Probability Analysis of Shallow Landslides in Varying Vegetation Zones with Random Soil Grain-Size Distribution
Hu Jiang,Qiang Zou,Yong Li,Yao Jiang,Junfang Cui,Bin Zhou,Wentao Zhou,Siyu Chen,Zihao Zeng
DOI: https://doi.org/10.1016/j.envsoft.2024.106267
IF: 5.471
2024-11-24
Environmental Modelling & Software
Abstract:The physically-based landslide susceptibility models are widely used to guide disaster prevention and mitigation in mountainous areas due to their significant predictive capability. However, this method faces limitations in regions with complex topography and vegetation types, primarily due to a lack of consideration for the spatial uncertainty of planted soil caused by variations in soil particle size composition. Therefore, a new model is established to predict shallow landslide occurrence considering the impact of the uncertainty of soil particle size composition on soil shear strength parameters. This model optimizes the assignment strategy for soil physical strength parameters with the support of the random soil grain-size field theory. Subsequently, it organically integrates the impact of plants on slope stability, involving root reinforcing, moisture regulation (preferential flow and root water uptake), and the canopy's interception and weight loading effects, based on the infinite slope model. The model is validated in a region with significant vegetation zonality in Sichuan Province, China. The results show: (i) the testing indicator AUC values range from 0.862 to 0.873, indicating that the model can effectively predict the spatial occurrence probability of shallow landslides, (ii) the proposed LSM-VEG-GSD model exceeds by 17.50% the traditional pseudo-static model according to the AUC score, and (iii) regardless of water height ratio interval, the probability of slope failure in different vegetation zones increases with slope angle, following an S-shaped curve regression pattern. Overall, the findings of this study contribute to predicting the stability of shallow landslides in terrain transition zones with high potential landslide concealment and uncertainty under the influence of vegetation.
environmental sciences,engineering, environmental,water resources,computer science, interdisciplinary applications