Influence of 3D Teleseismic Body Waves in the Finite-Fault Source Inversion of Subduction Earthquakes
Monday, 15 December 2014
Most large earthquakes are generated in subduction zones. To study the complexity of these events, teleseismic body waves offer many advantages over other types of data: they allow to study both the temporal and spatial evolution of slip during the rupture, they don't depend on the presence of nearby land and they allow to study earthquakes regardless of their location. Since the development of teleseismic finite-fault inversion in the 1980th, teleseismic body waves have been simulated using 1D velocity models to take into account propagation effects at the source. Yet, subduction zones are known to be highly heterogeneous: they are characterized by curved and dipping structures, strong seismic velocity contrasts, strong variations of topography and height of the water column. The main reason for relying on a 1D approximation is the computational cost of 3D simulations. And while forward simulations of teleseismic waves in a 3D Earth are only starting to be tractable on modern computers at the frequency range of interest (0.1Hz or shorter), finite-fault source studies require a large number of these simulations. In this work, we present a new and efficient approach to compute 3D teleseismic body waves, in which the full 3D propagation is only computed in a regional domain using discontinuous Galerkin finite-element method, while the rest of the seismic wave field is propagated in a background axisymmetric Earth. The regional and global wave fields are matched using the so-called Total-Field/Scattered-Field technique. This new simulation approach allows us to study the waveform complexities resulting from 3D propagation and investigate how they could improve the resolution and reduce the non-uniqueness of finite-fault inversions.