Monsoon Flood Mapping in Eastern Balochistan Using Machine Learning

Abstract

Eastern Balochistan, encompassing the low-lying districts of Nasirabad, Jaffarabad, and Jhal Magsi, is highly vulnerable to monsoon-induced flooding, which severely impacts agriculture, infrastructure, and livelihoods. In recent decades, the frequency and intensity of flood events have increased, necessitating advanced, data-driven approaches for flood hazard assessment and management. This study combines climate analysis and remote sensing–based flood mapping using Google Earth Engine (GEE) to assess flood risk from 1980 to 2024. Temporal and spatial trends of rainfall and temperature were analyzed using CHIRPS and TerraClimate datasets, revealing periods of extreme rainfall that contribute to heightened flood risk. Historical and recent flood events from 1988 to 2024 were further analyzed using Landsat and Sentinel satellite imagery processed in GEE. Two supervised machine learning algorithms—Random Forest (RF) and Support Vector Machine (SVM)— were applied to classify flooded and non-flooded areas using multispectral and Synthetic Aperture Radar (SAR) imagery. Model training and validation were conducted using benchmark flood years (2010 and 2022), followed by predictive assessments for subsequent events in 2012, 2013, 2023, and 2024. The RF classifier achieved overall accuracies of 88% and 82% (Kappa coefficients of 0.77 and 0.64) for the respective validation years, while the SVM model achieved 95% and 92% accuracy (Kappa = 0.91 and 0.84), effectively delineating flood extents with minimal errors. Both models consistently identified the southeastern floodplains as persistently vulnerable to inundation, reflecting the combined influence of topography, extreme rainfall events, and hydrological dynamics. The results demonstrate that machine learning–based flood mapping in GEE provides a reliable, scalable, and transferable framework for flood risk assessment in data-scarce regions, offering critical insights for disaster preparedness, infrastructure planning, and climate adaptation strategies in eastern Balochistan.