Direct reservoir property estimation based on prestack seismic inversion

Reservoir property estimation is an essential part for reservoir characterization. Most commonly-used estimation methods are implemented in two steps, seismic inversion and rock physics inversion. However, these indirect methods may increase the uncertainty and reduce the accuracy of estimation results. In this work, we propose a Bayesian inversion approach to estimate reservoir properties directly from prestack seismic data. Firstly, by combining the reflection coefficient equation and rock physics model, we derive a P-wave reflection approximation in terms of reservoir parameters, which establishes a direct link between seismic data and reservoir properties. Model examples illustrate the accuracy of the approximation comparing to the exact reflection coefficient equation, which satisfies the requirements of the prestack seismic inversion. Then in the framework of Bayesian inversion theory, a novel inversion method is presented to estimate porosity, mineral volume and water saturation directly from prestack seismic angle gathers. Direct estimation increases the stability and decreases the uncertainty. The synthetic test demonstrates the advantage of the proposed method on the accuracy and stability over indirect methods. The real data example verifies the feasibility of the proposed method in direct reservoir property estimation.