Article pubs.acs.org/est
Latitudinal Distribution of Persistent Organic Pollutants in Pelagic and Demersal Marine Fish on the Norwegian Coast Jan Ove Bustnes,*,† Katrine Borgå,‡ Tim Dempster,¶,§ Elisabeth Lie,∥ Torgeir Nygård,# and Ingebrigt Uglem# †
Norwegian Institute for Nature Research, FRAM − High North Research Centre on Climate and the Environment, NO-9296 Tromsø, Norway ‡ Norwegian Institute for Water Research, Gaustadalléen 21, 0349 Oslo, Norway, Norway § SINTEF Fisheries and Aquaculture, NO-7465 Trondheim, Norway ¶ Department of Zoology, University of Melbourne, Victoria 3010, Australia ∥ National Veterinary Institute and Norwegian School of Veterinary Science, P.O. Box 8156 Dep., NO-0033 Oslo, Norway # Norwegian Institute for Nature Research, Tungasletta 2, NO-7485 Trondheim, Norway ABSTRACT: The latitudinal distribution of persistent organic pollutants (POPs: legacy organochlorines [OCs], polybrominated diphenyl ethers [PBDEs,] and hexabromocyclododecane [HBCD]) was examined in livers of two species of marine fish, the pelagic saithe (Pollachius virens, n = 40) and the demersal cod (Gadus morhua, n = 40), along a south-north gradient (59°−70°N) on the Norwegian Coast. Cod had in general two to three times higher concentrations of POPs than saithe, probably because of higher exposure in the benthic food chain. The concentrations of heavy halogenated compounds were higher in the southernmost region than further north. Moreover, the POP pattern showed a gradual shift in the compositions from south to north, especially for OCs in cod: i.e. the relative importance of low-chlorinated polychlorinated biphenyl (PCB) congeners and some OC-pesticides (e.g., hexachlorobenzen [HCB]) in the contaminant burdens increased with latitude. The latitudinal fractionation signal was weaker in saithe, possibly due to its pelagic and nomadic behavior. Hence, this study shows not only a strong latitudinal fractionation in the compositional patterns of POPs in marine fish but also the effects of habitat use and fish behavior.
■
INTRODUCTION A number of studies have observed compositional shifts in the distribution of persistent organic pollutants (POPs) along latitudinal gradients in the abiotic environment.1−10 This has mostly been attributed to cold trapping, also referred to as global distillation. That is, when POPs of both high and low volatility are transported to remote areas such as the Arctic, the more volatile constituents become more important with increasing latitude and altitude, because the partitioning equilibriums of semivolatile chemicals shift as temperature drops.11,12 However, recently it also been suggested that no temperature gradient is required to observe fractionation patterns in POPs.13 In biota, the POP concentrations and patterns are determined by several individual uptake and elimination processes, including exposure through diet and biotransformation.14,15 Although some studies have observed compositional shifts in organochlorine (OC) patterns in biota in accordance with latitudinal fractionation,11,16,17 others have not.18,19 In addition, due to variation in POP recalcitrance and interspecies differences in uptake and elimination efficiencies, POPs differ in the bioaccumulation behavior; i.e. easily metabolized POPs are © 2012 American Chemical Society
eliminated whereas persistent ones are biomagnified. In general, the ability to metabolically eliminate POPs is higher in homeothermic predators than in poikilothermic species such as fish and crustaceans, which often reflect the abiotic composition of POPs.15 Moreover, compounds with high lipophilicity, e.g. highly chlorinated polychlorinated biphenyl (PCB) congeners, have higher affinity to particles and are thus more likely to adhere to settling particles which will result in a greater exposure through the benthic compared to the pelagic food web.20 We would expect that benthic feeding fish will both have higher concentrations of POPs and also compositions skewed more toward higher halogenated POPs compared to those in pelagic feeding fish.20 However, this might also occur through biomagnification if benthic fish feed at higher trophic levels than pelagic fish. Received: Revised: Accepted: Published: 7836
March 29, 2012 June 5, 2012 June 7, 2012 June 26, 2012 dx.doi.org/10.1021/es301191t | Environ. Sci. Technol. 2012, 46, 7836−7843
Environmental Science & Technology
Article
salmon farms to limit the possibility of sampling fish that had recently interacted with a farm, as this industrial activity is known to alter POP levels.31 The 4 km minimum limit was based on telemetry-derived observations of the predominant movements of wild cod and saithe in the vicinity of fish farms.32−34 All fish were caught with standardized hook and line fishing gear. Collection of fish was carried out at each location during summer and fall (June to October) 2007. Size, Diet, and Condition Indices. Upon capture, fish were placed on ice. Fish were later weighed and measured to the nearest 0.5 cm (fork length). In gadoid species, such as saithe and Atlantic cod, lipids are stored primarily in the liver.35 The hepatosomatic index (HSI) can therefore be regarded as an estimate of the individual status of energy reserves. HSI is the fraction of the total wet weight of the fish (without the stomach content), that is liver, expressed in percent. Stomach contents from the foregut were examined, and prey species were identified in the specimens analyzed here, in addition to a large number of additional fishes; see Dempster et al.36 for details of methods. Chemical Analysis. Analyses of lipid-soluble POPs (organochlorines OCs and brominated flame retardants BRFs) in fish livers were carried out at the Laboratory of Environmental Toxicology at the Norwegian School of Veterinary Science. All analytical procedures have been described in Bustnes et al.31 Percent recoveries of individual OCs and BFRs in spiked sheep blood varied from 79 to 140%, which are within the acceptable range set by the laboratory quality control system. Detection limits were set to 3× noise level and varied among species and analytes and ranged from 0.15 to 0.70 ng/g wet weight for the OCs and from 0.01 to 0.04 ng/g wet weight for BFRs. Nondetects were set at 0.5 x DL. Procedure blanks (solvents) were regularly prepared to control background contamination. Twenty-four PCB congeners (PCB-28, -52, -47, -74, -66, -101, -99, -110, -149, -118, -153, -105, -138, -187, -183, -128, -156, -157, -180, -170, -196, -189, -194, -206) and 11 OC-pesticides (hexachlorobenzene (HCB), trans-nonachlor, cis-nonachlor, oxychlordane, dichlorodiphenyltrichloroethane (p,p′-DDT), dichlorodiphenyldichloroetylene (p,p′-DDE), dichlorodiphenyldichloroethan (p,p′-DDD), mirex, and α- β-, and γ-hexachlorocyclohexane (HCHs) were determined. Finally, 5 polybrominated diphenyl ethers (BDE-28, -47, -100, -99, -154) and hexabromocyclododecane (HBCD) were analyzed. Lipids were analyzed gravimetrically. Data Analysis. Pollutant concentrations were Log (ln)transformed to approximate normal distributions. One-way ANOVAs and the Tukey HSD tests in SAS37 were used to test for differences in concentrations between locations. Two different methods were used to examine the latitudinal distribution of different POPs. First, direct (constrained) multivariate ordination analyses (redundancy analysis, RDA) were performed to analyze how the POP concentrations (logarithmic transformed wet weight concentrations, RDACONS) and patterns (relative contribution to the total POP [standardized to norm, RDAPATTERN]) varied between samples of different species (cod or saithe) and sampling location (Øksfjord, Hitra, or Ryfylke). The RDAs were carried out in CANOCO for Windows 4.5 (Microcomputer Power, Ithaca, USA, 2002). Significant explanatory variables (species, location, and covariables) were forward selected using Monte Carlo permutations38 (α = 0.05) iteratively including significant variables. In preliminary runs of RDACONS, all explanatory variables (fork length, species, location
To our best knowledge no studies have examined the distribution of POPs in pelagic and benthic marine fish simultaneously over latitudinal gradients. However, previous studies have documented latitudinal fractionation in the composition of polychlorinated biphenyl (PCB) congeners and pesticides in seawater in the northeastern Atlantic,3,21,22 suggesting that such patterns may also exist in fish from this region. In this study, we examined the geographical distribution of organochlorines (OCs) and brominated flame retardants in marine fish along the Norwegian Coast (59°−70°N, Figure 1).
Figure 1. Three study locations (Ryfylke, Hitra, and Øksfjord) along the Norwegian Coast (59°−70°N) where Atlantic cod (Gadus morhua) and saithe (Pollachius virens) were sampled from multiple sites.
We studied two commercially important species from the family Gadidae; the pelagic saithe (Pollachius virens) and the demersal coastal Atlantic cod (Gadus morhua). The life-history of the pelagic saithe in the Northeast Atlantic involves offshore spawning, an oceanic larval phase, recruitment to the coastal environment over a period of 2−4 years, and then offshore migration to spawning grounds.23,24 Saithe typically occur in pelagic schools during the coastal period and range extensively within specific fjords.24 Atlantic cod populations in coastal waters are typically sedentary and predominantly benthic.25−27 The diets of saithe and cod are diverse, but various fish dominate in adults of both species. In addition, pelagic crustaceans may be important for saithe, whereas benthos may dominate in cod.28,29 Estimated trophic levels for adult cod (4.83) and saithe (4.36) from the North Sea ecosystem are similarly high,30 and thus biomagnification of POPs should be relatively similar between the species. However, sedentary and benthic cod may have higher concentrations of highly halogenated POPs due to higher exposure in the benthic food chain.20 Moreover, if latitudinal fractionation plays a major role in determining the patterns of POP bioaccumulation in these fish species, we predicted a gradual compositional shift in organic pollutants and higher contribution of semivolatile compounds with high long-range transport potential in the north. Further, we predicted a stronger signal of latitudinal fractionation in cod compared to the pelagic and more nomadic saithe. We tested these predictions using two different analytical methods.
■
MATERIALS AND METHODS Study Locations. Saithe and cod were collected from 3 to 6 sites in three locations (Ryfylke [59° N], Hitra [62° N], and Øksfjord [70° N]) along the Norwegian Coast (Figure 1). Fish were sampled at sites 4 to 20 km distant from the nearest 7837
dx.doi.org/10.1021/es301191t | Environ. Sci. Technol. 2012, 46, 7836−7843
7838
17.5
1.6
0.71
0.7
55.9
18.6
13.9
2.59
5.69
31.5
HCB
α-HCH
β-HCH
γ-HCH
p,p′ -DDE
p,p′ -DDT
p,p′ -DDD
Mirex
oxy -chlordane
trans -nonachlor
240
0.28
BDE-28
PBDEs
1.04
3.62
PCB-170
PCB-194
8.44
PCB-180
24.4
PCB-128
2.84
33.3
PCB-138
PCB-183
44.1
PCB-153
5.96
13.8
PCB-149
PCB-187
2.06
8.32
PCB-105
PCB-141
17.6
PCB-118
ΣPCB
octa-CBs
hepta-CBs
hexa-CBs
13.8
PCB-99
45.5
PCB-74
PCB-101
4.04
PCB-52
penta-CBs
8.25
PCB-28
tetra-CBs
2.98
12.5
54.3
lipd content (%)
cis -nonachlor
1423.3
body weight (g)
tri-CB
PCBs
chlordanes
DDTs
HCHs
51.2
fork length (cm)
mean
0.25
204.9
0.98
3.06
7.26
2.45
4.75
21.9
26.8
35.3
11.2
1.59
6.96
14.4
11.9
42.9
3.43
7.08
2.79
11.0
25.4
4.99
2.36
11.9
15.4
44.9
0.69
0.79
1.58
15.8
55.3
1367.5
51.5
median
Std Dev
0.10
109.9
0.61
2.21
4.65
1.68
3.26
11.2
19.7
25.7
5.96
1.19
3.79
9.42
7.00
9.83
2.00
3.56
1.19
7.24
18.8
2.41
1.18
6.6
9.57
36.6
0.08
0.21
0.13
3.17
7.8
324.4
3.6
Øksfjord (n = 10)
0.45
326.6
1.84
6.12
13.7
4.54
10.3
29.3
52.4
63.1
16.8
3.33
10.3
22.7
18.1
56.9
5.07
8.55
3.53
15.0
38.4
5.24
2.92
21.2
24.6
95.3
0.7
0.61
1.59
18.9
46.3
869
39
mean
0.3
197.3
0.96
3.03
7.7
2.36
5.6
13.3
25.8
42.2
10.3
1.68
6.31
12.8
10.1
46.1
2.64
5.74
2.05
6.91
18.8
2.42
1.41
9.33
16.3
39.9
0.64
0.56
1.53
12.8
44.1
500
35
median
Hitra (n = 20)
saithe
0.49
418
3.29
10.4
20.5
7.21
15.7
52.2
83.3
77.9
21.5
5.26
14.5
30.2
24.5
33.5
7.50
7.97
4.92
28.5
69.1
8.65
4.74
41.1
27.1
190.7
0.17
0.19
0.32
22.1
14.8
1105
12
Std Dev
0.38
281.5
1.59
4.86
11.5
3.67
8.65
22.9
39.9
51.9
19.5
2.80
8.70
18.1
15.6
55.3
4.21
9.19
3.05
9.96
26.2
4.4
3.36
28.1
27.7
76.9
0.7
0.86
1.57
18.6
70.8
1585.5
52.3
mean
0.41
295.1
1.64
5.17
11.8
3.72
9.03
20.1
42.1
55
20.9
2.83
8.56
18.9
16.0
57
4.23
9.8
3.44
9.08
23.2
4.25
3.09
24.6
28.8
75.9
0.72
0.89
1.53
20.2
70.9
1510.5
50.5
median
Std Dev
0.10
73.4
0.42
1.36
2.79
0.91
2.5
8.48
10.8
13.4
5.07
0.75
2.50
4.39
3.67
17.5
0.85
2.41
0.89
3.58
9.97
1.33
1.12
12.6
8.14
24.5
0.11
0.23
0.09
4.85
4.4
267.7
3.9
Ryfylke (n = 10)
0.13
0.31
0.70
0.24
0.47
0.42
0.50
0.35
0.45
0.67
0.69
0.13
0.42
0.73
0.84
0.66
0.54
0.62
0.40
0.78
0.55
0.67
0.88
625.1
4.1
12.5
31.5
9.02
10.7
69.5
95.1
140.8
17.1
5.37
19.9
58.1
41.3
69.5
13.6
19.5
7.31
36.1
98.1
13.4
7.94
33.35
0.02b 0.06
40.9
156
0.75
0.68
1.38
0.34
0.17
0.94
0.06a
0.92
34.8
42.8