The effect of depth dependency on stochastic infinite slope stability analysis

Stochastic stability analysis is carried out for the unsaturated infinite slope during rainwater infiltration in this paper. The spatial variability of soil is considered by taking saturated hydraulic conductivity (k _s) as the only random variable. Since field results indicate that k _s often reduces with depth, the linear depth dependency of the mean trend of k _s is highlighted on the stochastic infinite slope stability analysis during infiltration. The spatial variability of k _s considering the depth dependency is generated using the random field theory. The finite-difference method is adopted to simulate the transient infiltration, while the analytical solution is used during steady-state infiltration. It is found that the depth dependency of the mean trend of k _s has a significant effect on the slope stability. For the case that the mean of k _s reduces linearly with depth during steady-state infiltration, an increase in the ratio of depth dependency can significantly increase the probability of failure. When transient infiltration is considered, three stages can be categorised through the location of the wetting front. If the depth dependency is not considered, the probability of failure is overestimated in both the early and middle stages, and it is highly underestimated in the late stage of the infiltration.