Seismic Lithofacies Classification over Turbidite Reservoir Offshore Brazil

An integrated workflow, including careful seismic data preconditioning, simultaneous elastic inversion and seismic lithology classification has been applied to predict the distribution of five lithofacies in a complex turbiditic reservoir offshore Brazil. Four partial angle stacks (2°-40°) with 6 key wells, have been used to derive elastic properties through inversion. A wellbased feasibility study shows that reservoir facies can be identified from a Vp/Vs ratio versus P-wave impedance (Ip) cross-plot. The main challenge for this seismic lithology discrimination was the complexity of the depositional system, which was tackled by incorporating a 3-D facies trend model in the seismic lithology prediction workflow. By applying a supervised Bayesian classification technique to our simultaneous elastic inversion results, litho-probability cubes were calculated for a total of 5 reservoir and non-reservoir facies. The inversion was performed in a layered stratigraphic framework constructed from interpreted horizons. This has yielded 3-D images of reservoir elastic properties and facies that better conform to the complex shapes of the channelized deposits.