Triggering mechanism of rainfall and reservoir water level dynamic change‐induced step‐like displacement for reservoir bank landslide

Reservoir bank landslides in the Three Gorges Reservoir (TGR) area frequently show step‐like displacement characteristics under coupled reservoir water level (RWL) fluctuations and rainfall, posing significant challenges to hazard early‐warning systems due to their abruptness and complexity. This study identifies three key trigger conditions for step‐like displacement by analysing the displacement characteristics of landslides. Using the Hejiabao landslide as a case study, the transient release‐inhalation method (TRIM) was employed to assess the unsaturated soil hydraulic properties of both the slide body and slide zone soils. Additionally, physical modelling tests were conducted under rainfall and RWL rise and fall conditions to simulate the triggering conditions. The results from TRIM and physical modelling tests reveal the underlying mechanisms of step‐like displacement in reservoir bank landslides. Drying‐dominated scenarios: RWL drawdown leads to abrupt increases in seepage forces and significant shear strength reduction in the sliding zone. Wetting‐dominated scenarios: Intense rainfall triggers pore water pressure increases and softening in the sliding zone. Hybrid scenarios: Cumulative wetting‐drying cycles and dual hydraulic gradients activate multi‐stage step‐like displacements with maximal magnitudes. Furthermore, the asymmetric hysteresis effect of the soil‐water characteristic curve (SWCC) governs the spatial–temporal distribution of pore water pressure. High hysteresis in the slide body delays deformation, while low hysteresis in the sliding zone accelerates instability. This study suggests optimizing early warning models by incorporating hydraulic hysteresis parameters and dynamic permeability thresholds, with particular attention to the synergistic effects of RWL drop rate and rainfall intensity. These findings provide a theoretical basis for risk assessment and early warning improvement in reservoir bank landslides, highlighting the importance of hydraulic hysteresis and dynamic coupling modelling for enhanced prediction accuracy.