Assessing Landslide Susceptibility Mapping in Shimla District, Himachal Pradesh, India: A Comparative Approach Using Fuzzy-AHP, and FR for Risk Prediction

Landslides are one of the most common and frequently occurring hazards that endanger the natural environment, human lives, and infrastructure in mountainous areas. The Himalayas, specifically the Shivalik and Greater Himalayan range, are a region prone to landslides, and recent incidents, such as in 2023 in the Summer Hill area of the Shimla district, have resulted in the loss of 14 lives and 51 lives from all over the Himachal Pradesh. A landslide susceptibility map is essential for planning, policy development, management and mitigation measures to reduce damages caused by landslides. Multiple approaches were used to develop a landslide susceptibility map with their respective effectiveness and drawbacks however in this study a fuzzy-AHP (F-AHP), and frequency ratio (FR) have been incorporated to develop the landslide susceptibility map for a part of Shimla district in Himachal Pradesh, India. 1011 landslide sites were randomly selected from a landslide inventory map and divided into testing and training groups in a 25:75 ratio. Slope, geomorphology, profile curvature, drainage density, aspect, lineament density, texture, soil, land use and land cover, lithology, and rainfall were used as model inputs. According to the findings, the occurrence of landslides is more frequent in the regions with high rainfall distribution, steep slopes, weaker bed rocks, sparse vegetation, and close association with roads. The result statistics portray that the south, and southwestern parts (around 21%) of the Shimla district are the most landslide-prone areas. The measurement of accuracy of produced (i.e. FAHP and FR) LSZ map using ROC-AUC method presents an excellent (0.915) and very good (0.807) accuracy of the produced LSZ map. This study might help the local and state government bodies with prediction and mitigation measures against landslides in the area.