Integrating Flood Susceptibility and Deforestation Mapping for Climate Vulnerability Assessment: A Geospatial and AI-Based Approach. (Papers Track)
Serah Akojenu (Data Scientists Network); Chinazo Anebelundu (Data Scientists Network); Godwin Adegbehinde (Data Scientists Network); Olamide Shogbamu (Data Scientists Network); Blessing Agboola (Data Scientists Network); Tochuckwu Abia (Data Scientists Network); Anthony Soronnadi (Data Scientists Network)
Abstract
Climate change has intensified global natural disasters, with floods and droughts posing the greatest threats to human settlements and economic systems. In Nigeria alone, flooding causes 80% of climate-induced deaths and inflicts catastrophic economic losses of 60 billion dollars annually, while persistent drought threatens the food security of millions in northern regions, which contribute significantly to the food consumed across the country. Despite these devastating impacts, existing machine learning disaster prediction studies focus on single hazards with limited scope and insufficient risk indicator integration, leaving communities vulnerable and unprepared. This study addresses this critical research gap by developing a comprehensive Climate Vulnerability Index (CVI) that integrates flood susceptibility and drought risk mapping across Nigeria using advanced Random Forest, XGBoost, and Light GBM algorithms with multi-criteria decision analysis. This framework represents a shift from fragmented, single-hazard approaches to a unified, multi-hazard assessment system that incorporates diverse risk indicators, creating a scalable tool essential for protecting lives and economic stability at national, state, and local levels.