Characterization of centimetre- to decimetre-scale lithological heterogeneity in reservoir models through stochastic seismic inversion and seismic facies probabilities

The lithological heterogeneities on centimetre to decimetre scale can significantly control the preferential migration of fluids such as CO<inf>2</inf>. However, it is challenging to model such thin layers in the area between the wells due to coarse seismic data resolution. This study employs stochastic seismic inversion to enhance seismic attribute resolution by combining them with wireline log data. This allows for representing centimetre (cm)-scale lithological heterogeneity in the inter-well regions along with the uncertainty in petrophysical property distribution as multiple equiprobable scenarios are generated in contrast to deterministic inversion. High-resolution seismic porosity generated from seismic inversion was subsequently used to derive maps of likely seismic facies distributions using supervised classification. By considering the seismic facies probabilities as trends for Sequential Indicator Simulation and including wireline log data, we developed detailed facies models for Parasequence-2 of the Paaratte Formation in the Otway Basin, as part of a field experiment on geological CO<inf>2</inf> storage. The facies models indicate that about 60 % of the reservoir section is comprised of proximal mouth bar facies while the carbonate-cements form about 15 % of the reservoir volume. The results obtained from the application of the approach were validated against well log and power spectrum data where a strong match between predicted and acquired seismic sections was observed from threshold value of −6 dB. The methodology significantly improves representation of fine layers with a thickness of a few 10s of centimetres in geological models and can be applied to highly heterogeneous reservoirs.