Look‐ahead walkaways using effective VTI models

Abstract

Summary Determining the long wavelength velocity trend ahead of an intermediate VSP receiver array is important for pore pressure prediction and walkaway imaging. A coherency optimization – based inversion of reflection moveout is presented to determine velocities in an anisotropic 1D model below a receiver array where multi-offset VSP (walkaway) data have been acquired. The forward modeling engine for the inversion is effective VTI (vertical transverse isotropy) anisotropic modeling, which replaces a stack of VTI layers with a single VTI layer. When used in conjunction with overburden calibration from walkaway direct arrival times, it is accurate out to very long offsets. A novel approach to handle Sv propagation is included. This approximate forward modeling approach is fast, allowing an efficient solution of the multi-parameter inverse problem. The long wavelength component of the velocity function below the receivers is parameterized as the sum of a gradient plus harmonic terms. Walkaway reflection data are then inverted using a downhill simplex algorithm, and since both P-p and P-s reflections can be handled, both Vp and Vs long wavelength trends can be estimated. Performance of the inversion is shown on anisotropic synthetic data, where estimates of both Vp and Vs are obtained below a receiver array.