Remote sensing and GIS-based landslide susceptibility mapping using frequency ratio method in Sikkim Himalaya

Abstract

Sikkim Himalaya is located in the North-Eastern Himalaya and is prone to landslides caused by rainfall, anthropogenic factors, earthquakes, and other natural disasters. This study aims to identify landslide susceptibility map for the Sikkim Himalaya, India, utilizing an integrated methodology of Geographic Information System (GIS), Remote Sensing (RS), and the Frequency Ratio (FR) method. Identification of landslide susceptibility map in this region has used multiple datasets as rainfall, earthquake's magnitude, slope angle, altitude, distance to drainages, topographic roughness index, geomorphology, geology, soil, gravity anomaly, distance to faults, stream transportation index, topographic witness index, stream power index, distance to roads, LULC, and NDVI. All the above-mentioned factor/thematic layers were generated using remotely sensed as well as ground data with the help of Arc GIS software. Therefore, the weights of these all-thematic layers were calculated using the FR model for the occurrence of landslides prone in the area. After that, calculated weights of all factor/thematic layers were integrated with Arc GIS software to generate a landslide susceptibility map. The landslide susceptibility zone of the study area has been divided into 5 different classes, namely ‘Very High (11.88%)’, ‘High (15.75%)’, ‘Medium (25.88%)’, ‘Low (25.30%)’, and ‘Very Low (21.19%)’. The accuracy assessment of the study area was 87.8%, which was done by the Receiver Operating Characteristic (ROC) curve method. This research investigates the possibility of using broader approaches to determine landslide-prone areas.