What Have We Learnt About CO2 Leakage in the Context of Commercial-Scale CCS?

The viability of Carbon Capture and Storage (CCS) depends on the reliable containment of injected CO2 in the subsurface. Robust and cost-effective approaches to measure monitor and verify CO2 containment are required to demonstrate that CO2 has not breached the reservoir, and to comply with CCS regulations. This includes capability to detect and quantify any potential leakage to surface. It is useful to consider the range of possible leak rates for potential CO2 leak pathways from an intended storage reservoir to surface to inform the design of effective monitoring approaches. However, in the absence of a portfolio of leakage from engineered CO2 stores we must instead learn from industrial and natural analogues, numerical models, and laboratory and field experiments that have intentionally released CO2 into the shallow subsurface to simulate a CO2 leak to surface. We collated a global dataset of measured or estimated CO2 flux (CO2 emission per unit area) and CO2 leak rate from industrial and natural analogues and field experiments. We then examined the dataset to compare emission and flux rates and seep style, and consider the measured emission rates in the context of commercial scale CCS operations. We find that natural and industrial analogues show very wide variation in the scale of CO2 emissions, and tend to be larger than leaks simulated by CO2 release experiments. For all analogue types (natural, industrial, or experiment) the emission rates show greater variation between sites than CO2 flux rates. Quantitation approaches are non-standardized, and that measuring and reporting both the CO2 flux and seep rate is rare as it remains challenging, particularly in marine environments. Finally, we observe that CO2 fluxes tend to be associated with particular emission characteristics (vent, diffuse, or water-associated). We propose that characteristics could inform the design and performance requirements for CO2 leak monitoring approaches tailored to detect specific emission styles