Comparative evaluation of information value and frequency ratio in landslide susceptibility analysis along national highways of Sikkim Himalaya

Landslide is the most common natural hazard in Himalaya that mainly triggered by the earthquake and rainfall. Landslide Susceptibility analysis is the spatial prediction of landslide occurrence based on local terrain condition. In Sikkim, every year occurrences of intense rainfall and earthquake cause landslide and related casualties along different parts of national highways. The present study is on GIS-based landslide probability analysis with 'Information Value' and 'Frequency Ratio Method.' These methods are used to derive the weighted value of causative factors and classes for Landslide Susceptibility Zonation (LSZ) along major roads. Eleven causative factors viz. slope, elevation, aspect, relative relief, land use, NDVI, soil, lithology, distance to drainage, distance to lineament, and rainfall used for landslide susceptibility analysis. 'Information value' and 'frequency ratio' methods have predicted the probability of landslide susceptibility for the study area. Receiver Operating Characteristic (ROC) curves for landslide susceptibility are drawn for both the methods and derived area under curve value are found 0.88 and 0.84 respectively. These results suggest that the 'Information Value' method has better performance than 'Frequency Ratio' Method. This statistical method ROC curve result validated by field survey and ground truth data used for accuracy assessment of LSZ map.