MR026-04
Time-dependent variations in velocity structure and seismogenic fault orientations in southern Kansas
Wednesday, 16 December 2020: 11:45
Virtual
Kayla Kroll1, Eric Matzel2, Christina Morency3, Elizabeth S Cochran4, Joshua A. White1 and Laura Chiaramonte5, (1)Lawrence Livermore National Laboratory, Livermore, CA, United States, (2)LLNL, Livermore, CA, United States, (3)Lawrence Livermore National Laboratory, Atmospheric, Earth and Energy Division, Livermore, CA, United States, (4)U.S. Geological Survey, Earthquake Science Center, Pasadena, CA, United States, (5)Electric Power Research Institute Palo Alto, Palo Alto, CA, United States
Abstract:
Seismicity rates in southern Kansas have increased dramatically since 2013 in conjunction with increased disposal of wastewater in Kansas and neighboring Oklahoma. Fluid injection increases pore-fluid pressures, which leads to reactivation of basement faults and may be associated with time-dependent changes in the seismic velocity structure. Here, we apply ambient noise tomography methods to dozens of stations across a 100 km
2 region of southern Kansas to generate a 3D image of the upper crustal seismic structure and to search for measurable changes in the subsurface that can be linked to injection. Small amplitude and arrival time variations of ambient noise Green functions indicate changes in the media, perhaps indicative of subsurface pressure changes. Measured amplitudes are particularly sensitive to changes related to the presence of fluids in the Earth.
Next, we employ virtual seismometer (VS) techniques to a matched filter seismic catalog (Cochran et al., 2018) to estimate spatiotemporal changes in the style of faulting. VS techniques are highly sensitive to differences in fault slip between closely spaced events. For an event pair, the relative moment tensor is defined with respect to the full moment tensor of a target event. Orientations of reactivated faults are compared to fault orientations for regional, historical events that occurred prior to the onset of significant fluid disposal. Results will further our understanding of the stress and pressure conditions responsible for inducing earthquakes in this region. For example, if faults are being reactivated solely by fluid pressure increases, we predict an increase in the range of fault orientations activated with time during injection. Conversely, if faults are reactivated due to a rotation in the stress field related to increased fluid pressures or occurrence of moderate-sized events, we expect to see a significant rotation in the mean seismogenic fault orientation. Prepared by LLNL under Contract DE-AC52-07NA27344.
