owusugeorge946-afk/Volta-Flood-Susceptibility-ML: v1.1.1– Volta River Basin Flood Susceptibility Assessment

Initial public release of the code and reproducible workflow supporting the study, "Spatially Validated Machine-Learning Flood Susceptibility Assessment in the Volta River Basin Using Extreme Rainfall and Multi-Source Geospatial Data." This release provides the Google Earth Engine and Google Colab workflows used for multi-source predictor preparation, machine-learning modelling, spatial holdout validation, model interpretation, and basin-wide flood-susceptibility assessment. The framework integrates terrain, drainage, extreme rainfall, vegetation, land-cover, and soil information and compares Random Forest and XGBoost models using spatially independent validation. The repository includes code for constructing 15 environmental predictors, implementing the 0.5° spatial-block validation framework, evaluating model performance, generating Random Forest predictor importance and SHAP interpretation, and producing basin-wide flood-susceptibility probability and classified susceptibility outputs. Authors: George Owusu Amoah, Francis Quayson, and Crispin Awodanzo Ajugu. Corresponding author: George Owusu Amoah, Department of Geography and Regional Planning, University of Cape Coast, Ghana. Release: v1.1.1 Year: 2026