Seismic imaging using microearthquakes induced by hydraulic fracturing

Abstract

Seismic imaging using microearthquakes induced by hydraulic fracturing produces a three-dimensional (3-D), S-wave velocity model of the fractured zone, improves the calculated locations of the microearthquakes, and may lead to better estimates of fractureplane orientations, fracture density, and water flow paths. Such information is important for predicting the amount of heat energy that may be extracted from geothermal reservoir. A fractured zone was created at the Los Alamos Hot Dry Rock Reservoir in north-central New Mexico within otherwise impermeable basement rock by injecting [Formula: see text] of water into a borehole under high pressure at a depth of 3.5 km. Induced microearthquakes were observed using four borehole seismometers. The P-wave and S-wave arrival times have been inverted to find the 3-D velocity structures and the microearthquake locations and origin times. The inversion was implemented using the separation of parameters technique, and constraints were incorporated to require smooth velocity structures and to restrict the velocities within the fractured region to be less than or equal to the velocities of the unfractured basement rock. The rms amval time residuals decrease by 11–15 percent during the joint hypocenter-velocity inversion. The average change in the microearthquake locations is 20–27 m, depending on the smoothing parameter used. Tests with synthetic data imply that the absolute locations may improve by as much as 35 percent, while the relative locations may improve by 40 percent. The general S-wave velocity patterns are reliable, but the absolute velocity values are not uniquely determined. However, studies of inversions using various degrees of smoothing suggest that the S-wave velocities decrease by at least 13 percent in the most intensely fractured regions of the reservoir. The P-wave velocities are poorly constrained because the P-wave traveltime perturbations caused by the fluid-filled fractures are small compared to the amval time noise level. The significant difference in the relative signal-to-noise levels of the P-wave and S-wave arrival time data, coupled with the limited ray coverage, can produce a bias in the computed [Formula: see text] ratios, and corresponding systematic rotation of the microearthquake cluster. These adverse effects were greatly reduced by applying a [Formula: see text] lower bound based on the [Formula: see text] ratio of the unfractured basement rock.