Failure Process of the September 22, 2017 Wrangell Mountain landslide in Alaska

Abstract:

Landslides can drastically change landscapes and affect regional ecosystems. Therefore, understanding landslide failure mechanisms has important societal relevance. Further, a better understanding of landslides can help to improve deciphering seismic signals associated with surficial processes. However, landslide processes remain poorly understood. Geodetic observations like satellite imageries can identify landslides but the sparse coverage and sporadic repeating time hamper detailed investigations of landslide dynamics. Alternatively, seismic observations can provide continuous and near-real-time monitoring of landslides, which could detect and locate events in remote regions. Yet, the lack of clear body wave phases causes landslide sources often missing in standard catalogs, but surface waves have proven useful in detecting and locating such unconventional events. In this study, we use surface waves from the dense USArray network and the Automated Event Location Using a Mesh of Arrays method to study landslides in Alaska. We detected a landslide event in the Wrangell Mountain region on September 22, 2017 that had not been reported in previous studies despite that the event generated a coherent surface wavefield across Alaska and was recorded by stations in the central US. We further use a frequency domain method to invert for a centroid single force (CSF) model to understand the slope failure event. The CSF model is relatively robust and can explain seismic records that were not used for the inversion. We found the landslide lasted for about 100s but only with an apparent surface wave magnitude (Ms) of 0.96. Intriguingly, the CSF model suggests the event involving at least three distinct episodes, which may have been modulated by the complex local topography. We plan to combine geodetic and seismic observations to estimate the total mass and infer the landslide trajectory. With the resolved trajectory and acceleration, we will further investigate and compare the dynamic frictional coefficients during different failure episodes. This landslide is likely a representative case of mass wasting events in the region, and we hope to gain improved understanding of landslide activity and their dynamic attributes in Alaska.