S015-06
Machine Learning Reveals Additional Regions of Hydraulic Fracture Induced Seismicity in the Eagle Ford Shale
Machine Learning Reveals Additional Regions of Hydraulic Fracture Induced Seismicity in the Eagle Ford Shale
Tuesday, 8 December 2020: 16:22
Virtual
Abstract:
Southern Texas has a history of active oil and gas production, hydraulic fracturing (HF), wastewater disposal, and seismicity, with some areas of pervasive faulting. With the advancements in horizontal drilling and HF, the Upper Cretaceous Eagle Ford shale has become a focus of shale-gas development since 2008. A recent study of the region generated a catalog of template matched events (2014-2018) and found that 86% of earthquakes were closely related in time and space to HF wells mostly in Karnes, Atascosa, and Gonzales counties. The number of earthquakes in the Eagle Ford has continued to increase, illuminating a new region of seismicity in Live Oak county. We sought to investigate how HF may have contributed to recent seismicity. We used times and locations of HF from the FracFocus catalog and enhanced the TexNet catalog of earthquakes from January 2019 to February 2020 by employing multi-station template matching. We detected over 1500 HF-induced earthquakes (80%), and of the seismicity in Live Oak county nearly 70% correlated with HF wells. We also sought to apply a repeating signal detector (RSD) that employs unsupervised machine learning to data from January 2015 to February 2020 in order to detect sequences of seismicity different from those in the template catalog. RSD identified over 900 new events revealing additional events during known HF-induced sequences as well as new bursts of seismicity. Most of these new bursts had earthquakes with a shorter SP time (~2 sec) than the previous catalog of earthquakes (> 4 sec) suggesting they occurred closer to the seismic station. We found there were HF stimulations close to the station at the times when those short S-P events occurred. The short S-P events have smaller magnitudes, consistent with the idea that HF-induced seismicity in the Eagle Ford is likely more pervasive than previously reported, but detection is limited by the density of seismic stations.