linking mercury exposure to habitat and feeding behaviour in beaufort sea beluga whales

13
Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales L.L. Loseto a, , G.A. Stern b,c , D. Deibel d , T.L. Connelly d , A. Prokopowicz e , D.R.S. Lean f , L. Fortier e , S.H. Ferguson a,b a Department of Zoology, University of Manitoba, 500 University Cres., Winnipeg MB, Canada R3T 2N2 b Freshwater Institute/Fisheries and Oceans Canada, 501 University Cres., Winnipeg MB, Canada R3T 2N6 c Department of Environment & Geography, University of Manitoba, 500 University Cres., Winnipeg MB, Canada R3T 2N2 d Ocean Sciences Centre, Memorial University, St John's NL, Canada A1C 5S7 e Dept. de Biologie, Université Laval, Pavillon Vachon, Quebec QC, Canada G1K 7P4 f Department of Biology, University of Ottawa, 30 Marie Curie Ottawa ON, Canada K1N 6N5 Received 16 April 2007; received in revised form 24 September 2007; accepted 12 October 2007 Available online 24 October 2007 Abstract Mercury (Hg) levels in the Beaufort Sea beluga population have been increasing since the 1990's. Ultimately, it is the Hg content of prey that determines beluga Hg levels. However, the Beaufort Sea beluga diet is not understood, and little is known about the diet Hg sources in their summer habitat. During the summer, they segregate into social groups based on habitat use leading to the hypothesis that they may feed in different food webs explaining Hg dietary sources. Methyl mercury (MeHg) and total mercury (THg) levels were measured in the estuarine-shelf, Amundsen Gulf and epibenthic food webs in the western Canadian Arctic collected during the Canadian Arctic Shelf Exchange Study (CASES) to assess their dietary Hg contribution. To our knowledge, this is the first study to report MeHg levels in estuarine fish and epibenthic invertebrates from the Arctic Ocean. Although the Mackenzie River is a large source of Hg, the estuarine-shelf prey items had the lowest MeHg levels, ranging from 0.1 to 0.27 μg/g dry weight (dw) in arctic cisco (Coregonus autumnalis) and saffron cod (Eleginus gracilis) respectively. Highest MeHg levels occurred in fourhorn sculpin (Myoxocephalus quadricornis) (0.5 μg/g dw) from the epibenthic food web. Beluga hypothesized to feed in the epibenthic and Amundsen Gulf food webs had the highest Hg levels matching with high Hg levels in associated food webs, and estuarine-shelf belugas had the lowest Hg levels (2.6 μg/g dw), corresponding with the low food web Hg levels, supporting the variation in dietary Hg uptake. The trophic level transfer of Hg was similar among the food webs, highlighting the importance of Hg sources at the bottom of the food web as well as food web length. We propose that future biomagnification studies incorporate predator behaviour with food web structure to assist in the evaluation of dietary Hg sources. © 2007 Elsevier B.V. All rights reserved. Keywords: Amundsen Gulf; Delphinapterus leucas; Diet; Biomagnification 1. Introduction Mercury (Hg) in the form of methyl mercury (MeHg) bioaccumulates in organisms over time, and biomagnifies at each trophic level (Morel et al., 1998). As a result, Hg Available online at www.sciencedirect.com Journal of Marine Systems 74 (2008) 1012 1024 www.elsevier.com/locate/jmarsys Abbreviations: Hg, Mercury; THg, total mercury; MeHg, methyl mercury; BMF, biomagnification factor. Corresponding author. Tel.: +1 204 984 2425; fax: +1 204 984 2403. E-mail address: [email protected] (L.L. Loseto). 0924-7963/$ - see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.jmarsys.2007.10.004

Upload: ll-loseto

Post on 07-Sep-2016

212 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

Available online at www.sciencedirect.com

74 (2008) 1012–1024www.elsevier.com/locate/jmarsys

Journal of Marine Systems

Linking mercury exposure to habitat and feeding behaviour inBeaufort Sea beluga whales

L.L. Loseto a,⁎, G.A. Stern b,c, D. Deibel d, T.L. Connelly d, A. Prokopowicz e,D.R.S. Lean f, L. Fortier e, S.H. Ferguson a,b

a Department of Zoology, University of Manitoba, 500 University Cres., Winnipeg MB, Canada R3T 2N2b Freshwater Institute/Fisheries and Oceans Canada, 501 University Cres., Winnipeg MB, Canada R3T 2N6

c Department of Environment & Geography, University of Manitoba, 500 University Cres., Winnipeg MB, Canada R3T 2N2d Ocean Sciences Centre, Memorial University, St John's NL, Canada A1C 5S7

e Dept. de Biologie, Université Laval, Pavillon Vachon, Quebec QC, Canada G1K 7P4f Department of Biology, University of Ottawa, 30 Marie Curie Ottawa ON, Canada K1N 6N5

Received 16 April 2007; received in revised form 24 September 2007; accepted 12 October 2007Available online 24 October 2007

Abstract

Mercury (Hg) levels in the Beaufort Sea beluga population have been increasing since the 1990's. Ultimately, it is the Hg content ofprey that determines beluga Hg levels. However, the Beaufort Sea beluga diet is not understood, and little is known about the diet Hgsources in their summer habitat. During the summer, they segregate into social groups based on habitat use leading to the hypothesis thatthey may feed in different food webs explaining Hg dietary sources. Methyl mercury (MeHg) and total mercury (THg) levels weremeasured in the estuarine-shelf, Amundsen Gulf and epibenthic foodwebs in the western Canadian Arctic collected during the CanadianArctic Shelf Exchange Study (CASES) to assess their dietary Hg contribution. To our knowledge, this is the first study to report MeHglevels in estuarine fish and epibenthic invertebrates from the Arctic Ocean. Although the Mackenzie River is a large source of Hg, theestuarine-shelf prey items had the lowest MeHg levels, ranging from 0.1 to 0.27 μg/g dry weight (dw) in arctic cisco (Coregonusautumnalis) and saffron cod (Eleginus gracilis) respectively. Highest MeHg levels occurred in fourhorn sculpin (Myoxocephalusquadricornis) (0.5μg/g dw) from the epibenthic foodweb. Beluga hypothesized to feed in the epibenthic andAmundsenGulf foodwebshad the highest Hg levels matching with high Hg levels in associated food webs, and estuarine-shelf belugas had the lowest Hg levels(2.6μg/g dw), correspondingwith the low foodwebHg levels, supporting the variation in dietaryHg uptake. The trophic level transfer ofHgwas similar among the foodwebs, highlighting the importance ofHg sources at the bottom of the foodweb aswell as foodweb length.We propose that future biomagnification studies incorporate predator behaviour with food web structure to assist in the evaluation ofdietary Hg sources.© 2007 Elsevier B.V. All rights reserved.

Keywords: Amundsen Gulf; Delphinapterus leucas; Diet; Biomagnification

Abbreviations: Hg, Mercury; THg, total mercury; MeHg, methylmercury; BMF, biomagnification factor.⁎ Corresponding author. Tel.: +1 204 984 2425; fax: +1 204 984 2403.E-mail address: [email protected] (L.L. Loseto).

0924-7963/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.jmarsys.2007.10.004

1. Introduction

Mercury (Hg) in the form of methyl mercury (MeHg)bioaccumulates in organisms over time, and biomagnifiesat each trophic level (Morel et al., 1998). As a result, Hg

Page 2: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

1013L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

concentrations at trace levels inwater can reach toxic levelsin top predators. For example, high Hg levels have beenfound in the Beaufort Sea beluga whale (Delphinapterusleucas) population that spend summers in the easternBeaufort Sea and Mackenzie Delta region of the Canadianwestern Arctic. Elevated Hg levels represent a risk tobeluga, and predators of beluga such as polar bears andInuvialuit subsistence hunters. In the 1990′s, liver Hglevels in this beluga population tripled in comparison to1980 levels (Lockhart et al., 2005a), and were the highestrelative to other Canadian Arctic beluga populations.Although, still higher than the 1980 levels, Hg concentra-tions have dropped and are nowcomparable to otherArcticpopulations (Lockhart et al., 2005a). Apart from of highHg levels, this is a large beluga population, numberingover 20,000 (Harwood et al., 1996).

Due to the biomagnifying properties of MeHg, Hgconcentrations in predators such as beluga largely reflectthe Hg levels in their diet (Mathers and Johansen, 1985).However, little is known about the Beaufort Sea belugadiet, thus dietary Hg sources cannot be accounted for.Typically in animal diet studies, stomach contents andfeces are used to identify diet items. This is not feasiblewith the Beaufort Sea beluga because harvested whalesoften have empty stomachs, and feces cannot be found,yet local hunters have observed beluga feeding duringsummer (Harwood and Smith, 2002).

Stable isotopes are useful diet biomarkers that provideinformation about animal feeding preferences and canovercome problems associated with conventional dietdetermination such as over-or under-representation ofprey that were recently eaten or quickly digested (Tollitet al., 1997). Stable isotopes have been used in previousstudies to examine diet and the trophic level transfer ofHg (Atwell et al., 1998; Dehn et al., 2006b). At highertrophic levels, variability in top predator Hg wasobserved and attributed to diversity in foraging behav-iour resulting from intra-species differences in sex orsize, in addition to variation in seasonal prey abundance(Atwell et al., 1998). Atwell et al. (1998) found Hg levelsin top predator muscle tissue may better reflect diet andto a lesser extent Hg bioaccumulation with age, sup-porting the need to understand dietary sources of Hgand biomagnification.

The Canadian Arctic Shelf Exchange Study (CASES)provided an opportunity to sample regions where theBeaufort Sea belugas spend summer, such as the easternBeaufort Sea, Amundsen Gulf, Franklin Bay and theMackenzie Delta. Habitat use of Beaufort Sea belugasdiffers with length, sex and reproductive status likelyreflecting social structure (Loseto et al., 2006). Variationin beluga habitat selection suggests diet may differ,

which can result in differing dietary Hg uptake. Due tothe diversity of possible Beaufort Sea beluga prey, avariety of food items as well as lower trophic levelorganisms, were collected during the CASES expedition.

Here we propose an approach to determine belugaprey and resulting Hg levels that incorporates belugabehaviour and the complexity of the eastern BeaufortSea ecosystem. First, we examine Hg levels in organ-isms in three food webs in the beluga summering rangeto establish if dietary sources of Hg differ. Second, wedetermine if Hg levels differ among beluga feedinggroups defined a priori by their length, sex and repro-ductive status. Finally, beluga feeding groups are pairedwith food webs that best fit their habitat use, and usingHg and stable isotopes, beluga Hg biomagnificationfactors (BMFs) are calculated and used as a tool to in-crease our understanding of beluga feeding.

2. Material and methods

2.1. Sample collection

Food web samples were collected during 2002, 2003,2004 and 2006 from the CCGS Amundsen, Nahidik andPierre Radisson in the Western Arctic region, encom-passing the Mackenzie Delta, Amundsen Gulf, FranklinBay, and the eastern Beaufort Sea (Fig. 1). All biotasamples were frozen at −20 °C after collection andshipped to Fisheries & Oceans Canada in Winnipegwhere they were stored at −20 °C until analysis.

2.1.1. ZooplanktonCalanus spp. and Themisto libellula were collected

from the CCGS Amunsden during CASES from Septem-ber 2003 to September 2004. The one-year samplingexpedition included a fall survey covering the entire studyarea, an over-wintering of the ship in Franklin Bay, and asummer/autumn survey of the Amundsen Gulf andMackenzie Shelf (Fig. 1). During the open water season,integrated vertical tows were taken with a double tuckertrawl (200 μm mesh and 500 μm mesh) and depth-stratified samples were taken with a Hydrobios multi-net(200 μm mesh). During the over-wintering period wesampled under the sea ice using a horizontal double tuckertrawl (200 μm mesh and 500 μm mesh).

Calanus spp. were picked from the samples and placedinto plastic (Whirlpak®) bags and frozen at −20 °C.Additional T. libellula samples were collected during the2002 fall cruise of the CCGS Pierre Radisson. T. libellulawere identified, sexed, measured, staged and dried at60 °C. Mercury analysis was completed for T. libellulacollected in 2002 and 2003/2004, and stable isotope

Page 3: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

Fig. 1. Study Area: Amundsen Gulf, Franklin Bay,Mackenzie Delta, eastern Beaufort Sea. The 20 m isobath north of theMackenzie outflow demarks theestuarine region, and the 200m isobath separates theMackenzie shelf from the eastern Beaufort Sea. The short horizontal lines represent the estuarine-shelffood web collection region, and the vertical lines represent the Amundsen Gulf food web region. Epibenthic food web organisms were collected from theepibenthos of both regions. Tuktoyaktuk, Northwest Territories, is the location of the beluga harvest and tissue sample collection.

1014 L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

analyses were completed only for 2002 samples. Zoo-plankton samples from the shallow Mackenzie Deltaregion (b20 m depth) were collected from the CCGSNahidik in the late summer of 2003 and 2004.Zooplankton were not sorted prior to being placed intoplastic (Whirlpak®) bags and frozen at −20 °C. Analysisof relative species abundance revealed that the twodominant zooplankton species were omnivorous Pseu-docalanus spp. and Calanus glacialis (65% and 26%respectively).

2.1.2. Epibenthic InvertebratesEpibenthic animals were collected in the fall of 2003

and spring/summer of 2004 with a modified MACER-GIROQ sled that collected organisms living directly onandwithin 60 cm of the seafloor (Choe andDeibel, 2000).The sled was equipped with a 500-μm net, a partiallyclosed cod end, and a door that opened when the sled wason the sea floor and closed while in the water column.As soon as the sled came on board, the contents of the codend were gently rinsed into coolers. Cooler contents werethen rinsed with surface water to remove mud from the

samples. The decapods Eualus spp. and Bythocaris spp.(which will be referred to as “shrimp”), the amphipodsAnonyx spp. and Acanthostepheia malmgreni, as well asfour pooled mysid genera (Psuedamma spp., Erythropsspp., Mysis spp., Michthyops spp.) were picked from thesamples and identified onboard the ship.

2.1.3. FishArctic cod (Boreogadus saida) were collected from

March 15 to May 27 2004. Most of the samples werecollected in Franklin Bay using a 90-m long gill net setvertically from the bottom (i.e., 230 m), where most fishcaptured were 100 m above the ocean floor. In addition,arctic cod were collected in September 2006 from theCCGS Nahidik between 10 and 100 m depths north ofthe Mackenzie River outflow. Only adult arctic cod wereselected for analysis (length N110 mm).

Fish in the brackish water of the Mackenzie Delta werecollected from the shoreline out to the 20 m isobath viacommunity-based sampling programs and the CCGSNahidik. Species collected included pacific herring(Clupea palasii), rainbow smelt (Osmerus mordax), arctic

Page 4: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

1015L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

cisco (Coregonus autumnalis), least cisco (Coregonussardinella) and saffron cod (Eleginus gracilis). Species char-acteristic of the epibenthic habitat, such as the starry flounder(Platichthys stellatus), arctic flounder (Pleuronectesglacialis) and fourhorn sculpin (Myoxocephalus quad-ricornis) were also collected in this region. For our studyall fish species will be referred to by their common names.Only adult fish were used (N200 mm).

2.1.4. BelugaBeluga tissue samples were collected from Tuktoyak-

tuk harvests in 2004 (n=26). Twenty-one belugas weremale and five were female, of which two were in theirfirst trimester of pregnancy and three were lactating,suggesting they had calves in attendance. Muscle tissuewas selected for Hg analysis because unlike liver tissuewhich contains largely elemental Hg, 97–100% of theTHg in muscle is MeHg (Wagemann et al., 1998). Moreimportantly, muscle may better reflect dietary sources ofHg, whereas liver is a site of MeHg demethylation(Wagemann et al., 1998). Ages were determined from athin section of a tooth by counting growth layer groups inthe dentine (Stewart et al., 2006).

2.2. Food webs: estuarine-shelf, Amundsen Gulf,epibenthic

Organisms were placed into an estuarine-shelf,Amundsen Gulf or epibenthic food web group based oncollection regions and their feeding ecology. Theestuarine-shelf food web combined samples that repre-sented an estuarine environment influenced by thebrackish water of theMackenzie River (samples collectedin the delta b20 m depth) and samples from deeperregions of the delta, reaching the shelf break (20 to 200 mdepth). Although processes occurring within the estuarymay be different than in the deeper shelf waters, theseareas were combined due to the important influence of theMackenzie River on energy and Hg sources to the localfood web (Leitch et al., 2007). Samples collected in theshallow estuary region included rainbow smelt, saffroncod, pacific herring, least cisco, arctic cisco and zoo-plankton (65% Pseudocalanus spp., and 26%C. glacialisand 26%). The estuarine habitat is diverse, offering a largegradient of salinity, temperature and depth. This food webincludes first level consumers and plankivours fishfollowed by beluga. Samples collected in deeper shelfwaters included the first level consumerCalanus spp., thepredator T. libellula and arctic cod (Fig. 1).

The Amundsen Gulf food web represents an offshelf,pelagic ecosystem. This food web is characterized byCalanus spp., T. libellula and arctic cod collected from

Franklin Bay, Amundsen Gulf and Eastern Beaufort Sea(Fig. 1). Thus, this food web is similar in structure to theshelf, beginning with copepods, and then carnivorousmacrozooplankton T. libellula, followed by planktivoursfish that may feed both zooplankton or just one,followed by beluga the top predator.

Lastly, the epibenthic food web included invertebratesand fish from the near bottom environment throughout theCASES study region. Sculpin, starry and arctic flounderwere collected in the estuary, but were placed within theepibenthic food web because their food Hg sources arespecific to the near bottom environment and differ fromthe estuarine group. The four epibenthic invertebratesincluded shrimp, two amphipod species and mysids(see above). We were unable to clearly identify a firstconsumer level (i.e. herbivores) in this food web, andselected the mysids because many species are known tofeed on fresh phytodetritus shortly after it sinks to thebottom (Richoux et al., 2004).

2.3. Beluga feeding groups and food web pairing

Animal length, age, sex and body condition shouldbe considered when examining predator Hg levelsbecause they influence feeding behaviour (Atwell et al.,1998; Dehn et al., 2006b). Based on sex, length,reproductive status, the satellite tagged (1993–1997),Beaufort Sea beluga segregated during summer intothree separate habitats that differed in sea ice concen-tration, bathymetry and distance from the coast (Losetoet al., 2006). Three beluga habitat groups were defined:1) shallow open-water near the mainland was selectedby females with and without calves and by small males(b4 m); 2) the sea ice edge was selected by mediumlength males (3.8–4.3 m) and a few females (N3.4 m)without neonates; and 3) heavy sea ice concentrations indeep, offshore waters were selected by the largest males(4–4.6 m). This type of habitat segregation suggests acomplex beluga social structure, and indicates where theBeaufort Sea belugas are feeding during the summerseason in the western Arctic.

Length of beluga among the three habitat groupsoverlapped. Thus, Beaufort Sea beluga growth curves(Luque and Ferguson, 2006) were used to definedifferences (Table 1). For females, the adult asymptoticlength of 3.7 m was used to demarcate smaller females(with and with out calves) using shallow open-water andlarger animals selecting the sea ice edge (Table 1). Themaximum length of the smallest males selecting theshallow open-water habitat was based on the mid-pointon the growth curve between juveniles and adults(3.8 m). Lastly, the adult asymptotic length of 4.2 m was

Page 5: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

Table 1Lengths of female and male belugas in the three habitats use groups taken from Loseto et al. (2006)

Length boundaries Number of whales (length range in m)

Habitat use groups Males Females Males Females Feeding groups

Shallow open water b3.8 m b3.7 3 (3.51–3.56) 4 (3.4–3.76) Estuarine–ShelfSea ice edge 3.8–4.2 N3.7 8 (3.89–4.16) 1 (3.81) Amundsen GulfClosed sea ice N4.2 n/a 10 (4.22–4.55) 0 Epibenthic

Number of whales from 2004 and the length ranges used to create feeding groups based on growth length curves (Luque and Ferguson, 2006). Belugafeeding groups were matched with food webs.

1016 L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

used to separate large whales selecting heavy sea iceconcentrations from smaller males selecting the sea iceedge (Table 1).

We paired the beluga feeding groups with hypothet-ical, habitat-specific food webs according to their spatialsegregation, i.e.: 1) beluga using shallow, open-waterhabitats likely use the Mackenzie Delta extensively andwere therefore paired with the estuarine-shelf food web;2) beluga selecting the ice edge are likely feeding on thesea ice associated arctic cod (Gradinger and Bluhm,2004) and/or in pelagic food webs, thus we paired themwith the Amundsen Gulf food web that includedorganisms collected offshelf in the pelagic environment;and 3) beluga selecting high levels of sea ice concen-tration do so in deep waters where they dive up to 800 m(Richard et al., 1997), likely feeding benthically, andtherefore were paired with the epibenthic food web. Theassigned feeding groups are an oversimplification of thetemporal and spatial complexities involved in belugamovement and seasonal feeding preferences. However,we consider this approach of pairing whale groups withhabitat-specific food webs an important development inevaluating Hg dietary sources.

2.4. Total and methyl mercury extraction and analysis

Belugamuscle and fishmuscle tissuewere subsampledforMeHg and THg analysis. Whole individual T. libellulawere dried at 60°C and those weighing b0.01 g werepooled by station and age class before analysis. Freeze-dried shrimp and amphipods were analyzed individuallywhen enough biomass was available, otherwise similarsmall mysids were pooled by station before analysis.Generally, ten whole dried Calanus spp. were pooled bysample station for analysis. Due to low biomass for Ca-lanus spp. and T.libellula collected from the shelf region,theywere not analyzed forMeHg, and values presented infigures are back calculated based on the percent MeHgfound in Amundsen Gulf Calanus spp. and T. libelulla.All samples were weighed to approx 0.15 g for THganalysis. Sampleswere digestedwith a hydrochloric/nitric

acid mixture (Aqua Regia) heated to 90 °C. The digestedsamples were analyzed for THg by Cold Vapour AtomicAbsorption spectroscopy (CVAAS) (Armstrong andUthe, 1971). The detection limit was 0.005 μg/g.

For MeHg analysis, beluga and fish muscle tissuesamples were extracted according to the procedure of(Uthe et al., 1972). Wet tissue samples were homoge-nized in an acidic bromide and copper sulphate solution.This was extracted with toluene and partitioned into athoisulfate solution where the addition of potassiumiodide allowed back-extraction into toluene. The extractwas analyzed on a gas chromatograph with an electroncapture detector. Due to the low biomass and unknownrange of MeHg levels in invertebrate organisms amodified method was used that required less mass. Theextraction was completed using a dichloromethanephase rather than toluene (Cai, 2000). Samples wereanalyzed by capillary gas chromotography — atomicfluorescence spectrometry at the University of Ottawa(D. Lean).

Certified standard reference materials (CRM 2976,TORT-2, DOLT-2) were analyzed in duplicate in everyrun. Recovery within ten percent of the certified valueswas used as a batch validation for samples. Averagedifferences in duplicates for beluga, fishmuscle tissue andCalanus spp. was five, five and ten percent respectively.

2.5. Stable isotope analysis and food web calculations

Based on relative isotopic fractionation processes,δ15N can be used to describe trophic levels (Cabana andRasmussen, 1994) and δ13C can be used to evaluatefood web carbon sources enter (France, 1995). Toprepare for stable isotope analysis, beluga muscle tissueand whole Calanus spp. and epibenthic invertebrateswere freeze dried, while fish muscle tissue samples andindividual whole T. libellula were oven dried at 60 °C.Carbon and nitrogen isotopic analyses on the muscle(protein) were accomplished by continuous flow,isotopic ratio mass spectrometry (CF-IRMS) using aGV-Instruments® IsoPrime attached to a peripheral,

Page 6: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

1017L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

temperature-controlled, EuroVector® elemental analyz-er (EA) (University of Winnipeg Isotope Laboratory,UWIL). One-mg samples were loaded into tin capsulesand placed in the EA auto-sampler along with internallycalibrated carbon/nitrogen standards. Carbon and nitro-gen isotope results are expressed using standard delta (δ)notation in units of per mil (‰). The delta values ofcarbon (δ13C) and nitrogen (δ15N) represent deviationsfrom a standard such as

dsamplex ¼ Rsample=Rstandard

� �� 1� �� 1000 ð1Þ

where R is the 13C/12C or 15N/14N ratio in the sampleand the standard. The standards used for carbon andnitrogen isotopic analyses are Vienna PeeDee Belemnite(VPDB) and IAEA-N-1 (IAEA, Vienna), respectively.Analytical precision, determined from the analysis ofduplicate samples, was ±0.16‰ for δ13C and ±0.18‰for δ15N. Accuracy was obtained through the analysis oflaboratory standards used for calibration of results.

To estimate Hg biomagnification factors in a food webthe slope determined from logarithmically transformedTHg concentrations against δ15N or trophic level is used(e.g. Campbell et al., 2005).Given that Hg can biomagnify

Table 2Summary of mean (±SE) MeHg, THg, stable isotopes and morphometrics of

Organisms (code) MeHg (μg/g dw) THg (μg/g dw) δ15N (‰)

Estuarine-shelfZooplankton (EZP) 0.010±0.001 0.035±0.005 9.1±0.2Calanus spp. (CAL) 0.007±0.001 0.025±0.003 9.6±0.6T. libellula (TLB) 0.066±0.002 0.087±0.007 9.4±0.2Pacific herring (PHR) 0.111±0.003 0.123±0.005 12.4±0.5Arctic cisco (ACS) 0.104±0.003 0.133±0.009 12.3±0.1Least cisco (LCS) 0.125±0.009 0.133±0.009 12.3±0.4Rainbow smelt (RSM) 0.163±0.009 0.181±0.028 13.1±0.3Saffron cod (SCD) 0.273±0.008 0.308±0.029 13.7±0.2Arctic cod (ACD) 0.158±0.008 0.163±0.016 13.1±0.2

Amundsen GulfCalanus spp. (CAL) 0.011±0.001 0.032±0.002 9.4±0.2T. libellula (TLB) 0.095±0.019 0.127±0.019 10.2±0.2Arctic cod (ACD) 0.301±0.023 0.377±0.029 14.7±0.2

EpibenthicMysids (MYS) 0.032±0.013 0.081±0.009 10.9±0.4Shrimp (SHP) 0.228±0.026 0.316±0.061 13.5±0.2Anonyx spp. (ANX) 0.160±0.083 0.291±0.050 11.9±0.3A. malmgremi (AMI) 0.052±0.028 0.116±0.011 12.7±0.1Arctic flounder (AFR) 0.234±0.055 0.255±0.044 11.6±0.2Starry flounder (SFR) 0.226±0.090 0.277±0.073 11.5±0.2Sculpin (SCP) 0.534±0.046 0.587±0.076 16.1±0.9

“n” is the number of samples taken for THg/stable isotopes/MeHg. Fish agewas not available for sculpin.a MeHg concentrations back calculated based on levels in the same speci

differently at each trophic level, biomagnification factors(BMFs) are often used to describe the transfer of Hgbetween prey and predator using the formula below:

BMF ¼ Hgpredator=Hgprey� �

= TLpredator=TLprey

� � ð2Þ

This formula requires an estimate of the trophic level(TL) of both predator and prey. This is typically done bynormalizing δ15N toCalanus spp. δ15N values, assumingthat Calanus represents a consumer that feeds only onphytoplankton (Fisk et al., 2001), with the followingequation:

TLconsumer ¼ 2:0þ d15Nconsumer � d15NCalanus

� �=3:8

ð3Þ

The denominator of 3.8 represents the nitrogenisotopic fractionation from one trophic level to the next(Hobson and Welch, 1992). We argue that using 3.8 orany other number for the nitrogen fractionation valuemay bias determination of the true trophic levels becausewe do not fully understand the nitrogen fractionationprocess in the food webs investigated here. Therefore wecalculated the biomagnification factor by replacing the

organisms collected from three food webs in the eastern Beaufort Sea

δ13C (‰) n % MeHg Length (mm) Age

−25.6±0.3 8/2/0 28.0 a

−25.6±0.2 8/5/3 28.0−26.0±0.2 40/28/0 75.0 a

−23.7±0.2 29/28/5 90.2 207.1±2.0 5.1±0.6−22.9±0.1 30/15/5 78.0 311.5±6.0 5.6±0.5−22.2±0.4 20/10/5 94.1 285.4±8.3 7.6±0.8−22.9±0.1 10/10/5 89.9 203.6±15.9 6.1±0.9−22.8±0.3 20/20/5 88.8 280.1±10.7 5.8±0.5−22.3±0.3 20/20/5 97.2 143.3±3.1 2.6±0.2

−25.1±0.2 33/32/5 34.3−26.1±0.2 57/28/4 75.0−22.0±0.2 60/39/20 79.9 158.3±3.4 3.4±0.1

−23.0±0.4 15/15/4 39.5−21.5±0.7 14/3/7 72.3−21.4±0.0 6/6/6 55.0−21.2±0.4 2/3/2 45.0−23.6±0.2 9/9/5 91.9 226.4±10.6 9.1±0.6−24.4±0.5 11/11/5 81.5 246.8±13.5 9.0±1.2−23.0±0.7 5/2/5 91.0 219.8±32.3

data was obtained from otiliths (see Materials and methods). Age data

es in the Amundsen Gulf food web.

Page 7: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

Fig. 2. Mean stable isotope values for δ15N (‰) and δ13C (‰) (±SE)in organisms from the estuarine-shelf, Amundsen Gulf and epibenthicfood webs. For species abbreviations see Table 2. Superscript “A”represents Amundsen Gulf prey.

1018 L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

trophic levels in Eq. (2) with the raw δ15N values asfollows:

BMF ¼ Hgpredator=Hgprey� �

= d15Npredator=d15Nprey

� �:

ð4ÞUsing δ15N instead of trophic levels reduces the need

to know the enrichment factor of nitrogen or to assume aTL of 2.0 for pooled Calanus spp.

2.6. Statistical Analysis

Differences in mean values of MeHg, THg and stableisotopes among fish species and invertebrates weredetermined with an analysis of variance followed bypair-wise a posteriori Tukey tests. Comparison betweenshelf and Amundsen Gulf Calanus spp., T. libellula andarctic cod THg, and stable isotopes were analyzedindividually with paired t-tests. To test if collection siteaffected arctic codHg levels an analysis of covariancewas

Table 3Mean and standard error values for MeHg, THg (μg/g dw), and stable isoto

Beluga feeding groups n( f )

MeHg(μg/g dw)

THg(μg/g dw)

Estuarine-shelf 7 (4) 2.59±0.73a 2.56±0.8a

Amundsen Gulf 9 (1) 4.42±0.64b 4.41±0.7b

Epibenthic 10 6.03±0.61b 6.53±0.7b

F-ratio 7.95 8.45P value 0.002 0.002

Beaufort Sea belugas from 2004 harvests were partitioned into three groups bThe beluga groups are referred to by their hypothesized feeding behaviour asas the Amundsen Gulf group; and heavy ice concentration as the epibenthic gdesignated by different letters (‘a’ and ‘b’ symbols).

used to control for age. A general linearmodel was used toassess the effects of length, age and sex on Hg levels inbeluga muscle. Differences among beluga feeding groupsin length, age, THg, MeHg concentrations and stableisotopes were tested with an analysis of variance followedby pair-wise Tukey tests. Analysis of covariance wasemployed to evaluate differences in slopes among thethree food webs and between MeHg and THg. Allstatistical tests were carried out using Systat 11® (SystatSoftware Inc., 2004, San Jose, CA) on log normalizeddata. All data is presented as mean +/− standard error andall significant statistical tests are Pb0.05.

3. Results

3.1. Methyl and total mercury in three food webs

In fish muscle, MeHg averaged 88% of the THg, rang-ing from 78% in arctic cisco to 94% in least cisco (Table 2).In shrimp, MeHg averaged 72% of THg and ranged from65 to 80%, whereas Anonyx spp. ranged considerablyfrom 12 to 90% MeHg. MeHg in T. libellula represented71 to 100% of THg, and Calanus spp. had the lowestpercent of THg present as MeHg (34±5%). Theproportion of MeHg in Calanus spp. did not differsignificantly between the shelf and Amundsen Gulf foodwebs.

Estuarine fish had mean MeHg levels ranging from0.1 μg/g dw in pacific herring to 0.27 μg/g dw in saffroncod. Pacific herring, arctic cod, arctic cisco, leastcisco and rainbow smelt had similar MeHg, THg andδ15N values, that were lower than concentrations insaffron cod (Pb0.01) (Table 2; Fig. 2). Pacific herringδ13C values (−23.7±0.2 ‰) were lower than thosefor arctic cod and least cisco (Fig. 2). Mean MeHgand THg concentrations in beluga prey items, normal-ized by δ15N values, reveal that the estuarine-shelfbeluga prey had half the Hg of the Amundsen Gulfand epibenthic beluga prey (δ15N normalized values

pes (δ15N ‰, δ13C ‰) in beluga muscle tissue

δ15N(‰)

δ13C(‰)

Length(m)

Age

16.2±0.3a −18.6±0.3a 3.55±0.1a 28.0±2.9a

16.9±0.2b −18.6±0.2a 4.08±0.1b 31.9±2.6a

16.6±0.2ab −18.4±0.2a 4.33±0.1c 26.1±2.4a

4.44 0.51 96.88 1.380.023 0.600 b0.001 0.300

ased on length and sex parameters associated with habitat use (Table 1).follows: shallow open-water as the estuarine-shelf group; sea–ice edgeroup (Table 1). Significant differences from a posterior Tukey tests are

Page 8: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

Fig. 3. Biomagnification of MeHg and THg in three food webs: a)Estuarine-shelf, b) Amundsen Gulf and c) Epibenthic. Mean (±SE)MeHg (dashed line) and THg (solid line) (μg/g dw) versus δ15Ndepicting potential trophic relationships of species. For speciesabbreviations see Table 2.

1019L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

not shown). Zooplankton from the near shore estuaryhad higher, but not significantly different THg levelsthan did Calanus spp. collected in the deeper shelfwater.

Amundsen Gulf arctic cod had higher MeHg, THgand δ15N levels than estuarine-shelf fish species(Pb0.01), but THg and MeHg concentrations were notsignificantly higher than levels in saffron cod (Table 2).Although Amundsen Gulf and shelf arctic cod were thesame length, those from the shelf were younger and hadlower THg, MeHg and δ15N values (Pb0.05), yet δ13Cvalues were similar (Fig. 2). THg differences among theAmundsen Gulf and shelf arctic cod were not significantwhen controlled for age (PN0.05).

Within the Amundsen Gulf food web, Calanus spp.collected from the eastern Beaufort Sea near theMackenzie Delta shelf break had the highest meanTHg concentrations (0.055 μg/g dw±0.005; Pb0.05)relative to those collected in Franklin Bay (0.026 μg/gdw±0.002) and the Amundsen Gulf (0.026 μg/g dw±0.003). Total Hg in T. libellula did not differ amongthe two years sampled (PN0.05), thus samples fromboth years were combined. T. libellula from Amund-sen Gulf had significantly higher THg levels than didshelf T. libellula, yet the trophic increase in Hg levelsrelative to Calanus spp. did not correspond with atrophic enrichment in δ15N (e.g. 3‰) (Table 2).

Prey in the epibenthic food web had high MeHg andTHg levels, in some cases invertebrate MeHg concen-trations exceeded estuarine-shelf fish levels (Table 2).Sculpin had the highest Hg, and the δ15N value wassimilar to values in beluga. Stomach content analysis ofsculpin revealed a diet of amphipods and small fishsupporting the high Hg and δ15N values. Amonginvertebrates shrimp had the highest MeHg, THg andδ15N levels of potential prey items (Table 2).

3.2. Mercury levels of beluga feeding groups

In beluga muscle 99% of the THg was present asMeHg. Beluga length was the best predictor of MeHgconcentrations in muscle (Pb0.01), whereas sex andage did not have significant effects (PN0.05). Sex wascorrelated with length (r=−0.62; Pb0.01), with femalesshorter than males. Beluga feeding groups weresignificantly different in length, but not age, and belugalength and age had a weak correlation (r=0.2; PN0.1).The longest whales were not the oldest whales, andmean ages in beluga feeding groups were within severalyears of one another (Table 3).

Concentrations of THg, MeHg and δ15N values weresignificantly different among the beluga feeding groups

(Pb0.05) (Table 3). Estuarine-shelf beluga had thelowest MeHg and THg levels that were significantlylower than levels of the Amundsen Gulf and epibenthicbeluga groups (Table 3). The longest beluga whales thatwe hypothesized to feed epibenthically had the highestHg and MeHg levels, but not significantly higher thanlevels of the Amundsen Gulf group. The δ15N valuesdid not follow the Hg and beluga length trend becausehighest δ15N values were not associated with largestbeluga with highest Hg levels (Table 3). δ13C did notdiffer among the three whale groups.

Page 9: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

1020 L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

3.3. Trophic level transfer of methyl and total mercury

MeHg and THg trophic magnification slopes did notdiffer between the three food webs (PN0.1, Fig. 3). Inall three food webs MeHg biomagnification slopes weregreater than the THg slopes, and within each food web,MeHg and THg slopes did not differ (PN0.1). TheAmundsen Gulf food web had the highest rate of MeHgbiomagnification (Fig. 3b). The MeHg slope in theestuary-shelf food web was influenced by the low δ15Nin T. libellula (Fig. 3a). The range in δ15N values acrosssimilar Hg concentrations in the epibenthic food webresulted in a shorter food web length and greaterunexplained variance in Hg values than for the other twohabitats (Fig. 3c).

A large difference between the THg and MeHg BMFswas present at the lower end of the food web (Fig. 3) thatconverged at higher tropic levels. This was due to theincrease in the percent of MeHg up the food web.Calculated BMFs from prey to beluga were higher forMeHg than THg because it is the form that biomagnifies(Table 4). BMFs from estuarine-shelf fish to belugaranged from 9.3 in saffron cod to 20.5 in arctic cisco(Table 4). Arctic cod BMFs from the Amundsen Gulffood web fell within the range for estuarine-shelf fish to

Table 4Total and methyl mercury biomagnification factors (BMFs) in threefood webs measured between beluga and potential prey items withintheir paired groups (see Eq. (4))

Biomagnification factors (BMFs)

Beluga/prey MeHg Beluga/prey THg

Estuarine-shelfACS 19.00 14.64LCS 15.74 14.64RSM 12.75 11.48SCD 8.03 7.06PHR 17.86 15.91ACD 13.32 12.79Average 14.45 12.76

Amundsen GulfACD 12.77 10.20

EpibenthicSHP 39.08 16.87ANX 26.67 15.88AMI 88.50 43.10AFR 18.01 20.07SFR 18.51 16.32SCP 10.88 10.81Average 22.63 15.99

Refer to Table 2 for species abbreviations.

beluga (Table 4). The MeHg BMFs from potential epi-benthic prey items to beluga ranged considerably from12.4 in sculpin to 88.5 in a small amphipod (A.malmgreni)(Table 4).

4. Discussion

4.1. Methyl and total mercury in three food webs

The importance of determining beluga diet andforaging behaviour is demonstrated by the variability ofHg concentrations among food webs in the westernArctic. The percent of MeHg in arctic cod and beluga areconsistent with previous studies (e.g. arctic cod 100% inCampbell et al., 2005; 97% beluga in Wagemann et al.,1998). To our knowledge, this study is the first to reportMeHg levels in estuarine fish and epibenthic inverte-brates collected from the Arctic Ocean. The percentagesof MeHg for the estuarine fish are similar to levelstypically measured in fish (Morel et al., 1998), and THgin pacific herring and least cisco are within the rangereported during the 1990′s (Lockhart et al., 2005b).Shrimp MeHg content was in the lower range reported inBarents Sea shrimp (ca. 77–100%) (Joiris et al., 1997).Generally, the high percent MeHg in shrimp, Anonyx andT. libellula is comparable to fish, and thus negates thepossibility of using the MeHg:THg ratio as an indicatorof piscivory in belugas as was suggested for seals (Dehnet al., 2005). The low percentage of MeHg in Calanusspp. is common for suspension-feeding zooplankton(Watras and Bloom, 1992) and comparable to Calanusspp. in the Beaufort Sea (Stern and Macdonald, 2005)and in the Northwater Polynya (Campbell et al., 2005).

Despite high elemental Hg and MeHg output fromthe Mackenzie River plume (Leitch et al., 2007); fish inthe estuary-shelf habitat (excluding saffron cod) had thelowest Hg concentrations. The diet of the estuary-shelffish species is not fully known, and although they arealmost twice the size of Amundsen Gulf arctic cod theirlow δ15N and Hg levels suggest that they prey on lowertrophic level organisms. Calanus spp., T. libellula andarctic cod collections from the Amundsen Gulf andMackenzie shelf provided a common means for spatialcomparison of food web Hg levels. Mercury levels inAmundsen Gulf Calanus spp. are comparable toprevious reports (Stern and Macdonald, 2005), whereasHg in T. libellula were higher than reported in Campbellet al. (2005). Arctic cod from the Amundsen Gulffood web had higher Hg levels than did others in thesame size range reported in the literature (Atwell et al.,1998; Campbell et al., 2005; Stern and Macdonald,2005). Hg levels in shelf arctic cod were similar to

Page 10: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

1021L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

previous literature reports. Although differences amongAmundsen Gulf and shelf arctic cod Hg were largelydriven by age, their dominant prey items Calanus spp.and T. libelulla (Lonne and Gulliksen, 1989) demon-strated spatial differences in Hg and δ15N levels, withhigher values in the Amundsen Gulf, Franklin Bay andeastern Beaufort Sea region. The higher Hg and δ15N inAmundsen Gulf T.libellula slightly lengthened thisportion of the food web possibly causing AmundsenGulf arctic cod to have higher values. Given that arcticcod in the Amundsen Gulf are older; they maypredominantly feed on the macrozooplankton T.libellulathan on Calanus spp., thus, bioaccumulation with age inaddition to food web structure may drive the regionaldifferences in arctic cod Hg levels. Alternatively,regional differences in Hg sources may have causedthe differences observed.

High Hg and relatively lower δ15N in several epiben-thic organisms suggests different Hg uptake or trophiclevel transfer processes in the epibenthic food web.Epibenthic species feed in or near sediments, where Hgmethylation occurs (Bloom et al., 1999), exposing themto 106 times higher Hg levels than those feeding in thewater column (Morel et al., 1998). Mobile epibenthicanimals have access to organic matter falling throughthe water column, re-suspended from the benthos anddeposited on the sea floor. The quality of organicmaterial may be highly variable and the heterogeneousnature of available food sources complicates interpreta-tion of Hg sources and trophic transfer. Dead orpreviously processed organic material may containhigh MeHg or elemental Hg that can increase Hg fluxto these organisms, and start the food web at a higher Hglevel (Lindberg and Harris, 1974). This may explain thehigh and variable Hg levels among the epibenthicinvertebrates observed here and in previous studies(Lawrence and Mason, 2001).

High Hg outputs from the Mackenzie Delta (Leitchet al., 2007) did not result in highest Hg concentrations inthe estuarine-shelf food web, yet Hg from theMackenzieRiver was found to be an important source to theBeaufort Sea marine food web (Stern and Macdonald,2005). Those findings, and the high Hg levels in theAmundsen Gulf food web (that includes the easternBeaufort Sea), suggests that Hg outputs from theMackenzie River may only become bioavailable afterleaving the delta and entering the Eastern Beaufort Seaand surrounding regions. The majority of MeHg leavesthe Mackenzie River in a dissolved phase, whereaselement Hg is bound to particulate (Leitch et al., 2007),influencing mobility and food web uptake dynamics thatrequire further investigation.

4.2. Variability in beluga mercury levels

Variation in Hg among the beluga feeding groupsmay be explained by differing Hg levels in their diets inthe various habitats. The lack of a relationship betweenbeluga age and muscle Hg concurs with previous marinemammal observations that muscle Hg concentrationsmay better reflect Hg biomagnification through diet, andto a lesser extent bioaccumulation over time (Wagemannet al., 1990; Atwell et al., 1998). However, liver tissuenot examined here, does have a strong positiverelationship with age due to the continuous demethyla-tion of MeHg and accumulation of mercuric selenide(Wagemann et al., 1998). The beluga length and muscleHg relationship supports our use of length data to createhabitat-specific beluga feeding groups and to evaluateHg exposure through diet. The weak relationshipbetween age and length was due the beluga largelybeing over the asymptotic length of the Gompertz curve(Luque and Ferguson, 2006).

Low Hg and δ15N levels in the estuary-shelf belugafeeding group corresponded with the low Hg levels inthe estuarine-shelf food web, supporting this pairing.Mercury levels in the estuarine-shelf beluga group weresimilar to Hg muscle levels in Point Lay, Alaska, beluga(ca. 1.1 μg/g ww THg) (Dehn et al., 2006a). Fiftypercent of belugas in the estuarine-shelf group werefemales, which may confound interpretation of dietaryHg effects because Hg elimination during birthing maydecrease Hg levels in mothers. However, lactatingfemales had higher Hg levels than did the other females,indicating minimal effect of birthing on muscle Hglevels, as has been found in previous studies (Lockhartet al., 2005a).

The lack of a significant Hg difference between theepibenthic and Amundsen Gulf beluga groups may be aresult of the similarly high Hg levels found in the twofood webs. Consequently, differentiating between foodweb Hg contributions to these two beluga groups andassessing Hg uptake differences in diet is not feasiblegiven the commonly high Hg in both beluga and prey.Mercury levels in the epibenthic beluga group wassimilar to Beaufort Sea beluga levels in 2001 (Lockhartet al., 2005a), and δ15N were similar to Alaskan belugapopulations (Dehn et al., 2006a).

Although food webs differed structurally, the differ-ences in δ15N values among the three beluga groupswere within a range of 3‰, suggesting that they all fed ata similar trophic level, or at a combination of trophiclevels (Cabana and Rasmussen, 1994). Highest Hg levelsin the epibenthic beluga group did not associate with thehighest δ15N values, suggesting food web Hg sources

Page 11: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

1022 L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

associated with the sediment interface may be moreimportant in driving Hg levels than food web length.

4.3. Biomagnification in three Arctic food webs

Similar food web biomagnification slopes showedthat Hg biomagnification processes do not differ amongfood webs. It also demonstrates the importance of Hg,and specifically MeHg concentrations at the bottom ofthe food web, as well as food web length. If thebetween-trophic-level transfer of Hg remains consistentupward through a food web, then the initial Hgconcentration at the bottom of the web will constrainHg levels in top predators. In addition, lengthening thefood web by increasing the number of trophic levelswould result in higher concentrations of Hg in toppredators (Cabana and Rasmussen, 1994).

Total Hg biomagnification slopes were comparableto, but slightly higher than those reported in previousArctic aquatic studies (e.g. slopes: 0.197 in Campbellet al., 2005; 0.2 in Atwell et al., 1998; 0.19 in Poweret al., 2002). If Hg biomagnification is relativelyconstant over space and time, then we can use BMFsas tools to test how well the beluga social groups pairwith their habitat-specific potential prey.

Methyl mercury BMFs are more relevant than THgBMFs because it is the MeHg fraction that biomagnifiesup a food web. However, to our knowledge this studyreports the first MeHg BMFs to beluga. Thus, we areunable to draw comparisons to previous work. Total HgBMFs from arctic cod to beluga in this study werecomparable to the THg BMF value of 10 calculated forLancaster Sound beluga feeding on arctic cod usingδ15N and THg levels in Atwell et al. (1998). The similarBMF values provide support for our Amundsen Gulf-beluga food web pairing. Despite significant differencesin THg concentrations between estuary-shelf fish andAmundsen Gulf arctic cod, their THg BMFs to belugawere comparable. This suggests that the Hg trophictransfer from fish to beluga is predictable, and supportsour estuarine-shelf pairing. However, the THg BMF forsculpin was also comparable (ca. 10), but is not likelyimportant in the beluga diet given the similar δ15Nvalues. High BMFs for items such as A. malmgremisuggest they are not important to beluga diet. Theoverall higher BMFs in the epibenthic food web mayresult from their association with high Hg environments(e.g. sediments and detritus), or perhaps trophic leveltransfer processes from crustaceans to beluga is differentas compared with fish to beluga. For example, differentenergy content among diet items may alter the totalamount of food consumed and influence the resulting

Hg biomagnifications and calculated BMF values. Werecommend these diet factors be considered in futurefood web and biomagnification studies of Hg.

5. Conclusions

Here we documented Hg variability within three foodwebs in the Canadian western Arctic where BeaufortSea beluga spend their summer. Mercury levels inorganisms locally exposed to high levels of Hg in theMackenzie River outflow had the lowest Hg concentra-tions. However, the higher Hg levels in biota collectedfrom the Amundsen Gulf, Franklin Bay and easternBeaufort Sea may suggest that Hg from the MackenzieRiver may only become available for biological uptakein offshelf environments. On the other hand, food webstructure may also be an important factor influencingbeluga Hg levels. The Amundsen Gulf food web lengthmay have caused Hg levels to be higher in arctic cod andbeluga relative to the estuarine-shelf, whereas theepibenthic association with high Hg level environmentssuch as sediment and detritus likely resulted in high Hglevels. These findings, along with similar Hg trophiclevel transfer slopes among food webs, suggest thatfuture research should focus on MeHg uptake processesat the bottom of the food web and food web structure.

This study demonstrated the importance of incorpo-rating predator habitat selection when examining Hguptake processes from prey. By incorporating informa-tion on potential prey in food webs associated with theirhabitat use, we found beluga Hg levels to correspondwith prey Hg levels. Although we were unable todistinguish between the two larger sized beluga groupsbased upon differences in food web dietary Hg sources,they were different from the smaller estuarine-shelfbeluga that appeared to feed on habitat associated lowHg level prey. Using BMFs from prey to belugaprovided insight for beluga-food web pairing and forexamining potential Hg diet sources and trophic leveltransfer. However, beluga diet was not directly deter-mined in this study, therefore we suggest future studiesincorporate dietary biomarkers such as fatty acidssignatures and compound-specific stable isotope anal-yses to confirm beluga prey items. Finally, we proposethat future studies evaluating contaminants in predatorsshould incorporate behaviour and subsequent foragingcharacteristics into their analysis.

Acknowledgements

This project was supported by the Natural Scienceand Engineering Research Council—Industrial Post

Page 12: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

1023L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

Scholarship sponsored by Devon Corporation Canada toLL. Funding and project support was provided byCASES (Canadian Arctic Shelf Exchange Study),ArcticNet, Fisheries Joint Management Committee,Northern Students Training Program, Northern Con-taminants Program and by NSERC Discovery Grants toDD and SHF. We thank J. DeLaronde, A. MacHutchon,D. Armstrong, B. Gemmill, E. Yumvihoze, E. Braeke-velt, B. Rosenberg, G. Boila, D. Leitch and C. Vickersfor field and technical support. We are grateful for thesupport of the science and technical crew of the CCGSAmundsen, Nahidik and Pierre Radisson, the partner-ships and support of Hunters and Trappers Committeesof Inuvik, Aklavik and Tuktoyaktuk for beluga samplecollection program. We thank A. Majewski, J. Sareault,W. Walkusz, J. Johnson, R. Wastle and the Tariuqcommunity sampling program for samples collected inthe M. Delta region. We also thank two anonymousreviewers for their helpful comments.

References

Armstrong, J.A.J., Uthe, J.F., 1971. Semi-automated determination ofmercury in animal tissue. At. Absorpt. Newsl. 10, 101–103.

Atwell, L., Hobson, K.A., Welch, H.E., 1998. Biomagnification andbioaccumulation of mercury in an Arctic marine food web: insightsfrom stable nitrogen isotope analysis. Can. J. Fish. Aquat. Sci. 55,1114–1121.

Bloom, N.S., Gill, G.A., Cappellino, S., Dobbs, C., Mcshea, L.,Driscoll, C., Mason, R.P., Rudd, J.W.M., 1999. Speciation andcycling of mercury in Lavaca Bay, Texas, sediments. Environ. Sci.Technol. 33, 7–13.

Cabana, G., Rasmussen, J.B., 1994. Modelling food chain structureand contaminant bioaccumulation using stable nitrogen isotopes.Nature 372, 255–257.

Cai, Y., 2000. Speciation and analysis of mercury, arsenic, and seleniumby atomic fluorescence spectrometry. Trends Anal. Chem. 19, 62–66.

Campbell, L.M., Norstrom, R.J., Hobson, K.A., Muir, D.C.G., Backus,S., Fisk, A., 2005. Mercury and other trace elements in a pelagicArctic marine food web (Northwater Polynya, Baffin Bay). Sci.Total Environ. 351/352, 247–263.

Choe, N., Deibel, D., 2000. Seasonal vertical distribution andpopulation dynamics of the chaetognath Parasagitta elegans inthe water column and hyperbenthic zone of Conception Bay,Newfoundland. Mar. Biol. 137, 847–856.

Dehn, L.A., Sheffield, G.G., Follman, E.H., Thomas,D.L., Bratton, G.R.,Taylor, R.J., O'Hara, T.M., 2005. Trace elements in tissues of phocidseals harvested in the Alaskan and Canadian Arctic: influence of ageand feeding ecology. Can. J. Zool. 83, 726–746.

Dehn, L.A., Follmann, E.H., Rosa, C., Duffy, L.K., Thomas, D.L.,Bratton, G.R., Taylor, R.J., O'Hara, T.M., 2006a. Stable isotopeand trace element status of subsistence-hunted bowhead andbeluga whales in Alaska and gray whales in Chukotka. Mar. Pollut.Bull. 52, 301–319.

Dehn, L.A., Follmann, E.H., Thomas, D.L., Sheffield, G.G., Rosa, C.,Duffy, L.K., O'Hara, T.M., 2006b. Trophic relationships in anArctic food web and implications for trace metal transfer. Sci. TotalEnviron. 362, 103–123.

Fisk, A.T., Hobson, K.A., Norstrom, R.J., 2001. Influence of chemicaland biological factors on trophic transfer of persistent organicpollutants in the Northwater Polynya marine food web. Environ.Sci. Technol. 35, 732–738.

France, R.L., 1995. Differentiation between littoral and pelagic foodwebs in lakes using stable carbon isotopes. Limnol. Oceanogr. 40,1310–1313.

Gradinger, R.R., Bluhm, B.A., 2004. In-situ observations on the distri-bution and behavior of amphipods and Arctic cod (Boreogadussaida) under the sea ice of the High Arctic Canada Basin. Polar Biol.27, 595–603.

Harwood, L.A., Smith, T.G., 2002. Whales of the Inuvialuit settlementregion in Canada's Western Arctic: an overview and outlook.Arctic 55, 77–93.

Harwood, L.A., Innes, S., Norton, P., Kingsley, M.C.S., 1996.Distribution and abundance of beluga whales in the MackenzieEstuary, southeast Beaufort Sea, and west Amundsen Gulf duringthe late July 1992. Can. J. Fish. Aquat. Sci. 53, 2262–2273.

Hobson, K.A., Welch, H.E., 1992. Determination of trophic relation-ships within a high Arctic marine food web using del carbon anddel nitrogen analysis. Mar. Ecol., Prog. Ser. 84, 9–18.

Joiris, C.R., Laroussi, M.N., Holsbeek, L., 1997. Mercury andpolychlorinated biphenyls in zooplankton and shrimp from theBarents Sea and the Spitsbergen Area. Bull. Environ. Contam.Toxicol. 59, 472–478.

Lawrence, A.L., Mason, R.P., 2001. Factors controlling the bioaccu-mulation of mercury andmethylmercury by the estuarine amphipodLeptocheirus plumulosus. Environ. Pollut. 111, 217–231.

Leitch, D.R., Carrie, J., Lean, D.R.S., Macdonald, R.W., Stern, G.A.,Wang, F., 2007. The delivery of mercury to the Beaufort Sea of theArctic Ocean by the Mackenzie River. Sci. Total Environ. 373,178–195.

Lindberg, E., Harris, C., 1974. Mercury enrichment in estuarine plantdetritus. Mar. Pollut. Bull. 5, 93–94.

Lockhart, L., Stern, G., Wagemann, R., Hunt, R., Metner, D.,DeLaronde, J., Dunn, B., Stewart, R., Hyatt, C., Harwood, L.,Mount, K., 2005a. Concentrations of mercury in tissues of belugawhales (Delphinapterus leucas) from several communities in theCanadian Arctic from 1981–2002. Sci. Total Environ. 351/352,391–412.

Lockhart, L., Stern, G., Low, G., Hendzel, M., Boila, G., Roach, P.,Evans, M., Billeck, B., DeLaronde, J., Friesen, S., Kidd, K.,Atkins, S., Muir, D., Stoddart, M., Stephens, G., Stephenson, S.,Harbicht, S., Snowshow, N., Grey, B., Thompson, S., DeGraff, N.,2005b. A history of total mercury in edible muscle of fish fromlakes in northern Canada. Sci. Total Environ. 351/352, 427–463.

Lonne, O.J., Gulliksen, B., 1989. Size, age and diet of Polar cod,Boreogadus saida (Lepechin 1773) in ice covered waters. PolarBiol. 9, 187–191.

Loseto, L.L., Richard, P., Stern, G.A., Orr, J., Ferguson, S.H., 2006.Segregation of Beaufort Sea beluga whales during the open-waterseason. Can. J. Zool. 84, 1743–1751.

Luque, S.P., Ferguson, S.H., 2006. Age structure, growth, andmortalityof eastern Beaufort Sea beluga (Delphinapterus leucas): a com-parison among Canadian populations. Fisheries Joint ManagementCommittee, report, p. 23.

Mathers, R.A., Johansen, P.H., 1985. The effect of feeding ecology onmercury accumulation in walleye (Stizostedion vitreum) and pike(Esox lucius) in Lake Simcoe. Can. J. Zool. 62, 2006–2012.

Morel, F.M.M., Kraepiel, A.M.L., Amyot, M., 1998. The chemicalcycle and bioaccumulation of mercury. Annu. Rev. Ecol. Syst. 29,543–566.

Page 13: Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales

1024 L.L. Loseto et al. / Journal of Marine Systems 74 (2008) 1012–1024

Power, M., Klein, G.M., Guiguer, K.R.R.A., Kwan, M.K.H., 2002.Mercury accumulation in the fish community of a sub-arctic lake inrelation to trophic position and carbon sources. J. Appl. Ecol. 39,819–830.

Richard, P., Martin, A.R., Orr, J., 1997. Study of summer and fallmovements and dive behaviour of Beaufort Sea belugas, usingsatellite telemetry: 1992–1995. Environ. StudiesRes. FundsNo. 134.Calgary, 26p.

Richoux, N.B., Deibel, D., Thompson, R.J., 2004. Population biologyof hyperbenthic crustaceans in a cold ocean environment(Conception Bay, Newfoundland) I. Mysis mixta (Mysidacea).Mar. Biol. 144, 881–894.

Stern, G.A., Macdonald, R.W., 2005. Biogeographical provinces oftotal and methyl mercury in zooplankton and fish from theBeaufort and Chukchi Seas: results from the SHEBA drift.Environ. Sci. Technol. 39, 4707–4713.

Stewart, R.E.A., Campana, S.E., Jones, C.M., Stewart, B.E., 2006.Bomb radiocarbon dating calibrates beluga (Delphinapterusleucas) age estimates. Can. J. Zool. 84, 1840–1852.

Tollit, D.J., Steward, M.J., Thompson, P.M., Pierce, G.J., Santos, M.B.,Hughes, S., 1997. Species and size differences in the digestions ofotoliths and beaks: implications for estimates of pinniped dietcomposition. Can. J. Fish. Aquat. Sci. 54, 105–115.

Uthe, J.F., Solomon, J., Grift, N., 1972. Rapid semi-micro method forthe determination of methyl mercury in fish tissue. J. Assoc. Off.Anal. Chem. 55, 583–589.

Wagemann, R., Stewart, R.E.A., Beland, P., Desjardins, C., 1990. Heavymetals and selenium in tissues of beluga whales, Delphinapterusleucas, from the Canadian Arctic and the St. Lawrence Estuary. In:Smith, T.D., Aubin, D. St., Geraci, J.R. (Eds.), Advances in researchon the beluga whale,Delphinapterus leucas. Can. Bull. Fish. Aquati.Sci. Department of Fisheries and Oceans, Ottawa, pp. 191–206.

Wagemann, R., Trebacz, E., Boila, G., Lockhart, L., 1998.Methylmercury and total mercury in tissues of arctic marinemammals. Sci. Total Environ. 218, 19–31.

Watras, C.J., Bloom, N.S., 1992. Mercury and methylmercury inindividual zooplankton: implication for bioaccumulation. Limnol.Oceanogr. 37, 1313–1318.