Evaluation of the Temporal and Spatial Relationship of Slip and Tremor during ETS events in Cascadia

Abstract:

We characterize the temporal and spatial relationship of tremor and slip during several ETS events in Cascadia using GPS, strainmeter, and tremor observations. In Cascadia, tremor and slip tend to be broadly coincident, both spatially and temporally. However, GPS-derived slip inversions tend to be offset slightly updip of tremor. It remains unclear whether this offset is real or an artifact of the inversion methodology due to limited data resolution and the imposition of spatial and temporal smoothing. Borehole strainmeters provide greater precision and temporal resolution than GPS and provide an independent means to assess slow slip. However, non-tectonic artifacts and other sources of error, coupled with the limited number of high-quality stations, have made using strainmeters difficult. Typically, GPS displacements used in the assessment of slow slip are daily averaged values, which limit our ability to evaluate the sub-daily temporal relationship of tremor and slip. We utilize the greater temporal resolution of a select group of reliable strainmeters to compare the propagation of tremor and slip. We also assess the increased resolution of incorporating reliable strainmeters into joint GPS and strainmeter inversions of slow slip.

Tremor observations are used as a proxy for where slip occurs on the plate interface, while the amplitude of slip and spatial smoothness are optimized. Tremor-informed slip models are used to forward model surface displacements and strains, which are then compared with the GPS and strainmeter observations, as well as independent inversion results. Results indicate that tremor-derived slip distributions can provide an adequate estimation of the spatial and temporal propagation of slip, although there exists some regional discrepancies that vary on an event-by-event basis. Tremor-derived slip distributions that are systemically shifted updip of tremor can provide a better fit to the surface displacements and strains for all but one of the events assessed, indicating that the observed discrepancy of GPS inversions and tremor observations may be real.