Chemosphere xxx (2014) xxx–xxx
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard Ida Beathe Øverjordet a,⇑, Geir Wing Gabrielsen b, Torunn Berg c, Anders Ruus d, Anita Evenset e,f, Katrine Borgå d,g, Guttorm Christensen e, Syverin Lierhagen c, Bjørn Munro Jenssen a a
Norwegian University of Science and Technology (NTNU), Department of Biology, N-7491 Trondheim, Norway Norwegian Polar Institute, N-9296 Tromsø, Norway c NTNU, Department of Chemistry, N-7491 Trondheim, Norway d Norwegian Institute for Water Research (NIVA), N-0349 Oslo, Norway e Akvaplan-niva, Fram Centre, N-9296 Tromsø, Norway f Faculty of Biosciences, Fisheries and Economics, UiT The Arctic University of Norway, Norway g Department of Biosciences, University of Oslo, Norway b
h i g h l i g h t s Biomagnification potential of Hg and Cd both in Arctic and Atlantic food webs. Trophic position explains Hg variation in kittiwakes but not in little auks. Selenium was associated with Hg in little auks, but not in kittiwakes. High Se:Hg molar ratios indicate low risk of Hg toxicity.
a r t i c l e
i n f o
Article history: Received 25 June 2014 Received in revised form 11 October 2014 Accepted 20 October 2014 Available online xxxx Handling Editor: Keith Maruya Keywords: Arctic Marine environmental pollution Food web Bioaccumulation Black-legged kittiwake Little auk
a b s t r a c t Hepatic concentrations of mercury (Hg), selenium (Se) and cadmium (Cd) were determined in blacklegged kittiwakes (Rissa tridactyla) and little auks (Alle alle) from two fjords in Svalbard (Kongsfjorden; 78°570 N, 12°120 E and Liefdefjorden; 79°370 N, 13°200 E). The inflow of Arctic and Atlantic water differs between the two fjords, potentially affecting element accumulation. Trophic positions (TP) were derived from stable nitrogen isotope ratios (d15N), and stable carbon isotope ratios (d13C) were assessed to evaluate the terrestrial influence on element accumulation. Mercury, Cd, TP and d13C varied significantly between locations and years in both species. Trophic position and feeding habits explained Hg and Cd accumulation in kittiwakes, but not in little auks. Biomagnification of Hg and Cd were found in the food webs of both the Atlantic and the Arctic fjord, and no inter-fjord differences were detected. The d13C were higher in the seabirds from Kongsfjorden than in Liefdefjorden, but this did not explain variations in element accumulation. Selenium concentrations were not influenced by Hg accumulation in kittiwakes, indicating baseline levels of Se in this species. In contrast, correlations between Hg and Se and lower Se:Hg ratios in little auks from Kongsfjorden than in Liefdefjorden indicate a more pronounced influence of Se–Hg complex formation in little auks feeding in Atlantic waters. Ó 2014 Published by Elsevier Ltd.
1. Introduction Even after extensive regulations of anthropogenic Hg emissions over the past decades, Hg levels are increasing in Arctic seabirds ⇑ Corresponding author at: SINTEF Materials and Chemistry, Environmental Technology, N-7465 Trondheim, Norway. Tel.: +47 951 76491. E-mail address: [email protected] (I.B. Øverjordet).
and mammals (Braune et al., 2005; Riget et al., 2011; Dietz et al., 2013). Despite low emission in Arctic areas, an annual input of 90–450 metric tons of Hg from atmospheric deposition is estimated (Ariya et al., 2004; Skov et al., 2004). A large fraction is reemitted to the atmosphere, but still, a net input of Hg to Arctic ecosystems has been reported (Poissant et al., 2008). High Hg deposition during the Arctic spring overlaps with the spring bloom of phytoplankton, potentially increasing the assimilation of Hg in
http://dx.doi.org/10.1016/j.chemosphere.2014.10.060 0045-6535/Ó 2014 Published by Elsevier Ltd.
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
2
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
algae (Stern and Macdonald, 2005; Douglas et al., 2012). In turn, this could facilitate Hg accumulation in Arctic marine food webs. By using fatty acid markers unique for the Arctic copepod Calanus glacialis, Routti et al. (2012) found a coherence between high concentrations of Hg in polar bears feeding on a diet with a high influence of this Arctic copepod. Selenium (Se) is an essential element and is therefore physiologically regulated within an optimal range. However, as Hg concentrations increase Se accumulation is enhanced by the formation of Hg–Se complexes, a process that protects cells and organisms against Hg toxicity (both organic and inorganic Hg; Khan and Wang, 2009). Similar to Hg, Cd accumulates in Arctic seabirds despite low emission within the Arctic (Dietz et al., 1996; Savinov et al., 2003). Arctic atmospheric Cd concentrations are low (Berg et al., 2008), suggesting that most Cd found in Arctic marine biota are derived from geochemical processes of the ocean (MacDonald et al., 2000). Cd is naturally more abundant in the North Pacific than in the North Atlantic (by a factor of five), resulting in a net flux of Cd from the Pacific to the Atlantic side of the Arctic Ocean (Bruland and Franks, 1983; MacDonald et al., 2000). This may lead to higher Cd concentrations in Arctic water compared to Atlantic water. Cadmium is easily adsorbed to the chitinous surface of zooplankton, facilitating uptake in fish and birds (Dietz et al., 1996). In a scenario of a changing climate, the species composition of the marine Arctic may shift towards Atlantic species (FalkPetersen et al., 2007; Eriksen and Dalpadado, 2011). Such a shift has been observed in Kongsfjorden (Svalbard), due to unusually high influxes of Atlantic water (Willis et al., 2008; Gabrielsen and Hop, 2009). The change has persisted, as shown by abnormally large amounts of Atlantic species (Atlantic cod (Gadus morhua), capelin (Mallotus villosus), euphausiids and Calanus finmarchicus) in Kongsfjorden during the years 2007–2011 (Gabrielsen and Hop, 2009; Buchholz et al., 2012). During parts of this period black-legged kittiwakes (Rissa tridactyla) fed on capelin (Mallotus villosus) rather than their traditionally more important prey, the polar cod (Boreogadus saida), in Kongsfjorden (Gasbjerg, 2010). Element concentrations correlate to trophic position (TP) in several Arctic seabird species (Atwell et al., 1998; Braune et al., 2005; Ricca et al., 2008). Studies of bioaccumulation in relation to diet commonly apply the ratios between stable isotopes of nitrogen (15N/14N) to determine TP as the heavy 15N is enriched relative to the lighter 14N in predators. (Minagawa and Wada, 1984; Hobson and Welch, 1992; Atwell et al., 1998; Power et al., 2002). The food web baseline is defined by d15N in organisms at the first trophic position (TP1), i.e. primary producers. Alternatively, d15N of primary consumers can be used to as a proxy for the second trophic position (TP2). Baseline d15N varies between locations and seasons, making it important to correct for the baseline when comparing accumulation of elements in different food webs (Cabana and Rasmussen, 1996). In Arctic food webs, the Calanus species may be used as TP2 organisms to estimate TP higher in the food web (Hobson and Welch, 1992). The carbon source of a food web can be derived from the ratio between stable isotopes of carbon (13C/12C) (DeNiro and Epstein, 1978; Post, 2002; Søreide et al., 2006). Enrichment of 13C in inshore environments reflects a carbon source of terrestrial origin or a stronger coupling to the benthic environment (Hobson, 1999). Mercury accumulation may in some cases be explained by the carbon source of marine food webs (Ricca et al., 2008; Bond and Diamond, 2009). Little auks (Alle alle) and black-legged kittiwakes are key species in Arctic marine ecosystems. The kittiwake is an opportunistic feeder preying on small pelagic fish like polar cod and capelin, as well as pelagic amphipods (Themisto spp.) and euphausiids (Thysanoessa spp.) (Mehlum and Gabrielsen, 1993). In contrast, little auks are specialist feeders, preying mainly on pelagic zooplankton like
copepods, Themisto spp., Thysanoessa spp. and Mysis spp. (Mehlum and Gabrielsen, 1993). Copepods of the Calanus genus dominate the diet of little auks in the chick rearing period (Gabrielsen et al., 1991; Steen et al., 2007). Hence, kittiwake and little auk represent different trophic levels in the partial food web of the Arctic pelagic ecosystem. The aim of the present study was to evaluate the influence of diet, feeding location and sampling year on accumulation of Hg, Cd and Se in Arctic pelagic seabirds. Samples of kittiwakes and little auks were collected in Liefdefjorden and Kongsfjorden at Svalbard. Kongsfjorden is dominated by Atlantic water while Liefdefjorden is dominated by Arctic water (Hallanger et al., 2011a). Furthermore, d15N and d13C were analyzed in muscle tissue of seabirds and in Calanus glacialis to examine the effects of TP and carbon source on element accumulation and biomagnification in the two fjords. 2. Methods 2.1. Field sampling Kittiwakes and little auks were collected in the vicinity of two breeding colonies at Svalbard, one in Kongsfjorden (78°570 N, 12°120 E) and the other in Liefdefjorden (79°370 N, 13°200 E; Fig. 1). Sampling was carried out during the last two weeks of July 2008 and 2009. Ten free-ranging individuals of each species were collected outside the breeding colony at each location each year using a shotgun with non-toxic ammunition (steel/bismuth). Permission for birds sampling was granted by the Governor of Svalbard. Samples of the left pectoralis muscle and liver tissue were dissected out for stable isotope and element analysis, respectively. Sex and body mass was recorded. Calanus glacialis were collected for food web baseline reference in both fjords each year using Method Isaac Kidd (MIK) nets (1000 lm, whole water column). Only C. glacialis at developmental stage copepodite V (CV) was used in the assessment of trophic position. All samples were stored in plastic bags at 20 °C before transportation and analysis. 2.2. Element determination Frozen liver tissue was lyophilized in acid washed polypropylene tubes (2 mL; Nalgene cryotubes, Thermo Scientific, Waltham, MA USA) with the screw cap partially open for a minimum of 24 h prior to digestion (CIT2, Leybold–Heraeus, Köln, Germany). Dry samples (0.15 g) were transferred to PTFE-vials (18 mL; UltraClave, MLS GmbH, Leutkirch, Germany) together with 2.3 g ultrapure water (Q-option, Elga Labwater, Veolia Water Systems LTD, UK) and 4.2 g concentrated nitric acid, HNO3 (Scanpure, equal to ultrapure grade, Chem Scan, Elverum, Norway). Digestion was carried out using a high-pressure microwave emitter (Milestone Ultra Clave, EMLS, Leutkirch, Germany). The temperature was gradually increased from room temperature up to 250 °C within 1 h and thereafter gradually cooled down to the initial value within the next hour. The digested samples were diluted with ultrapure water in acid washed polypropylene vials (BD Falcon 50 mL conical, BD Biosciences, Bradford, MA, US) to a final volume of 60 mL and a final HNO3 concentration of 0.6 M. Total Hg, Cd and Se were determined by high-resolution inductively coupled plasma mass spectrometry (HR-ICP-MS, Thermo Finnigan model Element 2, Bremen, Germany). Instrument settings were described in detail by Sørmo et al. (2011). Samples were randomized with respect to sampling location and species. Three blank samples containing ultrapure water and HNO3 were prepared in the same way as the samples (0.6 M in final solution). To account for potential carryover during digestion, the PTFE-vials
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
3
Fig. 1. Map of Svalbard showing the locations for sampling of black-legged kittiwakes (Rissa tridactyla), little auks (Alle alle) and Calanus glacialis.
containing blank samples were alternated between each run. The detection limits based on the standard deviation of blank samples were 0.144, 0.013 and 0.012 lg g1 for Se, Cd and Hg, respectively. No samples were below the level of detection. The average relative standard deviations (RSD) of multiple scans were below 3% for all elements. Certified standard reference materials (SRM) bovine liver (NIST 1577b), oyster tissue (NIST 1566b) and chicken (GBW 10018) were digested and analyzed at regular intervals (n P 6 of each SRM). The analytical values of SRM were within the range of the certified values. The recovery of Se was 114%, 123% and 102% in bovine liver, chicken and oyster tissue, respectively. Mercury was certified in oyster tissue, with a recovery of 105%. Cadmium recovery in bovine liver and oyster tissue was 100% and 101%, respectively.
centration of 4 ng g1. However, the Cd concentrations in these samples did not increase with increasing Bi concentration. Hence, we consider the potential contamination from ammunition to be negligible. 2.4. Stable isotopes Stable isotopes were analyzed at the Great Lakes Institute for Environmental Research, Canada, according to the procedure by Hussey et al. (2010). Lipids were removed from all samples, and additionally carbonates were removed from the C. glacialis samples. Stable isotope ratios are expressed in d notation as the deviation from standard in parts per thousand (‰):
dX‰ ¼
2.3. Quality assurance of tissue sampling for element analysis In the field, the tissue samples were dissected out in a laboratory on board r/v Lance or at the Kings Bay Marine Laboratory. The whole liver was removed before it was divided into subsamples on a clean cutting board. The tissue samples were frozen in zip-lock bags that were pre-tested for potential element leakage with negative results. (The bags were equilibrated with MQ-water that was subsequently analyzed for trace elements using HR-ICPMS.) Prior to the freeze drying procedure, the bulk tissue sample collected in the field was transferred to a plastic-covered cutting board, where the outermost layer of tissue was removed using a scalpel. Subsamples were taken from the inner tissue to prevent potential contamination from the field to be included in the analyzed sample. The scalpel blade and the plastic cover were changed between each sample. The clean subsample were transferred to acid washed (0.6 M HNO3) polypropylene tubes with screw caps, that were only partially opened during the freeze drying procedure. The dry subsample was weighted directly in the PTFE tube in which they subsequently were digested to prevent direct handling after drying. Potential contamination from ammunition was examined by correlation plots between Cd and Bi and between Cd and Fe. No samples had deviating Fe concentrations. A few samples contained higher Bi concentrations (max 2 lg g1) than the median Bi con-
Rsample 1 1000 Rstandard
ð1Þ
where dX is the delta value of 13C and 15N (‰) and R is the molar ratio of 13C/12C or 15N/14N in the sample and in an international standard, respectively. Vienna Pee Dee Belemnite served as standard for isotopic ratios of carbon, and atmospheric air was used as standard for isotopic ratios of nitrogen. Replicate measurements of internal laboratory standards (muscle tissue of fish) generally run for every 10 sample, indicated measurement errors ranging from ±0.05 to ±0.07 SD for d13C and from ±0.13 to ±0.27 SD for d15N. A trophic enrichment factor of 2.4‰ for 15N between seabirds and their prey was assumed, resulting in the following equation: d15Nbird = d15Nprey + 2.4 (Mizutani et al., 1991). For the rest of the food web an enrichment factor of 3.8‰ was applied (Hobson and Clark, 1992). The trophic position for kittiwake and little auk was calculated using the d15N values of Calanus glacialis as TP2 by the following equation (each fjord and sampling year separate):
TPbird ¼ 3 þ
d15 Nbird ðd15 NC:glacialis þ 2:4Þ 3:8
ð2Þ
2.5. Data analysis Statistical analysis was performed using R (v2.13.1; R Development Core Team). The distribution of variables was tested using QQ-plots and Shapiro–Wilk normality test. Variables that deviated
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
4
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
from a normal distribution were log-transformed (kittiwake: Hg and Se:Hg; little auk: Hg, Cd, Se, and Se:Hg). Equality of variance was tested using Levene’s test in the ‘‘lawstat’’ package. All variables had a homogenous variance within the four groups (two locations, two years), except trophic position in little auk. Based on these results, parametric tests were applied. The level of significance was set to p < 0.05. However, p < 0.1 and p > 0.05, were applied to indicate trends in the results. Examination of model plots of the regression between log(Hg) and d13C in kittiwakes resulted in one individual kittiwake from Liefdefjorden 2009 being considered as an outlier (d13C: 21.96‰, Cook’s D > 0.5). This individual was removed from the dataset prior to statistical analysis involving d13C. Element concentrations did not differ between sexes in either species (Hg, Cd, Se: p > 0.05), thus, both sexes are included in data analysis. Analysis of variance (ANOVA) with Tukey’s HSD (honest significant difference) post hoc test was used to determine differences in mean levels of Hg, Cd, Se, Se:Hg, d15N, d13C and TP between groups (location x year) for little auks (n = 38) and kittiwakes (n = 40). In kittiwakes, ANOVA analysis of Hg and Cd revealed significant between-group differences (p < 0.05), but the TukeyHSD test did not display which groups were different due to significant interactions between year and locality (F = 9.00, p = 0.005). t-tests were applied for these groups. Linear models and t-tests were run to evaluate differences between the fjords and years. The influence of location and sampling year on separate variables (y = TP, dC, Cd, Hg, Se, Se:Hg) were tested separately using Eq. (3). Letters a–d are regression constants.
y ¼ a þ b Location þ c Year þ d Location Year
ð3Þ
Eqs. (4) and (5) were used to examine potential differences in trophic magnification potentials of Hg and Cd between locations and years for parts of the food web including both avian species (n = 78).
log½element ¼ a þ b TP þ c Location þ d ðTP LocationÞ
ð4Þ
log½element ¼ a þ b TP þ c Year þ d ðTP YearÞ
ð5Þ
Literature values of elements given on a wet weight basis were recalculated to dry weight using the following formula: [Me]wet weight/0.3 = [Me]dry weight, where Me is the element in question, and 0.3 is the approximate portion of dry matter in liver samples of kittiwakes and little auks from the present study (31.6 ± 1.53 %, n = 70). 3. Results and discussion 3.1. Species differences The concentrations of Hg, Cd (Fig. 2) and Se (Table 1) were higher in kittiwakes than in little auks independent of location and sampling year (TukeyHSD, p < 0.001). Previously reported data on element accumulation confirm these inter-species differences (Dietz et al., 1996; Savinov et al., 2003; Poissant et al., 2008; Jæger et al., 2009). As expected from the inter-species difference in diet (Mehlum and Gabrielsen, 1993), the kittiwakes were feeding at a higher TP than the little auks (3.1 and 2.7, respectively; t-test, p < 0.001, Fig. 2). The TP of both species were lower than previously reported (Jæger et al., 2009; Hallanger et al., 2011b), indicating that the TPs may be slightly underestimated when calculated based on C. glacialis (Eq. (2)). Predatory species usually accumulate higher levels of Hg than primary consumers, increasing the dietary exposure of birds as they feed higher up in the food chain (Dietz et al., 1996; Jæger
et al., 2009). A difference in TP of less than 1 between the species indicates that the kittiwakes have included lower trophic level prey in their diet as shown by Mehlum and Gabrielsen (1993). Despite this, Hg concentrations were two times higher in kittiwakes than in little auks (Fig. 2). The Hg composition of prey items, i.e. the relative amount of methyl mercury (MeHg), may have enhanced the inter-species differences in total Hg concentrations (Fig. 2). Due to its physical and chemical properties, ingested MeHg is more readily absorbed than inorganic Hg, while inorganic Hg is more efficiently eliminated than MeHg (Scheuhammer et al., 2008). MeHg is therefore more efficiently transferred in the food chain, resulting in a larger percentage of MeHg in fish species like polar cod (90–100%) compared to the Calanus species (7.5–30%; Campbell et al., 2005; Loseto et al., 2008; Jæger et al., 2009). Hence, the opportunistic feeding habit of kittiwakes may result in a more diet-dependent variation in Hg accumulation relative to the specialist feeder little auk. 3.2. Annual and spatial differences Mercury concentrations in little auks were higher in Kongsfjorden than in Liefdefjorden in both years (2008: p = 0.025; 2009: p = 0.069; Fig. 2). In kittiwakes, the Hg concentrations were highest in Liefdefjorden in 2008 and in Kongsfjorden in 2009 (2008: p = 0.027; 2009: p = 0.080). Likewise, Cd concentrations varied between years in both species (Fig. 2). These between-year variations suggest that between-fjord differences are determined by other factors than simply higher Hg and Cd accumulation in food webs of Arctic waters (MacDonald et al., 2000; Routti et al., 2012). Mercury was positively correlated with TP in kittiwakes (Pearson, R = 0.352, p = 0.030), indicating that the inter-annual variation in Hg concentrations were linked to between-year shifts in the TP of kittiwakes (TukeyHSD, p < 0.01, Fig. 2). In contrast, in little auks, no between-year differences in Hg accumulation were observed, despite significant variation in TP between fjords and years (TukeyHSD, p < 0.001, Fig. 2). Stable isotope signals have been shown to vary inter-annually in Arctic seabirds as a consequence of variable prey availability (Moody et al., 2012). Inter-annual variation in d15N in kittiwakes was attributed to the ice cover dependent availability of polar cod (Moody et al., 2012). Similar variations in prey availability may explain the between-year variation in TP of the kittiwakes in the present study. Indeed, polar cod was less frequent in the diet of kittiwakes from Kongsfjorden, and in recent years their feed has been increasingly dominated by capelin (Mehlum and Gabrielsen, 1993; Gasbjerg, 2010). In little auks, significant correlation was found between Hg and d15N (Pearson, R = 0.404, p = 0.013), but not between Hg and TP (p = 0.435). Little auks prey exclusively on Calanus spp. during the breeding season (Gabrielsen et al., 1991; Steen et al., 2007). Lower d15N values were found in Calanus glacialis in Liefdefjorden 2008 compared to the other sampling groups (p < 0.001, Table 1), indicating that the differences in TP may rather be explained by a shift in baseline d15N. Hence, the inter-annual variation in Hg accumulation between the two fjords in kittiwakes may be explained by the shift in TP between the years, while in little auks this was not the case (Fig. 2). The elevated Hg concentrations in little auks from Kongsfjorden relative to Liefdefjorden (Fig. 2) may also be a result of local geological differences or long range oceanic transport resulting in higher input of Hg to Kongsfjorden (Lu et al., 2013). Correlations between Hg and d13C have been found in seabirds from areas where the terrestrial input of Hg is significant (Bond and Diamond, 2009). Little auks from Kongsfjorden were enriched in 13C compared to little auks from Liefdefjorden (both years: t-test, p = 0.003; Fig. 3). Indeed, a positive correlation was found between Hg and d13C in little auks (Pearson, R = 0.303, p < 0.001, n = 37, Fig. 4), indicating the influence of Hg of terrestrial origin. However, natural positive relationships
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
5
Fig. 2. Trophic position (TP), hepatic mercury (Hg) and cadmium (Cd) concentrations (lg g1 ± SE dw) in kittiwakes (Rissa tridactyla; left) and little auks (Alle alle; right) from Kongsfjorden and Liefdefjorden sampled in 2008 (light) and 2009 (dark). Differences between groups within each species (location x years) are denoted with different letters (ANOVA and TukeyHSD, p < 0.05; Hg and Cd in kittiwakes: t-test, p < 0.05).
between d15N and d13C make it challenging to assess the influence of d13C alone on element accumulation (Bond and Diamond, 2009). No correlations were found when the residuals from the regression
between d13C and d15N were regressed against Hg. Similarly, despite significant enrichment of d13C in kittiwakes from Kongsfjorden relative to Liefdefjorden (19.91‰ and 20.54‰, respectively;
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
6
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
Table 1 Hepatic concentrations of selenium (Se; lg g1 dry weight) and molar ratio of Hg:Se in kittiwakes (Rissa tridactyla) and little auks (Alle alle) from Svalbard. Stable isotopes (dC and dN, ‰) in muscle tissue of birds and whole Calanus glacialis. Mean ± SE. n = 10 unless otherwise stated. Statistically different groups are assigned different letters (TukeyHSD; p < 0.05). ANOVA: ⁄⁄ = p < 0.01, ⁄⁄⁄ = p < 0.001. Kongsfjorden
Kittiwake Se Se:Hg d13C d15N
A B C D
lg g1 dw molar ratio ‰ ‰
2009
2008
Mean ± SE 16.37 ± 2.01 25.58 ± 5.58 20.12 ± 0.07 11.46 ± 0.14
– – a a
Mean ± SE 16.48 ± 1.81 17.09 ± 1.75 19.71 ± 0.14 12.22 ± 0.12
– – b b
Mean ± SE 16.25 ± 1.21 17.37 ± 2.87 20.46 ± 0.04 12.25 ± 0.18
– – c b
Mean ± SE 13.62 ± 1.43 17.98 ± 1.74 20.63 ± 0.09 11.76 ± 0.21
– – c ab
p (ANOVA) 0.295 0.436
Contents lists available at ScienceDirect
Chemosphere journal homepage: www.elsevier.com/locate/chemosphere
Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard Ida Beathe Øverjordet a,⇑, Geir Wing Gabrielsen b, Torunn Berg c, Anders Ruus d, Anita Evenset e,f, Katrine Borgå d,g, Guttorm Christensen e, Syverin Lierhagen c, Bjørn Munro Jenssen a a
Norwegian University of Science and Technology (NTNU), Department of Biology, N-7491 Trondheim, Norway Norwegian Polar Institute, N-9296 Tromsø, Norway c NTNU, Department of Chemistry, N-7491 Trondheim, Norway d Norwegian Institute for Water Research (NIVA), N-0349 Oslo, Norway e Akvaplan-niva, Fram Centre, N-9296 Tromsø, Norway f Faculty of Biosciences, Fisheries and Economics, UiT The Arctic University of Norway, Norway g Department of Biosciences, University of Oslo, Norway b
h i g h l i g h t s Biomagnification potential of Hg and Cd both in Arctic and Atlantic food webs. Trophic position explains Hg variation in kittiwakes but not in little auks. Selenium was associated with Hg in little auks, but not in kittiwakes. High Se:Hg molar ratios indicate low risk of Hg toxicity.
a r t i c l e
i n f o
Article history: Received 25 June 2014 Received in revised form 11 October 2014 Accepted 20 October 2014 Available online xxxx Handling Editor: Keith Maruya Keywords: Arctic Marine environmental pollution Food web Bioaccumulation Black-legged kittiwake Little auk
a b s t r a c t Hepatic concentrations of mercury (Hg), selenium (Se) and cadmium (Cd) were determined in blacklegged kittiwakes (Rissa tridactyla) and little auks (Alle alle) from two fjords in Svalbard (Kongsfjorden; 78°570 N, 12°120 E and Liefdefjorden; 79°370 N, 13°200 E). The inflow of Arctic and Atlantic water differs between the two fjords, potentially affecting element accumulation. Trophic positions (TP) were derived from stable nitrogen isotope ratios (d15N), and stable carbon isotope ratios (d13C) were assessed to evaluate the terrestrial influence on element accumulation. Mercury, Cd, TP and d13C varied significantly between locations and years in both species. Trophic position and feeding habits explained Hg and Cd accumulation in kittiwakes, but not in little auks. Biomagnification of Hg and Cd were found in the food webs of both the Atlantic and the Arctic fjord, and no inter-fjord differences were detected. The d13C were higher in the seabirds from Kongsfjorden than in Liefdefjorden, but this did not explain variations in element accumulation. Selenium concentrations were not influenced by Hg accumulation in kittiwakes, indicating baseline levels of Se in this species. In contrast, correlations between Hg and Se and lower Se:Hg ratios in little auks from Kongsfjorden than in Liefdefjorden indicate a more pronounced influence of Se–Hg complex formation in little auks feeding in Atlantic waters. Ó 2014 Published by Elsevier Ltd.
1. Introduction Even after extensive regulations of anthropogenic Hg emissions over the past decades, Hg levels are increasing in Arctic seabirds ⇑ Corresponding author at: SINTEF Materials and Chemistry, Environmental Technology, N-7465 Trondheim, Norway. Tel.: +47 951 76491. E-mail address: [email protected] (I.B. Øverjordet).
and mammals (Braune et al., 2005; Riget et al., 2011; Dietz et al., 2013). Despite low emission in Arctic areas, an annual input of 90–450 metric tons of Hg from atmospheric deposition is estimated (Ariya et al., 2004; Skov et al., 2004). A large fraction is reemitted to the atmosphere, but still, a net input of Hg to Arctic ecosystems has been reported (Poissant et al., 2008). High Hg deposition during the Arctic spring overlaps with the spring bloom of phytoplankton, potentially increasing the assimilation of Hg in
http://dx.doi.org/10.1016/j.chemosphere.2014.10.060 0045-6535/Ó 2014 Published by Elsevier Ltd.
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
2
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
algae (Stern and Macdonald, 2005; Douglas et al., 2012). In turn, this could facilitate Hg accumulation in Arctic marine food webs. By using fatty acid markers unique for the Arctic copepod Calanus glacialis, Routti et al. (2012) found a coherence between high concentrations of Hg in polar bears feeding on a diet with a high influence of this Arctic copepod. Selenium (Se) is an essential element and is therefore physiologically regulated within an optimal range. However, as Hg concentrations increase Se accumulation is enhanced by the formation of Hg–Se complexes, a process that protects cells and organisms against Hg toxicity (both organic and inorganic Hg; Khan and Wang, 2009). Similar to Hg, Cd accumulates in Arctic seabirds despite low emission within the Arctic (Dietz et al., 1996; Savinov et al., 2003). Arctic atmospheric Cd concentrations are low (Berg et al., 2008), suggesting that most Cd found in Arctic marine biota are derived from geochemical processes of the ocean (MacDonald et al., 2000). Cd is naturally more abundant in the North Pacific than in the North Atlantic (by a factor of five), resulting in a net flux of Cd from the Pacific to the Atlantic side of the Arctic Ocean (Bruland and Franks, 1983; MacDonald et al., 2000). This may lead to higher Cd concentrations in Arctic water compared to Atlantic water. Cadmium is easily adsorbed to the chitinous surface of zooplankton, facilitating uptake in fish and birds (Dietz et al., 1996). In a scenario of a changing climate, the species composition of the marine Arctic may shift towards Atlantic species (FalkPetersen et al., 2007; Eriksen and Dalpadado, 2011). Such a shift has been observed in Kongsfjorden (Svalbard), due to unusually high influxes of Atlantic water (Willis et al., 2008; Gabrielsen and Hop, 2009). The change has persisted, as shown by abnormally large amounts of Atlantic species (Atlantic cod (Gadus morhua), capelin (Mallotus villosus), euphausiids and Calanus finmarchicus) in Kongsfjorden during the years 2007–2011 (Gabrielsen and Hop, 2009; Buchholz et al., 2012). During parts of this period black-legged kittiwakes (Rissa tridactyla) fed on capelin (Mallotus villosus) rather than their traditionally more important prey, the polar cod (Boreogadus saida), in Kongsfjorden (Gasbjerg, 2010). Element concentrations correlate to trophic position (TP) in several Arctic seabird species (Atwell et al., 1998; Braune et al., 2005; Ricca et al., 2008). Studies of bioaccumulation in relation to diet commonly apply the ratios between stable isotopes of nitrogen (15N/14N) to determine TP as the heavy 15N is enriched relative to the lighter 14N in predators. (Minagawa and Wada, 1984; Hobson and Welch, 1992; Atwell et al., 1998; Power et al., 2002). The food web baseline is defined by d15N in organisms at the first trophic position (TP1), i.e. primary producers. Alternatively, d15N of primary consumers can be used to as a proxy for the second trophic position (TP2). Baseline d15N varies between locations and seasons, making it important to correct for the baseline when comparing accumulation of elements in different food webs (Cabana and Rasmussen, 1996). In Arctic food webs, the Calanus species may be used as TP2 organisms to estimate TP higher in the food web (Hobson and Welch, 1992). The carbon source of a food web can be derived from the ratio between stable isotopes of carbon (13C/12C) (DeNiro and Epstein, 1978; Post, 2002; Søreide et al., 2006). Enrichment of 13C in inshore environments reflects a carbon source of terrestrial origin or a stronger coupling to the benthic environment (Hobson, 1999). Mercury accumulation may in some cases be explained by the carbon source of marine food webs (Ricca et al., 2008; Bond and Diamond, 2009). Little auks (Alle alle) and black-legged kittiwakes are key species in Arctic marine ecosystems. The kittiwake is an opportunistic feeder preying on small pelagic fish like polar cod and capelin, as well as pelagic amphipods (Themisto spp.) and euphausiids (Thysanoessa spp.) (Mehlum and Gabrielsen, 1993). In contrast, little auks are specialist feeders, preying mainly on pelagic zooplankton like
copepods, Themisto spp., Thysanoessa spp. and Mysis spp. (Mehlum and Gabrielsen, 1993). Copepods of the Calanus genus dominate the diet of little auks in the chick rearing period (Gabrielsen et al., 1991; Steen et al., 2007). Hence, kittiwake and little auk represent different trophic levels in the partial food web of the Arctic pelagic ecosystem. The aim of the present study was to evaluate the influence of diet, feeding location and sampling year on accumulation of Hg, Cd and Se in Arctic pelagic seabirds. Samples of kittiwakes and little auks were collected in Liefdefjorden and Kongsfjorden at Svalbard. Kongsfjorden is dominated by Atlantic water while Liefdefjorden is dominated by Arctic water (Hallanger et al., 2011a). Furthermore, d15N and d13C were analyzed in muscle tissue of seabirds and in Calanus glacialis to examine the effects of TP and carbon source on element accumulation and biomagnification in the two fjords. 2. Methods 2.1. Field sampling Kittiwakes and little auks were collected in the vicinity of two breeding colonies at Svalbard, one in Kongsfjorden (78°570 N, 12°120 E) and the other in Liefdefjorden (79°370 N, 13°200 E; Fig. 1). Sampling was carried out during the last two weeks of July 2008 and 2009. Ten free-ranging individuals of each species were collected outside the breeding colony at each location each year using a shotgun with non-toxic ammunition (steel/bismuth). Permission for birds sampling was granted by the Governor of Svalbard. Samples of the left pectoralis muscle and liver tissue were dissected out for stable isotope and element analysis, respectively. Sex and body mass was recorded. Calanus glacialis were collected for food web baseline reference in both fjords each year using Method Isaac Kidd (MIK) nets (1000 lm, whole water column). Only C. glacialis at developmental stage copepodite V (CV) was used in the assessment of trophic position. All samples were stored in plastic bags at 20 °C before transportation and analysis. 2.2. Element determination Frozen liver tissue was lyophilized in acid washed polypropylene tubes (2 mL; Nalgene cryotubes, Thermo Scientific, Waltham, MA USA) with the screw cap partially open for a minimum of 24 h prior to digestion (CIT2, Leybold–Heraeus, Köln, Germany). Dry samples (0.15 g) were transferred to PTFE-vials (18 mL; UltraClave, MLS GmbH, Leutkirch, Germany) together with 2.3 g ultrapure water (Q-option, Elga Labwater, Veolia Water Systems LTD, UK) and 4.2 g concentrated nitric acid, HNO3 (Scanpure, equal to ultrapure grade, Chem Scan, Elverum, Norway). Digestion was carried out using a high-pressure microwave emitter (Milestone Ultra Clave, EMLS, Leutkirch, Germany). The temperature was gradually increased from room temperature up to 250 °C within 1 h and thereafter gradually cooled down to the initial value within the next hour. The digested samples were diluted with ultrapure water in acid washed polypropylene vials (BD Falcon 50 mL conical, BD Biosciences, Bradford, MA, US) to a final volume of 60 mL and a final HNO3 concentration of 0.6 M. Total Hg, Cd and Se were determined by high-resolution inductively coupled plasma mass spectrometry (HR-ICP-MS, Thermo Finnigan model Element 2, Bremen, Germany). Instrument settings were described in detail by Sørmo et al. (2011). Samples were randomized with respect to sampling location and species. Three blank samples containing ultrapure water and HNO3 were prepared in the same way as the samples (0.6 M in final solution). To account for potential carryover during digestion, the PTFE-vials
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
3
Fig. 1. Map of Svalbard showing the locations for sampling of black-legged kittiwakes (Rissa tridactyla), little auks (Alle alle) and Calanus glacialis.
containing blank samples were alternated between each run. The detection limits based on the standard deviation of blank samples were 0.144, 0.013 and 0.012 lg g1 for Se, Cd and Hg, respectively. No samples were below the level of detection. The average relative standard deviations (RSD) of multiple scans were below 3% for all elements. Certified standard reference materials (SRM) bovine liver (NIST 1577b), oyster tissue (NIST 1566b) and chicken (GBW 10018) were digested and analyzed at regular intervals (n P 6 of each SRM). The analytical values of SRM were within the range of the certified values. The recovery of Se was 114%, 123% and 102% in bovine liver, chicken and oyster tissue, respectively. Mercury was certified in oyster tissue, with a recovery of 105%. Cadmium recovery in bovine liver and oyster tissue was 100% and 101%, respectively.
centration of 4 ng g1. However, the Cd concentrations in these samples did not increase with increasing Bi concentration. Hence, we consider the potential contamination from ammunition to be negligible. 2.4. Stable isotopes Stable isotopes were analyzed at the Great Lakes Institute for Environmental Research, Canada, according to the procedure by Hussey et al. (2010). Lipids were removed from all samples, and additionally carbonates were removed from the C. glacialis samples. Stable isotope ratios are expressed in d notation as the deviation from standard in parts per thousand (‰):
dX‰ ¼
2.3. Quality assurance of tissue sampling for element analysis In the field, the tissue samples were dissected out in a laboratory on board r/v Lance or at the Kings Bay Marine Laboratory. The whole liver was removed before it was divided into subsamples on a clean cutting board. The tissue samples were frozen in zip-lock bags that were pre-tested for potential element leakage with negative results. (The bags were equilibrated with MQ-water that was subsequently analyzed for trace elements using HR-ICPMS.) Prior to the freeze drying procedure, the bulk tissue sample collected in the field was transferred to a plastic-covered cutting board, where the outermost layer of tissue was removed using a scalpel. Subsamples were taken from the inner tissue to prevent potential contamination from the field to be included in the analyzed sample. The scalpel blade and the plastic cover were changed between each sample. The clean subsample were transferred to acid washed (0.6 M HNO3) polypropylene tubes with screw caps, that were only partially opened during the freeze drying procedure. The dry subsample was weighted directly in the PTFE tube in which they subsequently were digested to prevent direct handling after drying. Potential contamination from ammunition was examined by correlation plots between Cd and Bi and between Cd and Fe. No samples had deviating Fe concentrations. A few samples contained higher Bi concentrations (max 2 lg g1) than the median Bi con-
Rsample 1 1000 Rstandard
ð1Þ
where dX is the delta value of 13C and 15N (‰) and R is the molar ratio of 13C/12C or 15N/14N in the sample and in an international standard, respectively. Vienna Pee Dee Belemnite served as standard for isotopic ratios of carbon, and atmospheric air was used as standard for isotopic ratios of nitrogen. Replicate measurements of internal laboratory standards (muscle tissue of fish) generally run for every 10 sample, indicated measurement errors ranging from ±0.05 to ±0.07 SD for d13C and from ±0.13 to ±0.27 SD for d15N. A trophic enrichment factor of 2.4‰ for 15N between seabirds and their prey was assumed, resulting in the following equation: d15Nbird = d15Nprey + 2.4 (Mizutani et al., 1991). For the rest of the food web an enrichment factor of 3.8‰ was applied (Hobson and Clark, 1992). The trophic position for kittiwake and little auk was calculated using the d15N values of Calanus glacialis as TP2 by the following equation (each fjord and sampling year separate):
TPbird ¼ 3 þ
d15 Nbird ðd15 NC:glacialis þ 2:4Þ 3:8
ð2Þ
2.5. Data analysis Statistical analysis was performed using R (v2.13.1; R Development Core Team). The distribution of variables was tested using QQ-plots and Shapiro–Wilk normality test. Variables that deviated
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
4
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
from a normal distribution were log-transformed (kittiwake: Hg and Se:Hg; little auk: Hg, Cd, Se, and Se:Hg). Equality of variance was tested using Levene’s test in the ‘‘lawstat’’ package. All variables had a homogenous variance within the four groups (two locations, two years), except trophic position in little auk. Based on these results, parametric tests were applied. The level of significance was set to p < 0.05. However, p < 0.1 and p > 0.05, were applied to indicate trends in the results. Examination of model plots of the regression between log(Hg) and d13C in kittiwakes resulted in one individual kittiwake from Liefdefjorden 2009 being considered as an outlier (d13C: 21.96‰, Cook’s D > 0.5). This individual was removed from the dataset prior to statistical analysis involving d13C. Element concentrations did not differ between sexes in either species (Hg, Cd, Se: p > 0.05), thus, both sexes are included in data analysis. Analysis of variance (ANOVA) with Tukey’s HSD (honest significant difference) post hoc test was used to determine differences in mean levels of Hg, Cd, Se, Se:Hg, d15N, d13C and TP between groups (location x year) for little auks (n = 38) and kittiwakes (n = 40). In kittiwakes, ANOVA analysis of Hg and Cd revealed significant between-group differences (p < 0.05), but the TukeyHSD test did not display which groups were different due to significant interactions between year and locality (F = 9.00, p = 0.005). t-tests were applied for these groups. Linear models and t-tests were run to evaluate differences between the fjords and years. The influence of location and sampling year on separate variables (y = TP, dC, Cd, Hg, Se, Se:Hg) were tested separately using Eq. (3). Letters a–d are regression constants.
y ¼ a þ b Location þ c Year þ d Location Year
ð3Þ
Eqs. (4) and (5) were used to examine potential differences in trophic magnification potentials of Hg and Cd between locations and years for parts of the food web including both avian species (n = 78).
log½element ¼ a þ b TP þ c Location þ d ðTP LocationÞ
ð4Þ
log½element ¼ a þ b TP þ c Year þ d ðTP YearÞ
ð5Þ
Literature values of elements given on a wet weight basis were recalculated to dry weight using the following formula: [Me]wet weight/0.3 = [Me]dry weight, where Me is the element in question, and 0.3 is the approximate portion of dry matter in liver samples of kittiwakes and little auks from the present study (31.6 ± 1.53 %, n = 70). 3. Results and discussion 3.1. Species differences The concentrations of Hg, Cd (Fig. 2) and Se (Table 1) were higher in kittiwakes than in little auks independent of location and sampling year (TukeyHSD, p < 0.001). Previously reported data on element accumulation confirm these inter-species differences (Dietz et al., 1996; Savinov et al., 2003; Poissant et al., 2008; Jæger et al., 2009). As expected from the inter-species difference in diet (Mehlum and Gabrielsen, 1993), the kittiwakes were feeding at a higher TP than the little auks (3.1 and 2.7, respectively; t-test, p < 0.001, Fig. 2). The TP of both species were lower than previously reported (Jæger et al., 2009; Hallanger et al., 2011b), indicating that the TPs may be slightly underestimated when calculated based on C. glacialis (Eq. (2)). Predatory species usually accumulate higher levels of Hg than primary consumers, increasing the dietary exposure of birds as they feed higher up in the food chain (Dietz et al., 1996; Jæger
et al., 2009). A difference in TP of less than 1 between the species indicates that the kittiwakes have included lower trophic level prey in their diet as shown by Mehlum and Gabrielsen (1993). Despite this, Hg concentrations were two times higher in kittiwakes than in little auks (Fig. 2). The Hg composition of prey items, i.e. the relative amount of methyl mercury (MeHg), may have enhanced the inter-species differences in total Hg concentrations (Fig. 2). Due to its physical and chemical properties, ingested MeHg is more readily absorbed than inorganic Hg, while inorganic Hg is more efficiently eliminated than MeHg (Scheuhammer et al., 2008). MeHg is therefore more efficiently transferred in the food chain, resulting in a larger percentage of MeHg in fish species like polar cod (90–100%) compared to the Calanus species (7.5–30%; Campbell et al., 2005; Loseto et al., 2008; Jæger et al., 2009). Hence, the opportunistic feeding habit of kittiwakes may result in a more diet-dependent variation in Hg accumulation relative to the specialist feeder little auk. 3.2. Annual and spatial differences Mercury concentrations in little auks were higher in Kongsfjorden than in Liefdefjorden in both years (2008: p = 0.025; 2009: p = 0.069; Fig. 2). In kittiwakes, the Hg concentrations were highest in Liefdefjorden in 2008 and in Kongsfjorden in 2009 (2008: p = 0.027; 2009: p = 0.080). Likewise, Cd concentrations varied between years in both species (Fig. 2). These between-year variations suggest that between-fjord differences are determined by other factors than simply higher Hg and Cd accumulation in food webs of Arctic waters (MacDonald et al., 2000; Routti et al., 2012). Mercury was positively correlated with TP in kittiwakes (Pearson, R = 0.352, p = 0.030), indicating that the inter-annual variation in Hg concentrations were linked to between-year shifts in the TP of kittiwakes (TukeyHSD, p < 0.01, Fig. 2). In contrast, in little auks, no between-year differences in Hg accumulation were observed, despite significant variation in TP between fjords and years (TukeyHSD, p < 0.001, Fig. 2). Stable isotope signals have been shown to vary inter-annually in Arctic seabirds as a consequence of variable prey availability (Moody et al., 2012). Inter-annual variation in d15N in kittiwakes was attributed to the ice cover dependent availability of polar cod (Moody et al., 2012). Similar variations in prey availability may explain the between-year variation in TP of the kittiwakes in the present study. Indeed, polar cod was less frequent in the diet of kittiwakes from Kongsfjorden, and in recent years their feed has been increasingly dominated by capelin (Mehlum and Gabrielsen, 1993; Gasbjerg, 2010). In little auks, significant correlation was found between Hg and d15N (Pearson, R = 0.404, p = 0.013), but not between Hg and TP (p = 0.435). Little auks prey exclusively on Calanus spp. during the breeding season (Gabrielsen et al., 1991; Steen et al., 2007). Lower d15N values were found in Calanus glacialis in Liefdefjorden 2008 compared to the other sampling groups (p < 0.001, Table 1), indicating that the differences in TP may rather be explained by a shift in baseline d15N. Hence, the inter-annual variation in Hg accumulation between the two fjords in kittiwakes may be explained by the shift in TP between the years, while in little auks this was not the case (Fig. 2). The elevated Hg concentrations in little auks from Kongsfjorden relative to Liefdefjorden (Fig. 2) may also be a result of local geological differences or long range oceanic transport resulting in higher input of Hg to Kongsfjorden (Lu et al., 2013). Correlations between Hg and d13C have been found in seabirds from areas where the terrestrial input of Hg is significant (Bond and Diamond, 2009). Little auks from Kongsfjorden were enriched in 13C compared to little auks from Liefdefjorden (both years: t-test, p = 0.003; Fig. 3). Indeed, a positive correlation was found between Hg and d13C in little auks (Pearson, R = 0.303, p < 0.001, n = 37, Fig. 4), indicating the influence of Hg of terrestrial origin. However, natural positive relationships
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
5
Fig. 2. Trophic position (TP), hepatic mercury (Hg) and cadmium (Cd) concentrations (lg g1 ± SE dw) in kittiwakes (Rissa tridactyla; left) and little auks (Alle alle; right) from Kongsfjorden and Liefdefjorden sampled in 2008 (light) and 2009 (dark). Differences between groups within each species (location x years) are denoted with different letters (ANOVA and TukeyHSD, p < 0.05; Hg and Cd in kittiwakes: t-test, p < 0.05).
between d15N and d13C make it challenging to assess the influence of d13C alone on element accumulation (Bond and Diamond, 2009). No correlations were found when the residuals from the regression
between d13C and d15N were regressed against Hg. Similarly, despite significant enrichment of d13C in kittiwakes from Kongsfjorden relative to Liefdefjorden (19.91‰ and 20.54‰, respectively;
Please cite this article in press as: Øverjordet, I.B., et al. Effect of diet, location and sampling year on bioaccumulation of mercury, selenium and cadmium in pelagic feeding seabirds in Svalbard. Chemosphere (2014), http://dx.doi.org/10.1016/j.chemosphere.2014.10.060
6
I.B. Øverjordet et al. / Chemosphere xxx (2014) xxx–xxx
Table 1 Hepatic concentrations of selenium (Se; lg g1 dry weight) and molar ratio of Hg:Se in kittiwakes (Rissa tridactyla) and little auks (Alle alle) from Svalbard. Stable isotopes (dC and dN, ‰) in muscle tissue of birds and whole Calanus glacialis. Mean ± SE. n = 10 unless otherwise stated. Statistically different groups are assigned different letters (TukeyHSD; p < 0.05). ANOVA: ⁄⁄ = p < 0.01, ⁄⁄⁄ = p < 0.001. Kongsfjorden
Kittiwake Se Se:Hg d13C d15N
A B C D
lg g1 dw molar ratio ‰ ‰
2009
2008
Mean ± SE 16.37 ± 2.01 25.58 ± 5.58 20.12 ± 0.07 11.46 ± 0.14
– – a a
Mean ± SE 16.48 ± 1.81 17.09 ± 1.75 19.71 ± 0.14 12.22 ± 0.12
– – b b
Mean ± SE 16.25 ± 1.21 17.37 ± 2.87 20.46 ± 0.04 12.25 ± 0.18
– – c b
Mean ± SE 13.62 ± 1.43 17.98 ± 1.74 20.63 ± 0.09 11.76 ± 0.21
– – c ab
p (ANOVA) 0.295 0.436
