Predicting Lithology And Porosity Of Carbonate Reservoirs Using Geostatistical Stochastic Pre-stack Inversion For Geomodeling Constraint

This paper describes a methodology to characterize lithology variations impacting the porosity distribution in a carbonate reservoir. We performed a stochastic seismic inversion to obtain multiple realizations of elastic properties discriminating between the target lithologies. These multi-realizations were then used in a supervised lithoseismic classification for obtaining cubes of occurrence probability of lithologies. Finally, a pseudo porosity cube is obtained using a linear combination between the acoustic impedance and the lithology probabilities. The entire workflow is performed in a stratigraphic grid layering-consistent with geomodeling grid. Such capability is important for further use of the results as 3D constraint during the geomodel infilling. Lithology probability cubes show proportions in agreement with vertical facies proportions observed from well data.