Comparative analyses on susceptibility of cutting slope landslides in southern Jiangxi using different models

There are many landslide disasters in southern Jiangxi, with a wide area and a small scale, and are characterized by mass and suddenness. More than 90% of landslides are caused by artificial slope cutting. In order to study the applicability of the susceptibility evaluation model for cutting slope landslides caused by cutting slopes in southern Jiangxi, taking Yinkeng Town, Yudu County, Ganzhou City as an example, based on the results of field geological surveys, and using GeoDetectors, the slope, the slope structure, rock formation, fault, road, and vegetation, were selected to carry out landslide susceptibility assessment by using the information value model (I), artificial neural network model (ANN), decision tree model (DT) and Logic regression model respectively. The results show that the AUC values obtained from information value model, artificial neural network model, decision tree model and logistic regression model are 0.800, 0.708, 0.672 and 0.586, respectively. The susceptibility results obtained by the information value model are in good agreement with the actual distribution of landslides in the study area. The specific value of the proportion of landslides in high-prone areas and medium-prone areas exceeds 80%. The information model is more suitable for the landslide susceptibility assessment under cutting slope in southern Jiangxi than the other three models. The assessment results provide a reference for the selection of the assessment model for the geohazard susceptibility in this region.