Rainfall-Induced Shallow Landslides: Hydrological Response Analysis and Stability Prediction Model Based on Field Monitoring and Model Experiments

Shallow landslides triggered by heavy rainfall often occur in clusters in mountainous regions and pose serious hazards. Understanding how groundwater and soil moisture respond to rainfall is therefore crucial. This study draws on in-situ monitoring and capillary rise experiments to examine these processes. Using three years of continuous field observations, we analyzed the relationships among rainfall infiltration, soil volumetric moisture content, and groundwater level. In addition, 14 model tests were conducted to assess the influence of soil density and fine/coarse particle composition on capillary rise, quantify the correlation between rainfall intensity and groundwater level, and develop a predictive model for capillary rise height based on fine particle content in gravelly soils. Building on these results, a landslide stability prediction model that incorporates rainfall and groundwater dynamics were formulated. The findings indicate that: (1) Shallow groundwater on slopes shows periodic fluctuations, with each hydrological year comprising three phases: slow decline, rapid decline, and rapid rise. Groundwater depth is negatively and linearly correlated with rainfall, while the link between water-level rise after rainfall events and rainfall is weak; (2) The final stable height of capillary rise in residual gravelly soil follows a power function of fine-grained content. Higher fine-grained content produces greater stable heights, but all samples remained below 1.0 m, suggesting that groundwater has limited influence on upper soil moisture in these soils; (3) According to the stability prediction model, the critical rainfall threshold for slope failure is 81.8 mm/day, and the groundwater depth threshold is 0.73 m. The results provide a basis for early warning and risk mitigation of rainfall-induced shallow landslides in mountainous terrain.