Measurements of radon exhalation from a gravely sandy loam have been made in a semi‐arid climate by using a combination of closed accumulation, flow‐through accumulation, and 222Rn and 210Pb soil profiles. The meteorological factors that most affected the instantaneous value of exhalation of 222Rn were atmospheric pressure and rain. Effects due to other parameters such as wind or temperature were either comparatively small or undetectable. No evidence was seen for migration of radon from distant (≫10m) sources or for an effect on exhalation due to limited nearby seismic activity. Measurements for 220Rn indicated its exhalation was also sensitive to pressure variation but to a lesser extent than for 222Rn. While instantaneous exhalation of 222Rn could easily vary by a factor of 2 or more, time‐averaged exhalation was found to be close to that expected for pure diffusion. There is thus some indication that the time‐averaged effect of cyclic environmental variables is small for this soil. Comparison with transport equations indicates that it is difficult to explain the observed variation in surface flux density solely in terms of the radon concentration gradient in the top few decimeters of soil. A contribution to transport from direct flow, perhaps through inhomogeneities such as cracks or channels, is one possible explanation.

[1] There is a need for improved estimates of the radon ( 222 Rn) flux density from the ocean for use in the modeling and interpretation of atmospheric radon in global climate and air pollution studies. We use a modification of a frequently used model of gas transfer to generate global predictions of ocean radon flux density for each month of the year (climate averaged) on a 192 by 94 global grid. Compared with the often-used approximation of a constant radon flux from the ocean, the model's predictions indicate large variations over regions of the ocean (a factor of ten is not uncommon). For example, latitude bands near the equator and Southern Ocean are predicted to emit relatively high average radon flux compared with other latitude bands. The predicted annually-averaged flux density from the ocean is 0.0382 mBq m À2 s À1 (0.00182 atoms cm À2 s À1 ), smaller than some commonly-used estimates.

The accumulator technique was used to measure radon and thoron flux density at a variety of locations throughout Australia. This is the first such systematic study of Australia and, in the case of thoron, one of few such studies of any large land mass. Seasonally adjusted arithmetic mean flux densities from Australian soils were estimated to be 22 mBq m−2 s−1 (1.05 atom cm−2 s−1) for radon and 1.7 Bq m−2 s−1 (0.0135 atom cm−2 s−1) for thoron. Consideration of statistical sampling error, and systematic error with the accumulator method, leads to an error estimate of about ±20% for these numbers; projection of total flux to the atmosphere requires consideration of additional sources of error. Only modest correlations with variables easily measured in the field were observed. The strongest correlation was a positive one between flux density and gamma dose rate 1 m above ground. Weaker correlations were seen with soil temperature (positive) and soil moisture (negative at higher moistures). Radon and thoron flux density were strongly correlated, but only a weak correlation (negative) existed between them and vegetation. The amount of radon isotope released to the pore space seems particularly important for controlling the wide variation in observed flux densities, but it remains difficult to predict flux densities based on simple field measurements or information in conventional soil and geological maps.

A two‐dimensional finite element code was used to investigate the effect of changes in surface air pressure on radon flux from soil with parallel, partially penetrating cracks. A sensitivity analysis investigates the effects of various crack dimensions, soil characteristics, and surface air pressure on radon flux from the soil surface to the atmosphere. Simulation results indicate that radon flux is most sensitive to soil properties; the diffusion coefficient is most important, followed by permeability and porosity. Radon flux is also sensitive to changes in barometric pressure, which cause variations in radon flux above and below the average diffusive flux. Sinusoidal variations in barometric pressure cause a net increase in the average radon flux from the soil, because increases in flux during periods of decreasing pressure are greater than the decreases in flux during periods of decreasing pressure of equal magnitude. Cracks were found to significantly increase radon flux from soils of low permeability.

The transport of 222Rn from fractured rock has been studied in an abandoned mine. Pressure‐induced flow (both natural and artificial) is quite important and can easily cause an order of magnitude increase in the instantaneous transport of radon into the tunnel airspace compared to that due to flowfree diffusion. Permeability studies indicate that large‐scale cracks of surface density of the order of several cracks per square meter dominate flow. In first order, effective flux due to natural pressure variation follows an approximate dP/dt dependence. In higher order, there exists an enhancement of time‐averaged flux by typically a factor of 2 due to natural pressure variation. Mathematical modeling indicates that the relation between pressure and the strength and time dependence of the radon transport is difficult to explain solely with conventional models for semi‐infinite homogeneous porous media. A model of cul‐de‐sac chambers is proposed to account for some of this dependence.