Bayesian back analysis of unsaturated hydraulic parameters for rainfall-induced slope failure: A review

Rainfall-induced soil slope failures are among the most frequent and widespread landslides in the world. The infiltration of water plays a critical role in the instability of slopes under rainfall conditions. By employing the physical models, field monitoring data becomes instrumental for the back analysis of soil parameters, providing critical insights into the mechanisms of slope instability. In this review, the initial section investigates the hydraulic characteristics of unsaturated soils, encompassing the soil-water characteristic curve, permeability function, and their variability nature. Subsequently, a comprehensive overview of infiltration models for unsaturated soil, coupled with their numerical and analytical solutions, is provided. The subsequent section systematically introduces physical model-based Bayesian parameter inference, with a particular emphasis on the formulations of likelihood functions. The exploration then shifts towards investigating the impacts of prior knowledge, likelihood, model structure, and assimilation approach on an unsaturated soil slope monitoring case. Finally, the limitations of the current approaches and future outlooks in this field are presented.