Landslide Susceptibility Mapping Using Statistical Methods in East Jaintia Hills, Meghalaya, India

Landslides represent a significant natural hazard, particularly in vulnerable regions like the East Jaintia Hills in Meghalaya, which pose risks to livelihoods and civilization. This study is about developing a map illustrating the landslide susceptibility (LS) of the East Jaintia Hills. The first step was the development of a landslide inventory (LI) map, including a total of 121 landslide locations. Subsequently, the LI map was randomly partitioned into 70% and 30% for training and testing, respectively. In the second phase, nine landslide conditioning (LC) factors: slope, rainfall, normalized difference in vegetation index (NDVI), land use land cover (LULC), distance to drainage (Dtd), distance to road (Dtr), aspect, geology, and geomorphology, were used. Statistical-based methodologies such as the modified information value (MIV), statistical index (SI), and an ensemble of these approaches in eight distinct scenarios are used to develop the LS map. The prediction model’s accuracy was assessed by calculating the area under the curve (AUC) value of the receiver operating characteristic (ROC) curve. Results indicate that scenario 2 (individual SI model) performed better, with AUC values of 0.874 and 0.926 for successive and predictive rates, respectively. Scenario 2 results show that 21.5% of the region is classed as very low susceptible, 37% as low susceptible, 23% as moderately susceptible, and 15% and 3.5% as highly and very highly susceptible, respectively. Researchers can apply this methodology in similar hilly regions, which should signify the novelty of the study and the primary outcome beneficial for the society of the East Jaintia Hills region.