Holistic Landslide Risk Assessment Using Gis-Based Susceptibility Modeling and Socio-Economic Impact Ranking: A Case Study of Roshi Rural Municipality, Nepal

Abstract

Landslides pose a significant hazard in the Himalayan region of Nepal due to steep terrain, fragile geology, intense monsoon rainfall, and expanding rural infrastructure. This study evaluates landslide susceptibility in Roshi Rural Municipality, Kavrepalanchowk District, using two GIS-based approaches: the Frequency Ratio (FR) model and the Weighted Overlay Method (WOM) based on the Analytical Hierarchy Process (AHP). Key conditioning factors, including slope, elevation, geology, land use/land cover, rainfall, and proximity to roads and rivers, were analyzed. A landslide inventory from the 2024 extreme rainfall event was used for model training and validation through the Receiver Operating Characteristic (ROC) curve and Area under the Curve (AUC). Results show that the Frequency Ratio (FR) model achieved an AUC value of 0.671, indicating moderate predictive capability. In comparison, the Weighted Overlay Method (WOM) produced an AUC value of 0.572, indicating relatively lower prediction performance. These results suggest that the FR model provides a more reliable representation of landslide susceptibility in the study area.