TEMPORAL AND SPATIAL TRENDS IN ... - Springer Link

Environmental Monitoring and Assessment (2006) 113: 245–263 DOI: 10.1007/s10661-005-9083-7

c Springer 2006

TEMPORAL AND SPATIAL TRENDS IN CHLORINATED HYDROCARBON CONCENTRATIONS OF MINK IN CANADIAN LAKES ERIE AND ST. CLAIR PAMELA A. MARTIN1,∗ , TANA V. MCDANIEL1 and BRUCE HUNTER2 1

Canadian Wildlife Service, Environment Canada, Box 5050, 867 Lakeshore Road, Burlington, Ontario, Canada; 2 Department of Path-biology, University of Guelph, Guelph, Ontario, Canada (∗ author for correspondence, e-mail: [email protected])

Abstract. Mink (Mustela vison) carcasses were collected from local commercial trappers from fall 1998 to spring 2003 from tributaries and marshes within 4.8 km from the shores of Lakes Erie and St. Clair, including the Wheatley Harbour and St. Clair River Areas of Concern (AOCs), as well as from inland sites (8 to 40 km from shore). Liver concentrations, on a lipid weight basis, of chlorinated hydrocarbons were measured and compared among sites and to tissue concentrations of mink from two previous collections from similar sites over the past 25 years. Mink from the western Lake Erie sites, which included the Wheatley Harbour AOC, had significantly higher concentrations of sum PCBs and PCB Aroclors than those from the St. Clair corridor or inland sites, with concentrations from the rest of Lake Erie being intermediate. Dieldrin concentrations were also significantly higher in western Lake Erie than many other sites, and those of oxychlordane and mirex also tended to be higher (0.05 < p < 0.1). There were no differences in contaminant concentrations between the St. Clair River AOC and the downstream non-AOC Lake St. Clair site, with the exception of slightly higher levels of octachlorostyrene (OCS). Concentrations of PCBs and other chlorinated hydrocarbons in mink showed a general decrease over the past two decades. In contrast, PCBs and some organochlorine pesticides tended to increase, significantly so with oxychlordane, in western Lake Erie mink over the same time period. DDE declined at all sites. Currently, mink liver PCB concentrations are within the range associated with reproductive impairment, as determined from captive mink studies, in 11.7% of all animals collected from the Lakes Erie and St. Clair basin overall, but in almost 40% of individuals from western Lake Erie. Keywords: areas of concern, aquatic mammal, mink, PCBs

1. Introduction Mink (Mustela vison) are top mammalian carnivores in the Great Lakes ecosystem, and as such can accumulate relatively high concentrations of persistent contaminants. Avian aquatic predators including bald eagles (Haliaeetus leucocephalus), osprey (Pandion haleaetus) and double-crested cormorants (Phalacrocorax auritus) are known to be sensitive to organochlorine (OC) pesticides such as DDT and dieldrin. Mink however, appear to be less affected by these contaminants but extremely sensitive to the chlorinated industrial contaminants such as polychlorinated biphenyls (PCBs) in general and the PCB congeners that resemble 2,3,7,8-tetrachlorodibenzo-p-dioxin in particular. Because mink have relatively small home ranges (0.5–5.5 km), short life spans, and do not

246

P. A. MARTIN ET AL.

migrate as do their avian counterparts, they more accurately reflect current local contamination. Mink therefore, may provide the most sensitive biological indication of PCB contamination in upper trophic levels of the Great Lakes ecosystem. The sensitivity of mink to PCBs was initially detected in the early 1970s when ranch mink fed contaminated Great Lakes fish experienced reproductive failure, manifested as reduced litter size, increased kit mortality and lowered kit weight (Aulerich and Ringer, 1971; Hornshaw et al., 1983). Subsequent laboratory toxicity testing with ranch mink established a lowest observable adverse effect level (LOAEL) on kit growth of 0.25 ppm PCBs in the diet, corresponding to a maternal liver PCB concentration of 0.98 mg/kg wet weight; the LOAEL for kit survival was correlated to a maternal liver sum PCB concentration of 2.2 mg/kg wet weight (Restum et al., 1998). Relating laboratory results to effects on reproduction in wild animals is very difficult (Forsyth, 2001). Contaminant surveys of wild mink have provided some indication of PCB exposure in the lower Great Lakes basin. Mink trapped in some Ontario townships adjacent to Lake Erie during the 1970s and 1980s contained concentrations of PCBs exceeding the calculated LOAEL for growth and reproduction (Proulx et al., 1987; Haffner et al., 1998). PCB concentrations in animals trapped in most inland sites were lower than those trapped within 3 km of Lake Ontario or the St. Lawrence River shoreline, and it is generally acknowledged that exposure to PCBs and other chlorinated hydrocarbons is greatest in Great Lakes shorelines relative to inland tributaries (Addison et al., 1991). Direct studies of reproduction in wild mink have seldom been attempted and have not met with success. However, recent field studies in western North America have revealed negative correlations between liver PCB concentrations in wild mink and otters and baculum (penis bone) length and testes size (Harding et al., 1999) at considerably lower PCB concentrations than those measured in Great Lakes mink. In the current study, we compare contaminant concentrations in liver tissue of mink trapped from 1998 to 2003 to those of mink similarly obtained in 1978/79, and to data in the literature. Our objectives were to provide information on the current contaminant status of mink in the lower Great Lakes basin, to examine temporal contaminant trends in mink of the Lake Erie and Lake St. Clair basins, and to assess the potential for reproductive effects of these concentrations in wild mink as demonstrated from laboratory studies,.

2. Methods and Materials We received skinned mink carcasses from commercial fur trappers during the winters of 1998/1999 through 2002/2003. We attempted to obtain animals trapped from wetlands and tributaries on the Canadian shores of Lakes Erie and St. Clair,

TEMPORAL AND SPATIAL TRENDS IN CHLORINATED HYDROCARBON

247

Figure 1. Study area in Lakes Erie and St. Clair and the St. Clair River from which mink were trapped and the carcasses obtained from commercial trappers, 1998–2003. Stippled boxes indicate approximate areas in which collections took place.

as close to the shoreline as possible. We targeted townships near areas designated by the Great Lakes International Joint Commission as Areas of Concern (AOCs), as well as townships from which mink had been collected by the Canadian Wildlife Service in an earlier collection in the late 1970s. During both sampling periods, animals from Great Lakes sites were trapped within 4.8 km (3 miles) from the specific Lake or connecting channel; animals trapped 8–40 km from a Great Lake were considered to be inland. In the current sampling period, adequate samples of mink carcasses were obtained from six main areas (Figure 1). These included Walpole Island First Nations Reserve (42 56 N, 82 56 W), at the confluence of the St. Clair River into Lake St. Clair, comprising a portion of the St. Clair River AOC; marshes near the mouth of the Thames River into Lake St. Clair (42 20 N, 82 24 W) including Dover and Tilbury East townships; western Lake Erie (Mersea, Gosfield South and Colchester South townships), which included the Wheatley Harbour AOC (41 59 N, 82 53 W); Long Point in Walsingham township, at the junction of the east and central basins of Lake Erie (42 35 N, 80 25 W); marshes at the mouth of the Grand River in the eastern basin of Lake Erie (42 52 N, 79 33 W) within Dunn, Rainham and Moulton townships; five inland townships within the Lake Erie watershed, including Harwich, Howard, Orford, North Norwich and North Cayuga. Trappers were provided with kits in an attempt to ensure relatively uniform collection and storage of carcasses. Kits included aluminum foil, labels for recording identification, location and date of trapping, and detailed instructions on handling

248

P. A. MARTIN ET AL.

and storage of trapped specimens. Skinned carcasses were wrapped in foil and stored at −20C until pickup and dissection. Mink carcasses were thawed, sexed and assessed for gross abnormalities and organ lesions. Body weight (without pelt) was obtained and total length from nose to tip of tail measured. Tail length was also measured. One lower canine tooth was removed for dental cementum determination of age (Matson’s Laboratory, Milltown, MT). Livers were excised with a hexane-washed scalpel, weighed and placed in acetone-petroleum ether washed amber glass jars and stored at −20C until subsequent analysis for organochlorine pesticides and PCBs. Livers were sent to the National Wildlife Research Centre (NWRC, Hull, PQ) for analysis of chlorinated organic contaminants. The contaminants measured included 62 PCB congeners and the following organochlorine pesticides: p,p’-DDT; p,p’-DDE; p,p’-DDD; hexachlorobenzene (HCB); octachlorostyrene (OCS); photomirex (p-mirex) and mirex; dieldrin, cis- and oxy-chlordanes, cis- and transnonachlor; heptachlor epoxide, alpha-, beta-, and gamma-hexachlorocyclohexane (HCH). Samples were analyzed according to the methods of Norstrom et al. (1988). Essentially, thawed livers were homogenized and a 5 g aliquot was taken for analysis. The sample underwent neutral extraction, removal of lipids and biogenic compounds by gel permeation chromatography, followed by further cleanup using florisil column chromatography. Quantitative analysis of OCs and PCBs was performed using capillary gas chromatography, coupled with a mass selective detector operated in selected ion monitoring mode. Each sample extract was injected twice. The first injection was designed to determine OCs by using a 21 OC standard mixture, and the second injection was to determine PCBs using an Aroclor 1:1:1 PCB quantitation standard mixture. The detection limit was 0.0001 mg/kg ww, and was treated as 0.00005 in statistical analyses. Trace concentrations were between 0.0001 to 0.0009 mg/kg ww, and were treated as 0.0005 in statistical analyses. The percent recovery of 13 C12 -labelled internal standard PCBs ranged from 64.5% to 104.7%, with a mean percent recovery efficiency of 80.24%. The percent recovery of 13 C12 -labelled internal standard of tetra-, penta, and hexa-chlorobenzene ranged from 39.7% to 93.9%, with a mean percent recovery of 67.7%. Samples were not adjusted for recoveries. One or two quality assurance samples were analyzed with each run of mink livers. These samples consisted of 1989 diluted herring gull (Larus argentatus) egg pool reference material. Aliquots of whole body homogenates from the mink carcasses collected in 1978 and 1979 by Canadian Wildlife Service, Environment Canada (Proulx et al., 1987) had been archived at NWRC. Because analytical methods have changed considerably over the last two decades, we reanalyzed a subsample (n = 7) of these archived samples to determine adjustment factors to facilitate temporal comparisons among PCB Aroclor mixtures and OC pesticides. Adjustment equations derived to compare values from the 1970s collection to current values were as follows:


249

Aroclor 1254:1260 [current] = y= −0.026 + 0.762∗ Aroclor 1254:1260 [original], R 2 = 0.99 Aroclor 1260 [current] = 0.485 + 0.477∗ Aroclor 1260 [original], R 2 = 0.68; Sum PCBs [current] = −0.13 + 0.358∗ Aroclor 1254:1260 [original], R 2 = 0.98; DDE [current] = 0.041 + 0.806∗ DDE [original], R 2 = 0.97; DDD [current] = 0.006 + 0.765∗ DDD [original], R 2 = 0.95; Dieldrin [current] = 0.003 + 0.59∗ Dieldrin [original], R 2 = 0.65; Oxychlordane [current] = −3.968E-5 + 1.102∗ Oxychlordane [original], R 2 = 0.80; In the 1979 collection, whole mink carcasses (minus pelt, head, stomach contents and hind limbs), rather than livers, were measured for contaminants; percent lipid of these homogenates was somewhat higher than that of liver homogenates (mean of 3.1% for the livers versus a mean of 5.2% for whole body homogenates). Mason (1989) found PCB concentrations in muscle and liver tissue to be similar on a lipid weight basis. In an effort to increase comparability of the temporal data therefore, all statistical analysis was conducted on lipid corrected data. Contaminant analysis for mink collected in 1989 (Haffner et al., 1998) followed similar techniques to those currently used at NWRC and mean values provided in the literature could be compared to our values, on a lipid weight basis. Concentrations of PCDDs, PCDFs and non-ortho PCB congeners (IUPAC # 37, 77, 81, 126, 169, 189 IUPAC number) were measured in a single pooled sample of mink liver from each of the main collection areas: eastern Lake Erie, Long Point, western Lake Erie, Lake St. Clair and Walpole Island. Concentrations of PCDD/F and non-ortho PCBs were determined using capillary gas chromatography and mass spectroscopy (Simon and Wakeford, 2000). Sample preparation involved grinding a 5-g aliquot of the sample with anhydrous Na2 SO4 and extracting it with dichloromethane/hexane (1:1 v/v). The sample was spiked with a primary internal standard mixture (13 C12 )-labeled PCDD/PCDFs and non-ortho PCBs (Cambridge Isotope Laboratories, Andover, MA) at a concentration of 50–100 ng/kg on top of the column prior to extraction. Prior to adjustment to the final volume of 10 ml, a secondary internal standard was added (37 C117 )-2378-tetrachlorodibenzo-p-dioxin (TCDD). Samples were analyzed with a Hewlett-Packard 5987B GC/MS (highresolution GC, low-resolution MS) in SIM mode, using a 30-m DB-5, thin-film, capillary GC column. Concentrations of chlorinated hydrocarbons were highly skewed and the data were log transformed for parametric comparisons among sites, sex and age using three factor analyses of variance (ANOVA). If age or sex were not significant they were removed from the model. If the factor of site was significant ( p < 0.05), site means were compared using Tukey’s HSD tests. Kruskal-Wallis ANOVAs were used when the assumption of homogeneity of variance (Levene’s Test) was not met. Post hoc comparisons were conducted using non-parametric multiple contrast tests for unequal sample size (Zar, 1996). Temporal comparisons

250

P. A. MARTIN ET AL.

between the late 1970s data and the current data were made using t-tests; if the comparisons violated the assumption of equal variance, then comparisons between years were made using a non-parametric Mann Whitney U test.

3. Results A total of 83 mink received from trappers in the target areas of Canadian Lakes Erie and St. Clair and the St. Clair River between 1998 and 2003 had contaminant analysis completed on liver tissue. Because 74% of all animals obtained were males, 70% of individual liver analyses were of male animals. Nine to 18 livers were individually analyzed for contaminants from each Lake site; and a pool of 3–4 females was also analyzed at each of three sites. Although 14 carcasses were received from the inland townships, only four had analysis of all contaminants completed; the remaining 10 were analyzed only for sum PCBs. 3.1. G EOGRAPHIC

PATTERNS AMONG SITES

(1998--2003)

Percent lipid was significantly higher in mink from western Lake Erie relative to those from Long Point and there was overall a very large range in percent lipid (0.95 to 45.4%; Table I) which could affect contaminant concentrations. Therefore, all contaminants data were lipid normalized for geographic and temporal comparisons. There was no significant effect of age or sex ( p > 0.10) in the overall model predicting PCB concentrations, therefore concentrations of PCBs were compared among the five sites pooled across age and sex. Sum PCBs, as well as the Aroclor mixtures 1254:1260 and 1260, were significantly higher in western Lake Erie mink than those from Walpole Island in the St. Clair AOC, and from Lake St. Clair and the inland sites (Table I). Although there was no overall effect of age, we were interested in the effect of age class at individual sites so concentrations of sum PCBs were compared between juveniles and adults at each site separately. PCB concentrations did not differ between adults and juveniles at Walpole Island and Long Point ( p > 0.50). Adult concentrations were significantly greater ( p < 0.05) than those of juveniles at the Lake St. Clair (7.12 versus 1.40 mg/kg lipid weight), inland (3.57 versus 1.78 mg/kg) and western Lake Erie (42.3 versus 19.0 mg/kg) sites. Eastern Lake Erie was anomalous in having much higher PCB concentrations in juveniles than adults (51.6 versus 4.52 mg/kg lipid weight). The ratios of number of juveniles to number of adults trapped are shown in Table I. Concentrations of organochlorine pesticides and related compounds were one to three orders of magnitude lower than those of PCBs, but there were still significant differences among sites on a percent lipid basis. The factor of age tended towards significance in the overall model for DDT ( p = 0.06) and was retained in the model; site differences were not significant. Age and sex were not significant factors ( p values > 0.10) in overall models predicting mink liver concentrations of DDT


251

TABLE I Mean (standard error) and range of organochlorine contaminant concentrations (ug/kg) expressed on a wet weight basis, and mean (SE) on a lipid basis (mg/kg) in bold, in liver of mink trapped in watersheds of Lakes Erie and St Clair, 1999–2004. Animals from Great Lake sites were trapped 0.30). There was a strong decreasing trend in DDE concentrations in mink throughout Lake Erie, which was significant in western Lake Erie ( p = 0.0001: Figure 2b).

254

P. A. MARTIN ET AL.

TABLE II Concentrations of non-ortho PCBs, dioxins and furans with associated TEQs (ng/kg wet weight) in pooled livers of mink collected from Lake Erie and the St. Clair River Lake Erie

St. Clair Walpole Is.

Western basin

Long point

Eastern basin

5 3.59

6 13.8

4 8.64

4 4.08

0.89 5.34 80 6.35 261 8.1 67.2%

5.25 58.8 699 80.3 3,710 70.7 82.7%

1.01 6.78 156 19.8 756 15.8 75.6%

1.32 5.86 113 14.8 944 11.4 77.0%

N Lipid % Non-Ortho PCBs PCB-37 PCB-77 PCB-126 PCB-169 PCB-189 TEQ (non-ortho PCBs) %total TEQ PCDD Isomers 2,3,7,8-TCDD 1,2,3,7,8-PnCDD 1,2,3,4,7,8-HxCDD 1,2,3,6,7,8-HxCDD 1,2,3,7,8,9-HxCDD 1,2,3,4,6,7,8-HpCDD OCDD TEQ (dioxins) % total TEQ PCDF Isomers 2,3,7,8-TCCDF 1,2,3,7,8,PnCDF 2,3,4,7,8-PnCDF 1,2,3,4,7,8-HxCDF 1,2,3,6,7,8-HxCDF 1,2,3,7,8,9-HxCDF 2,3,4,6,7,8-HxCDF 1,2,3,4,6,7,8-HpCDF 1,2,3,4,7,8,9-HpCDF OCDF TEQ (furans) % total TEQ

0.52 0.48∗ 0.52 3.79∗ 0.63∗ 35.9 54.2 1.7 14.2%

2.49 1.21 0.72∗ 10.8 1.04∗ 33.6 145 5.1 6.0%

1.00 0.71 0.46∗ 2.54 0.55 8.17 16.2 2.0 9.6%

0.62 0.44 0.45 2.92 0.41 9.73 38.0 1.5 10.1%

0.15 < 0.10 4.13 0.54∗ 0.98 < 0.10 1.58 0.83 0.13∗ 1.76 2.2 18.3%

0.56 0.20∗ 18.4 1.70 2.46 < 0.10 3.96 1.53 0.47∗ 2.54 9.7 11.3%

0.16 < 0.10 5.79 0.46 0.75 < 0.10 1.06 0.60 0.14∗ 1.04 3.04 14.5%

0.12 < 0.10 3.57 0.28 0.59 < 0.10 1.03 0.56∗ 0.14∗ 1.20∗ 1.9 12.8%

Total TEQs

12.0

85.5

20.9

14.8

Note. the TEQs are based upon toxic equivalency factors developed by van den Berg et al. (1998) for mammals. ∗ These values are outside the 15% quality control limits around ion abundance and should be viewed with caution.


255

Figure 2. Trends in concentrations of a) PCBS (Aroclor 1254:1260), b) DDE, c) dieldrin, and d) oxychlordane in mink livers and herring gull eggs in 1979/80 (open bars), 1987/88 (filled bars) and 1998/03 (hatched bars). Note different scales between the two species. Mink data from 1987/88 from Haffner et al. (1998). Herring gull data from Bishop et al. (1992) and Jermyn-Gee et al. (2005).

256

P. A. MARTIN ET AL.

Dieldrin concentrations declined significantly in eastern Lake Erie ( p = 0.04), but not at the other sites. There was a general increasing trend between 1970s and 2000 in oxychlordane which was significant in western Lake Erie ( p < 0.05) and approached significance at Long Point ( p = 0.067; Figure 2d). While concentrations of Aroclor 1254:1260 in herring gull eggs were two orders of magnitude greater than those in mink livers in the late 1970s, they have declined substantially over the past two decades in both eastern and western Lake Erie (Figure 2a). Similarly, with all three OC pesticides, herring gull eggs have shown consistent decreases over the past two decades (Figures 2 b–d). Although mean egg concentrations of all contaminants remained higher than those of mink livers in both locations, declining trends seen in herring gull eggs do not necessarily corroborate those shown by mink. 3.4. PCB

CONGENERS

The distribution of PCB congeners differed somewhat between mink liver and herring gull eggs (Figure 3). The contribution of congener 153 typically exceeded 20% in herring gulls but accounted for