As anisotropic depth imaging is widely used in the industry, anisotropic velocity models are routinely being constructed and applied. However, the degree of accuracy of the anisotropic velocity model needed to correctly position the horizons is not well defined. To that end, we tested four anisotropic parameter estimation scenarios and performed depth imaging using a realistic synthetic sedimentary seismic data that included an irregularly shaped salt body. Although similar images derived from the data using those four different anisotropic input models were obtained, a quantitative analysis indicates that systematic differences do exist. They enable us to relate the degree of accuracy needed in the image to the effort needed to build the anisotropic velocity model.
In seismic exploration, common-image-point (CIP) tomography has become a standard tool for seismic depth imaging. CIP tomography continues to improve its capability including dealing with anisotropic VTI and TTI models as exploration demands increase (Woodward et. al, 2008). However, while determining an anisotropic velocity model in depth is essential to correctly position reflectors in depth and improve the clarity of the imaged reflectors, anisotropic model estimation becomes difficult in areas where complex geological structure may cause lateral anisotropic heterogeneity. Recently, Sengupta et al. (2009) showed how to generate anisotropic models in the presence of salt bodies using geomechanical principles. Geomechanical modeling indicated that significant saltinduced velocity anisotropy variations are expected in the sediments surrounding the salt body. In this paper, we explore the sensitivity of depth imaging to various anisotropic ( and ) models. We use a realistic 2.5D synthetic anisotropic velocity model where saltinduced lateral anisotropy changes are present. The model was generated using geomechanical methods, following the same technique presented by Sengupta et al. (2009). We simulated a seismic survey using elastic anisotropic finitedifference modeling, and imaged the data using the following anisotropy models that were constructed based on the following realistic scenarios; a 1D anisotropic model that is built based on a single profile (assumed to be known from checkshot data), a 2D anisotropic model, an interpolated profile derived from two wells at two different locations, and a anisotropic model generated by using inaccurate third-order elasticity (TOE) parameters. For the last model, the scenario assumed that we can simulate the stresses using geomechanics, but there are errors in mapping stress changes to velocity changes. As our goal is to investigate the role of lateral heterogeneity in the anisotropy and to avoid subsalt imaging issues, we focus our attention to the sediments above the salt.
Sengupta et al. (2009) presented an application of 3D geomechanical modeling for anisotropic parameter estimation. A synthetic model with an irregularly shaped salt body was created based on a real 3D seismic image from a survey in the Gulf of Mexico. Sengupta et al. showed that the presence of irregularly shaped salt gives rise to large deviatoric stresses that can generate significant anisotropic velocity anomalies in the sediments surrounding the salt. The 3D stress and the velocity anisotropy are related.
