Capturing high-resolution lithological heterogeneity in geological CO2 storage reservoir models using stochastic seismic inversion

Lithological heterogeneity at cm- to dm-scale in siliciclastic reservoirs dominantly controls CO2 migration and trapping. Such heterogeneity can exist in form of sedimentary structures such as cross and planar bedding and cemented layers and is typically characterised in core descriptions and in core logs. However, such variability in the inter-well region is often unknown as the scale of heterogeneity is below the resolution of geophysical data typically used for building static geological models. In this study, we downscale high-resolution seismic inversion data by integrating it with wireline logs using the stochastic seismic inversion approach. This allows predicting plausible distributions of dm-scale rock properties in the inter-well region. The approach has advantages over the conventional deterministic seismic inversion which is limited by data downscaling along with providing the opportunity to generate equiprobable realizations of heterogeneity distribution. High resolution supervised seismic facies and probabilities from seismic stochastic inversion and seismic porosity (after depth conversion) were used as trends for facies modelling using Sequential Indicator Simulation (SIS). The methodology is applied to develop a high-resolution facies model of Parasequence-2 of the Paaratte Formation, Otway Basin, which is a target for R&D CO2 injection experiment.