Role of the Andean structure in the Post-Seismic Deformation Following the 2014 Mw 8.1 Iquique Earthquake in Chile: New insights from a Finite Element Model Constrained by GNSS and InSAR Data

In the South American subduction zone, major earthquakes, thousands of kilometers apart, have occurred close in time (Maule 2010, Iquique 2014, Illapel 2015), and despite intriguing observations of surface velocity changes in distant regions, the mechanisms driving possible interactions between these events remain unclear. Viscoelastic relaxation is one phenomenon that can explain large-scale deformations, and modelling this process requires a good knowledge about the medium structure and the rheology. In this study, we first focus on the post-seismic deformation following the 2014 Mw 8.1 Iquique earthquake. The signal observed is used to add constrain on the structure and the rheology of the surrounding medium. Then, we extend the study to a larger scale. We investigate the influence of a rigid cold nose and of a Brazilian craton in the far field, as well as the role of the low viscosity mantle wedge.

Our dataset consists in 21 GNSS time series (processed with GipsyX) located in North Chile, Peru and Bolivia. The InSAR data consist in two Sentinel-1 time series (ascending and descending tracks) processed with the NSBAS chain, starting 7 months after the earthquake up to the end of 2019. We use a Finite Element Model (using the FEM software Pylith), firstly with a 2-dimensional model for the Iquique study, then extended it in 3D. Imposing a co-seismic slip on the interface, we investigate the relative contribution of afterslip and viscoelastic relaxation to explain the surface deformation. We vary the structure and the rheology of viscous zones and compare their impact on surface deformations, as well as the amplitude and location of the afterslip on the plate interface.

Our results reveal the necessity of a low viscosity area under the altiplano constrained by a long wavelength signal seen at long-term on the InSAR data, and confirmed by an uplift on the GNSS data. The location of this signal requires a rigid cold nose. We also show that a transient rheology with a lower viscosity at short term, using a Burgers rheology, is needed to fit the temporal evolution seen in GNSS. Tests on the viscosity with the 3D model highlight that the lower the viscosity, the more displacement is observed at large spatial scale. As it impacts a large area, relaxation in low viscosity zones could explain the far GNSS changes observations.