Susceptibility assessment of debris flow based on watershed units in Kangding City, Sichuan Province

To study the susceptibility of debris flow in Kangding City, the study area was divided into 421 watershed units. Spatial analysis tools in ArcGIS software and SPSS software were used to analyze the internal superposition of evaluation indicators and the correlation between evaluation indicators and debris flows disasters. By screening out the evaluation factors with a high degree of overlap and poor correlation, eight evaluation factors were selected for debris flow susceptibility assessment. These included watershed unit area, melton rate, form factor ratio, collapse and landslides density of catchment, average fractional vegetation cover of catchment, road density of catchment, average stream power index of catchment, and average rainfall during the multi-year flood season. The susceptibility of debris flow was quantitatively evaluated by combining the information value model and the entropy method. The weights of the evaluation indicators were quantitatively determined by the entropy method, and the evaluation factor weighted information quantity value was calculated. Based on this, the debris flow susceptibility in Kangding City was divided into four grades: extremely high, high, medium and low. The results of debris flow susceptibility assessment were tested using the frequency ratio model and the Receiver-Operating Characteristic (ROC) curve, with an AUC curve of 0.842, indicating high accuracy of the evaluation model.