Characterization of terrain-specific risk posed by potential future CO₂ release from orphan wells remains a major carbon storage technology gap. For such wells, risk analysis has focused on aggregate long-term future impacts of seepage at rates on the order of 1 g/m² per day on storage goals but not of potential future large-scale venting through overlooked unsealed orphan wells, which may dominate CO₂-loss scenarios. Flux measures made on Crystal Geyser, in Utah, indicate that such loss could range from 10 to 100,000 kg/day per site, and so is acutely hazardous in unfavorable meteorological and terrain circumstances. To characterize potential risk from such abrupt discharges, a prototype heavy-gas near-field dispersion model was coupled to a Geographic Information Systems (GIS) algorithm to evaluate terrain-related impacts on lethal downwind CO₂ concentration as a function of wind speed. By eliminating low-risk areas, application of the risk model to a potential sequestration site in west-central Indiana allowed a 30-fold gain in the cost effectiveness of surveying the site for unsealed orphan wells. After improvements to better model near-field CO₂ dispersion in complex terrains, this type of approach could be used to rapidly screen large areas proposed for future CO₂ sequestration for relative potential risk, and thereby prioritize them to design, or comply with, regulatory risk management goals.