Abstract

Inversions of geodetic data are regularly used to estimate the degree of locking in subduction zones. However, the ill-posed nature of these problems motivates us to include prior information, consistent with observed mechanical processes of rupture in the case of coseismic events. This results in these estimates being sensitive to the distribution of stations and the methods used to solve the inversions. To deal with this instability and find feasible solutions, we present an inversion method to estimate both coseismic slip and plate-locking, with associated uncertainties by inverting GNSS-derived displacements and velocities based on a Bayesian model selection procedure. Spatial locking correlations are also characterized for a better understanding of the rupture process. Our method allows us to impose positive constraints via a multivariate folded-normal distribution, with a specified covariance matrix. To validate our approach, we use the method to invert synthetic displacements from analytic plate-locking models. We apply our method to estimate the locking degree in Central Chile (28-39ºS) and its relation to the slip of above Mw 8.0 earthquakes on the Chilean subduction zone since 2010. Our results allow us to robustly characterize the locking degree and coseismic slip distribution along with their uncertainties, confirming the spatial correlation between locked asperities and the 2010 Mw 8.8 Maule and 2015 Mw 8.3 Illapel earthquake rupture zones. Future applications of this method include revisiting the degree of locking and slip of major earthquakes that have occurred in the last decades in Chile as well as inversions of transient signals.