LANDSLIDE SUSCEPTIBILITY MODELING USING REMOTE SENSING AND GIS TECHNIQUES IN ZENTI CATCHMENT, SOUTHERN ETHIOPIA

Abstract

The Zenti Catchment, situated in the Gofa zone of South Ethiopia, is experiencing notable landslide occurrences and slope instability, leading to significant destruction of agricultural land, crops, and housing. This study explores the primary factors that contribute to landslides, assesses their geographical distribution, and determines the most significant causes through remote sensing and Geographic Information System (GIS) techniques in combination with the Analytical Hierarchy Process (AHP) and Frequency Ratio (FR) models. Eleven major elements affecting landslides were analyzed: elevation, slope, aspect, curvature, drainage density, lineament density, proximity to rivers and roads, land use and land cover (LULC), rainfall, and lithology. A comprehensive dataset of 840 past landslide locations was recorded by conducting field surveys and utilizing Google Earth imagery, with 70% of this data allocated for model training and 30% set aside for validation. Causative factor maps were created using ArcGIS 10.8, while frequency ratio and AHP weight values were derived using Microsoft Excel 2010 and ArcGIS. From these analyses, landslide susceptibility indexes (LSI) were formulated for both FR and AHP models, categorizing AHP scores as very low (LSI = 0.61), low (LSI = 01.07), moderate (LSI = 1.98), high (LSI = 2.11), and very high (LSI = 2.93). The FR model yielded scores of (LSI = 0.49) indicating ‘very low’, (LSI = 1.32) for ‘low’, (LSI = 1.84) for ‘moderate’, (LSI = 2.26) for ‘high’, and (LSI = 3.11) for ‘very high’ susceptibility. The FR model exhibited greater accuracy (92.7%) in comparison to AHP (86.5%), mainly due to its direct connection with historical and current landslide events. These results offer critical insights for risk assessment, disaster preparedness, and effective land-use planning, providing local authorities and environmental management agencies with vital resources to manage prospective landslide risks