Geospatial Assessment and Mapping Landslide Susceptibility for the Garo Hills Division, Meghalaya, India

ABSTRACT Creating accurate and effective Landslide Susceptibility (LS) maps can aid disaster prevention and mitigation efforts and provide sufficient public safety. The primary aim of this study is to develop an LS map for the Garo Hills region in Meghalaya, India, using the weight of evidence (WoE), frequency ratio (FR), and Shannon entropy (SE) methods. A comprehensive landslide inventory catalogued 98 events from 2000 to 2023 for the analysis, and nine key geographical and environmental parameters were prepared. Conducted multicollinearity and correlation analysis to identify and mitigate collinearity issues between factors. The model's performance was analysed through the area under the curve (AUC) value of receiver operating characteristic (ROC) curves and three recent landslides. The results showed that FR method achieved the highest accuracy, with successive rate curve (SRC) AUC and predictive rate curve (PRC) AUC values of 0.860 and 0.940, respectively, and classified susceptibility at three sites as high, moderate, and low. The WoE method effectively identified three landslides site in high and very high susceptibility zones, achieving SRC AUC and PRC AUC values of 0.844 and 0.915, respectively. The SE method showed robust performance in predicting landslide‐prone areas, with PRC AUC comparable to other methods (0.913), though its SRC AUC (0.771) was lower. Developed maps revealed that high and very high susceptibility zones account for approximately 10% and 3% of the study area, predominantly near roads, steep slopes, and higher elevations. The information in this study is valuable for civilians and the government authorities involved in hazard monitoring and management.