Application of Geospatial and Frequency Ratio Techniques in Landslide Susceptibility Mapping: Case Study of Daramalo District, Ethiopia

Abstract

Daramalo district, located in the Gamo Zone of South Ethiopia, is one of the areas most affected by landslides. This study aims to assess the landslide susceptibility of the area and to develop a comprehensive landslide susceptibility map. To achieve this, a bivariate statistical frequency ratio model was employed. A detailed inventory of landslides was compiled through fieldwork and the interpretation of Google Earth imagery, identifying a total of 32 landslides. These were categorized into training landslides (70%) for model development and validation landslides (30%) for model evaluation. Eight causative factors slope, aspect, elevation, curvature profile, drainage density, lithology, lineament density, and land use/land cover (LULC) were integrated with the training landslide data to determine the frequency ratio values for each class of these factors. Relative frequency values were assigned to the appropriate factor classes, which were then summed using a raster calculator algorithm to produce the landslide susceptibility map. The final susceptibility map indicates that 44% (110 km²) of the study area is classified as low susceptibility, 36.8% (92 km²) as moderate susceptibility, and 19.2% (48 km²) as high susceptibility. This suggests that approximately 20% of the area is at significant hazard of landslides, while about 80% has relatively low to moderate susceptibility to this natural hazard. The performance of the frequency ratio model was validated using the receiver operating characteristic (ROC) curve, achieving a notable success prediction rate of 89.03%. Overall, the model demonstrated strong accuracy. The resulting map is anticipated to be a valuable resource for land use planning, site selection, and the formulation of effective landslide prevention and mitigation strategies.