A METHOD OF WAVEFORM INVERSION FOR EARTHQUAKE RUPTURE PROCESS

Abstract

We developed an inversion method to infer the earthquake rupture process from far-field body waves. The earthquake source is modeled by a space-time dependent slip on a fault plane. The slip motion at a point on the fault plane is assumed to be a ramp function in time, which is specified by the final dislocation, the rise time, and the time when the rupture front reaches the point. The fault plane is divided into many subfaults. The final dislocation and the rise time are assumed to be constant within each subfault. The rupture front motion is determined by the rupture starting time at the subfault corners using an interpolation formula. Thus the model parameters are the final dislocation, the rise time, and the rupture starting time assigned to each subfault. An algorithm to invert waveforms for the source model is formulated on the basis of a general approach of nonlinear inversion under prior constraints. Numerical experiments using simulated data show that the inversion method gives correct solutions. We also attempted a different approach using a constant rupture velocity model and found that we cannot estimate the true dislocation distribution under the assumption of constant rupture velocity.