Landslide susceptibility mapping based on the deformation intensity

Abstract Affected by the human engineering activities and extreme climate change, landslide disasters develop frequently in the channel of the Three Gorges Reservoir Area. The framework related to the extension of dynamic susceptibility modeling has largely not been explored. This work considered the Wanzhou channel of the Three Gorges Reservoir Area as the experimental site, which a transportation channel with significant economic value to carry out innovative research in two stages: (i) five machine learning models logistic regression (LR), multilayer perceptron neural network (MLPNN), support vector machine (SVM), random forest (RF) and decision tree (DT) were used to explore landslide susceptibility distribution based on detailed landslide boundaries; (ii) The PS-InSAR-based intensify factor was generated by the Persistent Scatterer Interferometric Synthetic Aperture Radar (PS-InSAR) technology. Then the intensify factor was combined with the proposed static factors and machine learning models to generate enhanced landslide susceptibility mapping (ELSM). The area under the receiver operating characteristic curve (AUC) was proposed as the evaluation indicator. Dynamic landslide susceptibility mapping has improved model accuracy, especially with DT models achieving 2% enhancement and the highest AUC value of 93.1%. The susceptibility results of introducing intensify factor are more in line with the spatial distribution of actual landslides. The research framework proposed in this study has important reference significance for the dynamic management and prevention of landslide disasters in the study area.