Abstract
The unsaturated permeability coefficient of granite residual soil (GRS) increases rapidly with rising moisture content, as the loss of matric suction enhances the continuity of the water phase within the soil pores. This can lead to slope instability and embankment collapse during rainfall. This study investigated the effects of wet-dry cycles on the hydraulic and microstructural evolution of GRS, introducing key innovations over prior research. First, microstructure changes were investigated using mercury intrusion porosimetry (MIP) tests, investigates the evolution of bimodal pore structure under cyclic wetting and drying. Second, the entire range of matric suction was comprehensively measured by integrating the pressure plate method (PPM), filter paper method (FPM), and vapor equilibrium method (VEM), capturing both low and high suction regimes comprehensively. Third, the Li model was applied to fit the bimodal SWCC across different wet-dry cycles, the unsaturated permeability coefficient was calculated using the Zhai model. The results indicate that the microstructure of GRS under different wet-dry cycles presents a clear bimodal pore size distribution (PSD), intra-aggregate pores peaked near 450 nm, while inter-aggregate pores ranged between 20,000-60,000 nm.. After six wet-dry cycles, the volume of the dominant intra-aggregate pores decreased by approximately 25%, while the larger inter-aggregate pores saw a reduction of about 15%, indicating a coarsening of the pore network. Meanwhile, there is a clear decrease in inter-aggregate pore distribution density. The combination of measurement methods can cover the entire matric suction range. The Li model is applied to fit the SWCC under different wet-dry cycles, and the correlation coefficient (R2) are all higher than 0.95. The unsaturated permeability coefficient of GRS exhibits a nonlinearly variation with saturation, in-creasing with the increase in saturation or the increase in wet-dry cycles. The unsaturated permeability coefficient of bimodal GRS was calculated based on the Zhai model and the lgk(s) and saturation can be expressed by a logarithmic function, with the correlation coefficient (R2) higher than 0.99 under different wet-dry cycles. The study contributes useful insights into the evolution of pore structure and hydraulic behavior of GRS under cyclic wetting and drying, which is important for slope stability and hydrological modeling in subtropical regions.