Application of the WesternGeco uncertainty workflow to a TTI 3-D synthetic model produced by BP revealed new insights on the scale dependency of the anisotropic nullspace and its impact on structural uncertainty. This analysis was done by using tilted transverse isotropy (TTI) offset ray tracing through the model for an ocean-bottom station (OBS) mirror acquisition geometry: no synthetic seismic data were involved.
Recent advances in imaging technology, as well as new full-azimuth and long-offset acquisition techniques, have resulted in greatly enhanced images in even the most complex structural settings. Highly accurate migration algorithms such as reverse-time migration (RTM) require Earth velocity models with ever-greater resolution. Incorporation of more non-seismic data and knowledge into the model building process is required to build these models. However, even with our best efforts to combine all available data, there is still ambiguity in our models, caused by the inherent non-uniqueness of the seismic experiment. Many different Earth velocity models can exist that explain the observed seismic data, and this ambiguity grows rapidly away from known control points such as well logs. The result is uncertainty of the true positions of events in our images. These uncertainties can lead to exploration risk (e.g., trap failure), drilling risk (e.g., dry wells), and volumetric uncertainties. While the underlying ambiguity can never be fully eradicated, a quantified measure of these uncertainties provides deeper understanding of the risks and related mitigation plans. Osypov et al. (2008a, 2008b, 2010) introduced an uncertainty analysis method that generates sample models from the posterior distribution by using eigendecomposition of the anisotropic tomography operator. A realistic synthetic model is an important tool to validate the uncertainty method because the ground truth is known and various “what if” scenarios can be explored. This study uses the 3-D TTI synthetic model of BP (Nolte et al, 2010) to gain insights about anisotropic parameter estimation ambiguity and the scale dependency of uncertainty. We studied a set of scenarios, for instance, one that corresponds to a hypothetical situation when tomography and uncertainty analysis are performed under an erroneous assumption of vertical transverse isotropy (VTI). In this paper, we present a scenario in which the TTI angle of symmetry is fixed and known, but the starting velocity field is a smoothed version of the true field.
Let us recap the uncertainty analysis method of Osypov et al. (2008a, 2008b, 2010). The uncertainty workflow is applied after tomography for a given scale length has converged and driven gather flatness to an acceptable level. It consists of 4 steps. In the first step, a geologically reasonable prior distribution for the anisotropic parameters is defined: i.e. plausible geologic concepts are considered in terms of shapes and patterns of the Earth’s anisotropic behavior, and also allowable ranges of velocity, e, and d perturbations are obtained from rock physics analysis. In the second step, multiple, equally likely sample models are pulled from the posterior distribution. Let us explain this process in more detail.
