Identification of groundwater potential zones of Idukki district using remote sensing and GIS-based machine-learning approach

Abstract Kerala's Idukki district, which is situated on the Western Ghats of India, susceptible to flooding and landslides. As a result 2018 Kerala floods, this disaster-prone region experienced drought conditions. In order lessen effects future disasters, it also necessary identify evaluate district's groundwater potential (GWP). This work used three machine-learning (ML) algorithms – Random Forest (RF), Adaptive Boosting (AdaBoost), Gradient (GB) model produce GWP zonation maps for district. Fourteen conditioning factors include elevation, slope, curvature, Topographic Roughness Index, lineament density, soil, geology, geomorphology, Wetness Sediment Transport drainage rainfall, land-use/land-cover (LULC), Normalised Difference Vegetation Index that were adopted as input parameters in modelling. All showed prominence when they examined feature importance using recursive elimination (RFE) method. The RF outperformed other two ML models terms fit, with an area under curve (AUC) value 0.92, while GB AdaBoost displayed less AUC values 0.90 0.88, respectively. produced by each reclassified into five zones very high low was discovered evenly spread throughout region.