Estimating Inland Ground Motions from Lake Turbidite Sequences, Northern Cascadia margin, USA.

Using cores collected from lakes in northern Oregon and Washington, we are attempting to estimate ground motions from plate boundary earthquakes at inland paleoseismic sites. Paleoseismic evidence in Cascadia comes largely from coastal and offshore sites, while population the main population centers of Seattle, Victoria, Vancouver and Portland are 100-180 km inland. Cores from Leland Lake on the Olympic Peninsula, Lake Sawyer, near Seattle, and Bull Run Lake, 65 km east of Portland contain sequences of event beds that are interpreted as internal lake turbidites. The number, timing based on 14C constrained age models, sequencing, and individual event characteristics correlated with physical properties and CT data are compatible with onshore and offshore paleoseismic records of plate boundary earthquakes. The likely correlative turbidite sequence at Bull Run is well-matched to the nearest offshore turbidite sequences at Hydrate Ridge and Oceanus Basin (see also Hausmann et al. this meeting). Similarly, the Washington lake sequences are well matched to the offshore Washington sequences (Goldfinger et al. 2016), with the likely inclusion of a single Seattle Fault earthquake 1000 cal BP. Bull Run Lake has several ashes, but otherwise, additional event beds related to crustal faulting or other events are not observed. Our strategy is to investigate lakes that have low sensitivity to subaquatic slope failures in order to explore the limits of stability. In this case, the minimum ground shaking required for slope failure will approach actual ground motions as stability increases. We mapped failure zones within the lakes, and collected shear vane measurements to estimate sediment cohesion. We then computed minimum ground motions for these sites. For Leland Lake, there are no mappable failures, indicating internal lake turbidites likely were generated by thin surface failures below mapping resolution. For Sawyer and Bull Run, the most stable failure sites require 0.2-0.3g PGA, which we consider to be minimum values representative of the most energetic events in the Holocene sequence. Given the lack of evidence for crustal earthquakes or other events in addition to the probable plate boundary sequence, we consider these values to be potentially representative of the largest Holocene plate boundary earthquakes.