INFLUENCE OF SEISMIC SITE RESPONSE ON LANDSLIDE REACTIVATION DURING THE NEXT CASCADIA EARTHQUAKE

Coseismic landslides are known to cause widespread damage and loss of life around the world—in some cases exceeding damage caused by the earthquake. The Oregon Coast Range (OCR) is likely to experience extensive landsliding during the next Cascadia Subduction Zone (CSZ) earthquake and for many subsequent years when winter storms saturate slopes weakened by strong shaking. These landslides can wreak havoc on evacuation, emergency response, and recovery efforts throughout Cascadia by destroying critical highway routes and infrastructure. Despite coseismic landslides being common in tectonically active areas, the triggering mechanisms have only recently received considerable attention and the interaction between seismic waves and topography is still poorly understood. The potential for coseismic landsliding is particularly pressing in the OCR, where thousands of large, deep-seated landslide deposits exist. Yet, historic activity of these features is ominously rare, suggesting CSZ earthquakes may regulate or exacerbate reactivation by amplifying ground motion at ridgetops or within the deposit. In a pilot study to quantify the site response of an ancient deep-seated (>3 m) landslide and likelihood of coseismic movement, we deployed 5 short-period seismometers on and adjacent to a deposit in the OCR. We analyzed the amplitude spectrum from weak motion events and ambient noise in order to calculate horizontal to vertical spectral ratios (HVSR). By comparing the HVSR to our local bedrock reference station, we found a 2-3x amplification in horizontal motion within the landslide with similar magnitude amplification at the ridgetop—implying topography can influence site response as much as substrate, contrary to previous assumptions. Comparing the shaking response to landslide depth inferred from seismic refraction data, we determined that the lowest amplification coincided with the shallowest landslide deposit, suggesting that thicker deposits may be more prone to coseismic reactivation. With additional data from other landslides in the area, we can determine a correction factor for probabilistic maps of coseismic landsliding based on input ground motion and landslide characteristics—information needed for modeling CSZ ground motion patterns and distribution of potential landslide reactivation.