The transfer of 137 Cs, Pu isotopes and 90 Sr to bird, bat and ground-dwelling small mammal species within the Chernobyl exclusion zone

Protected species are the focus of many radiological environmental assessments. However, the lack of radioecological data for many protected species presents a signi ﬁ cant international challenge. Further-more, there are legislative restrictions on destructive sampling of protected species to obtain such data. Where data are not available, extrapolations are often made from ‘ similar ’ species but there has been little attempt to validate this approach. In this paper we present what, to our knowledge, is the ﬁ rst study purposefully designed to test the hypothesis that radioecological data for unprotected species can be used to estimate conservative radioecolgical parameters for protected species; conservatism being necessary to ensure that there is no signi ﬁ cant impact. The study was conducted in the Chernobyl Exclusion Zone. Consequently, we are able to present data for Pu isotopes in terrestrial wildlife. There has been limited research on Pu transfer to terrestrial wildlife which contrasts with the need to assess radiation exposure of wildlife to Pu isotopes around many nuclear facilities internationally. Our results provide overall support for the hypothesis that data for unprotected species can be used to adequately assess the impacts for ionising radiation on protected species. This is demonstrated for a range of mammalian and avian species. However, we identify one case, the shrew, for which data from other ground-dwelling small mammals would not lead to an appropriately conservative assessment of radiation impact. This indicates the need to further test our hypothesis across a range of species and ecosystems, and/or ensure adequate conservatism within assessments. The data presented are of value to those trying to more accurately estimate the radiation dose to wildlife in the Chernobyl Exclusion Zone, helping to reduce the considerable uncertainty in studies reporting dose-effect relationships for wildlife. A video abstract for this paper is available from: © 2016 The Authors. Published by Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Th e t r a n sf e r of 1 3 7 C s, P u iso t o p e s a n d 9 0 S r t o bi r d, b a t a n d g Ti t l e Th e t r a n sf e r of 1 3 7 C s, P u iso t o p e s a n d 9 0 S r t o bi r d, b a t a n d g Protected species are the focus of many radiological environmental assessments. However, the lack of radioecological data for many protected species presents a significant international challenge. Furthermore, there are legislative restrictions on destructive sampling of protected species to obtain such data. Where data are not available, extrapolations are often made from 'similar' species but there has been little attempt to validate this approach.
In this paper we present what, to our knowledge, is the first study purposefully designed to test the hypothesis that radioecological data for unprotected species can be used to estimate conservative radioecolgical parameters for protected species; conservatism being necessary to ensure that there is no significant impact.
The study was conducted in the Chernobyl Exclusion Zone. Consequently, we are able to present data for Pu isotopes in terrestrial wildlife. There has been limited research on Pu transfer to terrestrial wildlife which contrasts with the need to assess radiation exposure of wildlife to Pu isotopes around many nuclear facilities internationally.
Our results provide overall support for the hypothesis that data for unprotected species can be used to adequately assess the impacts for ionising radiation on protected species. This is demonstrated for a range of mammalian and avian species. However, we identify one case, the shrew, for which data from other ground-dwelling small mammals would not lead to an appropriately conservative assessment of radiation impact. This indicates the need to further test our hypothesis across a range of species and ecosystems, and/or ensure adequate conservatism within assessments.
The data presented are of value to those trying to more accurately estimate the radiation dose to wildlife in the Chernobyl Exclusion Zone, helping to reduce the considerable uncertainty in studies reporting dose-effect relationships for wildlife.
A video abstract for this paper is available from: http://bit.ly/1JesKPc. © 2016 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).

Introduction
A necessary component of the tools (e.g. Brown et al., 2008;Copplestone et al., 2001Copplestone et al., , 2003USDoE, 2002) now established to estimate the exposure of wildlife to ionising radiations is an ability to predict wholebody activity concentrations of radionuclides in a wide range of biota. Although there are alternative approaches to predict transfer to wildlife in development, such as the use of taxonomic relationships (e.g. Beresford et al., 2013Beresford et al., , 2015, most of the available tools use concentration ratios (CR wo-media ) relating the activity concentrations in plants and animals to those in the appropriate environmental media (soil, air or water) (Beresford et al., 2008a). Whilst databases of CR wo-media values for wild species have been collated (e.g. Beresford et al., 2008b;Copplestone et al., 2013;Hosseini et al., 2008;Howard et al., 2013;Yankovich et al., 2013), data for many radionuclide-organism combinations are sparse or not available. Where data are unavailable, assumptions such as applying data for a 'similar organism' (e.g. mammal data for birds) are often made to provide default CR wo-media values for use in dose assessment tools (Beresford et al., 2008b;Brown et al., 2013).
Protected species are the focus of many assessments (e.g. Copplestone et al., 2005;Wood et al., 2008). For many protected species, transfer data are lacking and there are legislative restrictions on destructive sampling to obtain data (Wood et al., 2011). For some protected species, there are very few data for the overall taxonomic group appropriate to that species. A good example of this is chiroptera (bats), all species of which are protected in the European Union (HMSO, 1994). For some radionuclides there are many CR wo-soil data for other animals within the class mammalia and the extent to which these data are applicable to bats needs to be established. Similarly, at many ecologically important sites requiring assessment (e.g. Natura 2000 sites), the most prevalent protected organisms are aves (bird) species (Copplestone et al., 2003). However, there are very few CR wo-media values for birds (e.g. ICRP, 2009).
Previously, we have published data on the transfer of 137 Cs and 90 Sr to a range of bat species sampled from a variety of sites within the Chernobyl Exclusion Zone (CEZ) (Gaschak et al., 2010). The CEZ, which is the area established around the Chernobyl nuclear complex following the 1986 accident, is increasingly viewed as a natural laboratory, and more recently as a radioecological observatory (https://wiki.ceh.ac.uk/x/NoFsD). It provides an opportunity to study the transfer of radionuclides to different species of wildlife across different taxonomic groups (e.g. Beresford et al., 2005). In this paper we present a study where species of birds, bats and ground-dwelling small mammals were sampled from a site within the CEZ and analysed for 137 Cs, Pu isotopes and 90 Sr. To our knowledge, this is the first study purposefully designed to test the hypothesis that radioecological modelling parameters derived from the sampling and analysis of unprotected species (i.e. grounddwelling small mammals) result in a conservative dose assessment for protected species inhabiting the same site. The paper also makes an important contribution to the available database of Pu isotope data for terrestrial wildlife, few studies having been published previously (e.g. Johansen et al., 2014Johansen et al., , 2015.

Study sites
In Beresford et al. (2008c) we report a study to determine the exposure of small mammal species at three forest sites within the CEZ conducted during the summer of 2005. The sites were initially selected to have a range in radionuclide activity concentrations; animal samples from each site were collected within a 100 Â 100 m area. In the present study, samples have been collected from one of these sites (termed the 'Medium site' in Beresford et al., 2008c).
The Medium site was approximately 8 km to the west of the Chernobyl power plant complex. The woodland at the Medium site consisted mainly of Pinus sylvestris (Scots pine) and Quercus robur (Oak), with some Sorbus aucuparia (Rowan) and Tilia platyphyllos (Large leaved lime). The sparse understorey vegetation included Pteridium aquilinum (Bracken). The site had soddy pseudopodzolic sandy and boggy soils on modern alluvial deposits. Beresford et al. (2008c) describes the collection and analyses of soils (n ¼ 23) from the Medium site; soils were collected from an area extended to 50 m beyond the animal sampling area to encompass the likely home ranges of the animal species being trapped (i.e. soils were collected from an area of 200 m Â 200 m or 40000 m 2 ). Soil activity concentrations were reported in Beresford et al. (2008c) as: 43.3 ± 25.7, 0.83 ± 1.49, 18.6 ± 14.9 kBq kg À1 dry mass for 137 Cs,238,239,240 Pu and 90 Sr respectively. Whilst variable, there was no spatial pattern in soil activity concentrations across the sampling site.

Bird samples
A range of passerine species were collected by mist net at the Medium site during June 2005, euthanised and retained frozen. Species, sample numbers and information on feeding and home range are presented in Table 1.

Bat samples
Three species of bats were collected from the site during the period MayeJune 2008 using mist nets (Table 1). After being euthanised the samples were stored frozen whilst awaiting analyses.

Ground-dwelling small mammals
In Beresford et al., 2008c, live- reported. Some carcasses of A. flavicollis, M. glareolus and also Sorex araneus (Table 1), collected in Sherman humane traps (baited with cereals) during this study, were retained and stored frozen. Some of these have been analysed for the purposes of the present paper.

Analyses
Prior to analysis, samples were defrosted and prepared as follows: bird carcasses were plucked and had the gastrointestinal tract (GIT) removed; bat samples had the GIT removed; and ground-dwelling small mammals had the pelt and GIT removed. All prepared samples were then weighed and washed before analyses.
The wholebody 137 Cs and 90 Sr concentrations were determined using the method described by Bondarkov et al. (2011). Prior to counting, the carcasses were placed in a small, disposable, cardboard box (70 Â 40 Â 40 mm), the upper side of which was made from <0.1 mm thick polyethylene. The box was then placed inside a lead shielded counting container. The detectors comprised a hyper-pure germanium detector and thin-film (1 mm) NaI scintillation detector to measure 137 Cs and 90 Sr, respectively. The 137 Cs spectra were analysed using the Canberra Genie-2000 software package. The activity concentration of 90 Sr was determined from that of its daughter nuclide, 90 Y. The method has previously been calibrated against phantoms containing 137 Cs and 90 Sr; the methodology has been validated against traditional radiochemical extraction and analysis methodologies. Counting times varied from 150 to 1200 s depending upon the radioactivity in the animal. Counting errors were typically <3% for 90 Sr and <7% for 137 Cs.
To determine Pu isotopes, samples were initially dissolved in 65% HNO 3 and 242 Pu was added as a yield tracer. Following anion exchange separation (Bio Rad AG 1 Â 8, 100e200 mesh) and coprecipitation, the samples were counted using a planar ion implanted silicon detector. Counting errors were typically <20% for the Pu isotopes.

Results
Radionuclide activity concentrations summarised by species are presented in Table 2. Given the limited sample numbers of some species we have not attempted any statistical comparison at the species level, focussing instead on group-level comparisons ('bird', 'bat' and 'ground dwelling small mammal') ( Table 3). Recognising that radionuclide activity concentration data for wildlife generally follow a lognormal distribution , the data were log-transformed prior to analysis. The ground-dwelling small mammals had significantly higher activity concentrations of all three radionuclides compared to the other two groups (Generalised Linear Model; p < 0.05). Differences between birds and bats were not consistent between radioisotopes, with 137 Cs activity concentrations being higher in birds and 90 Sr concentrations higher in bats. The 137 Cs: 90 Sr activity concentration ratio in bats was significantly lower than that for both the other groups by a factor of c. 6.
The transfer of radionuclides to wildlife is most commonly described by the whole-organism concentration ratio (CR wo-soil ) (IAEA, 2014; Beresford et al., 2008a), where for terrestrial animals: As only a total 238,239,240 Pu activity concentration in soil was available, we calculated Pu CR wo-soil values by combining the 238 Pu and 239,240 Pu values for each animal. Concentration ratios are presented by species in Table 4 and by group in Table 5. Significant differences in CR wo-soil between the groups (Table 5) were the same as for activity concentrations (Table 3). For all groups the highest CR wo-soil values were for 90 Sr which were approximately double those determined for 137 Cs (p < 0.001; paired t-test); the Pu CR wosoil values were two to three orders of magnitude lower than the 137 Cs CR wo-soil values (p < 0.001; paired t-test).

Discussion
It is likely that diet contributes to the higher activity concentrations, for all radioisotopes, observed in the ground-dwelling small mammals compared to the birds and bats ( Table 2). The diet of all of these ground-dwelling small mammals includes ground living invertebrates. Previous studies have suggested that flying insects (which comprise the diet of study bat species and are an important component of the diet of most of the bird species sampled) generally have lower radionuclide activity concentrations than ground-dwelling invertebrate species collected from the same site Barnett et al., 2014). Ground-dwelling invertebrates may contain soil in their digestive tract or be externally contaminated by soil. There is also more potential for the grounddwelling small mammals to inadvertently ingest contaminated soil.
The CR wo-soil for Pu to shrews is 1e2 orders of magnitude higher than those for the other species studied in this paper (Table 4). A similar observation has been made for a number of toxic metals that have been shown to accumulate to higher levels in shrews than in other small mammals (Shore and Rattner, 2001;Tom a skov a et al., 2005). The high metabolic rate and diet of shrews have been suggested to be the reason for this (Hegstrom and West, 1989;Swiergosz-Kowalewska et al., 2005).
Although protected bat species are often the target of assessment (Copplestone et al., 2003), there are very few available data on the transfer of radionuclides to this group of mammals. The only other data that we are aware of are those reported by Gaschak et al. (2010) for bats sampled during 2007e2009 from different areas of the CEZ. Gashchak et al. present transfer of 137 Cs and 90 Sr relative to the deposition in soil rather than as CR wo-soil values. For comparison with our data, Table 6 presents CR wo-soil values calculated from the data of Gaschak et al. (2010); these data were extracted from the database described by Copplestone et al., 2013. The CR wo-soil values presented in the present paper are within the range of those calculated from the Gashchak et al. paper. The only potential exception is that the Pipistrellus pipistrellus CR wo-soil value for Cs is lower than the range of values presented in the larger dataset of Gashchak et al.. We should acknowledge that, whilst the size of our sampling area was appropriate for the home range of the birds and small ground dwelling mammals, the bat species generally forage over a larger area (Table 1).
The lower 137 Cs: 90 Sr ratio observed in bats compared to the other animal types implies a comparatively high transfer of Sr compared to Cs in bats. The diet of insectivorous bats is limited in calcium (Adams et al., 2003) a strong Sr analogue. It is therefore, likely that the dietary absorption of calcium, and consequently Sr, is higher in bats than in the other species. However, we acknowledge that the 137 Cs: 90 Sr ratio is not consistent across the CEZ (Kashparov et al., 2003) and the greater home range of the bats compared to the other species may contribute to our observation.
Data from previous studies both within the Chernobyl Exclusion Zone and in other areas support our observation that the transfer of Sr is comparatively high compared to Cs Gashchak et al., 2003Gashchak et al., , 20082009Maklyuk et al., 2007;Sheppard, 1991 1 ;Sheppard et al., 2004Sheppard et al., 1 , 20082010 1 Sheppard andEvenden, 1990 1 ). Data presented by Beresford et al. (2008c) for small ground-dwelling mammals sampled from the Chernobyl Exclusion Zone in 2005 suggest that the extent of this difference between Sr and Cs transfer was highest at sites with a generally low level of contamination (<10 kBq kg À1 dry mass 137 Cs). At the most contaminated site, which was located on the western trace (c.100 kBq kg À1 dry mass 137 Cs), the results of Beresford et al. (2008c) suggest that the transfer of Cs was higher than that for Sr. Gaschak et al. (2010) suggested that for bats sampled throughout the CEZ there was a decreasing transfer of 90 Sr with increasing deposition. Similar observations have been reported for amphibians, small birds and ungulates (Gashchak et al., 2008Gashschak, 2009;Gaichenko et al., 2001). This may be because Sr contamination is predominantly in particulate form at the highest contamination sites. Kashparov et al. (1999) showed that the dissolution rate of Sr from particles to the west of the reactor (where the most contaminated site in Beresford et al. was located) was lower than that from other areas.
The data presented in this paper provide an opportunity to test the hypothesis that sampling unprotected species (represented by two species of the ground-dwelling small mammals, M. glareolus and Sylvaemus flavicollis) and determining activity concentrations in these will allow a conservative estimate of activity  Table 6 Concentration Ratios (CR wo-soil ) for bats within the Chernobyl Exclusion Zone, calculated from the data discussed in Gaschak et al. (2010) (data extracted from the database described by Copplestone et al., 2013). Geometric mean values are presented with the geometric standard deviation given in brackets. concentrations in protected species to be obtained. Table 2 demonstrates that sampling M. glareolus and S. flavicollis would give higher activity concentration estimates for 137 Cs,238,239,240 Pu and 90 Sr than those measured in bats directly. Therefore, using data for these two small ground-dwelling mammals to make an assessment of dose to bats would result in a conservative estimate. Given that many species of birds are also protected, Table 2 also lends confidence to the use of a similar approach to conservatively assessing doses to birds. However, in the case of S. araneus, a protected species of ground-dwelling small mammal in the United Kingdom (HMSO, 1981), sampling the M. glareolus and S. flavicollis would underestimate activity concentrations and, therefore, dose from Pu isotopes given the higher transfer of Pu to this species as discussed previously.

Conclusions
The data presented in this paper significantly improves the available information on small bird and bat species and also Pu isotope levels in wildlife from the CEZ. The data provide a valuable addition to available transfer databases and will also facilitate more accurate dose assessments for wildlife in the Chernobyl exclusion zone perhaps aiding the debate on reported effects from the area (e.g. Beresford and Copplestone, 2011;Garnier Laplace et al., 2013).
Bats had a high 90 Sr: 137 Cs ratio compared to other species; a diet deficient in calcium may contribute to this observation.
The data support the hypothesis that sampling and analyses of small ground-dwelling mammals would provide a conservative estimate of exposure for birds and bats which are often protected and for which there are comparatively few data.
The CR wo-soil value for Pu transfer to shrews (a protected species in the United Kingdom) was 1e2 orders of magnitude higher than those for the other species. This should be taken into account when conducting assessments if using other small mammals as surrogates for shrews.