WHY INDOOR RADON DATA ARE NOT LOG-NORMAL, BUT MAY SOMETIMES BE APPROXIMATELY

It is often considered that indoor radon concentration data follow a log-normal distribution. The distribution of indoor radon data of the Walloon region (Belgium) is compared to the log-normal with the quantile-quantile plot, to show which kind of deviations can be observed. Simulated data provide a theoretical frame for understanding these deviations. The 3-component nature of indoor radon (radon from outdoor air, from building materials and from subsoil) generates deviations from the log-normal trend on the low- and high-concentration sides, but the low-C deviation can be compensated and even inverted by the effect of measurement uncertainties. The predicted shapes of low-C and high-C tails are well observed in the global distribution of experimental measurements. The deviation from log-normality is often strong in the low-C tail for which there is no interest, as it does not correspond to a radiological problem. Therefore, it is proposed to use the log-normal only for the high-C part of the distribution, above the median. This part of the distribution is expected to behave log-normally when a single component is dominant, e.g. the soil component in radon-affected areas, whereas more deviations should show up in unaffected areas, where the three components significantly contribute to the radon concentration. This result is illustrated with the data obtained for homogeneous geological units.

TONDEUR François;  CINELLI Giorgia; 

2015-01-23

Czech Technical University in Prague

JRC91712

978-80-01-05548-9,   

https://publications.jrc.ec.europa.eu/repository/handle/JRC91712,