The naturally-occurring radioactive gases ²²²Rn and ²²⁰Rn are widely used as atmospheric tracers in a variety of applications pertinent to climate and air quality studies, but their use in this context is currently limited by poor knowledge of the spatial and temporal pattern in emissions. The aim of this research was to improve knowledge of ²²²Rn and ²²⁰Rn emissions by investigating their spatial and temporal variation in emissions on various scales. Novel approaches to measuring ²²²Rn and ²²⁰Rn fluxes by the closed chamber method have been developed and compared with some existing methods, and studies of some controlling variables for which there is limited and conflicting information (water table depth, freezing of the soil and snow cover) have been carried out. Studies of the short-term variability in ²²²Rn emissions made with an automatic chamber showed that outside periods of rain the ²²²Rn flux was relatively constant (CV = ~ 25 %), but heavy rainfall may temporarily completely suppress the radon flux if the soil surface becomes saturated. On the seasonal time-scale, flux measurements made on a medium- moisture was found to be the most important factor controlling the variations in radon flux; air pressure and temperature were not important. Field measurements and a laboratory study using a soil monolith showed that water table depth was also an important factor for ²²²Rn flux, but not for ²²⁰Rn, due to its much shorter diffusion length. Freezing of the soil surface layer (~ 5cm depth) did not cause a significant reduction in ²²²Rn or ²²⁰Rn flux. Studies of the spatial variability of ²²²Rn emissions at the local scale showed that ²²²Rn flux is approximately normally distributed (CV = 55 %), and that soil moisture is an important factor. However, measurements made at 15 sites of different soil type and geology across North Britain showed ²²²Rn emissions to be log-normally distributed at this larger scale and highly variable (CV = ~ 200%). The ²²⁶Ra content of the soil, which ranged from <3 Bq kg^-1 to 55 Bq kg^-1 was found to account for a large proportion of the observed variability (~ 80 %). The median ²²²Rn flux was 9.7 Bq m^-2 h^-1, indicating that the average flux from this region is considerably lower than the global average, as would be expected given the large proportion of peat soils and generally high soil water content and shallow water tables. A ²²²Rn flux map produced for North Britain using a GIS and empirical model based on soil and geology classifications was able to reproduce the main spatial pattern in emissions, but on the whole, significantly under- predicted the magnitude of fluxes.