High-Resolution (30 m) Flood Susceptibility Mapping Data of Pakistan

Dataset Description This dataset provides national-scale, high-resolution (30-meter) Flood Susceptibility Maps (FSM) for Pakistan. These rasters were generated using an integrated framework combining Machine Learning (ML) and Explainable AI (EAI), as detailed in the research article "High-resolution flood susceptibility mapping and exposure assessment in Pakistan" published in the International Journal of Disaster Risk Reduction. → Project Updates & Code:For Google Earth Engine (GEE) applications, dataset details, and related source codes, please visit the official GitHub repository:https://github.com/waleedgeo/fsm-pk Files Included The dataset includes two raster layers representing flood susceptibility predictions derived from Light Gradient Boosting Machine (LGBM) and eXtreme Gradient Boosting (XGBoost) models. 1. fsm_lgbm_pakistan.tif Model: Light Gradient Boosting Machine (LGBM) Performance: This model demonstrated the highest reliability in the study, with an Adjusted Accuracy of 0.85 and an AUC of 0.81. Format: Unsigned 8-bit Integer (UInt8) GeoTIFF 2. fsm_xgboost_pakistan.tif Model: eXtreme Gradient Boosting (XGBoost) Performance: Adjusted Accuracy of 0.82 and an AUC of 0.81. Format: Unsigned 8-bit Integer (UInt8) GeoTIFF Data Values & Interpretation The pixel values in both raster files have been classified into 5 discrete susceptibility categories based on quantile distribution: 1: Very Low 2: Low 3: Moderate 4: High 5: Very High Methodology Note:The models were trained on a balanced inventory of 20,000 historical flood/non-flood points derived from major flood events (2010, 2014, 2022). Ten flood conditioning features were utilized, including FABDEM elevation, slope, aspect, curvature, TWI, distance to river/drainage, NDVI, rainfall frequency, and distance to roads. Citation If you utilize this data, please cite the original research article: Waleed, M., & Sajjad, M. (2025). High-resolution flood susceptibility mapping and exposure assessment in Pakistan: An integrated artificial intelligence, machine learning and geospatial framework. International Journal of Disaster Risk Reduction, 121, 105442. https://doi.org/10.1016/j.ijdrr.2025.105442 Contact Portfolio: https://waleedgeo.com/ Email: [e-mail]