A physics-informed data-driven model for landslide susceptibility assessment in the Three Gorges Reservoir area
Citation
Liu, S., Wang, L., Zhang, W., Sun, W., Fu, J., Xiao, T., Dai, Z. (2023). A physics-informed data-driven model for landslide susceptibility assessment in the Three Gorges Reservoir area. Geoscience Frontiers, 14(5): 101621. Link to paper
Abstract
Landslide susceptibility mapping is a crucial tool for analyzing geohazards in a region. Recent publications have popularized data-driven models, particularly machine learning-based methods, owing to their strong capability in dealing with complex nonlinear problems. However, an outstanding challenge for these approaches is the lack of awareness of the physical mechanism of landslide processes and poor interpretability. To address this, this study employed Scoops 3D as a physics-informed tool to qualitatively assess slope stability in the Hubei Province section of the Three Gorges Reservoir Area (TGRA). The random forest algorithm was employed for data-driven landslide susceptibility analysis, with the area under the receiver operating characteristic curve (AUC) serving as the model evaluation index. By considering the insights gained from the three-dimensional stability model, the physics-informed data-driven method improves the extraction of negative samples and the interpretability. The methodology demonstrates that negative samples should be distributed on slopes with factor of safety greater than 1.5. Compared to pure data-driven landslide susceptibility mapping, this approach improves the AUC performance by 29.49%. Additionally, the regional landslide susceptibility map generated by the physics-informed model can better delineate the actual distribution of the landslides in the study area.