The relation between radon in schools and in dwellings: A case study in a rural region of Southern Serbia

Recognized as a significant health hazard, radon (Rn) has been given increasing attention for years. Surveys of different kinds have been performed in many countries to assess the intensity and the geographical extent of possible Rn problems. Common surveys cover mainly dwellings, the indoor place with highest occupancy, and schools, where people spend a large fraction of their lifetime and which can also be considered exemplary for Rn exposure at workplaces; it has however been observed that relating them is difficult. It was unclear whether residential Rn at a location, or in a region, can be predicted by Rn at a school of that location, or vice versa. To current knowledge, no general rule seems applicable, as few models to describe the relationship between Rn in dwellings and in schools have been developed. In Southern Serbia, a Rn survey in a predominantly rural region was based on measurements in primary schools. The question arose whether or to which degree the results can be considered as indicative or even representative for residential Rn concentrations. To answer the question an additional survey of indoor Rn concentrations in dwellings was initiated, designed and performed in Sokobanja district in 2010–2012 in a manner to be able to detect a relationship if it exists. In the study region, 108 dwellings in 12 villages and towns were selected, with one primary school each. In this paper, we investigate how a relation between Rn in schools and dwellings could be identified and quantified, by developing a model and using experimental data from both the above main and additional surveys. The key criterion is the hypothesis that the relation dwellings – schools, if it exists, is stronger for dwellings closer to a school than for those dwellings further away. We propose methods to test the hypothesis. As result, the hypothesis is corroborated at 95% significance level. More specifically, on town level (typical size about 1 km), the Rn concentration ratio dwelling/school is about 0.8 (geometrical mean), with geometrical standard deviation (GSD) about 1.9. For dwelling and school hypothetically in the same location, the ratio is estimated about 0.7 with GSD about 1.5. We think that the methodology can be applied to structurally similar problems. The results could be used to create “conditional maps” of Rn concentration in dwellings, i.e., for example a map of probabilities that indoor Rn concentrations in dwellings exceed 100 Bq/m³, as function of Rn concentration in the local school.