Egg Concentrations of Polychlorinated Dibenzo-p-dioxins and

Jan 24, 2003 - Eggs of double-crested and pelagic cormorants were collected between 1973 and 1998 from colonies in the Strait of Georgia, BC, Canada, ...
6 downloads 0 Views 281KB Size
Download PDF
Environ. Sci. Technol. 2003, 37, 822-831

Egg Concentrations of Polychlorinated Dibenzo-p-dioxins and Dibenzofurans in Double-Crested (Phalacrocorax auritus) and Pelagic (P. pelagicus) Cormorants from the Strait of Georgia, Canada, 1973-1998 MEGAN L. HARRIS,† LAURIE K. WILSON,‡ ROSS J. NORSTROM,§ AND J O H N E . E L L I O T T * ,‡ Lorax Environmental, 111-1634 Carmi Avenue, Penticton, British Columbia, Canada V2A 8K5, Pacific Wildlife Research Centre, Canadian Wildlife Service, 5421 Robertson Road, RR1 Delta, British Columbia, Canada V4K 3N2, and National Wildlife Research Centre, Canadian Wildlife Service, 100 Gamelin Boulevard, Hull, Quebec, Canada K1A 0H3

Eggs of double-crested and pelagic cormorants were collected between 1973 and 1998 from colonies in the Strait of Georgia, BC, Canada, and assayed for concentrations of polychlorinated dibenzo-p-dioxins (PCDDs), dibenzofurans (PCDFs), and non-ortho- and mono-ortho-biphenyls (PCBs). Double-crested cormorant eggs contained (on average) up to 433 ng kg-1 wet weight 1,2,3,6,7,8-HxCDD, 151 ng kg-1 1,2,3,7,8-PnCDD, and 74 ng kg-1 2,3,7,8-TCDD, whereas pelagic cormorant eggs contained up to 300, 99, and 28 ng kg-1 wet weight of these respective congeners. The dominant non-ortho-PCB was CB-126, which ranged as high as 2263 ng kg-1 in double-crested cormorant eggs. Concentrations of PCDDs and PCDFs fell dramatically in the early 1990s, following both severe restrictions on the use of chlorophenolic wood preservatives and antisapstains and a switch from molecular chlorine bleaching to alternative bleaching technologies at pulp mills in the region. Concentrations of PCBs did not show similar marked declines over time. On the basis of total TEQs g 148 ng kg-1 and previously published documentation of effects in siblings of the cormorant eggs analyzed here, doublecrested cormorant young may have exhibited significantly elevated EROD activity and/or brain asymmetries at all colonies from 1973 to 1989 and even at some colonies during the 1990s. Pelagic cormorant eggs collected from a few colonies in 1988-1989 also contained total TEQs greater than the threshold value estimated for double-crested cormorants.

* Corresponding author telephone: (604)940-4700; fax: (604)9467022; e-mail: [email protected]. † Lorax Environmental. ‡ Pacific Wildlife Research Centre. § National Wildlife Research Centre. 822

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 5, 2003

Introduction Since the 1960s, when researchers first found that pesticides and other synthetic chemicals were causing unforeseen and deleterious effects in wildlife (1), government and academic scientists have quantified environmental impacts using indicator species. Despite the acknowledgment of widespread wildlife contamination, monitoring programs to track population changes over time, particularly during and after industrial and other remedial actions, have typically been short-lived and therefore of limited value. Where long-term monitoring has been undertaken, data are too often archived in unpublished, essentially inaccessible government databases. On the west coast of Canada, the environmental impacts of the initial usage of organochlorine pesticides such as DDT and dieldrin were probably relatively minor because of the economics of the region and a pocketed pattern of application [see Harris et al. (2) for a notable exception]. However, a strong reliance on forestry-related industries produced a different set of problems with chlorinated hydrocarbons, notably those associated with polychlorinated dibenzo-pdioxins (PCDDs) and dibenzofurans (PCDFs). The Canadian Wildlife Service began surveys of chlorinated hydrocarbons in marine bird populations in the late 1960s (3) and extended the work to monitoring for chlorinated hydrocarbons from forest industry sources in the late 1980s (4). It became clear early in that program and from subsequent surveys of fishes and shellfish that there was significant contamination of bird and fish species with PCDDs and PCDFs along the southern inner coastline, namely in the Strait of Georgia, the Juan de Fuca Strait, and connected waterways. Concentrations of PCDDs and PCDFs in waterfowl, fishes, and shellfish exceeded acceptable criteria and led to fishery closures and consumption advisories along the British Columbia coast (5, 6). Geographic patterns of contamination in fishes and great blue herons (Ardea herodias fannini) implicated pulp mills, either directly or indirectly, as the primary sources for the dominant dioxin and furan congeners (6, 7). Subsequent changes to bleaching technologies, in-mill processes, and effluent treatment at pulp mills occurred during the late 1980s to early 1990s, first as voluntary modifications made by individual mills and then by all mills enforced with Canada-wide legislation in 1993 (8). We have already reported the long-term trends in contamination of great blue heron eggs before, during, and after these major industrial changes (7). Here, we report similar data for double-crested cormorants (Phalacrocorax auritus albociliatus) and pelagic cormorants (Phalacrocorax pelagicus resplendens) collected over a slightly longer time frame (1973-1998). Like great blue herons, both species of cormorants are resident in the Strait of Georgia and both are piscivorous. However, cormorant diets are more specialized on prey from the littoral-benthic zone, and their movements during the nonbreeding season may be over a somewhat greater range of distances than herons (9; Elliott, unpublished data). Since dietary uptake was shown to be the main route of exposure for great blue herons, which feed opportunistically and predominantly on young fish (7), it was of interest to compare their contaminant burdens over time with other birds more dependent on bottom fish associated with sedimentary deposits of contaminants. Also, much of the value of both the heron and the cormorant data lies in the fact that it was collected over close to three decades; everchanging anthropogenic pressures on wildlife may be as10.1021/es0208613 CCC: $25.00

 2003 American Chemical Society Published on Web 01/24/2003

FIGURE 1. Locations of double-crested cormorant (squares) and pelagic cormorant (triangles) colonies in the Strait of Georgia, BC, Canada. Super-imposed symbols (triangle inside square) indicate islands where colonies of both species were present and monitored. Building symbols indicate pulp mill locations. sessed in a greater temporal context, and continued monitoring will help ensure that new issues of concern for wildlife are identified quickly (10).

Methods Cormorant Colony Locations. The one double-crested cormorant colony monitored over the entire time period (1973-1998) was that on Mandarte Island, located off of the southeastern tip of Vancouver Island (Figure 1). Although variable from year to year, it was typically the largest breeding colony for the species in British Columbia during the period of study, with a maximum size of 1463 pairs recorded in 1981 (9). It was not adjacent to any one particular contaminant source, but rather was chosen as a reflection of the general status of cormorant contamination within the Strait of Georgia resident population. Two other study colonies selected similarly on criteria of large size and generalized

location were those on Chain Islands and Five Finger Island. Double-crested cormorants nesting on pilings in the Fraser River Delta were monitored to assess the potential influences of Vancouver and Fraser River contaminant inputs to wildlife tissue burdens on the southern coast. Two colonies located adjacent to pulp mills were monitored; double-crested cormorants at Crofton nested on the superstructure of the bleached kraft pulp mill effluent discharge pipe, while those nesting on Christie Islet were within 10-15 km of two bleached kraft pulp mills in Howe Sound, an inlet connected to the Strait of Georgia. Pelagic cormorants often nest in cliff-face habitat adjacent to the rocky, relatively flat outcroppings preferred by doublecrested cormorants in the Strait of Georgia; consequently, there was some overlap in island colony locations for the two species. Eggs of both species were collected from Mandarte Island and Five Finger Island. Colonies on VOL. 37, NO. 5, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

823

Mitlenatch Island, Snake Island, Chrome Island, and McRae Islets were also sampled to provide a general indication of the extent of chlorinated hydrocarbon contamination within the strait (Figure 1). The colony at Chemainus was within 10 km of the aforementioned pulp mill at Crofton. A colony at Prospect Point was in the heart of Vancouver and thus potentially exposed to multiple industrial and municipal pollutants. Two relatively small pelagic cormorant colonies on the outer northwestern coast of Vancouver Island (Nipple and Volcanic Rocks, Figure 1) were sampled for comparison of Strait of Georgia birds with those from a more pristine, isolated region. Collection and Chemical Analysis of Eggs. Between 1973 and 1998, double-crested and pelagic cormorant eggs were collected in 14 and 9 yr, respectively. Collections were made during the incubation period in May-July. Eggs were collected by one or two persons walking through the colony. Often they were made over the course of more than one visit; the sensitivity of cormorants to disturbance, the vulnerability of eggs and chicks to predation by glaucous-winged gulls (Larus glaucensis) and Corvids during disturbance, and resultant poor hatching successes all made it necessary to make multiple, short collection trips. After 1991, collections were synchronized with local conditions such that most eggs were retrieved approximately midway through the incubation period; in earlier years there was no standard time, and some records indicate collections were made late in incubation or early in chick-rearing. One egg each was taken from a selection of four or more nests within a colony. Upon collection, eggs were quickly refrigerated (4 °C); within 2 weeks, egg contents were emptied into chemically rinsed glass jars sealed with aluminum foil-lined lids. Tissue samples were stored frozen (-40 °C) at the Canadian Wildlife Service National Specimen Bank [National Wildlife Research Centre (NWRC), Hull, Quebec] until chemical analysis. In 1985, 1987, and 1990, double-crested cormorant eggs collected from a few colonies were assayed individually; otherwise, eggs were pooled by colony each year, and the composites were assayed. Homogenates were normally analyzed within the year of collection; however, some early samples (1973, 1979, 1985) were retrieved from the specimen bank and analyzed in 1989, while 1995 samples were held until 2001 before analysis. All egg samples were analyzed for PCDDs and PCDFs at NWRC using a gas chromatography/mass spectrometry (GC/ MS) procedure that was modified little over the years. In the 1980s, egg samples were assayed using a low-resolution GC/ MS method, while in the 1990s, refinements led to a highresolution GC/MS replacement method. The low-resolution method has been described elsewhere (4), as have the preparatory procedures (neutral extraction, gel permeation chromatography, alumina column cleanup, Florisil column chromatography; 11-13) and the quantification of nonortho- and mono-ortho-PCBs (14). Until 1993, quantification was achieved using a Hewlett-Packard 5987B GC/MS in SIM mode and with a 30-m DB-5 thin film capillary column. After 1993, quantification was achieved using a VG AutoSpec double-focusing high-resolution MS linked to a HewlettPackard 5890 series II high-resolution GC with a 30-m DB-5 fused silica column. Isotopically labeled (13C12) internal standards were used for all PCDD and PCDF congeners measured, and corrections for percent recovery of each individual congener were made. Also, in all years, herring gull egg reference samples and replicates of cormorant samples were used to check analytical accuracy (15). Minimum detection limits were assessed for each analytical run prior to 1991 and for each sample in more recent years; they are reported in the results where relevant. Lipid and moisture content were determined using gravimetric methods (13). Coplanar PCBs are described in 824

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 5, 2003

the text using International Union of Pure and Applied Chemistry (IUPAC) numbers. Statistical Analysis. All statistical analyses were completed using SYSTAT 5.0 (16). Tissue concentrations of PCDDs, PCDFs, and coplanar PCBs were log-normally distributed. Values below the detection limit were submitted as half of the lowest detection limit reported. In cases where a trace concentration was identified in the chemistry report, the reported trace value was used. TCDD toxic equivalency concentrations (TEQs) were calculated using avian toxic equivalency factors (TEFs) (17). Coplanar PCBs were only measured in a fraction of the tissue samples (25 of 63 composites); therefore, a ∑PCB TEQ was estimated using the significant positive regression between PCB 153 and ∑PCB TEQs measured in 11 composites (from 1994, 1995, 1997, and 1998). The measured TEQs were the sum of TEFs for congeners 77, 81, 126, 169, 189, 105, and 118. We forced the regressions through the origin, under the assumption that 0 ng kg-1 PCB 153 should equate to 0 ng kg-1 total PCBs in environmental media from the British Columbia coast. The regional pattern of contamination is typically dominated by a mixture of Aroclor 1254 and 1260 sources, which have similar PCB 153 content on a weight percent basis; this assumption would not work in other areas such as Green Bay, Lake Michigan, where Aroclor 1242 (containing minor amounts of PCB 153) is prominent (18). The two species were considered independently, using the following regression equations to estimate PCB TEQs:

double-crested PCB TEQ ) 0.00021‚PCB 153 (n ) 11, r 2 ) 0.902, p < 0.00001) (1) pelagic PCB TEQ ) 0.00016‚PCB 153 (n ) 9, r 2 ) 0.847, p ) 0.0002) (2) We were principally interested in exploring for spatial and temporal trends in the cormorant data set and then assessing species differences (pelagic vs double-crested cormorants vs great blue herons). The preponderance of composite analyses compromised statistical power; therefore, to increase n for temporal trend analyses, we assumed that spatial differences in contamination (e.g., among colonies) were negligible. We based this on the results of two principal components analyses (PCAs, one per species), which showed no pattern of segregation among colonies, and a strong distinction among collection years. The PCAs explained a large majority of the variance in chemical concentrations (>90% in the first two components), but they produced a poor separation of congeners into different components, thus we did not include the results here. We tested for changes in egg contamination over time (without consideration of colony location) using multiple log-linear regressions of year and egg lipid content (19) with the dominant congeners, 2,3,7,8-TCDD, 1,2,3,7,8-PnCDD, 1,2,3,6,7,8-HxCDD, 2,3,4,7,8PnCDF, and PCBs 126, 77, 169, and 118. When a significant interaction existed between year and egg lipid content, concentrations were lipid-normalized and simple linear regressions were conducted with contaminant and year. Similar regressions were used to test for significant changes in TEQs over time. We used the small subset of replicated data to further consider spatial and temporal trends in cormorant contamination. Differences among colonies (within a given year) or among years (within a given colony) were assessed using analyses of covariance (ANCOVAs) on log-transformed wet weight values with percent lipid as the covariate. Where a significant difference was found at an R-level of 0.05, the specific placement of significance was determined using posthoc Tukey honestly significant difference (HSD) multiple

TABLE 1. Concentrations of PCDDs and TCDD-Equivalent TEQs (ng kg-1 Wet Weight) in Double-Crested Cormorant Eggs Collected from the Strait of Georgia, BC, Canadaa year

n

location

1973 1979 1985

Mandarte Island Mandarte Island Fraser River

14p 5p 5

1987

Mandarte Island Crofton

5p 5

Christie Islet Crofton Fraser River Christie Islet Mandarte Island

7p 6p 7p 10p 20b

Chain Island

11

Christie Islet

7

Crofton

6

Fraser River Mandarte Island Crofton Five Finger Island Fraser River Mandarte Island Christie Islet Crofton Five Finger Island Fraser River Mandarte Island Crofton Five Finger Island Chain Island Crofton Five Finger Island Fraser River Mandarte Island Chain Island Crofton Five Finger Island Mandarte Island Crofton Mandarte Island

7p 11p 10p 10p 5p 10p 4p 8p 10p 10p 10p 10p 10p 10p 10p 11p 9p 15p 8p 10p 10p 10p 10p 14p

1988 1989 1990

1991

1992

1993 1994

1995

1997 1998

lipid (%)

water (%)

2378-TCDD

12378-PnCDD

123678-HxCDD

∑PCDD + PCDF TEFs

6.7 4.7 3.9 (0.4) 5.3 4.7 (0.3) 4.6 4.4 5.2 5.0 4.6 (0.1) 3.0 (0.3) 3.7 (0.4) 3.0 (0.5) 4.2 4.7 3.8 4.5 3.9 4.2 4.8 4.8 4.5 4.7 4.3 5.2 4.9 5.4 5.1 5.4 4.7 5.3 4.2 3.5 5.3 4.0 7.8 6.0

74.0 83.6 83.7 (0.5) 83.2 80.7 (1.1) 83.5 83.6 83.3 84.0 83.9 (0.1) 85.3 (0.6) 84.3 (0.5) 84.7 (1.2) 84.0 84.2 84.1 83.8 84.0 84.1 84.3 84.7 84.2 84.9 84.4 83.9 85.0 83.7 83.7 84.1 84.7 83.4 79.1 80.0 76.2 79.1 78.2 84.2

13 14 20 (2) 11 74 (8) 68 55 74 30 25 (6) 16 (4) 39 (6) 26 (2) 11 21 12 13 18 6 7 9 10 8 13 11 6 7 5 5 3 6 7 5 6 7 4 4

58 54 83 (9) 107 151 (39) 101 150 34 23 44 (0.5) 37 (8) 42 (4) 42 (3) 5 26 18 20 10 14 31 23 22 14 29 25 27 26 18 10 8 20 24 18 21 18 15 18

63 98 198 (32) 259 433 (155) 237 289 52 36 60 (14) 53 (10) 69 (11) 67 (9) 7 49 37 39 23 17 16 39 38 11 31 39 43 40 33 15 11 20 29 29 29 21 27 22

101 83 118 (10) 138 268 (56) 229 241 127 70 83 (5) 70 (13) 98 (12) 82 (5) 18 59 43 46 37 29 39 42 33 24 53 45 46 42 28 19 14 35 46 34 39 44 26 30

∑PCB TEFs 423 67 (11) 123 50 (7) 99 369 47 87 140 98 (17) 50 (7) 66 (15) 38 56 34 32 35 90 80 40 42 35 98 42 106 158c 47c 48c 36c 111c 122c 39c 55c 107c 44c 76c

a Values are arithmetic means with standard errors in parentheses or a pooled value (indicated by a ‘p’ after the sample size n). ∑TEFs for PCDDs + PCDFs are measured and based on the TEFs for birds (17), whereas ∑TEFs for PCBs are estimated, on the basis of the significant linear relationship between PCB 153 and a subsample of measured values (see Methods). b Values are arithmetic means and standard errors of 2-3 pooled samples. For PCDD + PCDF analysis, each pool contained the contents from 10 eggs (2 pools), while for PCB analysis, each pool contained the contents from 5 eggs (3 pools). The individual eggs analyzed for PCDDs, PCDFs, and PCBs were the same, even though their organization into composites for analysis was different. c These ∑PCB TEFs were calculated from measured concentrations of PCBs 77, 81, 126, 169, 189, 105, and 118 (form the basis for the regression used to estimate ∑TEFs in all other samples).

comparisons with Tukey-Kramer adjustments for unequal sample sizes. To assess species differences at proximate colonies (e.g., those located on the same offshore island or otherwise within a few hundred meters of each other), we used paired t-tests. Pairs were comprised of either great blue heron and doublecrested cormorant or double-crested cormorant and pelagic cormorant samples collected from proximate colonies in the same year. Differences in lipid content of eggs between species were first tested and, where significant, paired data were lipid-normalized prior to testing.

Results PCDDs and PCDFs in Cormorants. All egg samples from both species of cormorant (n ) 94) contained 1,2,3,7,8PnCDD and 1,2,3,6,7,8-HxCDD. These were also the most concentrated congeners: double-crested cormorant (DCCO) eggs contained 7-962 ng kg-1 wet weight HxCDD and 5-275

ng kg-1 PnCDD (Table 1), whereas pelagic cormorant (PECO) eggs (not including the two West Vancouver Island samples) contained 11-300 ng kg-1 HxCDD and 5-99 ng kg-1 PnCDD (Table 2). Ninety-nine percent of samples contained 2,3,7,8TCDD (3-100 ng kg-1 for DCCO and 1-28 ng kg-1 for PECO), and 91% contained 2,3,4,7,8-PnCDF (2-51 ng kg-1 for DCCO and 2-22 ng kg-1 for PECO). Concentration ranges (ng kg-1) for other common PCDD and PCDF congeners not listed in tables were 1,2,3,7,8,9-HxCDD (DCCO ) 1-125, PECO ) 1-58, in 65% of samples), 1,2,3,4,6,7,8-HpCDD (DCCO ) 0.4-14, PECO ) 0.5-42, in 43% of samples), and 2,3,7,8TCDF (DCCO ) 0.1-15, PECO ) 0.2-2, in 35% of samples). Congeners detected in 1-25% of samples were 1,2,3,4,7,8HxCDD; 1,2,3,4,6,7,9-HpCDD; OCDD; and 1,2,3,4,7,8-, 1,2,3,6,7,8-, 1,2,3,7,8,9-, and 2,3,4,6,7,8-HxCDF. Double-crested cormorants nesting adjacent to pulp mills (Crofton, Christie Islet) contained moderately higher concentrations of some congeners in the early years of monitoring VOL. 37, NO. 5, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

825

TABLE 2. Concentrations of PCDDs and TCDD-Equivalent TEQs (ng kg-1 Wet Weight) in Pelagic Cormorant Eggs Collected from the Strait of Georgia and Northwest Vancouver Island, BC, Canadaa year

location

n

1985 1988

Nipple Rock Chemainus Mitlenatch Island Five Finger Island Mandarte Island

10p 7p 9p 9p 20b

Prospect Point Snake Island Chrome Island McRae Islets Mitlenatch Island Mitlenatch Island Chrome Island Five Finger Island McRae Islets Mitlenatch Island Volcanic Rock Five Finger Island Chrome Island Mandarte Island McRae Islets Mitlenatch Island Prospect Point Chrome Island Five Finger Island McRae Islets Mitlenatch Island

9p 10p 6p 8p 9p 10p 10p 10p 10p 10p 15p 10p 10p 15p 7p 10p 8p 10p 10p 9p 10p

1989

1990 1991 1992

1993 1994

1995

lipid (%)

water (%)

2378-TCDD

12378-PnCDD

123678-HxCDD

∑PCDD + PCDF TEFs

4.4 5.0 5.0 4.1 4.9 (0.04) 5.2 5.3 4.8 4.5 4.7 4.8 4.6 5.6 4.2 4.8 6.8 5.3 5.4 5.6 5.3 4.6 4.7 4.6 4.5 3.7 4.1

83.9 83.3 83.0 83.5 83.5 (0.1) 84.0 82.7 84.0 83.7 83.9 84.0 83.3 84.0 84.5 83.9 83.7 84.2 83.5 83.7 84.0 83.2 84.0 81.5 80.5 79.6 83.1

8 23 20 21 14 (2) 28 19 12 16 14 6 5 5 4 9 0.8 3 1 3 35%) appeared to increase (Figure 2). Coplanar PCBs in Cormorants. All eggs tested for nonortho-PCBs (DCCO n ) 20 composites, PECO n ) 13 composites) contained the congeners 126 and 169, and all but one pelagic cormorant sample contained congener 77. Between 1973 and 1998, double-crested cormorant eggs contained 183-2263 ng kg-1 wet weight PCB 126, 27-161 ng kg-1 PCB 169, and 11-150 ng kg-1 PCB 77. Between 1992 and 1995, pelagic cormorant eggs contained 73-493, 5-29, and 0.2-9 ng kg-1 of these same respective congeners. PCBs 81 826

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 5, 2003

and 189 were measured in samples collected from 1994 onward; concentrations in double-crested cormorants ranged from 20 to 117 ng kg-1 for PCB 81 and from 1990 Crofton, 1987 > 1990

year year year location

log TCDD (ww) log PnCDD (ww) log HxCDD (ww) log TCDD (ww) log PnCDD (ww) log HxCDD (ww) log 1,2,3,7,8,9HxCDD (ww) log HpCDD (ww) log TCDF (lw) log PnCDF (ww) log TCDD (ww)

11 11 11 24

0.735 0.632 0.588 0.451

0.001 0.003 0.006 0.002

location location

log HpCDD (ww) log TCDF (lw)

24 24

0.262 0.607

0.041 Crofton 1990, Chain Island < Crofton ∼ Christie Islet log TCDD ) (-0.17‚year) + 344 log PnCDD ) (-0.12‚year) + 249 log HxCDD ) (-0.13‚year) + 261 log PnCDF ) (-0.09‚year) + 177 DCCO > PECO DCCO > PECO DCCO > PECO DCCO > PECO GBHE > DCCO

year year year year year year year

0.617 0.439

a DCCO, double-crested cormorant; PECO, pelagic cormorant; GBHE, great blue heron; TCDD, 2,3,7,8-TCDD; PnCDD, 1,2,3,7,8-PnCDD; HxCDD, 1,2,3,6,7,8-HxCDD; TCDF, 2,3,7,8-TCDF; PnCDF, 2,3,4,7,8-PnCDF; ww, wet weight; lw, lipid weight.

from the same island colonies (Five Finger and Mandarte Islands) in the same years (1989-1995), double-crested cormorants almost always contained higher concentrations of PCDDs and PCDFs than pelagic cormorants. Species differences were significant for PnCDD, TCDD, and PnCDF but not for 1,2,3,6,7,8-HxCDD or TCDF (Table 3). Although the contrast between species in PCB concentrations appeared more marked (1-3 orders of magnitude, refer to Figures 2 and 3), no significant differences were detected; unfortunately, sample size for the latter comparisons was low (n ) 3). Estimated ∑PCB TEFs were significantly greater in doublecrested cormorants as compared to pelagic cormorants from Mandarte Island in 1989 (n ) 10, Table 3). A similar test with PCDD + PCDF TEFs found marginal significance (p ) 0.053, n ) 4). In a comparison of double-crested cormorant and great blue heron eggs collected in the same years from Crofton (n ) 6, 1987-1994), there were no species differences in concentrations of the main PCDDs, PCDFs, or coplanar PCBs (p > 0.05). Clearly, however, concentrations of TCDD were far higher in herons as compared to cormorants during the late 1980s, before industrial restrictions were implemented (Figure 5). A test using replicated data from 1987 alone found a significant difference between the two species in TCDD egg concentrations (Table 3). In terms of temporal trends, the timing of the TCDD concentration peak in eggs, around 1989, was equivalent for both double-crested cormorants and great blue herons (Figure 5). For other congeners, heron monitoring generally began too late in the 1980s to compare patterns among species.

Discussion Patterns, Trends, and Sources of PCDDs, PCDFs, and Dioxin-like PCBs in Cormorants. The congener pattern of PCDDs and PCDFs in cormorants from the Strait of Georgia matched a pattern documented for other avian piscivores, waterfowl, and raptors from the region (5, 7, 20). Unlike most other published contaminant profiles for cormorants and

herons in North America (21-25), 1,2,3,7,8-PnCDD and 1,2,3,6,7,8-HxCDD dominated tissue burdens. Concentrations of these as well as TCDD, TCDF, and 2,3,4,7,8-PnCDF were generally higher in resident cormorants from the Strait of Georgia than in eggs of cormorants collected from other areas of the continent in comparable years (22-26). As late as 1998, the measured PCDD + PCDF TEQ for double-crested cormorants nesting on Mandarte Island (30 ng kg-1) remained higher than that calculated for a population in the Great Lakes (18 ng kg-1 for North Channel, Lake Superior; 22). The prevalence of PnCDD and HxCDD in wildlife from coastal British Columbia is likely the result of the extensive past production and use of polychlorinated phenols in the local forest industry (4, 5, 7, 20, 27, 28). Historically, pentachlorophenol (PCP) was commonly applied to undried lumber as a wood preservative, while tetrachlorophenol (TeCP) was used primarily in wet coastal areas as an antisapstain agent. Wood chips pulped in the coastal mills were often contaminated with both compounds. Historically, the main source for TCDD and TCDF in wildlife from the Strait of Georgia was the pulp mills that used molecular chlorine bleaching (29). TCDF was one of the few congeners exhibiting some spatial segregation in double-crested cormorant eggs, as it was typically not detected anywhere other than in eggs from colonies directly adjacent to pulp mills (e.g., Crofton and Christie Islet). There is less certainty about the sources of PnCDF in wildlife from the strait. In Europe, cormorants (Phalacrocorax carbo) contained proportionately high levels of 2,3,4,7,8PnCDF, which researchers associated with PnCDF contamination of PCB mixtures (30). Since PnCDF did not show a clear peak and decline in eggs of double-crested cormorants, it is possible that there were multiple sources for this congener along the coast, some not associated with the forest industry. Undoubtedly, pulp mills were responsible for a portion of the bioavailable PnCDF via the chlorination of furan precursors (31); however, PCB sources might have also contributed to overall tissue burdens. VOL. 37, NO. 5, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

827

FIGURE 2. Temporal trends in 2,3,7,8-TCDD, 1,2,3,6,7,8-HxCDD, and 2,3,4,7,8-PnCDF in eggs of double-crested cormorants (crosses, solid line) and pelagic cormorants (squares, hatched line) collected from the Strait of Georgia, BC, Canada, from 1973 to 1998. Data are lipidnormalized for comparison of species, but Table 3 refers to regression details in wet weight terms as there was no effect of lipid on contaminant values within either species. Where non-ortho-PCBs were measured in cormorant eggs, PCB 126 dominated the TEQ profiles, as has been found elsewhere in North America in cormorant tissues (22, 24, 26, 32). Assessment of PCB congener patterns in great blue herons from the strait indicated that contamination was relatively high in urban colonies, with some small amount of variation in pattern also found at urban sites, but overall few differences in contamination profiles among colonies in the region (33). 828

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 5, 2003

There may be some minor sources of varying Aroclor composition in the strait, but PCB contamination is low as compared to regions such as the Great Lakes (22, 23, 25). The temporal trend in PCDD and PCDF contamination of cormorants and other wildlife in the Strait of Georgia indicates a clear link between the implementation of environmental upgrades at pulp mills and a tissue-level response of wildlife to source reductions. Between 1988 and 1992, mills in the region made several process changes that included screening of wood chips for chlorophenol contamination prior to pulping, changes to defoamer products (to limit precursor PCDDs and PCDFs; 34), conversion of bleach plants to chlorine dioxide substitution, adjustment of bleaching sequences, and upgrades to secondary treatment of effluent. We noted that concentrations of TCDD in cormorant eggs peaked around 1989, just before most of these process changes came into effect. Similarly, a significant drop in TCDD contamination of great blue heron eggs occurred between 1989 and 1991, and the decline was associated with changes in contamination of fish prey captured near pulp mills (7). Also, production of PCP and TeCP was stopped in British Columbia in 1989, and their use was phased out quickly thereafter. The pattern of peak concentrations of PnCDD and HxCDD in cormorant eggs closely tracked the sales records for PCP, which register declines in use from the mid 1980s. Comparisons of Great Blue Herons and Cormorants. Of the many wildlife species monitored for chlorinated hydrocarbon contamination in the Strait of Georgia, resident great blue herons and cormorants have been the most consistently studied. Although they have many niche similarities, cormorants are more dependent on benthic fish prey than herons, which spend part of the winter feeding on voles and otherwise take a greater variety of aquatic species (35). Thus, cormorants might be expected to retain elevated concentrations of dioxins and furans for longer after the reduction of source inputs to waterways. However, our results do not support such an hypothesis. Peak periods of contamination were coincidental across species, and declines occurred over roughly equivalent and relatively rapid time spans. The tendency for great blue heron PCDD and PCDF concentrations to range higher than cormorant concentrations may be a reflection of the habitat in which the three piscivores forage. Although there is considerable overlap in the prey species taken by herons and cormorants along the coast (35-37), differences in the microhabitat from which their prey are collected may account for varying degrees of contamination among the three piscivores. The slow moving, stagnant water where herons prefer to forage favors the deposition of suspended, contaminated particulates over the open, faster moving waters where cormorants prefer to hunt. Also, the differences in PCB contamination between doublecrested and pelagic cormorants may relate to differences in the degree of residency of the two species. A telemetry study of double-crested cormorants from the southern-most colonies in the strait (e.g., Chain Islands, Mandarte Island, Crofton) found that about one-third of radio-tagged birds remained in the general vicinity of their breeding colony throughout the year, while the other two-thirds of birds made trips southwards into the Puget Sound area during some of the year (Elliott, unpublished data). There are reports of congregations of up to 600 double-crested cormorants on Washington State inland waterways during releases of salmonid smolts (K. Kramer, personal communication); presumably, some of these may be Strait of Georgia birds. Double-crested cormorant individuals moving down into Puget Sound for some of the year could elevate the relative tissue burden of PCBs in that species. Toxicological Significance of Tissue Concentrations. Sibling hatchlings of the cormorant eggs analyzed from pulp

FIGURE 3. Temporal trends in coplanar PCBs (nos. 77, 169, 126, and 118) in eggs of double-crested cormorants (crosses) and pelagic cormorants (squares) collected from the Strait of Georgia, BC, Canada, from 1973 to 1998.

FIGURE 4. Temporal trends in TCDD-equivalent TEQs for doublecrested cormorants nesting at Mandarte Island and Crofton colonies, 1985-1998. The PCDD + PCDF TEFs were measured, whereas the PCB TEFs were largely estimated (see Methods and Tables 1 and 2 for details). The Crofton colony was located at a pulp mill. mill colonies, Crofton and Christie Islet, and reference colony, Chain Island, were assessed for morphological and biochemical indicators of toxicity in 1990 (24, 38). Hepatic EROD activity in hatchlings from all three British Columbia colonies was significantly elevated as compared to hatchlings from a reference colony in Saskatchewan (24). There were no gross abnormalities, edema, or changes in morphological measurements in hatchlings from the pulp mill colonies. Henshel et al. (38) reported 3-7 hatchlings exhibiting at least one type of brain asymmetry from each of the three British Columbia colonies (sample sizes ranged from 6 to 15

FIGURE 5. Comparison of great blue heron (GBHE) and doublecrested cormorant (DCCO) egg concentrations of 2,3,7,8-TCDD (µg kg-1 lipid weight) at colonies in the Strait of Georgia, BC, Canada, 1983-1998. Both species nested within a few hundred meters of each other at Crofton, while the GBHE colony at UBC and the DCCO colony on the Fraser River Delta were within ∼15 km of each other and both residents foraged in the Fraser River Estuary. individuals, abnormality rates of 0.4-0.5). The frequency of width asymmetries was significantly greater in hatchlings from Crofton (∼0.17) as compared to hatchlings from the Saskatchewan reference colony (0.0). All of these effects were also found in resident great blue herons (39, 40), and they all regressed strongly with measured TEQs (24, 38-40). In a separate study, Henshel and associates also showed that chicken embryos and hatchlings dosed in ovo with TCDD VOL. 37, NO. 5, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

829

exhibited similar types of brain asymmetry to wild cormorant and heron chicks (41). TCDD TEQs measured in sibling eggs from these cormorant colonies were 168, 148, and 148 ng kg-1 wet weight for Chain Island, Christie Islet, and Crofton, respectively (Table 1). Although the reference cormorants at Chain Island exhibited the highest TEQ, the greatest contribution was from PCBs (estimated 58% on average), whereas 66% (Christie Islet) and 55% (Crofton) of TEQs at the pulp mill colonies were from PCDD + PCDF contributions. On the basis of these values for total TEQs without differentiating between PCB and PCDD + PCDF contributions, all cormorant eggs collected prior to 1990 and three sets collected later in 19921994 (from Mandarte Island, Five Finger Island, and Chain Island) contained equivalent or greater TEQs. That suggests double-crested cormorant young could have exhibited significantly elevated EROD activity and/or brain asymmetries at all colonies from 1973 to 1989 and even in some colonies during the 1990s. Pelagic cormorant eggs collected from Mitlenatch Island and Prospect Point (Vancouver) in 1988-1989 also contained total TEQs greater than 148 ng kg-1, but the differences between species in sensitivity are unknown. Understanding the relative contributions of PCDDs, PCDFs, and PCBs to observed TCDD-like toxicity in British Columbia cormorants is confounded by the unique nature of the pattern of contamination in these birds. Most other cormorant colonies where TCDD-like toxicity has been documented are heavily contaminated with PCBs; PCDD and PCDF concentrations are nominal or even undetectable (22, 24, 42, 43). The studies of brain asymmetry and inferred effects on the early development of the nervous system in cormorants and herons by Henshel (40) seem to suggest that dioxins and TCDD in particular are potent, if not the most potent, inducers of neurological effects. Otherwise, the effects observed and presumably mediated through the Ah receptor (primarily EROD induction) may be variably caused by PCB, PCDD, and PCDF congeners. There is some evidence that the relative weight associated with individual congener toxicity is not well-estimated in cormorants by using TEFs derived from chicken experiments. In particular, Powell et al. (44) used egg injection studies to calculate a TEF for PCB 126 in double-crested cormorants of 0.02; the avian (chicken)derived TEF recommended by the World Health Organization (WHO) and most frequently applied in recent published literature is 0.1 for PCB 126 (17). This difference of almost an order of magnitude for the dominant nonortho-PCB in many cormorant tissue samples could substantially alter the estimation of effect levels in wild populations. Our limited direct measurements suggest that doublecrested cormorants from the Strait of Georgia were experiencing some biochemical and developmental effects due to TCDD-like exposure up to the late 1980s (24, 38). On the basis of injection studies and a TCDD LC50 of 4000 ng kg-1 egg (44), embryolethality due to PCDD or PCDF exposure was not occurring. Given the relatively high thresholds for effects of PCBs, the continuing elevated concentrations of PCBs in some eggs during the 1990s probably do not significantly impact reproduction in cormorants along the coast of British Columbia.

Acknowledgments We thank I. Moul, A. Breault, P. Whitehead, and S. Lee for their assistance in the collection of eggs. Also, thanks to M. Simon, H. Won, and M. Mulvihill for conducting the bulk of chemical analyses. Thanks to P. Whitehead for drafting the map.

Literature Cited (1) Ratcliffe, D. A. Nature 1967, 215, 208-210. 830

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 5, 2003

(2) Harris, M. L.; Wilson, L. K.; Elliott, J. E.; Bishop, C. A.; Tomlin, A. D.; Henning, K. V. Arch. Environ. Contam. Toxicol. 2000, 39, 205-220. (3) Ohlendorf, H. M.; Risebrough, R. W.; Vermeer, K. Exposure of Marine Birds to Environmental Pollutants; Report 9; U.S. Fish & Wildlife Service: Washington, DC, 1978. (4) Elliott, J. E.; Butler, R. W.; Norstrom, R. J.; Whitehead, P. E. Environ. Pollut. 1989, 59, 91-114. (5) Elliott, J. E.; Martin, P. A. Environ. Monit. Assess. 1998, 53, 337362. (6) Colodey, A. G.; Wells, P. G. J. Aquat. Ecosyst. Health 1992, 1, 201-206. (7) Elliott, J. E.; Harris, M. L.; Wilson, L. K.; Whitehead, P. E.; Norstrom, R. J. Ambio 2001, 30, 416-428. (8) Government of Canada. Can. Gazette Part II. Registration SOR/ 92-269 1992, 126, 1967-2006. (9) Campbell, R. W.; Dawe, N. K.; McTaggart-Cowan, I.; Cooper, J. M.; Kaiser, G. W.; McNall, M. C. E. The Birds of British Columbia. Volume One-Nonpasserines. Introduction, Loons Through Waterfowl; Royal British Columbia Museum: Victoria, Canada, 1990. (10) Fox, G. A. Water Qual. Res. J. Can. 2001, 36, 233-251. (11) Norstrom, R. J.; Simon, M.; Mulvihill, M. J. Int. J. Environ. Anal. Chem. 1986, 23, 267-287. (12) Norstrom, R. J.; Simon M. In Polychlorinated Dibenzo-p-dioxins and Dibenzofurans; Rappe, D., Buser, H. R., Dodet, B., O’Neill, I. K., Eds.; World Health Organization: Lyon, France, 1991; pp 281-297. (13) Environment Canada. National Wildlife Research Centre (NWRC) Laboratory Services Methods Manual; MET-CHEM-OC-04A; Canadian Wildlife Service: Hull, Canada, 1997. (14) Ford, C. A.; Muir, D. C. G.; Norstrom, R. J.; Simon, M.; Mulvihill, M. J. Chemosphere 1993, 26, 1981-1991. (15) Elliott, J. E.; Norstrom, R. J. Environ. Toxicol. Chem. 1998, 17, 1142-1153. (16) Wilkinson, L. SYSTAT: The System for Statistics; SYSTAT: Evanston, IL, 1990. (17) Van den Berg, M.; Birnbaum, L.; Bosveld, A. T.; Brunstro¨m, B.; Cook, P.; Feeley, M.; Giesy, J. P.; Hanberg, A.; Hasegawa, R.; Kennedy, S. W.; Kubiak, T.; Larsen, J. C.; van Leeuwen, F.; Djien Liem, A. K.; Nolt, C.; Peterson, R. E.; Poellinger, L.; Safe, S.; Schrenk, D.; Tillitt, D.; Tysklind, M.; Younes, M.; Wærn, F.; Zacharewski, T. Environ. Health Perspect. 1998, 106, 775-792. (18) Ankley, G. T.; Niemi, G. J.; Lodge, K. B.; Harris, H. J.; Beaver, D. L.; Tillitt, D. E.; Schwartz, T. R. Giesy, J. P.; Jones, P. D.; Hagley, C. Arch. Environ. Contam. Toxicol. 1993, 24, 332-344. (19) Hebert, C. E.; Keenleyside, K. A. Environ. Toxicol. Chem. 1995, 14, 801-807. (20) Elliott, J. E.; Norstrom, R. J.; Smith, G. E. J. Arch. Environ. Contam. Toxicol. 1996, 31, 354-367. (21) Custer, T. W.; Hines, R. K.; Melancon, M. J.; Hoffman, D. J.; Wickliffe, J. K.; Bickham, J. W.; Martin, J. W.; Henshel, D. S. Environ. Toxicol. Chem. 1997, 16, 260-271. (22) Kannan, K.; Hilscherova, K.; Imagawa, T.; Yamashita, N.; Williams, L. L.; Giesy, J. P. Environ. Sci. Technol. 2001, 35, 441447. (23) Ryckman, D. P.; Weseloh, D. V.; Hamr, P.; Fox, G. A.; Collins, B.; Ewins, P. J.; Norstrom, R. J. Environ. Monit. Assess. 1998, 53, 169-195. (24) Sanderson, J. T.; Norstrom, R. J.; Elliott, J. E.; Hart, L. E.; Cheng, K. M.; Bellward, G. D. J. Toxicol. Environ. Health 1994, 41, 247265. (25) Yamashita, N.; Tanabe, S.; Ludwig, J. P.; Kurita, H.; Ludwig, M. E.; Tatsukawa, R. Environ. Pollut. 1993, 79, 163-173. (26) Buck, J.; Sproul, E. Organochlorine Contaminants in DoubleCrested Cormorants from Lewis and Clark National Wildlife Refuge in the Columbia River Estuary; Final Report, Project 13420-1261-1N14; U.S. Fish and Wildlife Service: Portland, OR, 1999. (27) Bright, D. A.; Cretney, W. J.; Macdonald, R. W.; Ikonomou, M. G.; Grundy, S. L. Environ. Toxicol. Chem. 1999, 18, 1097-1108. (28) Luthe, C. E.; Berry, R. M.; Voss, R. H. Tappi J. 1993, 76, 63-69. (29) Macdonald, R. W.; Cretney, W. J.; Crewe, N.; Paton, D. Environ. Sci. Technol. 1992, 26, 1544-1550. (30) Van den Berg, M.; Craane, B. L.; Sinnige, T.; Van Mourik, S.; Dirksen, S.; Boudewijn, T.; Van der Gaag, M.; Lutke-Schipholt, I. J.; Spenkelink B.; Brouwer, A. Environ. Toxicol. Chem. 1994, 13, 803-816. (31) LaFleur, L. E.; et al. Chemosphere 1990, 20, 1731-1738. (32) Meadows, J. C.; Tillitt, D. E.; Schwartz, T. R.; Schroeder, D. J.; Echols, K. R.; Gale, R. W.; Powell, D. C.; Bursian, S. J. Arch. Environ. Contam. Toxicol. 1996, 31, 218-224.

(33) Harris, M. L.; Elliott, J. E.; Butler, R. W.; Wilson, L. K. Environ. Pollut. 2003, 121, 207-227. (34) Berry, R. M.; Fleming, B. I.; Voss, R. H.; Luthe, C. E.; Wrist, P. E. Pulp Pap. Can. 1989, 90, 48-58. (35) Butler, R. W. Ph.D. Dissertation, University of British Columbia, Vancouver, Canada, 1991. (36) Robertson, I. Condor 1974, 76, 346-348. (37) Harfenist, A.; Whitehead, P. E.; Cretney, W. J.; Elliott, J. E. Food Chain Sources of Polychlorinated Dioxins and Furans to Great Blue Herons (Ardea herodias) Foraging in the Fraser River Estuary, British Columbia; Technical Report Series 169; Canadian Wildlife Service: Delta, Canada, 1993. (38) Henshel, D. S.; Martin, J. W.; Norstrom, R. J.; Elliott, J.; Cheng K. M.; DeWitt, J. C. J. Great Lakes Res. 1997, 23, 11-26. (39) Sanderson, J. T.; Elliott, J. E.; Norstrom, R. J.; Whitehead, P. E.; Hart, L. E.; Cheng, K. M.; Bellward, G. D. J. Toxicol. Environ. Health 1994, 41, 435-450.

(40) Henshel, D. S. Environ. Toxicol. Chem. 1998, 17, 88-98. (41) Henshel, D. S.; Martin, J. W.; DeWitt, J. C. Environ. Health Perspect. 1997, 105, 718-725. (42) Custer, T. W.; Custer, C. M.; Hines, R. K.; Stromborg, K. L.; Allen, P. D.; Melancon, M. J.; Henshel, D. S. Arch. Environ. Contam. Toxicol. 2001, 40, 89-100. (43) Van den Berg, M.; Craane, B. L.; Van Mourik, S.; Brouwer, A. Ardea 1995, 83, 299-313. (44) Powell, D. C.; Aulerich, R. J.; Meadows, J. C.; Tillitt, D. E.; Kelly, M. E.; Stromborg, K. L.; Melancon, M. J.; Fitzgerald, S. D.; Bursian, S. J. Environ. Toxicol. Chem. 1998, 17, 2035-2040.

Received for review August 2, 2002. Revised manuscript received November 20, 2002. Accepted December 2, 2002. ES0208613

VOL. 37, NO. 5, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

831