Imaging the next Cascadia earthquake: Optimal design for a seafloor observation network given realistic data uncertainties

Abstract:

An outstanding question in the Cascadia subduction zone is the degree and spatial extent of interseismic strain accumulation on the subduction megathrust. Seafloor geodetic methods are capable of imaging this strain accumulation on the offshore portion of the subduction zone and therefore anticipating the potential size and rupture pattern of a future earthquake. However, the high cost of seafloor geodesy means that only a limited number of stations may be deployed and monitored. To facilitate expansion of current geodetic networks offshore, we develop a quantitative recommendation of optimal locations for future seafloor geodetic measurements, based on the amount of new information provided by that observation. The optimal network may be different for observing different physical processes (e.g., subduction zone locking rates, coupling rates, total moment rate deficit, etc.), and will also depend on a number of modeling and data uncertainty assumptions. In particular, data uncertainty assumptions will change over time, as more position observations reduce velocity uncertainties. We quantify how well different network geometries can resolve different physical processes using differential entropy, a concept from information theory that measures the uncertainty in the value of a random variable. We find that near-trench observations on the megathrust hanging wall, distributed along-strike, consistently provide significant reduction in differential entropy over a large suite of assumptions, and that a well-placed seafloor observation can provide up to ~30 times the information gain of the most optimal onshore observation.