Landslide Assessment and Mapping Using Frequency Ratio and Analytical Hierarchy Process at Maze Catchment, Gamo Zone, Southern Ethiopia

Abstract

Landslides are one of the natural threats that cause loss of life and destruction of property in Ethiopia. This study was conducted in the Maze watershed in the Gamo zone, southern Ethiopia. The objective of this study was to assess and map the landslide susceptibility of the area using frequency ratio and analytic hierarchy process (AHP) approaches. To achieve this objective, landslides inventory data and their causative factors were identified. About 793 landslide polygons were collected by using remote sensing and field observations. These landslide inventory datasets were classified into two datasets, landslide training dataset (70%) for mapping landslide susceptibility and validation landslide dataset (30%) for validating landslide susceptibility maps. Ten landslide causative factors were selected for this study including slope, aspect, curvature, lithology, land use/land cover, normalized vegetation index and distance to fault, distance to river, and distance to road. The landslide susceptibility map was produced by superimposing the weights of all landslides causative factors using the raster calculator in the spatial analysis tools in Arc GIS. The final landslide susceptibility map is reclassified into susceptibility classes of “very low,” “low,” “moderate,” “high,” and “very high” using both frequency ratio and analytical hierarchy process approaches. The results of the landslide susceptibility map created using the frequency ratio approach show that the very low, low, medium, high, and very high susceptibility classes account for 25%, 20%, 18%, 18%, and 19% of the area, respectively. The landslide susceptibility map created by the analytical hierarchical process approach shows that the very low, low, medium, high, and very high susceptibility classes cover 3%, 7%, 26%, 36%, and 28% of the area, respectively. Both susceptibility maps were validated using the area under the receiver operating characteristic curve. The AUC results for the frequency ratio and analytical hierarchy process approaches were 0.873 and 0.87 for success rates, respectively, and 0.81 and 0.80 for prediction rates, respectively. Therefore, the landslide susceptibility map using the validation dataset provided acceptable results. These maps will play an important role in land use planning efforts.