Abstract. Because of its ubiquitous release on land and well-characterized atmospheric loss, radon-222 has been very useful for deducing fluxes of greenhouse gases such as CO₂, CH₄, and N₂O. It is shown here that the radon-tracer method, used in previous studies to calculate regional-scale greenhouse gas fluxes, returns a weighted-average flux (the flux field F weighted by the sensitivity of the measurements to that flux field, f) rather than an evenly-weighted spatial average flux. A synthetic data study using a Lagrangian particle dispersion model and modeled CO₂ fluxes suggests that the discrepancy between the sensitivity-weighted average flux and evenly-weighted spatial average flux can be significant in the case of CO₂, due to covariance between F and f for biospheric CO₂ fluxes during the growing season and also for anthropogenic CO₂ fluxes in general. A technique is presented to correct the radon-tracer derived fluxes to yield an estimate of evenly-weighted spatial average CO₂ fluxes. A new method is also introduced for correcting the CO₂ flux estimates for the effects of radon-222 radioactive decay in the radon-tracer method.