DEVELOPMENT OF PREDICTIVE MODEL USING RANDOM FOREST FOR COMMUNITY-LED WATER SAFETY IN FLOOD-PRONE AREAS OF NIGERIA

Flood-induced deterioration of water quality remains a persistent challenge across many flood-prone communities in Nigeria, where limited monitoring infrastructure constrains timely risk assessment and response. This study developed a Random Forest-based predictive framework for assessing flood-related water quality risks through the integration of multi-source geospatial and environmental datasets, including Sentinel-1 synthetic aperture radar imagery, Sentinel-2 optical imagery, Global Precipitation Measurement (GPM) rainfall data, topographic variables, historical flood records, and Internet of Things (IoT) sensor observations collected between 2018 and 2024. Data preprocessing involved z-score normalization, Principal Component Analysis, and Box–Cox transformation to improve model stability and predictive performance. The model achieved an overall accuracy of 88.4% and an Area Under the Curve (AUC) of 0.93. Feature importance analysis identified rainfall intensity and flood proximity as dominant predictors of water quality risk. The proposed framework offers a scalable decision-support tool for strengthening flood preparedness, water safety management, and climate resilience in data-constrained environments.