Joint Inversion Methods For Source Parameter Determination Under Point Source Approximation

Reliable centroid depth is fundamental to many seismological problems. In this paper, we present a method to jointly invert for earthquake centroid depths with local (distance< 5°) seismic waveforms and regional (distance of 5°～15°) Rayleigh-wave amplitude spectra in sparse networks. During the inversion, earthquake source parameters are first retrieved from the Cut-and-Paste (CAP) method, and the mismatch between synthetic and observed Rayleigh wave amplitude spectra is computed for a range of depths. Then the joint error is defined by combining amplitude spectra mismatch and local waveform mismatch, and a grid search is performed to find the optimal centroid depth. As case studies, this method is applied to determine depth of the well-studied 2008 Mw6.0 Nevada earthquake and a 2013 Mw5.6 outer-rise earthquake to the east of Japan Trench. Uncertainty of centroid depth is estimated with a bootstrap resampling approach for different number of stations. The centroid depths are constrained well with the joint inversions, and are also confirmed with teleseismic depth-phase data.Rapid and accurate determination of full moment tensor plays an important role in seismic source studies. In this paper, we developed a joint inversion method for full moment tensor determination that combines both local and teleseismic waveform data. This method, called "gCAPjoint", is based on CAP method which enables time shift for synthetics, thus is robust to inaccurate structural velocity models. Meantime, this method samples both local and teleseismic phases, thus make use of better azimuthal and ray parameter coverage to provide improved constraints on source parameters. To test the accuracy and robustness of gCAPjoint method, we computed synthetic waveforms stacked applied station noise on them, and performed the method for inversion. The inverted source parameters are similar to input values. As case study, we applied the method on two moderate to strong earthquakes occurred in Hindu Kush intermediate-depth earthquake zone. We obtained similar focal depth and full moment tensor with solutions from several catalog projects including Global CMT. This reveals that gCAPjoint provides reliable source parameter determination on these two events. However, moment tensor of these two intermediate-depth earthquakes shows different double couple proportion, though the two events are located very close to each other. This fact may suggest intrinsically different rupture process and dynamics for these two events.