Assessing Landslide susceptibility of Bangad Kupinde municipality, Salyan, Karnali Province, Nepal

Abstract This Paper examines the landslide susceptibility mapping of Bangad Kupinde Municipality which lies in Salyan district, Karnali Province Nepal.The hills are regarded as one of the most vulnerable areas in the world so, this study was conducted in order to know about the susceptibility of this municipality with respect to landslide. Google Earth Engine (GEE) and ArcMap were used during the image processing and analysis. Besides, ArcMap toolbox Arc-Spatial Data Modeler was used for map validation. Sentinel 2 surface reflectance product i.e., MSI Level-2A product obtained from GEE database was used in the study. Similarly, a higher resolution ALOS PALSAR DEM (12.5m) was Alaska Satellite Facility`s distributed active archive center. Pre-processing was done before further processing and analysis of satellite images to enhance their quality and to bring them to a desired output. This study used pixel-based classification approach using Random Forest classifier for image classification. Landslide susceptibility map was obtained by integrating various landslide causative factors using Analytical Hierarchal Process (AHP) and weighted overlay technique. Weighted overlay was used to integrate the factor whereas AHP was used to assign weight to the factors. Altogether seven factors i.e., land cover, NDVI, altitude, slope, aspect, Topographic Wetness Index (TWI) and stream network were used in producing the landslide susceptibility map. A total of 41 locations of landslides were determined by Google Earth digitization and field surveys which was used to create a landslide inventory map of the study area. Finally, a series of landslide susceptibility index were prepared based on AUC. The obtained Landslide Susceptibility Index was then classified into five categories of susceptibility levels namely: very high, high, moderate, low and very low. The results show that a total of 11.20% of area are under very high susceptible zone, 21.65% area under high susceptibility,28.22% area under moderate susceptibility, 25.70% area under low susceptibility and 13.23% area under very low susceptible zones.