Stability analysis of unsaturated soil slopes under random rainfall patterns

The significance of rainfall pattern in the assessment of rainfall-induced landslide is widely recognized. However, much work so far is limited to several simplified typical rainfall patterns. In this study, the random rainfall pattern (RRP) is introduced and generated using random cascade model based on the rainfall event characterized by average rainfall intensity and duration. The stability of unsaturated slope considering RRPs is studied from three perspectives: deterministic analysis by means of safety factor under different generated RRPs, probabilistic analysis through conditional failure probability considering the diversity of generated RRPs based on Monte Carlo method and risk assessment analysis by introducing annual failure probability (AFP) considering also the occurrence frequencies of rainfall events. Three typical rainfall patterns are introduced for comparison analysis. The results show that slope stability is sensitive to the RRP and is strongly depend on the temporal distribution of rainfall intensity in RRP. High likelihood of slope failure may occur considering the variety of RRPs even though the slope is in a stable state in terms of deterministic analysis. The AFP considering RRPs increases rapidly with increasing rainfall duration and is significantly different from those under typical rainfall patterns. The findings lead to the conclusion that RRPs should be considered in the estimation of unsaturated slope stability