Effect of the interface curvature on the reflections for long offset data

Analysis of seismic reflection amplitudes versus offset (AVO) is one of common techniques widely exploited in the industry for reservoir characterization. For the last two decades a lot of approaches to analysis, inversion and interpretation of AVO data have been developed. Existing modifications valid for weak-contrast interfaces were successfully employed for conventional reservoirs. The growing interest of the industry to unconventional reservoirs, such as stiff-carbonate reservoirs, heavy oil traps and reservoirs close to salts domes - associated with strong-contrast interfaces and critical angles - implies the development of AVO techniques valid prior and beyond the critical angle. It has been reported in literature that near- and post-critical reflections have a potential to be employed as an additional source of information about the media. However, the use of these reflections is limited by the inability of well-known Zoeppritz equations to explain phenomena observed around and beyond the critical angle. The aim of the thesis is to investigate phenomena observed at the reflected data around and beyond the critical angle, understand their potential from the AVO analysis and inversion point of view and develop a long-offset AVO inversion approach valid for strong-contrast interfaces. The theory of effective reflection coefficients is exploited as a mathematical apparatus providing an adequate description of phenomena observed at near- and post-critical reflections. The thesis consists of five papers, where four major issues are addressed. The sensitivity of the reflection coefficient to isotropic and HTI media parameter changes prior to and beyond the critical angle is studied. The long-offset AVO inversion approach valid prior to and beyond the critical angle, strong-contrast and curved interfaces is developed and tested on synthetic data obtained for models with a single interface of various curvatures. Frequency effects in pre- near- and post-critical domains observed on the data of physical modeling are studied from the point of view of potential exploiting. Finally, the sensitivity of long-offset AVO inversion to errors related to overburden velocity misinterpretation is analyzed.