Fig. 3 on July 15, 2003 (northern summer)

Fig. 6 Same as Fig. 3 on October 15, 2003 (northern autumn)

Noersomadi and N. Cholianawati

Comparison of CHAMP GPS Radio Occultation Dry Temperature …

773

the altitude of minimum temperature in the tropics is relatively constant in all seasons at ~17 km.

3.2 Synoptic and Planetary Scale Waves We collected all temperature profiles within 90–160 E and 15S–15 N in January 2004, then we calculated the daily mean temperature of all profiles at 20–32 km, which is known as the stable region of the stratosphere layer. The synoptic scale waves are often observed in this stable region [26, 27]. We demonstrated the temperature fluctuations in Fig. 7, derived by subtracting the monthly mean profile. The fluctuations consist of a mixture of multi-scale vertical wavelengths. We observed short vertical scales in the RISH profile up to 30 km related to the altitude limit of applying wave optics for the retrieval process [28]. To separate between long and short vertical scale waves, we used a fast fourier transform method for each profile. Therefore, we reconstructed temperature perturbations for λz > 1 km and λz < 1 km (Fig. 8). All three products indicate fluctuations from synoptic waves with λz > 1 km with amplitude ±2 K. CHAMP RISH product is able to demonstrate the amplitude of synoptic wave with λz ~2 km. For λz < 1 km, the amplitude of UCAR and WEGC profiles are close to zero. However, the

Fig. 7 Temperature fluctuations on January 1, 2004, over the 90–160 E, 15 S–15 N ¯

774

Noersomadi and N. Cholianawati

RISH profile shows a significant amplitude of ±0.3 K. We conclude that wave optics applied by RISH could resolve long and short vertical scales of synoptic waves. We analyzed the daily temperature profiles from all products within January 1, 2004, to April 25, 2004. The results of all three products are consistent. We show the height versus time variation of temperature perturbations from CHAMP UCAR in Fig. 9. The temperature perturbations were defined as the deviation from the mean temperature profile an entire period. To support the analysis, we also described the zonal wind anomaly over the same area. CHAMP temperature perturbations demonstrate clear downward cold and warm, indicating the propagation of planetary waves in March 2004. The downward temperature propagation is also associated with the westward and eastward wind. This propagation seems related to equatorial trapped wave activity [29, 30].

Fig. 8 Reconstructed temperature fluctuations T* of long (λz > 1 km) and short (λz < 1 km) vertical wavelength ranges from the temperature perturbations in Fig. 7

Comparison of CHAMP GPS Radio Occultation Dry Temperature …

775

Fig. 9 Height versus time-cross-section of daily temperature perturbations (top) and zonal wind anomaly (bottom) over 90–160 E, 15S–15 N in the period of January 1 to April 25 2004

3.3 Zonal Mean Temperature Fluctuations in the Stratosphere Figures 10, 11, and 12 show the height versus time cross-section of monthly temperature perturbations, static stability, and potential energy (Eq. 2) of synoptic-scale waves with three retrieval products. We overlaid the temperature perturbations and potential energy with the monthly zonal mean zonal wind at 70 hPa (~18.6 km) to 10 hPa (~32 km). For the static stability (Eq. 1), we superimposed it with the time derivative of zonal wind to analyze the relationship between vertical derivative of temperature and acceleration or deceleration of wind variation in the stratosphere. Top panel of each figure depicts the downward propagation of an oscillation between warm and cold temperature associated with an alternating westward and eastward wind. We noticed that a warm anomaly, particularly in the mid-2006, reached up to 6 K, and occurred when the wind direction changed from westward to eastward. On the other hand, a cold anomaly of about −6 to −4 K collocated with the change from eastward to westward. The annual oscillation of N2 is observed in the lower stratosphere as documented by [26, 28]. The downward propagation of the wind acceleration from 32 km altitude increased static stability starting from 28 km altitude. Meanwhile, when the wind decelerated, the strong static stability was seen trapped at ~22 km. We noted the high potential energy increased when the wind direction changed. All parameters shown

776

Noersomadi and N. Cholianawati

Fig. 10 Height versus time cross-section of CHAMP UCAR monthly temperature perturbations (top); Brunt Vaisala frequency square (middle); and potential energy of synoptic-scale waves (bottom). The top and bottom figures are overlaid by a contour line of zonal mean zonal wind anomaly with solid (eastward), dashed (westward), and thick solid (0 m.s–1 ). The contour line ranges from –25 to 20 m.s–1 with an interval is 3 m.s–1 . The middle panel is overlaid by a contour line –0.2 to 0.4 m.s–2 of the time derivative of zonal mean zonal wind with solid (acceleration) and dashed (deceleration)

by three retrieval products describe consistency of clear downward propagation in the stratosphere, linked with the Quasi-Biennial Oscillation.

4 Summary We investigated three CHAMP temperature product retrievals by UCAR, RISH, and WEGC. All CHAMP temperature products show fluctuations in the stratosphere

Comparison of CHAMP GPS Radio Occultation Dry Temperature …

777

Fig. 11 Same Fig. 10 for CHAMP RISH

layer. The wave optics retrieval applied by RISH provides very high vertical resolution up to 30 km altitude. RISH wave optics profiles indicate small-scale fluctuations in the lower stratosphere. We found the altitude of local minimum temperature in the polar summer region is 9–11 km, in mid-latitude spring and autumn is ~15 km, and in the tropics is at ~17 km for all seasons. We analyzed the vertical wavelengths of the temperature perturbations for λz > 1 km (long) and λz < 1 km (short) of synoptic scale waves. All three products show perturbation from the long vertical wavelengths. RISH wave optics profile was able to observe an amplitude ± 0.3 K of the short vertical scale. Hence, we summarize that wave optics profiles could resolve long and short vertical fluctuations. For the planetary-scale waves, CHAMP temperature describes downward cold and warm anomaly, associating with the westward and eastward wind. The height versus time cross section of monthly temperature perturbations, static stability, and potential energy shows downward propagation in the stratosphere. A

778

Noersomadi and N. Cholianawati

Fig. 12 Same Fig. 10 for CHAMP WEGC

warm (cold) anomaly of about ±6 K occurred when the wind direction changed from westward to eastward (eastward to the westward). Strong static stability is associated with downward wind acceleration. These are linked to the well-known phenomena in the tropical stratosphere called the Quasi-Biennial Oscillation. Acknowledgements This research is supported by National Research and Innovation Agency of Indonesia (BRIN) Grant Number 15/III/HK/2022.

Comparison of CHAMP GPS Radio Occultation Dry Temperature …

779

References 1. Wickert, J., et al.: GPS radio occultation with CHAMP: Initial results, pp. 4–7 (2001) 2. Hasebe, F., Aoki, S., Inai, Y., Nakazawa, T., Sugawara, S., Ikeda, C., Honda, H., Yamazaki, H., Halimurrahman., Komala, N., Putri, F.A., Budiyono, A., Soedjarwo, M., Ishidoya, S., Toyoda, S., Shibata, T., Hayashi, M., Eguchi, N., Nishi, N., Fujiwara, M., Coordinated uppertroposphere-to-stratosphere balloon experiment in Biak, (1213–1230) (2018) 3. Ho, S.-P., Zhou, X., Shao, X., Zhang, B., Adhikari, L., Kireev, S., He, Y., Yoe, J.G., Xia-Serafino, W., Lynch, E.: Initial assessment of the COSMIC-2/FORMOSAT-7 Neutral atmosphere data quality in NESDIS/STAR using in situ and satellite data. Remote Sens. 12, 4099 (2020). https:// doi.org/10.3390/rs12244099 4. Butchart, N.: The brewer-dobsoncirculation. Rev. Geophys. 52, 157–184 (2014). https://doi. org/10.1002/2013RG000448 5. Ho, S., Peng, L., Vömel, H.: Characterization of the long-term radiosonde temperature biases in the upper troposphere and lower stratosphere using COSMIC and Metop-A/GRAS data from 2006 to 2014. (4493–4511) (2017) 6. Bell, T.M., Klein, P.M., Lundquist, J.K., Waugh, S.: Remote-sensing and radiosonde datasets collected in the San Luis Valley during the LAPSE-RATE campaign. Earth Syst. Sci. Data 13, 1041–1051 (2021). https://doi.org/10.5194/essd-13-1041-2021 7. Yoneyama et al.: (2013). https://doi.org/10.1175/BAMS-D-12-00157.1 8. Anthes, R.A.: Exploring earth‘s atmosphere with radio occultation: contributions to weather, climate and space weather. Atmos. Meas. Tech. 4(6), 1077–1103 (2011) 9. Gorbunov, M., Irisov, V., Rocken, C.: The influence of the signal-to-noise ratio upon radio occultation retrievals. Remote Sens. 14(2742) (2022). https://doi.org/10.3390/rs14122742 10. Schreiner, W., Sokolovskiy, S., Hunt, D., Rocken, C.: Analysis of GPS radio occultation data from the FORMOSAT-3/COSMIC and Metop/GRAS missions at CDAAC. Atmos. Meas. Tech. 4(10), 2255–2272 (2011) 11. Sokolovskiy, S., Rocken, C., Schreiner, W., Hunt, D.: On the uncertainty of radio occultation inversions in the lower troposphere. J. Geophys. Res. Atmos. 115(22), 1–19 (2010) 12. Jensen, A.S., Lohmann, M.S., Benzon, H.H., Nielsen, A.S.: Full spectrum inversion of radio occultation signals. Radio Sci. 38(3), 1040 (2003) 13. Melbourne, W.G.: Radio occultation using earth satellites. Wiley (2004) 14. Jensen, A.S., Lohmann, M.S., Nielsen, A.S., Benzon, H.H.: Geometrical optics phase matching of radio occultation signals. RadioSci. 39(3), 1–8 (2004) 15. Tsuda, T., Lin, X., Hayashi, H., Noersomadi.: Analysis of vertical wave number spectrum of atmospheric gravity waves in the stratosphere using COSMIC GPS radio occultation data, Atmos. Meas. Tech. 4(8) (2011) 16. Wickert et al.: GRAS SAF workshop on applications of GPSRO measurements, 16–18 (2008) 17. Noersomadi., Tsuda, T.: Comparison of three retrievals of COSMIC GPS radio occultation results in the tropical upper troposphere and lower stratosphere 2. Aeron. Earth Planet. Sp. 69(1) (2017) 18. Schreiner, W., Rocken, C., Sokolovskiy, S., Hunt, D.: Quality assessment of COSMIC/FORMOSAT-3 GPS radio occultation data derived from single—and doubledifference atmospheric excess phase processing. GPS Solut. 14, 13–22 (2009). https://doi.org/ 10.1007/s10291-009-0132-5 19. https://data.cosmic.ucar.edu/gnss-ro/champ/repro2016/ 20. http://database.rish.kyoto-u.ac.jp/arch/iugonet/data/GPS/champ/ 21. Angerer, B., Ladstädter, F., Scherllin-Pirscher, B., Schwärz, M., Steiner, A.K., Foelsche, U., Kirchengast, G.: Quality aspects of the Wegener Center multi-satellite GPS radio occultation record OPSv5.6. Atmos. Meas. Tech. 10, 4845–4863 (2017). https://doi.org/10.5194/amt-104845-2017 22. https://wegcwww.uni-graz.at/data-store/WEGC/OPS5.6:2019.1/

780

Noersomadi and N. Cholianawati

23. Baldwin, M.P., Gray, L.J., Dunkerton, T.J., Hamilton, K., Haynes, P.H., Randel, W.J., Holton, J.R., Alexander, M.J., Hirota, I., Horinouchi, T., Jones, D.B.A., Kinnersley, J.S., Marquardt, C., Sato, K., Takahashi, M.: The quasi-biennial oscillation. Rev. Geophys. 24. Zeng, Z., Sokolovskiy, S., Schreiner, W.S., Hunt, D.: 2019: Representation of vertical structures by radio occultation observations in the upper troposphere and lower stratosphere: comparison to high-resolution radiosonde profiles. J. Atmos. Oceanic Tech. (2019). https://doi.org/10.1175/ JTECH-D-18-0105.1 25. Schmidt, T., Heise, S., Wickert, J., Beyerle, G., Reigber, C.: GPS radio occultation with CHAMP and SAC-C: global monitoring of thermal tropopause parameters. Atmos. Chem. Phys. 5(6), 1473–1488 (2005) 26. Grise, K.M., Thompson, D.W.J., Birner, T.: A global survey of static stability in the stratosphere and upper troposphere. J. Clim. 23(9), 2275–2292 (2010) 27. Duran, P., Molinari, J.: tropopause evolution in a rapidly intensifying tropical cyclone: a static stability budget analysis in an idealized axisymmetric framework. J. Atmos. Sci. 76, 209–229 (2019) 28. Noersomadi and Tsuda, T.: Global distribution of vertical wavenumber spectra in the lower stratosphere observed using high-vertical-resolution temperature profiles from COSMIC GPS radio occultation. Ann. Geophys. 34, 203–213. https://doi.org/10.5194/angeo-34-203-2016 29. Randel, W.J., Wu, F.: Kelvin wave variability near the equatorial tropopause observed in GPS radio occultation measurements. J. Geophys. Res. 110, D03102 (2005). https://doi.org/10.1029/ 2004JD005006 30. Kim, J., Son, S.W.: Tropical cold-point tropopause: climatology, seasonal cycle, and intraseasonal variability derived from COSMIC GPS radio occultation measurements. J. Clim. 25(15), 5343–5360 (2012)

Evaluation of ERA5 Precipitation Reanalysis Data in Indonesia Sigit Kurniawan Jati Wicaksana and Iis Sofiati

Abstract Rainfall data is used widely in essential applications such as flood/drought monitoring, water management, and climate monitoring. This information is critical in Indonesia, where rainfall has a significant socioeconomic impact. The selection of a precipitation product can have a significant impact on the performance of such an application that mentioned above. Rainfall products are classified into three types: direct measurement, satellite based, and reanalysis. The purpose of this study is to evaluate the performance of ERA5 precipitation data against Global Satellite Mapping of Precipitation (GSMaP) and Climate Forecasting System (CFSv2) and analyze the annual trends and spatial patterns of reanalysis data in Indonesia to find out which areas have a better level of accuracy. The study area was in Indonesia, between 12°S and 10°N, and 90°E and 50°E. The dataset used in this study were ERA5, GSMAP, and CFSv2 from 2011to 2021. The research shows that ERA5 has a very good level of accuracy in the southern region of Indonesia. The results of this study can be used for further information for climate change and the hydrological cycle in Indonesia, especially in the use of precipitation data.

1 Introduction Indonesia is a tropical country with a lot of rainfall each year. The availability of accurate precipitation data, both temporally and spatially, has become an essential requirement in various fields [1, 2]. In meteorology, precipitation is typically defined as rain, snow, sleet, or hail falling from a cloud toward the ground [3]. As a result, precipitation is one of the most critical variables in weather and climate studies. Accurate and reliable precipitation data are essential for determining climate trends S. K. J. Wicaksana (B) Research Center for Oceanography, National Research and Innovation Agency, Jakarta, Indonesia e-mail: [email protected] I. Sofiati Research Center for Climate and Atmosphere, National Research and Innovation Agency, Bandung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_72

781

782

S. K. J. Wicaksana and I. Sofiati

and variability, including impact indices like floods and droughts, as well as water resource management in agriculture, forestry, and energy sectors, as well as weather, climate, and hydrologic forecasting [4–6]. A deep understanding of these processes and phenomena requires precipitation data with high precision and spatial–temporal resolution. At various spatial and temporal scales, various global rainfall products are available. To calculate rainfall estimates, these products take very different approaches. They have classified into three types: (1) reanalysis, which is based on a numerical weather prediction model and data assimilation; (2) gage-only products, which are derived solely from gage data; and (3) satellite-based products, which are based solely or partially on satellite data. All of these products differ in some way. They each have distinct advantages and disadvantages [7]. As a result, precipitation products based on satellite imagery and reanalysis have become the most promising data, outperforming calibrated meteorological data and weather radar data. Reanalysis datasets generally display greater variability than other types of datasets, such as satellite- or rain gage-based datasets, with the degree of variability varying by region [4, 6, 8, 9]. Satellite-based precipitation retrievals and precipitation estimates from models/reanalysis systems have become increasingly appealing and accessible in recent decades. Some of satellite based and reanalysis precipitation products, such as Tropical Rainfall Measuring Mission (TRMM) Multi-Satellite Precipitation Analysis (TMPA), the Global Satellite Mapping of Precipitation (GSMaP), the National Oceanic and Atmospheric Administration (NOAA), the Integrated Multisatellite Retrievals for Global Precipitation Measurement, and a new global MultiSource Weighted-Ensemble Precipitation (MSEWP), have been developed. Reanalysis precipitation data, which is simulated using new atmospheric models combined with advanced data assimilation systems, are an important alternative [4, 10, 11]. Because observed datasets are not always available or present, reanalysis data has emerged as a viable alternative. Reanalysis is not dependent on the density of surface observational networks and can provide surface variables in areas with little surface coverage. As a result, reanalysis systems provide nearly realistic atmospheric circulation fields, allowing many things to be done, including understanding precipitation changes through the lens of atmospheric dynamical mechanisms. There are several different reanalysis products available, but their quality is known to vary with frequent upgrades. The most popular reanalysis precipitation datasets include the National Centers for Environmental Prediction Remote Sens. 2020, 12, 2902 3 of 25 reanalysis (NCEP1 and NCEP2), NCEP Climate Forecast System Reanalysis (CFSR), ECMWF Reanalysis-5 (ERA-5) and ERA-Interim, the Japanese 55-year Reanalysis Project (JRA-55), and the National Aeronautics and Space Administration (NASA) Modern Era Reanalysis for Research and Applications (e.g., MERRA-2) [11, 12]. The purposes of this study are to (1) evaluate the performance of ERA5 precipitation data against GSMAP and CFSv2, (2) analyze the annual trends and spatial patterns of ERA5 precipitation reanalysis data in Indonesia to find out which areas

Evaluation of ERA5 Precipitation Reanalysis Data in Indonesia

783

have a better level of accuracy. The results are critical for the continued rational application of reanalysis data and information on climate change and the hydrological cycle in Indonesia.

2 Data and Methodology 2.1 Study Area The study area was in Indonesia, between 12° S and 10° N, and 90° E and 50° E. Indonesia’s tropical climate is influenced by monsoon winds that change every six months. Indonesia has a rainy season from November to April and a dry season from May to October [13].

2.2 Dataset For this study, three satellite-based rainfall data products and reanalyzes were chosen: ERA5, CFSV2, and GSMAP. Because of the availability of data, the monthly precipitation from 2011 to 2021 was selected to be evaluated, and to prevent missing information and change the values of the other products, a nearest neighbor interpolation method with a spatial resolution of 0.25° was used. Precipitation data converted to one scale accumulations in monthly means of daily means has been scaled to include units such as “per day,” respectively. The data used in this study are detailed below. ERA5 ERA5 is the Copernicus Climate Change Service’s latest generation ECMWF atmospheric reanalysis of global climate. Using advanced modeling and data assimilation systems, it transforms vast amounts of historical data into global estimates. Since 1979, and eventually, since 1950, estimates of a large number of atmospheric, land, and oceanic climate variables have been available on a 0.25° × 0.25° global grid. For the high-resolution realization of ERA5, which assimilates observations twice daily and provides hourly output forecast steps ranging from 0 to 18 h, the model time step is 12 min. Precipitation observations are not assimilated into the model; instead, forecast accumulations for hourly precipitation fields are generated. Data from ERA5 can be downloaded from (https://cds.climate.copernicus.eu/) [5, 14]. GSMaP GSMap is one type of SPP developed by Japanese scientists. Initially, this project was developed by the Japan Science and Technology Agency (JST). However, the Japan Aerospace Exploration Agency is currently carrying it on (JAXA). GSMaP provides global rainfall data from microwave and infrared radiometers. These data cover the

784

S. K. J. Wicaksana and I. Sofiati

entire globe (from 60° N to 60° S) and have a temporal resolution of one hour. The data is available from 2000 to the present. GSMaP data can be downloaded from the JAXA website or by visiting (ftp://hokusai.eorc.jaxa.jp/standard/v7/) [4, 15, 16]. CFSv2 NCEP CFSv2 is the primary operational sub-seasonal forecast system used in the United States. CFSv2 is a fully coupled atmosphere–ocean modeling system. The spatial coverage spans the globe (90° N to 0° S) and generates hourly data with a ½° horizontal resolution (approximately 56 km). The data is available from 2011 to present. CFSv2 includes 16 runs per day, but only four runs were initialized daily during the reforecast period. These consist of four 9-month runs initialized (at 0000, 0600, 1200, and 1800 UTC) every five days, and a 0000 UTC one-season (3-month) run with three subsequent 45-day runs (0600, 1200, and 1800 UTC) initialized during the intervening four days. CFSv2 data can be downloaded from the NCAR website or by visiting (https://rda.ucar.edu/datasets/ds094.2/) [17].

2.3 Evaluation Method and Criteria In this study, the evaluation of precipitation products was performed with calibration and validation for each month (January-December). The calibration phase uses regression analysis with a scatter plot. At this stage, it can be seen the relationship between reanalysis data and observation data. Meanwhile, the calibration model’s accuracy is tested at the validation stage using the parameters root-mean-square error (RMSE 0, + ∞) and Pearson correlation coefficient (CC –1, 1). The correlation coefficient (CC) and root mean square error (RMSE) measure the quantitative accuracy of the precipitation products. The correlation coefficient (CC) represents the relationship between observed and estimated precipitation. The magnitude of the difference between observed and estimated rainfall is denoted by the RMSE, which assigns greater weight to significant errors. These performance statistics are expressed in the following ways:   N 1  RMSE =  (Si − G i )2 N i=1

(1)

N 

  Si − S G i − G CC = 2  N  2 N  S − S i i=1 i=1 G i − G i=1

(2)

where N is the total number of daily records, Si and G i are the estimated and gaged precipitation, respectively, S and G are their mean values.

Evaluation of ERA5 Precipitation Reanalysis Data in Indonesia

785

3 Results 3.1 Comparison of Precipitation Products Comparison of precipitation products with each other can help identify the consistency and differences in precipitation estimates among other precipitation products. Figure 1 shows the spatial–temporal distribution of the three precipitation products (ERA5 (A), GSMaP (B), and CFSv2 (C)). Overall, the spatial pattern of the mean annual rainfall of these products is similar. The results show that in general the performance of ERA5 with GSMaP, it looks underestimated in the waters of the Indian Ocean west of Sumatra, waters north of Kalimantan, and waters north of Papua. As for other regions, ERA5 can capture spatial patterns that are more similar to GSMaP than CFSv2. Furthermore, the ERA5 and CFSv2 data show underestimation (alongside the waters of Sumatra, critical Sulawesi, and maximum of the waters of Papua) and overestimation (withinside the Papuan plains). Meanwhile, for CFSv2 compared to GSMaP, the dominant overestimate is in almost all mainland (large islands) such as Sumatra, Kalimantan, parts of Sulawesi, Java, and most of Papua. This phenomenon is closely related to the influence of the highlands on each island which will produce high-intensity orographic rainfall. However, for waters in the central region and some islands (such as West Nusa Tenggara, East Nusa Tenggara, and Maluku) the performance of CFSv2 looks closer to GSMaP [4].

3.2 Basic Statistics and Distribution of the Precipitation Products The accuracy of ERA5 data against GSMAP and CFSv2 data was further analyzed statistically by calculating Root Mean Square Error (RMSE) and determination (R2) (Fig. 2). The linear correlation between ERA5 and GSMaP looks better (with a lower RMSE value of 3.949, a higher R2 of 0.740) compared to CFSv2 (RMSE 4.799, and R2 of 0.725). However, ERA5 has good (strong) linearity to (GSMaP) and (CFSv2) with the correlation coefficient (r) being (0.86) and (0.85), respectively. There are three categories of grouping on the R2 value, namely, the strong category, moderate category, and weak category. [18] stated that the R2 value of 0.75 belongs to the strong category, the R2 value of 0.50 belongs to the moderate category and the R2 value of 0.25 belongs to the weak category. R2 can not only be used in regression, but can use the R2 formula in all models to determine whether or not the model is in addition to strengthening the model that has been obtained [19]. Figure 3 shows the temporal pattern of annual rainfall from the three ERA5, GSMAP, and CFSv2 data were analyzed for 11 years, average for all of Indonesia, and the results are shown in Fig. 3. Generally, the annual pattern for ERA5, GSMAP, and CFSv2 look the same, with only slightly different intensity. The highest

786

S. K. J. Wicaksana and I. Sofiati

Fig. 1 Multiyear mean daily total precipitation from 2011 to 2021. (A) ERA5, (B) GSMAP, and (C) CFSv2

average rainfall occurs during the wet season (December-January–February, DJF) at 10.3 mm/day, and the lowest during the dry season (June-July–August) at about 6.2 mm/day. The annual cycle of ERA5 and GSMaP is nearly identical, except for January, June, July, September, October, and December, when ERA5 is slightly (1 mm/day) lower than GSMaP. Meanwhile, the annual cycle of CFSv2 is above ERA5 and GSMaP. Based on the results of the previous processing and analysis (Fig. 1), this also supports the finding that the CFSv2 data is overestimated for both ERA5 and GSMaP. To make things apparent, annual cycle analysis was applied to several areas by following the type of rainfall from [20]. Three regions were studied marked with

Evaluation of ERA5 Precipitation Reanalysis Data in Indonesia

787

Fig. 2 Scatter plot of (a) ERA5 vs. GSMAP and (b) ERA5 versus. CFSv2 over the period 2011– 2021

Fig. 3 Annual cycle of daily total precipitation rates in Indonesia for ERA5, GSMAP, and CFSv2 region averaged over the period 2011–2021

red boxes: Region.1 (6–10) °S, (105–110) °E, Region II (2–4) °N, (95–100) °E, and Region III (4–5) °S, (125–130) °E, as shown in Fig. 4. In Region I, the highest average rainfall occurs during the wet season (DecemberJanuary–February, DJF) at 14.1 mm/day, and the lowest during the dry season (JuneJuly–August, JJA) at around 1.1 mm/day. The annual cycle of ERA5 and GSMaP is nearly identical, except for January, June, July, September, October, and December,

788

S. K. J. Wicaksana and I. Sofiati

Fig. 4 location of rainfall type based on [20]

when ERA5 is slightly (1 mm/day) lower than GSMaP. Meanwhile, the annual cycle of CFSv2 is above ERA5 and GSMaP. For Region II, there are two peaks of rainfall, namely in April and November as shown in Fig. 6. The highest average rainfall occurs twice, during the wet-to-dry transition season (March–April-May, MAM) of 13.1 mm/day, and the dry-to-wet transition (September–October-November, SON) was about 18.3 mm/day. Meanwhile, the lowest average was in June which only reached 6.7 mm/day. Furthermore, for Region III as shown in Fig. 7, the highest average rainfall value occurs in the month (May, June, July, MJJ) at 13.8 mm/day, and the lowest average is in the month (August–September-October, ASO) about 2.5 mm/day. From the three

Fig. 5 Annual cycle of daily total precipitation rates in the region I (6–10) °S, (105–110) °E

Evaluation of ERA5 Precipitation Reanalysis Data in Indonesia

789

Fig. 6 Annual cycle of daily total precipitation rates in region II (2–4) °N, (95–100) °E

selected regions, the intensity of the annual cycle of ERA5 and GSMaP is almost the same, while the CFSv2 is above ERA5 and GSMaP. In Fig. 8 can be seen that for both ERA5 versus. GSMAP and ERA5 versus. CFSv2, a solid and positive correlation (correlation coefficient value can reach one) mainly occurs in the southern and northern parts of Indonesia, for the equator area itself a strong correlation on average occurs in inter-island waters. Although there are some points in the Papua region that show a negative correlation, it is not as strong.

790

S. K. J. Wicaksana and I. Sofiati

Fig. 7 Annual cycle of daily total precipitation rates in region III (4–5) °S, (125–130) °E

Fig. 8 Spatial Correlation total precipitation over the period 2011–2021 a ERA5 versus. GSMAP and b ERA5 versus. CFSv2

Evaluation of ERA5 Precipitation Reanalysis Data in Indonesia

791

4 Conclusion Three precipitation products were evaluated over the Indonesia region from 2011 to 2021. The main findings are as follows. In general, the spatial pattern for ERA5, GSMAP, and CFSv2 looks similar, only slightly different intensity, and the value of CFSv2 has a higher value (overestimate) than the other two products. The evaluation results show that ERA5 is well-validated against two other products. The RMSE between ERA5 to GSMAP and CFSv2, respectively, is the value of 3.949 and 4.799 mm/day. The temporal pattern of annual rainfall from the three looks the same, only the intensity is slightly different. A comparison also made in three locations, and it shows that the three products had the same pattern as the research conducted by [20]. In general, the ERA5 precipitation product has a high level of accuracy and can be used as hydrological data in Indonesia, particularly in the southern region.

References 1. Arrokhman, N.A., Wahyuni, S., Suhartanto, E.: Evaluasi Kesesuaian Data Satelit untuk Curah Hujan dan Evaporasi Terhadap Data Pengukuran di Kawasan Waduk Sutami. J. Teknol. dan Rekayasa Sumber Daya Air. 1(2), 904–916 (2021) 2. Lee, H.S.: General rainfall patterns in Indonesia and the potential impacts of local seas on rainfall intensity. Water (Switzerland). 7(4), 1751–1768 (2015) 3. Siepielski, A.M., Morrissey, M.B., Buoro, M., Carlson, S.M., Caruso, C.M., Clegg, S.M., Coulson, T., DiBattista, J., Gotanda, K.M., Francis, C.D., Hereford, J., Kingsolver, J.G., Augustine, K.E., Kruuk, L.E.B., Martin, R.A., Sheldon, B.C., Sletvold, N., Svensson, E.I., Wade, M.J., MacColl, A.D.C.: Precipitation drives global variation in natural selection. Science (80-). 355 (6328), 959–962 (2017) 4. An, Y., Zhao, W., Li, C., Liu, Y.: Evaluation of six satellite and reanalysis precipitation products using gauge observations over the yellow river basin, China. Atmosphere (Basel), 11 (11) (2020) 5. Hassler, B., Lauer, A.: Comparison of reanalysis and observational precipitation datasets including era5 and wfde5. Atmosphere (Basel), 12 (11) (2021) 6. Sun, Q., Miao, C., Duan, Q., Ashouri, H., Sorooshian, S., Hsu, K.L.: A review of global precipitation data sets: data sources, estimation, and intercomparisons. Rev. Geophys. 56(1), 79–107 (2018) 7. Le Coz, C., Van De Giesen, N.: Comparison of rainfall products over Sub-saharan Africa. J. Hydrometeorol. 21(4), 553–596 (2020) 8. Bai, P., Liu, X.: Evaluation of Five Satellite-Based Precipitation Products in Two Gauge-Scarce Basins on the Tibetan Plateau. Remote Sens. 2018, Vol. 10, Page 1316. 10 (8), 1316 (2018) 9. Okamoto, K., Takahashi, N., Iwanami, K., Shige, S., Kubota, T.: High precision and high resolution global precipitation map from satellite data. 2008 Microw. Radiom. Remote Sens. Environ.—10th Spec. Meet. Proceedings, MICRORAD. (2008) 10. Dee, D.P., Uppala, S.M., Simmons, A.J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M.A., Balsamo, G., Bauer, P., Bechtold, P., Beljaars, A.C.M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A.J., Haimberger, L., Healy, S.B., Hersbach, H., Hólm, E.V., Isaksen, L., Kållberg, P., Köhler, M., Matricardi, M., Mcnally, A.P., Monge-Sanz, B.M., Morcrette, J.J., Park, B.K., Peubey, C., de Rosnay, P., Tavolato, C., Thépaut, J.N., Vitart, F.: The ERA-Interim reanalysis: configuration and performance of the data assimilation system. Q. J. R. Meteorol. Soc. 137(656), 553–597 (2011)

792

S. K. J. Wicaksana and I. Sofiati

11. Sun, S., Shi, W., Zhou, S., Chai, R., Chen, H., Wang, G., Zhou, Y., Shen, H.: Capacity of satellitebased and reanalysis precipitation products in detecting long-term trends across Mainland China. Remote Sens. 12(18) (2020) 12. Wanzala, M.A., Ficchi, A., Cloke, H.L., Stephens, E.M., Badjana, H.M., Lavers, D.A.: Assessment of global reanalysis precipitation for hydrological modelling in data-scarce regions: a case study of Kenya. J. Hydrol. Reg. Stud. 41, 101105 (2022) 13. Aldrian, E., Gates, L.D., Widodo, F.H.: Variability of Indonesian Rainfall and the Influence of ENSO and Resolution in ECHAM4 Simulations and in the reanalyses. MPI Rep. (346) (2003) 14. Qin, S., Wang, K., Wu, G., Ma, Z.: Variability of hourly precipitation during the warm season over eastern China using gauge observations and ERA5. Atmos. Res. 264(September), 105872 (2021) 15. Kubota, T., Shige, S., Hashizume, H., Aonashi, K., Takahashi, N., Seto, S., Hirose, M., Takayabu, Y.N., Ushio, T., Nakagawa, K., Iwanami, K., Kachi, M., Okamoto, K.: Global precipitation map using satellite-borne microwave radiometers by the GSMaP project: production and validation. IEEE Trans. Geosci. Remote Sens. 45(7), 2259–2275 (2007) 16. Pratama, A., Agiel, H.M., Oktaviana, A.A.: Evaluasi satellite precipitation product (GSMaP, CHIRPS, dan IMERG) di Kabupaten lampung selatan. J. Sci. Appl. Technol. 6(1), 32 (2022) 17. Saha, S., Moorthi, S., Wu, X., Wang, J., Nadiga, S., Tripp, P., Behringer, D., Hou, Y.T., Chuang, H.Y., Iredell, M., Ek, M., Meng, J., Yang, R., Mendez, M.P., Van Den Dool, H., Zhang, Q., Wang, W., Chen, M., Becker, E.: The NCEP climate forecast system version 2. J. Clim. 27(6), 2185–2208 (2014) 18. Joseph, F., Hair, J.R.E.A.R.L.T., W.C.B.: Multivariate Data Analysis Fifth Edition. Prentice Hall, Inc. (1998) 19. Ghozali, I.: Aplikasi Analisis Multivariate Dengan Program IBM SPSS 25. Universitas Diponegoro, Semarang (2018) 20. Aldrian, E., Dwi Susanto, R.: Identification of three dominant rainfall regions within Indonesia and their relationship to sea surface temperature. Int. J. Climatol. 23(12), 1435–1452 (2003) 21. Liu, C.Y., Aryastana, P., Liu, G.R., Huang, W.R.: Assessment of satellite precipitation product estimates over Bali Island. Atmos. Res. 244(January), 105032 (2020) 22. Wati, T., Hadi, T.W., Sopaheluwakan, A., Hutasoit, L.M.: Statistics of the Performance of Gridded Precipitation Datasets in Indonesia. Adv. Meteorol. 2022 (2022)

Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis of Atmospheric Conditions During Tropical Cyclone Dahlia Dendi Rona Purnama , I. Nyoman Agus Astina Putra, Dewangga Palguna, and Gandhi Mahendra Abstract The decay of tropical cyclone Cempaka into a tropical depression was followed by the emergence of a new tropical low in the southwest of Bengkulu. This tropical low status was upgraded on 29th November 2017 at 19.00 LT into tropical cyclone Dahlia. In this study, an analysis of several weather parameters was carried out during the occurrence of tropical cyclone Dahlia, including wind conditions, divergence, sea surface temperature, geopotential height, air temperature, as well as the movement of water vapor transport on 30th November–1st December 2017. The influence caused by the presence of tropical cyclone Dahlia on atmospheric conditions is described in this study. This research utilizes ERA-Interim reanalysis data from the European Center for Medium-Range Weather Forecast (ECMWF). The method used is the descriptive method. The results showed that the sea surface temperature in the study area supports the formation of a tropical cyclone. The significant water vapor transport causes high water vapor intensity and low geopotential height during the research period. The cyclonic pattern is visible in the streamline, and the divergence conditions during the research period are generally negative in the lower tropospheric layers and positive in the upper tropospheric layers. The presence of MJO, Equatorial Rossby Waves, and Kelvin Waves in the same location and time supported the formation of TC Dahlia. D. R. Purnama (B) Public Weather Services, Indonesian Agency for Meteorology Climatology and Geophysics, Angkasa I St. No. 2, Kemayoran, Central Jakarta 10720, Indonesia e-mail: [email protected] I. N. A. A. Putra Kendari Marine Meteorological Station, Indonesian Agency for Meteorology Climatology and Geophysics, Jendral Sudirman St. No. 158, Kendari 93127, Indonesia D. Palguna Komodo Meteorological Station, Indonesian Agency for Meteorology Climatology and Geophysics, Yohanes Sehadun St., West Manggarai 86554, Indonesia G. Mahendra Sentani Meteorological Station, Indonesian Agency for Meteorology Climatology and Geophysics, Yabaso St., Sentani, Jayapura Region 99111, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_73

793

794

D. R. Purnama et al.

1 Introduction One of the most destructive hydrometeorological disasters on Earth is tropical cyclones (TCs). In some cases, 90% of tropical cyclones have a terrible impact. Tropical cyclones are capable of causing enormous damage to the areas they pass through [1]. This damage can reach a radius of 250 miles from the center of a tropical cyclone [2]. Effects resulting from tropical cyclones can change geomorphology and ecology, especially in water areas [3]. Significant wave heights around tropical cyclones will increase as the intensity of the tropical cyclones increases [4]. Technically, a tropical cyclone is defined as a synoptic-scale non-frontal lowpressure system that grows over warm waters with areas of convective cloud and a maximum wind speed of at least 34 knots in more than half of the area surrounding its center, and lasts at least six hours [5, 6]. This strong cyclonic circulation is formed in the season when the water conditions warm, and this circulation is in the mature stage of the tropical disturbance [7, 8]. A cyclonic circulation can develop into a tropical cyclone due to the Coriolis force, which is supported by several parameters. Gray divides these parameters into two categories, namely thermal parameters (thermal energy of the ocean with sea temperatures > 26 °C to a depth of 60 m, the equivalent potential temperature difference between the surface layer with 500 mb layer > 10 °C, and wet air with humidity > 70% in the 700–500 mb layers) and dynamic parameters (strong low-level vorticity, weak vertical wind shear, and latitude Coriolis parameter > 3°) [9]. Most tropical cyclones form at latitudes 10°–20° from the equator (65%), and only ± 13% of tropical cyclones occur at latitude 22° [2]. With an average tropical cyclone occurrence of 10 times per year, the Indian Ocean region and the waters of Western Australia are the areas where tropical cyclone growth occurs most frequently in the world [10]. The early appearance of most tropical cyclones in the waters of Western Australia or the Southeast Indian Ocean can be found around the intertropical convergence zone (ITCZ). McBride and Keenan [11] estimated that approximately 85% of cyclones in the Southeast Indian Ocean have their origin location near the ITCZ. Tropical cyclone activity around the territory of Indonesia can also be influenced by the presence of El Nino and La Nina phenomena [1, 12] as well as influenced by Indonesian Throughflow conditions [13]. Tropical cyclones have an essential role in atmospheric dynamics, especially in the upper troposphere and lower stratosphere [14]. Several studies on the atmospheric conditions during tropical cyclones have been widely studied using various kinds of data. Fibriantika and Alhaqq [15] have analyzed the vertical profile of the atmosphere on the occurrence of tropical cyclones Cempaka and Dahlia using radiosonde data, with the results showing that tropical cyclone Cempaka has a more significant effect on the condition of the atmospheric vertical profile on Java Island compared to tropical cyclone Dahlia, especially at vertical profile parameters of air temperature, dew point temperature, and humidity. Analysis of tropical cyclone Nock-Ten by Ismail et al. [16] using Himawari satellite data shows that at the time of the occurrence, a high convective index and strong winds were recorded at category five on the

Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis …

795

Saffir–Simpson scale. Using the Weather Research and Forecasting (WRF) model, the occurrence of tropical cyclone Haiyan was analyzed on the parameters of wind speed, air humidity, sea surface temperature, surface pressure, and rainfall [17]. In the study of Samrin et al. [18], ECMWF ERA-Interim reanalysis data, Himawari-8 satellite data infrared channel 1 (IR1), and Global Satellite Mapping of Precipitation (GSMaP) data were used to analyze atmospheric conditions and rainfall in tropical cyclone Cempaka. Various types of modeling data have been widely used to simulate atmospheric conditions, both for analysis of past events, re-create past conditions (hindcast or back-cast), as well as for predicting future atmospheric conditions (forecast). The development of reanalysis data products started in the 1990s by the European Center for Medium-Range Weather Forecast (ECMWF) and the National Centers for Environmental Prediction (NCEP) has played a vital role in simulating future trends and scenarios related to climate change studies [19]. In de Lima and Alcântara’s research [20], reanalysis data from the National Centers for Environmental Prediction/National Center for Atmospheric Research (NCEP/NCAR), reanalysis data from Climate Forecast System Reanalysis (CFSR), and ECMWF ERA-Interim reanalysis data were compared to find out how much error the three data generated against the observed data. The results show that the ERA-Interim ECMWF reanalysis data gives the smallest error values for the precipitation and temperature parameters. The presence of tropical cyclone Dahlia seed also coincided with the collapse of tropical cyclone Cempaka on 29th November 2017 into a tropical depression. This tropical cyclone seed was upgraded to a tropical cyclone status at 19.00 LT and was at a position of 8.2 °S and 100.8 °E. Because the seeds of the cyclone appeared in the area of responsibility of the Tropical Cyclone Warning Center (TCWC) Jakarta, TCWC Jakarta has the right to give a name to the tropical cyclone, hereinafter referred to as tropical cyclone Dahlia. The purpose of this study was to analyze the atmospheric conditions related to the tropical cyclone Dahlia phenomenon by utilizing the ERA-Interim ECMWF reanalysis data. The effect caused by the presence of tropical cyclone Dahlia on the conditions of the lower and upper layers of the atmosphere will be described in this study.

2 Data and Method 2.1 Time and Location The research location is the waters of the Southeast Indian Ocean in the southern part of Sumatra and Java at coordinates 0°–20°S and 90°E–130°E (Fig. 1). Research time is limited to 30th November–1st December 2017.

796

D. R. Purnama et al.

Fig. 1 Research location

2.2 Data The data used in this study is the ECMWF ERA-Interim reanalysis data downloaded from https://www.ecmwf.int, which has a spatial resolution of 0.125° × 0.125° and a temporal resolution of 6 hours [21, 22]. Some of the parameters used to analyze the atmospheric conditions at the time of the tropical cyclone Dahlia were the zonal wind component (u), the meridional wind component (v), divergence, air temperature, and geopotential height at 850, 700, 500 and 200 mb layers. In addition, this study also used data on specific humidity of 1000, 700, 500 and 300 mb layers as well as sea surface temperature from ECMWF ERA-Interim. All data was downloaded to coincide with 30th November–1st December 2017.

2.3 Method The descriptive method was used in this research. The ERA-Interim ECMWF reanalysis data was processed using the Grid Analysis and Display System (GrADS) application [23] to display weather parameters spatially during the tropical cyclone Dahlia. Furthermore, descriptive analysis was carried out on several weather parameters, including analyzing streamline, divergence values, sea surface temperature conditions, analyzing geopotential height and air temperature, and water vapor transport movements. The distribution of water vapor transport is calculated by dividing it into several layers, namely 1000–300, 1000–700, 700–500 and 500–300 mb.

Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis …

797

Fig. 2 Track of tropical cyclone Dahlia [24]

3 Result and Discussion 3.1 Tropical Cyclone Dahlia Movement Tropical cyclone Dahlia began with the formation of a low-pressure pattern in the Indian Ocean west of Sumatera on 25th November 2017 and moved the southeast (Fig. 2). A tropical cyclone began to form on 29th November 2017, which was named tropical cyclone Dahlia by TCWC Jakarta. This tropical cyclone moved southeast to extinction in the waters south of Java on 3rd December 2017.

3.2 Sea Surface Temperatures Analysis The sea surface temperature that supports the growth of a tropical cyclone is more than 26 °C [2, 9, 25]. In Fig. 3, the sea surface temperature on 30th November 2017 to 1st December 2017 in the southern waters of Java has temperatures ranging from 27.5–29 °C. This indicates that the region has warm sea surface temperatures and supports the occurrence of a tropical cyclone.

798

D. R. Purnama et al.

Fig. 3 Sea surface temperature on a 30th November 2017 and b 1st December 2017

3.3 Streamline Analysis Streamlines are lines that are parallel to the horizontal wind speed vector at a certain height and time [26]. Figure 4 shows the streamline of tropical cyclone Dahlia that occurred in the southern hemisphere (SH) on 30th November 2017 and 1st December 2017 at 00.00 UTC or 07.00 LT at 850, 700, 500 and 200 mb layers. On 30th November 2017 at the 850, 700 and 500 mb layers, a slightly elongated cyclonic circulation pattern was seen in the Southeast Indian Ocean (Fig. 4a). The wind speed around the center of the cyclone tends to increase with increasing altitude, while the wind speed at the center of the cyclone is relatively the same in all layers. At the 200 mb layer, the cyclonic wind circulation pattern has changed to anticyclonic winds. Because it is an anticyclonic region, the 200 mb layer is the layer where the air mass exits tropical cyclone Dahlia. On 1st December 2017 which was the peak of activity for the tropical cyclone Dahlia, the cyclonic wind circulation pattern was not much different from the previous day, but became more visible (Fig. 4b). The intensity of the wind speed around the center of the cyclone also increased more than the previous date in all layers.

3.4 Geopotential Height and Air Temperature Analysis Atmospheric pressure can be expressed as a function of altitude with the following equation [27]: dp = ρgdz or

(1)

Fig. 4 Streamline in 850, 700, 500 and 200 mb layers at 00.00 UTC on a 30th November 2017 and b 1st December 2017. The unit of wind speed is meters per second (mps)

Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis … 799

800

D. R. Purnama et al.

dp = −ρd∅

(2)

where ∅ = gz is called the geopotential and shows the magnitude of the potential energy at a height z. In the gas equation, ρ = p/RT , which shows that p is a function of pressure and T is a function of temperature. Thus, geopotential changes in pressure at a certain altitude layer are related to changes in temperature [27]. Figure 5 shows the geopotential height and air temperature on 30th November–1st December 2017. On 30th November 2017, in the 850 mb and 700 mb layers, two nearly converging low geopotential areas were seen in the southwest of Java Island (Fig. 5a). In addition, there is another low geopotential area to the east. In the 500 mb layer, there is only one low geopotential area. Temperatures tend to be warmer around low geopotential areas. The low geopotential height indicates that the area is a low-pressure area, and the air movement tends to move up to the upper layers. Two low geopotential areas in the 850 and 700 mb layers merged into one on 1st December 2017, with the geopotential height becoming progressively lower (Fig. 5b). Likewise, at the 500 mb layer, the geopotential height is getting lower. However, the temperature tends to be the same as the previous day. This condition can indicate the movement of air moving up, and the area is getting lower pressure. In contrast to the layer below, at the 200 mb layer, there is no closed isogeopotential lower than the surrounding area. This indicates that the air movement is no longer converging in this layer, but tends to show a strong divergence of air masses.

3.5 Divergence Analysis The existence of divergence and convergence air mass flow patterns can cause vertical movement [28]. Divergence is an area of air mass distribution caused by air currents spreading over an area. The condition of the area with a negative divergence indicates the presence of air mass compression (convergence) in the area and supports the growth of convective clouds that cause heavy rains [29]. Figure 6 shows the condition of the divergence value on 30th November–1st December 2017. Based on Fig. 6a, on 30th November 2017, in the 850 mb and 700 mb layers, the divergence value around the tropical cyclone Dahlia area, which is in the south of Java Island, is generally negative (−), which indicates the opportunity for the growth of rain-producing clouds in the region is quite large. At the 500 mb layer, there is a positive (+) divergence value which is squeezed by negative (−) divergence values, this indicates that there is a pattern of air movement that is starting to spread. The 200 mb layer is dominated by positive (+) divergence values, which are quite large and indicate that the air in this layer has been dominated by a diffused air pattern. In the 850 mb and 700 mb layers on 1st December 2017, in the area where tropical cyclone Dahlia was formed, there was a low divergence value with a wider area coverage than the previous day (Fig. 6b). This indicates that there is a stronger compression of the air masses moving upwards than the previous day. Furthermore, in the 500 mb layer, the tropical cyclone Dahlia area begins to be dominated by

Fig. 5 Geopotential height (solid line) and air temperature (shaded) in the 850, 700, 500 and 200 mb layers at 00.00 UTC on a 30th November 2017 and b 1st December 2017. The unit of geopotential height is meter

Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis … 801

Fig. 6 Divergence of layers of 850 mb, 700 mb, 500 mb, and 200 mb at 00.00 UTC on a 30th November 2017 and b 1st December 2017

802 D. R. Purnama et al.

Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis …

803

positive (+) divergence values. This indicates that the upward-moving air mass begins to spread in the 500 mb layer. In the 200 mb layer, the tropical cyclone Dahlia area shows a positive divergence value which indicates that in the 200 mb layer the air masses are spreading.

3.6 Water Vapor Transport Analysis Water vapor transport can have different values and directions in each layer of the atmosphere [30]. Water vapor transport has an important role in analyzing the impact of some weather disturbances on rainfall. One widely adopted approach to determine the characteristics of water vapor transport is to calculate vertically integrated water vapor transport fluxes (IVT) between the pressure in the surface layer and the pressure limit at the highest altitude from a radiosonde measurement or other upper air condition measurement [31]. IVT can be calculated by ⎛ ⎞2 ⎛ ⎞2  p p  1 ⎝ 1 qudp ⎠ + ⎝ qvdp ⎠ IVT =  g g p0

(3)

p0

where q: u: v: p0 : p:

specific humidity (kg/kg) zonal wind component (m/s) meridional wind component (m/s) pressure in the lower boundary layer or surface layer pressure (mb) pressure in the upper boundary layer or the pressure of the highest layer to be analyzed (mb).

The product components q × u × dp/g and q × v × dp/g at each grid point are summed vertically from the surface layer (bottom layer pressure limit) to a certain layer (top layer pressure limit) which are then combined into a horizontal transport vector, where the unit of IVT is kg m¯1 s¯1 [31, 32]. For example, if you want to calculate IVT in 1000–500 mb layers, then p0 can be replaced with 1000, and p can be replaced with 500. In this study, the water vapor transport analyzed is between layers 1000–300 mb (overall troposphere), 1000–700 mb (lower troposphere), 700– 500 mb (middle troposphere), and 500–300 mb (upper troposphere). Figure 7 shows water vapor transport on 30th November–1st December 2017. In Fig. 7a, water vapor transport in the 1000–300 mb layer on 30th November 2017 shows a movement of water vapor from north to south which indicates negative meridional winds and positive zonal winds. This indicates the presence of Asian monsoon winds moving to SH carrying water vapor of 800 to more than 1600 kg m¯1 s¯1 to the southeastern part of Java Island. The high intensity of water vapor has an effect on increasing rainfall around the tropical cyclone Dahlia area. In the 1000–700

Fig. 7 Water vapor transport between 1000–300 mb, 1000–700 mb, 700–500 mb, and 500–300 mb at 00.00 UTC on a 30th November 2017 and b 1st December 2017

804 D. R. Purnama et al.

Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis …

805

mb layer, the movement of water vapor tends to be the same as the 1000–300 mb layer, namely from the northwest of Sumatra Island to the southeast. Water vapor transport in this layer ranges from 400 to 1000 kg m¯1 s¯1 . The cyclonic pattern has been formed as in the previous layers. The distribution of water vapor in the 700–500 mb layers shows a fairly high water vapor, with the intensity of water vapor moving from along the northwest of Sumatra Island to the southeast in the range of 400–1200 kg m¯1 s¯1 . Likewise, in the 500–300 mb layer, the water vapor transport pattern is the same as the previous layer, but with less intensity (100–800 kg m¯1 s¯1 ). In Fig. 7b, in the 1000–300 mb layer, the water vapor is concentrated in several areas. Water vapor is concentrated mainly around the center of the cyclonic pattern (south of Java Island) and west of Sumatra Island with an intensity reaching more than 1600 kg m¯1 s¯1 . The water vapor mass movement pattern is the same as the previous day. The condition of the water vapor transport pattern in the 1000–700 mb layer is similar to the 1000–300 mb layer. The intensity in this layer tends to be lower than in the previous layer, which is 400–1000 kg m¯1 s¯1 in the tropical cyclone Dahlia area. In the 700–500 mb layer, the pattern water vapor transport moves from the northeast (in the South China Sea) and meets in the western area of Sumatra Island, flowing southeast to enter the territory of Indonesia. The cyclonic pattern is still visible in the south of Java Island. The concentration of water vapor occurs in the Sunda Strait and around the center of a tropical cyclone with an intensity of 400– 1000 kg m¯1 s¯1 . In the 500–300 mb layer, the conditions are similar to the 700–500 mb layer with a lower intensity than the previous three layers, which is 200–1000 kg m¯1 s¯1 in the area of concentration of water vapor.

3.7 Influence of Tropical Waves In addition to the thermal and dynamic parameters required for TCs growth [9], another potential factor causing TCs formation is the modulation by tropical waves. In conditions before, during, and after TC Dahlia, the Madden–Julian Oscillation (MJO) was active in quadrant 4 (maritime continent). Equatorial Rossby (ER) was also observed in the area where TC Dahlia grew, both before and during the occurrence of TC Dahlia. In addition, Kelvin Waves were also observed in the western part of Indonesia accompanied by the presence of low-frequency waves (Fig. 8). The presence of MJO and ER is known to have a significant large modulation of TC genesis over the south Indian Ocean [35–37]. Meanwhile, Kelvin Waves have a significant small modulation [36]. With the presence of MJO and ER at the same time, the modulation was best attributable to its vorticity, convection, and shear fields. Thus, the presence of MJO, ER, and Kelvin Waves in the same location and time could be a supporting factor in the formation of TC Dahlia.

806

D. R. Purnama et al.

Fig. 8 Outgoing Longwave Radiation (OLR) in the 850 mb layer combines with line contours of MJO, Equatorial Rossby waves, Kelvin waves, and the low-frequency waves [33, 34]

4 Conclusion The results of the analysis of the occurrence of tropical cyclone Dahlia using the ECMWF ERA-Interim reanalysis data showed that, in general, the data was able to represent atmospheric conditions that were in accordance with the conditions required for the formation of a tropical cyclone. The use of this data is able to describe atmospheric conditions, both vertically and horizontally, especially related to atmospheric conditions at the occurrence of the tropical cyclone Dahlia. The presence of MJO, ER, and Kelvin Waves in the same location and time supported the formation of TC Dahlia. The analysis of the atmospheric conditions at tropical cyclone Dahlia showed that both cyclonic patterns and convergence conditions were found in the lowerto-middle troposphere layer (500 mb layer and below), both in streamline analysis, air temperature, isogeopotential, and divergence. Meanwhile, the upper troposphere shows an anticyclonic and divergence pattern. This means that in the upper troposphere, the movement of air masses is spread, especially in the 200 mb layer. In addition, the water vapor transport analysis showed a mass accumulation of water vapor around tropical cyclone Dahlia. The distribution of water vapor from the South China Sea indicates the activity of the Asian monsoon winds moving to the southern hemisphere.

References 1. Purnama, D.R., Zulistyawan, K.A., Christian, B., Veanti, D.P.O.: Dampak terjadinya el nino/la nina terhadap intensitas, masa hidup dan frekuensi siklon. Jurnal Meteorologi Klimatologi dan Geofisika 5(2), 10–21 (2018) 2. Tjasyono, B.: Klimatologi Umum. Penerbit ITB, Bandung (2004)

Utilization of ECMWF ERA-Interim Reanalysis Data for Analysis …

807

3. Done, T.J.: Effects of tropical cyclone waves on ecological and geomorphological structures on the Great Barrier Reef. Cont. Shelf Res. 12(7–8), 859–872 (1992) 4. Haryanto, Y.D., Riama, N.F., Purnama, D.R., Sigalingging, A.D.: The effect of the difference in intensity and track of tropical cyclone on significant wave height and wave direction in the Southeast Indian Ocean. The Scient. World J. 5492048, 11 (2021) 5. Badan Meteorologi Klimatologi dan Geofisika.: http://web.meteo.bmkg.go.id/id/component/ content/article/37-siklon-tropis/268-siklon-tropis. Last Accessed 01 Nov 2021 6. Zakir, A., Sulistya, W., Khotimah, M.K.: Perspektif operasional cuaca tropis. Pusat Penelitian dan Pengembangan BMKG, Jakarta (2010) 7. Saha, K.: Tropical Circulation Systems and Monsoons. Springer, Heidelberg (2010) 8. Kepert, J.D.: Tropical cyclone structure and dynamics. In: Global Perspectives on Tropical Cyclones, pp. 3–53. (2010) 9. Gray, W.M.: Hurricanes: their formation, structure and likely role in the tropical circulation. Meteorol. Over the Tropical Oceans 155, 218 (1979) 10. Haryani, N.S., Zubaidah, A.: Dinamika siklon tropis di Asia Tenggara menggunakan data penginderaan jauh. Majalah Ilmiah Widya 29(324), 54–58 (2012) 11. McBride, J.L., Keenan, T.D.: Climatology of tropical cyclone genesis in the Australian region. J. Climatol. 2(1), 13–33 (1982) 12. Camargo, S.J., Sobel, A.H.: Western North Pacific tropical cyclone intensity and ENSO. J. Clim. 18(15), 2996–3006 (2005) 13. Pradita, N., Ismah, S.F., Purnama, D.R.: Pengaruh arlindo terhadap durasi dan intensitas siklon tropis di Pasifik Barat Laut dan Samudra Hindia Selatan. In: Seminar Nasional Bumi dan Atmosfer, pp. 7–15. Sekolah Tinggi Meteorologi Klimatologi dan Geofisika, South Tangerang (2019) 14. Ratnam, M.V., Babu, S.R., Das, S.S., Basha, G., Krishnamurthy, B.V., Venkateswararao, B.: Effect of tropical cyclones on the stratosphere-troposphere exchange observed using satellite observations over the north Indian Ocean. Atmos. Chem. Phys. 16(13), 8581–8591 (2016) 15. Fibriantika, E., Alhaqq, R.I.: Profil vertikal atmosfer selama aktifitas siklon tropis Cempaka dan Dahlia. Jurnal Meteorologi dan Geofisika 19(2), 49–58 (2019) 16. Ismail, P., Hidayat, N.M., Siadari, E.L.: Analisis siklon tropis Nock-Ten berbasis data Satelit Himawari. Jurnal Meteorologi Klimatologi dan Geofisika 4(3), 16–23 (2017) 17. Perawiska, E., Muliadi, M., Adriat, R.: Analisis unsur cuaca pada saat kejadian siklon tropis Haiyan menggunakan model WRF (weather research and forecasting). Prisma Fisika 6(2), 129–136 (2018) 18. Samrin, F., Irwana, I., Hasanah, N.: Analysis of the meteorological condition of tropical cyclone Cempaka and its effect on heavy rainfall in Java Island. In: IOP Conference Series: Earth and Environmental Science vol. 303(1), pp. 012065. (2019) 19. Houle, D., Bouffard, A., Duchesne, L., Logan, T., Harvey, R.: Projections of future soil temperature and water content for three southern Quebec forested sites. J. Clim. 25(21), 7690–7701 (2012) 20. de Lima, J.A.G., Alcântara, C.R.: Comparison between ERA Interim/ECMWF, CFSR, NCEP/NCAR reanalysis, and observational datasets over the eastern part of the Brazilian Northeast Region. Theoret. Appl. Climatol. 138(3), 2021–2041 (2019) 21. Berrisford, P., Dee, D., Poli, P., Brugge, R., Fielding, K., Fuentes, M., Kållberg, P., Kobayashi, S., Uppala, S., Simmons, A.: The ERA-interim archive version 2.0, European Centre for Medium Range Weather Forecasts, Reading (2011) 22. Dee, D.P., Uppala, S.M., Simmons, A.J., Berrisford, P., Poli, P., Kobayashi, S., Andrae, U., Balmaseda, M.A., Balsamo, G., P. Bauer, Bechtold, P., Beljaars, A.C.M., van de Berg, L., Bidlot, J., Bormann, N., Delsol, C., Dragani, R., Fuentes, M., Geer, A.J., Haimberger, L., Healy, S.B., Hersbach, H., Hólm, E.V., Isaksen, L., Kållberg, P., Köhler, M., Matricardi M., McNally, A.P., Monge-Sanz, B.M., Morcrette, J.-J., Park, B.-K., Peubey, C., de Rosnay P., Tavolato, C., Thépaut, J.-N., Vitart, F.: The ERA-Interim reanalysis: Configuration and performance of the data assimilation system. Quart. J. The Royal Meteorol. Soc. 137(656), 553–597 (2011)

808

D. R. Purnama et al.

23. Doty, B.E., Kinter III, J.L., Fiorino, M., Hooper, D., Budich, R., Winger, K., Schulzweida, U., Calori, L., Holt, T., Meier, K.: The grid analysis and display system (GrADS): An Update for 1997. In 13th International Conference on IIPS for Meteorology, Oceanography and Hydrology, American Meteorology Society, vol. 117 (1997) 24. Bureau of Meteorology.: http://www.bom.gov.au/cyclone/history/Dahlia.shtml. Last Accessed 01 Nov 2021 25. Wang, Z.: Tropical cyclogenesis. In: Encyclopedia of Atmospheric Sciences, vol. 6(2), pp. 57– 64. (2015) 26. Wallace, J.M., Hobbs, P.V.: In: Atmospheric Science: an Introductory Survey. Elsevier (2006). 27. Wirjohamidjojo, S., Swarinoto, Y.S.: Meteorologi sinoptik. Pusat Penelitian dan Pengembangan BMKG, Jakarta (2013). 27. Wirjohamidjojo, S., Swarinoto, Y.S.: Meteorologi sinoptik. Pusat Penelitian dan Pengembangan BMKG, Jakarta (2013) 28. Trewartha, G.T., Horn, L.H.: An Introduction to Climate, 5th edn. McGraw-Hill Company, New York (1980) 29. Ariyanti, R.D., Sugihartati, F.M., Retharanita, W., Winarso, P.A.: Uji keakuratan data GFS menggunakan model WRF-ARW pada siklon tropis Flamboyan. J. Ilmu dan Inovasi Fisika 3(2), 62–74 (2019) 30. Xiaoxia, Z., Yihui, D., Panxing, W.: Moisture transport in the Asian summer monsoon region and its relationship with summer precipitation in China. J. Meteorol. Res. 24(1), 31–42 (2010) 31. Zhu, Y., Newell, R.E.: A proposed algorithm for moisture fluxes from atmospheric rivers. Mon. Weather Rev. 126(3), 725–735 (1998) 32. Neiman, P.J., Ralph, F.M., Wick, G.A., Lundquist, J.D., Dettinger, M.D.: Meteorological characteristics and overland precipitation impacts of atmospheric rivers affecting the West Coast of North America based on eight years of SSM/I satellite observations. J. Hydrometeorol. 9(1), 22–47 (2008) 33. North Carolina Institute for Climate Studies.: https://ncics.org/portfolio/monitor/mjo/. Last Accessed 01 Nov 2021 34. Wheeler, M., Weickmann, K.M.: Real-time monitoring and prediction of modes of coherent synoptic to intraseasonal tropical variability. Mon. Weather Rev. 129(11), 2677–2694 (2001) 35. Camp, J., Wheeler, M.C., Hendon, H.H., Gregory, P.A., Marshall, A.G., Tory, K.J., Watkins, A.B., MacLachlan, C., Kuleshov, Y.: Skilful multiweek tropical cyclone prediction in ACCESSS1 and the role of the MJO. Q. J. R. Meteorol. Soc. 144(714), 1337–1351 (2018) 36. Bessafi, M., Wheeler, M.C.: Modulation of south Indian Ocean tropical cyclones by the Madden–Julian oscillation and convectively coupled equatorial waves. Mon. Weather Rev. 134(2), 638–656 (2006) 37. Wang, B., Moon, J.Y.: An anomalous genesis potential index for MJO modulation of tropical cyclones. J. Clim. 30(11), 4021–4035 (2017)

Evaluation of Detecting and Tracking Algorithms of Reflectivity Area Based on Rain Scanner Observation Data Tiin Sinatra, Edy Maryadi, Syahrul, Syukri Darmawan, Ginaldi Ari Nugroho, and Asif Awaludin

Abstract A rain scanner for rainfall monitoring had been developed using marine radar. Rainfall monitoring as a part of atmospheric observation was conducted using this rain scanner in near real time. Rain scanner could produce spatial rain conditions in reflectivity every 2 min on a radius of 44 km. This study utilizes the output of the rain scanner and generates the rainfall movement trajectory. Preliminary processing using object detection is conducted to separate the rainfall object from its background. Six case studies of the single rainfall cell in Bandung in January 2021 are used. The results show that, in general, using edge detection, find contour, and frame difference methods can identify rain objects well. The tracking results can carry in the same direction. However, the detailed rain tracking results have differences. Verified with other data, this method could serve as a reliable and effective system for the trajectory of the movement of rain in urban areas.

1 Introduction A weather radar is an essential instrument for providing rainfall distribution information over a large area within a certain time interval. Using the radar output, vectors showing the direction and speed of the precipitation trajectory can be determined [1]. The quantitative nowcast algorithm is mainly based on extrapolation of the rainfall distribution. The apparent movement of the rainfall distribution from the radar is extrapolated toward time at each grid or pixel. One of the extrapolation methods is Lagrangian persistence which allows extrapolation to obtain the next precipitation with the assumption that the precipitation features and movement are persistent [2]. The Lagrangian method consists of three steps: identifying, tracking, and forecasting [3]. The first step is identifying the rainfall object from the current radar data. The T. Sinatra (B) · G. A. Nugroho · A. Awaludin BRIN—Research Organization for Earth Sciences and Maritime, Bandung, Indonesia e-mail: [email protected] E. Maryadi · Syahrul · S. Darmawan BRIN—Research Organization for Electronics and Informatics, Bandung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_74

809

810

T. Sinatra et al.

second step is tracking the motion by identifying the same object in successive radar data. Later, forecasting using the extrapolation method [4]. The early steps of identification and tracking are essential to provide valuable information for the next forecasting calculation. Several methods have been proposed. A field-based method or cell-based method is often used in the identification method. The field-based advection considers all the pixels or grids of the radar data for calculating the motion vector. The cell-based method identifies a rainfall object from a connected domain within a certain threshold [5]. The tracking method based on cell-based identification (referred to as cell tracking) is often used since it is more efficient than field-based. Several studies of cell tracking have been done using various methods, such as cross-correlation [6], tracking barycenter of object cells [7], and motion vector calculation using optical flow [5]. The basic idea of this cross-correlation is to spatially shift the radar image taken at time t and overlay it with the image taken at time t − dt. The next step is to look for the best possible relative position of the two images that correspond to the best cross-correlation coefficient [3]. On the barycenter method, it calculates the coordinate from the center line of the precipitation cell based on two consecutive radar data. Confidence factors were analyzed as a separate measurement with other neighborhood cells. Another study tracked the results of radar monitoring using the optical flow method [8]. One method for tracking is moving object detection (MOV), a method for recognizing or determining the presence of an object and its scope and location on a moving image or video [9, 10] based on the pixel values over time [11] by separating an object from its background. The techniques that can be used to get the value of these pixels are temporal frame difference, background subtraction, optical flow, double difference, statistical methods, and others. Identification and tracking are well-established and studied using common weather radar. Implementing those two methods into a rain scanner is different challenge. A rain scanner developed based on navigation radar is often used as the filling gap of the weather radar network [12–14]. The advantage of this rain scanner is small, compact size, and ease of installation. The rain scanner can detect rain with a daily pattern close to the daily pattern of rainfall observed by the rain gage [15, 16]. The output of the rain scanner is monitoring spatial rain conditions in a series of images every 2 min. This research utilizes the output of this rain scanner for further analysis in order to provide information in the form of a rain movement trajectory. Tracking an object can be easier if the object does not change much in both feature and shape. This condition can be obtained if the temporal resolution between images from rain observations is very high and the observed rain comes from a single cloud. In fact, the observed rain can dissociate into several rain cells or, conversely, aggregate of several cells into a unique cell. Even for complex cases, the three conditions above can be observed in a one-time frame. A comprehensive method is needed to cover all three conditions. This study only tested the method for single-cell cases. The rain scanner data used in this study has 2 min temporal where the area and intensity change significantly within 2 min. This study aims to apply moving object detection to determine the movement of precipitation.

Evaluation of Detecting and Tracking Algorithms of Reflectivity Area …

811

2 Methodology Rain scanner is installed in Bandung (6.894941°S, 107.58648°E), West Java, with 44 km effective radius. It can cover rain observation over Bandung. Bandung is an urban area that consists of mountain area in the north and southern part (Fig. 1). The data used in this study are rain scanner observation data during January 2021, which was randomly selected. The characteristics of the rain object are summarized in Table 1. The rain scanner is able to produce rain observations spatially with a temporal resolution of 2 min and a spatial resolution of 120 m × 20 m at a radius of 44 km. Sequence data of rain scanner observations are then further processed for rain tracking. Data processing begins with data collection for the six rain events. The rain object selected for each rain event is a single cell. In general, the steps in this study are shown in Fig. 2. In order to determine whether the trajectory result is good or not, we use three criteria. The first is the ability to identify the rain objects. The second is that the trajectory result could follow the rain object propagation based on Table 1. The third is that no backward trajectory path exist, only forward movement is captured. Method-1 The first method uses the optical flow method. Optical flow Lukas-Kanade was used for estimating the direction of moving rain. Lucas-Kanade applies a local differential approach which calculate the flow a point by identifying the intersection of all flow path constraints to the image pixels within a frame. This method assumptions that

Fig. 1 Map of research location; black circle is coverage area of rain scanner

Table 1 Characteristics of the selected events Event no

Start

End

Decaying type

Propagation

Event 1

1 Jan 2021 13:28

1 Jan 2021 14:20

Develop-decay

Southeast

Event 2

1 Jan 2021 14:38

1 Jan 2021 15:08

Develop

East

Event 3

18 Jan 2021 17:52

18 Jan 2021 18:34

Develop

Southeast

Event 4

20 Jan 2021 02:42

20 Jan 2021 03:18

Decay

Southeast

Event 5

30 Jan 2021 12:02

30 Jan 2021 12:28

Develop

Southeast

Event 6

30 Jan 2021 13:20

30 Jan 2021 13:38

Decay

Southeast

812 Data Collection

T. Sinatra et al. Read data

Rain Area Detection

Rain area identification between frames

Motion Tracking

Fig. 2 Tracking rain algorithm

the brightness of the image is constant, where the movement at a point in the image is accompanied by no change in brightness level [17]. The first step in this method is to select objects in the frame using edge detection. Next, read the initial frame (t frame) to determine the interesting point formed on the detected rain object. The interesting point is determined using the Shi-Tomasi Corner Detection algorithm [18]. This algorithm determines the interesting point by comparing the eigenvalue with a certain value using (1). R = min(λ1 , λ2 )

(1)

Dot blue in Fig. 3a shows the interesting point determined. Then, estimate the movement of the interest point in the next frame (t + 1) until the frame-n (t + n) uses the Lukas-Kanade algorithm. The result of object tracking in frame-n is shown in Fig. 3b. Method-2 Figure 4 shows algorithm of method 2. In this method, the first thing to do is to subtract the reflectivity value of the object in the frame-t with frame t − 1. The results obtained are positive value (new rain areas appeared-red color) and negative value (old rain areas disappeared—blue color). Position x and y for all red and blue areas are calculated using mean formula to get the center point of each area. The center point is then used to determine the direction of rain movement. The illustration of the application of this method is shown in Fig. 5. Fig. 3 Simulation of rain detection and rain tracking uses Lukas-Kanade

Evaluation of Detecting and Tracking Algorithms of Reflectivity Area …

813

Fig. 4 Method 2 algorithm

Fig. 5 a Previous frame (t − 1), b base frame (t), c the result of subtracted reflectivity from base frame and previous frame

Method-3 and 4 Methods 1 and 2 identify the selected rain object for tracking by cutting other rain objects that are not the focus of monitoring in one frame. On the other hand, methods 3 and 4 identify the focus rain objects without eliminating other rain objects in one frame. Figure 6b shows the results of tracking the position of the rain area on one rain scanner data using method 3. Rain area detection is a process to find the rain area in each spatial rain data automatically, with an algorithm depicted in Fig. 7. Rain area detection will mark areas in the spatial rain data and will treat them as rectangular

814

T. Sinatra et al.

Fig. 6 a The result of rain detection. b The result of rain tracking. The rain area in the spatial rainfall data is marked with a green box

Spatial rain data

Read data

Find Contour

Evaluate size of rain area

Fig. 7 Rain area detection algorithm of method 3

objects, with the center point as an attribute of the object. The rain area resulting from the rain area detection process is shown in Fig. 6a. Identifying the rain area between the spatial rainfall data frame-t and frame t + 1 is calculated using the K-Nearest Neighbor (K-NN) method. Identification of the rain area is carried out to find the equivalent of each rain area on frame t with the rain area on spatial rain data to t + 1. K-NN calculates the distance between objects based on object attributes, in this case, is the center point, and selects as many as K objects that have the smallest distance or known as neighbors. Method 3 uses the value of K = 1 and the threshold large of rain area is 200 and 700, and the distance between objects is less than 3000. The improvement is conducted in this algorithm by applying a threshold on the Euclidean distance. Threshold of distance is used to prevent failed selected objects that can be occurred. Different from method 3, method 4 identifies the equivalent of the rain object in the next frame based on the closest distance without a specified range. In addition, method 4 identifies rain objects using edge detection, whereas method 3 uses one of shape identification techniques that is find contour.

3 Results and Discussion Figure 8 is the result of observing rain with a rain scanner on January 18, 2021. It is seen that the object of rain changes in shape and reflectivity. Proper detection of rain objects needs to be considered for tracking rain. Figure 9 shows the results of single-cell rain tracking with the four methods for each rain event. In general, it can be seen that the four methods show the movement of rain in the same direction, namely to the southeast for events 1, 3, 4, 5, and 6 and to the east for event 2.

Evaluation of Detecting and Tracking Algorithms of Reflectivity Area …

815

Fig. 8 Time variation of the rainfall event observed by rain scanner on January 18, 2021

Detailed analysis of the trajectory showed different characteristics. The trajectory from method 1 and method 2 in several events such as events 2, 3, 5, and 6 are quite different from methods 3 and 4. On the other hand, methods 3 and 4 show almost the same rain tracking (even coincide). This is because the tracking methods 3 and 4 are almost similar. Although, a small difference can be found between these two methods because of the difference in the identification method (edge detection and find contour), such as in event 2. The deviation of the trajectory is very visible in the result of method 1 (event 2, 3, 5, and 6). Although in the identification process utilizes edge detection same as method 4, the tracking system depends on the determination of the interest point. The calculation of the interest point will affect the determination of the trajectory. All the trajectory results did not show any backward path, which shows that all of the methods are able to identify and track the rain object movement. Another interesting characteristic is the start and endpoint of each method. The start points of the rain trajectory of method 2 is different from the other methods (events 1, 2, and 6), although the position of the endpoint is almost the same. This is because when determining the center point of rain, there is a part of another object (noise seen as speckle in Fig. 8) which is included in the calculation. Only events 1 and 4 that give a better conformity of the start point (except method 2) and endpoint. This shows that the complexity of the rain object also affected the tracking result. Overall result confirms that the object identification still needs to be improved to increase tracking accuracy.

4 Conclusion Detection and tracking applied on a single rain cell have been carried out for six different objects. The results show that in general, the four methods are able to identify rain objects well. Identification of rain objects using edge detection (method 1 and 4), frame difference (method 2), and find contour (method 3) gives almost the same results. The tracking results from the four methods as a whole are able to carry in the same direction. More detailed analysis shows that method 2 is less able to handle

816

T. Sinatra et al.

Fig. 9 Time variation of the rain tracking for a event 1, b event 2, c event 3, d event 4, e event 5, f event 6

speckle (noise) than other methods (1, 3, 4) because it is more sensitive to small/noise objects. In tracking rainfall, all methods are able to generate trajectory with almost similar direction to the rain object propagation. No backward trajectory paths were found within the result of those four methods. However, in the detailed analysis of the tracking result, Shi-Tomasi corner on method 1 showed a wider deviation of trajectory compared with other methods. Based on the conditions described, methods 3 and 4 give the optimum results for detecting and determining rain trajectory compared

Evaluation of Detecting and Tracking Algorithms of Reflectivity Area …

817

with other methods. Further evaluation is needed to compare the trajectory result with the reference data such as the automatic weather station network. Acknowledgements The authors would like to thank the National Research and Innovation Agency for providing the funding number 15/III/HK/2022 for this research. This research also is partly funded by research grant of Riset Inovasi Untuk Indonesia Maju number 65/II 7/HK/2022. We would also like to express our gratitude to our senior researchers who gave us guidance and support.

References 1. Ayzel, G., Heistermann, M., Winterrath, T.: Optical flow models as an open benchmark for radar-based precipitation nowcasting (rainymotion v0.1). Geosci. Model Dev. 12, 1387–1402 (2019) 2. Germann, U., Zawadzki, I.: Scale-dependence of the predictability of precipitation from continental radar images. Part I: description of the methodology. Mon. Weather Rev. 130, 2859–2873 (2002). https://doi.org/10.1175/1520-0493(2002)1302.0.CO;2 3. Austin, G.L., Bellon, A.: The use of digital weather radar records for short-term precipitation forecasting. Q. J. Roy. Meteor. Soc. 100, 658–664 (1974) 4. Austin, G. L., Bellon, A.: Guidelines for Nowcasting Techniques (WMO-No. 1198). Geneva (2017). 5. He, T., Einfalt, T., Zhang, J., Hua, J., Cai, Y.: New algorithm for rain cell identification and tracking in rainfall event analysis. Atmosphere 10(9), 532 (2019) 6. Brémaud, P.J., Pointin, Y.B.: Forecasting heavy rainfall from rain cell motion using radar data. J. Hydrol. 142(1–4), 373–389 (1993) 7. Raaf, O.: Efficient method for detecting and tracking rainfall clouds in non-Doppler radar images. J. Appl. Remote Sens. 8(1), 083547 (2014) 8. Nugroho, G.A., Maryadi, E.: Identifikasi Pergerakan area Reflektivitas Hujan Menggunakan Metode optical flow Berdasarkan data Pengamatan RDH. In: Prosiding Seminar Nasional Sains Atmosfer 2015, pp. 93–99. LAPAN, Bandung (2015) 9. Dewi, S.R.: Deep Learning Object Detection pada Video Menggunakan Tensorflow dan Convolutional Neural Network. Universitas Islam Indonesia, Yogyakarta (2018) 10. Kumar, S., Agarwal, Y.K.: Computer vision based moving object detection and tracking. Int. J. Adv. Sci. Technol. 30(1), 50–55 (2021) 11. Piccardi, M.: Background subtraction techniques: a review. In: 2004 IEEE International Conference on Systems, Man and Cybernetics (IEEE Cat. No.04CH37583), vol. 4, pp. 3099–3104 (2004) 12. Awaludin, A., Nugroho, G.A., Alam, S., Kurniawan, D.F. and Yuwono, R.: Development of marine radar signal acquisition and processing system. In: Proceeding of 2013 International Conference on Information Technology and Electrical Engineering (2013). https://doi.org/10. 1109/ICITEED.2013.6676219 13. Sinatra, T., Nugroho, G.A., Awaludin, A.: Optimasi data radar Hujan dan Perluasan Jangkauannya Menggunakan Jaringan radar. Jurnal Sains Dirgantara 18(1), 43–54 (2020) 14. Awaludin, A., Sinatra, T., Nugroho, G.A., Nauval F.: Clutter removal improvement of marine radar for weather observation. In: AIP Conference Proceedings, vol. 2366, p. 060023 (2021). https://doi.org/10.1063/5.0060062 15. Sinatra, T., Awaludin, A., Nauval, F., Purnomo, C.: Calibration of spatial rain scanner using rainfall depth of rain gauges. In: IOP Conference Series: Earth and Environmental Science, vol. 893, p. 012064 (2021). https://doi.org/10.1088/1755-1315/893/1/012064 16. Awaludin, A., Nugroho, G.A., Rahayu, S.A.: Analisis Kemampuan Radar Navigasi Laut Furuno 1932 Mark-2 untuk Pemantauan Intensitas Hujan. Jurnal Sains Dirgantara, vol. 10 (2013)

818

T. Sinatra et al.

17. Lucas, B., Kanade, T.: An iterative image registration technique with an application to stereo vision. In: Proceedings of Seventh International Joint Conference on Artificial Intelligence, pp. 674–679. Vancouver, Canada (1981) 18. Abdulameer Kadhim, H., Abdullah Araheemah, W.: A method to improve corner detectors (Harris, Shi-Tomasi & FAST) using adaptive contrast enhancement filter. Period. Eng. Nat. Sci. 8(1), 508–515 (2020)

Total Suspended Solids Concentration Estimation in Coastal Wasters Using Remote Sensing Data and Machine Learning Approach D. N. Lintangsasi, A. Rahmadya, I. Ridwansyah, and F. Setiawan

Abstract Total suspended solids (TSS) concentration is an important water quality parameter generally used to investigate the watershed ecosystem health related to erosion and sedimentation process. It is crucial to routinely and rapidly monitor the TSS concentration, particularly in coastal regions. We propose a model for estimating TSS concentration from remote sensing data using a machine learning approach. We develop the model using in situ TSS data collected from three Indonesian coastal areas (TSS values 26.20–134.67 g/m3 , n = 43). First, we carried out Dark Object Subtraction (DOS) atmospheric correction on Landsat-8 OLI Level-1 images and distinguished water and non-water using the Normalized Different Water Index (NDWI). Second, we tested all possible band and band ratio combinations as TSS predictors and four machine learning methods, i.e., linear model (LM), k-nearest neighbor (kNN), support vector machine (SVM), random forest (RF). Our results showed that the two-band ratios (Red/Ultra Blue and NIR/Ultra Blue) with the RF method gave the best performance (R2 of 0.81 and the root mean square error of 10.02 g/m3 ) and reasonable spatial TSS distribution. Our TSS estimation model well captured the TSS values and spatial distribution in three coastal regions, further valuable to support integrated watershed-river-coastal sustainable management.

1 Introduction Sedimentation is a major problem often encountered in river estuary [1], which can trigger river siltation and pollution in water bodies. Sediment deposition in water bodies can be influenced by the concentration of suspended sediment, which is

D. N. Lintangsasi Undergraduate School of Faculty of Geography, Gadjah Mada University, Yogyakarta, Indonesia A. Rahmadya · I. Ridwansyah · F. Setiawan (B) Research Center for Limnology and Water Resources, National Research and Innovation Agency (BRIN), Jakarta Pusat, Indonesia e-mail: [email protected] © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_75

819

820

D. N. Lintangsasi et al.

distributed by the river stream and finally deposited [2, 3]. Total suspended sediment or suspended solid (TSS) is a suspended material with a diameter of >1 µm which is retained on a 0.45 Millipore diameter filter [4]. TSS can consist of mud, fine sand, and microorganisms that most of which are formed from soil erosion entered into water bodies [5]. A conventional point-based TSS measurement on water bodies is limited to the number of sampling points; as the result, the spatial distribution of TSS remains unknown [6]. However, performing a routine monitoring of TSS and spatial distribution is urgently required to observe the possibility and direction of sedimentation in a water environment. Satellite remote sensing provides high spatial and temporal resolution data of the earth’s surface that has high potential to efficiently monitor the water quality such as TSS. Compared to manual monitoring, remote sensing technology has the advantages in the effectiveness of time, effort, and cost required. Several studies have proven that remote sensing imagery (Landsat-8 OLI imageries) can be used to observe TSS concentration in water. The image processing is carried out using a spectral transformation method using a TSS estimation algorithm [7–12]. However, some of these algorithms only effectively estimated TSS in water conditions within a specific TSS range as the model was built and further limited the application to certain water conditions. Indonesia has various water conditions, as well as the TSS concentrations range. It is necessary to build a TSS estimation model that can cover a wide range of TSS concentrations. The development of the model can be done using a machine learning approach, which has the advantage of being able to handle more complex problems, especially when compared to physical models because the working principle does not require other information where limited assumptions are required such as physical model [13–15]. The objective of this study is to develop a model for estimating a TSS concentration from Landsat-8 OLI data. The developed model is expected to be more applicable to various water conditions in Indonesia.

2 Data and Method 2.1 In Situ TSS Data We collected in situ TSS data from three conducted field surveys in three different coastal waters i.e., Batanghari Estuary (Jambi Province), Jakarta Bay (Jakarta Province), and Cimandiri Estuary (West Java Province) (Fig. 1a). In total, we have 43 points of data, with the minimum, maximum, and average values (mg/L) for Batanghari Estuary is 58.67, 134.67, and 85.80, respectively, for Jakarta Bay is 26.20, 100.80, and 58.59, respectively, and for Cimandiri Estuary is 58.00, 102.00, and 82.58, respectively, as shown in Fig. 1b.

Total Suspended Solids Concentration Estimation in Coastal Wasters …

821

Fig. 1 a Locations of the three Indonesian coastal areas, b distribution of TSS concentrations in the in situ dataset used in this study (yellow points are the outliers)

2.2 Landsat Data Collections and Pre-processing We collected satellite images that were acquired by Landsat-8, from the United States Geological Survey (USGS) earth explorer website [https://earthexplorer.usgs. gov/]. Landsat-8 OLI Level-1 images were atmospherically corrected to get the surface reflectance (SR) value as input data for algorithm processing. We carried out atmospheric correction using Dark Object Subtraction (DOS) method [16]. The working principle of this atmospheric correction method is to eliminate the lower pixel value which is assumed as the darkest object in the image (Table 1). We cropped the atmospherically corrected image using shapefile boundaries of each site that is digitized manually from the true color composite of the image. We removed the non-water pixels using the Normalized Different Water Index (NDWI) threshold [17], where a pixel with an NDWI value less than or equal to 0 is assumed as non-waters. NDWI =

Table 1 Three Landsat-8 OLI images corresponding with in situ TSS dataset

(blue − NIR) (blue + NIR )

(1)

Location

Landsat Path and row

Acquisition date

Days gap (days)

Batanghari

125–61

September 21, 2021

17–22

Jakarta

122–65

November 3, 2021

15–17

Cimandiri

122–64

May 11, 2021

13–15

822

D. N. Lintangsasi et al.

Fig. 2 Model development work flow

where blue and NIR are the atmospherically corrected values at the blue band and the near-infrared band, respectively.

2.3 Model Development We used in situ TSS and the pre-processed Landsat Dataset to develop TSS estimation models. We obtained a total of 43 pairs. We then used the in situ TSS values as dependent variables and various combinations of the corresponding Landsat bands (e.g., using single bands, band ratios, band ratios and single bands, and two-band ratios) as independent variables. We used four machine learning approaches, i.e., linear model (LM), k-nearest neighbor (k-NN), support vector machine (SVM), and random forest (RF). We perform machine learning training and testing using the “caret” package in the R language [18]. We split the data into 75% for model training (calibration) and 25% for data testing (validation). We provided the workflow in Fig. 2.

2.4 Accuracy Assessment We used the root means square error (RMSE) to assess the models’ accuracy. The index is defined as follows:  N 2 i=1 (X esti,i − X meas,i ) RMSE = (2) n where X esti, i and X meas, i are the estimated and measured TSS values, respectively; n is the number of samples. We also calculated the correlation between the measured and estimated values (R2 ). We perform visual comparison between TSS estimation result image and the true color image to evaluate the TSS spatial distribution pattern.

Total Suspended Solids Concentration Estimation in Coastal Wasters …

823

3 Result and Discussion 3.1 The Machine Learning Models for Estimating the TSS from Landsat-8 OLI We built and tested 68 models using our dataset, which consists of seven “single bandbased models,” 15 “two single band-based models,” 15 “band ratio-based models,” and 30 “two-band ratios-based models.” None of the single band-based models has a high coefficient of determination, even though a complex machine learning approach has been applied. Either, the two single band-based model has a low coefficient of determination. This worse performance revealed that the use of a single band both as one band and two bands format are not robust enough as a TSS predictor. Table 2 shows the selected developed TSS estimation models and their performances; we excluded the TSS estimation models with the worse performance by using thresholds of R2 values = 150 mm between model outputs and satellite. Rainfall characteristics/characteristics are obtained by comparing rainfall output from models and satellites, in the wet and dry seasons, with the average monthly rainfall for the 2014–2018 period. The results of seasonal predictions using CCAM include the values of rainfall reconstructed, rainfall probability >= 150 mm from time-lagged, and multi-sea surface temperature (multi-SST) ensemble. Seasonal predictions produce eight months of predictions, so the timelagged ensemble uses predictions from eight members, while multi-SST ensembles are carried out using 11 simulations with 11 different SST inputs. The results show that the seasonal predictions output from rainfall reconstruction, probabilistic timelagged, and multi-SST ensemble can capture monthly rainfall patterns >= 150 mm. However, in some parts of Indonesia, it is still far from being well predicted. Probabilistic rainfall >= 150 mm from a multi-SST ensemble has a better spatial pattern than the time-lagged ensemble compared to observation in almost all regions of Indonesia. H. Satyawardhana (B) · Risyanto · E. Yulihastin · Gammamerdianti · C. N. Ihsan · E. P. Wulandari · L. Q. Avia · I. Sofiati Research Center for Atmospheric Science and Technology, Indonesia National Agency of Research and Innovation, Bandung, Indonesia e-mail: [email protected] M. Arif Setyo Aji Bandung Institute of Technology, Bandung, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_77

837

838

H. Satyawardhana et al.

1 Introduction Indonesia is a maritime country between two oceans (the Pacific and Indian) and two continents (Asia and Australia). This condition makes the Indonesian sea the only route connecting the Pacific and Indian Oceans in the Equatorial zone [1, 2]. This flow, known as the Indonesian Throughflow (ITF), carries warm water from the Pacific to the Indies and vice versa through the oceans in the inter-island region of Indonesia [3, 4]. This condition causes the movement of monsoon winds and affects the variability of rainfall and seasons in Indonesia both temporally and spatially [5]. Seasonal predictions are mid-range forecasts of the average weather on a time scale of several months to a year in the future in several regions [6]. The need for seasonal predictions, especially related to temperature and rainfall information, is increasing along with the public awareness of the effect of seasonal variability in important sectors [7]. This condition encourages innovation development to get more accurate seasonal predictions, one of which is by comparing and verifying the results of seasonal predictions to produce high predictions. Among the many seasonal prediction models, one of the atmospheric models with a high resolution is the Conformal Cubic Atmospheric Model (CCAM) developed by CSIRO, Australia. CCAM is a conformal cubic grid-based global model that uses the Schmidt transform for regional and local forecasts. In addition, the multiple nesting technique is implemented for downscaling and integrates topographic and land use data [8, 9]. The conformal cubic coordinate system allows CCAM to be used as a global and regional prediction model. This feature is an advantage of CCAM compared to other global models in general. In this study, Global Satellite Mapping of Rainfall (GSMaP) data from the Japan Aerospace Exploration Agency (JAXA) was used as a verifier for CCAM output rainfall prediction data due to its high spatial and temporal resolution [10, 11]. In addition, GSMaP products have been evaluated and verified through comparisons with observational data from several previous studies and showed promising results in capturing rainfall patterns and variability [12–16]. The purpose of this study is to compare the rainfall from the results of seasonal predictions using CCAM and satellites in the dry and wet seasons in the Indonesian Territory. In addition, the results were obtained to verify the rainfall prediction from several seasonal prediction simulations. Research on verifying seasonal prediction data is urgently needed, given the importance and challenge of seasonal predictions.

2 Data and Methods 2.1 Data This study uses rainfall data from the results of seasonal predictions using CCAM and GSMaP. Prediction using CCAM is obtained with initial condition input from GFS

Ensembles Simulation on the Seasonal Rainfall Characteristics Over …

839

and SST prediction from POAMA. CCAM seasonal prediction data contains eight months of prediction periods from the initial condition and has 32 km spatial and a daily temporal resolution, although processed monthly. The CCAM seasonal prediction data are time-lagged, multi-SST ensemble, and prediction rainfall reconstruction. These three data were obtained from the simulation of the seasonal prediction, which is explained further in subsection 2.2. Seasonal prediction simulations were carried out in two periods: June 2019, which represented the dry season, and December 2018 represented the wet season, with the research area being the Indonesian Maritime Continent (BMI) region. CCAM is an atmospheric model developed by the Commonwealth Scientific and Industrial Research Organization (CSIRO), where CCAM was the development of DARLAM in 1994 [9]. This study uses CCAM to obtain medium-term predictions because CCAM can generate atmospheric prediction data using SST data guidance. The input data used are atmospheric data as initial conditions obtained from GFS data and SST prediction data obtained from the Predictive Ocean Atmosphere Model for Australia (POAMA) in the form of a surface temperature prediction ensemble. POAMA is a state-of-the-art seasonal to interannual forecast system based on a coupled ocean/atmosphere model developed in a joint project involving the Bureau of Meteorology Research Center (BMRC) and CSIRO Marine Research [18]. The comparison data for the rainfall parameter is GSMaP rainfall data. GSMaP rainfall data used is GSMaP_mvk which has 0.1° of spatial resolution. Before the comparative analysis, the two data (CCAM output and GSMaP) require spatial and temporal resolution adjustment.

2.2 Methods This study uses CCAM to perform several simulations of seasonal predictions (medium-term forecasts). Initial condition data uses Global Forecasting System (GFS) data for the first date of each month and uses forcing SST predictions for the next eight months from POAMA. This simulation description had a similar scheme to our previous studies [17]. On post-processing, CCAM output data and simulations in this study were built by three ensemble prediction systems: time-lagged ensemble, multi-SST ensemble, and predicted rainfall reconstruction (Fig. 1). The time-lagged ensemble system consists of an ensemble member, which results from CCAM seasonal predictions on the rainfall variable for the next eight months with differences in the time of the initial condition. The simulation results from this system are then cut (intersection) in June 2019 (representing the peak of the dry season) and December 2018 (the peak of the wet season), as described in Fig. 1a. Therefore, the time-lagged ensemble consists of eight members for each season simulation. While in Fig. 1b, the multi-SST ensemble system consists of an ensemble member that uses two initial conditions that have been given 11 different SST inputs. Finally, the predicted rainfall reconstruction is carried out using the output data from the multi-SST ensemble, whose work is depicted in Fig. 1c. Further explanation

840

H. Satyawardhana et al.

Fig. 1 a Seasonal prediction simulation scheme for time-lagged ensemble, b multi-SST ensemble, and c steps to reconstruct the rainfall prediction data. The date in the picture only illustrates the prediction scheme, not the date used for this study

regarding rainfall reconstruction can be seen in previous studies [20]. Thus, the timelagged and multi-SST ensemble results were probabilistic values, and the rainfall reconstruction results were deterministic. Rainfall data from these three simulations are classified based on the rainfall characteristics by the Indonesian Agency of Meteorology, Climatology, and Geophysics (BMKG). This rainfall characteristic is the result of comparing the rainfall intensity during the specified period and the average rainfall in the same period on a longterm scale (baseline) [21]. The long-term scale is defined as the arithmetic means of each climate (such as rainfall and temperature) for 30 years [22]. However, due to data availability, this study determined normal rainfall from the average rainfall only during 2014–2017. The characteristics of rain, according to BMKG, are divided into three categories, namely.

Ensembles Simulation on the Seasonal Rainfall Characteristics Over …

841

1. Above Normal (AN)—if rainfall values are more than 115% of baseline data. 2. Normal (N)—if rainfall values are between 85% - 115% of baseline data. 3. Below Normal (BN)—if rainfall values are less than 85% of baseline data [21].

3 Results and Discussion This study compared the characteristics of monthly rainfall between the predicted results and satellite rainfall. The monthly rainfall characteristics used in this study describe rainfall above 150 mm/month. Figure 2a shows that rainfall above 150 mm/month in June 2019 mainly occurred north and around the equator, while rainfall in southern Indonesia, such as in Java, Bali, Nusa Tenggara, and Timor, was less than 150 mm/month. The rainfall of less than 150 mm/month is caused by the movement of the Australian monsoon, which brings dry air in from southern Indonesia through Java and its surroundings. The several prediction simulations show similar results, as shown in Figs. 2b and c, that the probability value of rainfall events of more than (less than) 150 mm/month occurring at the equator and northern BMI (southern BMI). It shows that the probabilistic value of rainfall of more than 150 mm/month generated by the simulation results can describe the occurrence of rainfall of more than 150 mm/month. Although the probability value of rainfall is more than 150 mm/month in some areas, it is still not well predicted, such as in the western coast of Sumatra, the Malacca Strait, the South China Sea, and the eastern part of the Philippine Sea. The rainfall probability value of more than 150 mm/month for prediction simulation using multi-SST ensembles is more similar to GSMaP data than time-lagged ensembles. Figure 3 shows the spatial pattern of GSMaP rainfall of more than 150 mm/month and a simulated probability of rainfall of more than 150 mm/month, as predicted by CCAM in December 2018. In December 2018, most areas of Indonesia had entered the wet season, so rainfall of more than 150 mm/month occurs evenly in almost all parts of Indonesia (Fig. 3a). However, the probability of more than 150 mm/month rainfall shown by the CCAM seasonal prediction simulation (Fig. 3b and c) shows that only areas south of the equator get a high probability value for rainfall of more than 150 mm/month. It shows that the seasonal prediction simulation for the wet season in IMC was still not good enough for areas north of the equator following Schiemann et al. (2014) statement that the sensitivity model is not significantly associated with orographic conditions in the Indonesian Maritime Continent [23]. In addition, Satyawardhana et al. (2022) stated that the results of seasonal predictions in the areas on the west coast of Sumatra, Central Kalimantan, South Sulawesi, and the west coast of southern Papua have a negative correlation value with the observation data [20]. Therefore, research is still needed to get more accurate seasonal predictions by comparing and verifying the results of seasonal predictions to produce high-resolution seasonal predictions. Figures 4 and 5 show a comparison of rainfall characteristics in Indonesia in June 2019 and December 2018 based on satellite data and model results. Figure 4a shows

842

H. Satyawardhana et al.

Fig. 2 Comparison of rainfall above 150 mm/month and probability of rainfall above 150 mm/month in June 2019

Fig. 3 Comparison of rainfall above 150 mm/month and probability of rainfall above 150 mm/month in December 2018

Ensembles Simulation on the Seasonal Rainfall Characteristics Over …

843

the rainfall characteristics using GSMaP in June 2019, which were below normal for southern Indonesia and the Sulawesi Sea. In addition, the western sea of Sumatra and the South China Sea were below-normal compared to the 5-year average rainfall. The results of the rainfall reconstruction (Fig. 4b) show that areas in the Java Sea have different results from the GSMaP data, which should be below normal. In addition, the below-normal rainfall characteristics described by satellite data in the Malacca Strait, West Pacific, and the South China Sea are also described reverse by the CCAM rainfall reconstruction results. Similar results are shown by the probabilistic value of time-lagged (Fig. 4c) and multi-SST ensemble (Fig. 4d). Even in the probability value of rainfall characteristics, the time-lagged ensemble has a low accuracy of rainfall characteristics in the Sumatra area. In contrast, the rainfall characteristics of the multi-SST ensemble are well predicted. Figure 5a shows the spatial pattern of rainfall characteristics from GSMaP data which illustrates that the rainy season in Indonesia (December 2018) is classified as below-normal to normal. However, all the predictions in reconstructed rainfall, time-lagged, and multi-SST ensemble cannot capture normal rainfall (Fig. 5b, c, and d). Therefore, this seasonal prediction for December 2018 can only capture below and above-normal rainfall characteristics. This condition is due to the range of values for normal rainfall, which tends to be small compared to other categories.

Fig. 4 Comparison of rainfall characteristics in June 2019 with the climatological average of rainfall in June, between a GSMaP, b CH reconstruction results, c time-lagged ensemble, and d multi-SST ensemble

844

H. Satyawardhana et al.

Fig. 5 Same as Fig. 2, but in December 2018

4 Summary and Conclusions We investigated the ensemble simulations on the seasonal rainfall prediction during 2018–2019 over the Indonesia Maritime Continent. The wet (dry) seasons were represented by December 2018 (June 2019). To explore the methods, we compared a time-lagged ensemble, a multi-SST ensemble, and no-ensemble simulations, then validated the results using GSMaP satellite data. The main results could be described as follows: • A multi-SST ensemble predicted shows a better qualitative agreement in rainfall characteristics compared with observed rainfall than a time-lagged ensemble during wet and dry season periods. • The simulation results suggest that the methods have skillful predictions in the dry season, despite a low skill for the wet season. However, it might cause the model not to capture the normal rainfall observed characteristics. • This study noticed that for the dry season in the southwestern IMC, a multi-SST consistently improved the rainfall characteristics over land (Sumatra-Java) and ocean (Java Sea) compared to a time-lagged ensemble. The southwestern IMC was mentioned in numerous studies [24–26] as a prominent region responsive to local and remote ocean forcing. In the dry season case study, the predicted improvement was depicted by a reduction of above-normal rainfall over the Java sea, whereas an enhancement of below-normal rainfall inland. It suggests that a multi-SST method that considers various SST from 11 ensemble members successfully elaborates physics and dynamics processes over boundary layers associated with increased rainfall over land during the dry season, which was proposed in a previous study [27].

Ensembles Simulation on the Seasonal Rainfall Characteristics Over …

845

• It should be noted that in the wet season, both ensembles method gave underestimated prediction results over southwestern IMC. Furthermore, it proposed that the simulations have a systematical bias due to meso-to-regional scales atmospheric processes that could not represent well in the CCAM model. However, considering results that have almost similar predicted rainfall between a time-lagged and a multi-SST ensemble, this study suggests that a time-lagged ensemble should be chosen as a better strategy to reduce the time-running consumption of the model. Acknowledgements The authors would like to thank the Japan Aerospace Exploration Agency (JAXA) for the GSMaP data and the Australian Bureau of Meteorology (BOM) for providing POAMA data. The authors also want to show our gratitude to CSIRO for developing CCAM that we can use to improve our capability in atmospheric modeling.

References 1. Koch-Larrouy, A., Madec, G., Bouruet-Aubertot, P., Gerkema, T., Bessières, L., Molcard, R.: On the transformation of Pacific water into Indonesian throughflow water by internal tidal mixing. Geophys. Res. Lett. 34(4) (2007) 2. Sprintall, J., Révelard, A.: The Indonesian throughflow response to Indo-Pacific climate variability. J. Geophys. Res.: Oceans 119(2), 1161–1175 (2014) 3. Wijffels, S., Schmidt, R.W., Bryden, H.L., Stigebrandt, A.: Transport of freshwater by the Oceans. J. Phys. Oceanogr 22, 156–162 (1992) 4. Ganachaud, A., Wunsch, C., Marotzke, J., Toole, J.: Meridional overturning and large-scale circulation of the Indian Ocean. J. Geophys. Res. 105, 26117–26134 (2000) 5. Dayem, K.E., Noone, D.C., Molnar, P.: Tropical western Pacific warm pool and maritime continent precipitation rates and their contrasting relationships with the walker circulation. J. Geophys. Res. Atmos. 112(6), 1–12 (2007). https://doi.org/10.1029/2006JD007870 6. Jan van Oldenborgh, G., Balmaseda, M.A., Ferranti, L., Stockdale, T.N., Anderson, D.L.T.: Evaluation of atmospheric fields from the ECMWF seasonal forecasts over a 15-Year period. J. Clim. 18(16), 3250–3269 (2005) 7. Schepen, A., Wang, Q.J., Robertson, D.E.: Seasonal forecasts of Australian rainfall through calibration and bridging of coupled GCM outputs. Mon. Weather Rev. 142(5), 1758–1770 (2014). https://doi.org/10.1175/MWR-D-13-00248.1 8. Thatcher, M., McGregor, J.L.: A technique for dynamically downscaling daily-averaged GCM datasets using the conformal cubic atmospheric model. J. American Meteorol. Soc. 139, 79–95 (2011) 9. Thatcher, M., McGregor, J., Dix, M., Katzfey, J.: A new approach for coupled regional climate modeling using more than 10,000 Cores. In: Denzer, R., Argent, R.M., Schimak, G., Hˇrebíˇcek, J. (eds.) Environmental Software Systems. Infrastructures, Services, and Applications. ISESS 2015. IFIP Advances in Information and Communication Technology, vol. 448. Springer, Cham (2015). https://doi.org/10.1007/978-3-319-15994-2_61 10. Ushio, T., Sasashige, K., Kubota, T., Shige, S., Okamoto, K.I., Aonashi, K., Inoue, T., Takahashi, N., Iguchi, T., Kachi, M., Oki, R., Morimoto, T., Kawasaki, Z.I.: A Kalman filter approach to the global satellite mapping of precipitation (GSMaP) from combined passive microwave and infrared radiometric data. J. Meteorol. Soc. Japan Ser II 87A, 137–151 (2009). https://doi.org/ 10.2151/jmsj.87A.137

846

H. Satyawardhana et al.

11. Aonashi, K., Liu, G.: Passive microwave precipitation retrievals using TMI during the Baiu period of 1998. Part I: algorithm description and validation. J. Appl. Meteorol. 39(12), 2024– 2037 (2000) 12. Priyambodho, B.A., Kure, S., Yagi, R., Januriyadi, N.F.: Flood inundation simulations based on GSMaP satellite rainfall data in Jakarta Indonesia. Prog. Earth Planet Sci. 8, 34 (2021). https://doi.org/10.1186/s40645-021-00425-8 13. Setiawati, M.D., Miura, F.: Evaluation of GSMaP daily rainfall satellite data for flood monitoring: case study–Kyushu Japan. J. Geosci. Environ. Prot. 4(12), 101–117 (2016). https://doi. org/10.4236/gep.2016.412008 14. Admojo, D.D., Tebakari, T., Miyamoto, M.: Evaluation of a satellite-based rainfall product for a runoff simulation of flood event: a case study. J. Japan Soc. Civil Eng. Ser. B1 74(4), I_73–I_78 (2021) 15. Pakoksung, K., Takagi, M.: Effect of satellite-based rainfall products on river basin responses of runoff simulation on flood event. Model Earth Syst. Environ. 2(3), 143 (2016). https://doi. org/10.1007/s40808-016-0200-0 16. Rusmanansari, A.H., Suwarman, R., Djamil, Y.S., Firdaus Permadhi, Y.: GSMaP seasonal rainfall verification over Western Java. In: Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, 2021. Springer Proceedings in Physics, vol. 275. Springer, Singapore (2022). https://doi.org/ 10.1007/978-981-19-0308-3_60 17. Satyawardhana, H., Gammamerdianti (2019) Comparison of seasonal prediction outputs based on dynamic atmosphere model and observations (case study: seasonal prediction in 2016−2017). In: IOP Conference Series: Earth Environment Science, vol. 303. pp. 012050. (2017) 18. Wang, G., et al.: POAMA: an Australian ocean-atmosphere model for climate prediction. Bull. Am. Meteorol. Soc. 1, 4559–4563 (2004) 19. Kachi, M., Kubota, T., Aonashi, K., Ushio, T., Shige, S., Yamamoto, M.: Recent improvements in the global satellite mapping of precipitation (GSMaP). In: 2014 IEEE Geoscience and Remote Sensing Symposium, pp.3762−3765 (2014). https://doi.org/10.1109/IGARSS.2014. 6947302. The GFDL Global Atmospheric Model Development Team.: The New GFDL Global Atmosphere and Land Model AM2-LM2: Evaluation with Prescribed SST Simulations. J. Climate 17(24), 4641–4673 (2004) 20. Satyawardhana, H., Yulihastin, E., Gammamerdianti, Ihsan N.I., Wulandari, E.P.: Evaluation of CCAM seasonal prediction by GSMaP satellite rainfall data in Indonesia. In: Yulihastin, E., Abadi, P., Sitompul, P., Harjupa, W. (eds.) Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, 2021. Springer Proceedings in Physics, vol. 275, Springer, Singapore (2022). https://doi.org/ 10.1007/978-981-19-0308-3_14 21. BMKG: Prakiraan Musim Hujan 2022/2023 di Indonesia. BMKG 2022) 22. Muharsyah, R.: Penentuan Batas Atas Normal dan Bawah Normal Curah Hujan Bulanan Setara Tercile dengan Koefisien Regresi Linier Sederhana. Jurnal Meteorologi dan Geofisika 15(1), (2014) 23. Reinhard, S., Marie-Estelle, D., Mizielinski, M., Malcolm, R., Len, S., Strachan, J., Vidale, P.L.: The sensitivity of the tropical circulation and Maritime Continent precipitation to climate model resolution. Climate Dynam. 42 (2013). https://doi.org/10.1007/s00382-013-1997-0 24. Hamada, J.-I., Mori, S., Kubota, H., Yamanaka, M.D.: Interannual rainfall variability over northwestern Jawa and its relation to the Indian Ocean Dipole and El Niño-Southern Oscillation events. SOLA 8, 069–072 (2012) 25. Xu, Q., Guan, Z., Jin, D., Hu, D.: Regional characteristics of interannual variability of summer rainfall in the Maritime Continent and their related anomalous circulation patterns. J. Clim. 32, 4179–4192 (2019)

Ensembles Simulation on the Seasonal Rainfall Characteristics Over …

847

26. Yulihastin, E., Putrantom M.F., Suaydhi, Sofiati, I.: Oceanic effect on precipitation development in the Maritime Continent during anomalously-wet dry seasons in Java. Indones. J. Geogr. 53(3), 328–339 (2020) 27. Yulihastin, E., Nuryanto, D.E., Trismidianto, Muharsyah, R.: Improvement of heavy rainfall simulated with SST adjustment associated with mesoscale convective complexes related to severe flash flood in Luwu, Sulawesi, Indonesia. Atmosphere 12,1445 (2021). https://doi.org/ 10.3390/atmos12111445.

Application and Analysis of Remote Sensing for the Initial Baseline of the Quantitative Comfort Index in Ibukota Nusantara (IKN) and Its Surroundings Indah Susanti, Lilik S. Supriatin, Sinta B. Sipayung, Edy Maryadi, Adi Witono, Martono, Laras Toersilowati, and Amalia Nurlatifah Abstract Convenience is an important aspect to optimize work productivity and performance, so it needs to be considered in development planning, especially in determining the location of the capital city of a country. This is important because the level of comfort will affect the level of energy use, especially electrical energy. The location of the capital city of a country should be chosen which can minimize the level of energy use in accordance with the principles of sustainable development (SDGs). Based on air temperature and humidity data from the Atmospheric Infrared Sounder (AIRS), the quantitative comfort level (Temperature–Humidity Index formulation) was processed and analyzed in Balikpapan City, Gunung Mas Regency, Katingan Regency, Kutai Kartanegara Regency, Palangkaraya City, and Tanah Bumbu Regency. The time span of the data processed and analyzed is January 2003 to December 2018. The results of statistical analysis show that the monthly average THI in the six regions is in the range of 24–27 °C, or in the partially comfortable category. Gunung Mas Regency is an area that has an average THI of 24.8, while other areas have a THI above 25. This means that the level of comfort in Gunung Mas Regency is the best compared to other areas analyzed. In addition, in the time span analyzed, the trend of THI in Gunung Mas Regency shows a decline, which means toward a better level of comfort. This has implications for the level of energy use which is also better.

L. S. Supriatin · S. B. Sipayung · A. Witono · Martono · L. Toersilowati · A. Nurlatifah Research Center of Climate and Atmosphere, National Research and Innovation Agency, West Java, Indonesia I. Susanti (B) Directorate of Repository, Multimedia and Scientific Publication, National Research and Innovation Agency, West Java, Indonesia e-mail: [email protected] E. Maryadi Research Center for Data and Information Sciences, National Research and Innovation Agency, West Java, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_78

849

850

I. Susanti et al.

1 Introduction A few moments ago the issue of moving the capital city of Indonesia arose. The idea of relocating the nation’s capital was based on traffic and flood that occurred in DKI Jakarta. Based on traffic conditions, Jakarta is ranked as the fourth worst city, and this has implications for real financial losses [2]. The loss in 2013 reached Rp 56 trillion per year, and in 2019 it was close to Rp 100 trillion [2]. From the hydrological aspect, the Jakarta area which is included in the flood-prone category reaches 50% and has a flood vulnerability level of under 10 years. Moving the nation’s capital is not an easy matter. In this case, careful consideration is needed in planning, both from the economic, social, and environmental aspects. The city is an ecosystem formed by various types of land cover, vegetation, and various types of land use from a very complex landscape [5]. Therefore, relocating the nation’s capital means that an unavoidable urbanization process will occur. The trend is that urbanization puts a lot of pressure on the environment, including reducing the size of the habitat and changing the spatial pattern of the dynamics of living things. Several studies related to urbanization have shown changes in abiotic conditions in the remaining parts of the habitat, which are caused by increases in temperature, rainfall, humidity, and changes in the nitrogen cycle [6, 14]. Physical characteristics generally affect the ecological functions that support life. The level of primary productivity that determines the sustainability of an ecosystem is largely determined by temperature, rainfall, radiation, and carbon dioxide levels. Likewise, the level of human comfort is determined by temperature, water content in the air, wind, and several other factors. In areas with a low level of comfort, it requires higher facilities and energy to obtain an adequate level of comfort. The farther an area is from the optimal comfort level, the higher the additional energy required to obtain the required level of comfort. A clear example of this is the use of air conditioning (AC) in areas with low comfort levels. Planning the needs and use of facilities, infrastructure and energy in an area, needs to consider the carrying capacity and capacity of the environment in order to have sustainability. This consideration is in accordance with the demands of the development paradigm and international agreements that are currently developing, namely the sustainable development goals (SDGs). Consideration of environmental comfort in urban development becomes important, because the level of health of humans and other living creatures is strongly influenced by environmental conditions, including the level of heat stress. Heat stress is influenced by air temperature, humidity, water movement, solar radiation, and precipitation [8]. Several studies have been conducted, such as by Basu [1], Hajat and Kosatky [7], Morabito, et al. [11], Tobias, et al. [15], Wu et al. [18] showed a relationship between high environmental temperatures and human mortality rates in some areas. Previous studies use air temperature variables such as daily average temperature, maximum, and minimum air temperature, as environmental predictors that affect mortality rates [11]. In addition, other authors use an alternative temperature metric that combines several meteorological parameters (humidity, temperature,

Application and Analysis of Remote Sensing for the Initial Baseline …

851

wind speed, and solar radiation) into a single number, one of which is the Temperature–Humidity Index (THI). THI is a single value that describes the integrated effect of temperature and humidity on the degree of heat stress. Several studies use this index to show the heat load and the level of environmental comfort felt by humans that affect health and mortality rates, or indicate heat-free livestock that affects their productivity levels. The THI index has been used for various studies, one of which is to analyze the level of comfort in Jakarta as the current state capital [17]. The results of the analysis conducted by Wati and Fatkhuroyan [17], using data on air temperature and humidity from the Meteorology, Climatology, and Geophysics Agency (BMKG) meteorological stations in Kemayoran, Tanjung Priok, Halim Perdanakusuma, Cengkareng, and Pondok Betung, revealed that most areas of Jakarta have a THI of 25–27 C, which means that in the study area only some feel comfortable. This condition causes in general, in big cities such as Jakarta, the use of energy becomes relatively higher to obtain the optimum level of comfort. Basically, there are difficulties in obtaining the appropriate accuracy and validity of heat measurements to analyze environmental influences on the level of ecosystem productivity or the level of human health. The significance of the application of an environmental indicator, whether using air temperature or thermal index, cannot be fully ascertained, especially if it is used in different geographical areas. However, an approach is needed to analyze the effect of environmental temperature pressure on certain aspects. In this study, the THI thermal index is used to analyze the level of comfort felt by humans to live in urban areas, although this is relative, because it is influenced by the psychological condition of the population. People who tend to live in cold areas will find it difficult to feel comfortable living in hot areas. In this case, it is necessary to adapt to a different time for each individual. However, in general, there are patterns of perception of comfort, as Nieuwolt has studied in assessing comfort level based on THI value. In the process of calculation and analysis, in this study, the application of THI is carried out for six cities that are candidates for the capital city of Indonesia. The selected city was carried out based on a literature study of several mass media which provided information related to the opinion of the Head of State and Bappenas regarding the candidate for the capital city. Analysis of the comfort level in this study is expected to be one of the considerations in the selection of the national capital in terms of the environment and can support sustainable development goals. This consideration is also expected to complement other considerations in terms of land availability, land status, disaster-prone status, geographical position, security and politics, as well as the presence of residents as supporting activities. Remote sensing technology, that has developed today, has had a major impact on environmental studies. This is because remote sensing technology can provide data in a wide spatial range, including hard-to-reach areas. In addition, remote sensing data has a temporal resolution that can be adjusted to the needs of the analysis. The research uses Atmospheric Infrared Sounder (AIRS) sensor data which is an instrument on NASA’s Aqua satellite. In general, studies of the air comfort index have used AWS data which provides limited spatial analysis. By using AIRS data, the comfort index can be analyzed spatially with a wider area coverage and more

852

I. Susanti et al.

consistent time stages. In this case, the study of air comfort index using AIRS data is still very limited.

2 Data and Method The data used in this study is remote sensing data from the Aqua satellite Atmospheric Infrared Sounder (AIRS) sensor with spatial resolution 1°. The working principle of this sensor is to capture infrared radiation emitted by the earth’s surface and its atmosphere globally every day. The AIRS sensor can provide profile data for 20 levels of atmospheric altitude from the surface to the tropopause. The resulting data includes temperature, relative humidity, charge of carbon dioxide, methane, carbon monoxide, and several other parameters. At this time, with the development of existing algorithms, AIRS sensors can also detect volcanic smoke and forecast drought. The AIRS data used for this study is monthly data for the period January 2003 to January 2018. The use of data with the length of the period is expected to show a more general trend of changes due to certain factors, including climate change. The coverage area analyzed is Kalimantan, which is then focused on several selected areas, namely Balikpapan City, Gunung Mas Regency, Katingan Regency, Kutai Kartanegara Regency, Palangkaraya City, and Tanah Bumbu Regency. The AIRS data, then, analyzed the monthly THI pattern spatially for the Kalimantan region and temporally for the selected area. In addition, it is also analyzed using a time series to see trends in changes that occur, so that future comfort conditions can be estimated. The comfort index (THI) is calculated using the formula compiled by Nieuwolt [12], namely THI = 0.8T {(RH × T)/500} where T is the surface air temperature (˚C), RH is the relative humidity, and THI is the comfort index (˚C). As for the comfort limit, using a comfort interval based on Nieuwolt [12] and Emmanuel [4] was modified for a tropical climate [3]. Table 1 shows comfort index category based on Temperature–Humidity Index (THI) Table 1 Comfort index category based on THI

THI (˚C)

Categorized

21–24

Comfort

25–27

Comfort enough

>27

Not comfort

Application and Analysis of Remote Sensing for the Initial Baseline …

853

3 Results and Discussion Figure 1 is a map of the island of Borneo with seven potential new capitals of Indonesia. Most of the areas on the island of Borneo are lowlands, which can be seen from the dominant green color on the topographic map of the island of Borneo (Fig. 1). The average temperature throughout the year in the lowlands is 23–28 °C. Several options circulating in the community regarding the selection of the national capital include Pontianak, Palangkaraya, Katingan, Gunung Mas, Tanah Bumbu, Balikpapan, and Kutai Kartanegara. Today, the Ministry of State Secretary announced and designated Kutai Kartanegara as a candidate for the state capital. The question is, when viewed from the climatic conditions, is Kutai Kartanegara Regency quite comfortable as the nation’s capital? The analysis of this research will answer it. The results of remote sensing data processing show that THI in Kalimantan has quite large temporal and spatial variations, with a range of values between 21 and 26.8 °C. This is an advantage of using AIRS data, especially for monthly data, because it shows good data coverage, both temporally and spatially. In situ measurement data at this time have not been able to provide data that is spatially and temporally consistent. In January which represents the rainy season, the average THI on the island

Pontia

Kutai a Kartanegara Balikpa pan Katingan Palangkaraya

Fig. 1 Kalimantan map of topography and seven potential location, New Capital City, Indonesia

854

I. Susanti et al.

of Kalimantan is 24.5–26 °C where in this category the comfort index in Kalimantan is categorized as a partly comfortable area. In April which represents the transition season, the average THI in Kalimantan Island is 25–26.5 °C, including the category as a partly comfortable area with a comfort index of one level above January. The THI value in July and October which represents the dry season and relative transition is the same as the conditions in April, with the THI value being 25–26.5 °C which means that the conditions are partially comfortable. Kutai Kartanegara Regency itself, as the strongest candidate for the national capital, is in the range of THI values of 25.5–26.5 in four different months (January, April, July and October) which means it is in the partially comfortable category (Fig. 2). The choice of moving the capital, of course, is expected to be an area with the best comfort. However, in most areas of Kalimantan in January, March, July, and October the conditions are partly comfortable. Efforts to optimize the THI value into the comfortable category must be done to be able to reduce air temperature and increase humidity. Spatially, northern Kalimantan generally shows better comfort than southern Kalimantan (Fig. 2). This is influenced by the topography of the region, where the northern area tends to be higher than the southern area. Temporarily, variations in THI are influenced by monsoon patterns, where January which represents the rainy season period (December–January–February) is the period with the lowest THI. The difference in THI in January is about 1° lower compared to other periods. This difference has major implications for the various processes that occur on the earth’s surface. Humans can tolerate temperature increases of up to 2–3 °C. However, some types of vegetation and animals have high sensitivity to changes in temperature, so that a difference of 1° in comfort level can affect the level of ecosystem productivity. In general, seven (7) areas analyzed show high THI, which indicates a partially comfortable condition. Based on the cycle, the monthly pattern shows the highest value in May. The THI value in May is in the range of 25.4–26.2 °C, while the lowest THI value occurs in January with a range of values between 24.2 and 25.4 °C. However, Gunung Mas Regency is an area that has the lowest THI compared to six (6) other regions with differences about 1° for the average of all periods (January–December) (Fig. 3). The difference of 1 °C for the THI value in Gunung Mas Regency compared to other areas has major implications in various aspects, especially implications for energy needs to achieve the level of comfort required for humans. The use of electrical energy for air conditioning and food and beverage cooling machines in Gunung Mas Regency will be lower than in the other six alternative areas. In fact, in the December– January period comfortable conditions in Gunung Mas Regency formed naturally due to the support of weather and climate factors, so that energy use can be minimized, although in this case it has not been analyzed from the topography of the area. In addition, areas with comfortable air conditions are areas that have potential in the tourism sector. The level of air comfort in Gunung Mas Regency is better, in general it can also be seen from the average THI value for all analyzed periods as can be seen in Table 1.

Application and Analysis of Remote Sensing for the Initial Baseline …

855

Fig. 2 THI in Kalimantan for monthly average in the range 2003–2018

Based on the trend of developing climatic conditions, in addition to showing a lower THI value compared to other regions, THI in Gunung Mas Regency also shows a decrease, which means that changes in temperature and humidity as elements that determine THI values lead to better comfort conditions. Figure 4 shows that from January 2003 to December 2018, the THI value in Gunung Mas Regency has decreased even though the decline is at a low rate, which is 2 × 10–5 per month or 2.4 × 10–4°/10 year. However, this is better than what happened in Pontianak City, which experienced a significant increase in the THI value, which was 9 × 10–5 per

856

I. Susanti et al. THI (derajat Celcius) 27 26 25 24 jan feb mar apr may jun jul aug sep oct nov dec tanah bumbu palangka raya kangan

ponanak kutai kertanegara gunung mas

Fig. 3 Climatological average monthly THI (2003–2018) for the seven prospective capital cities

month or 1.08 °C/10 years. This tendency to increase the value of THI also shows the potential for increasing energy needs, in the value of seasonal rainfall, are more clearly seen when the maximum value is identified. The THI value in Kutai Kartanegara as the strongest candidate for the national capital is predicted to decrease by 1 × 10–5 per month or 1.2 × 10–3 °C/10 years (Fig. 4). This trend of decreasing THI value is expected to be able to anticipate the energy needs of the community in adjusting to the comfortable conditions in Kutai Kartanegara. However, to be able to stay in perfect climatically comfortable conditions, it takes a long time for the THI value to be in the overall comfortable category (Table 2).

Fig. 4 Trend of change in THI value per month according to the slope value of the regression equation with respect to time from January 2003 to December 2018 in °C

Application and Analysis of Remote Sensing for the Initial Baseline … Table 2 Average THI for the entire analysis period (2003–2018) for the seven prospective capital cities

857

Locations

Mean Temperature–Humidity Index (THI, °C)

Tanah Bumbu

25.6

Pontianak

25.6

Palangkaraya

25.8

Kutai Kartanegara 25.8 Katingan

25.8

Gunung Mas

24.9

Balikpapan

25.8

The downward trend in the value of THI in Gunung Mas is basically in line with the trend that occurred in the middle of the island of Borneo. In Fig. 4, it can be seen the trend of changes in the THI value per month according to the slope value of the regression equation with respect to time from January 2003 to December 2018 in °C. From the figure, it can be seen that in areas far from the coast, there is a trend of decreasing THI values, while in coastal areas, especially on the west coast of Kalimantan, there is a significant trend of increasing THI values, reaching more than 0.0025° per month, or 0.3°/10 years. From the existing values and the changing trends that occur in Kalimantan, especially for areas targeted to become the prospective capital city of Indonesia, in terms of climate and comfort level, Gunung Mas Regency is the optimum area to be developed as the nation’s capital. Apart from minimizing the use of energy, the level of comfort is related to the optimization of human resources. People will tend to be more productive if the surrounding environmental conditions are quite supportive. The trend of environmental development that occurred in the analyzed period cannot be used as a benchmark that the change is a permanent trend. There is a possibility that the downward trend in the THI value will change if the character of the region changes which has implications for changes in the energy balance, including air temperature. The conversion of forest land into a built-up area will have a major impact on the character of the area, including changes in its ecosystem. These changes have been widely proven through various studies which state that rapid urbanization leads to an imbalance between needs and needs resource availability, infrastructure, and climate change [9]. This is the implication of the acceleration of population growth and its needs. Consideration of environmental changes and availability of resources in determining the national capital is important. In the development of infrastructure and other functions as a supporter of the capital, it is unavoidable to avoid changes in the character of the region, which in general is a change from the natural environment to the built environment and built area. It also forms the dominance of artificial elements that have a positive role in promoting energy consumption [10,16]. Grimm et al. [6] and Pickett et al. [13] also demonstrated abiotic changes in the remaining parts of the habitat, caused by increases in temperature, rainfall, humidity, and changes in the nitrogen cycle [6, 13].

858

I. Susanti et al.

4 Conclusion The AIRS data obtained shows consistent data coverage, both spatially and temporally, so that the full THI value can be obtained for the entire period and the entire scope of the analyzed area. Based on the calculation of the comfort index value represented in the Temperature–Humidity Index (THI), Gunung Mas Regency has the potential for better climatic comfort compared to the other six districts. Kutai Kartanegara Regency as a prospective capital city also has a fairly good comfort index in the category of partially comfortable (THI value 25.5–26.0), but it is also predicted that the THI value will decrease by 0.0012 °C per 10 years and will be more quickly if the flow of infrastructure development increases.

References 1. Basu, R.: High ambient temperature and mortality: a review of epidemiologic studies from 2001 to 2008. Environ. Health 8(1), 40 (2009) 2. Bonasir, R.: Rencana pemindahan ibu kota dari Jakarta: Berapa lama waktu yang diperlukan dan apasajasyarat-syaratibu kota?. Internet:www.bbc.com/indonesia/indonesia-48093451. BBC News Indonesia (2019). Diakses Sept (2019) 3. Effendy, S.: Keterkaitan ruang terbuka hijau dengan urban heat island wilayah Jabotabek. IPB (Bogor Agricultural University), Disertasi (2007) 4. Emmanuel, R.: Thermal comfort implications of urbanization in a warm-humid city: the Colombo Metropolitan Region (CMR), Sri Lanka. Build. Environ. 40(12), 1591–1601 (2005) 5. Foresman, T.W., Pickett, S.T., Zipperer, W.C.: Methods for spatial and temporal land use and land cover assessment for urban ecosystems and application in the greater BaltimoreChesapeake region. Urban ecosystems 1(4), 201–216 (1997) 6. Grimm, N.B., Faeth, S.H., Golubiewski, N.E., Redman, C.L., Wu, J., Bai, X., Briggs, J.M.: Global change and the ecology of cities. Science 319 (5864), 756–760 (2008) 7. Hajat, S., Kosatky, T.: Heat-related mortality: a review and exploration of heterogeneity. J. Epidemiol. Community Health 64(9), 753–760 (2010) 8. Igono, M., Bjotvedt, G., Sanford-Crane, H.: Environmental profile and critical temperature effects on milk production of Holstein cows in desert climate. Int. J. Biometeorol. 36(2), 77–87 (1992) 9. Liu, X., Peng, D.: Study on the threshold effect of urbanization on energy consumption. Theoret. Econom. Lett. 8(11), 2220 (2018) 10. Ma, H., Xu, J., Wang, P.: Water resource utilization and China’s urbanization. Resour. Sci. 36(2), 334–341 (2014) 11. Morabito, M., Crisci, A., Moriondo, M., Profili, F., Francesconi, P., Trombi, G., Bindi, M., Gensini, G.F., Orlandini, S.: Air temperature- related human health outcomes: current impact and estimations of future risks in central Italy. Sci. Total Environ. 441, 28–40 (2012) 12. Nieuwolt, S.: In: Tropical Climatology. An introduction to the Climates of the Low latitudes. Wiley (1977) 13. Pickett, S.T., Cadenasso, M.L., Grove, J.M., Boone, C.G., Groffman, P.M., Irwin, E., Kaushal, S.S., Marshall, V., McGrath, B.P., Nilon, C.H.: Urban ecological systems: scientific foundations and a decade of progress. J. Environ. Manage. 92(3), 331–362 (2011) 14. Pickett, S.T.A., Burch, W.R., Dalton, S.E. et al.: A conceptual framework for the study of human ecosystems in urban areas. Urban Ecosys 1, 185–199 (1997). https://doi.org/10.1023/ A:1018531712889

Application and Analysis of Remote Sensing for the Initial Baseline …

859

15. Tobias, A., Armstrong, B., Zuza, I., Gasparrini, A., Linares, C., Diaz, J.: Mortality on extreme heat days using official thresholds in Spain: a multi-city time series analysis. BMC Public Health 12(1), 133 (2012) 16. Wang, L., Wei, H.: An empirical study on the impact of urbanization on energy consumption in China. Resour. Sci. 36(2), 1235–1243 (2014) 17. Wati, T., Fatkhuroyan, F.: Analisis Tingkat Kenyamanan Di DKI Jakarta Berdasarkan Indeks THI (Temperature Humidity Index). Jurnal Ilmu Lingkungan Undip 15(1), 57–63 (2017) 18. Wu, W., Xiao, Y., Li, G., Zeng, W., Lin, H., Rutherford, S., Xu, Y., Luo, Y., Xu, X., Chu, C.: Mortality on extreme heat days using official thresholds in Spain: a multi-city time series analysis. Sci. Total Environ. 449, 355–362 (2013)

A Model to Describe Elastic Light Scattering of a Single Sphere in a Non-homogeneous Illuminating Light Measured in an Optical Particle Counter Moch S. Romadhon

Abstract Optical particle counters (OPCs) are widely used in the studies of atmospheric particles. The main principle of OPCs is to illuminate particles and retrieve particle size based on the light scattered by the particles. Ideally, illuminating light should be homogeneous so that scattered light is independent of particle position. However, this requirement is practically difficult. This paper demonstrates the effect of the non-homogeneity of illuminating light on the relation between scattered light and particle size. The effect was studied theoretically and experimentally. In this study, scattered light was measured by a Photomultiplier Tube (PMT). The scattered light can be characterised by two parameters: PMT pulse width and depth. The width indicates the residence time of particles in a sensing volume. The depth indicates the intensity of scattered light. The measurements result in multi-modal distributions of pulse depth as a function of particle size. Based on the distribution, the retrieval of particle size using PMT pulse depth can be misleading. This feature should be considered in designing an OPC or interpreting OPC outputs.

1 Introduction Optical particle counters (OPCs) are widely used in the continuous measurements of atmospheric particle size distribution to study aerosols impact on climate and public health [1–3]. Their measurement principle is to illuminate particles in their sensing volume and develop relationship between the light scattered by the particles and the particle size [4]. Ideally, the illuminating light is homogeneous across the sensing volume so the scattering light is independent of particle position. However, one of the main problems undermining the accuracy of OPCs is the non-homogeneity of the illuminating light [5, 6]. In a particular case where the shape of illuminating light M. S. Romadhon (B) Center for Climate and Atmosphere, National Research and Innovation Agency, Jakarta Pusat, Indonesia e-mail: [email protected] URL: https://www.aerosols.web.id/ © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_79

861

862

M. S. Romadhon

is Gaussian, the character of the light scattering has been studied theoretically [7]. However, there is no study for the case of arbitrary non-homogeneous plane wave of illuminating light either theoretically or experimentally. Pragmatically, OPCs manufacturers set their instruments sampling rate in a level that particles are aerodynamically focused on a specific position in the sensing volume. In a particular OPC design that requires relatively low sampling rate that particles can not be aerodynamically focused, measurement bias due the non-homogeneity can be significant. This paper presents the study for that case based on measurements using SPARCLE: an OPC to measure particle size and refractive index [8].

2 Design and Calibration A part of SPARCLE design is shown in Fig. 1. One of the detectors used in the design is Photomultiplier Tube (PMT). The PMT can be used both to indicate the presence of particles in the sensing volume and to estimate particle size. This type of measurement is commonly used by commercial OPCs. The PMT is installed with its area vector direction along the x axis. The distance of the PMT from the centre of reference is around 6 mm. The diameter of the PMT window is around 25 mm. In SPARCLE, particles are transported from the sampling pipe to the sensing volume and removed through exhaust pipe. The path of particles in the sensing volume is modelled to be parallel to the y axis. In the sensing volume, particles are illuminated by a compressed laser beam. The beam intensity varies across the x–y plane. The residence time of particles in the sensing volume is expected around 1,000 ns by setting SPARCLE sampling flowrate around 0.2 cm3 s−1 .

Fig. 1 Part of SPARCLE design. Particles are sampled through the sampling pipe and transported to the sensing volume. In the sensing volume, particles are illuminated by a compressed laser beam. The compression is done by a cylinder lens

A Model to Describe Elastic Light Scattering of a Single Sphere …

863

Fig. 2 One example of PMT signals. The pulse is plotted as a red line containing noises in the form of very high-frequency signals. The noise can be removed by applying a Savitzky–Golay filter. The filtered signals are plotted as a green line. Also shown in the figure, trigger signals which are generated when PMT signals reached a threshold level. The threshold level was set at 0.4 V lower than the mean background

The design was calibrated using four monodisperse solid aerosols. The aerosols were generated in two steps: first, by nebulising suspensions containing polystyrene latex (PSL) beads into droplets and, second, drying the droplets to form solid particles. The mean particle sizes of the four aerosols are 1100, 1800, 2000 and 3000 nm. The refractive index of the PSL beads is 1.59. One example of PMT pulses as a response to a PSL bead is shown in Fig. 2. As shown in the figure, PMT pulses can be characterised by two parameters: pulse depth and pulse width. The width indicates the residence time of particles in a sensing volume. The depth indicates the intensity of scattered light. Due to background in the sensing volume, a threshold level was applied to differentiate between scattering signals and the background. PMT pulses shallower than the threshold were omitted. In the suspension nebulisation, some droplets might not contain PSL beads and, when they were dried, residual particles were formed from solidification of surfactants contained in the suspensions. The surfactant was added by the suspension manufacturers to stabilise the suspension. Measurements by commercial GRIMM 1.108 OPC indicated that the size of residual particles were normally distributed peaked between 600 and 700 nm.

864

M. S. Romadhon

3 Model Description The intensity of scattering light can be calculated using Mie theory assuming its shape is spherical [9]. In spherical coordinate, the intensity is expressed as λ2 I = 4π 2

φ2 θ2



 I (α, m, θ, φ) + I⊥ (α, m, θ, φ) G(θ, φ) dθ dφ,

(1)

φ1 θ1

where λ is the wavelength of the illuminating light; φ is azimuthal angles defined as the angle between the plane oscillation of the electric vector of the illuminating light and scattered light directions; θ is scattering angles defined as the angle between the propagation direction of the illuminating light and scattered light directions; α is a size πd parameter defined as α = λ p with dp as the particle diameter; G(θ, φ) is geometrical factor specific to the instrument’s design; I and I⊥ are scattered light intensities propagated in parallel and perpendicular, respectively, relative to the plane oscillation of the electric vector of the illuminating light. The intensities are expressed as I = |S1 (α, m, cos θ j )|2 sin2 φ j I⊥ = |S2 (α, m, cos θ j )|2 cos2 φ j

(2)

where S1 and S2 are Mie amplitude functions [10]; scattered light I is collected over a solid angle represented by an arc of angle φ1 – φ2 swept around a sphere by θ1 – θ2 . In a particular case where particles are illuminated in the centre of reference, scattered light collected by the PMT is the light over a solid angle represented by an arc of azimuth angles from φ1 = 50◦ to φ2 = −50◦ swept around a sphere from θ1 = 20◦ to θ2 = 160◦ . Since the beam intensity is distributed in the x and y axis, the intensity of scattering light measured by the PMT is a function of particle position along the x and y axis. The scattered intensity denoted as I1 (x, y, α, m) can be expressed as I1 (x, y, α, m) = f (x, y) I0 (α, m),

(3)

where f (x, y) is normalised distribution function of the beam intensity; I0 (α, m) is the intensity of scattering light when particles are illuminated by a beam with the intensity APbb where Pb is the beam power and Ab is the beam cross section. The power of scattered light measured by the PMT, denoted as P0 (α, m), can be calculated numerically by dividing the area of the PMT window into N small areas over which I1 is assumed to be homogeneous and adding up all the power of scattering light over the small areas. The power P0 (α, m) can be expressed mathematically as

A Model to Describe Elastic Light Scattering of a Single Sphere …

865

  N 1 Pb  |S1 (α, m, cos θ j )|2 sin2 φ j Re Ab k 2 j=1 r 2j  2 2 ˆ + |S2 (α, m, cos θ j )| cos φ j rˆ j . A j A j .

P0 (α, m) =

(4)

where k is wave number calculated by 2π ; A j is the small area indexed by j; Aˆ j λ is the unit vector of A j ; r j is the distant of A j from the centre of reference. The conversion of the power of scattered light into PMT outputs can be calculated as V (x, y, α, m) = P1 (x, y, α, m)

λ η

PMT

1.24



G PMT Rlo G amp ,

(5)

where ηPMT is the quantum efficiency of the photocathode on the PMT window; λ ηPMT is the conversion factor of light power into electrical currents with the unit of 1.24 λ is nm; G PMT is the gain of the PMT; Rlo is the load resistor installed on the PMT outlet; G amp is the amplification factor of the amplifier. The depth of PMT pulses can be regarded as the difference between PMT outputs when y = 0 and the background in SPARCLE scattering chamber. The depth is expressed mathematically as VD (x, α, m) = V (x, y = 0, α, m) − Vb ,

(6)

where VD is PMT pulse depth; V (x, y = 0, α, m) is PMT outputs when y = 0; Vb is the background. In the calibration, Vb = 3.1 V. If PMT pulses can be assumed as normally distributed then the width of PMT pulses can be calculated as

T (x, z, α, m) =

  2 σ y 2 ln VD (x,α,m) VB v(x, z)

.

(7)

where T is PMT pulse width; σ y is the standard deviation of PMT pulses; v(x, z) is the speed of air at particle position (x, z). A modelling of the speed of air in the sensing volume using OpenFoam indicates that the speed is normally distributed around the symmetrical axis of the sampling pipe. The speed is expressed as v(x, z) =

Q −(x 2 +z2 2 ) e 2 σv , π

(8)

where Q is SPARCLE sampling flow rate; σv is the standard deviation of the normally distributed air speed and can be determined from OpenFoam simulation. The probability of a particle to be at a particular position is proportional to the speed of air in which the particle suspended. The probability can be expressed as

866

M. S. Romadhon

p(x, z) = B v(x, z),

(9)

where B is a normalisation coefficient. The probability of the PMT to generate pulses with the depth of VD as responses to particles of size parameter αo and refractive index m can be calculated by integrating the probability of particle position where the light they scattering leads to PMT pulse depth equal to VD . The calculation can be expressed mathematically as p(VD (αo , m)) =



p(xi  , z i  ),

(10)

i

where (xi  , z i  ) is particle position where VD (xi  , α, m) = VD (αo , m). Similarly, the probability of pulse width T  as responses to the same particles can be calculated as p(T  (αo , m)) =



p(xi  , z i  ),

(11)

i 

where (xi  , z i  ) is particle position where T (xi  , z i  , α, m) = T  (αo , m).

4 Results and Discussion The distribution of PMT pulse depth and width as measured in SPARCLE calibration are shown as blue bars in Fig. 3. The average sampling rates of measurements are also indicated in the figure. As shown in the figure, the shape of pulse width distribution is skewed right with the range wider for slower sampling rate. The shape might be due to the variation of SPARCLE sampling flow rate resulted in various residence time of particles in the sensing volume. Meanwhile, the shape of pulse depth distributions is a multi-mode distribution with more than one peak identified. This results lead to a misleading interpretation when particle size is associated with PMT pulse depth. Equations (10) and (11) are used to calculate the distributions of PMT pulse depth and width. The results of the calculation for PSL beads and residual particles are shown as orange and magenta dashed lines, respectively, in Fig. 3. Particle sizes used in the calculation are indicated in the figure. As shown in the figure, the shape of the calculated and measured distributions are similar with some discrepancies in fine structures of the distributions. The similarity indicates that the model is sufficient to describe the scattering light measured by the PMT.

A Model to Describe Elastic Light Scattering of a Single Sphere …

867

Fig. 3 Distribution of pulse widths and pulse depths. The blue areas are the measured distributions generated from the calibration data. The dashed lines are the distributions calculated from the simulations. The simulated distributions have been scaled up to fit with the measured distributions. The scaling factor for each simulation was searched manually and the fit was evaluated visually. The simulations assumed that the size distributions of aerosol particles are bimodal distributions which peaked at two sizes. The measured distributions can be approximated as the combination of the simulated distributions for the two sizes. The two sizes are listed in the figure

5 Conclusion A model to describe the distribution of PMT pulse depth and width generated from the light scattered by particles illuminated in non-homogeneous beam intensity was developed. The comparison between the model and experimental results indicates the

868

M. S. Romadhon

similarity of the shape of the distributions. The measurements result in multi-modal distributions of pulse depth as a function of particle size. Based on the distribution, the retrieval of particle size using PMT pulse depth can be misleading. Since most commercial OPCs use the relationship between the two, it is important to consider the bias that is introduced from the non-homogeneity of beam intensity used in the commercial OPCs. Acknowledgements This research is funded by Lembaga Pengelola Dana Pendidikan, Kementerian Keuangan, Indonesia. This research was performed at Aerosol Lab. University of Oxford under supervision of Prof. Don Grainger, Dr. Dan Peters and Dr. Simon Proud.

References 1. Jianmin, C., Li, C., et al.: A review of biomass burning: Emission and impacts on air quality, health and climate in China. Sci. Total Environ. 579, 1000–1034 (2017). https://doi.org/10. 1016/j.scitotenv.2016.11.025 2. Myhre, G., et al.: Anthropogenic and natural radiative forcing. In: Stocker, T.F., et al. (eds.) Climate Change 2013: The Physical Science Basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York, NY, USA (2013) 3. Wooster, M., et al.: New tropical peatland gas and particulate emissions factors indicate 2015 Indonesian fires released far more particulate matter (but less methane) than current inventories imply. Remote Sens. 10, 495 (2018). https://doi.org/10.3390/rs10040495 4. Pramod, K., Baron, P.: Aerosol Measurement Principles, Techniques and Applications, chapter 5. https://doi.org/10.1016/0021-8502(80)90037-3, http://linkinghub.elsevier.com/retrieve/pii/ 0021850280900373 (2011). ISBN 9780470387412 5. Renliang, X.: Light scattering: a review of particle characterization applications. Particuology 18, 11–21 (2015). https://doi.org/10.1016/j.partic.2014.05.002 6. Romanov, A.V., Yurkin, M.A.: Single-particle characterization by elastic light scattering. Laser Photonics Rev. 15 (2021). https://doi.org/10.1002/lpor.202000368 7. Zhai, C., Cao, Z.: Elastic light scattering of a single sphere in a Gaussian light sheet. Optik-Int. J. Light Electron. Opt. 194 (2019). https://doi.org/10.1016/j.ijleo.2019.163083 8. Romadhon, M.S.: SPARCLE: an instrument for aerosol size and refractive index measurement (2021). https://ora.ox.ac.uk/objects/uuid:3f586dca-7d46-42e9-a468-d8d26f894ce2 9. Mie, G.: Beiträge zur Optik trüber Medien, speziell kolloidaler Metallösungen. Ann. Phys. 330(3), 377–445 (1908) 10. Hulst, H.C.: Light Scattering by Small Particles. Dover Publications Inc., New York (1957)

Analysis of Atmospheric Conditions on Hail Events in Bandung on March 8, 2022 Elfira Saufina, Tiin Sinatra, Anis Purwaningsih, Dita Fatria Andarini, Fahmi Rahmatia, Fadli Nauval, Ina Juaeni, Asif Awaludin, Aisya Nafiisyanti, Farid Lasmono, Adi Witono, Arief Suryantoro, Eddy Hermawan, and Acep Catur Nugraha Abstract Hail is one of the atmospheric phenomena generally caused by weather anomalies due to convective storms with adequate updraft strength, sufficient availability of supercooled liquid water content, and conducive temperature and optimal time. This paper aims to analyze the atmospheric mechanism when hail occurred in Bandung on March 8, 2022. We used data analysis of the rain observation system by rain scanner to observe rain’s distribution and temporal evolution. Also, we used analysis of satellite imagery data to observe the cloud type and cloud height during hail phenomenon in Bandung. Moreover, the atmospheric profile during the hail precipitation is analyzed using ERA5 data for several supporting parameters such as ice water content (ICW), liquid water content (LCW), moisture, vertical wind, and temperature profiles. The results indicate that heavy rain in the middle of Bandung area (study case: Antapani, Cicadas) was observed since 14:14 Local Time (LT) with rapid rain growth for approximately 1 h. The rain propagates and expands from south to north Bandung and converges with another cloud system developed in the northern area of Bandung. The rain is indicated to come from height clouds based on Himawari imagery. Persistent rainfall with a wide coverage lasts up to 15:30 LT and indicates cloud growth that reaches the height of the tropopause and stratosphere. This high cloud cluster is suspected to form ice precipitation falling to the study area at 14:40 LT. An indication of strong wind convergence since 14:00 LT is also one of the prominent features before the hail precipitation occurred over the study area.

E. Saufina (B) · T. Sinatra · A. Purwaningsih · D. F. Andarini · F. Nauval · I. Juaeni · A. Awaludin · A. Nafiisyanti · A. Witono · A. Suryantoro · E. Hermawan · A. C. Nugraha Research Center for Climate and Atmosphere, National Research and Innovation Agency (BRIN), Bandung, West Java 40173, Indonesia e-mail: [email protected] F. Rahmatia Agam Space and Atmospheric Observation Station, National Research and Innovation Agency, Bukit Kototabang, Koto Rantang Palupuh Agam, West Sumatera, Indonesia F. Lasmono Center for Data and Information, National Research and Innovation Agency (BRIN), Jakarta 10340, Indonesia © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2023 A. Basit et al. (eds.), Proceedings of the International Conference on Radioscience, Equatorial Atmospheric Science and Environment and Humanosphere Science, Springer Proceedings in Physics 290, https://doi.org/10.1007/978-981-19-9768-6_80

869

870

E. Saufina et al.

1 Introduction Extreme weather is a natural disaster that has a detrimental impact on human life. Extreme weather can be characterized by heavy rain, lightning, and/or ice. An extreme weather event in the form of heavy rainfall with hail has been reported to occur on Tuesday, March 8, 2022, in the city of Bandung around the Cicadas and Antapani areas (central Bandung). This hail event occurred around 14:40–15:00 LT as reported by the media [1]. Based on World Meteorological Organization (WMO) definition, hail is the precipitation of either transparent or partly or completely opaque particles of ice (hailstones), usually spheroidal, conical, or irregular in form and of diameter generally between 5 and 50 mm which falls from a cloud either separately or agglomerated into irregular lumps. In general, hail occurs much in extra-tropical areas because it has a relatively lower freezing level. The freezing level is the height of the atmospheric layer with a temperature of 0 °C so that water drops can freeze [2]. Hail can occur in the tropics if the ice particles that fall from Cumulonimbus clouds are very large and fixed in the form of ice particles despite friction in the clouds. The temperature at the top of the cloud can reach −60 °C, and water vapor is generally found in the form of ice crystals [3]. Hail forms as a result of collision and coalescence processes from cloud drops that are supersaturated and produce large raindrops in the frozen phase [4]. Hail is usually local and uneven and occurs suddenly [5, 6]. In addition, hail generally occurs during the transition period and is accompanied by heavy rains accompanied by lightning and strong winds [7]. In Indonesia, the hail phenomenon rarely occurs because of the higher freezing level, so the ice crystals formed in this layer will melt into water when they fall to the ground [6]. Hail events can be triggered by overshooting phenomena or strong updraft activity in Cumulonimbus cloud cells. Cloud overshooting occurs when clouds that normally only grow in the troposphere then penetrate the stratosphere [2, 8]. There are several cases of hail in Indonesia that have been recorded and analyzed for atmospheric conditions, including the case of hail in Bogor on September 23, 2020 [9], heavy rain with ice in Bandung on March 17, 2017, which concluded that local convective activity and wind convergence in the study area during hail events [10], hail case in Jakarta on November 22, 2018, hail in Bandung on May 3, 2017 [2], and hail in Bandung on April 19 and 23, 2017 [3]. Of the various hail cases that occurred in the western part of Java Island, Bandung is the area that experiences the most hail events. Bandung is often one of the locations where extreme weather is possible due to its geographical location and topographical contours. Regional and local factors such as the effects of tropical cyclones, eddies, and wind bends (shear line) causing convective clouds to become one of the causes of extreme weather over Bandung [11]. Therefore, this paper aims to analyze more in the case of hail that occurred in Bandung, in particular, to determine the atmospheric conditions during the extreme weather phenomenon of hail in Bandung on March 8, 2022.

Analysis of Atmospheric Conditions on Hail Events in Bandung …

871

2 Data and Method 2.1 Precipitation Analysis The temporal evolution of rain was analyzed descriptively based on the precipitation reflectivity data captured by the rain scanner developed from the x-band marine radar as part of atmospheric observation. The coverage of this radar can reach 44 km [12]. The temporal resolution of the rain scanner data is 2 min with a spatial resolution of 120 × 120 m2 . With a sufficiently high temporal and spatial resolution, rain scanners can be used to see the evolution of rain during hail events. The reflectivity value indicates the presence of an object that produces an echo in the atmosphere. The higher the reflectivity value, the larger the size of the object [13]. In general, objects that can reflect high echo values are convective clouds containing large particles such as ice grains [14]. Rain scanner data installed in Sumedang (6.912°S, 107.838°E) was used to analyze hail occurrence on March 8, 2022.

2.2 Cloud Identification and Analysis This study utilized the cloud properties/physical derived from Himawari level-2 products to identify the cloud characteristics. Himawari-8 is a satellite that was launched in 2015 with 10 min temporal resolution and has 16 channels as the latest generation of Multi Transpose Satellite-2 (MTSAT-2) observing the growth of convective clouds in more detail [15]. We obtained the cloud type, height, and temperature from Himawari product cloud properties with results based on the High-Resolution Cloud Analysis Information (HCAI) algorithm developed by the Japan Meteorological Agency. Furthermore, the vertical structure of the cloud was analyzed by identifying the fraction of cloud cover (CC), cloud liquid water content (CLWC), and cloud ice water content (CIWC). We obtained those parameters from the European Center for Medium-Range Weather Forecasts (ECMWF) atmospheric reanalysis of the global climate (ERA5) on pressure-levels data, for twelve different altitudes (from 1000 to 100 hPa) over 6.6–7.2 S and 107.4–107.9 E [16]. The data can be downloaded from http://www.ecmwf.int/. In addition, the surface wind data at 925 hPa pressure level from ECMWF was used to analyze the wind convergence and wind direction before over the study area before and during ice precipitation.

872

E. Saufina et al.

3 Result and Discussion 3.1 Temporal Evolution of Precipitation Based on Rain Scanner Observation Figure 1 shows the reflectivity distribution during the hail event on March 8, 2022, from a rain scanner located at Sumedang. The figure shows that it has rained around the city of Bandung including Antapani and Cicadas areas which were reported to experience heavy rain with ice on the day [1]. Rain scanner monitoring results show that light rain began to be detected in the east of Antapani and Cicadas at 13:52 LT (blue circle) and at 13:54 LT, it detected rain in the north of Antapani and Cicadas (red circle). Rainfall in both places is widespread as the intensity increases, which is indicated by an increase in the reflectivity value. At 14:14 LT the two rain objects began to unite (black circle). Meanwhile, rain objects in the southern area move closer to Antapani and Cicadas (green circle). At 14:24, the rain in the southern part split into two parts, one of these split parts began to merge with rain objects in the Antapani and Cicadas areas. At 14:30 LT, a fairly wide object of rain that extends vertically with a maximum reflectivity value (> 35 dBZ) is above the hail incident location. It was seen that the reflectivity value had started to decrease over Antapani and Cicadas at 14:46 LT. This continued until the rain stopped at around 15:28 LT.

Fig. 1 Precipitation scanner observation during hail event in Bandung on March, 8, 2022

Analysis of Atmospheric Conditions on Hail Events in Bandung …

873

3.2 Wind Analysis Wind convergence and wind direction over the study location were analyzed before and during the ice precipitation (Fig. 2). A convergence pattern shown by positive value (red contour) was detected from the morning at 10:00 LT (not shown). The convergence strengthened in time over the study area (black rectangle in Fig. 2) and reached its strongest magnitude (>9 × 10–5 s−1 ) at 13:00 LT on March 8, 2022, approximately two hours before the hail event over Bandung (Fig. 2). The air masse’s confluence forces air’s upward motion over this area, thus supporting the formation of convective clouds. Relatively stronger convergence activities were still persistent over some locations, and this wind convergence continuously formed higher convective clouds that potentially resulted in ice precipitation. Moreover, during the ice precipitation (around 14:00–15:00 LT), the wind vectors supported the propagation of the rain object shown by the rain scanner in Fig. 1. Overall, convergence is one of the indications of the beginning of ice precipitation, as stated by BMKG that convergence is one of the causes of hail in Bandung on March 8, 2022 [17]. The identification of clouds formed by this process is discussed in the following subsection, by looking at cloud temperature, cloud top height, and the vertical structure of the clouds.

Fig. 2. 925 hPa wind convergence and direction at 12:00–15:00 LT March, 8, 2022

874

E. Saufina et al.

3.3 Cloud Identification and Vertical Analysis Cloud Identification. The analysis of cloud identification was begun by investigating the cloud evolution from the spatial distribution of brightness temperature/temperature black body (TBB) of Himawari-8 Satellite over Bandung Area (black rectangle). Figure 3 shows the convective clouds observed in the Bandung area indicated by a cool TBB (= 5 mm/h and the standard deviation > 1.5 mm/h, while stratiform rain is classified with standard deviation