3.2.2 Viscoelastic Relaxation Simulation
Although afterslip is likely to be the dominant process in the early period of the post-seismic deformation, for an Mw8.2 event it is necessary to consider the viscoelastic contribution to the observed post-seismic signal as well (Sun and Wang, 2015. However, it is known to be difficult to separate the two post-seismic processes in the early time period Sun and Wang, 2015. Therefore, we consider a range of viscoelastic relaxation models based on different assumed viscoelastic structures, subtract these from the data, and then estimate the best-fitting afterslip model for each.
The numerical simulation we use is the software package VISCO2.5D, which imposes source and domain boundary conditions on a 2-D structure to approximate the 3-D equations of quasi-static equilibrium by the spectral element method (Pollitz, 2014). We use the Slab2.0 slab geometry at the center of the coseismic rupture and assume a laterally homogeneous viscosity structure (Figure 4a, b). The elastic slab separates the sub-oceanic asthenosphere from the mantle wedge. We vary the viscosity in the mantle wedge, and include a 90km thick elastic slab (including its mantle lithosphere). We also varied the continental lithosphere thickness and the presence of a cold nose to the mantle wedge (e.g., Luo and Wang, (2021)). All of these model assumptions affect our model predictions (see section 4.2).
(a) (b)