Joint Seismic and Petrophysical Inversion Constrained by Facies in a Bayesian Framework

Abstract

Summary We present a methodology for simultaneous seismic inversion and reservoir characterization, and apply it to a Norwegian Sea gas field of Jurassic age. Our Bayesian approach allows for joint probabilistic assessment of lithology/fluids, reservoir properties (porosity and saturation) and elastic attributes (acoustic impedance and Vp/Vs), and is demonstrated on a seismic section intersecting the Yttergryta gas discovery well. The data used in the inversion includes seismic AVO data (calibrated near, mid and far) and well log data (Vp, Vs, and density). For the rock physics likelihood estimates, empirical rock physics models extracted from observed well log data or first-order Taylor expansions of known rock-physics models (e.g. constant cement model) could be utilized in the inversion. We demonstrate the methodology on seismic AVO data from a Norwegian Sea gas field, and obtain a good match with the observed lithology/fluid transitions at a selected well location. Stratigraphic constraints (BCU horizon) allow for reliable inversion results away from the well location.