The injection of carbon dioxide (CO 2 ) captured at large point sources into deep saline aquifers can significantly reduce anthropogenic CO 2 emissions from fossil fuels. Dissolution of the injected CO 2 into the formation brine is a trapping mechanism that helps to ensure the long-term security of geological CO 2 storage. We use thermochronology to estimate the timing of CO 2 emplacement at Bravo Dome, a large natural CO 2 field at a depth of 700 m in New Mexico. Together with estimates of the total mass loss from the field we present, to our knowledge, the first constraints on the magnitude, mechanisms, and rates of CO 2 dissolution on millennial timescales. Apatite (U-Th)/He thermochronology records heating of the Bravo Dome reservoir due to the emplacement of hot volcanic gases 1.2-1.5 Ma. The CO 2 accumulation is therefore significantly older than previous estimates of 10 ka, which demonstrates that safe long-term geological CO 2 storage is possible. Integrating geophysical and geochemical data, we estimate that 1.3 Gt CO 2 are currently stored at Bravo Dome, but that only 22% of the emplaced CO 2 has dissolved into the brine over 1.2 My. Roughly 40% of the dissolution occurred during the emplacement. The CO 2 dissolved after emplacement exceeds the amount expected from diffusion and provides field evidence for convective dissolution with a rate of 0.1 g/(m 2 y). The similarity between Bravo Dome and major US saline aquifers suggests that significant amounts of CO 2 are likely to dissolve during injection at US storage sites, but that convective dissolution is unlikely to trap all injected CO 2 on the 10-ky timescale typically considered for storage projects.geological carbon storage | thermochronology | noble gases | porous media convection | carbon sequestration C arbon capture and storage has been identified as a potential technology for reductions in carbon dioxide (CO 2 ) emissions from coal-and natural gas-fired power plants. CO 2 that would otherwise be released into the atmosphere is captured at power plants and injected into porous geological formations for permanent storage. Carbon capture and storage has the potential for significant reductions of anthropogenic CO 2 emissions, because deep saline aquifers provide large storage volumes (1, 2) and existing operations have demonstrated that CO 2 injection and monitoring in saline aquifers are feasible (3).The leakage of CO 2 from the storage formation into potable aquifers or back into the atmosphere is an inherent risk of largescale geological CO 2 storage (4-6). Long-term storage security is therefore enhanced by physical and chemical processes that increasingly trap the injected CO 2 in the subsurface over time. Injected CO 2 can be trapped by capillary forces through the formation of disconnected ganglia or by precipitation as solid phases (7-9). Dissolution of CO 2 into the brine not only is a required first step for the subsequent permanent trapping in stable minerals, but also is considered a trapping mechanism itself. The density of the brin...