Insights Into Debris Flow Susceptibility From Spatial–Temporal Evolution of Vegetation, Precipitation and Temperature (a Case of Pingwu County, China)

Following the Wenchuan earthquake, the frequency of debris flow outbreaks in China's southwestern mountains increased significantly. Precipitation, vegetation and temperature are evaluation factors for debris flow susceptibility and exhibit consistent temporal changes. To eliminate the influence of time effects in the debris flow susceptibility evaluation (DFSE) process and improve the predictive ability of the evaluation results, this study selected Pingwu County, China, as its research subject. Using the variation trend slope model and the Theil‐Sen median model, we obtained the spatial–temporal evolution of the evaluation factors and the debris flow susceptibility based on ensemble models of minimum–maximum normalization (MMN), logistic regression (LR), and artificial neural networks (ANN). From 2015 to 2022, annual precipitation and average temperature increased, and the area with an increased annual normalized difference vegetation index (NDVI) increased by about 96%. The coverage area of high and extremely high susceptibility for different ensemble models was ≥ 25%. During the verification process, the MMN‐LR model of the slope factors had the highest area under the curve (AUC) (0.847). Therefore, the MMN‐LR of slope factors is valuable for guiding debris flow prevention in Pingwu County. Additionally, susceptibility evaluations based on the spatial–temporal evolution of evaluation factors can effectively inform debris flow warnings and provide new research ideas for post‐earthquake DFSE worldwide.