GeoConvention 2013: Integration 1 Velocity model calibration effects on microseismic locations J. Akram*, University of Calgary akramj@ucalgary.ca and D. W. Eaton, University of Calgary Summary Accurate microseismic event locations are critical for microseismic monitoring, since they form the basis for interpretation of stimulated regions of an unconventional reservoir during a hydraulic fracture treatment. Calculated event locations depend on the velocity model, which in general is not well known. In this paper, we use finite-difference synthetic seismograms to assess the importance of seismic anisotropy (weak transverse isotropy, up to 20%, with a vertical symmetry axis) for microseismic event location uncertainty. Using a velocity model based on the Barnett Shale, we consider systematic position errors that arise when the velocity model is assumed to be isotropic. In particular, the starting model uses the correct vertical isotropic velocities and is adjusted to fit arrival times from a single calibration point (e.g., a perforation shot or string shot). As expected, we find systematic location errors. In areas where strong anisotropy is known to occur, it is important to incorporate anisotropy into the hypocentre location method. Introduction Accurate microseismic event locations are the basic requirements from a microseismic monitoring survey to interpret the stimulated regions of an unconventional reservoir during a hydraulic fracture treatment. The accuracy of microseismic event locations depend on the accuracy of arrival time picking and velocity model, the source-receiver geometry and the suitability of the location estimation technique (Ge and Kaiser, 1992; Maxwell, 2009). The velocity model, in general, is not well known and is calibrated locally using the arrival time picks from a single perforation shot for improved microseismic event locations. In this paper, we use finite-difference synthetic seismograms to evaluate the significance of vertical transverse isotropy (VTI) for microseismic event location uncertainty. We use a velocity model based on the Barnett Shale (Maxwell, 2009) and consider systematic position errors that arise when the velocity model is assumed to be isotropic. Method We generate synthetic seismogram using finite-difference modeling for the Barnet Shale model (Maxwell, 2009), which is used as initial isotropic model. Figure 1 shows the initial isotropic model with a low velocity shale layer surrounded by high velocity layers. Sources are distributed in the offset range (200 – 400m). Receivers are positioned as a vertical array at 0m offset and in a depth range (2377 – 2432m), with inter-receiver spacing of ~6m. We relocate all the sources using the exhaustive grid search algorithm to validate the microseismic event location algorithm accuracy.