QUANTIFYING THE EFFECTS OF SPATIAL UNCERTAINTY IN FRACTURE PERMEABILITY ON CO2 LEAKAGE THROUGH CAP ROCKS DURING GEOLOGIC CO2 STORAGE IN CONTINENTAL FLOOD BASALTS

Recent studies suggest that continental flood basalts may be suitable for geologic CO₂ sequestration on the basis of fluid-rock geochemical reactions resulting in permanent mineral trapping over short time-scales. In addition, flood basalts consist of layered assemblages of high-permeability flow margins and low-permeability flow interiors, the latter of which may provide physical trapping potential on time-scales needed for mineralization. However, little is known about fracture-controlled CO₂ leakage through the flow interior as supercritical CO₂ undergoes phase change during decompression. We simulate CO₂ leakage within a fractured flow interior, using a fracture network derived from high-resolution LiDAR scans of a Columbia River Basalt outcrop near Starbuck, WA. The model is two-dimensional with 5 x 5 m areal extent and bimodal heterogeneity defined on the basis of fracture or matrix properties. The code selection for this research is TOUGH3 (beta), compiled with equation of state ECO2M, which accounts for CO₂ phase change from supercritical fluid to subcritical liquid or gas. Initial conditions consist of a hydrostatic gradient corresponding to 750-755 m below ground surface, and a constant temperature of 32° C; under these conditions, the critical point for CO₂ occurs 1.5 m above the bottom of the model domain. CO₂ leakage is simulated as a constant overpressure of 0.5 MPa at the base of the model. Matrix permeability is based on hydraulic tests of similar basalt, and fracture permeability is estimated on the basis of fracture aperture distributions taken from literature. To account for the uncertainty associated with the spatial distribution of in situ fracture apertures, 50 simulations are completed within the same fracture network with spatially random fracture permeability constrained by the aperture distribution. Model results suggest that leaking CO₂ migrates up through the fractures on the time scale of 10-20 years; however, the fractures must be adequately connected to allow leakage through the 5 m model domain. Moreover, even when such a path is available, gas-phase CO₂ may be unable to overcome the capillary entry pressure into the fractures above. Therefore, capillary effects are shown to be an important component of early-time CO₂ trapping during decompression within a basalt flow interior.