The invention of the portable iGrav superconducting gravimeter has provided unique precision and stability to monitor the subtle gravity field variations due to injection of CO2 into deep geological storages. According to the data collected over six months of continuous measurements by the iGrav001 at the University of Calgary, an accuracy of better than 1 µGal is achievable after corrections for environmental interferences, while the instrumental drift remains negligible. The combination of temporal observations of the iGrav along with observations of relative spring gravimeters, obtained in special network configurations, can lead to accuracies of 1-2 µGal for the gravity variations between two epochs before and after CO2 injection.
In almost all previous CO2 storage modelling studies, the porosity change effect on the gravity signal has been assumed to be negligible. In this study, geological storages are divided into two major categories: unconfined and confined reservoirs. In the unconfined reservoirs, injected CO2 replaces the ambient fluid, causing a negative density anomaly in the reservoir and a negative gravity signal with negligible deformation at the ground surface. Confined reservoirs, on the other hand, experience a volumetric change and the injected CO2 fills the excessive porosity, causing a positive gravity signal representing the attraction of the injected CO2, and a positive ground surface deformation signal along with the associated free-air gravity effect. Dependency of gravity and ground deformation signals on variable properties, such as injected CO2 mass and density, along with reservoir depth and size, is examined by forward modelling for both confined and unconfined CO2 reservoirs. In addition, while inversion of the unconfined CO2 reservoirs obeys the conventional gravimetric modelling formulation, a novel inverse modelling formulation is established for confined reservoirs. A relationship is developed between the density changes and the fractional volumetric changes due to the CO2 injection in the confined reservoirs, and a linear inversion is implemented to estimate fractional volumetric changes using both ground deformation and gravity observations jointly. The developed inverse modelling methodology is evaluated using several synthetic case studies.