Spatiotemporal analysis and threshold modeling of rainfall-induced geological disasters in Anhui Province

Rainfall-induced geological disasters are widespread in the Jianghuai region of China, endangering human lives and socioeconomic activities. Anhui Province, a hotspot for these disasters, warrants a thorough analysis of the temporal and spatial distribution of geological disasters and their correlation with rainfall for effective forecasting and warning. This study divides Anhui Province into the Dabie Mountains, southern Anhui Mountains, and other areas based on different background conditions, and establishes effective rainfall threshold warning models for each. We reconstructed the collection of geological disaster precipitation records and rainfall data in Anhui from 2008 to 2023. Using binary logistic regression, we analyzed the correlation between rainfall factors and geological disasters, selected the optimal effective rainfall attenuation parameters for the study area, and determined the critical effective rainfall for different warning levels. Results show: (1) Landslides and collapses are the main types, mostly occurring in high altitude areas like the Dabie and southern Anhui Mountains, and are concentrated in the rainy season of June - July each year; (2) Rainfall is the main inducer, with both single heavy rainfall processes and sustained rainfall influencing geological disaster occurrence, mostly through their combined effect; (3) Effective rainfall is significantly correlated with the day of and previous 8 days rainfall. The optimal attenuation coefficients in the Dabie Mountains, southern Anhui Mountains, and other regions are 0.60, 0.66, and 0.61, respectively. The study shows that setting fine tuned critical rainfall threshold models for different regions is better than a province wide threshold. With a 79% forecast accuracy, it can provide a scientific basis for geological disaster meteorological risk forecasting and warning in Anhui Province.