Joint Inversion for Earthquake Depths Using Local Waveforms and Amplitude Spectra of Rayleigh Waves

Abstract

Reliable earthquake depth is fundamental to many seismological problems. In this paper, we present a method to jointly invert for centroid depths with local (distance < 5°) seismic waveforms and regional (distance of 5°–15°) Rayleigh wave amplitude spectra on sparse networks. We use earthquake focal mechanisms and magnitudes retrieved with the Cut-and-Paste (CAP) method to compute synthetic amplitude spectra of fundamental mode Rayleigh wave for a range of depths. Then we grid search to find the optimal depth that minimizes the joint misfit of amplitude spectra and local waveforms. As case studies, we apply this method to the 2008 Wells, Nevada Mw6.0 earthquake and a Mw5.6 outer-rise earthquake to the east of Japan Trench in 2013. Uncertainties estimated with a bootstrap re-sampling approach show that this joint inversion approach constrains centroid depths well, which are also verified by independent teleseismic depth-phase data.