6. Conclusion
We generated a suite of coseismic slip models for the 29 July 2021, Mw 8.2 Chignik earthquake by inverting seismic and geodetic data, and varying the assumed downdip end of coseismic plane. Models with a narrower allowed downdip width have slip located farther offshore, and have a more abrupt downdip termination of slip, while models that allow a wider rupture have slip that tapers to zero more gradually with depth.
For each model, we computed stress driven afterslip models and viscoelastic relaxation models, and compared the postseismic model predictions to observed postseismic displacements over the first 3 months after the earthquake. As long as the frictional properties are uniform, the spatial pattern and relative magnitudes of postseismic afterslip displacements are determined entirely by the coseismic slip distribution, while the rate of early postseismic slip and its time decay depend on the values of the frictional parameters chosen. The predicted afterslip displacement pattern is significantly different for the different coseismic models, but the predicted viscoelastic relaxation deformation is not.
We find that the coseismic data alone cannot resolve the details of slip downdip end of the rupture, but the postseismic displacement provide important new information. By limiting the model fault plane width 120km downdip, the observed post-seismic signal is much better explained with stress-driven afterslip for all reasonable viscoelastic relaxation contributions considered. A model with a narrower downdip extent of slip, and a more abrupt downdip termination of slip, produces afterslip located farther offshore, and this is necessary to match the orientations of the observed postseismic displacements. This finding holds for all reasonable contributions from viscoelastic relaxation, including models where the viscoelastic signal is negligible. However, overall data fit is improved when a combination of stress-driven afterslip and viscoelastic relaxation is included in the model, and this combination makes it possible to explain the vertical displacements as well as the horizontal.
The preferred coseismic rupture plane with a 10km fault width also has a much better model-data fit for the 3-week coseismic + post-seismic deformation at GPS campaign sites , in which time period the coseismic signal should be the dominant, which further confirms that the spatial pattern of stress-driven afterslip brings new information of the coseismic rupture of the 29 July 2021, Mw 8.2 Chignik earthquake.
Our result indicate an abrupt instead of gradual downdip termination of coseismic slip. The lack of deep aftershocks further supports this conclusion and suggests that there was limited interseismic slip deficit deeper than the coseismic rupture, even accounting for stress-shadowing effects, in line with the Xiao et al. (2021) coupling model for this region.