We study transport in a three-dimensional, Eulerian tracer model that uses a subhemispheric domain. Because radon decays radioactively, with an e-folding time of 5.5 days, it provides a test of transport away from its surface, continental source. In particular, we evaluate the importance of two vertical transport processes: (1) parameterizations of cloud transport based on precipitation and based on relative humidity and (2) vertical advection resolved by winds with a horizontal scale of 190 km and a temporal scale of 6 hours. We compare simulated and observed concentrations and distributions of concentration, as well as results from different model versions, such as those using the finite difference advection schemes of Smolarkiewicz (1983) and of Prather (1986). Meteorological variability, which affects the ratio of the standard deviation of concentration &sgr; to the mean concentration c¿, determines the number of independent events that must be sampled to observe or simulate a statistically significant c¿. We did not allow lateral import of radon into our domain, but our simulations could be corrected to estimate the effect of a realistic input by multiplying the simulated c¿ by α and dividing &sgr;/c¿ by √α, with α≈1.04 in the planetary boundary layer (PBL) and 1.5 in the upper free troposphere (FT).

The best simulations over the United States suggest that less than 10% of the radon above 2 km and up to 30% of the radon above 5.5 km may come from Asia. The importance of imported radon and the statistical significance of our simulations limit the reliability of comparisons in the upper FT. The observed c¿ and the modeled c¿, using a 15-day simulation period, may only be accurate to about a factor of 2. Unfortunately the parameterization of cloud transport that gives the best simulations in the upper FT over the United States does not give the best results in the remote marine PBL. Nonetheless, above 6--8 km, cloud transport is probably more important than both the mean vertical advection and the finite difference advection scheme in determining the average radon concentration over the United States. ¿ American Geophysical Union 1989