23This paper presents the initial results of a scientific drilling project to recover core 24 and pressurized fluid samples from a natural CO 2 reservoir, near the town of Green River, 25Utah. The drilling targeted a stacked sequence of CO 2 -charged Jurassic sandstone reservoirs 26 and caprocks, situated adjacent to a CO 2 -degassing normal fault. This site has actively 27 leaked CO 2 from deep supercritical CO 2 reservoirs at depth >2km within the basin for over 28 Geyser constrain mixing models which show that, within the Navajo Sandstone, the 49 reservoir fluids are undergoing complex mixing of: (i) CO 2 -saturated brine inflowing from 50 the fault, (ii) CO 2 -undersaturated meteoric groundwater flowing through the reservoir and 51 (iii) reacted CO 2 -charged brines flow through fracture zones in the overlying Carmel 52Formation caprock, into the formations above. Such multi-scale mixing processes may 53 significantly improve the efficiency with which groundwaters dissolve the migrating CO 2 .

No abstract

Carbon capture and geological storage represents a potential means of managing atmospheric carbon dioxide levels. Understanding the role of faults, as either barriers or conduits to the flow of carbon dioxide, is crucial for predicting the long-term integrity of geological storage sites. Of particular concern is the influence of geochemical reactions on the sealing behaviour of faults and the impact of seismicity and stress regime on fault stability. Here, we examine a 135,000-year palaeorecord of carbon dioxide leakage from a faulted, natural carbon dioxide reservoir in Utah. We assess the isotope and trace-element composition of U-Th-dated carbonate veins, deposited by carbon-dioxide-rich fluids. Temporal changes in vein geochemistry reveal pulses of carbon dioxide injection into the reservoir from deeper formations. Surface leakage rates increase by several orders of magnitude following these pulses. We show that each pulse occurs around 100-2,000 years after the onset of significant local climatic warming, at glacial to interglacial transitions. We suggest that carbon dioxide leakage rates increase as a result of fracture opening, potentially caused by changes in groundwater hydrology, the intermittent presence of a buoyant gas cap and postglacial crustal unloading of regions surrounding the fault