Impact of velocity model calibration on microseismic locations

Abstract

Accurate locations of the microseismic events form the basis for interpretation of hydraulically stimulated regions. The velocity model, which in general is not well known, has a significant impact on the accuracy of located events. Correct knowledge of the anisotropy parameters and lithological heterogeneity, accompanied with a precise model calibration will reduce errors in microseismic locations. In this paper, we investigate the velocity model calibration process for multi-stage frac monitoring, in particular the use of single calibration shot. We generate synthetic microseismic events using elastic finite-difference modeling, assuming an initial isotropic velocity model based on the Barnett shale. We calibrate the initial model to best-fit arrival times from the shot(s) at known location(s). The predicted arrival times are based on a minimum traveltime tree algorithm. A similar approach is repeated for the initial model containing shale layer with vertical transverse isotropy. We then evaluate the behavior of errors in microseismic locations obtained with each calibrated velocity model using grid-search algorithm. As expected, the use of individual calibration shots results in significant microseismic location errors. Calibrating model with multiple known shots together improves the accuracy of microseismic locations. For the anisotropic case, the resulting errors are larger than in the isotropic case. It is therefore, important to calibrate velocity model with maximum number of available perforation shots, in particular for areas with strong anisotropy and lateral heterogeneity.