Elliptical dip moveout (EMO) for 3D seismic imaging in the presence of azimuthal anisotropy

Abstract

Seismic reflection data exhibiting the effects of azimuthal velocity variation can be generated by waves propagating through geologic layers containing intrinsic horizontal transverse isotropic (HTI) velocity anisotropy, or through isotropic layers deformed into complex 3D velocity structures, or through both. Most current approaches to processing such data attempt to compensate separately for one physical mechanism or the other, but not both simultaneously. Many azimuthal anisotropy (AA) processing flows initially assume zero structural dip to calculate an azimuthal residual moveout (RMO) correction that flattens CMP gathers after standard isotropic normal moveout. Other approaches initially assume no intrinsic HTI velocity anisotropy in order to generate isotropic migration image gathers, which subsequently are flattened through an azimuthal RMO operator. Both approaches address intrinsic anisotropy and complex 3D structure separately, where in fact both effects may be coupled in the data. This leads to velocity estimation errors and degraded image quality, especially for large offsets, strong velocity anisotropy, and steep geologic dips. The most accurate solution to this problem is to perform iterative full anisotropic prestack depth migration velocity analysis, but this is impractical with current computational resources. For elliptical HTI media and complex 3D structure, we develop a computationally efficient anisotropic elliptical moveout (EMO) operator to precondition and regularize seismic wavefields by incorporating azimuthal velocity profile ellipticity. Forward and adjoint elliptical DMO operators are cascaded together to form a single EMO operation,which has a skewed saddle-like impulse response resembling that of the isotropic azimuthal moveout (AMO) operator, and can correct structural dip image errors commonly 10-20◦ or more for typical far offset and ellipticity values. EMO can be used to improve imaging results by data regularization to interpolate the azimuthally anisotropic seismic wavefield, and/or by applying data preconditioning to create an approximately equivalent isotropic data set.