Stochastic Seismic Inversion to Estimate the Pore Volume for CO2 Injection in the St. Flavien Reservoir, Quebec, Canada

The estimation of the volume of pores available is of prime importance when evaluating the potential of a site for the geological storage of CO2. Sequential Bayesian simulations, using well logs and acoustic impedance cube are completed to model the porosity field and its variability in the Beekmantown Group reservoir of the Saint-Flavien region, Québec, Canada. We use stochastic seismic inversion to better recover the whole bandwidth of acoustic impedance distribution observed at well logs, needed to reproduce the distribution of porosity observed on well logs. This methodology is of prime importance in the Saint-Flavien context, as the average porosity is extremely low, and only a fraction of the reservoir, identified as Family 3 by Gaussian mixture models, exhibits higher porosity where CO2 injection would be possible. Comparison of the porosity fields simulated using the deterministic acoustic impedance cube and one realisation of a stochastic inverted acoustic impedance cube clearly show that zones of higher porosity are better reproduced by the latter option. The use of stochastic seismic inversion is needed to obtain a more realistic estimate of available pores for CO2 injection in the Beekmantown Group reservoir in the Saint-Flavien region.