GIS based flood susceptibility mapping in the Keleghai river basin, India: a comparative assessment of bivariate statistical models

Flooding in the Keleghai River Basin, India, threatens communities with severe economic and environmental consequences. This study develops a scalable model to map flood-prone areas, providing crucial insights for disaster management using bivariate statistical methods (Frequency Ratio, Shannon Entropy, and Information Value), with analyzed eight independent variables: slope, elevation, drainage density, rainfall, topographic wetness index, normalized difference vegetation index, land use, and stream power index. In this regard, to prepared flood susceptibility maps 418 the flood point’s inventory data was extracted from ‘BHUVAN’ geospatial platform from which 293 were used for model building and 125 were used for model validation. The resulting flood susceptibility map categorizes the basin into zones of varying flood susceptibility levels, ranging from very low to very high. Approximately 38%, 52.49% and 49.82% of the basin falls within the high and very high susceptibility categories according to the Frequency Ratio, Shannon Entropy and Information Value methods respectively and the accuracy of the models validated by the ROC curve. The success rates are 0.828, 0.844 and 0.829, whereas prediction rate are 0.816, 0.815 and 0.811 for the models of Frequency Ratio, Shannon Entropy and Information Value respectively. The generated flood susceptibility map will serve as a valuable tool for local authorities and decision-makers, enabling informed land-use planning, emergency response preparedness, and the implementation of appropriate flood mitigation strategies. Furthermore, the methodology employed in this study can be adapted and replicated in other river basins, contributing to enhanced flood risk management efforts at regional and national scales.