Two-way traveltime analysis for seismic reservoir characterization

Two-way traveltime (TWT) is one of the most important seismic attributes for reservoir characterization. Erroneous analysis of TWT can lead to incorrect estimates of velocity models resulting in improper structural interpretation of the subsurface. TWT analysis starts with the most fundamental step of seismic data processing, namely, Normal Moveout (NMO) correction. NMO correction is generally performed in the offset-time (X-t) domain, by fitting a hyperbolic curve to the observed traveltime corresponding to each reflection event. The performance of NMO correction depends on the quality of the data in the prestack domain and the underlying geology.;When ideal data sets are available (high signal to noise ratio), and underlying geology is simple (flat layers), the NMO correction can still be erroneous due to (1) its long offset non-hyperbolic behavior, and (2) due to the presence of seismic anisotropy. Even though in the X-t domain several equations have been developed to account for seismic anisotropy induced non-hyperbolic move out, they are prone to error, when multiple anisotropic and isotropic layers are present. The non-hyperbolic equations for moveout corrections are also approximate as they are some form of truncated Taylor series and can only estimate effective root mean square (rms) parameters for each reflection event.;In the plane wave (tau-p) domain, the estimation of layer parameters can be done using an exact equation for delay-time free from the approximation errors present in the X-t domain. In this domain a layer striping approach can also be used to account for the presence of multiple anisotropic and isotropic layers. Thus it is lucrative to develop NMO correction equation in the tau-p domain for an anisotropic medium, which in its limiting case can be useful for the isotropic medium as well.;The simplest anisotropic media are Transversely Isotropic (TI) media which are also common in exploration seismology. One of the TI media, with a vertical axis of symmetry (VTI) shows non-azimuthal anisotropy whereas another TI medium, with a horizontal axis of symmetry (HTI) shows azimuthal anisotropy. Since quantifying VTI is easy (as it does not have azimuthal effect) most of the initial researches have focussed on VTI media. Satisfactory NMO correction equations are also available for the VTI media in the tau-p domain. In this thesis I have developed a tau-p domain NMO correction equation for the HTI media, which in its limiting cases can be used for VTI and isotropic media.;A more complicated anisotropic medium than TI media is an orthorhombic medium. Orthorhombic media are common when two sets orthogonal vertical fractures sets are observed in the subsurface. Estimation of model parameters in such media from traveltime analysis is more complicated and no work in the tau-p domain for such media has been reported. To fill this gap, this thesis reports on the development of a tau-p domain NMO correction equation for the orthorhombic medium as well. In the limiting case this equation takes the form of a pervious tau-p domain TI media NMO correction equation.;For the purpose of subsurface parameter estimation based on traveltime, it is important to determine the presence of anisotropy in the data set. For that I developed a critical angle reflectometry scheme in the tau-p domain. This method analyzes critical slowness at each azimuth and relates it with the anisotropy parameters. Satisfactory critical slowness analysis for detecting anisotropy is presented in this thesis.;Even though most of the rocks are anisotropic 'until proved otherwise', data processing schemes commonly employed in the industry are based on isotropic assumption of the earth model. This can lead to large residuals in the NMO corrected traveltime. This residual traveltime can be used for anisotropic parameter estimation. In this thesis I present a case study from the Gulf of Mexico, where residual traveltime from seismic data and anisotropy parameter information from well logs are used with the help of geostatistics to generate fracture maps showing spatial variations of fracture orientation.;Finally I present a case study from the Gulf of Mexico, where estimation of structural uncertainty is made by combining traveltimes from seismic data, and well log derived marker depths in the time domain. I used a Markov-Bayes stochastic simulation to estimate structural uncertainty. Occurrence of significant uncertainty in several parts of the subsurface model may be attributed to using erroneous velocity model, presence of faults and fractures in the subsurface, and ignoring seismic anisotropy during data processing.