Spatial and temporal variability of radon concentrations and resulting statistical issues in radon research

Radon (Rn) is an environmental agent which is of high interest for two reasons. At the one hand it represents a health hazard; mainly through inhalation of its progeny; on the other hand it can be used as tracer for environmental processes, such as transport in and exhalation rates from the ground, and atmospheric dispersion studies. These phenomena are relevant in climatic and seismic research. All applications require reliable measurement of Rn concentrations in different media and quantification of derived values, such as spatial or temporal means and other statistics. Owing to the complex mechanisms by which Rn is generated, transported and diffused into various environmental media and compartments, its concentrations can be very variable as a reflection of the variability of the physical controls and their interactions. The observation process, on the other hand, i.e., sampling and measurement, is prone to error and uncertainty related to the observation procedures and to the stochastic character of radioactive decay. Uncertainty resulting from the observation process has to be kept controlled by quality assurance, while natural variability must be accounted for by appropriate sampling design and modelling steps. In this presentation, natural variability will be addressed on some examples. With regard to spatial variability of long-term mean Rn levels, we shall discuss physical controls which contribute to variability, and the consequences for planning indoor Rn surveys. Concerning temporal variability, we shall discuss the problem of reliability of short term measurements to assess long-term means. (author)