The Detectability of Offshore Slow Slip Earthquakes with Seafloor Pressure Sensors: A Sensitivity Analysis for the Cascadia Subduction Zone

Seafloor pressure records have been used to detect slow slip earthquakes (SSEs) in the updip portions of several subduction zones. The standard method for detection relies on differencing pressure records to eliminate oceanographic signals that are assumed to be spatially correlated throughout the sensor network. We evaluate this assumption and explore how the geometry of pressure sensor networks can be optimized to detect SSEs and minimize oceanographic signals. Our analysis focuses on the Cascadia Subduction Zone, where offshore SSEs have not yet been observed. We use a half-space fault model to calculate seafloor displacements for SSE scenarios over a range of magnitudes and locations and merge these displacements with hindcast seafloor pressure time series from regional oceanographic circulation models to simulate the pressure records for offshore SSEs. The oceanographic models show that circulation-related bottom pressure is strongly depth dependent with RMS amplitudes of tidally filtered, detrended records varying from >5 cm of water on the shelf to <2 cm on the abyssal plane. This compares to an observed 1-5 cm of peak vertical displacement from offshore SSEs in other settings. A preliminary model of an SSE event near the deformation front, that is similar in size to those observed in Japan and New Zealand, predicts that the oceanographic signals can be effectively removed by differencing when sensors are spaced up to 100 km apart along strike on isobaths; but the variability in the oceanographic signals with bathymetric depth makes the slow slip signal difficult to observe when the sensors are spaced more than a few tens-of-kilometers apart across strike. These differences are exacerbated for a similarly sized SSE beneath the shelf. We also directly compare time series from absolute pressure sensors from the Cascadia Initiative to the hindcast oceanographic models, finding that the real oceanographic pressure signals are somewhat higher amplitude and less spatially correlated than the oceanographic models predict. Therefore an experiment with pressure sensors along isobaths is needed to determine how well this approach of along-strike seafloor sensors actually improves the detectability of SSEs.