HYDROLOGIC PROPERTIES OF THE DEEP BASEMENT FROM PRESSURE HISTORY MATCHING AND IMPLICATIONS FOR INDUCED SEISMICITY IN THE US MIDCONTINENT

The rapid increase in injection of wastewater, coproduced from unconventional oil and gas production processes, has caused seismicity rates to surge in the US midcontinent over the last decade. Although the target disposal zone in this region is the Cambrian-Ordovician Arbuckle Group (and age equivalent reservoirs), the seismicity has mainly occurred in the underlying Precambrian crystalline basement, many kilometers below the injection interval. The earthquakes have revealed complex fault networks in the subsurface, however, the hydrologic properties of the basement remain poorly understood, due mostly to a lack of data. Constraints on these parameters are critical for modeling fluid flow and pressures across the entire Arbuckle-basement system, and are necessary to accurately evaluate and predict injection-induced earthquakes. This study uses newly compiled pressure data from Class I wells in Kansas to constrain the reservoir-scale properties of the basement such as permeability, diffusivity, and specific storage. To do this, we construct a detailed, three-dimensional geological and pressure history-matched numerical model for the Arbuckle and basement. The model, based on data from more than 400 wells in south-central Kansas, is used to simulate injection data from 319 wastewater disposal wells indicate that Arbuckle. The results suggest that pressures in the Arbuckle have increased to 1.1 MPa in high injection rate areas and overpressures of <0.1 MPa are the cause of basement seismicity. Pressure-history matching also yields the likely range in porosity (0.3-7%), permeability (0.1-0.7 md), and diffusivity (0.004-0.07 m²/s) for the basement. The results indicate that the reservoir-scale properties of the basement are enhanced by faults and fractures. Importantly, the diffusivities determined in this study are lower than estimates derived from Kansas earthquake triggering fronts, and suggest that such seismicity-based techniques may have limitations, particularly where space-time patterns between injection and seismicity are complex.