A comparative study on the integrative ability of the analytical hierarchy process, weights of evidence and logistic regression methods with the Flow-R model for landslide susceptibility assessment

Abstract

The main purpose of the present study is to compare the results of the integrated models, the weights of evidence (WoE), analytical hierarchical process (AHP) and logistic regression (LR) models combine with a Flow-R model for landslide susceptibility assessment in Ha Giang city, Vietnam and the surrounding areas. First, three landslide susceptibility index (LSI) maps were calculated using thirteen landslide conditioning factor maps. Then, a runout map was generated by Flow-R model using the digital elevation model and the slope map of the study area. Secondly, the success rate curve, prediction rate curve and area under the curves (AUC) were calculated to assess prediction capability. The final landslide susceptibility maps were produced by the integration of the LSI maps with the runout map that was extracted from FlowR model. The validation results showed that the AUC of LSI maps in the integrated models, AHP-FlowR, WoE-FlowR and LR-FlowR, were 82.89%, 88.15% and 86.53% with prediction accuracies of 80.41%, 84.94% and 85.71%, respectively. The combined models of WoE-FlowR (88.15%) and LR-FlowR (86.53%) have shown the highest prediction capability. In general, the integrated models provide reasonable results that may be aided in prevention planning and land use planning purposes.