Coseismic slip distribution of the 2015 Mw7.8 Gorkha, Nepal, earthquake from joint inversion of GPS and InSAR data for slip within a 3-D heterogeneous Domain

Abstract

We derive a coseismic slip model of the 2015 Mw7.8 Gorkha earthquake on the basis of GPS and line-of-sight displacements from ALOS-2 descending interferograms, using Green's functions calculated with a 3-D finite element model (FEM). The FEM simulates a nonuniform distribution of elastic material properties and a precise geometric configuration of the irregular topographical surface. The rupturing fault is modeled as a low-angle and north dipping surface within the Main Frontal Thrust along the convergent margin of the Himalayas. The optimal model that inherits heterogeneous material properties provides a significantly better solution than that in a homogenous domain at the 95% confidence interval. The best fit solution for the domain having a nonuniform distribution of material properties reveals a rhombus-shaped slip zone of three composite asperities. Slip is primarily concentrated at a depth of 15 km with both dip-slip (maximum 6.54 m) and strike-slip (maximum 2.0 m) components, giving rise to a geodetic-based moment of 1.09 × 1021 Nm in general agreement with the seismological estimate. The optimal relative weights among GPS and interferometric synthetic aperture radar (InSAR) are deduced from a new method, MC-HVCE which combines a Monte Carlo search and a Helmert Method of Variance Components Estimation. This method determines the relative weights in a systemic approach which preserves the intrinsic solution smoothness. The joint solution is significantly better than those inverted from each individual data set. This methodology allows us to integrate multiple data sets of geodetic observations with seismic tomography, in an effort to achieve a better understanding of seismic ruptures within crustal heterogeneity.