Impact study for landslide contributing factors using a multi-criterion approach for landslide susceptibility

Monitoring landslide-prone areas is necessary in Manipur, India, as this region is a landlocked state, and the highways, which served as lifelines, are often blocked during the rainy seasons every year causing a huge loss in the economy, many casualties, and even life threats. Analytic hierarchy process approach was employed in this study to develop nine different scenarios of landslide susceptibility, and the impacts of each causative factor, i.e., topographic slope (%), elevation, curvature, normalized difference vegetation index, aspect, land-use land-cover, rainfall data, and soil map, were analyzed by comparing the success rate and prediction rate of each scenario. The scenario consisting of all the eight causative factors gave a success accuracy of 81.61% and a prediction accuracy of 79.88%. The factor, land-use land-cover has the highest impact as the success, and the prediction rate was increased by 7.23% and 6.42% when this factor was considered while the topographic factor, elevation has the least impact on the landslide susceptibility as the success and prediction accuracy were increased by 5.67% and 6.68% when this factor was excluded. This study reveals that considering more factors does not result in higher accuracy of the model. © 2021, Saudi Society for Geosciences.