Seismic Shear Wave Velocity Response During Seasonal Initiation and Cessation of Earthflow Motion

Oak Ridge Earthflow in California's Diablo Range is a clay-rich earthflow that has moved downhill at an average rate of ~2 m/year for nearly a century, with acceleration typically occurring during the wet winter months followed by deceleration during the dry summer. Observations of stick-slip motion and shear localization at Oak Ridge suggests that currently observed displacement is frictional. That said, fence posts displaced over several decades show a component of internal deformation like a fluid, something also suggested by the flow-like morphology of the landslide. We monitored seismic Rayleigh wave velocity (VR), which is sensitive to shear stiffness (and hence the propensity to flow) as well as pore pressure in order to explore the possibility that some motion at Oak Ridge is linked to fluid flow rather than frictional sliding. Three passive, short-period seismic stations were installed in 2020, which we used to measure relative changes in VR by computing ambient noise cross-correlations between each station pair. Based on similar previous studies, we chose a 7-9 Hz bandpass filter to target the depth of the shear zone, which is saturated year-round. We also utilize three GPS stations, four piezometers, repeat airborne LiDAR surveys, and a weather station to monitor earthflow movement, pore pressure, and rainfall.

In general we observe that increases in landslide pore pressure are accompanied by decreases in seismic shear wave velocity (dV/V), an observation that could reflect either changes in pore pressure or shear stiffness. That said, we also found that dV/V response varies spatially in concert with landslide velocity (as measured via deformation of the LiDAR point cloud between 2018 and 2020): the largest dV/V reductions are associated with areas of the earthflow that experience greater motion. This observation implies one of two possibilities: 1) landslide acceleration is driven in part by a loss of shear stiffness, which is suggestive of fluid motion, or 2) faster moving sections of the landslide experience greater variation in pore fluid pressure, perhaps due to differences in material properties. Further work is needed to 1) identify the frequency-dependence of the dV/V reductions to bracket response depth and 2) characterize the spatial heterogeneity in landslide material properties at Oak Ridge.