Comparison analytical hierarchy process (AHP) and frequency ratio (FR) method in assessment of landslide susceptibility. A case study in Van Yen district, Yen Bai province

Abstract

Landslides are a natural disasters that frequently occur in the northern mountainous region of Vietnam. This study aims to compare the efectiveness of the Analytical Hierarchy Process (AHP) and Frequency Ratio (FR) modeling in mapping susceptibility to landslides with the support of a Geographic Information System (GIS). The study area is Van Yen district in Yen Bai province, which experiences a high frequency of landslides annually. Ten factors were used as variables in the model, including the lithology map, slope, aspect, plan curvature, profile curvature, topographic wetness index, fault network, river network, road network, and land cover data. The study used a landslide statistical report that including 211 landslide points to create the frequency ratio model, while the pairwise comparison method based on expert opinion was used to establish the weights for the AHP method. The results produced a spatial distribution of landslide susceptibility with five levels: very low, low, moderate, high, and very high. The study used the Area Under the Curve (AUC) to evaluate the performance of both models. The results indicated that the model using the Frequency Ratio method outperformed the Analytical Hierarchy Process model by 4.7% in addition to the similarity between landslide susceptibility maps and past landslide locations.