Topographic and hydrological controls on a shallow landslide: Prefailure evidence from a forested headwater catchment
Wei‐Li LiangAbstract
Comprehensive prefailure datasets combining soil depth distribution, topography and porewater pressure at the soil–bedrock interface are rarely available for natural shallow landslides. This study presents a case in which such datasets were collected before a shallow landslide in a forested headwater catchment, providing a unique opportunity to examine the dominant topographic and hydrological features triggering the landslide and to evaluate the predictability of its location and timing using prefailure information. The analysis revealed that the landslide area was characterized by thinner soil layers, steeper slopes, larger contribution areas and higher topographic wetness index values. A logistic regression analysis confirmed that bedrock topographic features better explained the landslide distribution than surface topographic features, with soil depth being the primary predisposing factor. The landslide event began under wetter‐than‐average conditions; at failure, widespread saturation developed across the site with a record‐high porewater pressure that exceeded the predicted value from the long‐term rainfall–porewater pressure relationship. This record‐high porewater pressure was attributable to both vertical rainwater percolation and groundwater exfiltration from bedrock layers. Integrating in situ information into a simple slope stability analysis improved the identification of locations with relatively high or low susceptibility; however, predicting the exact timing remained challenging. Notably, neither the amount nor the intensity of the triggering rainfall event was at a historical maximum, highlighting the limitations of rainfall‐only approaches and the added value of subsurface hydrological monitoring. Our field evidence demonstrates that porewater pressure exceeding a site‐specific threshold is a critical indicator for landslide occurrence. Establishing site‐specific rainfall–porewater pressure relationships and monitoring groundwater seepage from bedrock layers at susceptible locations can substantially improve early‐warning capabilities for landslides not triggered by extreme rainfall.