Probabilistic Evaluation of Heterogeneous Landslide Influence Zones Accounting for Stratigraphic Dips With Borehole Data

Accurate prediction of landslide runout and hazard zones is crucial for effective disaster risk management. Current studies often overlook complex soil structures by using stationary isotropic or transversely anisotropic unconditional random fields (RFs) in landslide post‐failure modeling. Limited attention has been given to using borehole data for enhancing the accuracy of post‐failure behavior predictions in slopes. To address these issues, this study presents a novel framework to evaluate landslide hazard zones using conditional RFs. It integrates enhanced Bayesian Updating with Structural Reliability Methods (BUS) to infer soil parameter distributions, identify dips and capture soil nonstationarity. The process of landslide is simulated using the generalized interpolation material point (GIMP) method. Additionally, an automated strategy is proposed for the borehole location selection to ensure that predictions align with actual conditions. Results indicate that this method improves prediction accuracy for runout and influence distances by using borehole data and reduces uncertainty compared to unconditional RFs, thereby simplifying decision‐making and lowering control costs. Moreover, fewer boreholes are required for predicting runout distances compared to influence distances. This study highlights the necessity of considering complex heterogeneity and integration of borehole data in landslide risk management.