Estimating pore fluid pressures during the Youngstown, Ohio earthquakes

Abstract:

Several months after fluid injection began in December 2010 at the Northstar 1 well in Youngstown, Ohio, low-magnitude earthquakes were detected in the Youngstown area, where no prior earthquakes had been detected. Concerns that the injection might have triggered the earthquakes lead to shutdown of the well in December 2011. Earthquake relocation analysis by Kim (2013, J. Geophy. Res., v 118, p. 3506-3518) showed that, from March 2011 to January 2012, 12 earthquakes with moment magnitudes of 1.8 to 3.9 occurred at depths of 3.5 to 4 km in the Precambrian basement along a previously unmapped vertical fault. The 2.8 km deep Northstar 1 well, which penetrated the top 60 m of the basement, appeared to have been drilled into the same fault. The earthquakes occurred at lateral distances of 0 to 1 km from the well. The present study aims to estimate the fluid pressure increase due to injection. The groundwater flow model MODFLOW is used to simulate fluid pressure propagation from the well injection interval into the basement fault and two permeable sandstone layers above the basement. The basement rock away from the fault is assumed impermeable. Reservoir properties (permeability and compressibility) of the fault and sandstone layers are estimated by calibrating the model to match injection history and wellhead pressure recorded daily during the operational period. Although the available data are not sufficient to uniquely determine reservoir properties, it is possible to determine reasonable ranges. Simulated fluid pressure increases at the locations and times of the earthquakes range from less than 0.01 MPa to about 1 MPa. Pressure measurements in the well after shut-in might enhance the estimation of reservoir properties. Such data could also improve the estimation of pore fluid pressure increase due to injection.