Probabilistic Stability Analysis of Slopes under Rainfall Accounting for Spatial Variability of Hydromechanical Properties

Understanding the interaction between soil hydraulic conductivity and shear strength parameters, as well as their influence on rainfall infiltration analysis in spatially variable soils, is particularly important for probabilistic slope stability analysis. Current probabilistic slope stability analyses rarely account for the combined spatial variabilities of hydromechanical properties, and many studies rely on computationally intensive methods, which limit their practical application. Therefore, this study develops an improved Green-Ampt model to elucidate the failure mechanisms of soil slopes, efficiently incorporating the spatial variabilities of both saturated hydraulic conductivity and shear strength parameters. The probabilistic slope stability analyses across various rainfall durations reveal that slopes characterized by low variability in hydraulic conductivity exhibit a reduced probability of failure (pf) when the factor of safety (FS) exceeds 1.0. However, when the rainfall infiltration reduces the FS below 1.0, these slopes show a significantly higher pf. The findings indicate that, during the initial stages of rainfall infiltration, the spatial variability in the shear strength predominantly influences the slope stability. Failures may occur not only along the wetting front but also along an impermeable bed or within the weaker layers due to soil spatial variability. As rainfall infiltration advances, the wetting front becomes the dominant factor in slope stability, and the slope failure predominantly occurs along the critical slip surface near this front. Notably, at this stage, the pf of slopes appears to be relatively insensitive to the spatial variability of shear strength parameters, highlighting a transition in the controlling failure mechanisms.