No abstract

The transversely isotropic model with a horizontal axis of symmetry (HTI) has been used extensively in studies of shear-wave splitting to describe fractured formations with a single system of parallel vertical penny-shaped cracks. Here, we present an analytic description of longspread reflection moveout in horizontally layered HTI media with arbitrary strength of anisotropy. The hyperbolic moveout equation parameterized by the exact normal-moveout (NMO) velocity is sufficiently accurate for P-waves on conventional-length spreads (close to the reflector depth), although the NMO velocity is not, in general, usable for converting time to depth. However, the influence of anisotropy leads to the deviation of the moveout curve from a hyperbola with increasing spread length, even in a single-layer model. To account for nonhyperbolic moveout, we have derived an exact expression for the azimuthally dependent quartic term of the Taylor series traveltime expansion [t 2 (x 2)] valid for any pure mode in an HTI layer. The quartic moveout coefficient and the NMO velocity are then substituted into the nonhyperbolic moveout equation of Tsvankin and Thomsen, originally designed for vertical transverse isotropy (VTI). Numerical examples for media with both moderate and uncommonly strong non-hyperbolic moveout show that this equation accurately describes azimuthally dependent P-wave reflection traveltimes in an HTI layer, even for spread lengths twice as large as the reflector depth. In multilayered HTI media, the NMO velocity and the quartic moveout coefficient reflect the influence of layering as well as azimuthal anisotropy. We show that the conventional Dix equation for NMO velocity remains entirely valid for any azimuth in HTI media if the group-velocity vectors (rays) for data in a commonmidpoint (CMP) gather do not deviate from the vertical incidence plane. Although this condition is not exactly satisfied in the presence of azimuthal velocity variations, rms averaging of the interval NMO velocities represents a good approximation for models with moderate azimuthal anisotropy. Furthermore, the quartic moveout coefficient for multilayered HTI media can also be calculated with acceptable accuracy using the known averaging equations for vertical transverse isotropy. This allows us to extend the nonhyperbolic moveout equation to horizontally stratified media composed of any combination of isotropic, VTI, and HTI layers. In addition to providing analytic insight into the behavior of nonhyperbolic moveout, these results can be used in modeling and inversion of reflection traveltimes in azimuthally anisotropic media.

No abstract

Horizontal transverse isotropy (HTI) is the simplest azimuthally anisotropic model used to describe vertical fracturing in an isotropic matrix. Assuming that the subsurface is laterally homogeneous, and using the elliptical variation of P-wave NMO velocity with azimuth measured in at least three different source‐to‐receiver orientations, we can estimate three key parameters of HTI media: the vertical velocity, anisotropy, and the azimuth of the symmetry axis. Such parameter estimation is sensitive to the angular separation between the survey lines in 2-D acquisition or, equivalently, to source‐to‐receiver azimuths in 3-D acquisition and the set of azimuths used in the inversion procedure. The accuracy in estimating the azimuth, in particular, is also sensitive to the strength of anisotropy, while the accuracy in resolving vertical velocity and anisotropy is about the same for any strength of anisotropy. To maximize the accuracy and stability in parameter estimation, it is best to have the azimuths for the source‐to‐receiver directions 60° apart when only three directions are used. This requirement is feasible in land seismic data acquisition where wide azimuthal coverage can be designed. In marine streamer acquisition, however, the azimuthal data coverage is limited. Multiple survey directions are necessary to achieve such wide azimuthal coverage in streamer surveys. To perform the inversion using three azimuth directions, 60° apart, an HTI layer overlain by an azimuthally isotropic overburden should have a time thickness, relative to the total time, of at least the ratio of the error in the NMO (stacking) velocity to the interval anisotropy strength of the HTI layer. Having more than three source‐to‐receiver azimuths (e.g., full azimuthal coverage), however, provides a useful data redundancy that enhances the quality of the estimates, thus allowing acceptable parameter estimation at smaller relative thicknesses.

No abstract