Multiphysics Bayesian rock physics inversion for CO2 storage monitoring

ABSTRACT Carbon capture and storage is acknowledged as one of the key technologies in the energy system transition toward low greenhouse gas emissions, as it simultaneously meets the demand for energy supply. Large-scale and safe CO2 storage requires monitoring plans to address operational and regulatory requirements as well as public acceptance. This work aimed to quantify the impact of various combinations of geophysical inputs to the accurate quantification of static and dynamic reservoir parameters in CO2 storage sites. Geophysical attributes were used as input data for multiphysics Bayesian rock physics inversion. This study quantified the contribution of P- and S-wave velocity, density, and resistivity measurements to estimate CO2 saturation, pore pressure, elastic rock properties, porosity, and a parameter related to the fluid distribution at the pore scale (patchiness). By considering case studies representative of Sleipner, Snøhvit, and Smeaheia storage sites, it was demonstrated that a well-constrained and low-uncertainty estimation of CO2 saturation requires the use of resistivity input data. Accurate pore pressure estimation is difficult and has to take advantage of all possible input data and, if available, prior information to be able to be discriminated from saturation effects. The patchiness parameter cannot be accurately recovered regardless of data inputs considered in this study, but it does not prevent a proper estimation of CO2 saturation. Porosity estimation requires density data, whereas rock frame mechanical moduli estimation requires S-wave velocity information. Use of prior information is always beneficial to better constrain the estimates. Quantifying the dynamic reservoir parameters with uncertainty estimation is crucial for conformance monitoring where observed and modeled behaviors of the CO2 injection and migration in the reservoir need to be verified by the operator.