Carbon capture and storage reservoir properties from poroelastic inversion: A numerical evaluation

We investigate the prospect of estimating carbon capture and storage (CCS) reservoir properties from P-wave intrinsic attenuation and velocity dispersion. Numerical analogues for two CCS reservoirs are examined: the Utsira saline formation at Sleipner (Norway) and the coal-bed methane basin at Atzbach-Schwanestadt (Austria). P-wave intrinsic dispersion curves in the field-seismic frequency band, obtained from theoretical studies based on simulation of oscillatory compressibility and shear tests upon representative rock samples, are considered as observed data. We carry out forward modelling using poroelasticity theories, making use of previously established empirical relations, pertinent to CCS reservoirs, to link pressure, temperature and CO2 saturation to other properties. To derive the reservoir properties, poroelastic inversions are performed through a global multiparameter optimization using simulated annealing. We find that the combination of attenuation and velocity dispersion in the error function helps significantly in eliminating the local minima and obtaining a stable result in inversion. This is because of the presence of convexity in the solution space when an integrated error function is minimized, which is governed by the underlying physics. The results show that, even in the presence of fairly large model discrepancies, the inversion provides reliable values for the reservoir properties, with the error being less than 10% for most of them. The estimated values of velocity and attenuation and their sensitivity to effective stress and CO2 saturation generally agree with the earlier experimental observation. Although developed and tested for numerical analogues of CCS reservoirs, the approach presented here can be adapted in order to predict key properties in a fluid-bearing porous reservoir, in general.