Assessment of probability of failure on rainfall-induced shallow landslides at slope scale using a physical-based model and fuzzy point estimate method

Shallow hillslope failure caused by rainfall is characterized by complex soil hydrology and mechanical behavior. It is important to understand the hydraulic behavior of hillslopes and quantify the effect of the uncertainty of mechanical parameters on hillslope stability for forewarning and hillslope management. Intra-hole deformation and displacement were record for the hillslope of the Babaoliao collapse site in the Chiayi County, as a case study. The fuzzy point estimation method and physical-based model were combined with the local factor of safety (LFS) theory to calculate the internal local factor of safety of the hillslope. A reliability analysis was then performed to determine the failure probability at different depths. Historical rainfall events were used to validate the model and predict the development of the failure probability for different rainfall patterns with the same warning rainfall. The results revealed that the failure probability model could effectively predict the area of hillslope instability and its changes over time and space. Different rainfall patterns affected the infiltration flux, leading to the difference in hillslope failure time. The delayed rainfall pattern had a significant impact on the time of slope instability, and shallow collapse was most likely to occur earlier. This study can be used as a reference for developing future hillslope warnings.