SENSITIVITY OF GLACIAL TERRACE STABILITY TO VARIABLE STRATIGRAPHY, PRECIPITATION, AND FAILURE HISTORY

Along the glacial terraces that line valleys west of the Cascade foothills in northwest Washington State, landslide volume distributions and failure styles vary strongly between adjacent valleys. What controls the differences in location and size of these slope failures? Here we use a coupled modeling approach to explore how variations in stratigraphic sequencing, precipitation rate, topographic relief, and failure history potentially affect groundwater hydrology and landslide size within glacial terraces.

We use a three-dimensional limit-equilibrium slope-stability analysis (Scoops3D) to systematically explore how the relative thickness and position of strata within a hillslope affect predicted landslide size. We find that landslide volumes vary greatly with the position and thickness of the weakest strata, with peak volumes expected for geometries similar to that of the Whitman Bench along the North Fork Stillaguamish River valley. For a given stratigraphic sequence, changes in terrace relief result in changes in failure pattern as scale-dependent factors like cohesion alter slope stability.

Variably saturated groundwater-flow simulations through a range of terrace architectures reveal that groundwater can either amplify the pattern of instability set by stratigraphy or spatially shift instability toward zones of high pore-water pressure. Similarly, terrace architecture can modulate the hydrologic response to increased precipitation. Pore-water pressures in perched water tables near the ground surface increase, while those deeper in the hillslope may decrease as groundwater preferentially exfiltrates above. Thus, an increase in precipitation can lead to relative stability for deeper-seated slides in some cases.

Landslide events prior to the 2014 Oso landslide also suggest that previous landslide activity may increase large-volume landslide occurrence in glacial terraces. We model groundwater flow through terraces with and without remnant landslide deposits and show that the presence of these deposits can enhance pore-water pressures at the hillslope base. The presence of older landslide deposits thereby leads to a competition between the stabilizing effects of topographic buttressing and the destabilizing effects of groundwater flow.