. We thank J. Apps and A. Cortis for constructive review of this report.
2
EXECUTIVE SUMMARYThe injection and storage of anthropogenic CO 2 in deep geologic formations is a potentially feasible strategy to reduce CO 2 emissions and atmospheric concentrations. While the purpose of geologic carbon storage is to trap CO 2 underground, CO 2 could migrate away from the storage site into the shallow subsurface and atmosphere if permeable pathways such as well bores or faults are present. Large-magnitude releases of CO 2 have occurred naturally from geologic reservoirs in numerous volcanic, geothermal, and sedimentary basin settings. Carbon dioxide and natural gas have also been released from geologic CO 2 reservoirs and natural gas storage facilities, respectively, due to influences such as well defects and injection/withdrawal processes. These systems serve as natural and industrial analogues for the potential release of CO 2 from geologic storage reservoirs and provide important information about the key features, events, and processes (FEPs) that are associated with releases, as well as the health, safety, and environmental consequences of releases and mitigation efforts that can be applied.We describe a range of natural releases of CO 2 and industrial releases of CO 2 and natural gas in the context of these characteristics. Based on this analysis, several key conclusions can be drawn, and lessons can be learned for geologic carbon storage. First, CO 2 can both accumulate beneath, and be released from, primary and secondary reservoirs with capping units located at a wide range of depths. Both primary and secondary reservoir entrapments for CO 2 should therefore be well characterized at storage sites. Second, many natural releases of CO 2 have been correlated with a specific event that triggered the release, such as magmatic fluid intrusion or seismic activity. The potential for processes that could cause geomechanical damage to sealing cap rocks and trigger the release of CO 2 from a storage reservoir should be evaluated. Third, unsealed fault and fracture zones may act as fast and direct conduits for CO 2 flow from depth to the surface. Risk assessment should therefore emphasize determining the potential for and nature of CO 2 migration along these structures. Fourth, wells that are structurally unsound have the potential to rapidly release large quantities of CO 2 to the atmosphere. Risk assessment should therefore be focused on the potential for both active and abandoned wells at storage sites to transport CO 2 to the surface, particularly at sites with depleted oil or gas reservoirs where wells are abundant. Fifth, the style of CO 2 release at the surface varies widely between and within different leakage sites. In rare circumstances, the release of CO 2 can be a self-enhancing and/or eruptive process; this possibility should be assessed in the case of CO 2 leakage from storage reservoirs. Sixth, the hazard to human health has been small in most cases of large surface releases of CO 2 . This could be due to imp...