Landslide Susceptibility Mapping Using Information Value Model with GIS in Wegeda Area, Northwestern Ethiopia

Abstract The geological and physiographical conditions have made Guna Mountain one of the area’s most vulnerable to disaster since it is a part of the northwestern highlands of Ethiopia. Parts of the Wegeda area are characterized by weathered volcanic rocks, rugged morphology with deeply incised gorges, heavy rainfall, and active surface processes. Many landslide events are frequently occurring in the studied area, mostly triggered by heavy seasonal rainfall and human activity, causing deaths and severe infrastructure damage to the local people. In this study, Information Value (IV) Model was applied to evaluate the landslide causative factors and generate a landslide susceptibility map (LSM). To do this, intensive fieldwork and Google Earth image interpretations were done on active and passive landslide scarps and produced a reliable landslide inventory map. These landslide locations were randomly divided into 80% training and 20% validation datasets. Seven landslide governing factors were combined with a training dataset using GIS tools to generate the LSM of the study area. Then the area was divided into five landslide susceptibility classes: very low, low, moderate, high, and very high. Later, the resulting map has been validated by using the area under the curve (AUC) and landslide density index methods for both the training and validation landslide datasets. According to the AUC graph, the IV model had a success rate and a predictive rate value of 82.4% and 78.3% respectively. This indicated that the produced LSM showed reasonable performance. Finally, the LSM produced by the IV model can be used by decision-makers for land use planning and landslide mitigation purpose.