Exploiting seismic signal and noise in an intracratonic environment to constrain crustal structure and source parameters of infrequent earthquakes

In many regions of the world characterized by a relatively low rate of seismicity, the determination of local and regional seismic source parameters is often restricted to an analysis of the first onsets of P waves (or first motion analysis) due to incomplete information about Earth structure and the small size of the events. When rare large earthquakes occur in these regions, their waveforms can be used to model Earth structure. This, however, makes the nature of the earthquake source determination problem circular, as source information is mapped as structure. Presented here is one possible remedy to this situation, where through a two-step approach we first constrain Earth structure using data independent of the earthquake of interest. In this study, we focus on a region in Western Australia with low seismicity and minimal instrument coverage and use the CAPRA/LP temporary deployment to demonstrate that reliable structural models of the upper lithosphere can be obtained from an independent collection of teleseismic and ambient noise datasets. Apart from teleseismic receiver functions (RFs), we obtain group velocities from the cross-correlation of ambient noise and phase velocities from the traditional two-station method using carefully selected teleseismic earthquakes and station pairs. Crustal models are then developed through the joint inversion of dispersion data and RFs, and structural Green's functions are computed from a layered composite model. In the second step of this comprehensive approach, we apply full waveform inversion (three-component body and surface waves) to the 2007 ML= 5.3 Shark Bay, Western Australia, earthquake to estimate its source parameters (seismic moment, focal mechanism, and depth). We conclude that the full waveform inversion analysis provides constraints on the orientation of fault planes superior to a first motion interpretation.