Analysis of bi-variate statistical and multi-criteria decision-making models in landslide susceptibility mapping in lower Mandakini Valley, India

Abstract

Landslide is recurrent phenomena in the Mandakini valley of Uttarakhand, India. This study concentrates on the analysis of landslide susceptibility mapping using Frequency Ratio (FR) and Analytical Hierarchical Process (AHP) models in the lower Mandakini valley. The models are applied by integrating eleven causative factors and 160 past landslides. Both models are validated and compared using Receiver Operating Characteristics and Seed Cell Area Index method. The validation result shows that the FR model gives better success rate and prediction rate than AHP model. Seed cell index values of high and very high susceptibility classes are more in the case of the FR model than AHP model. Thus, the landslide prediction capability of the FR model is more reliable in the study area. The study will contribute to understand future landslide risk and help in disaster reduction planning in the region.