Landslide Susceptibility Mapping of Kinnaur District in Himachal Pradesh, India Using Probabilistic Frequency Ratio Model

Landslides are a recurring phenomenon in hilly areas that destroy the natural environment, causing enormous property damage and human loss every year. Geological rock structure plays an important role in the occurrence of landslides. Concentration of dolomite, phyllite, schist and quartzite make this area more susceptible to failure. Landslides are prompted with the aid of using aggregate of things like slope, soil type, lithology, land use, land cover, drainage density, relative relief, curvature, etc. The main aim of this study is to delineate landslide susceptibility zone. In the study, the primary concern was to evaluate the contribution of controlling factors to landslides and to produce a landslide susceptibility map of the examination area. This study uses the landslide susceptibility model using frequency ratio in Kinnaur district, Himachal Pradesh, India. A landslide inventory was prepared with the help of Landsat, google earth images from 2010–2020 period. Digital elevation model (DEM) was used to prepare slope, aspect, curvature, and relative relief maps. Other maps like soil, land use land cover, was prepared using LISS-iii and LISSiv data to analyze landslide controlling factors. An aggregate of 379 landslide polygons was digitized and then divided into a preparation set (70%) with 265 polygons and a test set with 114 polygons (30%). The connection between landslides and controlling factors was statistically assessed with FR analysis. Values were utilized to create the landslide susceptibility index (LSI) and the investigation region was divided into five zones of relative landslide susceptibility which were divided as very high, high, moderate, low, and very low. The findings of the analysis were validated by calculating the area under curve (AUC), which reveals a success rate accuracy of 73.01% and a prediction rate curve accuracy of 0.73%, indicating a high-quality susceptibility map derived from the FR model.