Landslide Susceptibility Mapping Using Frequency Ratio And Dempster-Shafer Theory For Palakkad District, Kerala, India

Abstract

Landslides are a significant natural hazard in the Palakkad district of Kerala, India, causing widespread damage to life and property. This study aims to develop a landslide susceptibility map (LSM) for the region using the Frequency Ratio (FR) and Dempster-Shafer Theory of Evidence (DST) models. A comprehensive landslide inventory map was prepared, consisting of 593 landslide locations, which were divided into two subsets: 70% for model training and 30% for validation.The analysis incorporated several landslide causative factors including slope, aspect, drainage density, rainfall, geology, geomorphology, soil, land use/land cover, curvature, and road network density. These factors were integrated into Geographic Information System (GIS) platforms to produce LSMs using both the FR and DST models.The model validation was conducted using receiver operating characteristic (ROC) curves. The results demonstrated that the FR model achieved an accuracy of 85%, while the DST model exhibited a higher prediction accuracy of 95%. This highlights the superior performance of the DST model in predicting landslide prone areas.The finding provides valuable insights for disaster management and planning in the Palakkad district. The DST-based LSM, with its high predictive capability, can serve as a reliable tool for mitigating landslide risks and guiding sustainable land-use planning in the region.