Abstract Accurate assessment of natural CO2 occurrence and its saturation state with respect to the mineral buffering effect could better guide CO2 storage site selection, injection performance and capacity estimation. The primary control on naturally occurring CO2 in the Earth’s crust results from its origin: Mantle and magmatic sources, thermal decomposition of carbonate minerals, thermal maturation of organic matter, thermochemical sulfate reduction, and bacterial activity. The concentration and distribution of the generated CO2 is further controlled by the migration processes and trapping mechanisms, and eventually, buffered by mineral-gas-fluid interactions. We investigated the controls on natural occurring CO2, including its concentration and distribution as a function of key geological variables, such as temperature, pressure, mineralogy, and fluid chemistry. An example from Mobile Bay (e.g., ), is compared with other field studies including Gulf Coast reservoirs and geothermal areas from Iceland. We coupled CO2 thermodynamic modeling with the field data to describe the geological controls on CO2 occurrence, and refined a theoretical correlation for estimating CO2 concentrations in geological formations. The primary observation used for estimating CO2 volumetrics is that the partial pressure of CO2 (p CO2) covaries with temperature, while the reservoir mineralogy plays a significant role in determining the p CO2–T correlation. This trend can be related to pH buffering associated with aluminosilicate and carbonate minerals and pore water interactions based on fundamental thermodynamic principles. The theoretical correlation matches the field data reasonably well, and should apply to a wide range of CO2 concentrations for reservoirs with constrained access to CO2 rich fluids. We applied state of the art CO2 injection reactive transport modeling (RTM) to investigate how natural occurring CO2 concentration and buffering state affect CO2 injection performance and capacity. Our RTM simulates large-scale CO2 injection into subsurface reservoirs, with the ability to capture the complex interplay of multiphase flow, capillary trapping, diffusion, convection, and chemical reactions, that may have significant impacts on both injection performance and storage security. The simulation results suggest that reservoirs with initial CO2 concentration below the buffering capacity tend to facilitate CO2 storage by increasing the structural, residual and solubility trapping. The under saturation conditions also help to avoid potential mineral precipitation and formation damage near the injection well and therefore, is a better choice for CO2 storage over reservoirs with initial CO2 exceeding the buffering capacity. We plan to investigate CO2 injections in different types of siliciclastic and carbonate reservoirs. This study expanded our fundamental understanding of CO2 occurrence, buffering and sequestration processes at multiple spatial and temporal scales in nature. The results helped us make better judgments of CO2 storage capacity and site selection that led to better design decisions from appraisal to development to monitoring.