Carbon Dioxide-Enhanced Oil Recovery (CO2-EOR) has the potential for well blowouts that could cause casualties and environmental damage. To assess the consequence of such accidents, a reduced-scale field experiment of CO2 release was performed based on scaling rules instead of a full-size field test that was economically infeasible. A series of scaling rules was introduced to upscale the reduced-scale field experiment to full-size. To validate the scaling rules, numerical simulation was carried out based on the k-epsilon turbulence model which proved to be an effective way to predict the concentration field for heavy gas dispersion. For concentration variation, the general tendencies of the simulation and experimental observations remained identical except nearby the jet nozzle where the measured CO2 concentration from the experiment was obviously higher than that in the simulation. Statistical performance indicators were introduced to verify the consistency between the scaled results and the simulated ones, and the results showed that using the scaling rules to scale the field experiment exhibited acceptable accuracy at small flow rates and these scaling rules appear applicable for field experiments of accidental releases. (C) 2013 Elsevier Ltd. All rights reserved.