Use of radon for evaluation of atmospheric transport models: sensitivity to emissions

We present comparative analyses of atmospheric radon (Rn) distributions simulated using different emission scenarios and the observations. Results indicate that the model generally reproduces observed distributions of Rn but there are some biases in the model related to differences in large-scale and convective transport. Simulations presented here use an off-line three-dimensional chemical transport model driven by assimilated winds and two scenarios of Rn fluxes (atom cm−2 s−1) from ice-free land surfaces: (A) globally uniform flux of 1.0 within ±60° and 0.5 within 60°N–70°N and (B) uniform flux of 1.0 between 60°S and 30°N followed by a sharp linear decrease to 0.2 at 70°N. We considered an additional scenario (C) where Rn emissions for case A were uniformly reduced by 28%. Results show that case A overpredicts observed Rn distributions in both hemispheres. Simulated Northern Hemisphere Rn distributions from cases B and C compare better with the observations, but are not discernible from each other. In the Southern Hemisphere, surface Rn distributions from case C compare better with the observations. We performed a synoptic-scale source–receptor analysis for surface Rn to locate regions with ratios B/A and B/C less than 0.5. Considering the maximum uncertainty in regional Rn emissions of a factor of 2, our analysis indicates that additional measurements of surface Rn, particularly during April-October and north of 50°N over the Pacific as well as Atlantic regions, would make it possible to determine if the proposed latitude gradient in Rn emissions is superior to a uniform flux scenario.