Open-File Report O-24-11, Improved Cascadia Earthquake Source Models for Tsunami Hazard Assessment

This report describes the development of a suite of new Cascadia Subduction Zone (CSZ) megathrust source scenarios. Compared to previous Cascadia tsunami source models, several improvements are made. These improvements have been aided by the recent release of new data from the Cascadia Seismic Imaging Experiment 2021 (CASIE21), reinterpreted legacy seismic data, seismicity, and low-frequency earthquakes (LFE) in Episodic Tremor and Slip (ETS) events. 

PUBLICATION DOWNLOADS

This report describes the approaches used to construct megathrust source scenarios by integrating multiple datasets, the overarching goal being the eventual development of new probabilistic tsunami hazard analyses for Cascadia. Compared to previous Cascadia tsunami source models, several improvements are made. The megathrust geometry model is much improved, owing mainly to new data from the Cascadia Seismic Imaging Experiment 2021 (CASIE21). Constraints in addition to the depth of CASIE21 imaging presently include earlier seismic surveys, seismicity, and low-frequency earthquakes (LFE) in Episodic Tremor and Slip (ETS) events. Offshore of northern California, which was not imaged by CASIE21, the geometry is constrained by reinterpreted legacy seismic data and seismicity. In the downdip direction, we include three main types of rupture scenarios: buried rupture, splay faulting rupture, and trench-breaching rupture. In all cases, the downdip rupture limits are the same as assumed in the USGS National Seismic Hazard Model. CASIE21 imaging, combined with high-resolution sparker imaging, does not support the interpretation of the presence of a continuous active mega-splay fault along the margin. Although these data yield limited evidence to allow a hypothetical mega splay from 43°N–47°N, which will likely have some implications for tsunami hazard in Washington state and British Columbia. To comply with local geological evidence, we incorporate Holocene turbidite, estuary, and lake records, and known structural heterogeneity to guide the design of rupture lengths and recurrence intervals. Finally, to account for aleatoric uncertainties, we also construct random rupture scenarios that are not based on geologic evidence.