CASCADIA MEGATHRUST TSUNAMI SOURCE SCENARIOS FOR HAZARD ASSESSMENT AND EARLY WARNING

For tsunami hazard assessment and early warning along the Cascadia margin, we construct full-margin megathrust rupture models using a three-dimensional dislocation model. We employ a megathrust geometry based on a combination of previously published geometrical models and recently available low-frequency-earthquake hypocenter locations. Models created involve buried rupture, splay faults, frontal-thrust and back-thrust. Like previous studies, the buried rupture and splay-faulting rupture scenarios show that there can be large seafloor uplift and coastal subsidence, which will lead to tsunamis that seriously affect the coastal area. We also design rupture scenarios involving coseismic rupture of frontal-trusts at the deformation front, analogous to the trench-breaching rupture in the 2011 Tohoku-oki earthquake. Our full-margin splay fault scenarios are modified from one that was used for tsunami hazard assessment for Oregon. They result in the worst-case tsunami scenarios due to the larger seafloor uplift at the trace of the splay fault. However, if the whole margin were to rupture, we think the more likely scenario may be a combination of different types of tsunami-genic rupture in different segments along strike. We speculate the buried rupture may be the dominant mode for most of the margin, with splay-faulting and frontal-thrust ruptures being locally important for some segments. The dislocation models indicate that a properly configured land-based network of real-time Global Navigational Satellite System (GNSS) stations can recognize the type and length of a rupture along the Cascadia megathrust and therefore can effectively contribute to real-time tsunami early warning. In northernmost Cascadia, such a network can also recognize non-tsunami-genic strike-slip rupture of the Nootka fault zone and hence prevent potential false alarms. However, our modeling results also imply that the GNSS land-based measurements are not sensitive enough to determine the coseismic fault behaviour of the shallow portion of the megathrust far offshore, validating the urgent need for more seafloor observations near the deformation front such as the real-time seismic and pressure data being collected by the Ocean Networks Canada.