GSA Connects 2021 in Portland, Oregon

Paper No. 231-11
Presentation Time: 4:25 PM


PRIEST, George, Virginia Institute of Marine Science, William & Mary, PO Box 738, Newport, OR 97365-0055 and ALLAN, Jonathan, Newport Coastal Field Office, Oregon Dept. of Geology and Mineral Industries, PO Box 1033, Newport, OR 97365-0055

The most important constraint on local tsunami runup and inundation is peak slip directly offshore, so establishing maximum peak slip is key to probabilistic tsunami hazard assessment (PTHA). Turbidite and paleotsunami data from the Cascadia subduction zone (CSZ) offer minimum slip constraints. Plate convergence rate provides a maximum slip constraint. Combining the two data sets for the last 10,000 yrs of paleoseismic observations offers a “window” of most likely peak slips for potential CSZ earthquakes. These inputs to a PTHA are essential for critical facility design, risk assessments, and development of inundation maps for evacuation and land use planning.

For example, a recent PTHA undertaken by AECOM for use in the International Building Code governing critical facility design uses what appear to be unreasonably large peak slip (asperity) scenarios relative to inferences from CSZ paleoseismic data and tectonic slip budgets. Approximately 40% of published AECOM CSZ sources have peak slip deficit releases >1,100 yrs of CSZ convergence, whereas paleotsunami simulations at Bradley Lake by Witter et al. (2012) and Priest et al. (2017) show that it is difficult within tectonic slip budget constraints to have even one such event in the last ~5,000 yrs. The only way that such a large slip event could have occurred would be if most of the 12 Cascadia tsunami inundations of the lake had occurred during times of maximum erosion of paleo-beach and dunes. Turbidite mass data from ~10,000-5,000 yrs B.P. (i.e., before the Bradley Lake record) are, however, compatible with at most two such extra-large (XL) events. One or two out of 19 full- margin CSZ ruptures and ~20+ partial ruptures inferred by Goldfinger et al. (2012) is far less than the ~40% used by AECOM. For this reason, Witter et al. (2011, 2013) assumed that only one XL event occurred in 10,000 yrs when constructing their logic tree. This disparity in CSZ sources is why tsunami simulations by AECOM of 2,475-yr exceedance wave heights are similar to those of the XL scenario rather than the expected Large (L) scenario of Witter et al. (2011, 2013) and Priest et al. (2013).