A combined analysis of frequency ratio and analytical hierarchy process for landslide susceptibility assessment in Tenta, South Wollo, Ethiopia

Abstract

Landslide is most common geo environmental disaster in the study area Tenta, which is parts of Northern Ethiopian highland. It has caused a great deal of human sorrow in this area, affecting farms, villages, graves, and infrastructures. The main objective of this study was to evaluate the potential susceptibility to landslides using integrated approach involving FR and AHP models. To achieve this, field survey and satellite image was used to identify 328 landslides, which were separated into training landslides (75%) and validation landslides (25%). Slope, aspect, elevation, distance to stream, distance to spring, slope material, distance to lineament, and land use and land cover (LULC) are the parameters that were taken into account as contributing factors to landslide. Ultimately, utilizing GIS based Frequency Ratio (FR) and Analytical Hierarchy Process (AHP) models, landslide susceptibility zones were generated, and validated using overly, landslide density index (R-index), and the receiver operating characteristics (ROC) curve. Landslide susceptibility maps examined using overly and R-index infers training and validation landslides are depleted in very low and low susceptibility classes and concentrated in high and very high susceptibility classes. Based on ROC curve, FR model map scores success rate of 0.828 which is little higher than AHP model map success rate 0.826. Similarly, Prediction rate of FR model map is 0.835 which is again little higher than 0.832 AHP model map prediction rate. Overall, both models performed well but, the FR model gives little more accurate result than the AHP model. Decision-makers can use these landslide susceptibility maps produced by both model for regional land use planning, site selection, and landslide prevention and mitigation efforts.