Causes, Failure Mechanisms and Susceptibility Zonation of Landslides along Gedo-Dilb Road Corridor, Northern Ethiopian

Abstract

Abstract The hilly and mountainous areas of the highlands and rift escarpments of Ethiopia are frequently affected by first time as well as reactivated landslides of different types and sizes. One of the areas affected by such hazards, a focus of this study, is the Gedo-Dilb Asphalt road and its corridor in northern Ethiopia. The objective of this research was to evaluate the characteristics and influencing factors of landslides, determine the stability conditions of critical slope sections, produce susceptibility map of the road corridor for development planning and recommend mitigation measures. The research involved: (a) detailed inventory and characterization of landslides and associated ground failures, (b) field and laboratory investigation of the engineering properties of soils and rocks, (c) slope stability analysis of selected slope sections, (d) evaluations of the various possible causative factors, defining the most influential parameters and producing landslide susceptibility map of the road corridor, and (e) recommend appropriate mitigation measures. Results of this study show that the area is affected by 103 active landslides covering an area of 8.44 km2. In steep slopes, rock-falls are common. The debris/earth slides facilitated by gully erosion have depths which range from 0.8m to 12m. Out of the a total 16Km asphalt road length, 9.5Km road section was found to be affected by various interrelated hazards including landslides, siltation of road hydraulic structures, and erosion/scouring. Based on the comprehensive field and laboratory investigations as well as slope stability analysis the causes of landslides in the study area include: (a) preconditions: lithology, slope angle, slope shape; (b) preparatory: land use/land cover, gully erosion, road excavations; and (c) triggering: rainfall. Analytical hierarchy process (AHP) approach was used to develop landslide susceptibility map and the model was validated using area under curve (AUC). The AUC value was found to be 0.82, indicating very good prediction of landslide susceptibility: 75% of the inventoried landslides fall in the high and very high susceptibility zones. This study has revealed that landslides in the area are associated with several interrelated factors which require integrated solutions including gully rehabilitation/stabilization, drainage based water management (surface and sub-surface water), and bio-engineering solutions for slope stabilization.