PHYSICAL MODEL STUDY OF REFLECTION PITFALLS AND ANISOTROPY PARAMETER ESTIMATION IN A TRANSVERSELY ISOTROPIC MEDIUM

Imaging problems due to velocity anisotropy were investigated using forward modelling. Zero-offset and constant-offset reflection experiments were performed to image spherical domes that were carved into the bottom of VTI and HTI models, respectively, along three different layout directions. Subject to velocity anisotropy, the spherical dome was projected as laterally stretched images. The curvature radius of the spherical dome was increasingly imaged in the seismic profiles due to polar velocity variation in the VTI model. An ellipsoidal dome was detected in the HTI model, instead of a spheroidal dome, due to azimuthal velocity variation. To address the problems of erroneous images in reflections, laboratory data were processed by anisotropic 2D Kirchhoff migration. Although true images were not satisfactorily restored, the effects of "lateral stretch" were significantly reduced. Due to azimuthal velocity variations, Common Depth Point (CDP) reflections acquired from different azimuths on the horizontal symmetry axis plane of the HTI model cannot be well-tied. Making use of the "mis-tie" in the CDP, which was observed in the profiles, the feasibility of estimating the Thomsen's P-wave anisotropy parameter (epsilon) was evaluated. Although the physical models used are relatively simple, erroneous images caused by velocity anisotropy were confirmed, and the prospect of estimating the P-wave anisotropy parameter (epsilon) from arrival time difference was addressed.