We present an improved method for modeling 3D acoustic wavefields scattered at smooth curved interfaces. The approach is based on a high-frequency approximation of surface integral propagators and a correct description of their boundary values in terms of transmission operators. The main improvement is a uniform local approximation of these operators in the form of effective reflection and transmission coefficients. We show that the effective coefficients represent a generalization of the plane-wave coefficients widely used in conventional seismic modeling, even for the case of curved reflectors, nonplanar wavefronts, and finite frequencies. The proposed method is capable of producing complex wave phenomenas, such as caustics, edge diffractions, and head waves. Seismograms modeled for even simple models reveal significant errors implicit in the plane-wave approximation. Comparison of modeling based on effective coefficients with the analytic solution reveals errors less than 4% in peak amplitude at seismic frequencies.

Plane-wave reflection coefficients (PWRCs) are routinely used in amplitude-variation-with-offset analysis and for generating boundary data in Kirchhoff modeling. However, the geometrical-seismics approximation based on PWRCs becomes inadequate in describing reflected wavefields at near- and postcritical incidence angles. Also, PWRCs are derived for plane interfaces and break down in the presence of significant reflector curvature. Here, we discuss effective reflection coefficients (ERCs) designed to overcome the limitations of PWRCs for multicomponent data from heterogeneous anisotropic media. We represent the reflected wavefield in the immediate vicinity of a curved interface by a generalized plane-wave decomposition, which approximately reduces to the conventional Weyl-type integral computed for an apparent source location. The ERC then is obtainedas the ratio of the reflected and incident wavefields at each point of the interface. To conduct diffraction modeling, we combine ERCs with the tip-wave superposition method (TWSM), extended to elastic media. This methodology is implemented for curved interfaces that separate an isotropic incidence half-space and a transversely isotropic (TI) medium with the symmetry axis orthogonal to the reflector. If the interface is plane, ERCs generally are close to the exact solution, sensitive to the anisotropy parameters and source-receiver geometry. Numerical tests demonstrate that the difference between ERCs and PWRCs for typical TI models can be significant, especially at low frequencies and in the postcritical domain. For curved interfaces, ERCs provide a practical approximate tool to compute the reflected wavefield. We analyze the dependence of ERCs on reflector shape and demonstrate their advantages over PWRCs in 3D diffraction modeling of PP and PS reflection data.

A conventional amplitude variation with offset (AVO) inversion is based on geometrical seismics which exploit plane-wave reflection coefficients to describe the reflection phenomenon. Widely exploited linearizations of plane-wave coefficients are mostly valid at pre-critical offsets for media with almost flat and weak-contrast interfaces. Existing linearizations do not account for the seismic frequency range by ignoring the frequency content of the wavelet, which is a strong assumption. Plane-wave reflection coefficients do not fully describe the reflection of seismic waves at near-critical and post-critical offsets, because reflected seismic waves are typically generated by point sources. We propose an improved approach to AVO inversion, which is based on effective reflection coefficients (ERCs). ERCs generalize plane-wave coefficients for seismic waves generated by point sources and therefore more accurately describe near-critical and post-critical reflections where head waves are generated. Moreover, they are frequency-dependent and incorporate the local curvatures of the wavefront and the reflecting interface. In our study, we neglect the effect of interface curvature and demonstrate the advantages of our approach on synthetic data for a simple model with a plane interface separating two isotropic half-spaces. A comparison of the inversion results obtained with our approach and the results from an AVO inversion method based on the exact plane-wave reflection coefficient suggests that our method is superior, in particular for long-offset ranges which extend to and beyond the critical angle. We thus propose that long offsets can be successfully exploited in an AVO inversion under the correct assumption about the reflection coefficient. Such long-offset AVO inversion shows the potential of outperforming a conventional moderate-offset AVO inversion in the accuracy of estimated model parameters.