Impact of topography on coseismic modeling and earthquake static slip inversions

Abstract

After an earthquake occurs, slip models of the event may be produced by inverting geodetic data for slip on a finite fault. This process generally requires coseismic Green's functions, which must be calculated in advance. The vast majority of such studies use Green's functions that ignore any 3D structure that is present in the region. In this dissertation, I investigate the effects of 3D structure, particularly topography, on forward models of coseismic deformation and on earthquake static slip inversions.

First, we develop a new software package, SPECFEM-X, which can model seismic deformation in complex domains by using Gauss-Legendre-Labotto quadrature to solve a quasi-static equation of motion. SPECFEM-X implements infinite elements, which allow us to simulate far-field boundary conditions. Earthquake sources are implemented using a split-node approach or with moment-density tensors.

We then use SPECFEM-X to explore the impact of topography on forward models of coseismic deformation. We show that the presence of a hill in a synthetic domain changes the shape of predicted surface deformation patterns due to a given slip model. Then we compute surface deformation for slip models of two real events: the 2010 Maule and 2015 Gorkha earthquakes. In both cases, we find that the presence of topography alters the shape of the surface deformation pattern.

Given these findings, we would like to know whether the use of topographic Green's functions would result in a different estimated slip model. We compute coseismic Green's functions for the 2010 Maule and 2015 Gorkha earthquakes in domains with and without topography, and perform Bayesian inversions using geodetic data from these events. In both cases, we find that the use of Green's functions with topography yields a different distribution of slip. The difference between the models is much greater for Maule, which has poorer data coverage.

Our findings suggest that the presence of topography may alter the shape of observed surface deformation patterns and can have a significant effect on earthquake static slip inversions. Therefore, Green's functions with topography should be used when inverting for coseismic slip in regions with strong topographic gradients, particularly when data coverage is poor.