Figure 2. Average slip along strike of different existing coseismic
models. Solid lines show the average coseismic slip along strike of
published models, dashed line show the average coseismic slip along
strike of our preferred coseismic model with 120km fault width.
Stress-driven afterslip provides a physically based model for
post-seismic afterslip (Wang and Fialko, 2018). In this kind of model,
the slip distribution and time history of afterslip is determined by the
coseismic slip and the frictional properties of the fault plane. Under
the assumption of frictional homogeneity on the plane, the spatial
pattern of the stress-driven afterslip is determined by the coseismic
slip distribution, while the frictional parameters control the time
evolution of slip and displacement.
In this study, we first compare the stress-driven afterslip predictions
for three published slip models (Elliottet al., 2022; Liu et al., 2022;
Ye et al., 2022), and find that most of the models show significant
misfit to the postseismic displacements (section 4.1). We find that the
azimuthal misfits in some models result from the peak afterslip being
located too close to the coastline; the location of the peak afterslip
is determined mainly by the downdip end of the slip distribution. Using
the insights gained from these comparisons, we then systematically vary
the coseismic slip inversion to identify how the slip model must change
to best predict the afterslip, considering a range of models for the
contribution of viscoelastic relaxation. Finally, as an additional test
of our conclusions about the slip model, we compare the model
predictions to the data from GPS campaign measurements collected three
weeks after the earthquake, which contain both coseismic and postseismic
signal.
2 Data