Identifying Physics‐Based Thresholds for Rainfall‐Induced Landsliding

Abstract Most regional landslide warning systems utilize empirically derived rainfall thresholds that are difficult to improve without recalibration to additional landslide events. To address this limitation, we explored the use of synthetic rainfall to generate thousands of possible storm patterns and coupled them with a physics‐based hydrology and slope stability model for various antecedent soil saturation scenarios to analyze pore water pressure and factor of safety metrics. We used these metrics to generate two‐tiered alert thresholds that can be employed to assess shallow landslide potential for any given combination of storm and antecedent wetness. When applied to the San Francisco Bay region (CA, USA), the results are consistent with events that caused widespread landsliding. Our deterministic modeling approach, which accounts for plausible ranges in soil hydraulic and mechanical properties, can inform the development of the next generation of warning systems for rainfall‐induced landsliding.