Against the backdrop of global climate change, extreme rainfall-induced clustering landslides have caused severe losses and posed significant hazard risks. Addressing the limitation of spatiotemporal separation in traditional early warning models, this study takes the 2020 extreme rainfall event in the Bailong River Basin, China, as a case to construct a three-dimensional meteorological risk early warning model integrating “geological background susceptibility-critical rainfall threshold-short-term forecasted rainfall”. Based on the first post-hazard interpretation of inventory data for 4236 landslides, a genetic algorithm-optimized random forest model was employed for susceptibility assessment. The study revealed the key controlling roles of geomorphic factors such as slope aspect (20.95%) and topographic relief (10.38%), and effectively improved the accuracy of susceptibility assessment under extreme rainfall conditions (AUC = 0.879) by incorporating hydrogeological environmental factors including stream power index, distance to rivers, and soil types. Through spatial correlation analysis between 27 different time-lag rainfall scenarios and clustering landslide distributions, the characteristic rainfall parameters were determined as the 4-day antecedent effective rainfall (α = 0.9) and 6-h triggering rainfall. Nonlinear functional relationships between cumulative probability of landslide surface density and characteristic rainfall amount were established (R 2 = 0.925, 0.984), breaking through the limitations of traditional threshold models that rely on historical random data and subjective rainfall parameter selection. The results show that the same-day triggering rainfall is significantly more sensitive to landslide initiation than antecedent effective rainfall. Under the comprehensive effect of antecedent effective rainfall, a 6-h triggering rainfall of 12.4 mm triggers a medium-level warning. A three-dimensional early warning matrix was finally established, enabling medium-risk early warning for extremely high-susceptibility areas (accounting for 10.07% of the area) under low rainfall scenarios and identifying extremely high-risk areas covering 30.46% of the Bailong River Basin under extreme rainfall scenarios. The research results provide a new method for refined early warning of extreme rainfall-induced clustering landslides and hold important scientific significance for enhancing regional hazard prevention and mitigation capabilities.
Liu et al. (Sun,) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: