Applied Radiation and Isotopes 94 (2014) 328–337

Contents lists available at ScienceDirect

Applied Radiation and Isotopes journal homepage: www.elsevier.com/locate/apradiso

High variability of indoor radon concentrations in uraniferous bedrock areas in the Balkan region Z.S. Žunić a, P. Ujić a,n, L. Nađđerđ a, I.V. Yarmoshenko b, S.B. Radanović c, S. Komatina Petrović d, I. Čeliković a, M. Komatina d, P. Bossew e a

Institute of Nuclear Sciences “Vinča”, University of Belgrade, P.O. Box 522, 11001 Belgrade, Serbia Institute of Industrial Ecology, Ural Branch of Russian Academy of Sciences, Sophy Kovalevskoy 20, 620990 Ekaterinburg, Russia c Economic Association for Production, Processing and Transport, Kolubara Mining Basen Ltd., Lazarevac, Serbia d Association of Geophysicists and Environmentalists of Serbia, Dimitrija Avramovića 38, 11000 Belgrade, Serbia e Bundesamt für Strahlenschutz (German Federal Office for Radiation Protection), Postfach 10 01 49, D-38201 Salzgitter, Germany b

H I G H L I G H T S







The excessive radon values are explained from geological point of view. The excessive radon values must not be rejected, which is often performed. Rejection of excessive concentrations cause serious errors in the dose estimation. Method to estimate the annual mean concentration when some seasons are missing.

art ic l e i nf o

a b s t r a c t

Article history: Received 21 August 2013 Received in revised form 20 August 2014 Accepted 30 August 2014 Available online 22 September 2014

In this work the strong influence of geological factors on the variability of indoor radon is found in two of three geologically very different regions of South-Eastern Europe. A method to estimate the annual mean concentration when one seasonal measurement is missing is proposed. Large differences of radon concentrations in different rooms of the same house and significant difference in radon concentrations in one season comparing it to the others are noted in certain cases. Geological factors that can lead to such behavior are discussed. & 2014 Elsevier Ltd. All rights reserved.

Keywords: Indoor radon Indoor radon variations Uraniferous bedrock areas Geological structure Radon/groundwater migration

1. Introduction In the reports of the United Nations Scientific Committee on Exposure to Atomic Radiation (UNSCEAR, 2000) it has been estimated that excluding doses from radiotherapy and nuclear accidents, the global average annual effective dose to a member of the public is about 3.0 mSv. The largest contribution of 52% to the annual effective dose is due to the exposure to radon and its progenies. Recent re-evaluation of Rn (we put Rn in short for 222 Rn) dosimetry suggests that the contribution of Rn might be even much higher (Harrison and Marsh, 2012). This important exposure to radon is mainly due to radon in indoor environments

n

Corresponding author. Tel.: þ 381 11 3408149; fax: þ 381 11 2458 681. E-mail address: [email protected] (P. Ujić).

http://dx.doi.org/10.1016/j.apradiso.2014.08.018 0969-8043/& 2014 Elsevier Ltd. All rights reserved.

such as homes, schools and workplaces. At the same time, the radon concentration is affected by spatial and temporal variability at different scales, depending on geogenic and meteorological factors (Porstendörfer et al., 1994). Overall, radon is perhaps the most variable contributor to the total dose. Indoor radon concentration depends on natural and on anthropogenic factors. The latter include construction style of houses, position of a dwelling within a house, building materials, living habits etc. The consequence of this convolution of different influences is that the exactly same building, built on the same geological ground can show very different indoor radon concentrations due to different living habits of inhabitants (Gruber et al., 2013). The geogenic potential for high indoor radon concentrations depends on many factors, which includes the 226Ra content of the soil underneath buildings and the soil permeability. Therefore, indoor radon has a tendency to be correlated with local geology (Appleton and Miles, 2010; Kemski et al., 2009;

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

Scheib et al., 2009). Accordingly besides mapping indoor radon concentrations, which may serve as proxy to exposure, a mapping of the geogenic radon potential (RP) can be very useful in the identification of radon prone areas, i.e. regions in which for geogenic reasons elevated indoor radon concentrations can be expected, depending on anthropogenic factors. On an European level the project of mapping Rn has been underway for several years. Chronologically, the geogenic RP map, still in its initial phase, was preceded by a European map of indoor radon concentrations, which is not yet completed in all European countries. Since the RP depends on the radionuclide content and the permeability of the soil, it is not influenced by anthropogenic factors, and thus it is considered more reliable for the identification of radon prone areas (Gruber et al., 2013). For instance, in the Czech Republic, radon characterization of the ground of every new building site is required by law; in France priority areas are defined where radon surveys are obligatory for several types of public buildings (Ielsch et al., 2010). Sometimes airborne gamma spectrometry referring to an equivalent uranium content in soil is used construct RP maps (Appleton et al., 2011; Smethurst et al., 2008). In order to map RP and to investigate the geological influence on the indoor radon concentration, indoor radon mapping can be used in conjunction with geological boundaries (Appleton et al., 2011). The RP can vary significantly between different and within the same geological unit. Together with the variability of the anthropogenic factors this results in high variability of indoor radon concentration, so that only relatively little of it can be explained by the bedrock and superficial geology. Variability explained by geology ranges from a few percent (Kemski et al., 2009; Bossew et al., 2008) up to 25% (Appleton and Miles, 2010). In this paper, only the relation between indoor radon and the bedrock of the building site is investigated. Other radiometry, such as measurement of the radionuclide content in soil or airborne gamma measurement, was not performed in this investigation. Some authors identified geology as a key factor in increased indoor radon concentration levels (Zhu et al., 2001; Gillmore et al., 2005; Miles and Appleton, 2005). Moreover, some authors suggest that the correlation between bedrock and radon concentration can be used for prediction of the radon risk, especially for regions where few measurements are available (Kemski et al., 2009, Sundal et al., 2004). The radon potential depends on the geological features such as lithological variations and geochemistry and accordingly these features are of high importance in radon mapping (Kemski et al., 2005; Shi et al., 2006; Sundal et al., 2004). Moreover, a large-scale variability could be the consequence of geological features like different kind of faults (Ciotoli et al., 1998; King et al., 1996). The soil constitutes an interface cover between the solid geology (basement rocks) and the atmosphere and may also influence the vertical radon migration, according to its permeability. The water saturation of the shallow part of the soil can be responsible for seasonal variation of soil-gas radon concentration (King and Minissale, 1994). In Serbia a strong dependence between bedrock and indoor radon concentration is already noticed in Niška Banja (Žunić et al., 2007a; Zunic et al., 2007b). The strict proof was not obtained, however all the houses with high indoor radon concentration are built on a travertine, which is permeable and holds a high content of 226Ra. High correlation between soil gas radon concentration, gamma dose rate and 226Ra content (which spatially corresponds to travertine bedrock) in soil were found. In this paper, the correlation between indoor radon concentration and the geological and geochemical environment (i.e., rocks, soil and water) is investigated. Indoor radon measurements were performed in time series of different durations and sometimes longer than one year. A method to estimate average annual radon

329

concentration is presented in the case of an interrupted time series. A method of non-linear regression is proposed to estimate the annual mean concentration when one seasonal measurement is missing. The goal of this study consists of a qualitatively investigation of the relationship between the indoor radon concentration and the geological background, in order to outline geogenic radon prone areas. Variation of an order of magnitude of indoor radon concentration measured in several houses in one season is explained from a geological point of view.

2. Material and methods In Serbia, which is a part of the Balkan region, there are uranium deposits of different types, therefore radioactive geological anomalies are mainly caused by enhanced uranium and thorium concentrations in rock and soil. Several hundred anomalous zones have been identified, mostly due to elevated uranium or thorium concentrations or both (Jankovic, 1990). Since radon is a member of the uranium radioactive decay series, geogenic conditions for enhanced exposure of population to radon in closed spaces exist in Serbia (Zunic et al., 2007b). The reasons for occurrence of high indoor radon levels are many and complex: they include geological factors, building characteristics, usage patterns, etc. (Komatina, 2004). Geological studies in Serbia have clearly identified regions, which have high levels of uraniferous material, but little or no investigations on the exposure of humans to natural radiations in these areas have heretofore taken place. However, results of radon measurements in dwellings in the Balkan region show a close dependence between radon distribution and geological features (Komatina, 2004; Komatina-Petrovic, 2011). In the period from 1997– 1999, a survey of natural radiation sources was performed (Zunic et al., 2001) in three investigated areas, two of which are rural ones: Gornja Stubla, situated in the southwest part of Kosovo, in Kalna a former uranium mine in the Eastern part of Serbia, and in the Montenegrin coast of the Adriatic sea. 2.1. Geological structure of the study areas and migration of groundwater and radon 2.1.1. Kalna Kalna is a former uranium mining district which was exploited from 1948 to 1966, when the mine was closed. It is located in the Stara Planina mountain region in the eastern part of Serbia, where the discovered uranium mineralization is mostly limited to the western slopes of its central part, draining towards the river Trgoviski Timok through its tributaries, the Crnovrska, the Inovska and Gabrovnicka rivers. The most important mineralization and uranium ore deposits occur between 300 and 1000 m above sea level, limited to the granite massive of Janja. The hydrothermal lentice-like layers of deposits and the appearance of uranium are connected to a number of time-activated cleavage faults. Deposits of Gabrovnica are clay-like, crumbling, near-surface cleavage zones. Generally, the mean uranium content in the granite ranges between 6 ppm and 20 ppm. In the area of Stara Planina which is a significant metal-genetic zone, there are four uranium deposits which were temporarily exploited (Jankovic, 1965). The most significant uranium deposit near Kalna is concentrated in the granite massive of Janja (Fig. 1). Hydrothermal veins and saline deposits with the presence of uranium are linked to active brittle zones. The deposit of Mezdreja belongs to the same genetic type. The deposits of Gabrovnica are tied to clay and porous surface layers of brittle zones. The granite massive of Janja (“Janja–Inovo– Gabrovnica” – see Fig. 1) is elongated in the NW–SE direction. It is

330

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

Fig. 1. Geological map of Kalna area. 1. Conglomerate, sand, sandstone, limestone; 2. Conglomerate, sandstone, claystone, limestone (upper urgon facies); 3. Limestone, clayey limestone, shale, (lower urgon facies); 4. Conglomerate and sandstone (Liassic); 5. Medium-grained pyroxene gabbro; 6. Granites of Janja; 7. Red sandstone formation: conglomerate, sandstone, alevrolite; 8. Green rocks, green schists; 9. Sandstone and slate, limestone; 10. Granites of Ravno Bučje; 11. Faults, certain (solid line), inferred (dashed line).

about 19 km long and about 2.5 km wide. Along the perimeter, an increased content of biotite and rocks is distinguished, as well as a decrease in the quantities of K-feldspar and quartz. Deformations and secondary changes (schistose textures, crushing of materials and crystallization of secondary minerals), noticed for the massive as a whole, and are particularly present along its perimeter. They appear as schistose textures, crushed materials, and crystallization of secondary minerals, such as quartz, oligoclase, K-feldspar, and biotite. Subordinate constituents are sphene, apatite, zirconium, and magnetite and secondary are sericite, chlorite, epidote, calcite, limonite, and clayey substance. In the massive of Janja all transitions have been observed, from simple crushing of feldspar and quartz, then bending of mica leaves and feldspar laminae, crumbling of feldspar and filling of cracks by the crumbled feldspar, chlorination of biotite and formation of cataclastic structures, up to the formation of schistose textures with sericite along schistose surfaces. Thus a rock looks like gneiss. These schistose granites are particularly well known in Inovo and along the southern part of the perimeter towards Balta-Berilovac. The central part of Kalna area is located on the Janja granite massif, extending in a NW–SE direction. The massif is extruded into Paleozoic schist (Fig. 1). Because of granite extrusion into the schist, uranium mineralization occurred in contact zones, forming several small deposits. The very interesting geological environment here includes Permian red sandstones, forming a long belt at the SW, and partly at the SE. To the west, mentioned formations are broken by a long fault structure, consisting of Miocene conglomerate, and the western part is made of Lower Cretaceous limestone (Fig. 1).

2.1.2. Gornja Stubla This rural area lies over Paleocene rocks, lower cretaceous flysch and granite. This region represents the northern branches of Skopska Crna Gora at the extreme south of Serbia. It borders a NW–SE fault

zone filled with brecciate rock masses, bearing the secondary uranium mineral autinite. This mountainous area has an altitude ranging from 500 up to 1000 m above sea level. The terrain is made up of diabasic formation (Jurassic), lower Cretaceous chalk flysch, Paleocene layers of trachyte and trachyte tuffs. The diabase-chert formation is formed by magma and sedimentary rocks. Trachytes occur as veins penetrating deeply into diabase formations. The flysch facies is composed of sand with intercalations of clays and lapors. The Gornja Stubla area is predominantly made of alkaline volcanic products, trachyte and trachyte–leucitite tuff (Fig. 2). During geological investigations, uranium-bearing veins were identified not only near Stublovaca hill, but also in the village. It was found that trachyte is characterized by high thorium and uranium concentrations (in one soil sample 238U, 226Ra and 232Th concentrations of about 400, 250 and 200 Bq/kg were found, respectively (Zunic et al.,1999). From the hydrogeological point of view, trachyte and tuff are known as low-permeable to impermeable formations, hence, radon migration from the adjacent deposit and uranium occurrences to the terrain surface through fracture systems is difficult. That is why the area with high radon values is homogenous (isotropic) in this case, and the anomalous zone is autochthonous. The zone of Stublovaca (984 m above sea level) is a part of the northern hillsides of the mountains of Skopska Crna Gora. The upper parts are pastures and the lower are plowed fields and forests. The hydrographic network is well developed. A number of creeks pour from the hill of Stublovaca radially down towards the river Binacka Morava. Springs are very frequent particularly at joints of trachyte and sedimentary rocks. Geology of the area is characterized by diabase formations, lower Cretaceous chalk flysch, Paleocene sediments of trachyte and trachyte tuff. The diabase formation is represented by magmatic and sediment rocks. In the Stublovaca region diabase is more abundant than clays, sandstones, marl, limestones and chalk. In the structure of sandstone, material formed by diabase destruction predominates. The southwest part

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

331

Fig. 2. Geological map of Gornja Stubla area. 1. Alluvium; 2. Marly-sandy clay and sandstone; 3. Andesite breccia; 4. Andesite tuff; 5. Trachyte; 6. Diabase; 7. Gabbro; 8. Sandstone, marl, alevrolite, marly limestone, conglomerate, heterogeneous breccia (Senonian); 9. Limestone-alevrolite formation; 10. Ophiolite mélange with olistolith limestone, diabase, sandstone, chert, etc.; 11. Clay, marl, conglomerate, pyroclastic rocks, reef limestone; 12. Faults, certain (solid line), inferred (dashed line).

of the area is built of Paleocene sediments made of a mixture of gray marl and clay. 2.1.3. Montenegrin coast The geological structure of the Montenegrin coast is significantly different from the two localities presented above. Three geotectonic units are distinguished (see Fig. 3): (A) the autochthone, (B) the Budva or Cukali zone and (C) the high karst zone. Autochthone is made of Cretaceous limestone and Eocene flysch, forming a narrow belt along the coast (Fig. 3). The Budva zone is characterized by an heterogeneous structure (complex of limestone, diabase, trachyte, tuffite, volcanic breccia and Eocene flysch). The investigated houses, as well as the majority of the settlements, are located in the deposits of the Budva (Cukali) zone, partly on autochthone limestone. Within these two geotectonic units, but also in the so-called High karst zone, which is of limestone structure, geological conditions for forming U-formations and uranium deposits were absent – that is the reason why the measured radon concentrations are low. 2.2. Data acquisition Indoor radon measurements were conducted in three very different geological regions, briefly summarized in the previous section: Montenegrin coast (limestone), Kalna (with uranium deposits) in Serbia and Gornja Stubla (with the secondary uranium mineral

autinite) in Kosovo and Metohija. Radon indoor measurements have been conducted in three or four periods of measurements in selected houses for a total period of about one year: three periods at the Montenegrin coast (see Table 1 for the dates specifications), four periods at Gornja Stubla (see Table 4) and four periods longer than one year at Kalna (see Table 9). Indoor radon measurements were carried out at Gornja Stubla in 65 houses selected on the basis of previous geological and geochemical mapping (Jakupi et al., 1997), in 75 houses at Kalna, a former uranium mining district in eastern Serbia, and in 82 buildings along the Montenegrin coast. The measurements were carried out by using the SSI/NRPB design closed passive diffusion alpha-track type radon detectors, where the alpha track registration medium is CR39 and the housing shell is a conducting plastic. Both seasonal variations and annual average radon levels in the dwellings were determined. The radon calibration factors, in the range of 2–3 tracks cm  2 kBq  3 h, were determined by participation in a number of NRPB (UK) radon intercomparison exercises. Track counting was made by means of both automatic image analysis and manual optical microscopy. In order to obtain a realistic value of the mean annual indoor radon concentration in the chosen dwellings, a sequential series of detectors were placed in one or more living areas (usually living and bedroom) over a twelve month period. Each single detector remained exposed in a dwelling for a period of three months and was then immediately replaced by a new detector. In this way four detectors were placed at each chosen site thus

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

332

Fig. 3. Geological map of the study area in Montenegrin coast. A. Autochthone; B. Budva (Cukali) zone; C. High karst zone. 1. Flysch, conglomerate, sandstone, shale and marl; 2. Numulite limestone; 3. Thick-bedded and bedding limestone with dolomite interlayers and lenses; 4. Transition beds in flysch bottom and flysch-marly limestone, calcarenite and marl (Lower Eocene); 5. Calcarenite and micrite with chert interlayer; 6. Calcarenite, micrite, oolitic limestone, chert and dolomite; 7. Dolomite, dolomitic limestone and limestone; 8. Calcarenite, micrite with dolomite interlayers (Upper Triassic); 9. Porphyrite and diabase; 10. Thick-bedded to massive, rarely bedding limestone; 11. Flysch: conglomerate, graywacke and marl. 12. Faults, certain (solid line), inferred (dashed line).

Table 1 The parameters of linear regression for Montenegrin coast. The uncertainties are expressed at the confidence level of 95%. No.

Measurements period

A (slope)

B (intercept)

R2

I II III

26.05.1997–29.09.1997 29.09.1997–19.02.1998 19.02.1998–28.05.1998

1.17 0.5 0.687 0.10 0.75 7 0.10

107 12 0.3 7 4.0 127 3

0.42 0.86 0.86

yielding both season and annual average radon concentration values. In a small number of dwellings when a detector could not be recovered at the end of the exposure period, due to absence of a resident, the unrecovered detector was allowed to remain for a further 3 months until it was recovered. The detectors were deployed usually on cupboards, shelves, at least 20–30 cm far from the wall, and in a very few cases on the wall due to the lack of furniture. Attention was paid also to choosing rooms on ground floors, to avoid closeness of the deployed detectors to electrical devices or to windows or other locations where higher air exchange rates can be expected. 2.3. Data evaluation When the measurements did not last one year exactly, two types of correction were applied: one in the case when the total campaign of measurement lasted less than one year and another one when the campaign lasted longer than one year. When the total duration of the measurement campaign lasted for more than one year like in case of Kalna, weighting factors were used to determine mean values of different periods (seasons) of measurements in order to correctly estimate average annual concentration. In the case of Kalna the total period of the measurement campaign

is longer than one year, whereby the fourth period closes the total measurement of one year, but also extends into March, April, May, and therefore partly covers the same season of the year as the first period. (see Table 9 for the periods duration). Thus, the overlapping part of first and fourth period concentration is estimated as the weighted mean of the fourth and first period concentrations, where the weight is the number of days of the overlapping of these two periods. In certain measurements the detectors were lost and we have no value of the mean radon concentration of the corresponding season. In these cases a simple average value of the available season measurements would not give reliable estimation of the average annual concentration due to the fact that the indoor radon concentration is season dependent. In order to evaluate annual radon concentration we assume a linear relation between mean annual radon concentration Cann (calculated for the cases in which measurements from all four seasons were available) and the radon concentration Ci measured during the season i only (i¼ {I, II, III, IV} – see Table 9 for instance). The relation is estimated by regression of Cann against Ci, and used as a model for estimating Cann from available Ci, if not all four Ci are available. Since the concentrations are strongly positively skewed distributed, performing simple linear regression is statistically not optimal, because high values act as high leverage data and can introduce a bias in the estimated regression coefficients. Also, homoscedasticity of the residuals (a condition for Gaussian LSQ regression) is not fulfilled. Therefore we chose to define y¼ln(Cann) of the dependent variable and perform non-linear regression according to the model y¼ln(AxþBþs), with x¼ Ci, the independent variable. The logarithm rescales the variable to approximately normal. Regression was performed by the quasi-Newton method and LSQ as implemented in Statistica software. R as given in Fig. 4 is the coefficient of determination (fraction of variance explained by the model,

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

333

Table 2 The results of measurements of the parameters of the annual indoor radon concentration in Montenegrin coast houses. The percentage above 200 Bq/m3 has been calculated from the lognormal distribution whose parameters are given in the table. The superscript and subscript of the geometric mean presents the standard deviation (one sigma) of the lognormal distribution i.e. there is 68.2% of the probability that a measured radon concentration will be in this interval. Number of samples

Arithmetic mean

Geometric mean

Min Max 1st 3rd Percent quartile quartile above 200 Bq/m3

82

45

28 36 þ  15

13

177

25

54

0.37

Table 3 The statistic parameters for samples of individual houses and public buildings at Montenegrin coast presented separately on the basis of the results of t-test.

Fig. 4. Estimation of the annual concentration depending on the measured seasonal concentration. Example shown for data of Gornja Stubla and i ¼third season (autumn ) Uncertainties of A and B are given7 95% confidence limits (in brackets).

accounting for degrees of freedom) and the confidence interval estimated from the assumption of asymptotic Gaussian distribution of residuals. The annual concentration Cann in one room is estimated as the weighted mean of all Cann(i) estimated from all available season measurement Ci, whereby the weighting factor is the square of the inverse of the total confidence interval of the given fit of Cann(i). An example is given in Fig. 4, which shows the linear regression between Cann and CIII (the concentrations for season III – autumn) for Gornja Stubla (see later). It was found that this correction is very important since it was realized that simple averaging could lead to errors of up to 100%.

3. Results In this section the measurements carried out at the Montenegrin coast, Kalna and Gornja Stubla will be described in three subsections. 3.1. Montenegrin coast The radon survey in houses at the Montenegrin coast was carried out in 82 dwellings in 11 settlements. Three measurements together cover a completed year-long period, 26.05.1997–28.05.1998. The annual radon concentration is estimated by these three measurements, accounting for the duration of the exposure of the detectors. In this way the annual radon concentrations were estimated for 62 dwellings. Some measurements were missing in 20 dwellings. In order to evaluate annual radon concentration in these 20 dwellings the dependence of assessed annual radon concentration Cann on seasonal measurements Ci was estimated as has been explained in Section 2.3. The results of linear regression parameters (A and B) estimation and coefficients of determination R are presented in Table 1. The annual radon statistics for all 82 data points are shown in Table 2. Statistical parameters for samples of individual houses and other buildings are presented in Table 3 separately. In order to evaluate the differences in means between the two groups: “individual houses” and “public buildings”, the t-test is used. The test showed a statistically significant difference between two means of logarithms, meaning that these two groups must be considered separately.

Sample, Number of samples

Arithmetic mean

63 Individual houses, N¼ 18 40 Public buildings, N¼ 64

Geometric mean

Min Max 1st 3rd Percent quartile quartile above 200 Bq/m3

32 50 þ  18

18

141

33

71

2.2

26 32 þ  14

14

178

20

40

0.12

Table 4 The parameters of linear regression for Gornja Stubla for the periods from April 21, 1997 to April 16, 1998. The uncertainties are expressed at the confidence level of 95%. No.

Measurements period

A (slope)

B (intercept)

R2

I II III IV

21.04.1997–18.07.1997 18.07.1997–30.10.1997 30.10.1997–09.01.1998 09.01.1998–16.04.1998

1.2 70.3 0.94 70.16 0.65 70.11 0.93 70.12

20 760 90 740 60 740 20 730

0.72 0.87 0.86 0.92

Table 5 The parameters of the lognormal distribution of the indoor radon concentration in Gornja Stubla depending on the type of bedrock. The geometric means of the samples are shown with their uncertainties defined as σ/√N where N is the number of measurements. The results of Fisher LSD test are shown, where the significant differences of means with the probabilities better than 0.05 level of confidence are bolded and underlined. No. Bedrock

Geometric Mean (Bq/ m3)

40 370 þ  30

(1)

Argillite

(2)

Diabase

(3)

Trachyte 680 þ 240  180 Flysch 320 þ 30

(4)

18 116 þ  16

 20

Means difference probability (Fisher LSD test) (1)

(2)

–

o 0.001 o 0.001 0.60

o 0.001 –

(3)

o 0.001 0.009

o 0.001 o 0.001 – 0.60

0.009

(4)

o0.001

o 0.001 –

The statistical hypothesis that observed data of radon survey follow the log-normal distribution was tested by a Chi-square test which is not significant (p4 0.05, χ2: 24.8, probability: 0.47). Due to the small statistics, it was not possible to perform a Chi-square test for individual houses, while from the test of public houses one could accept the statistical hypothesis that the data follow a lognormal distribution.

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

334

in Section 2.3. The parameters of this linear regression are presented in Table 4. The average annual radon concentration in houses was estimated by averaging the values for rooms (65 houses). The houses are grouped according to the bedrock type, and results are shown in Table 5. The one-way analysis of variances (ANOVA) Kruskal–Wallis test shows a very high influence of geological factors on the variability of indoor radon. The Kruskal–Wallis test gave H¼108 and Prob(4H) o0.0001, rejecting the null hypothesis that the samples are coming from the same population with a 0.05 level of confidence. The Fisher LSD test shows that means of flysch and of argillite do not differ at the 0.05 level of confidence (Table 5). All other pairs of the means are significantly different. This difference between bedrocks is the most probable reason why the indoor radon distribution in Gornja Stubla does not follow the lognormal distribution as explained by a chisquare test (χ2: 26.6, probability:0.014). Although in the majority of houses no difference between radon concentrations in different rooms of the same house was found (ANOVA test provided F¼0.81, p¼0.52), several exceptions are remarkable (Table 6). Undoubtedly, the radon concentrations in the rooms of the same house can differ by more than one order of magnitude, even if they are on the same floor. Possible influence of thoron on the radon measurements could not explain such high difference considering the thoron concentrations (Zunic et al., 2000) and the thoron calibration coefficient of the SSI/NRPB detectors (Tokonami, 2005). We noticed a similar situation in the case of Kalna (see next subsection).

3.2. Gornja Stubla The measurements in Gornja Stubla were conducted in three different groups of houses during different periods of time (50 rooms in the period from April 21, 1997 to April 16, 1998; 54 rooms in the period from October 30, 1997 to February 20, 1999 and 68 rooms in the period from March 14, 1998 to December 21, 1998). The measurements of the last group of houses did not cover a period of one year. In order to assess annual radon concentrations in these rooms the weighted mean of linear regression between seasonal radon concentrations (Ci) measured in the other two groups and their estimated annual radon concentrations (Cann) was used as is explained Table 6 Radon concentrations in houses in Gornja Stubla where the concentrations in the rooms in the same period of measurement differ for order of magnitude. The values printed in bold are the values of concentration which are at least one order of magnitude higher than the minimum value. All rooms are on the same floor of the given house. House Room No

Period of exposure 21.04.0 97– 18.07.0 97 (Bq/m3)

18.07.0 97– 30.10.0 97. (Bq/m3)

30.10.0 97– 09.01.0 98. (Bq/m3)

09.01.0 98.– 16.04.‘98. (Bq/m3)

Annual mean (Bq/m3)

(1)

1 2 3 4 5 6

893 307 319 481 646 2674

468 494 283 111 808 2909

871 490 203 589 3348 8864

638 459 321 456 1332 9591

718 438 282 409 1534 6010

(13)

1 1 2 3

2033 2198 132 188

3524 2263 84 153

3467 3434 265 301

2418 2410 168 214

2861 2576 162 214

(16)

1 1 2

303 404 888

343 266 804

440 217 1493

612 168 2164

425 264 1337

(32)

1 2

1289 448

867 230

1164 137

1064 589

1096 351

(48)

1 2 3

248 524 2096

517 416 1673

N.A.

N.A.

383 470 1885

3.3. Kalna In the Kalna region, 11 settlements were investigated (Fig. 1): Inovo, Kalna, Stara Kalna, Belevica, Mezdreja, Ravno Bucje, Janja, Gabrovnica, Donje Polje, Balta-Berilovac and Vrtovac. The measurement covers in total 454 days (the periods of the measurement are shown in Table 9). The average radon concentration is estimated in 103 rooms of 75 houses. Among these 75 houses, there are 27 houses where more than one room was surveyed. The descriptive statistics of the measurement in the Kalna region is presented in Table 7. In 21 rooms some measurements were skipped or the exposure of a detector continued during two periods. Since the total period

Table 7 The parameters of the indoor radon concentrations for Kalna rural region. Number of samples

Arithmetic mean

Geometric mean

Min

Max

1st quartile

3rd quartile

Perc. above 200 Bq/m3

Perc. above 400 Bq/m3

75

178

120 150 þ  70

29

673

97

214

30.7

5.0

Table 8 The parameters of the lognormal distribution of indoor radon in Kalna depending on the type of bedrock. The results of Fisher LSD test are shown, where the significant differences of means with the probabilities better than 0.05 level of confidence are bolded and underlined. Herein, the geometric means of the samples are shown with their uncertainties (superscript and subscript) defined as σ/√N where N is the number of measurements. No.

Bedrock

(1)

Conglomerates, limestone, sandstone, claystone (upper urgon facies)

(2)

Conglomerate, limestone, clayey limestone, shale (lower urgon facies)

(3)

Alluvium on the conglomerates riverbed

(4)

Granites of Ravno Bučje

(5)

Gabbroids

(6)

Granites of Janja

(7)

Alluvium on argilo-schist and sandstone riverbed

(8)

Sandstone and limestone

Geometric Mean (Bq/m3)

40 140 þ  30 30 140 þ  20 60 200 þ  50

30 60 þ  20

50 150 þ  40 20 214 þ  18 22 150 þ  19 13 75 þ  11

Means difference probability (Fisher LSD test) (1)

(2)

(3)

–

0.66

0.30

0.66

–

0.22

0.30

0.22

–

0.046

0.12

0.71 0.043

(4)

(5)

(6)

(7)

(8)

0.046

0.71

0.046

0.35

0.06

0.12

0.96

0.046

0.50

0.13

0.015

0.32

0.58

0.46

0.02

0.015

–

0.21

0.002

0.03

0.90

0.96

0.32

0.21

–

0.14

0.59

0.20

0.046

0.58

0.002

0.14

–

0.11

0.005

0.35

0.50

0.46

0.03

0.59

0.11

–

0.045

0.06

0.13

0.02

0.90

0.20

0.005

0.045

–

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

of the campaign of the measurement is longer than one year the weighted mean was applied in order to estimate correctly the average annual concentration. The procedure has been described in Section 2.3. There is no significant difference between radon concentration in living rooms and bedrooms (except in two houses). The indoor radon distribution in the Kalna area follows the lognormal distribution (χ2:14.64, probability: 0.19). The results of statistical analysis of indoor radon annual concentration of Kalna region depending on different bedrocks are presented in Table 8. A significant influence of geological factors on the variability of indoor radon was found. The Kruskal– Wallis test resulted in a conclusion that the populations of indoor radon concentration on different bedrock are significantly different are significantly different at the 0.05 level (H ¼91 and Prob (4 H) o0.0001). The Fisher LSD test shows that there is a means difference at the 0.05 level of confidence between the pairs of bedrock shown in Table 8. In the case of 8 rooms of 7 houses, unusual excesses in the radon concentration in one period compared to the other three were found. The concentrations were an order of magnitude higher than the others. In a first evaluation (Zunic et al., 2001) these results were disregarded. However, in this work the values have been included for the following reasons. All detectors (from Gornja Stubla, Montenegro and Kalna) are the same type of the detectors and were handled, transported and etched under the same conditions, whereby all these measurements were mainly conducted during the period 1997–1998. In all of these measurements (Kalna 406, Gornja Stubla 541 and Montenegro 296) these 8 (or 7 if we count houses, not only rooms) excesses were found in three settlements of the Kalna region: Stara Kalna, Belevica and Balta-Berilovac; where in total 97 measurements were conducted, where two of these villages (Stara Kalna and Belevica) are placed at the southern and northern side of the same hill – Jankin plast. If the occurrence of all of these excesses was uncorrelated, the probability to have them all only in these three villages would be less than 10  6. Additionally, in house number 23 two detectors were placed in two adjacent rooms – the living room and the niche, being effectively placed in the same space. Both these detectors measured the same unusually high concentration in the same time period. The anomalous increase of indoor radon concentrations are shown in Table 9. The excesses in one period of measurement is an order of magnitude higher, when compared with the others periods. Such behavior is not noticed on the Montenegrin coast, nor in Gornja Stubla where the maximum excess was a factor of 4 or less, and only in one case was a difference factor of 5 noticed. The discrepancies between the radon concentrations in different rooms of the same house were noticed (see Table 9) in several cases, as in the case of Gornja Stubla.

335

4. Discussion High radon values were recorded within the terrains with uranium bearing formations, near Kalna and Gornja Stubla, while the limestone region of the Montenegrin coast shows low radon concentrations. The distributions of permeable and impermeable formations, as well as the presence of enhanced collectors of groundwater play an important role on the radon potential of the area. In the Kalna and Gornja Stubla areas, which show different geological characteristics, the indoor radon concentrations seem to be correlated with type of bedrock over which the houses were built. In the case of Gornja Stubla high indoor radon concentrations occurs on trachyte bedrock (Table 5). Trachyte is relatively rich in alkali feldspar and moderately rich in quartz, i.e. moderately felsic (acidic). It is the analog to syenite for plutonic rocks. In the course of their formation, they also develop a porosity, which could lead to higher permeability. Uranium has a moderate affinity to these rocks for geochemical reasons (in the course of magmatic differentiation, as a lithophile element U tends toassociate with the silica-rich phases, i.e. quartz and alkali-feldspar). Considering the position of trachyte in the QAPF/Streckeisen diagram, field 7, e.g. www.quartzpage.de/app_qapf.html, leads to the conclusion (See also Cothern and Smith (1987, J. Michel, chapter 4.2, p.83ff)). Diabase on the other hand is a mafic rock of andesite, the analog

Fig. 5. Distributions of indoor radon concentrations in three geologically diverse background regions. Kruskal–Wallis test shows a significant difference between these three populations (H ¼108 and Prob(o H) ¼0.0001).

Table 9 The houses with anomalous seasonal excess of indoor radon concentration. The concentrations are in Bq/m3. The excesses are printed as bold. The measurements are performed in living rooms (L), bedrooms (B) and in niche (N). House no.

19 20 22 23 66 70 74

Period of exposure

Annual mean (Bq/m3)

I 08.03.1997-06.07.1997 (Bq/m3)

II 06.07.1997-23.10.1997 (Bq/m3)

III 23.10.1997-13.02.1998 (Bq/m3)

IV 13.02.1998-05.06.1998 (Bq/m3)

L 1740 2220 47 1050 102 250 140

L 150 115 80 73 42 2010 150

L 340 180 86 170 330 310 200

L 150 61 75 46 50 90 120

B

N

150 820

96

B

N

1500 100

1340

B

N

68 100

140

B

N

34 36

85

617 673 249 312 131 647 277

336

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

to diorite and gabbro for plutonites. These are usually low in uranium for geochemical reasons and possibly less permeable. This may explain the significantly higher RP of trachyte vs. diabase. Argillites are difficult to predict in terms of RP, because clay (¼ feldspar) can contain relatively high U concentrations, on the other hand they are often quite impermeable. Similarly, flysch (compacted clayey marine sediment) shows an intermediate high variable RP as it can be derived from different rocks. Data collected in the Montenegrin coast area, mainly consisting of different types of limestone (marly limestone, karstified limestone, conglomerates of limestone, sandstone and shales, river sediment) do not suggest any dependence between indoor radon and lithology. Fig. 5 highlights the differences among the lognormal statistical distributions of indoor radon concentrations of the three investigated regions, Montenegrin coast, Kalna and Gornja Stubla. It is known that the indoor radon concentration can change by more than one order of magnitude during a short period of time (hourly, daily) (Robé et al., 1992; Miles, 2001). These changes are even more pronounced in granite regions (Robé et al., 1992). Usually monthly variation does not exceed a factor of 3 (GrovesKirkby et al., 2006). In rare cases the absence of inhabitants due to vacation can cause an increase of the indoor radon concentration by several times (Miles, 2001). However, the seasonal variation winter/summer reaches a factor of 3–5 (Cohen and Gromicko, 1988; Krewski et al., 2004; Miles et al., 2012). In some cases, the seasonal change of the indoor radon concentration could be caused by the habits of the inhabitants. On the other hand, such high seasonal variation, as found in our survey, is very unusual. Such a measurement must be treated carefully and should not be easily rejected under an assumption of unreliable measurement. For instance, in the case of house number 20 in Kalna (see Table 9), if the increased value in one period of measurement would be erroneously rejected the estimated effective dose of inhabitants would have been 3.5 mSv/year instead of the real 26.5 mSv/year which is more than 10 times higher than the average annual dose. Moreover, both in the case of Gornja Stubla and Kalna, it was noticed that the radon concentration in the rooms of the same house can differ by more than an order of magnitude, even if they are on the same floor. For example in the case of Gornja Stubla, if the indoor radon concentration would not be measured in all rooms of house number 1 (see Table 6), the estimated effective dose for the inhabitants of this house would have been at least 5 times lower than the real one 28.5 mSv/year instead of 133 mSv/year. We also suggest the need to conduct further investigations in houses where this phenomenon occurs. Also remediation should be considered given radon levels which are intolerably high in some cases, from the point of view of radiation protection. Once a physical reason has been identified, one may also think on particular, more intensive monitoring of houses of a similar characteristic. Such unusual phenomena could be a consequence of many natural and anthropogenic causes: different airing behaviors of inhabitants, ventilation type, heating system, existence or nonexistence of microcracks in building materials and their connectivity with an enhanced radon collector in the bedrock. Nonetheless, the behavior of inhabitants can explain only partially the seasonal excesses and hence we propose to explain it by the geological aspects of the bedrock. After all, it is unlikely that living habits are so heterogeneous over the region. In the Kalna area, migration of radioactive water and radon occurrences (from the U deposit and adjacent rocks) formed by the erosion of several U deposits, as well as radon occurrences from the contact of the Janja granite and Paleozoic schist, were impossible or difficult because of the barrier of impermeable schist. That is why the fault zone to the NW, within the permeable Miocene conglomerate and Mesozoic limestones, is defined as an enhanced collector

for radioactive water and radon drainage (Fig. 1). Groundwater and radon from the collector are discharged into the Stara Kalna fault, within the limestone formation, as well as into the Balta Berilovac fault, which is known as a tectonic contact between schist and red sandstone (Fig. 1). The anomalous high indoor concentrations, during one season, were measured in houses located in faulted zones. Two villages Stara Kalna and Belevica are located along the same NNW–SSE fault, while Balta-Berilovac is located on a NW–SE fault (Fig. 1). The causes of the occurrence of these seasonal peaks in radon concentration are a very complex interaction among geological structure, hydrogeological characteristics of rocks, conditions of recharge, movement and discharge of groundwater, change of precipitation, temperature and atmospheric pressure during the year, changes in temperature and chemical composition of ground water over time. It is possible that the mentioned faults (one in Stara Kalna and Belevica and the other in Balta-Berilovac) constitute the main collector of radioactive waters from uranium deposits. Dependence of the radon presence in the shallow environment and thus in the indoor environment on the mentioned parameters is known from the basics of geology and hydrogeology. For each case, the most important parameters are the amount of precipitation and recharge of the aquifer in the area of the U deposit, as well as precipitation distribution during the year. Peak concentrations of radon are related to intensive precipitation after the dry season. In this case the increase of seasonal indoor radon concentration could be the consequence of a pressure shock inside the main collector (possibly the fault itself), when the discharge of the deepest reserves of the radon enriched water occurs. This event can be induced by a sudden breach of artesian (mining) water caused by some of the meteorological factors like intensive precipitation, fast circulation through the mentioned enhanced collectors or degradation of the uranium deposits. If the floor of the building, which is in contact with the ground, has small cracks which connect the fault and a certain room in the house, it could explain the high radon concentration in that room compared it to the concentration in other rooms. This hypothesis can be tested by conducting soil gas investigations, which are able to detect such spots of radon entry. If the geogenic RP is low, as at the Montenegrin coast, spots of preferential radon infiltration would have little effect and would therefore probably not lead to the effects as observed in the high-radon areas.

5. Conclusion In this paper, a method to estimate the annual mean concentration when one seasonal measurement is missing. In such cases, the parameters of linear regression between seasonal and annual average concentration were used for the estimation of average annual concentration. In fact, in the case of Kalna, it was realized that a simple averaging of available seasonal measurement could lead to errors of up to 100% in an estimation of the average annual concentration. In this paper, the correlation between the type of the bedrock and indoor radon level could be used as a tool of a prediction of areas prone to high indoor radon concentration. This fact is very important because it means that the regional radon hazard can to some extent be predicted from geology, which is however in line with what is known from increasing literature about the subject. As a practical consequence denser indoor Rn surveys should be performed in such regions. The finding also suggests that geological information should be exploited more as a means of identifying radon prone areas. Another important conclusion is that anomalously high seasonal increase of indoor radon concentration of an order of magnitude is

Z.S. Žunić et al. / Applied Radiation and Isotopes 94 (2014) 328–337

possible. This phenomenon has been noticed only in Kalna, but not in the Montenegrin coast or Gornja Stubla, although Gornja Stubla is also an uraniferous area. The existence of such excesses indicates that in such a region indoor radon has to be monitored for a longer period of time of at least one year. The high difference of an order of magnitude of the indoor radon in the different rooms of the same floor proved that this phenomenon is possible, and it was noticed both in Kalna area and in Gornja Stubla. This behavior was noticed in both uranium bearing regions – Gornja Stubla and Kalna. We want to stress that such results should be taken seriously and not be rejected as implausible outliers. The fact implies that in order to estimate correctly the exposure of inhabitants in uraniferous regions every room with significant occupation time (living and sleeping rooms and kitchens mostly) should be investigated. This study should be considered as a kind of first screening which revealed certain problems and anomalies. Further elucidation requires specialized investigations about the characteristics of the affected houses, of (hydro-)geological peculiarities of the region and possibly on the dependence of weather (precipitation, atmospheric pressure etc.).

Acknowledgments The authors greatly appreciate the financial support within the current Projects 171018, 171028 and 41028 of the Ministry of Education, Science and Technological Development of Republic of Serbia. References Appleton, J.D., Miles, J.C.H., 2010. A statistical evaluation of the geogenic controls on indoor radon concentrations and radon risk. J. Environ. Radioact. 101 (10), 799–803. Appleton, J.D., Miles, J.C.H., Young, M., 2011. Comparison of Northern Ireland radon maps based on indoor radon measurements and geology with maps derived by predictive modelling of airborne radiometric and ground permeability data. Sci. Total Environ. 409 (8), 1572–1583. Bossew, P., Dubois, G., Tollefsen, T., 2008. Investigations on indoor radon in Austria, part 2: geological classes as categorical external drift for spatial modelling of the radon potential. J. Environ. Radioact. 99, 81–97. Ciotoli, G., Guerra, M., Lombardi, S., Vittori, E., 1998. Soil gas survey for tracing seismogenic faults: a case study in the Fucino basin, Central Italy. J. Geophys. Res.: Solid Earth (1978–2012) 103 (B10), 23781–23794. Cohen, B.L., Gromicko, N., 1988. Variation of radon levels in U.S. homes with various factors. JAPCA 38 (2), 129–134. Environmental Radon. In: Cothern, C.R., Smith, J.E. (Eds.), Environmental Science Research, vol. 35. Plenum Press, New York and London (ISBN 0-306-42707-9). Gillmore, G.K., Phillips, P.S., Denman, A.R., 2005. The effects of geology and the impact of seasonal correction factors on indoor radon levels: a case study approach. J. Environ. Radioact. 84, 469–479. Groves-Kirkby, C.J., Denman, A.R., Crockett, R.G.M., Phillips, P.S., Woolridge, A.C., Gillmore, G.K., 2006. Time-integrating radon gas measurements in domestic premises: comparison of short-, medium- and long-term exposures. J. Environ. Radioact. 86, 92–109. Gruber, V., Bossew, P., De Cort, M., Tollefsen, T., 2013. The European map of the geogenic radon potential. J. Radiol. Protect. 33 (1), 51. Harrison, J.D., Marsh, J.W., 2012. Effective dose from inhaled radon and its progeny. Ann. ICRP 41 (3–4), 378–387. Ielsch, G., Cushing, M.E., Combes, P., Cuney, M., 2010. Mapping of the geogenic radon potential in France to improve radon risk management: methodology

337

and first application to region Bourgogne. J. Environ. Radioact. 101 (10), 813–820. Jakupi, B., Krstic, G, Tonic, M, Ilic, R., 1997. Radon in mines and dwellings in Kosovo and Metohia. Radiat. Meas. 28 (1-6), 691–694. Jankovic, S., 1965. Pb–Zn deposits in Yugoslavia. Rud. Glas. 2, 31–62 (Beograd (in Serbian)). Jankovic, S., 1990. Ore Deposits of Serbia, Special edition Faculty of Mining and Geology, Belgrade, Serbia, p. 760 (in Serbian). Kemski, J., Klingel, R., Siehl, A., Stegemann, R., 2005. Radon transfer from ground to houses and prediction of indoor radon in Germany based on geological information. In: McLaughlin, J.P., Simopoulos, S.E., Steinhäusler, F. (Eds.), Radioactivity in the Environment. The Natural Radiation Environment, vol. 7. Elsevier, Amsterdam, pp. 820–832. Kemski, J., Klingel, R., Siehl, A., Valdivia-Manchego, M., 2009. From rad on hazard to risk prediction-based on geological maps, soil gas and indoor measurements in Germany. Environ. Geol. 56, 1269–1279. King, C.Y., Minissale, A., 1994. Seasonal variability of soil-gas radon concentration in central California. Radiat. Meas. 23 (4), 683–692. King, C.Y., King, B.S., Evans, W.C., Zhang, W., 1996. Spatial radon anomalies on active faults in California. Appl. Geochem. 11 (4), 497–510. Komatina, M., 2004. Medical geology. Effects of Geological Environments on Human Health, vol. 2. Elsevier, London, UK (488 pp.). Komatina-Petrovic, S., 2011. Environmental Geophysics. Geophysics and Environmental Protection. AGES, Beograd, pp. 1–362 (in Serbian). Krewski, D., Mallick, R., Zielinski, J.M., Létourneau, E.G., 2004. Modeling seasonal variation in indoor radon concentrations. J. Expo. Sci. Environ. Epidemiol. 15 (3), 234–243. Miles, J.C.H., 2001. Temporal variation of radon levels in houses and implications for radon measurement strategies. Radiat. Prot. Dosim. 93 (4), 369–375. Miles, J.C.H., Appleton, J.D., 2005. Mapping variation in radon potential both between and within geological units. J. Radiol. Prot. 25, 257–276. Porstendörfer, J., Butterweck, G., Reineking, A., 1994. Daily variation of the radon concentration indoors and outdoors and the influence of meteorological parameters. Health Phys. 67 (3), 283–287. Robé, M.C., Rannou, A., Bronec, J.L.e., 1992. Radon measurement in the evironment in France. Radiat. Prot. Dosim. 45, 455–457. Scheib C., Appleton J.D., Miles J.C.H., Green B.M.R., Barlow T.S., Jones D.G., 2009, Geological controls on radon potential in Scotland. Scot. J. Geol. 45(2):147–160. Shi, X., Hoftiezer, D.J., Duell, E.J., Onega, T.L., 2006. Spatial association between residential radon concentration and bedrock types in New Hampshire. Environ. Geol. 51, 65–71. Sundal, A.V., Henriksen, H., Soldal, O., Strand, T., 2004. The influence of geological factors on indoor radon concentrations in Norway. Sci. Total Environ. 328, 41–53. Tokonami, S., 2005. Summary of dosimetry (radon and thoron) studies. Int. Congr. Ser. 1276, 151–154. UNSCEAR, 2000. Sources and Effects of Ionizing Radiation. United Nations Scientific Committee on the Effects of Atomic Radiation, New York. Zhu, H.C., Charlet, J.M., Poffijn, A., 2001. Radon risk mapping in southern Belgium: an application of geostatistical and GIS techniques. Sci. Total Environ. 272, 203–210. Zunic, Z.S., Mc Laughlin, J.P., Walsh, C., Benderac, R., 1999. The use of SSNTDS in the retrospective assessment of radon exposure in high radon rural communities in Yugoslavia. Radiat. Meas. 31 (1–6), 343–346. Zunic, Z.S., Fujimoto, K., Mclaughlin, J.P., Birovljev, A., 2000. Field experience of indoor thoron gas measurements in a stable rural community in Yugoslavia. In: Proceedings of the 10th International Congress of the International Radiation Protection Association Harmonization of Radiation, Human Life and the Ecosys tem, IRPA10, pp. 14–19. Zunic, Z.S., McLaughlin, J.P., Walsh, C., Birovljev, A., Simopoulos, S.E., Jakupi, B., Gordanic, V., Demajo, M., Trotti, F., Falk, R., Vanmarcke, H., Paridaens, J., Fujimoto, K., 2001. Integrated natural radiation exposure studies in stable Yugoslav rural communities. Sci. Total Environ. 272 (1–3), 253–259. Žunić, Z.S., Kozak, K., Ciotoli, G., Ramola, R.C., Kochowska, E., Ujić, P., Čeliković, I., Mazur, J., Janik, M., Demajo, A., Birovljev, A., Bochicchio, F., Yarmoshenko, I.V., Kryeziu, D., Olko, Olko, P., 2007a. A campaign of discrete radon concentration measurements in soil of Niška Banja town, Serbia. Radiat. Meas. 42 (10), 1696–1702. Zunic, Z.S., Yarmoshenko, I.V., Birovljev, A., Bochicchio, F., Quarto, M., Obryk, B., Paszkowski, M., Celikovic, I., Demajo, A., Ujic, P., Budzanowski, M., Olko, P., McLaughlin, J.P., Waligorski, M.P.R., 2007b. Radon survey in the high natural radiation region of Niska Banja, Serbia. J. Environ. Radioact. 92 (3), 165–174.