Parameter inversion for seismic anisotropy

An important objective of reflection seismology is to transform a seismic dataset into a reflectivity image of the subsurface. Reflectivity imaging involves removal of propagation effects by creating an inverse propagation operator using an estimated macro velocity model. The accuracy of a reflectivity image depends on the accuracy of the velocity model used. Hence, the estimation of a good velocity model is crucial. Flat sedimentary layers, though laterally homogeneous over small-scale, often exhibit a simple kind of anisotropy called vertical transverse isotropy (VTI). Negligence of anisotropy in migration may result in significant deviation of an imaged reflector from its true position. Our ability to estimate anisotropic parameters using limited aperture, limited bandwidth seismic data is restricted. Several recent developments in data acquisition including more intense acquisition (e.g., 3-D), broader bandwidths, and new survey methodologies (e.g., Ocean Bottom Seismics and vertical cables) have improved our ability in dealing with anisotropy. In light of these recent developments the objective of my research is three-fold - (i) finding better algorithms for anisotropy estimation that makes use of enhanced data acquisition, more powerful computers and graphical interfaces, (ii) understanding and re-evaluating our limitations/capabilities of anisotropy estimation in light of the new algorithms, and (iii) using the results of analysis as an aid to constrain anisotropy estimates. The Common Focus Point (CFP) technology defines a recent method of prestack migration velocity analysis which has several advantages over other popular methods. One remarkable advantage is that the imaging error is given directly in time allowing easier model updates through inversion. The CFP technique has been used in recent past for estimating isotropic heterogeneous velocity model for geologically complex subsurface. For my research, I extend this method to anisotropic parameter inversion by suitably implementing the space-time domain version of CFP method for picking the traveltime error resulting from the incorrect guess model. I successfully implement this approach to estimate anisotropy of the shale thrust-sheets in the Canadian Foothills model. This model has targets of interest below transversely isotropic shale thrust-sheets and estimating anisotropic parameters is important for proper depth imaging of target zone. The synthetic P-mode data used for the analysis has been generated using a finite difference algorithm. In order to quantitatively assess the advantages or limitations of CFP domain velocity analysis in estimating the anisotropic parameters, I have performed sensitivity analysis under different experimental settings like different observation apertures, layer thicknesses, tilt in symmetry axis, picking error and the use of mode converted data. The results quantitatively establish the advantage of joint inversion of P-P and P-Sv over the conventional inversion of solely P-P data in constraining the depth and anisotropic parameters. Also the constraint on parameter estimation improves with increase in angle of tilt of symmetry axis with respect to the reflector. The CFP domain analysis also has advantages in dealing with mode converted P-Sv data as separate focusing at the receiver side and source side allows it to deal with one wave type at a time. In complex media, this helps avoid the cumbersome calculation of common conversion point (CCP) location. The CFP analysis has been applied for estimating anisotropic parameters as well as layer thickness by joint inversion of P-P and P-Sv synthetic data for the geology encountered in Blackfoot field.