GIS-based landslide susceptibility assessment in Nepal Himalayan: a comparison of heuristic and bivariate methods

Landsliding is one of the most hazardous natural disasters in the Himalaya region causing hundreds of deaths in Nepal every year. Government and research institutions have attempted for years to assess the susceptibility of landslides, and estimate their risk and spatial distribution. The purpose of this study is to evaluate and compare heuristic (index based) and bivariate (statistical index) landslide susceptibility assessment procedures in the Kulekhani watershed in Central Nepal. Two methods are applied and two separate landslide susceptibility maps are produced. The resulting landslide susceptibility maps delineate the study area into four susceptibility classes: very high, high, moderate, and low. Each susceptibility map is validated by correlating the landslide frequencies of different classes with existing field observations. When the agreed area between landslide susceptibility maps of the index-based method and the statistical index method is determined as the total area having the same landslide susceptibility in both landslide susceptibility maps, it is found that the landslide susceptibility zonation of the two maps agrees for 68.4% of the study area. Moreover, the percentage of observed landslide falling into the agreed very high susceptibility zone is 47.67% and 17.05% for the agreed high susceptibility zone, which proves the predictive capacity of the methodology.