Journal of Environmental Radioactivity 124 (2013) 57e67

Contents lists available at SciVerse ScienceDirect

Journal of Environmental Radioactivity

journal homepage: www.elsevier.com/locate/jenvrad

An approach to define potential radon emission level maps using

indoor radon concentration measurements and radiogeochemical data
positive proportion relationships
Jean-Philippe Drolet a, *, Richard Martel a, Patrick Poulin b, Jean-Claude Dessau c,
Denis Lavoie d, Michel Parent d, Benoît Lévesque b
a
Institut national de la recherche scientifique, Eau Terre Environnement Research Centre (ETE-INRS), 490 de la Couronne, G1K 9A9 Quebec, Canada
b
Institut national de santé publique du Québec (INSPQ), 945 Avenue Wolfe, G1V 5B3 Quebec, Canada
c
Agence de la santé et des services sociaux des Laurentides, 1000 rue Labelle, J7Z 5N6 Saint-Jérome, Canada
d
Geological Survey of Canada, 490 de la Couronne, G1K 9A9 Quebec, Canada

a r t i c l e i n f o a b s t r a c t

Article history: The aim of this paper is to present the first step of a new approach to make a map of radonprone areas
Received 28 September 2012 showing different potential radon emission levels in the Quebec province. This map is a tool intended to
Received in revised form assist the Quebec government in identifying populations with a higher risk of indoor radon gas exposure.
11 April 2013
This map of radon-prone areas used available radiogeochemical information for the province of Quebec:
Accepted 12 April 2013
(1) Equivalent uranium (eU) concentration from airborne surface gamma-ray surveys; (2) uranium
Available online
concentration measurements in sediments; and (3) bedrock and surficial geology. Positive proportion
relationships (PPR) between each individual criterion and the 1417 available basement radon concen-
Keywords:
Radon mapping
trations were demonstrated. It was also shown that those criteria were reliable indicators of radon-prone
ANOVA areas. The three criteria were discretized into 3, 2 and 2 statistically significant different classes
Geology respectively. For each class, statistical heterogeneity was validated by KruskaleWallis one way analyses
Airborne survey of variance on ranks. Maps of radon-prone areas were traced down for each criterion.
Uranium Based on this statistical study and on the maps of radon-prone areas in Quebec, 18% of the dwellings
Sediment located in areas with an equivalent uranium (eU) concentration from airborne surface gamma-ray sur-
veys under 0.75 ppm showed indoor radon concentrations above 150 Bq/m3. This percentage increases
to 33% when eU concentrations are between 0.75 ppm and 1.25 ppm and exceeds 40% when eU con-
centrations are above 1.25 ppm. A uranium concentration in sediments above 20 ppm showed an indoor
radon concentration geometric mean of 215 Bq/m3 with more than 69% of the dwellings exceeding
150 Bq/m3 or more than 50% of dwellings exceeding the Canadian radon guideline of 200 Bq/m3. It is
also shown that the radon emission potential is higher where a uranium-rich bedrock unit is not covered
by a low permeability (silt/clay) surficial deposit.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction levels. Because radon gas has a higher density than air, it tends to
accumulate in basements, the lowest dwelling level and also the
Radon is a naturally occurring noble gas which is colourless, least ventilated (Dessau et al., 2004).
odourless and tasteless. It is found in soil and rocks and drifts up- There are three natural radon isotopes, but radon-222 (222Rn) is
ward from the ground to the outdoor air. As a highly volatile gas, the most stable and decays with a half-life of 3.823 days. 222Rn is
radon is quickly diluted to harmless levels in exterior environ- produced by the radioactive decay of radium-226 (226Ra), both
ments. However, it can also infiltrate dwellings through cracks or daughter elements of uranium-238 (238U). 220Rn and 219Rn (decay
holes in the foundations and can reach high indoor concentration products of thorium-234 and uranium-235 respectively) are of no
concern from an air quality point of view because their half-lives
are short (219Rn ¼ 3.96 s and 220Rn ¼ 55.6 s) (Nambi and Aitken,
* Corresponding author. Tel.: þ1 418 654 2530; fax: þ1 418 654 2600. 1986) and as they cannot move very far. Because of its longer
E-mail address: jean-philippe.drolet@ete.inrs.ca (J.-P. Drolet). half-life, 222Rn can relocate from soil to indoor air before emitting a

0265-931X/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.jenvrad.2013.04.006
58 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67

high energy alpha particle of 5.49 MeV (Nero et al., 1990). Daughter Table 1
products of 222Rn are solid elements and may attach to dust and Redundancy of the most used indoor radon mapping criteria surveyed from 62 in-
ternational studies (surveyed studies are listed in Drolet, 2011).
aerosol particles.
This radioactive dust may be deposited on epithelial tissue of Radon mapping criteria Redundancies: proportion of
the lungs when inhaled. Radon progeny, especially polonium-218 studies using the criteria (%)

and polonium-214, also emit high energy alpha particles Indoor radon concentrations 89%
(6.00 MeV for 218Po and 7.69 MeV for 214Po) and decay with short Bedrock units 66%
Surficial deposits 35%
half-lives (218Po ¼ 3.04 min and 214Po ¼ 164 ms) (Nero et al., 1990).
Equivalent uranium (eU) 34%
Epidemiologic studies concluded that alpha particles emitted from concentration from surface
radon gas and from radon progeny can alter DNA of the cells lining gamma-ray measurements
lungs and may lead to lung cancer. Radon is considered to be the Soil gas radon concentration 31%
Building characteristics 26%
second leading cause of lung cancer after tobacco smoking (WHO,
Geochemistry (uranium 23%
2009). concentration in sediment
The World Health Organization estimates that 3%e14% of lung samples)
cancers are attributable to radon (WHO, 2009) which result in Temperature 6%
21,000 radon related deaths per year in the United States (USEPA, Precipitation 2%
2003) and proportionally to more than 600 deaths per year in
Quebec (CCSSC, 2009). The United States Environmental Protection environment (Fertl and Chilingar, 1988). Such conditions are
Agency (USEPA) classified radon as a “A class” cancer-causing especially met during the formation of black shales in marine en-
substance (Dessau et al., 2004) and the International Agency for vironments. Organic-rich black shales have uranium concentra-
Research on Cancer (IARC), as a “group 1” substance (IARC, 1988). tions ranging from 8 ppm to 168 ppm (aluminium-rich shale)
Those agencies concluded that “radon and its decay products are (Wedepohl, 1969). Fig. 1 shows the correlation between uranium
carcinogenic to humans” (IARC, 1988). concentrations and total organic carbon (TOC) in the Ohio Shale.
In agreement with recent epidemiological studies, Health Can- For igneous and metamorphic rocks, uranium concentration is
ada lowered the Canadian guideline from 800 to 200 Bq/m3 in 2007 usually associated with silica (SiO2) content. Felsic rocks (silica content
(Health Canada, 2007) and Quebec intersectorial radon committee greater than 65% by weight) such as granites have generally higher
prepared an Action plan about radon, effective in 2008. One of the uranium concentrations than mafic and ultramafic rocks (Wedepohl,
main objectives of the Action plan was to develop a map of radon- 1969). Felsic rocks have the highest uranium content, but intermedi-
prone areas for the entire province to help public health authorities ate rocks (silica content between than 52% and 65% by weight) can also
identify radon potential zones and inform populations living in have high values. With uranium content one or two orders of
these regions. magnitude higher than that of mafic and ultramafic rocks, felsic and
The objective of this paper is to show that radiogeochemical intermediate volcanic rocks are considered as likely radon sources.
data and indoor radon concentration measurements PPR can be There is no direct link between uranium-rich bedrock and
used to define different radon potential levels. Such a study was elevated indoor radon concentrations. In fact, interactions are
initiated in the 90’s by Martel (1991) and Lévesque et al. (1995). numerous and have been identified previously in the literature
Recent indoor radon measurements and radiogeochemical data (Nazaroff, 1992; Nazaroff et al., 1987; Nguyen et al., 2011). Radon
enable us to confirm the relationships between radiogeochemical emanation and migration depends on porosity, fractures, perme-
data and indoor radon concentrations and to map different levels of ability of geological units, depth of the water table and saturation,
indoor radon exposure.
Therefore, the paper presents the conclusions from a previous
literature survey of the mostly used criteria for mapping radon-
prone areas (Drolet, 2011), the available dataset used in Quebec
to make such a map, a PPR and a statistical analysis between indoor
radon concentrations and radiogeochemical information and,
finally, a map for each selected criterion showing different levels of
radon emission potential.

2. Survey of criteria used for mapping radon-prone areas

A redundancy analysis of radon mapping criteria (Drolet, 2011)

(Table 1) shows that 89% of 62 studies published during the last 25
years used indoor radon concentrations to map radon-prone areas.
The presence of a radon source is an essential condition for
having high indoor radon concentrations in dwellings. Uranium
rich rocks (and sediments derived from them) are generally the
main source of radon in the Quebec area (Martel, 1991). Many
mapping projects used geology as a primary criterion for locating
radon-prone areas (Alexander and Devocelle, 1997; Apte et al.,
1999; Gundersen and Schumann, 1996; Kemski et al., 2001, 2008;
Miles and Ball, 1996; Zhu et al., 2001) and 66% of the surveyed
international works employed this approach.
Uranium-rich sedimentary rocks are commonly associated with
high organic matter content (Harrell et al., 1991; Martel, 1991;
Poirson and Pagel, 1990). Uranium adsorption by organic matter is Fig. 1. Scatter plot and linear regression for uranium concentration and total organic
effective when carbonaceous materials are present in a reducing carbon (TOC) in the Ohio Shale (adapted from Harrell et al., 1991).
 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67 59

dissolution/migration/precipitation of uranium and radium. Radon a uranium concentration in sediments above 4 ppm to identify
gas emitted from the underlying bedrock must be able to migrate radon-prone areas.
through bedrock to near-surface soil before reaching house base- Even if they are often used in international radon studies, soil
ments and decaying into its daughter product 218Po. Therefore, soil gas radon concentrations and building characteristics were not
permeability is the main surficial deposit characteristic reported in included in our study because of the overall lack of reliable data.
35% of the surveyed studies. Sediment permeability controls radon Other criteria such as temperature and precipitation were also
mobility and hence the distance that radon can travel upward ignored (Table 1).
(Alexander and Devocelle, 1997). Coarse grained surficial sediments The next section shows the available radiogeochemical data that
like sand or gravel have high permeabilities and provide excellent were used to define the radon-prone areas in Quebec.
pathways for upward radon migration. Finer grained materials such
as clay and silt, with significantly lower permeabilities, can form 3. Available datasets and selection of criteria for the Quebec
near-impermeable barriers to radon migration (Alexander and study
Devocelle, 1997; Miles and Ball, 1996). The silt/clay barriers are
effective only when they are thick enough to avoid being pene- In our study, five datasets were used to map radon-prone areas:
trated by buildings foundations (Miles and Ball, 1996). While some (1) indoor radon concentrations, (2) equivalent uranium (eU)
marine silt and clay sediments may be derived from uranium-rich concentrations from surface gamma-ray measurements, (3) ura-
rocks or glacial sediments and may thus constitute radon sources nium concentrations in sediments, (4) bedrock units and (5) sur-
themselves, the scarcity of data on uranium concentrations in Late ficial deposits.
Quaternary marine sediments of southern Quebec does not allow
us to further investigate this possibility. However, only low ura- 3.1. Indoor radon concentration measurements
nium concentrations in glacial and glaciomarine sediments have
been reported in geochemical surveys from uranium-rich Shield 1417 basement radon measurements were collected by different
terrains north of Gatineau (Kettles and Shilts, 1989), which suggests organizations between 2008 and 2011 (Fig. 2). 90 measurements were
that generally low uranium concentrations may be expected in collected by Radioprotection inc. (a private company) and Health
southern Quebec fine-grained marine sediments. Canada in one specific area in north-western Quebec in 2008. 585
One third of the surveyed international studies used airborne alpha-track detectors sampled indoor radon concentrations in 65
surface gamma-ray surveys to predict radon-prone areas. In the schools from eastern and central-western Quebec during winter 2010.
literature, this criterion is considered effective to identify radon- These measurements were made on different floors, (Poulin and
prone areas (Akerblom, 1995; Ball et al., 1995; Chen, 2009; Leclerc, 2010) but only the 48 basement radon measurements were
Christensen and Rigby, 1995; Doyle et al., 1990; Drolet, 2011; Duval, included in the dataset. Quebec Lung Association (QLA) provided the
1989; Ford et al., 2001; Heincke et al., 2008; Jackson, 1992; Kline remaining 1279 basement radon concentrations measured with
et al., 1989; Lévesque et al., 1995; Martel, 1991; Mose et al., 1990; alpha-track detectors. Basement radon concentration measurements
Pellerin et al., 1999; Savard et al., 1998; Smethurst et al., 2008). are the only ones used in this study. Basement radon concentrations
Earlier studies carried out by Martel (1991) and Lévesque et al. are generally higher than values typical of the first floor of the same
(1995) in Quebec set 2 ppm in eU as the threshold for regions building. The goal of the exclusion of the first floor data was to
with a high indoor radon potential. generate the worst case scenario. Radon measurements were scat-
Also, uranium concentration in sediments (geochemical data) tered all over Quebec from volunteers who had chosen to test radon
has been used as a proxy for radon potential in 23% of the surveyed concentrations in their dwellings. Naturally, higher populated regions
international studies. Martel (1991) and Lévesque et al. (1995) used had more radon tested dwellings. For the purpose of mapping, the

Fig. 2. Locations of the 1417 basement radon concentration measurements across Quebec.
60 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67

Fig. 3. Location of the low permeability silt/clay barrier across Quebec.

civic addresses of tested dwellings were systematically converted into engine of the Quebec Ministère des Ressources naturelles et de la
latitude/longitude coordinates to be included into a geo-referenced Faune (MRNF). SIGEOM is a complete database that contains up-to-
database. Web mapping services such as Google Earth, Yahoo! Maps date information on mining rights, geology, geochronology,
and MapQuest were used to locate the dwellings. When two or more geophysics, geochemistry, deposits, drillings, Quaternary geology
concentrations were measured for a specific address, the maximum and hydrogeology in Quebec. There are 278,841 uranium point
measurement was kept in the dataset. All the radon measurements samples that provide uranium concentrations from analysed
are owned by the Ministère de la Santé et des Services sociaux (MSSS) stream and lake sediments, till or soil. Samples were analysed by
and were obtained with the homeowner’s consent. different analytical methods such as inductively coupled plasma
mass spectrometry, fluorometry, neutron activation, atomic ab-
3.2. Equivalent uranium (eU) concentrations from airborne surface sorption, paper chromatography and plasma emission. Uranium
gamma-ray measurements concentrations from geochemical surveys were interpolated to a
regular grid using the inverse distance weighted (IDW) method
Airborne surface gamma-ray surveys provide concentrations of implemented in ESRI’s ArcGIS (ESRI, 2012; Drolet, 2011). This
natural radioactive elements (uranium, thorium and potassium) interpolation technique was used with success by Lamothe (2009)
present in rocks or sediments from the analysis of the energy spec- and Trépanier (2009) to identify uranium geochemical anomalies
trum. Airborne surface gamma-ray surveys were downloaded from (uranium-bearing sediments) for mineral exploration purposes.
Natural Resources Canada web site. Airborne surface gamma-ray The grid dataset covers 80% of the studied territory (Fig. 7B), with
measurements mainly cover the southern reach of Quebec (Fig. 7A). the exception of the highly populated Saint-Lawrence Lowlands
Typically, airborne surface gamma-ray surveys are flown with flight region of southern Quebec. Also, north-western Quebec is not
line spacing ranging from 200 to 5000 m, at a 120 m terrain clearance. covered by grid dataset. However, this area is largely uninhabited.
Higher density surveys (200e500 m spacing) were done for detailed
mapping purposes while lower density surveys (5000 m) were flown 3.4. Bedrock units
for reconnaissance scale purposes (NRCan, 2007).
Uranium concentrations are generally low and measured in parts The bedrock geology of Quebec is relatively well known.
per million (ppm). Airborne surface gamma-ray measurements do Uranium-rich sedimentary, igneous and metamorphic rocks were
not measure uranium-238 concentrations directly. The term equiv- tentatively identified on geological maps available online via
alent uranium concentration (eU in ppm) is used because bismuth- SIGEOM. Those were maps available at 1:20,000 or 1:50,000 scales
214, a 238U daughter product, is measured by the gamma surveys and cover the entire province. It is the only radon-prone mapping
and it is assumed to be in equilibrium with its parent isotope 238U. By dataset that covers the entire province of Quebec (Fig. 7C).
measuring a radon-daughter element (bismuth-214) emitted few In southeastern Quebec (Saint-Lawrence Lowlands and the
minutes after the radon-222 decay (Nero et al., 1990), the airborne Appalachians), most bedrock units are grouped into formations
surface gamma-ray measurements are a fairly direct method of based on their lithological characteristics. Sedimentary black shales
establishing the amount of radon in the upper part of the ground. and intermediate/silicic volcanic rock units in those two geological
and physiographic domains were identified from descriptions
3.3. Geochemistry found in Globensky et al. (1993).
In northern and north-western Quebec (Canadian Shield), rock
Uranium concentrations in sediment samples were available units are rarely grouped into formations. Radon-prone igneous and
online via SIGEOM (GEOMining Information System), a web search metamorphic rocks were identified from their mineralogy as
 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67 61

documented in Moureau and Brace (2000). A Streckeisen double exchanges also control indoor radon concentrations. Therefore, a
triangle was also used to identify intermediate/silicic rocks small increase in eU concentration is not necessarily associated
(Streckeisen, 1976). Potentially uranium-rich bedrock units across with a proportional increase in indoor radon concentration.
Quebec were more specifically identified in Drolet (2011). Moreover, all three criteria are obtained from interpolated data and
are thus approximated. These errors justify working with classes
3.5. Surficial deposits instead of continuous data. Each proposed class had to respect
three conditions: (1) classes cover approximately an equal range of
Silt/clay surficial sediments were identified on nine Quaternary eU concentrations, (2) classes should encompass approximately the
geological maps (Dredge, 1983; Lamarche, 2011; Lasalle and same number of radon measurements and (3) there are at least 50
Tremblay, 1978; Parent, 2012; St-Onge, 2009; Veillette, 1996; radon measurements per class.
Veillette and Cloutier, 1993, 2012; Veillette et al., 2003). These cover The same process was performed with geochemical data. The
only 30% of the studied territory. However, close to 85% of these 1417 geo-referenced dwellings were superimposed on the layer of
maps are from areas invaded by postglacial seas (Occhietti et al., uranium concentration interpolated from geochemical surveys.
2001); the marine fine grained sedimentation likely resulted in Uranium concentrations ranging from 0 to 76 ppm were extracted
deposition of irregular silt/clay barriers to radon migration. Paper for the 448 basement radon concentration measurements covered
maps were digitized prior to identifying deep water marine, gla- by the interpolated geochemical layer. Uranium concentrations
ciomarine and glaciolacustrine sediments. Surficial sediments interpolated from geochemical surveys generated six classes that
inferred to be thinner than 3 m were not included in the silt/clay respect the previously stated three conditions.
barrier because most dwelling foundations break through it when Before superimposing basement radon concentrations on
the soil is excavated. The silt/clay barrier is shown in Fig. 3. bedrock units and surficial deposits, both geological datasets were
merged to create a new variable herein identified as the geology
4. Methodology for PPR and statistical analysis criterion. Radon-prone areas based on geology consist of zones
where there is an assumed radon source in the underlying
4.1. PPR between basement radon concentrations and other bedrock unit (presence of a uranium-rich rock or unconsolidated
selected criteria unit) that is not covered by a silt/clay barrier. The potential to
measure a high indoor radon concentration associated to the
The PPR (positive proportion relationship) between equivalent geology criterion is low when there is no radon source in the
uranium concentration from airborne surface gamma-ray mea- subsurface. It is also low when a potentially uranium-rich
surements and indoor radon concentration measurements was bedrock unit is overlain by a silt/clay barrier which prevents
evaluated to verify the capacity of the eU concentration dataset to radon gas migration from deeper in the ground upwards into
predict radon-prone areas in Quebec. The 1417 basement radon dwelling basements.
concentrations were compared to eU concentrations from airborne The 1417 geo-referenced dwellings were superimposed on the
surface gamma-ray measurements, but only 1077 of the radon geology criterion. Two output classes were created from this cri-
measurements coincide with the eU dataset. The 1077 basement terion because there are only two situations considered in this
radon concentrations were linked to eU concentrations between paper: 1) potential for high radon emissions based on geology
0 and 7 ppm. Equivalent uranium concentration from airborne (uncapped uranium-rich bedrock unit) and 2) potential for low
surface gamma-ray measurements were discretized into five clas- radon emissions based on geology (all other combinations of
ses. The discretization process was carried out because the rela- bedrock units and surficial deposits).
tionship between indoor radon concentrations and eU PPR between basement radon concentration measurements and
concentrations (and the other two criteria) are not straight forward. each of the criteria were evaluated by calculating, for each previ-
Factors such as building structure and indoor/outdoor air ously proposed class:

Fig. 4. Graph of equivalent uranium concentration measured by airborne surface gamma-ray surveys discretized into 5 classes as a function of percentage of dwellings having
indoor radon concentrations above three existing radon guidelines (made from 1077 indoor radon measurements in basements).
62 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67

Fig. 5. Graph of uranium concentration measured in sediments discretized into 6 classes as a function of percentage of dwellings having indoor radon concentrations above three
existing radon guidelines (made from 448 indoor radon measurements in basements).

- the percentage of dwellings exceeding the United States radon 4.2. Statistical study of relationships between basement radon
guideline (150 Bq/m3); concentrations and the selected criteria
- the percentage of dwellings exceeding the actual Canadian
radon guideline (200 Bq/m3); The KruskaleWallis one way analysis of variance on ranks
- the percentage of dwellings exceeding the previous Canadian (ANOVA) were performed on indoor radon concentrations for each
radon guideline (800 Bq/m3); predetermined class of the three selected criteria (equivalent ura-
- the geometric mean (GM); nium concentration from surface airborne gamma-ray measure-
- the geometric standard deviation (GSD). ments, uranium concentration in sediments and geology). This was
done in order to test if there is a statistically significant difference
Geometric mean (GM) and geometric standard deviation (GSD) between each class. A p-value, calculated by the ANOVA, lower than
allowed us to interpret classes in terms of log-normal distributions 0.05 was set as the threshold for a statistically significant difference.
and our basement radon concentrations dataset is log-normally A p-value above 0.05 allowed us to merge classes based on statis-
distributed. tical similarities. This analysis was carried out for each of the three

Fig. 6. Graph of the geology criterion discretized into 2 classes as a function of percentage of dwellings having indoor radon concentrations above three existing radon guidelines
(made from 1417 indoor radon measurements in basements).
 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67 63

Table 2 Table 4
KruskaleWallis one way ANOVA on five classes of equivalent uranium concentration KruskaleWallis one way ANOVA on six classes of uranium concentration interpo-
from airborne surface gamma-ray measurements. A p-value >0.05 means that the lated from geochemical surveys. A p-value >0.05 means statistically significant
hypothesis that they share a common mean cannot be rejected at the 95% confi- similarities between the tested classes (bold characters). The median of indoor
dence level (bold characters). The median of indoor radon concentration in base- radon concentration in basement, in Bq/m3, is included in parenthesis for each class.
ment, in Bq/m3, is included in parenthesis for each class.
Compared class 1 (median) Compared class 2 (median) p-value
Compared class 1 (median) Compared class 2 (median) p-value
0e1 ppm (81) 1e3 ppm (78) 0.602
0e0.5 ppm (63) 0.5e0.75 ppm (59) 0.814 0e1 ppm (81) 3e10 ppm (78) 0.479
0e0.5 ppm (63) 0.75e1.00 ppm (81) 0.007 0e1 ppm (81) 10e15 ppm (81) 0.470
0e0.5 ppm (63) 1.00e1.25 ppm (85) 0.001 0e1 ppm (81) 15e20 ppm (80) 0.422
0e0.5 ppm (63) 1.25 ppm (122) <0.001 0e1 ppm (81) 20 ppm (212) <0.001
0.5e0.75 ppm (59) 0.75e1.00 ppm (81) 0.004 1e3 ppm (78) 3e10 ppm (78) 0.987
0.5e0.75 ppm (59) 1.00e1.25 ppm (85) <0.001 1e3 ppm (78) 10e15 ppm (81) 0.839
0.5e0.75 ppm (59) 1.25 ppm (122) <0.001 1e3 ppm (78) 15e20 ppm (80) 0.642
0.75e1.00 ppm (81) 1.00e1.25 ppm (85) 0.233 1e3 ppm (78) 20 ppm (212) <0.001
0.75e1.00 ppm (81) 1.25 ppm (122) <0.001 3e10 ppm (78) 10e15 ppm (81) 0.945
1.00e1.25 ppm 1.25 ppm (122) 0.065 3e10 ppm (78) 15e20 ppm (80) 0.765
3e10 ppm (78) 20 ppm (212) <0.001
10e15 ppm (81) 15e20 ppm (81) 0.863
criteria individually with SigmaStat’s functions featured into Sig- 10e15 ppm (81) 20 ppm (212) <0.001
maPlot 12 (from Systat Software Inc.). 15e20 ppm (80) 20 ppm (212) <0.001

5. Results KruskaleWallis one way analyses of variance on ranks (ANOVA).

Table 2 shows p-values calculated for all combinations of compared
5.1. PPR between basement radon concentrations and other classes (compared class 1 versus compared class 2 in Table 2) for the
selected criteria gamma-ray spectrometry criterion.
The “0e0.5 ppm” and “0.5e0.75 ppm” classes are merged
An increase in the percentage of dwellings exceeding three because they are not statistically different. Also, statistical similar-
North American radon guidelines is observed from lower to higher ities exist between “0.75e1.00 ppm” and “1.00e1.25 ppm” classes,
classes of equivalent uranium concentration from airborne surface and also between “1.00e1.25 ppm” and “1.25 ppm” classes. The
gamma-ray measurements (Fig. 4). The geometric mean increase of choice to regroup the “1.00e1.25 ppm” class with the “0.75e
indoor radon concentration in the basement of dwellings is in 1.00 ppm” class instead of grouping with the “1.25 ppm” class was
accordance with increasing eU concentrations classes, thus strongly based on higher p-values (0.233 and 0.065 respectively).
suggesting a statistically significant PPR between those two KruskaleWallis one way ANOVA analyses were performed on
parameters. the three new groups (“0e0.75 ppm”, “0.75e1.25 ppm” and
The increase in the percentage of dwellings that exceed North “1.25 ppm”) of equivalent uranium (eU) concentration from
American radon guidelines from low uranium concentrations airborne surface gamma-ray measurements to ensure statistical
interpolated from geochemical surveys to higher levels (Fig. 5) is significance (Table 3). Percentages of dwellings having their indoor
not as obvious as it is for the gamma-ray spectrometry data. The radon concentrations in basements exceeding the three North
highest class (>20 ppm; Fig. 5) shows higher percentages of American radon guidelines of 150, 200 and 800 Bq/m3 are shown as
dwellings exceeding the three guidelines than for the highest class a function of three eU concentration groups.
of eU (Fig. 4). Moreover, a higher geometric mean of indoor radon Statistics presented in Table 3 show that the probability of
concentration is associated with this class compared to that of the encountering a higher proportion of elevated radon concentrations
highest class of eU. The first five classes might be statistically ho- increases when eU concentration from airborne surveys is above
mogeneous and a statistical study on the similarities of classes is 0.75 ppm. Regions located in areas with an eU concentration above
needed (see section 5.2). 0.75 ppm were considered radon-prone areas because Duval (1989)
The PPR between indoor radon concentrations in basements showed that there is a high health risk when 30% of the dwellings in
and the geology criterion is shown on Fig. 6. The percentage of a region exceed 150 Bq/m3. When the eU concentration is above
dwellings exceeding all three North American radon guidelines and 1.25 ppm, indoor radon emission potential increases and over 40%
the geometric mean of indoor radon concentration are higher for of dwellings exceed the 150 Bq/m3 guideline.
the potentially high radon emission level class compared with the The same validation procedure of classes was made for the
potentially low radon emission level class. geochemistry criterion (Table 4). The classes “0e1 ppm”, “1e
3 ppm”, “3e10 ppm”, “10e15 ppm” and “15e20 ppm” were merged
5.2. Statistical study because it was impossible to reject the hypothesis that they share a
common mean at the 95% confidence level (a p-value above 0.05).
Statistically significant differences and similarities between Two new groups were created: “0e20 ppm” and “20 ppm”. The
each class for all three above selected criteria were validated by statistical dissimilitude between the “0e20 ppm” group and the

Table 3
KruskaleWallis one way ANOVA on three groups of equivalent uranium concentration from airborne surface gamma-ray measurements.

Group n Median 25th percentile 75th percentile GM GSD % [Rn]  150 Bq/m3 % [Rn]  200 Bq/m3 % [Rn]  800 Bq/m3
(x in ppm of eU) (Bq/m3) (Bq/m3) (Bq/m3) (Bq/m3) (Bq/m3)

0 < x < 0.75 342 59 33 115 64 2.5 17.833.3 10.8 0

0.75  x < 1.25 450 81 41 185 89 2.8 33.3 22.2 0.9
x  1.25 285 122 56 219 116 2.8 40.7 27.7 2.5

c2 ¼ 53,375 with 2 degrees of freedom. (p-value between three groups  0.001; statistically significant difference).
64 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67

Table 5
KruskaleWallis one way ANOVA on one group and one class of uranium concentration interpolated from geochemical surveys.

Group n Median 25th percentile 75th percentile GM GSD % [Rn]  150 Bq/m3 % [Rn]  200 Bq/m3 % [Rn]  800 Bq/m3
(x in ppm) (Bq/m3) (Bq/m3) (Bq/m3) (Bq/m3) (Bq/m3)

0 < x < 20 366 80 39 167 84 2.8 28.4 20.2 1.1

x  20 82 212 136 358 215 2.7 69.5 53.7 8.5

c2 ¼ 52,384 with 1 degree of freedom. (p-value between two groups  0.001; statistically significant difference).

Table 6
KruskaleWallis one way ANOVA on two potential radon emission levels based on the geology criterion.

Group n Median 25th percentile 75th percentile GM GSD (Bq/m3) % [Rn]  150 Bq/m3 % [Rn]  200 Bq/m3 % [Rn]  800 Bq/m3
(Bq/m3) (Bq/m3) (Bq/m3) (Bq/m3)

Potentially low radon emission 1110 74 37 144 74 2.6 24.1 14.1 0.7
level based on geology
Potentially high radon emission 307 115 44 281 111 3.1 43.0 34.5 2.0
level based on geology

c2 ¼ 33,097 with 1 degree of freedom. (p-value between two groups  0.001; statistically significant difference).

“20 ppm” class was validated by another KruskaleWallis one way airborne surface gamma-ray survey is needed to increase radon
ANOVA test (Table 5). mapping accuracy.
The “20 ppm” class of uranium in sediments has 70% of the The synthesized provincial map of radon-prone areas is not
dwellings exceeding the 150 Bq/m3 guideline and is considered to presented herein because a more complicated approach than the
represent a high health risk as suggested by Duval (1989). Also, the simple addition of potential radon emission level maps has been
geometric mean of indoor radon concentration of this class is used. The approach used (KruskaleWallis one way ANOVA on ranks
215 Bq/m3 which is higher than the actual Canadian indoor radon to combine radon emission potential maps of radiogeochemical
guideline. The presence of such uranium concentrations interpo- information) is explained in details in another paper (Drolet et al.,
lated from geochemical surveys in a region indicates that it is a 2013). This map shows four different levels of indoor radon expo-
radon-prone zone. sure for the whole province of Quebec.
The last KruskaleWallis one way ANOVA on ranks was carried Fig. 8 shows a map of radon potential based on basement radon
out in order to determine if there is a significant statistical differ- measurements only. This is a direct way of estimating the radon
ence between measured basement radon concentrations with potential. It is useful to evaluate the radon model predictions pre-
potentially low and high radon emission levels as defined by the sented in Fig. 7A, B and C. This radon potential map based on direct
geology criterion (Table 6). A p-value lower than 0.001 confirms indoor radon measurements was generated by calculating the
that the populations of radon values assigned to the low and high proportion of dwellings above the action level of 200 Bq/m3 for
radon potential classes are statistically different and that the high each municipality. A minimum of 5 basement radon measurements
radon potential class can be considered as a radon-prone area. This was required to assign a radon potential to the municipalities.
class has 43% of dwellings exceeding an indoor radon concentration
of 150 Bq/m3.
6. Discussion

5.3. Maps Indoor radon concentration sampling has not been done
randomly. It is plausible that a significant proportion of the mea-
Fig. 7 shows the three proposed criteria used in Quebec for surements were performed because homeowners had some reason
radon mapping (A: equivalent uranium concentration from to suspect their home might have a high radon concentration. This
airborne surface gamma-ray measurements with three classes; B: could result in an overrepresentation of high values in the dataset
uranium concentration in sediment geochemistry with two classes; (Burke and Murphy, 2011). To illustrate this effect, it is interesting to
C: geology with two classes). Based on the statistical study with the note that the proportion of homes above 200 Bq/m3 is 10.8% in
three criteria and by combining radiogeochemical information areas where eU concentration is between 0 and 0.75 ppm (Table 3),
from Figs. 7A, 7B and 7C, it is possible to determine a radon which is comparable to the proportion recently published by
emission potential map. Regions where each criterion indicates a Health Canada for the entire province (10.1%; Health Canada, 2012).
high potential radon emission level (Map A: “1.25 ppm”; B: These considerations emphasize the need to obtain new data from
“20 ppm”; C: “high potential of measuring an elevated indoor all regions of the province, if possible randomly, to improve the
radon concentration”) were considered as radon-prone areas. external validity of the model.
Dwellings built in south-western Quebec could potentially have Also all three criteria contain uncertainties and may create
high basement radon concentration measurements based on radi- variability in the results obtained in the PPR and the statistical
ogeochemical information. Fig. 7A and C show high potential radon studies. Uranium concentrations from geochemical surveys were
emission levels for south-eastern Quebec not covered by the interpolated by the inverse distance weighted technique on ArcGIS
geochemistry criterion (see Fig. 7B) and is considered as radon- creating variability in the model. Equivalent uranium concentra-
prone. If a region has only one criterion with a high radon emis- tions from gamma-ray surveys and the bedrock units were inter-
sion potential, the zone only has a medium potential for high in- polated by kriging also showing variability. Also, bedrock units
door radon concentrations. Identification of potential radon come from many local geological maps that were all merged
emission level is less accurate in zones only covered by the geology together. Some non-geological boundaries can be seen in Fig. 7C
criterion. The Montréal region and the region north of Montréal are because the ways rock types have been identified by different ge-
the only highly populated zones in this situation; where an ologists. The less precise variable was the silt/clay barrier. The
 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67
Fig. 7. Discretized criteria across Quebec based on the statistical study to define radon emission potential maps. Map A: Indoor radon emission potential based on equivalent uranium concentration from airborne surface gamma-ray
measurements; Map B: Indoor radon emission potential based on uranium concentration in sediments from geochemistry surveys; Map C: Indoor radon emission potential based on bedrock geology combined with low permeability
silt/clay surficial deposits.

65
66 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67

Fig. 8. Radon potential map based on direct basement radon measurements.

geology of surficial sediments source datasets were preliminarily Acknowledgements

interpolated by kriging causing some noise on the output datasets.
Also, the thickness and the mineralogy of the surficial deposits The Ministère de la Santé et des Services sociaux (MSSS) funded
were not detailed in the datasets for such a huge territory causing this study. Authors are grateful to Professor André St-Hilaire (INRS-
uncertainties on the areal extent of the “thick enough (3 m) non ETE) and Suzanne Gingras (INSPQ) for reviewing the statistical
radon emitting silt/clay barrier”. Nonetheless, the three criteria are methodology. We also want to thank Jean-Marc Leclerc (INSPQ),
reliable enough to identify high indoor radon concentrations and Drs. Guy Sanfaçon and Albert Daveluy (MSSS) for constructive
show that variability in the model is not large enough to invalidate discussions.
the PPR.

7. Conclusion References

Radiogeochemical information is useful to generate maps of Akerblom, G., 1995. The use of airborne radiometric and exploration survey data
and techniques in radon risk mapping in Sweden. Application of uranium
radon-prone areas. PPR (positive proportion relationships) be-
exploration data and techniques in environmental studies. Int. At. Energy
tween three selected criteria and radon concentrations in base- Agency Tech. Document 827, 159e180.
ments demonstrate that these criteria are statistically reliable Alexander, W.G., Devocelle, L.L., 1997. Mapping indoor radon potential using geol-
ogy and soil permeability. In: 1997 International Radon Symposium.
predictors for potential radon emission levels. A statistical study
Apte, M.G., Price, P.N., Nero, A.V., Revzan, K.L., 1999. Predicting New Hampshire
proposes thresholds where potential radon emission levels in- indoor radon concentrations from geologic information and other covariates.
crease for each criteria. This methodology allows mapping radon- Environ. Geol. 37, 181e194.
prone areas in regions where no actual indoor radon measure- Ball, T.K., Cameron, D.G., Colman, T.B., Roberts, P.D., 1995. The use of uranium
exploration data for mapping radon potential in the UKdadvantages and pit-
ments are available. While radon potential is based on radio- falls. Application of uranium exploration data and techniques in environmental
geochemical information, accuracy of the potential radon emission studies. Int. At. Energy Agency Tech. Document 827, 139e149.
level increases with the number of available criteria. New gamma- Burke, O., Murphy, P., 2011. The use of volunteer radon measurements for radon
mapping purposes: an examination of sampling bias issues. J. Radiol. Prot. 31,
ray surveys and/or uranium concentration measurements in sedi- 319e328.
ments would improve the model accuracy. There is always at least Canadian Cancer Society’ Steering Committee (CCSSC), 2009. Canadian Cancer
one criterion that can be used to determine radon potential because Statistics 2009. Canadian Cancer Society, Toronto.
Chen, J., 2009. A preliminary design of a radon potential map for Canada: a multi-
the geology criterion is available over the entire study area. A tier approach. Environ. Earth Sci. 59, 775e782.
combined map of radon-prone areas in the province of Quebec Christensen, L.G., Rigby, J.G., 1995. GIS applications to radon hazard studies e an
based on the three criteria (airborne gamma-ray spectrometry, example from Nevada. Nev. Bur. Mines Geol.. Special Publication 17.
Dessau, J.C., Gagnon, F., Lévesque, B., Prévost, C., Leclerc, J.M., Belles-Isles, J.C., 2004.
uranium concentrations in sediment samples and geology) and on Le radon au Québec: Évaluation du risque à la santé et analyse critique des
indoor radon measurements still has to be made. Putting the three stratégies d’intervention. Québec. Institut national de santé publique du
criteria together into one total radon potential requires more in- Québec.
Doyle, P.J., Grasty, R.L., Charbonneau, B.W., 1990. Predicting geographic variations in
door radon measurements to determine the model efficiency.
indoor radon using airborne gamma-ray spectrometry. Paper 90-1A. In:
Public health authorities may use such a combined map of radon- Geological Survey of Canada, Current Research, Part A, pp. 27e32.
prone areas to inform populations living in potentially high radon Dredge, L.A., 1983. Surficial geology, Sept-Iles, Quebec/Géologie de surface, Sept-
concentrations regions. Indoor radon measurements and mitiga- Îles, Québec. Geological Survey of Canada. In: “A” Series Map 1575A.
Drolet, J.P., 2011. Revue des données géoscientifiques existantes pour identifier les
tion measures need to be promoted and conducted in such areas to zones potentielles à la présence de radon domiciliaire au Québec, Canada. In:
help lowering the occurrence of radon-related lung cancers. Université du Québec (ETE-INRS), p. 263. (Master thesis).
 J.-P. Drolet et al. / Journal of Environmental Radioactivity 124 (2013) 57e67 67

Drolet, J.P., Martel, R., Poulin, P., Dessau, J.C., 2013. An Approach to Define map of Moureau, M., Brace, G., 2000. Dictionnaire des Sciences de la Terre/Comprehensive
Radon-prone Areas Based on Indoor Radon Concentration Measurements and Dictionary of Earth Science. technip, Paris.
Radiogeochemical Data Positive Proportional Relationships. Nambi, K.S.V., Aitken, M.J., 1986. Annual dose conversion factors for TL and ESR
Duval, J.S., 1989. Indoor radon prediction using gamma-ray spectrometric data. EOS, dating. Archaeom 28, 202e205.
Trans. Am. Geophys. Union 70, 496. Natural Resources Canada (NRCan), 2007. Radioactivity Data. Airborne Gamma-ray
Environmental Systems Research Institute (ESRI), 2012. ArcMap 10.1. ESRI, Red- Spectrometry. Available at: http://gdr.nrcan.gc.ca/gamma/about_e.php (accessed
lands, CA. 12.03.12.).
Fertl, W.H., Chilingar, G.V., 1988. Total organic carbon content determined from well Nazaroff, W.W., 1992. Radon transport from soil to air. Rev. Geophys. 30, 137e160.
logs. Soc. Pet. Engineers (SPE) Formation Eval. 3, 407e419. Nazaroff, W.W., Doyle, S.M., Nero, A.V., Sextro, R.G., 1987. Potable water as a source
Ford, K.L., Savard, M., Dessau, J.C., Pellerin, E., Charbonneau, B.W., Shives, B.K., 2001. of airborne 222Rn in U.S. dwellings: a review and assessment. Health Phys. 52,
The role of gamma-ray spectrometry in radon risk evaluation: a case history 281e295.
from Oka. Quebec. Geosci. Can. 28, 59e64. Nero, A.V., Gadgil, A.J., Nazaroff, W.W., Revzan, K.L., 1990. Indoor Radon and Decay
Globensky, Y., et al., 1993. Lexique stratigraphique canadien. Volume V-B: région Products: Concentrations, Causes, and Control Strategies. Lawrence Berkeley
des Appalaches, des Basses-Terres du Saint-Laurent et des Îles de la Madeleine. Laboratory. Technical Report, LBL-27798.
Gouvernement du Québec, Ministère de l’Énergie et des Ressources et Direction Nguyen, D.C., Dulinski, M., Jodlowski, P., Nowak, J., Rozanski, K., Sleziak, M.,
générale de l’exploration géologique et minérale, 327 p. DV 91e23. Wachniew, P., 2011. Natural radioactivity in ground water. Isot. Environ. Health
Gundersen, L.C.S., Schumann, R.R., 1996. Mapping the radon potential of the United Stud. 47, 415e437.
States: examples from the Appalachians. Environ. Int. 22, S829eS837. Occhietti, S., Chartier, M., Hillaire-Marcel, C., Cournoyer, M., Cumbaa, S.L.,
Harrell, J.A., Belsito, M.E., Kumar, A., 1991. Radon hazards associated with outcrops Harrington, C.R., 2001. Paléoenvironnements de la Mer de Champlain dans la
of Ohio Shale in Ohio. Environ. Geol. Water Sci. 18, 17e26. région de Québec, entre 11 300 et 9750 BP: le site de Saint-Nicolas. Geogr. Phys.
Health Canada, 2007. Radon: What You Need to Know. Information sheet, HC Quat. 55, 23e46.
Publication, p. 4, 4928. Parent, M., 2012. Géologie des formations superficielles du Québec méridional.
Health Canada, 2012. Cross-Canada Survey of Radon Concentrations in Homes. Final Geological Survey of Canada. Preliminary map under revision.
Report, p. 23. Pellerin, E., Savard, M., Dessau, J.C., 1999. Évaluation des risques de surexposition au
Heincke, B.H., Smethurst, M.A., Bjørlykke, A., Dahlgren, S., Rønning, J.S., radon dans un secteur des collines de Saint-André-Est. Direction régionale de la
Mogaard, J.O., 2008. Airborne gamma-ray spectrometer mapping for relating santé publique, p. 17.
indoor radon concentrations to geological parameters in the Fen region, Poirson, C., Pagel, M., 1990. Distribution séquentielle de l’uranium dans les bassins
southeast Norway. In: Slagstad, T. (Ed.), Geology for Society. Geological Survey sédimentaires: exemples du Permien de Lodève (Hérault) et de Rodez
of Norway, pp. 131e143. Special Publication 11. (Aveyron), du Carbonifère de Baden Baden (Bade Wurttemberg) et du Lias
International Agency for Research on Cancer (IARC), 1988. Monographs on the lorrain. Université de Nancy, p. 163. (Ph.D. thesis).
Evaluation of Carcinogenic Risks to Humans. In: Man-made Mineral Fibres and Poulin, P., Leclerc, J.M., 2010. Projet de dépistage du radon dans des écoles primaires
Radon, vol. 43. IARC, Lyon. du Québec situées en zone d’intervention prioritaire. Institut national de santé
Jackson, S.A., 1992. Estimating radon potential from an aerial radiometric survey. publique, p. 37.
Health Phys. 62, 450e452. Savard, M., Dessau, J.C., Pellerin, E., 1998. Le radon à oka e Rapport d’intervention de
Kemski, J., Siehl, A., Stegemann, R., Valdivia-Manchego, M., 2001. Mapping the santé publique. Régie régionale de la santé et des services sociaux des Lau-
geogenic radon potential in Germany. Sci. Total Environ. 272, 217e230. rentides, Direction de la santé publique, p. 146.
Kemski, J., Klingel, R., Siehl, A., Valdivia-Manchego, M., 2008. From radon hazard to Smethurst, M.A., Strand, T., Sundal, A.V., Rudjord, A.L., 2008. Large-scale radon
risk prediction-based on geological maps, soil gas and indoor measurements in hazard evaluation in the Oslofjord region of Norway utilizing indoor radon
Germany. Environ. Geol. 56, 1269e1279. concentrations, airborne gamma ray spectrometry and geological mapping. Sci.
Kettles, I., Shilts, W.W., 1989. Geochemistry of Drift Over the Precambrian Grenville Tot. Environ. 407, 379e393.
Province, southeastern Ontario and southwestern Quebec. Geological Survey of St-Onge, D.A., 2009. Géologie des formations en surface, basse vallée de l’Outaouais,
Canada, pp. 97e112. Paper 89e20. Ontario-Québec (compilation). Geological Survey of Canada, p. 27. Current
Kline, S.W., Mose, D.G., Zietz, I., 1989. The use of gamma aeroradioactivity in Research 2001eD12.
regional indoor radon prediction: a case study from Fairfax County, Virginia, Streckeisen, A.L., 1976. Classification of the common igneous rocks by means of
USA. J. Radiol. Prot. 9, 121e128. their chemical composition: a provisional attempt. Neues Jahrbuch Fuer Min-
Lamarche, L., 2011. Évolution paléoenvironnementale de la dynamique quaternaire eralogie, Monatshefte 1, 1e15.
dans la région de Québec: application en modélisation tridimensionnelle et Trépanier, S., 2009. Évaluation du potentiel minéral en uranium et en Cu-Au-U, et
hydrogéologique. Université du Québec (ETE-INRS), p. 221. Philosophiae Doctor cartographie prévisionnelle d’intrusions mafiques-ultramafiques dans le
thesis. Grenville. Géologie Québec, p. 46. document EP 2009-03.
Lamothe, D., 2009. Cartes géochimiques ModelBuilder et cibles anomales de l’en- US Environmental Protection Agency (USEPA), 2003. EPA Assessment of Risks from
vironnement secondaire pour le Québec. Ministère des Ressources naturelles et Radon in Homes, 402-R-03e003.
de la Faune, GM, 64290, 33 maps and other numerical data. Veillette, J.J., Cloutier, M., 1993. Géologie des formations en surface, Gaspésie,
Lasalle, P., Tremblay, G., 1978. Dépôts meubles-Saguenay-Lac-Saint-Jean. Ministère Québec. In: “A” Series Map 1804A. Geological Survey of Canada.
des richesses naturelles. In: Direction générale de la recherche géologique, Veillette, J.J., 1996. Géomorphologie et géologie du Quaternaire du Témiscamingue,
p. 61. Rapport géologique 191. Québec et Ontario. Geological Survey of Canada, p. 269. Report 476.
Lévesque, B., Gauvin, D., McGregor, R.G., Martel, R., Gingras, S., Dontigny, A., Veillette, J.J., Paradis, S.J., Thibaudeau, P., 2003. Les cartes de formations en surface
Walker, W.B., Lajoie, P., 1995. Étude d’exposition au radon-222 dans les rési- de l’Abitibi, Québec. Geological Survey of Canada. Open File 1523.
dences de la province de Québec. Centre de santé publique de Québec, p. 46. Veillette, J.J., Cloutier, M., 2012. Géologie du Quaternaire dans le Bas-Saint-Laurent,
Martel, R., 1991. Zones proposées pour l’échantillonage du radon dans les habita- Québec. Geological Survey of Canada. Preliminary map under revision.
tions québécoises. Ministère de l’environnement du Québec, Direction des Wedepohl, K.H., 1969. Handbook of Geochemistry, vol. I/II. Springer-Verlag, Berlin.
écosystèmes urbains, p. 77. Division des eaux souterraines. World Health Organization (WHO), 2009. Fact Sheet No 291: Radon and Cancer.
Miles, J., Ball, K., 1996. Mapping radon-prone areas using house radon data and Available at: http://www.who.int/mediacentre/factsheets/fs291/en/ (accessed
geological boundaries. Environ. Int. 22, S779eS782. 17.07.12.).
Mose, D.G., Chrosniak, C., Kline, S.W., Mushrush, G.W., 1990. Prediction of indoor Zhu, H.C., Charlet, J.M., Poffijn, A., 2001. Radon risk mapping in southern Belgium: an
radon by aeroradioactivity. Pure Appl. Geophys. 133, 213e227. application of geostatistical and GIS techniques. Sci. Tot. Environ. 272, 203e210.