Comparison of Facies Models Based on AVO Stochastic Inversion and AVO Deterministic Inversion

Summary Incorporating AVO inversion results into the modelling workflow is supposed to give a better delineation of the facies in comparison with modelling based just on well data. Therefore, a more precise estimation of the Gross Rock Volume is anticipated, which is one of the most important parameters for hydrocarbon volume estimation. Two methods of facies modelling are compared: facies simulation guided by deterministic AVO inversion and guided by stochastic AVO inversion. The workflows are applied to a dataset of the Oseberg Sor field in the Northern North Sea provided by Equinor. The study shows that thin layers below seismic resolution of facies models based on deterministic AVO inversion may lead to an over- or underestimation of the facies volumes. Stochastic inversion provides a better estimation of the facies fractions but cannot predict the position of the thin layers. Consequently, stochastic inversion is a tool for estimating the uncertainty of the reservoir facies and their P10, P50 and P90 sandstone volume values more reliably compared to Gauss Indicator Simulation guided by the facies probabilities derived from the deterministic AVO inversion.