Building geologically plausible anisotropic models using well data and horizon-guided interpolation

Abstract

Depth imaging with anisotropic velocity models has been shown to deliver more accurate images than traditional data processing methods. Accounting for anisotropy is particularly important for complex geologies but should ideally be included in all imaging projects as the Earth is inherently anisotropic. While imaging in complex settings may require tilted transversely isotropic (TTI) models, anisotropic depth imaging with vertical transversely isotropic (VTI) models has become the dominant practice in the seismic industry. A VTI model velocity field requires three parameters: symmetry-axis (vertical) velocity (VP0 ) and Thomsen parameters e and δ. The challenge of building such models is that it is not feasible to rely on seismic tomography to derive multiple parameters in an entire 3D volume because inversion of seismic data alone for all three parameters is highly non-unique (Tsvankin, 2001). The current industry practice typically involves deriving a single smooth profile of Thomsen parameters e and δ based on limited well control in areas of flat-layered geology. This profile is propagated throughout an entire volume by hanging it off the water bottom or another shallow horizon. Thomsen parameters are kept fixed while velocity along the symmetry axis is updated by tomographic inversion of reflection seismic data (Woodward et al., 2008). This process represents an improvement compared to the isotropic models used in the past, but suffers from several limitations, including: 1. The use of a single anisotropy profile disregards lateral variation of anisotropy in the subsurface. 2. Anisotropic volumes that follow water bottom topography do not represent the structure of subsurface geology. 3. The single anisotropy profile is likely to be overly smoothed in the vertical direction because fine details cannot be accurately propagated in a 3D volume with complex geology. The goal of the presented approach is a workflow that overcomes these limitations to build more geologically plausible