UNRAVELLING THE DEEP CRUSTAL STRUCTURE OF OKLAHOMA THROUGH 3D PROCESSING OF LOCAL EARTHQUAKE WAVEFORMS

Limited information exists about the nature of the deeper crust in the midcontinent region, particularly in Oklahoma. Not being accessible to direct sampling, the lower crust is usually investigated through teleseismic earthquakes and ambient noise tomography, which provide models with low lateral and vertical resolution. Based on such techniques, recent studies suggest high-velocity lower crust in some parts of the midcontinent region, but the spatial resolution of the models is comparably low and does not allow for detailed interpretation on local and regional scales.

Our aim is to use local earthquake waveforms to generate a 3D velocity model of the upper and lower crust to investigate the presence of high velocity crustal layer and what that implies in context of Oklahoma’s geological history. Local earthquakes in Oklahoma are of lower magnitude and significantly higher frequency than teleseismic events, thus offering the possibility of generating velocity models with higher resolution. A total of 27,582 local earthquake events from 2010-2017 are used in this study. The data are obtained from 10 seismic networks active during this time. The network and event coverage resemble an irregular 3D active seismic data acquisition, allowing for the adaption and application of wavefield-based processing techniques conventionally used in exploration seismology. Our approach is based on common mid-point (CMP) sorting and stacking of the local earthquake waveforms. This methodology simplifies the wavefields and increases the S/N ratio, especially at larger offsets allowing for deeper crustal investigation. For each CMP bin, we stack the waveforms to obtain 1D travel time curves representative of the velocity-depth function of the CMP bin. These travel time curves are then inverted for 1D velocity depth models. Finally, a 3D velocity model is derived from combination and interpolation of all 1D velocity models.

At this stage, our research evaluates velocity model based on stacking and inversion of the first arrivals of the earthquakes (Pg phase). Further, our aim is to apply similar techniques to shear wave arrivals (Sg) and Moho-related phases (PmP, Pn phases). This is part of our ongoing study to develop Moho depth and upper mantle velocity models for Oklahoma as well as to characterize Poisson’s ratio in the crust.