Decoding inter-seismic deformation: Insights from viscoelastic modeling

GPS positioning offers millimetric precision in measuring deformation of the lithosphere during the seismic cycle. In particular, during the post-seismic phase, long-lasting and large-scale deformation are measured. They result from the viscoelastic relaxation in the asthenosphere. Consequently, the post-seismic phase is currently modeled using viscoelastic rheologies (e.g., Maxwell or Burgers viscous models). On the other hand, the inter-seismic phase is mainly modeled using purely elastic models. In particular, coupling models, widely used to quantify the accumulation of deformation on the subduction fault, are therefore used to evaluate earthquake hazard. However, such elastic models fail to explain mid-field deformation without the use of an external hypothesis (e.g., a third plate called sliver).

The study of post-seismic deformation has provided important insights into the rheological properties of the asthenosphere during the post-seismic phase. For example, viscous creep has been found Newtonian since the cumulative post-seismic displacements normalized by the co-seismic offset, as a function of distance to the trench, superimpose very well for earthquakes of different magnitudes [Boulze et al. 2022].

By incorporating these different results and using the backslip theory [Savage 1983], we model the inter-seismic phase using viscoelastic models. We explore the impact on coupling distribution along the Chilean subduction zone, in particular discussing differences with the elastic model in terms of depth and lateral extension. We also examine the impact of viscoelastic models in a region of Chile (Taltal region, 25.2°S) where elastic models currently fail to reproduce deformation in the near-field [Klein et al. 2018]. Finally, we show that a 2-Burgers viscous model is necessary to reproduce deformation in Argentina in 2010, before the Maule earthquake.