Effects-based marine ecological risk assessment at a polychlorinated biphenyl-contaminated site in Saglek, Labrador, Canada



Although the presence and distribution of polychlorinated biphenyls (PCBs) in Arctic marine environments has been well documented, the implications for the health of biota are poorly understood. In the present study, multiple lines of evidence, including site-specific effects data, were used to assess PCB-related risks to marine biota at a contaminated military site in Saglek Bay, Labrador, Canada, from 1997 to 1999. Risks were evaluated for three components of the ecosystem: benthic invertebrates, a bottom-feeding fish (shorthorn sculpin, Myoxocephalus scorpius), and a diving seabird (black guillemot, Cepphus grylle). Average sediment PCB concentrations exceeded the Canadian interim sediment quality guideline level by 41-fold. However, sediment toxicity testing and a benthic community survey showed no evidence of adverse effects. In contrast, shorthorn sculpin and black guillemot PCB exposures (measured as sum of 55 congeners) were elevated enough to pose risks to survival or reproduction. Based on the collective evidence, the authors estimated that risks were posed by sediment PCB concentrations greater than 77 ng/g dry weight for black guillemots and 750 ng/g dry weight for shorthorn sculpins. The present study, along with two parallel studies, provided information to support the management decisions concerning potential remedial action on the contaminated sediments. This ecological risk assessment describes the steps and rationale taken to evaluate the risk posed by an area of PCB-contaminated marine sediments in an otherwise relatively pristine northern coastal environment. Environ. Toxicol. Chem. 2013;32:453–467. © 2012 SETAC


During an environmental site assessment in 1996, significant amounts of polychlorinated biphenyl (PCB)-contaminated soil were found at Saglek Bay, Labrador, Canada. A U.S. Air Force communication station operated at Saglek Bay for a period of 20 years ending in 1971 1. In the 1980s, the old site was demolished to make way for the construction of a North Warning System Long Range Radar facility (LAB-2) 2. The 1996 assessment also found preliminary evidence that contaminated soil close to the site's sea-lift landing beach had eroded into the adjacent marine environment, resulting in significant PCB contamination of nearshore sediments 2. The presence of such compounds in the sediments was of particular concern because of their tendency to biomagnify to high concentrations in Arctic marine food webs.

In conjunction with the excavation and removal of the PCB-contaminated soil from the terrestrial environment from 1997 to 1999, a series of studies was conducted to establish a scientific basis for selecting an appropriate management strategy for the contaminated marine sediments. The studies addressed questions involving the extent, distribution, and fate of contaminated sediments in Saglek Bay 1, 3, 4, the ecological risk posed by the contaminated sediments 5, and the feasibility of using existing remediation technologies to treat the contamination 6.

Although generic ecological risk assessment (ERA) and management frameworks for contaminated sites exist, their application to Arctic sites is difficult because of the lack of guidance documents for conducting ERAs in a northern setting. Environmental quality guidelines (e.g., Canadian Council of Ministers of the Environment [CCME] Canadian Tissue Residue Guidelines for the Protection of Wildlife Consumers of Aquatic Biota and CCME Canadian Sediment Quality Guidelines for the Protection of Aquatic Life) are generally based on surrogate species, interspecies extrapolations, bioaccumulation models, and safety factors and therefore have inherent uncertainties 7. Toxicity reference values (TRVs) for PCBs have been established for temperate species, but no known TRVs exist for Arctic species.

In addition to being an Arctic site, Saglek possesses certain unique geographical features that further highlight the value of conducting a site-specific effects-based ERA. Although contaminated sites typically possess highly complex mixtures of contaminants, the marine sediments at Saglek contained only one major contaminant (PCBs) in an otherwise relatively pristine environment 1, 8. The contaminated sediments within Saglek Bay are primarily well-sorted medium sands with very low organic carbon content, reflecting the high energy of this coastal environment 1, 3.

Most other studies of PCBs in marine environments have been conducted in industrial harbors where energy from waves and currents is limited and sediments consist of fine sands, silt, and clays, with substantial levels of organic carbon 9. Threshold and probable effects levels designed to guide assessments in such temperate sites 10 may therefore have limited value when applied to coarse sedimentary environments such as those encountered at Saglek Bay. Furthermore, the derivation of sediment quality guidelines is based on the protection of benthic invertebrates 10 and may not be protective of upper-trophic-level species such as seals or seabirds.

The protection of multiple trophic levels was important in the case of Saglek, because this area has been a popular hunting and fishing area for centuries. Since 2005, Saglek has formed part of the Labrador Inuit Settlement Area and now represents the southern boundary for Torngat Mountains National Park.

Here, we present the results of the studies that were conducted to assess the risk of the PCB-contaminated marine sediments in Saglek Bay as they existed during the study period (1997–1999). To our knowledge, this ERA represents one of the first assessments to measure the effects of PCBs on wildlife species at several different trophic levels in a natural (northern) setting 11, 12. Furthermore, it was one of the first assessments of a contaminated site in northern Canada that informed the development and led to the implementation of a site-specific risk management strategy designed to protect the marine ecosystem. This approach and subsequent management strategy are particularly relevant today, because numerous environmental and ecological assessments in support of northern development projects are underway.


Study site characterization

Saglek Bay lies along the northern Labrador coast at approximately 58°23′N, 62°35′W (Fig. 1). The contaminated beach at the radar station, which is believed to have provided the source of PCB contamination to the marine environment, lies in Saglek Anchorage along the south shore of Saglek Bay (Fig. 1). The surficial sediment distribution varies from sand interspersed in a matrix of boulders and cobbles in the nearshore area in the anchorage (less than 15–20 m deep), to a poorly sorted lag deposit on the slope, and homogeneous mud (silt and clay) in the central parts of the deep basins (water depths greater than 70 m) 3, 4. Relative sea level in the vicinity of Saglek has fallen by about 54 m since deglaciation, bringing parts of the seabed up into the wave base, allowing reworking of previously deposited marine sediments along slopes and on sills 3. Saglek Anchorage is a highly energetic environment, with significant resuspension and redistribution of sediments by wave action, especially when large (8 m), long-period (15 s) waves enter the mouth of the Bay from the northeast and east (Labrador Sea) during storm events 3.

Figure 1.

Distribution of polychlorinated biphenyl (PCB)-contaminated sediments in Saglek Bay, Labrador, Canada, during the study period (1997-1999) and location of benthic (n = 22, squares), shorthorn sculpin (n = 13, circles), and black guillemot (n = 7, triangles) sample sites for site-specific studies of PCB exposure and effects.

After the discovery of the PCB-contaminated soil at the radar station beach in 1996, the extent of contaminated marine sediments was delineated through a sediment grab sampling and coring program 1, 3. Polychlorinated biphenyl concentrations (sum of 51 congeners) in 191 sediment samples collected within 25 km of the contaminated beach ranged from 0.24 to 62,000 ng/g dry weight, with an exponential decrease in concentration with distance from the beach (r2 = 0.79, n = 191, p = 0.0001 1; Fig. 1). Polychlorinated biphenyl concentrations near the beach and on the surrounding nearshore platform were in the order of 1,000 to 10,000 ng/g dry weight during the 1997 to 1999 period, and concentrations in the intertidal area near the source were an order of magnitude greater before the clean-up of contaminated beach soil started in 1997. For comparison, PCB concentrations in marine sediments in northern Canada, remote from local point sources of PCBs, typically fall below 0.3 ng/g dry weight 13.

Rationale, objectives, and conceptual model

The aim of the Saglek Bay ERA, which was identified through consultation with stakeholders, including the Labrador Inuit Association (LIA) and the provincial and federal governments 2, was the protection of the health of the Saglek Bay marine ecosystem from the harmful effects of PCBs. As a means of informing decision-making concerning management measures and possible remediation, an ERA for the PCB-contaminated marine sediments was designed based on the PCB concentrations found in the environment during the study period (1997–1999). The ERA was conducted following the basic approach outlined by the Canadian Council of Ministers of the Environment 14. However, multiple lines of site-specific, scientific evidence from multiple trophic levels were used to assess both the exposure and the associated adverse effects of the PCB-contaminated sediments within the local food web.

To aid in the selection of receptors for the assessment, a conceptual model was developed to describe the relationship between source, transport pathway, and receptor in the transfer of PCBs through the marine ecosystem of Saglek Bay (Fig. 2). The model is site specific in that it depicts the primary modes of PCB transfer from contaminated sediments through the subarctic marine food web at Saglek Bay. Because of the hydrophobic and lipophilic nature of PCBs, they are not readily dissolved into the water column but are readily sorbed onto particles and especially organic material 15. Of relevance here also is that the PCBs introduced into Saglek Bay were largely a highly chlorinated mixture resembling Aroclor 1260 1 and thus were extremely hydrophobic 15. Furthermore, the PCBs likely entered the marine environment already associated with soil particles (via erosion of contaminated soil from the station's beach and backshore). Thus, in the conceptual model (Fig. 2), PCBs are transferred from sediment into the base of the food chain primarily by direct contact with or ingestion of contaminated sediments, whereas uptake of dissolved PCBs from the water column is considered negligible 1. Because PCBs are persistent and resistant to metabolism, biological uptake frequently exceeds all routes of elimination, with the result that PCBs bioaccumulate in organisms (reaching higher concentrations than in the surrounding environment) and are effectively transferred through the food web via the diet, biomagnifying to higher concentrations with each successive step in the food web 16. Thus, the extent to which organisms are associated with contaminated sediment, either directly or indirectly (via the food web), is important in determining PCB exposure, as is the trophic level of the organism.

Figure 2.

Conceptual model depicting the primary polychlorinated biphenyl exposure routes from contaminated sediments in Saglek Bay, Labrador, Canada (adapted from Environmental Sciences Group 5).

Preliminary food chain investigations conducted from 1997 to 1999 provided supporting evidence for this conceptual model 1. Polychlorinated biphenyl concentrations (sum of 55 congeners) were determined in several sediment-associated species collected from Saglek Bay, including benthic invertebrates (bivalves and sea urchins), bottom-feeding fish (shorthorn sculpins, Myoxocephalus scorpius), and diving seabirds (black guillemots, Cepphus grylle). At the same time, PCBs were also determined in pelagic migratory species, including arctic char (Salvelinus alpinus), ringed seals (Phoca hispida), and great black-backed gulls (Larus marinus). These latter species undergo migrations and therefore are not limited to PCB exposures within the study area. The sediment-associated biota (invertebrates, sculpins, and guillemots) collected near the contaminated beach at Saglek were found to have PCB levels in their tissues several orders of magnitude above background levels (i.e., the levels in samples of the same species, collected remote from the site) 1. In contrast, arctic char sampled in Saglek showed no evidence of PCB accumulation from the contaminated sediments, whereas great black-backed gulls and ringed seals had variable PCB levels, including occasional high levels. These elevated levels appeared to relate more to their high trophic level than to uptake from the local contaminant source 1. Benthic invertebrates, shorthorn sculpins, and black guillemots were therefore chosen as the receptors to evaluate the ecological risk posed by the PCB-contaminated sediments at Saglek (Fig. 2).

Receptor characterization

The benthic invertebrate community represents the first step by which sediment-associated PCBs are introduced into a marine food web. Several factors resulted in our choosing to use a community approach (rather than, for example, a species-specific approach) for assessing risk to these receptors. First, although there is very little information on the benthic invertebrate communities of northern Labrador, generally, the variety and abundance of benthic fauna in northern regions is very high, in some instances greater than at similar water depths in temperate or tropical zones 17. The type of benthic community present varies in relation to water depth and the nature of the substrate 17. Thus, there may not be one species of benthic invertebrate found ubiquitously throughout the region of PCB contamination in Saglek Bay. Second, there is a wide range of vulnerability in exposure to sediment-associated contaminants among benthic invertebrate species. Differing life history and feeding habits can substantially affect levels of PCB accumulation, with filter-feeders typically showing the lowest levels of accumulation and deposit feeders and scavengers showing greater accumulation 18. Similarly, species that live in direct contact with sediment, rather than living in constructed tubes, also show greater accumulation 19. The low sediment organic carbon levels at Saglek (averaging 0.2% in the sand-sized sediments to 0.7% in muddy sediments from deeper waters 1) increase the potential vulnerability of some species, such as benthic infauna, to PCB accumulation. Finally, some benthic invertebrates tend to be relatively sensitive to organic contaminants such as PCBs, exhibiting adverse effects such as reduced survival, abnormal development, and impaired reproduction at relatively low levels, whereas other species are less sensitive 20.

Shorthorn sculpins are sluggish, benthic fish, abundant all along the Labrador coast and throughout the Arctic. They are generally limited to nearshore areas, with very little movement or migration 21. A related species, the fourhorn sculpin (Myoxocephalus quadricornis), which is less abundant in Saglek Bay, has been recommended as an indicator species for the monitoring of Arctic estuaries 22. Generally, shorthorn sculpins are bottom-feeders that will opportunistically consume any prey of an appropriate size 21. Because prey is often taken directly off the bottom, it is likely that some sediment is also ingested. Consumption of both benthic invertebrates and sediment may represent significant exposure pathways for PCB accumulation. As a result, shorthorn sculpins are excellent indicators of localized sediment contamination.

Black guillemots are diving seabirds that nest in small colonies or alone along rocky shorelines, including Saglek Bay 23. Although guillemots are present in Saglek Bay only during the ice-free period, they are one of the few seabirds that remain in northern waters (and thus remote from most contaminant sources) year round 23. Pairs nest in cliff cracks and talus slopes and return to the same nesting site in subsequent years 11, 23. Generally, black guillemots forage on littoral and benthic fish and invertebrates close to their nesting sites, diving to depths of up to 40 m 11. Their foraging habits, including the consumption of many sediment-associated organisms, combined with their affinity for near-shore areas and fidelity to specific sites throughout their life suggest that black guillemots are an ideal upper-trophic-level receptor in which to evaluate the effects of local sources of sediment contamination.

Exposure characterization

Benthic invertebrates

Twenty-two sites within Saglek Bay were selected for the benthic invertebrate analyses based on PCB concentrations in the sediments (Fig. 1). These included five reference sites (<5 ng/g dry wt PCB anticipated in sediments), five sites with anticipated PCB concentrations from 5 to 250 ng/g dry weight, six sites with anticipated PCB concentrations between 250 and 2,000 ng/g dry weight, and six sites with PCB concentrations anticipated to exceed 2,000 ng/g dry weight. The exposure assessment for benthos was based on the PCB concentrations measured in the sediment samples at these 22 sampling sites. The sampling program was designed to minimize the influence of abiotic parameters, such as grain size and organic carbon content, by sampling sites where these factors would be relatively constant. Samples were also screened to rule out the presence of other organochlorine contaminants of potential toxicological concern. Sampling sites were at depths of 1.5 to 24.4 m. At each site, six sediment–benthic invertebrate samples were collected with a ponar grab (30 × 30 cm) that was rinsed with seawater between uses. These discrete samples were combined and homogenized in a polyethylene container to create one composite sample for each site. Subsamples from each homogenized composite sample were then taken for toxicity testing (3.5–4 L), contaminant analysis (1 L), and determination of geochemical parameters such as grain size and organic carbon content (0.5 L). All sediment samples were placed in I-Chem amber jars with Teflon-lined lids. The remainder of the composite sample was sieved through 0.5-mm mesh, and benthic invertebrates were preserved in 10% buffered formalin.

Sediment samples were stored at 4°C prior to toxicity testing. Subsamples were collected for triplicate analysis of sulfide, redox potential, and ammonia by specific ion electrodes. One day prior to beginning the toxicity tests, sediments were homogenized again, subsampled into five 1-L mason jars filled with seawater, and aerated overnight with oil-free compressed air.

Shorthorn sculpin

Shorthorn sculpins (n = 101) were collected from 13 locations in Saglek Fjord (Fig. 1)—12 (locations 23–34) located within 9 km of the contaminated beach (i.e., within the boundaries identified for the ERA 5), and the 13th from reference sites (location 35) located between 13 and 25 km from the beach. The mean sculpin tissue (whole body minus liver) PCB concentrations for the 12 Saglek Bay locations and the reference sites are presented in Table 1. Whole body minus the liver was used for PCB analysis because liver tissue was used for biomarker measurements 12.

Table 1. Polychlorinated biphenyl concentrations in biota (ng/g wet wt) from the Saglek Bay, Labrador, Canada, area
MatrixNo.aLocationNLipid mean (%)Sculpin Geometric mean (range; ng/g wet wt)Sediment geometric mean (ng/g dry wt)
  • a

    No. corresponds to the sample location indicated in Figure 1.

  • b

    Polychlorinated biphenyls include the following congeners: 31/28, 52, 74, 70/76, 66, 95, 84/92, 89/90/101, 99, 97, 87, 85, 110, 118, 105, 136, 151, 144/135, 149, 134, 146, 153, 141, 138/163/164, 158, 129, 128, 156, 157, 179, 176, 178, 175, 187/182, 183, 185, 174, 177, 171, 172, 180, 193, 191, 170/190, 189, 201, 197, 198, 199, 196/203, 195, 194, 205, 208, and 206.

Shorthorn sculpinb (whole body minus liver)23Beach at LAB-2111.32,660 (217–13,070)1,207
 24East of beach (1 km)101.8748 (376–1,168)54.2
 25Major Point East (3 km)51.6121 (16–744)10.6
 26First Bay West (4 km)141.5595 (313–11,797)43.7
 27Second Bay West (4 km)51.8548 (422–787)11.6
 28Marker Point (5 km W)61.3114 (33–313)6.6
 29Site Summit (5 km E)51.6391 (276–752)4.4
 30Big Island (6 km E)62.043 (19–79)1.8
 31Bluebell Island (6 km E)71.151 (27–158)1.6
 32Big Island (6 km NW)61.834 (34–56)2.9
 33Shuldham Island (8 km NW)51.424 (6.7–26)1.3
 34St. John's Harbour (9 km W)41.243 (36–44)1.9
 35Reference sites (13–25 km W)171.57.8 (4.8–40)0.4
Black guillemot nestlingb (liver)36–37Beach113.91,099 (295–6,480)156 (10–500)
 38–39Islands103.995 (34–186)2.6 (1.3–3.9)
 40–42Reference154.032 (11–70)0.4
Black guillemot nestlingb (egg)36–37Beach89.18,356 (333–47,900)102 (27–264)
 38–39Islands11121,028 (388–4,540)2.0 (1.3–3.9)
 40–42Reference611404 (218–702)0.4

Because the ERA results were intended to support decision-making about the need to consider remediation of the contaminated sediments at Saglek, it was important to develop a means of linking organism PCB exposures (measured as concentrations of PCBs in sculpin or guillemot tissues) with concentrations of PCBs in sediments. To this end, an average value was derived from sediment concentrations collected within a 500-m radius of each sculpin sampling location (see Table 1) and compared with the concentrations in sculpin tissues. A 500-m-radius foraging area was selected based on previous studies observing sculpin behavior (V.A. Pepper, 1974, MSc Thesis, Memorial University of Newfoundland, St. John's, NL, Canada) and was considered appropriate for Saglek, given the significant differences in PCB concentration between sculpin collected from adjacent sites located less than 1 km apart (Table 1). The relationship was assessed by linear regression analysis of mean sediment PCB concentrations and concentrations of PCBs in individual sculpin (whole body minus liver) on a lipid-weight basis (Fig. 3). The sculpin–sediment PCB relationship (R2 = 0.79, p < 0.0001, n = 101) yielded the following equation: equation image

Figure 3.

Relationship between polychlorinated biphenyl (PCB) concentrations in sediment (ng/g total organic carbon [TOC]) and shorthorn sculpin (whole body minus liver; ng/g lipid) in Saglek Bay, Labrador, Canada (1997-1999; adapted from Environmental Sciences Group 5).

The strong relationship indicates that sculpin exposures to PCBs closely reflect PCB concentrations in the sediment. Thus, for the purposes of the ERA, the relationship was used for describing and predicting total PCB concentrations in sculpin, given PCB concentrations in sediment, and vise versa.

Black guillemots

Active black guillemot nests (n = 54) were identified in July and August, 1999, at distances of 1 to 18 km from the contaminated beach 11. Black guillemot chicks (n = 36) were monitored every 2 or 3 d, weather permitting, from hatch to sample collection date, when the nestlings were about 22 d old 11. Body size measurements (head–bill length, tarsus length, wing length, and body mass) were taken to determine the age of the chicks (Burgess, unpublished data). Black guillemot eggs were collected from 14 of the 54 nest locations prior to nestling collection. Eggs were frozen and shipped to the National Wildlife Research Centre (NWRC) in Hull, Quebec, Canada, for further processing and archiving until analysis. The distribution of nests fell into three groups: (1) the beach group (locations 36 and 37; Fig. 1), consisting of nests located on the rocky shoreline within 4 km of the contaminated beach; (2) the islands group (locations 38 and 39; Fig. 1), located on two unnamed islands (hereafter called Rocky and Little Bluebell Islands) 5 to 6 km from the beach; and (3) the reference group (locations 40–42; Fig. 1), located on Glitsch Island and two unnamed islands (hereafter called Black and Gull Islands) 16–18 km from the beach (Fig. 1) 8, 11. The mean black guillemot nestling liver and egg PCB concentrations for each group are presented in Table 1.

To assess the relationship between sediment and black guillemot nestling or egg PCB concentrations, an average value derived from sediment concentrations collected within a 500-m radius of each nest was calculated and compared with guillemot liver or egg concentrations on lipid-weight basis. A 500-m radius corresponds to the low end of the range of foraging distances reported in previous studies 24 but was considered appropriate given the steep gradient in guillemot PCB concentrations with distance from the contaminated beach and supported by observations of foraging guillemots at Saglek 5. The relationships were assessed by linear regression analysis of mean sediment PCB concentrations and concentrations of PCBs in individual guillemot nestlings or eggs (Fig. 4). The guillemot nestling–sediment PCB relationship (R2 = 0.85, p < 0.0001, n = 36) and egg–sediment PCB relationship (R2 = 0.72, p < 0.0001, n = 25) yielded the following equations: equation image; and equation image

Figure 4.

Relationship between polychlorinated biphenyl (PCB) concentrations in sediment (ng/g organic carbon [TOC]) and black guillemot nestling livers (lower line; ng/g lipid) and eggs (upper line; ng/g lipid) in Saglek Bay, Labrador, Canada (1997-1999; adapted from Environmental Sciences Group 5).

The strong relationships found for the nestling livers and eggs indicate that the guillemot exposures closely reflect the average PCB concentrations in the sediment. Therefore, the sediment–guillemot nestling and sediment–egg relationships were used for predicting PCB concentrations in the livers of the guillemot nestlings and eggs given PCB concentrations in the sediment, and vise versa. The mean sediment PCB concentrations that correspond to the mean guillemot nestling and egg PCB concentrations are shown in Table 1.

Hazard identification

General assessment and measurement endpoints were developed to guide the assessment for each receptor (Table 2). The assessment endpoint assigns a target level of protection for each receptor by describing the impact or fate from which the receptor must be protected. Measurement endpoints are measurable changes in an assessment endpoint that can be used to infer a measure of protection or to evaluate risk to the assessment endpoint 25. For each measurement endpoint, effects thresholds were developed based on literature toxicity thresholds (Table 2). This is routinely performed for effects-based risk assessments. For example, several studies have identified a 20% reduction in fish survival, growth, or reproduction as potentially significant 26. When a literature-based effects threshold could not be identified, a statistical comparison between biological endpoints measured in sculpin or black guillemot nestlings from the Saglek reference areas and the contaminated beach area was used as the criterion for determining significance.

Table 2. List of the biological receptors, assessment endpoints, measurement endpoints, and criteria used to determine significance
ReceptorAssessment endpointsMeasurement endpointsCriteria for determining significance/effect threshold
Benthic invertebratesPrevent significant adverse impact to invertebrate communities due to toxicitySediment toxicity tests 
  • Amphipod mortality (10-d test)• Significant decrease in survival results for beach sediments compared with reference and control sediments
  • Bacterial luminescence (Microtox)• Samples above the Microtox solid-phase toxic hazard level (EC50 < 1,000 mg/L)
  Benthic community surveyCommunity composition vary significantly among contaminated sites and reference sites
  Species abundance and diversityNumber of individuals, taxa, and diversity varies significantly between contaminated and reference sediments
Shorthorn sculpinPrevent reductions in the abundance or productivity of local sculpin populations due to toxicityBiochemistry and physiology measurements 
  • Mixed function oxygenase (MFO) induction• Significant increase in the beach area compared with the reference area
  • Body condition• Significant differences in the beach area compared with the reference area
  • Liver mass• Significant differences in the beach area compared with the reference area
  • Lipid content• Significant differences in the beach area compared with the reference area
  Literature toxicity data 
  • Reported polychlorinated biphenyl toxicity thresholds for risk to fish survival and reproduction

• >100,000 ng/g wet weight, concentration associated with lethality in fish 37

   • >1,000–5,000 ng/g wet weight, concentration range associated with reduced reproductive success
Black guillemotsPrevent reductions in the abundance or productivity of local black guillemot populations due to toxicityField observations 
  • Hatching success• Statistically significant reduction of at least 10% compared with the reference area (G. Fox, personal communication, Canadian Wildlife Service, National Wildlife Research Centre, Environment Canada, Ottawa, ON)
  • Nestling growth/development

• 30% decrease in chick growth in the beach area; indicative of risks to survival and reproduction; 20% decrease for fledgling guillemots 50

indicative of risks of sublethal effects
  • Immune response (skin test)• Immune suppression of 50% or more in the beach area; indicative of risks to survival and reproduction; 25% was the threshold indicative of significant risk of sublethal effects
  Biochemistry, physiology, and histology measurements 
  • MFO induction

• Threefold MFO induction 51

; threshold indicative of sublethal effects
  • Organ enlargement (liver)

• 25% enlargement of liver 52

; threshold indicative of sublethal effects
  • Decrease in organ mass• Sublethal effects

 30% 53


 25% 34


 30% 54


 50% 55


 50% 55


 50% 55

  • Malic enzymes

• 30% reduction in malic enzyme levels 56

; threshold indicative of sublethal effects
  • Thyroid activity

• 50% increase or a twofold increase in thryoxine and triiodothyronine levels 57

; indicative of sublethal effects
  • Retinoid levels

• 40% decrease in retinoid levels 34

; threshold indicative of sublethal effects
  • Testosterone and estradiol levels

• 20% difference in testosterone and estradiol levels 58

; threshold indicative of sublethal effects. 40% difference in testosterone and estradiol levels; threshold indicative of risks to survival and reproduction
  • Corticosterone levels

• 40% decrease in corticosterone levels 50

; threshold indicative of sublethal effects
  • Gonad histology

• Survival/reproduction: presence of any abnormalities that could compromise reproductive capacity 59

; threshold indicative of significant risks to survival and reproduction
  • Liver porphyrin content

• Two-fold (200%) increase in porphyrin levels 51

; threshold indicative of sublethal effects

Benthic invertebrates

To analyze the risks of effects from PCB contamination on the benthic invertebrate community, two sediment toxicity tests (10-d infaunal amphipod survival [Amphiporeia virginiana] and Microtox [Vibrio fischeri] solid-phase) and a benthic community survey were carried out across a gradient of sediment PCB concentrations (range, <1–7,000 ng/g dry wt). Use of multiple benthic invertebrate endpoints provides the best estimation of the actual toxicity of the sediments 27.

Amphipod (A. virginiana) toxicity testing was conducted according to the procedures outlined by Environment Canada 28. Amphipods were acclimated to 10 ± 2°C and added to the test jars containing sediments from each site (five replicates per site) for 10 d. A 24-h photoperiod, 10 ± 1°C temperature range, and sufficient aeration were maintained and monitored daily. One replicate per site was monitored three times per week for dissolved oxygen, pH, and salinity. At the end of the 10-d test period, the contents of each jar were sieved with a 0.5-m mesh, and any missing or immobile amphipods were counted as dead. In addition to the reference sites from Saglek, two control samples from the amphipod collection site (Martinique Beach, Nova Scotia) were also analyzed for quality assurance/quality control purposes. To ensure that the population of amphipods used in the test was of normal sensitivity, a reference toxicant test was also conducted with cadmium chloride using a water-only exposure for 96 h. It was determined that the 96-h LC50 (concentration calculated to cause 50% mortality) was within standard operating limits of the Shewhart Control Chart. Results were analyzed by comparing amphipod survival among the Saglek reference samples (n = 5, PCB concentrations <5 ng/g) with survival among the more highly contaminated sediments.

The Microtox test was conducted according to Environment Canada 29 and Microbics Corporation 30 methods. A dilution series of 12 concentrations and three controls was prepared from diluted wet sediment. Bacterial reagent was added to the dilutions, with incubation at 15°C for 20 minutes and then filtering. The filtrate was transferred to the Microtox analyzer, and bioluminescence was recorded after 10 min. Median effective concentration (EC50) values were calculated as the concentration at which the light output by the bacteria was reduced by 50%, in comparison with controls. Three aliquots of sediment were dried to determine moisture content and to convert EC50 values to milligrams dry sediment/liter of diluents. The distributions of all variables were examined. The variables with skewed distributions were transformed to ensure normality. Polychlorinated biphenyl concentrations were log transformed, and numbers of taxa and individuals were square-root transformed. Analyses were performed on both nontransformed and transformed data, and the interpretation of results did not differ. Linear regression analysis was used to determine relationships between data from the amphipod and Microtox toxicity tests and PCB sediment concentrations. All samples for the Microtox toxicity test were also compared with the Microtox solid-phase toxic hazard level (EC50 < 1,000 mg/L) 31.

For the benthic community analyses, preserved benthos were identified to the lowest taxon level possible (species level in the vast majority of samples) and enumerated. To ensure sorting accuracy, 10% of the samples were resorted and counted with an accuracy of at least 98%. Two samples (TS99-01 and TS99-03) were discarded because the samples had been compromised during transport. Linear regression analysis was used to determine the relationships between the numbers of taxa and individuals and the Shannon–Wiener diversity index versus PCB sediment concentrations. For benthic community analyses, benthic samples were categorized according to actual PCB concentrations in the sediments where they were collected: high (2,350–7,000 ng/g), medium (1,360–2,140 ng/g), low (33–732 ng/g), and reference (<3.3 ng/g) groups. Benthic community analysis was performed in Plymouth Routines Multivariate Ecological Research (PRIMER) Version 6.1.10 (PRIMER-E Ltd.). Abundance data were square-root transformed. Using 100 restarts and a minimum stress formula (Kruskal) of 0.01, nonmetric multidimensional scaling (MDS) was used to illustrate the degree of difference in the community structure among the locations sampled. To test the null hypothesis of no differences in the benthic community structure of the groups, a one-way analysis of similarity (ANOSIM) with 999 permutations was carried out on the Bray–Curtis similarity matrix.

Shorthorn sculpin

To assess the effects of local PCB contamination on shorthorn sculpins, we compared PCB concentrations in Saglek Bay sculpins with values reported to affect fish reproduction and survival and with values for four site-specific indicators applied at Saglek. These latter indicators include fish body condition (condition factor), lipid content, relative liver mass (hepatosomatic index), and activity of a cytochrome P450 1A (CYP1A) enzyme (ethoxyresorufin-O-deethylase [EROD]) in sculpin liver tissue. Condition measurements and activity of inducible liver enzymes, such as mixed function oxygenase (MFO or CYP1A), have been used in previous studies as overall indicators of exposure to PCBs and fish health 32. Induction of MFO enzymes can alter the metabolism of important compounds such as vitamins and steroid hormones, which may have secondary effects on development, growth, and reproduction. The site-specific biological responses were measured in five shorthorn sculpin samples from each of seven sites, including a reference site 12. Polychlorinated biphenyl and EROD data were log-transformed to meet normality assumptions for parametric analyses. Each endpoint was examined using linear regression (p < 0.05), in keeping with the gradient approach of the overall study. Ethoxyresorufin-O-deethylase activity and other measures of biological data were also compared among locations using analysis of variance (ANOVA), and differences between the individual groups were tested using Tukey's comparisons (see Kuzyk et al. 12 for details of methods and data analysis). Threshold levels identified in the literature for risks to reproduction and survival were applied to sculpin tissue residue measurements and translated into sediment PCB concentrations, using the sediment–sculpin PCB relationship (Fig. 3).

Black guillemots

To analyze the impacts of the local source of PCB contamination on black guillemot nestlings, we compared measured PCB concentrations in Saglek Bay guillemot nestling livers with 13 site-specific biological indicators (Table 2). The design of the guillemot effects study was based on programs that have been conducted for more than a decade to monitor contaminants and associated effects in fish-eating birds of the Great Lakes 33. Responses at Saglek were measured across a wide range of PCB exposures. The present study focused on guillemot nestlings still in the developmental stage because birds are more sensitive to the effects of PCBs during this time. Furthermore, we were able to minimize the effects of age as a confounding factor. The endpoints—MFO induction, organ enlargement, malic enzyme levels, thyroid activity, liver porphyrin content, organ atrophy, retinoid (vitamin A) levels, steroid hormone levels, genetic variation, gonad histology, growth, immune function, and hatching success—represent a range of physiological responses that have a range of implications for guillemot health. At one extreme are indicators merely indicating that an organism has responded biologically to the PCB exposure (e.g., MFO induction), whereas other endpoints indicate potential for impairment to reproduction, development, or survival (e.g., hatching success). The assessment endpoint for black guillemots is protecting the abundance and productivity of the local population, ensuring successful reproduction and survival.

A phytohemagglutinin (PHA) immune test was conducted on the guillemot nestlings when they reached a head–bill length of 68 mm, following the methods of Grasman et al. 34. After the PHA test, 36 nestlings were collected from the beach (n = 11), islands (n = 10), and reference (n = 15) groups. Using a heparinized syringe, blood samples were collected from each nestling and centrifuged immediately. Plasma was immediately frozen in liquid nitrogen. After collection of the blood, the nestlings were necropsied and identified by gender according to gonadal examination, and tissues were collected for laboratory analyses. Liver samples were subdivided and frozen for the analysis of PCBs and biomarkers (vitamin A levels, porphyrin levels, MFO induction, and malic enzyme activity). See Kuzyk et al. 11 for specific field methods detailing measurements as well as tissue collection for liver biomarkers and preservation. Left and right thymus and thyroid glands were separated for freezing and for histological preservation in formalin. Bursa, heart, spleen, kidneys, gastrointestinal tracts, and reproductive systems were preserved in formalin for histological analysis. Prior to submission for laboratory analysis, each sample was given a blind number that did not provide any information about its sampling location.

To evaluate the risk to reproduction and survival, results for the multiple endpoints were evaluated together in a weight-of-evidence approach. Although not all endpoints relate directly to effects as extreme as impaired reproduction or survival, they were useful in placing guillemot health along the overall continuum between good and impaired health. Figure 5 illustrates how the endpoints were evaluated in a weight-of-evidence approach to determine whether there was a significant risk to guillemot reproduction or survival. The horizontal lines represent each endpoint; a line was used rather than a point because each endpoint has a range of associated responses. The endpoints were ranked along the vertical axis according to how closely they related to guillemot survival or reproduction. For example, as hatching success is a direct measurement of reproduction, it was ranked highest and positioned at the top of the axis. The horizonatal axis corresponds to the interpreted significance of the endpoints in terms of guillemot health, which is a function of the way in which they were ranked on the vertical axis, as well as the results or responses found in the ERA. The two vertical lines represent thresholds, points at which the endpoints were considered indicative of significant risk for the purpose of this ERA. The endpoints intersecting with the red vertical line are indicators of significant risk to guillemot survival or reproduction. Each endpoint had to exceed the established threshold, described in Table 2, to indicate the presence of risk. The endpoints intersecting with the blue vertical line indicate a threshold for sublethal effects. This threshold was used to provide an additional means of determining the overall health of the Saglek guillemot population.

Figure 5.

Weight-of-evidence approach for interpreting the guillemot nestling endpoints and evaluating the risk of sublethal effects and risk to survival or reproduction (adapted from Environmental Sciences Group 5).

The thresholds for this ERA were estimated using the scientific literature, the professional judgment of wildlife toxicologists, and statistical analyses. The literature was first reviewed to identify previous studies in which these biomarkers were investigated. The study results were then examined to identify the level of response observed, whether other effects were correlated or co-occurred, and the significance of the study findings. The distributions of all variables were examined. Variables with skewed distributions were log-transformed to meet normality assumptions. Each endpoint was examined using linear regression (p < 0.05), in keeping with the gradient approach of the overall study. Measures of biological endpoints were also compared among the three PCB concentration groups using ANOVA. Differences between the individual groups were tested using Tukey's comparisons. Only those endpoints showing a significant response according to at least one of the statistical tests were included in further analyses. The endpoint responses were then evaluated to determine whether they exceeded the identified threshold (Table 2), by comparing all measured responses for each endpoint to the average responses measured in reference-group birds. If responses equaled or exceeded these thresholds, the sediment PCB exposures associated with the thresholds were determined using the sediment–guillemot PCB regressions (Fig. 4). When relationships were nonlinear, thresholds were evaluated using the mean responses (relative to the reference group) for the beach and islands groups; PCB exposures associated with any equaled or exceeded thresholds were estimated from the average group PCB exposures. The identified threshold values, presented in Table 2, are believed to reflect the present state of knowledge at the time of the study (1999) concerning PCB effects on birds.


Risk characterization

Benthic invertebrates

Mean survival of the marine amphipods (A. virginiana) in the 22 Saglek sediments ranged from 72 to 96%, including 72 to 87% in the reference sites (n = 5, PCB concentrations <5 ng/g) and 78 to 96% in the more contaminated sites (n = 17; Supplemental Data, Table S1). The mean survival of amphipods for two control sediments—which were collected together with the amphipods at Martinique Beach, Nova Scotia, Canada—was 94% and 85%. Based on the regression analysis, the results from the amphipod survival test and Microtox (V. fischeri) assay versus sediment PCB concentrations were not statistically different (p > 0.05). All samples were below the Microtox solid-phase toxic hazard level (EC50 < 1,000 mg/L) 31.

In total, 96 taxa were found in the benthic community analysis. Samples were numerically dominated by the polychaetes Capitella capitata, Eteone longa, Marenzellaria viridis, Microspio sp., Ophelia limacina, and Spio filicornis; the amphipods Oediceros saginatus and Protomedia fasciata; and the nemertean Cerebratulus sp. Based on the regression analysis, the number of individuals, taxa, and diversity versus sediment PCB concentrations were not statistically different (p > 0.05). Analyses of similarity among reference and contaminated sites were not significantly different (R = 0.06, p > 0.05). Analysis of the benthic community structure using nonmetric multidimensional scaling (nMDS) also showed reference sites interspersed among, or lying close to, contaminated sites. Thus, the results of sediment toxicity testing and a benthic community survey provide no evidence of adverse effects on benthos associated with the PCB-contaminated sediments in Saglek Bay.

The finding of no adverse effects on benthos from the contaminated sediments at Saglek may seem surprising, especially in view of the relatively high PCB concentrations (average PCB concentrations within 1.5 km of the contaminated beach = 884 ng/g dry wt 12) compared with the Canadian sediment quality guideline (21.5 ng/g dry wt) and probable effects levels (189 ng/g dry wt), which are designed to be protective of benthic invertebrates 10. However, there is limited information about the toxicity of PCB-contaminated sediments to benthos under conditions in which cocontaminants are present at negligible levels, as they are at Saglek. It is possible that toxicity is significantly reduced 20. Furthermore, slight differences in sediment characteristics (e.g., grain size or total organic carbon) can have an effect on the survival of certain benthic organisms 35. Therefore, the effects levels established from previous studies in areas of fine-grained sediments (e.g., harbors) might not be transferable to the sand-dominated (with low organic carbon content) benthic community at Saglek. A study of benthic community health in relation to PCB contamination in San Francisco Bay also found no relationship between species diversity and PCB concentrations within the range of 9 to 1570 ng/g, which is a range similar to that at Saglek 36.

Shorthorn sculpin

Literature-derived toxicity values. Polychlorinated biphenyls have been associated with toxicity related to a variety of endpoints in fish, including lethality, immune suppression, liver lesions, impaired reproduction, growth inhibition, cellular changes, and biochemical effects 37. Two endpoints (lethality and reproductive success) were selected, for which a literature-derived PCB threshold concentration served to guide the interpretation of PCB risks to sculpin health. An exhaustive review of the literature concerning the response of aquatic organisms to PCB exposure 37 proposed a threshold level for lethality in adult male and female fish and reproductive effects in adult females of about 100,000 ng/g wet weight. For progeny, Niimi 37 proposed a threshold level for reproductive effects of about 50,000 ng/g wet weight. Other laboratory studies reviewed by Kime 38 showed adverse effects on reproduction, including impaired ovarian growth and hormonal changes, at concentrations ranging from 50 to 5,000 ng/g wet weight. Establishing a threshold of risk to fish survival and reproduction based on the literature data is complicated by various studies reporting contradictory results for PCB effects. Also, thresholds vary among different species, among tissues analyzed, and with the mix of PCB congeners analyzed. However, based on the literature, the threshold for risk to reproductive effects in fish appears to be in the range of 1,000 to 5,000 ng/g wet weight, with some risk of effects occurring with as low as 100 to 1,000 ng/g wet weight. The former part of the range coincides with the average sculpin PCB exposures (2,660 ng/g wet wt) in the beach group (Table 1). Based on the sediment–sculpin PCB relationship established for the exposure assessment (Fig. 3), the sediment PCB concentration associated with this sculpin exposure is approximately 750 ng/g dry weight. The higher sculpin PCB concentrations measured at the other sites as far as 3 km east or west of the beach approach the lower end (1,000 ng/g wet wt) of this threshold concentration (Table 1).

Site-specific indicators

Results for lipid content (%), condition factor (CF), hepatosomatic index (HIS), and EROD activity have been reported by Kuzyk et al. 1. Lipid content, CF, and HIS were similar between sexes and among the four groups, and EROD activity was significantly different among the four groups (p < 0.0001). The average activity in sculpins from the beach group was nearly fourfold greater than average levels at the next most contaminated site (west 1 km) and 25-fold greater than the average levels at the reference sites 1. Furthermore, there was a strong correlation between EROD activity and tissue PCB concentrations (r2 = 0.60, p < 0.05). To interpret the observed biological responses in the context of ecological risk, Kuzyk et al. 1 used an upper estimate of background EROD activity as a threshold for EROD induction, above which there may be an increase in EROD activity associated with PCB concentrations. The determined threshold for EROD induction for Saglek Bay sculpin was 3,620 ng/g lipid weight and 50 ng/g wet weight in the whole body minus liver tissue 1. Average concentrations of PCBs in sculpin tissue 4.5 km from the beach of Saglek Bay substantially exceeded the established threshold 1. Based on the sediment–sculpin PCB relationship established for the exposure assessment (Fig. 3), the sediment PCB concentration associated with this sculpin exposure is approximately 5 ng/g dry weight 1.

Jorgensen 39 related EROD induction to PCB exposure in a northern pelagic fish population (Arctic char) and proposed a threshold level of between 100 and 1,000 ng/g wet weight in the carcass tissue. For shorthorn sculpin, EROD results were generally consistent with the lower end of this proposed threshold range 1. The threshold determined by Kuzyk et al. 1 was also in agreement with the lower end ranges reported for other fish species 40. Ethoxyresorufin-O-deethylase induction is one of the best known sublethal responses to dioxins, polyaromatic hydrocarbons, and PCBs and is used often as an early-warning indicator for levels that might have adverse effects on fish health 12. The biological significance of EROD induction in sculpin from Saglek Bay remains unknown; however the results suggest that sculpin collected within 4.5 km of the beach elicit a sublethal response to elevated PCB levels in the area.

Black guillemots

The results for the guillemot endpoints studied are summarized in Table 3. Results for liver biomarkers (EROD, liver retinol concentration, porphyrin concentrations, and malic enzyme activity) have been reported by Kuzyk et al. 11. Among the 13 endpoints, several showed no relationship to PCB exposure: malic enzyme; thyroid mass and activity; mass of the bursa, spleen, heart, and adrenal glands; chick growth; and hatching success (Table 3). Malic enzymes, the thyroid, the heart, and the adrenal are all direct or indirect physiological indicators of metabolic functioning, and growth is a functional endpoint strongly linked to metabolism. Therefore, the nonsignificant response for this set of endpoints is biologically consistent and provides quite convincing evidence that guillemot metabolism has not been significantly affected by PCB exposure at Saglek.

Table 3. Summary of biological effects data for the black guillemots exposed to polychlorinated biphenyls in Saglek Baya
  Polychlorinated biphenyl (PCB) relationshipEstimated PCB concentrations associated with risk
EndpointEffect thresholdSignificantSignificant but subthresholdSublethal effectsSurvival or reproduction
Nestling liver (ng/g wet wt)Sediment (ng/g dry wt)Nestling liver (ng/g wet wt)Sediment (ng/g dry wt)
  • a

    Nestling liver and associated sediment PCB concentrations associated with risk are provided for statistically significant endpoints.

1. MFO inductionSublethal effects threefold ×    
2. Adaptive organ enlargement (liver)Sublethal effects 25%× 75050  
3. Malic enzymeSublethal effects 30%      
4. Thyroid activitySublethal effects twofold increase or 50% increase      
5. Liver porphyrinSublethal effects twofoldIndeterminate     
6. Change in organ massSublethal effects      
 Thyroid −30%      
 Thymus −25%× 1,09087  
 Bursa −30%      
 Spleen −50%      
 Heart −50%      
 Adrenal −50%      
7. Retinoid levelsSublethal effects 40%× 1,05082  
8. Steroid hormones       
 a. Testosterone and estradiolSublethal effects 20%      
Survival/reproduction 40%×   1,00077
 b. CorticosteroneSublethal effects 40%      
9. Genetic variationN/A rejected due to data quality      
10. Gonad histologySurvival/reproduction; presence of any abnormalities that could compromise reproductive capacity×   ××
11. GrowthSublethal effects 20%      
 Survival/reproduction 30%      
12. Immune functionSublethal effects 25%× 27511  
 Survival/reproduction 50%×   2,750355
13. Hatching successSurvival/reproduction; statistically significant reduction of at least 10%      

In contrast, statistically significant effects related to nestling PCB exposure were found for liver porphyrin content, MFO induction, liver enlargement, thymus mass, retinoid levels, steroid hormone levels, and the PHA skin test immune response (Table 3). Liver porphyrin content showed unexpected decreases with increasing PCB concentrations, given the results of previous studies 41, so these results were considered to be indeterminate in terms of risk to guillemot health. The MFO induction results were significantly related (R2 = 0.26, p = 0.004, n = 31) to PCB exposure, but the magnitude (1.5-fold increase for the beach compared with reference) of the responses was not great enough to be considered indicative of risk (Table 3). However, for the rest of the endpoints (liver enlargement, decrease in thymus mass, retinoid levels, steroid hormone levels, and PHA skin test immune response), the results strongly indicated the presence of PCB-related adverse effects.

Among the male guillemots, qualitative analysis of the gonads revealed three significant testis abnormalities. Two males—one from the beach group and one from the island group—had disconnected medullar tissue that extended outside the cortex 5. Normal medullar tissue is located and contained within the cortex. Another male from the island group had seminiferous tubules extending through the cortex to the outside of the testis. If this tubule persisted during sexual maturity, growing sperm that traversed it would eventually be displaced to the outside of the testis within the abdominal cavity 5. No medullary abnormalities were observed in nestlings from the reference group. However, males of all groups showed a relatively high frequency of disrupted capsules (2 of 6 reference males, 4 of 7 island males, and 4 of 5 beach males, including the two males that had disconnected medullary tissue). Among the quantitative parameters examined, only the number of Sertoli cells varied significantly among the exposure groups. Major testicular abnormalities are extremely rare, and their presence in any member in a sample of wild, prefledgling birds is typically considered indicative of some developmentally disruptive factor, such as an estrogenic contaminant 42. This sample size, however, was very small and limited the statistical power of the analysis. Among females, histological analysis revealed neither significant structural abnormalities nor qualitative cellular changes related to PCB concentration. Relative ovarian weight results were inconclusive.

The results of the multiendpoint guillemot effects study were biologically consistent with each other and were also consistent with previous results indicating that the immune system and the developing reproductive system are the most sensitive systems in terms of contaminant-related effects. Liver enlargement, thymic atrophy, and retinoid depletion are all believed to share a common mode of toxicity (aryl hydrocarbon [Ah] receptor mediated) 43–45. Mixed function oxygenase induction also shares the Ah receptor-mediated mode of toxicity 43. Although MFO induction results were not high enough to be indicative of risk, they did show a significant relationship with PCB exposure. These results were also consistent with the observed effects on the PHA skin test immune response and the observed abnormalities in nestling gonads. The PHA skin test measures the T-cell–mediated portion of the immune response; the thymus is the site of T-cell maturation in the body. Therefore, the results show corresponding structural (thymus mass) and functional (immune response) changes in relation to PCB exposure. Contaminants that have caused atrophy and (or) histological changes in immune organs are usually associated with altered immunological function 46.

The observed depletion of retinoids is also consistent with this set of responses in that inadequate retinoid levels are related to immunosuppresion as well as overall susceptibility to disease 47. Retinoids are also associated with reproductive and developmental effects such as changes in secondary sexual characteristics, testes weight, spermatogenesis, and embryo survival 48. Elevated concentrations of estradiol and testosterone were observed in the guillemot nestlings at the beach area. These results are consistent with the presence of other reproductive effects and suggest that there is some potential for risk to guillemot reproduction. Abnormal concentrations of sex hormones are believed to be a causal factor related to developmental gonad abnormalities 42.

For each of the statistically significant endpoints, the relationship between the observed response and nestling PCB exposure was used to estimate the nestling liver PCB concentrations associated with risk to guillemot health. Nestling liver PCB concentrations were then translated into estimates of sediment PCB concentration using the exposure equation (Fig. 4) and are presented in Table 3. In terms of sublethal health risks, nestling liver PCB concentrations ranged from 750 to 1,090 ng/g wet weight, and corresponding sediment PCB concentrations were estimated at 50 to 87 ng/g dry weight. The PCB concentrations associated with sublethal immune effects were estimated at 275 ng/g wet weight and 11 ng/g dry weight for nestling livers and sediment, respectively. In terms of risk to guillemot reproduction or survival, nestling liver PCB concentrations were 1,000 and 2,750 ng/g wet weight for two endpoints, with corresponding sediment PCB concentrations at 77 and 355 ng/g dry weight. These values represent the best estimates of the sediment PCB concentrations associated with the effects observed in the present study; however, the values may either underestimate or overestimate risk.


Definition of the ecological risk zone

The results of this ERA indicated that both shorthorn sculpin and black guillemots are potentially at risk from the contaminated sediments in Saglek Bay as they existed during the assessment period of 1997 to 1999. Although thresholds for biological responses (e.g., EROD induction) occurred at PCB exposures as low as 50 ng/g wet weight in sculpins, the minimum PCB exposures associated with risk to sculpin and guillemot reproduction and survival appear to be in the order of 1,000 ng/g wet weight or more. Among sculpin, these levels of exposure (>1,000 ng/g wet wt in whole fish, excluding livers) occurred within 3 km of the contaminated beach area; however, they occurred frequently only in the immediate coastal area surrounding the contaminated beach, where the overall average sediment PCB concentration is ∼750 ng/g dry weight. Among the black guillemots, only nestlings from the beach group experienced PCB exposures at concentrations likely to be associated with risk to survival or reproduction (>1,000 ng/g wet wt in nestling livers). The average sediment PCB concentrations associated with these exposures are ∼77 to 355 ng/g dry weight. About 50% (n = 22) of the guillemot nests located within 2.1 km of the former source were surrounded by sediments with PCB concentrations averaging 77 ng/g dry weight or more, whereas about 20% of the nests were associated with sediment PCB concentrations averaging 355 ng/g dry weight or more. Because guillemot nest sites are limited in this area, every active nest that could be accessed within 2.1 km of the site was included in the study. Therefore, these 50 and 20% estimates likely reflect the actual risk for almost all black guillemot nestlings within 2.1 km of the former source.

A secondary objective of this ERA, provided that the contaminated sediments represented a risk as they existed during 1997 to 1999, was to provide information that would support the development of site-specific sediment-remediation objectives to ensure the protection of local wildlife. The results of this ERA suggest that average sediment PCB concentrations in the foraging areas of shorthorn sculpin would have to be reduced to less than 750 ng/g dry weight to protect the local sculpin population. For black guillemots, the ERA identified a lower sediment threshold of 77 to 355 ng/g dry weight. The conservative end of this range, 77 ng/g dry weight, was therefore the best estimate of a PCB concentration in sediments that would be protective of the black guillemots. Because black guillemots occupy a relatively high trophic level, have a relatively small home range, and forage locally, the derived site-specific level (77 ng/g dry wt) should be protective of most bird and fish species in the area. The lower sediment threshold values observed in our study, compared with an ERA conducted in Lower Fox River and Green Bay, Wisconsin 49, likely are due to the degree of variability among different fish and bird species in response to PCBs, as well as the more comprehensive site-specific data that were acquired in our study. Such differences further support the utility of using multiple lines of evidence, including site-specific effects data, to assess PCB-related risks to marine biota.

As with shorthorn sculpin, black guillemots tend to be associated with shallow water, where benthic prey species are relatively abundant. They forage most commonly in waters less than 18 m deep and very rarely in water more than 40 m deep 23. Ecologically, these conditions occur throughout the shallow subtidal portions of Saglek Anchorage, around the former contaminated beach and to the west and east for a distance of 1 to 1.5 km. Much of this area had sediment PCB concentrations greater than 77 ng/g dry weight during 1997 to 1999. The specific area represented by these two parameters—PCB concentrations greater than 77 ng/g dry weight and water depths less than 40 m—is illustrated in Figure 6. It is within this area that decreases in sediment PCB concentrations (e.g., as a result of sediment remediation) would directly reduce the risks to guillemots. Because black guillemots were the most sensitive receptor examined in the ERA, this area also represents the best estimate of an ecological risk zone in Saglek Bay, wherein sediment contamination represented a significant ecological risk during the study period (1997–1999).

Figure 6.

Area defined by sediment polychlorinated biphenyl (PCB) concentrations greater than 77 ng/g dry weight and water depths less than 40 m, which defined the ecological risk zone related to the contaminated sediments in Saglek Bay, Labrador, Canada, during the assessment period (1997-1999).


The present study along with three parallel studies addressing the transport and fate of the contaminated sediments 3, 4, existing sediment remediation technologies 6, and the potential for human health risks provided information to support the Canadian Department of National Defence in making a decision on the management of the contaminated sediment in Saglek Bay. The final management decision, after consultation with a broad stakeholder group including government regulators and the Labrador Inuit Association (now the Nunatsiavut Government), was to allow natural recovery to occur while conducting monitoring to ensure that this process occurred as expected 3, 4. A followup study 8 demonstrated that natural ecosystem recovery was indeed occurring and that, as of 2006, the PCB concentrations in the sediments near the former source were rapidly approaching the site-specific threshold for ecological risk (77 ng/g dry wt). Companion studies also have shown that the declines in PCB concentration in the sediments are associated with decreases in biological effects (manuscripts in preparation).


Table S1. (66 KB DOC)


The present study was funded primarily by the North Warning System and Director General Environment, Canadian Department of National Defence, and through a grant from one of the Networks of Centres of Excellence, ArcticNet, to K.J. Reimer. This study represents the collective effort of members of the Department of National Defence's North Warning Systems Office, the Director General Environment, the Nunatsiavut Government (formerly the Labrador Inuit Association), Environment Canada, Fisheries and Oceans Canada (St. John's), and the Newfoundland and Labrador Department of Environment and Labour. We thank G. Worthman, K. Doe, K.-L. Tay, P. Stewart, P. Hodson, D. Jeffrey, G. Fox, A. Lorenzen, S. Trudeau, K. Grasman, E. Kelly, J. Wingfiled, A. McNabb, and P.-Y. Daoust for their contribution to the assessment of risks to the benthic invertebrates, shorthorn sculpin, and black guillemots. We also thank C. Webb, J. Webb, the crew of the Viola Dee, Nanuk Diving, and ESG staff for assistance in the field. We are extremely grateful for the support and advice that we have received from J. Rowell.