Reconstructing Hydrodynamic Flow Parameters of the 1700 Tsunami at Ecola Creek, Cannon Beach, Oregon

Coastal communities in the western U.S. face risks of inundation by distant tsunamis that travel across the Pacific Ocean as well as local tsunamis produced by great (M >8) earthquakes on the Cascadia subduction zone. In 1964 the M 9.2 Alaska earthquake generated a distant tsunami that flooded Cannon Beach, a small community (population 1640) in northwestern Oregon, causing over $230,000 in damages. However, in the wake of the 2004 Indian Ocean tsunami, renewed concern about the potential impacts of a local Cascadia tsunami, has motivated a need for closer examination of the hazard. This study applies a simple sediment transport model, TsuSedMod (Jaffe and Gelfenbaum, 2007), to reconstruct the flow speed of the most recent Cascadia tsunami that flooded the region in 1700 using the thickness and grain size of sand layers deposited by the waves. Sand sheets recording the 1700 tsunami were sampled in the field and analyzed in the laboratory to produce model inputs. TsuSedMod calculates tsunami flow speed from the shear velocity required to suspend the quantity and grain size distribution of the observed sand layers. The model assumes a steady, spatially uniform tsunami flow and that sand deposits form from sediment falling out of suspension when the flow stops. Assuming sensitivity analyses test the appropriate parameter values found in nature, flow speeds estimated for the 1700 tsunami range from about 5 to 9 m/s. Using flow depths constrained by tsunami simulations for Cannon Beach, the sediment model calculated flow speeds of 6.5 to 7.6 m/s for sites within 0.3 km of the beach and higher flow speeds (7.4 to 8.8 m/s) for sites 0.6 to 1.2 km inland. The higher flow speeds calculated for the two sites furthest landward contrast with much lower maximum velocities (<3.8 m/s) predicted by the simulations. Grain size distributions of sand layers from the most distal sites are inconsistent with deposition from sediment falling out of suspension. We infer that rapid deceleration in tsunami flow caused convergences in sediment transport and, therefore, the higher flow speeds calculated by the sediment model may overestimate the actual wave velocity. Key recommendations for future research include investigations focusing on sites with low-relief and simple geography and multidisciplinary studies that couple tsunami sediment models with inundation models to more accurately estimate flow parameters from tsunami deposits.