MODELING INJECTION INDUCED STRESS CHANGES IN THE FORT WORTH BASIN, TEXAS

Ongoing seismicity within the crystalline basement in the Fort Worth Basin (FWB) has been causally linked to the injection of wastewater into the thick, directly overlying, dolomitic limestone Ellenburger formation. The majority of basement earthquakes are located within 10 km of at least one fluid injection well, but some active sequences are not near wells. Spatiotemporal correlation between earthquake and injection activities has led to the conclusion that pore fluid pressure and/or poroelastic stress changes induce seismicity on pre-existing, well-oriented normal faults, even at 10s of km distance. Here, we create a 3D coupled geomechanical model to examine injection related time-dependent stress changes along faults within the FWB. Prior analyses of injection related stress changes observed within the basin lacked high-resolution geologic and fault geometry data. In this model we incorporate: updated 3D geometries and parameters of the main geologic formations, fault geometries created using earthquake and 3D seismic reflection data, and fluid injection data collected by the Texas Railroad Commission. The model focuses on the propagation of stress changes associated with basin-wide injection activities. Over 200 injection wells were active in the FWB between 2004 and 2019 that collectively injected ~357 million m³ of fluids into the Ellenburger formation.

Fall-off testing results in 2015 indicated that the Ellenburger formation was overpressured by 0.9-4.8 MPa in the NE Johnson County area. Initial modeling of the NE FWB region indicates that if the primary seismogenic Venus fault is hydraulically conductive, then injection related stress changes on the order of 0.5-2.5 MPa would be reasonable. The 3D modeling incorporates both currently seismically active and inactive faults to explore compartmentalization within the system in greater detail. Altogether, the modeling results to date confirm that injection related stress changes have been large enough to induce fault slip along the certain faults, consistent with other findings based on space and time relationships and on fault slip potential calculations.