Developing a regional 3-D velocity model in southwest Texas for monitoring seismicity in the Eagle Ford Shale Play

The State of Texas has commissioned the Bureau of Economic Geology to install a seismic network (TexNet) which, when complete, will employ 22 permanent and 33 portable new stations. In the region of southeast and coastal Texas, where it consists of a rich oil field - the Eagle Ford Shale, the felt seismicity has recently increased. Prior to the operation of TexNet, a seismogenic zone has been identified using data from the Transportable Array (2009-2011). In this seismogenic zone, seismicity has been reported as early as in the year 1973, while the largest event occurred on October 20^th, 2011 with moment magnitude of 4.8. Since January 2017, TexNet has also detected a total of nearly 100 earthquakes in the area of Eagle Ford Shale, suggesting that this region is still seismically active.

Despite that a number of studies have been focused on the velocity structure of Texas with a large dimension consisting of very coarse grid spacing and extended as deep as a several hundred kilometers depth, there is a need of a detailed 3-D velocity model focusing on the region of Eagle Ford Shale to better locate earthquakes on a daily basis. In this study we were based on data of local earthquakes detected by the TexNet and previous studies, along with the teleseismic earthquakes, to perform a joint tomographic inversion. Based on the initial hypocentral information determined by the TexNet's routine process, we further relocated these earthquakes using the double-difference relocation method (i.e., hypoDD) based on an initial 1-D velocity model obtained from reflection seismic profiles. At the same time, we employed statistic regression (i.e., the Wadati diagram) to constrain the origin times of these relocated earthquakes, while their hypocentral locations have been better constrained by hypoDD relocation. The constrained origin times and relocated local earthquake hypocenters, along with the teleseismic earthquakes, were used to perform raytracing to update the 1-D velocity model obtained from seismic profiles. This procedure help reduce the possible bias of the starting model, while the resulting 3-D model from raytracing can be used as a new starting model for the subsequent full iterations of tomographic inversion.

Compared to the initial hypocentral locations, preliminary results of tomographic relocation have shown a tightly clustered linear seismicity that runs in an overall northeast-southwest direction which comprises several secondary earthquake clusters. Some of the secondary seismic clusters were recently detected by the TexNet, where no seismicity was detected during the deployment of the Transportable Array in 2009-2011. Our next steps are to validate robustness of the tomographic model and to incorporate it into the TexNet's routine seismicity detection process.