Decreases in Root Strength and Hydraulic Conductivity Promote Reactivation of Coastal Bluff Landslides, Puget Sound, WA

To improve landslide susceptibility maps and forecasts of landslide occurrence we need to better understand the behavior and triggering conditions of reactivated and recurring landslides. The inherent unpredictability and destructiveness of landslides, however, makes it difficult to observe changes that happen during failure. Initial landslides may cause substantial alterations to subsurface hydraulic and mechanical properties, as well as topography and vegetation, often increasing the likelihood that the remaining landslide deposits will fail again in the future. In this contribution, we capitalized on an existing dataset combining field and laboratory measurements from a reactivated landslide along the coastal bluffs of Puget Sound to examine which changes were most influential in promoting landslide recurrence. We developed a numerical model of coupled hydro-mechanical processes for idealized hillslopes parameterized with the existing field data. Using a nearby unfailed hillslope as a control, we performed a series of virtual experiments to test the relative importance of disturbances to vegetation, strength, and hydraulic properties for determining the timing and location of slope failures. We found that shifts in root reinforcement and saturated hydraulic conductivity had the largest impacts in our simulations, followed to a lesser degree by properties of the soil water retention curve. We conclude that the loss of root strength and decrease in conductivity which occurred during the initial slope failure made the landslide site more likely to fail in the future relative to the unfailed control site. This suggests that practitioners should prioritize collecting data on hydraulic conductivity and the extent and type of vegetative cover when evaluating how susceptible existing landslides are to future failure. The critical role of root reinforcement in our model also highlights the need for comprehensive measurements of root strength variability through space and time to incorporate broadly into future landslide susceptibility estimates.