Relationships between ecological variables and four organochlorine pollutants in an artic glaucous gull (Larus hyperboreus) population

Environmental Pollution 136 (2005) 175e185 www.elsevier.com/locate/envpol

Relationships between ecological variables and four organochlorine pollutants in an artic glaucous gull (Larus hyperboreus) population Jan Ove Bustnesa,*, Øystein Milanda, Magnus Fjelda, Kjell Einar Erikstada, Janneche Utne Skaareb a

Norwegian Institute for Nature Research, Division for Arctic Ecology, The Polar Environmental Centre, N-9296 Tromsø, Norway b National Veterinary Institute, P.O. Box 8156 Dep., N-0033 Oslo, Norway Received 28 May 2004; accepted 17 September 2004

In arctic glaucous gulls, nesting behaviour, early chick growth and adult return rate were negatively related to blood concentration of organochlorines. Abstract The Arctic has become a sink for organochlorine contaminants (OCs) from lower latitudes, and relatively high levels have been found in different biota. Recent studies of the glaucous gull, Larus hyperboreus, a top predator in the arctic food web, have documented that high blood residues of various OCs are related to lower reproductive performance and reduced adult survival. Here we provide additional evidence that OCs are having ecological effects in the glaucous gull population at Bear Island in the Norwegian Arctic, and compare the effects of the four major OCs found in the glaucous gulls: HCB, oxychlordane, DDE and PCBs, which made up O95% of measured OCs. Firstly; it has previously been shown that gulls with high levels of PCBs in their blood spent more time away from the nest site during incubation than gulls with low levels. Here we reanalyzed the data and found that PCBs (P!0.02) and oxychlordane (P!0.05) were positive and significantly related to time away from the nest site, while DDE and HCB were not related to this trait. Secondly, among females which bred in an area where fish dominated the diet, and thus had high flight costs during feeding, early chick growth was negatively related to maternal levels of all four OCs, especially HCB and DDE (P!0.01). On the contrary, among females breeding in an area where the diet was dominated by eggs and young from nearby seabird colonies, and thus feeding costs were low, there were no effects of OC levels on early chick growth. This indicates that additional stress may be fundamental in causing reproductive effects of OCs in this population. Finally, during three breeding seasons we examined the probability of adults returning to the breeding grounds in the subsequent season, as a function of blood concentration of the four OCs. Overall, return rate from one year to the next was negatively related to blood residues of oxychlordane (PZ0.02), but not significantly related to the other three compounds. Further support for the importance of oxychlordane was that a 60% drop in the blood levels between 1997 and 2000 led to a significant increase in return rate between these two years. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Arctic; POPs; Organochlorines; PCB; Nesting behavior; Chick growth; Adult return rate

1. Introduction * Corresponding author. Tel.: C47 77 75 04 07; fax: C47 77 75 04 01. E-mail address: [email protected] (J.O. Bustnes). 0269-7491/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.envpol.2004.09.026

Organochlorine contaminants (OCs) consist of a large number of manufactured chemicals of which many, e.g., PCBs, DDT, and HCB, have become global

176

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185

contaminants. Due to atmospheric processes, such as the cold-condensation effect (Mackay and Wania, 1995), various organochlorines have reached the Arctic, often occurring in relatively high levels. Many OCs are known to have serious detrimental effects on humans and wildlife, including endocrine disruption, impairment of enzyme activity, and reduced immune function (Vos and Luster, 1989; Colborn et al., 1993; Kelce et al., 1995; Dewailly et al., 2000). Since the early 1970s, relatively high levels of various OCs have been found in glaucous gulls, Larus hyperboreus, at Bear Island in the European Arctic (Bogan and Bourne, 1972; Gabrielsen et al., 1995), but the residues recorded in dead and dying birds have been below concentrations known to be lethal (Gabrielsen et al., 1995). Recently, however, high blood residues of different organochlorines have been linked to impaired behavior, reduced reproductive performance, poor survival, and other adverse effects (see Sagerup et al., 2000; Bustnes et al., 2001a, 2002, 2003a, 2004; Verrehault et al., 2004). In this paper we present additional evidence that birds with high concentrations of OCs in this population are suffering reduced fitness, by presenting previously unpublished data collected between 1997 and 2002. It has been difficult to decisively attribute lowered fitness, i.e., impaired reproduction and lowered survival in wild birds and mammals, to pollutants (see Hose and Guillette, 1995; Bustnes et al., 2003a). There are several reasons for this. Firstly, experimental manipulations of OC burdens, involving poisoning of individuals, is very difficult to carry out and considered unethical in wild populations. This means that much of the documentation of ecological effects must rely on observational data, with potential confounding factors influencing the results. One way to mitigate this problem is to control for such confounding factors by including them into statistical models. Secondly, few studies have measured individual OC burdens non-destructively, and then related individual fitness components to such OC levels. In most studies, it has thus been impossible to follow individuals over extended time periods, which is a great advantage in ecological studies where reproductive traits and survival are measured. In this study we used blood samples to measure OCs and were able to relate individual fitness components to OC levels. Individual OC measurements are also a prerequisite for proper statistical control for possible confounding variables, such as annual environmental variation, body condition, and different breeding locations. A third problem is that OCs accumulate simultaneously in biota, often resulting in very high correlations between different compounds; i.e., individuals having high levels of polychlorinated biphenyls (PCBs) also have high levels of other persistent compounds such as p,p#-dichlorodiphenyldichloroethylene (DDE) (Mora

et al., 1993; Jones and Voogt, 1999; Bustnes et al., 2001b). As a consequence it is very difficult to extract the effects of particular chemicals in such mixtures in observational field studies (Jones and Voogt, 1999; Bustnes et al., 2003a). However, correlations between different OCs vary, and by correlating various fitness components of individuals to the concentrations of different compounds some OCs may be more consistently strong predictors of reduced fitness. In glaucous gulls at Bear Island, hexachlorobenzene (HCB, 3e4%), oxychlordane (3e4%), DDE (15e20%) and PCBs (73e74%) are the most common organochlorines, making up more than 95% of the measured OCs in blood (Bustnes et al., 2003a). In these analyses we therefore focus on these four components of the OC burdens, and aim at comparing the strength of the effects they are having on various fitness components. We analyze the following fitness components in relation to blood residues of the four OCs: (1) the time that breeding individuals spent away from the nest when not incubating (mostly feeding time) in 1998; (2) early chick growth in 1997, 1998 and 2000; and (3) adult return rate from one year to the next, for 1997, 2000 and 2001. In this paper, we also reanalyzed data that we had reported in another paper (Bustnes et al., 2001a). Our reanalysis enabled us to compare the relative strength of the relationships between concentrations of the four groups of OCs and time spent away from the nest. We did not attempt to make such comparisons in our first paper. We have previously also shown that the adult survival rate from 1997 to 1998 was significantly reduced in birds with high blood concentrations of OCs, especially oxychlordane, by observing these birds annually between 1997 and 2000 (Bustnes et al., 2003a). In this study we have added data on birds caught in 2000 and 2001 and observed until 2002 whether OC levels in their blood were correlated to the probability of returning from one year to the next.

2. Materials and methods 2.1. Ecological data Reproduction of glaucous gulls was studied at Bear Island (74  30#N, 19  01#E) from 1997 to 2001. Individuals were subsequently followed until 2002 for analyses of adult return rate. The study area is described in detail in Bustnes et al. (2000, 2001a). Glaucous gulls were studied in two different breeding areas: seabird cliffs and areas near the sea level, and we have previously documented large differences in feeding ecology between the two types of areas. Gulls at seabird cliffs fed mostly on other seabirds (eggs, young and adults) in the seabird colonies close to their nests and thus had relatively low feeding costs in terms of energy expenditure.

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185

The average feeding trip of females on seabird cliffs lasted only 1.6 h (2 h in males), while in sea level areas adjacent to the seabird cliffs the birds had high feeding costs because they fed much on fish at sea, and the average feeding trips of females lasted 6.4 h (3.8 h in males) (Bustnes et al., 2000, 2001a). Birds were caught on their nests using a nest trap (Bustnes et al., 2001a) and individually marked with PVC leg bands and numbered steel bands. Blood (10 ml) was sampled from the wing vein with a syringe. We measured bill length, bill height, skull length (head and bill) and wing length (all morphological values measuredG0.5 mm). Body condition was indexed using body mass (to nearest 10 g), controlling for body size (head and bill) (Garcı´ a-Berthou, 2001). Sex was determined by size, males being larger than females (Cramp and Simmons, 1983). As there was no overlap between males and females in either bill length or skull length (bill and head) on Bear Island, we assumed that birds with a bill longer than 61.5 mm and skull longer than 142 mm were males (Bustnes et al., 2001a). Previously, several behavioral traits of incubating glaucous gulls were analyzed in relation to PCBs, only (Bustnes et al., 2001a). Here we reanalyzed and compared the effects of different OCs on the time spent away from the nest when not incubating (assumed to be mostly feeding time), because this behavior was found to be strongly negatively related to PCBs (Bustnes et al., 2001a). All details about the incubation behavior and nest attendance are found in Bustnes et al. (2001a). The behavioral observations were carried out on 27 gulls (11 males and 16 females) in the two breeding areas, between 31 May and 5 June 1998. Glaucous gull usually incubate for longer continuous periods of several hours. The non-incubating partner, when not on feeding trips, sits close to the nest responding aggressively to external threats. Records of whether the marked birds were incubating, attending the nest-site or absent from the nest-site (nest-site attentiveness) were made every hour, for 48 h. The area surrounding the nest was surveyed if the marked birds were not observed by the nest. This produced a data set where non-incubating individuals were present at the nest (1) or absent (0), which was used to analyze the proportion of time spent away from the nest in relation to blood concentrations of the four OCs. We measured chick growth in three years (1997, 1998, and 2000). Nests were followed from the start of egg laying and eggs were marked and checked every third day until predation occurred or they hatched. In the hatching period, nests were checked daily and chicks were weighed to the nearest gram and tagged with a numbered web tag. The whole brood was then weighed between 10 and 14 days after the first chick hatched. The mean of the daily mass increase among chicks in a brood (growth of one chick, or the mean of 2-3 chicks) was used as a measure of daily chick growth.

177

We expanded the data to include birds caught in 2000 and 2001 and tested if blood levels of OCs were negatively related to the probability of returning over several years. Because the project was terminated in 2002 we could not follow the birds for sufficient time to estimate survival for each of the three years. The breeding areas were thoroughly searched for returning birds in all breeding seasons from 1998 to 2002. Glaucous gulls usually occupy the same, or nearby, nest sites for several years. Thus, we feel confident that we were able to record a very high percentage of returning individuals.

2.2. Chemical analysis The OC analyses were carried out at the Environmental Toxicology Laboratory at the Norwegian School of Veterinary Science/National Veterinary Institute. Samples of whole blood (ca. 8 g) were weighed, and an internal standard (CB-29 and CB-112) was added. The methods used for extraction (cyclohexane and acetone), clean-up (with sulfuric acid) and quantification (gas chromatography) of the samples were first described in Brevik (1978), and modifications are described in Andersen et al. (2001). Percent extractable fat was determined gravimetrically. Aliquots of the final extracts were injected on an Agilent gas chromatograph/electron capture detector (GC-ECD) (6890 Series, Agilent Technologies, USA). The GC was equipped with two capillary columns (SPB-5 and SPB-1701, Supelco, Inc., Bellefonte, PA). Quantification was done within the linear range of the detector. Detection limits were defined as three times the background noise. The following PCB-congeners were determined for this study (IUPAC nos. 99, 118, 138, 153, 170, and 180). Other compounds analyzed included HCB (hexachlorobenzene), oxychlordane, and DDE (p,p#-dichlorodiphenyldichloroethylene). The detection limits for the individual PCBs were from 0.01 to 0.02 ng/g wet weight, for HCB from 0.005 to 0.01 and for oxychlordane and DDE from 0.01 to 0.02 ng/g. GC conditions, temperature program and quality assurance procedures are described in Andersen et al. (2001). Analytical standards of the laboratory are certified by participation in international intercalibration tests. Certified international reference materials (CRM 349 and 350, ICES cod liver oil and mackerel oil) are analyzed regularly with results within the given ranges. The laboratory is accredited for these analyses according to the requirements of NS-EN 45001 (Norwegian and European standard) and ISO/IEC Guide 25. The applicability of whole samples has previously been documented as dosimetric for glaucous gull at Bear Island (Henriksen et al., 1998; Bustnes et al., 2001b).

178

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185

2.3. Statistical analyses We conducted separate analyses for the relationship between all fitness components and the four OCs, to test if some compounds were more consistently related to fitness effects than the others. For PCBs, all the measured congeners (CB: 99, 118, 138, 153, 170 and 180) are persistent in food chains (Norstrom, 1988; Boon et al., 1997), and were highly correlated to the sum of these congeners (all R2, values between 0.96 and 0.99). Consequently, no relationship between fitness components and these congeners, not demonstrated for the sum of persistent PCBs, could be found. We therefore only report analyses carried out using the sum of these PCB congeners. Statistical analyses were carried out using SAS, version 8e (SAS, 1999). OC values were Log10 transformed to approximate a normal distribution. Standard error (SE) is given for all means and estimates. Behavior was studied in the same year as blood samples were taken. When comparing the effects of the different OCs on time away from the nest between incubation bouts we only controlled statistically for the two factors that strongly influenced the feeding time of incubating gulls: sex and breeding area, even if other factors also had minor effects in models testing the effects of contaminant levels (see Bustnes et al., 2001a). This was done to make the comparison between the compounds more easily interpretable. Behavior was analyzed using logistic regression in the GENMOD PROCEDURE in SAS (SAS, 1993). Since there was some over-dispersion in the data we used Quasi-Likelihood modeling (McCullagh and Nelder, 1989). The early chick growth and the parent blood levels of OCs were measured in the same year. Early chick growth has not been analyzed in this population before and we controlled for a set of variables and interactions, including; year, breeding area and body condition using the GENMOD PROCEDURE, Type 1 and 3 sum of squares (SAS, 1993). We used Type 1 when interaction terms were included in the models and otherwise Type 3. The difference in feeding ecology made us include breeding area into all statistical models, to control for the effects of this variable on the fitness components. Blood levels of OCs were measured in one breeding season and return rate was evaluated in the year following. To test if some compounds were more related than the others to the probability of returning from one year to the next, we used logistic regression in the GENMOD PROCEDURE (LR test, Type 1 statistics) and controlled for year and sex in the models for the different OCs. We also tested if breeding area and body condition affected the return rate, but these variables were removed if not significant (P!0.05). Each individual appeared only once in the return rate dataset, i.e., if a bird was captured, and blood sampled in more

than one year, only the first blood level and return observation was used to avoid pseudo-replication.

3. Results 3.1. Levels of organochlorines During the four years of this study, OCs were measured in a total of 331 blood samples (Table 1). Mean HCB levels varied between 12.5 and 20 ng/g (wet weight); oxychlordane between 10 and 23.5 ng/g; DDE between 58 and 98 ng/g; and PCBs between 326 and 431 ng/g (Table 1). 3.2. Time absent between incubation bouts in relation to blood concentration of OCs All descriptive data on the incubating behavior of the glaucous gulls studied here can be found in Bustnes et al. (2001a). When controlling for sex and breeding area, PCBs was the strongest predictor of the time away from the nest site when not incubating (PZ0.0193) followed by oxychlordane (PZ0.0415). The effects of HCB and DDE were not significant (Table 2). 3.3. Early chick growth The number of chicks in a brood had no effect on the mean chick growth, either in females (PZ0.98) or in males (PZ0.33), and this variable was excluded from further analysis as it did not change the effects of the other predictor variables in the statistical models. Two chicks with daily growth lower than 10 g were excluded as probable outliers, since the mean daily chick growth was 36.1 g (range 19.2e58.6 g). The relationships between OCs and chick growth were different in the two breeding areas; i.e., there were statistically significant interactions between OC levels in blood of females for all four OCs and breeding area (P values between 0.0069 and 0.0661), but not in males (P values between 0.27 and 0.51). The two areas were thus analyzed separately for females, but not for males. There was a positive relationship between chick growth and body condition in the breeding area close to the sea Table 1 Concentration (mean and SE) of different organochlorine compounds (ng/g, wet weight) in the blood of glaucous gulls from Bear Island, Barents Sea, in different years Year N

HCB

Oxychlordane DDE

1997 111 16.85 (1.48) 23.55 1998 33 17.52 (2.39) 14.73 2000 86 12.50 (0.82) 9.99 2001 101 20.11 (1.14) 13.42

(2.44) (1.73) (0.75) (0.86)

84.64 85.64 58.09 97.73

PCBs (6.70) (10.57) (4.15) (7.31)

431.13 369.49 325.87 376.55

(46.23) (49.71) (29.08) (32.54)

179

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185

Effect

DF

c2

P value

Estimate

SE

Sex Breeding area

1, 23 1, 23

7.61 4.71

0.0058 0.0299

ÿ1.35 ÿ1.31

0.53a 0.80

HCB Oxychlordane DDE PCBs

1, 1, 1, 1,

0.17 4.16 0.56 5.48

0.68 0.0415 0.45 0.0193

ÿ2.20

1.16

ÿ2.27

1.04

23 23 23 23

Logistic regression, GENMOD, quasi-likelihood estimation (Dscale option), Type 1 statistics (SAS, 1993). a Males vs. females.

level (Table 3, Fig. 1), while no such relationship was found in females at the seabird cliffs (PZ0.47, NZ15). After controlling for body condition, there were negative relationships between chick growth and blood concentration of HCB (PZ0.0057), oxychlordane (PZ0.03), DDE (PZ0.0091) and PCBs (PZ0.025) in the sea level breeding area (Table 3, Fig. 2). In the area with low feeding costs there were no significant relationships (P values between 0.18 and 0.44, Fig. 2). Among males there were no negative effects of OCs on chick growth (P values between 0.68 and 0.93). 3.4. Adult return rate Of 262 different birds marked in 1997, 2000 and 2001, and included in the return rate analyses, 196 (74.8%) Table 3 The effects of different organochlorine compounds on the early chick growth in female glaucous gulls with high feeding costs, controlled for year (1997, 1998 and 2000; estimates not given for year) and body condition (body mass controlled for body size: head and bill) Effect

DF

F

P value

Estimate

SE

Year Body size Body mass HCB

2, 1, 1, 1,

30 30 30 30

7.61 0.09 9.15 8.88

0.0058 0.77 0.0051 0.0057

0.000332 ÿ0.1103

0.00011 0.0370

Year Body size Body mass Oxychlordane

2, 1, 1, 1,

30 30 30 30

4.27 0.0002 7.48 5.19

0.0233 0.99 0.0104 0.03

0.000314 ÿ0.0740

0.00012 0.0325

Year Body size Body mass DDE

2, 1, 1, 1,

30 30 30 30

5.01 0.003 8.65 7.79

0.0133 0.99 0.0063 0.0091

0.000327 ÿ0.0991

0.00011 0.0355

Year Body size Body mass PCB

2, 1, 1, 1,

30 30 30 30

4.10 0.0007 7.76 5.56

0.0267 0.98 0.0092 0.0251

0.000319 ÿ0.0751

0.00012 0.0319

SAS: PROC GENMOD, Type 3 sum of squares.

2.0

Daily chick growth (Log10)

Table 2 Relationships between time away from the nest site when not incubating and blood concentration of four different organochlorines in glaucous gull, after controlling for sex and breeding area

1.8

1.6

1.4

1.2

1.0 1200

1300

1400

1500

1600

1700

Female body mass (g) Fig. 1. The mean daily chick growth in broods of glaucous gull females with high feeding costs in relation to female body mass. Data from Bear Island, 1997, 1998, and 2000.

were observed the following seasons. Males had a higher return rate than females (80.4% vs. 68.5%), with considerable variation between the years (Table 4). We consequently controlled for year and sex in the models for all four OCs (Table 5). Overall, only oxychlordane was significantly related to return rate (PZ0.02, Table 5, Fig. 3), while the other compounds were not significant (P values between 0.09 and 0.20, Table 4). Breeding area (PZ0.98) and body condition (PZ0.64) had no significant effects on returning probabilities when controlling for year and sex (LR test, Type 1 statistics). To further test the effect of OCs on return rate we investigated whether the four OCs were differently related to return rate in the 3 years (statistical interaction between OC levels and year). There were indications that the relationship differed between years (P values between 0.0554 and 0.0911), and we further tested if some years were different from other years. There was a significantly different relationship between blood concentrations of oxychlordane and return rate between 1997 and 2000 (PZ0.0453, Table 5); i.e., oxychlordane was negatively related to return rate in 1997, but not in 2000. The same relationship was near significant for PCBs (PZ0.0589), while for HCB (PZ0.12) and DDE (PZ0.17) it was not significant. No relationships between 2000 and 2001 were significant (P values between 0.58 and 0.78). Between 1997 and 2001 there were significant differences in the relationship between blood concentrations of HCB (PZ0.0332) and DDE (PZ0.0364) on the returning probability (Table 5), but this was not found for oxychlordane (PZ0.20) and PCBs (PZ0.0976). Coinciding with the changes in the relationship between oxychlordane and return rate between 1997 and 2000, was a 58% drop in the mean blood concentration of oxychlordane, and it remained low also in 2001 (43% lower than 1997 level; Fig. 4). In comparison,

180

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185

High feeding costs

Low feeding costs

1.9

HCB

1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.9 1.8

Oxychlordane

1.7

Daily chick growth (Log10, g)

1.6 1.5 1.4 1.3 1.2 1.9

DDE

1.8 1.7 1.6 1.5 1.4 1.3 1.2 1.9

PCB

1.8 1.7 1.6 1.5 1.4 1.3 1.2 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

0.0

0.5

1.0

1.5

2.0

2.5

3.0

Blood level of organochlorines (Log10) Fig. 2. The mean daily chick growth in broods of glaucous gull females with high feeding costs and low feeding costs, in relation to blood concentration of HCB, oxychlordane, DDE and PCB (ng/g, wet weight). In the figure, chick growth is controlled for year and female body condition. Data from 1997, 1998, and 2000.

the drop between 1997 and 2000 in the other compounds were much smaller (24e31%), and in 2001 HCB and DDE were higher than in 1997 (15e20%), while PCBs was only 13% lower than in 1997 (Fig. 4). Thus only the variation in oxychlordane levels shows an annual pattern that coincides with the changes seen for the return rate. This study thus lends support to the hypothesis that oxychlordane is a major agent in reducing the return rate of glaucous gulls at Bear Island.

4. Discussion The results from this study corroborate earlier findings from the glaucous gull population at Bear Island (Sagerup et al., 2000; Bustnes et al., 2001a, 2002, 2003a, 2004; Verreault et al., 2004), and the four measured OCs were negatively related to one or more of the fitness components studied. However, a major problem in ecotoxicology is that free-living animals are exposed to mixtures of OCs of which the different

181

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185 Table 4 Return rate of adult male and female glaucous gulls at Bear Island, 1997e2002 Females N (%)

49 37 40 e

60 43 33 e

0.6

(76.7) (65.1) (63.6) (68.5)

0.4

N. total number marked each year; (%), % returning the subsequent season.

components are highly correlated. Hence elucidating which OCs are actually causing the effects is very difficult (Jones and Voogt, 1999; Bustnes et al., 2003a). It is therefore important to be cautious when interpreting results obtained from different statistical tests run on the same samples since a significant result of one OC may only be a result of high correlations with the OCs that are actually causing effects. However, if some compounds keep coming out as the strongest predictors of different effects, it may suggest that they are more important in causing adverse effects than others. In this study, we were, however, unable to show that any of the compounds in general were of greater importance than others since the three effect parameters examined were best predicted by three different compounds (time away from the nest site by PCBs; lowered early chick growth

Table 5 The probability that glaucous gulls returned to the breeding grounds in the preceding breeding season in relation to year, sex (controlled for year) and blood concentration of each of four different OCs (controlled for year and sex); then the interaction terms that showed significant differences between individual years; data from 1997 to 2002 Effect

DF

c2

P value

Estimate

SE

Year Sex

2, 259 1, 258

1.51 4.18

0.47 0.0409

ÿ0.78

0.31

Oxychlordane HCB DDE PCB

1, 1, 1, 1,

5.36 1.89 0.49 2.95

0.0206 0.17 0.49 0.09

ÿ0.85

0.32

Significant interactions Year!Oxychlordane 1997 2001 2000 Year!HCB 1997 2000 2001 Year!DDE 1997 2000 2001

257 257 257 257

1, 255 1, 255 0

4.01 0.31 0

0.0453 0.58

1, 255 1, 255 0

4.52 0.22 0

0.0364 0.64

1, 255 1, 255 0

4.38 0.23 0

0.0362 0.63

ÿ2.33

ÿ2.70

ÿ2.43

1.17

1.27

1.16

Estimates are given for significant variables only. GENMOD PROCEDURE, logistic regression, LR test Type 1 statistics.

Predicted return rate

(79.6) (86.5) (75.0) (80.4)

0.8

0.2

Males

0.0 1.0 0.8 0.6 0.4 0.2

Females

0.0 0.0

0.5

1.0

1.5

2.0

Blood concentration of Oxychlordane (Log10) Fig. 3. Adult return rate of glaucous gulls at Bear Island (statistics in Table 5). Graphs present predicted return rate for males and females from 1997 to 1998 (NZ109), 2000 to 2001 (NZ80) and 2001 to 2002 (NZ73) in relation to blood concentration of oxychlordane (ng/g, wet weight). Graphs have not been controlled for year.

by HCB; adult return rate by oxychlordane). Previous studies of the four compounds, both laboratory and field studies, have documented that they all may be involved in biochemical, neurotoxic, immunological, physiological and ecological effects (reviewed by Hoffman et al., 1996; Blus, 1996; Wiemeyer, 1996).

Mean percent change from the 1997 level

1997e1998 2000e2001 2001e2002 Mean

Males N (%)

1.0

60 HCB Oxychlordane DDE PCB

40 20 0 -20 -40 -60 -80

1997

1998

1999

2000

2001

Year Fig. 4. The percentage change in mean blood concentrations of four different organochlorines (ng/g, wet weight) in 1998 (NZ33), 2000 (NZ86) and 2001 (NZ101), compared to the mean levels in 1997 (NZ111).

182

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185

Effects may also be caused by several compounds working synergistically or additively (e.g., Bocquene et al., 1995; Payne et al., 2001; Walker et al., 2001). The levels of OCs found in glaucous gull at Bear Island are high by arctic standards, but not particularly high if compared to equivalent species in industrialized areas (e.g., Herbert et al., 1999). However, Brunstro¨m and Halldin (2000) pointed out that the range in estimated levels of PCBs in glaucous gull eggs from the Svalbard region, including Bear Island, were well within the concentration expected to have adverse effects on chick development. We are aware of no other studies that have measured individual OCs levels non-destructively in blood of adult birds, and subsequently recorded individual fitness components. Thus, comparison to most other studies of ecological effects of OCs is difficult because they so frequently have depended on comparisons between locations. That is, ecological variables have been measured in birds in contaminated areas and then compared to ‘‘uncontaminated’’ control areas, usually without knowing individual contaminant levels (e.g., McCarty and Secord, 1999; Gill et al., 2003; Janssens et al., 2003). Such comparisons between areas is potentially misleading, and a much more powerful way to study effects of OCs is to measure individual levels and then record the fitness components in question (see Bustnes et al., 2001a). In herring gulls, Larus argentatus, from the Great Lakes in North America, where endocrinal, immunological and also ecological effects have been demonstrated (Herbert et al., 1999; Grasman et al., 2000), pooled hepatic levels of PCBs ranged between 1.8 and 23.8 ppm and DDE between 0.6 and 7.4 ppm, between colonies in the early 1990s (Grasman et al., 2000). In comparison, the mean hepatic levels of DDE and PCB in glaucous gulls at Bear Island were 1.44 and 4.41 ppm, respectively (Henriksen et al., 2000); the highest PCBs values recorded exceeded 17 ppm (w.w.), and DDE exceeded 4 ppm. 4.1. Time absent from the nest site during incubation Experimental studies of birds exposed to various OCs have documented changes in operant behaviors and ability to conduct complex behavioral patterns (Hoffman et al., 1996; Burger et al., 2001; Walker, 2003), but under field conditions there is little direct evidence for such adverse effects on parental behavior (Peakall, 1996; Burger et al., 2001). However, the nesting period is the most critical phase in avian reproduction, and glaucous gulls usually suffer high rates of nest predation (Bustnes et al., 2001a, 2003a). Even small OC mediated changes in the ability to protect the nest may thus lead to massive reproductive failures in glaucous gull populations. Recent findings also indicate that nesting success (the probability of hatching young in a nest) is negatively

related to blood residues of OCs in glaucous gull males (J. O. Bustnes et al., unpublished data). Fox et al. (1978) found a positive relationship between total OC content in eggs and total time that a nest was not incubated in herring gulls. However, at the time when Fox et al. (1978) did their study, the OC levels (e.g., PCBs and DDE) in the Great Lakes were extremely high (Gilman et al., 1977), which may explain aberrant behaviors. In comparison, glaucous gull eggs were always incubated by one of the parents, and the potential effects of contaminants may be more subtle than those found in the Great Lakes and may only be unraveled by detailed observations of nesting birds for which OC levels are known. Even if HCB and DDE were not related to time away from the nest, we cannot exclude that they are important in creating a stress that is fundamental for behavioral effects to appear. Moreover, PCB was the strongest predictor for time away from the nest site, but oxychlordane may still be important in creating such effects since the mechanism through which these toxins influence behavioral patterns of glaucous gulls is not known. Adverse effects of OCs may be due to neurological effects of intoxication (Walker, 2003) or through endocrine disruption (Burger et al., 2001). Several compounds are known to have endocrinal effects (McKinney et al., 1985; Petersen et al., 1993). For example, PCBs within the range of concentrations found in glaucous gulls are known to impair the corticosterone response, which means that the birds may be less capable of responding to environmental stress (reviewed by Love et al., 2003). Moreover, PCBs are known to disrupt thyroid and steroid hormone action (e.g., Colborn et al., 1993) and possibly such disruption may influence breeding motivation in the birds. 4.2. Early chick growth Among females, the relationship between early chick growth and OCs was different in the two breeding areas, suggesting that factors additional to OCs were influencing their ability to raise chicks. We have previously documented large differences in feeding ecology between the two adjacent breeding areas; i.e., gulls at seabird cliffs fed mostly in the seabird colonies close to the nests and had low feeding costs while birds in the sea level areas had long flights at open sea (Bustnes et al., 2000, 2001a). Moreover, the positive relationship between chick growth and female body condition in the sea level breeding area also indicates high feeding costs. This may suggest that there are interactions between the energy expenditure and different OCs, and females with high OC levels may have fewer resources available to provide for their chicks. However, the evidence that coping with the toxic effects of pollutants has high metabolic costs,

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185

e.g., enzyme production, is poor. Calow (1991) suggested that there were considerable metabolic costs related to a set of pathways through which the damage from toxins could be eliminated. However, Calow (1991) discussed mostly metal exposure in aquatic or soil invertebrates, fish and plants and not OCs in birds. Walker et al. (2001) acknowledged that there may be energetic costs associated with enzyme induction, the main detoxification pathway of PCBs and some other OCs in biota, but concluded that such energetic costs were small compared to the total energetic stress encountered by organisms in natural environments. In birds very few studies of the interactive effects of pollutants and natural stress have been carried out. A central study of captive ringed turtle doves, Streptopelia risoria, however, found that the negative effects of food deficiency on reproductive performance were much stronger when the birds were exposed to DDE (Keith and Mitchell, 1993). However, DDE levels were much higher than those documented in glaucous gulls from Bear Island (Henriksen et al., 2000). Moreover, Rattner and Franson (1984) found that cold intensified the toxicity of methyl parathion in American kestrel, Falco sparverious, suggesting some trade-off between combating toxins and cold tolerance in birds. Despite lower levels of contaminants in nature compared to doses used in these experiments, the effects may not be directly comparable since lower levels may have comparably stronger effects in the wild compared to laboratory tests (reviewed by Grue et al., 2002). All four OCs were negatively associated with chick growth and it is thus difficult to draw conclusions about which compounds that were most important in causing such effects. However, we note that HCB and DDE were somewhat more related to the effects than oxychlordane and PCBs. In summary we believe that it is premature to conclude both about the reason why the effect of OCs on chick growth was only found in females with high energetic costs, and about which compounds are most the important in causing such effects. 4.3. Adult return rate In glaucous gulls, adult return rate between years is a strong predictor of adult survival, even if some birds survive, but fail to show up in the colony in subsequent seasons (e.g., of 109 birds marked in 1997, 85 birds were found in 1998; only 5 new birds were found in 1999, and two more in 2000; J.O. Bustnes, unpublished data). In long-lived birds adult survival probability is the parameter to which population growth rate is most sensitive (Lebreton and Clobert, 1991; Wooller et al., 1992). In our previous analysis of the survival rate between 1997 and 1998 (survival estimates were based on surveys of the study area for birds marked in 1997 in

183

all years from 1998 to 2000) we found that PCB, DDE and HCB concentrations were related to declining survival probabilities, but oxychlordane clearly had a stronger effect (Bustnes et al., 2003a). In this extended analysis of adult return rate between subsequent years, only oxychlordane remained significant when combining all three years, which strengthens the conclusion that this compound is essential in causing failure to return. Moreover, the dramatic drop in oxychlordane level between 1997 and 2000 (nearly 60%), and the significantly different relationship between oxychlordane and return rate between 1997 and 2000, further indicate that this compound may play a primary role in OC related mortality in the glaucous gull. Why oxychlordane? Oxychlordane is a metabolite from cis- and trans-chlordane which are major compounds in chlordane, a cyclodiene insecticide previously used commonly in agriculture (Wiemeyer, 1996). Oxychlordane has a lethal hazard level that begins near 5 ppm in brain tissue of passerines (Wiemeyer, 1996), and has a very steep dose-response curve (Bondy et al., 2003). In comparison, lethal concentrations of PCB and DDE levels are much higher (reviewed by Blus, 1996; Hoffman et al., 1996; Wiemeyer, 1996). However, failure to return may again depend on additive or synergistic effects of all these compounds, of which oxychlordane may be the most important. Between 1997 and 2003 our study population has declined (Bustnes et al., 2003a; H. Strøm, personal communication) and our studies may suggest that organochlorine pollution has the potential to influence glaucous gull populations in the Arctic. The effect of OCs on the population will, however, depend on the proportion of the population with sufficiently high levels of contamination (Bustnes et al., 2003a). Exposure to contaminants seems to be related to the trophic level at which the birds feed (Bustnes et al., 2000, 2003b). As a result, the ecological effects of OCs may largely depend on the feeding ecology of a given glaucous gull population. There are also other confounding factors that may influence return rates, e.g., age, but this is not a likely explanation for the relationship between return rate and OCs (see Bustnes et al., 2003a,b for discussions).

Acknowledgements We are grateful to Kjetil Sagrup and Jonathan Verreault for valuable help during field work, and Anuschka Polder and her team for conducting the OC analyses. We also wish to thank two anonymous reviewers for comments that greatly improved an earlier draft of the manuscript. The study was funded by the Norwegian Research Council and the Norwegian Ministry for Environment.

184

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185

References Andersen, M., Lie, E., Derocher, A.E., Belikov, S.E., Bernhoft, A., Bolutov, A.N., Garner, G.W., Skaare, J.U., Wiig, O., 2001. Geographic variation of PCB congeners in polar bears (Ursus maritimus) from Svalbard east to the Chukchi Sea. Polar Biology 24, 231e238. Bocquene, G., Bellanger, C., Cadiou, Y., Galgani, F., 1995. Joint action of combinations of pollutants on the acetyl cholinesterase activity of several marine species. Ecotoxicology 4, 266e279. Blus, L.J., 1996. DDT, DDD and DDE in birds. In: Beyer, W.N., Heinz, G.H., Redmon-Norwood, A.W. (Eds.), Environmental Contaminants in Wildlife. Interpreting Tissue Concentrations. CRC Press, Boca Raton, pp. 49e71. Bogan, J.A., Bourne, W.R.P., 1972. Organochlorine levels in Atlantic seabirds. Nature 240, 358. Boon, J.P., van den Meer, J., Allchin, C.R., Law, R.J., Klungsøyr, J., Leonards, P.E.G., Spliid, H., Storr-Hansen, E., Mckenzie, C., Wells, D.E., 1997. Concentration-dependent changes of PCB patterns in fish-eating mammals: structural evidence for induction of cytochrome P450. Archives of Environmental Contamination and Toxicology 33, 298e311. Bondy, G., Armstrong, C., Coady, L., Doucet, J., Robertson, P., Feeley, M., Barker, M., 2003. Toxicity of the chlordane metabolite oxychlordane in female rats: clinical and histopathological changes. Food and Chemical Toxicology 41, 291e301. Brevik, E.M., 1978. Gas chromatographic method for determination of organochlorine pesticides in human milk. Bulletin of Environmental Contamination and Toxicology 19, 281e286. Brunstro¨m, B., Halldin, K., 2000. Ecotoxicological risk assessment of environmental pollutants in the Arctic. The Science of the Total Environment 112, 111e118. Burger, J.K., Giesy, J.P., Grue, C.E., Gochfeld, M., 2001. Effects of environmental pollutants on avian behaviour. In: Dell’Omo, G. (Ed.), Behavioural Ecotoxicology. John Wiley & Sons, West Sussex, pp. 337e376. Bustnes, J.O., Erikstad, K.E., Bakken, V., Mehlum, F., Skaare, J.U., 2000. Feeding ecology and the concentration of organochlorines (OCs) in glaucous gulls. Ecotoxicology 9, 175e182. Bustnes, J.O., Bakken, V., Erikstad, K.E., Mehlum, F., Skaare, J.U., 2001. Patterns of incubation and nest-site attentiveness in relation to organochlorine (PCB) contamination in glaucous gulls. Journal of Applied Ecology 38, 791e801. Bustnes, J.O., Skaare, J.U., Erikstad, K.E., Bakken, V., Mehlum, F., 2001. Whole blood concentrations of organochlorines as a dose metric for studies of the glaucous gull (Larus hyperboreus). Environmental Toxicology and Chemistry 20, 1046e1052. Bustnes, J.O., Folstad, I., Erikstad, K.E., Fjeld, M., Miland, Ø.O., Skaare, J.U., 2002. Blood concentration of organochlorine pollutants and wing feather asymmetry in glaucous gulls. Functional Ecology 16, 617e622. Bustnes, J.O., Erikstad, K.E., Bakken, V., Skaare, J.U., Mehlum, F., 2003a. Ecological effects of organochlorine pollutants in the arctic: a study of the glaucous gull. Ecological Applications 13, 504e515. Bustnes, J.O., Bakken, V., Skaare, J.U., Erikstad, K.E., 2003b. Age and accumulation of persistent organochlorines: a study of arctic breeding glaucous gulls. Environmental Toxicology and Chemistry 22, 2173e2179. Bustnes, J.O., Hanssen, S.A., Folstad, I., Erikstad, K.E., Hasselquist, D., Skaare, J.U., 2004. Immune function and organochlorine pollutants in arctic breeding glaucous gulls. Archives of Environmental Contamination and Toxicology 47, 530e541. Calow, P., 1991. Physiological costs of combating chemical toxicants: ecological implications. Comparative Biochemistry and Physiology 100, 3e6.

Colborn, T., vom Saal, F.S., Soto, A.M., 1993. Developmental effects of endocrine-disrupting chemicals in wildlife and humans. Environmental Health Perspectives 101, 378e384. Cramp, S., Simmons, K.E.L., 1983. The birds of the Western Palearctic. Oxford University Press, Oxford, UK. Dewailly, E., Ayotte, P., Bruneau, S., Gingras, S., Belles-Isles, M., Roy, R., 2000. Susceptibility to infections and immune status in Inuit infants exposed to organochlorines. Environmental Health Perspectives 108, 205e211. Fox, G.A., Gilman, A.P., Peakall, D.B., Anderka, F.W., 1978. Behavioral abnormalities of nesting Lake Ontario herring gulls. Journal of Wildlife Management 42, 477e483. Gabrielsen, G.W., Skaare, J.U., Polder, A., Bakken, V., 1995. Chlorinated hydrocarbons in Glaucous gulls (Larus hyperboreus) in the southern part of Svalbard. The Science of the Total Environment 160/161, 337e346. Garcı´ a-Berthou, E., 2001. On the misuse of residuals in ecology: testing regression residuals vs. the analysis of covariance. Journal of Animal Ecology 70, 708e711. Gill, H., Wilson, L.K., Cheng, K.M., Elliott, J.E., 2003. An assessment of DDT and other chlorinated compounds and reproductive success of American Robins (Turdus migratorius) breeding in fruit orchard. Ecotoxicology 12, 113e123. Gilman, A.P., Fox, G.A., Peakall, D.B., Teeple, S.M., Carroll, T.R., Haymes, G.T., 1977. Reproductive parameters and contaminant levels of Great Lakes herring gull. Journal of Wildlife Management 41, 458e468. Grasman, K.A., Scanlon, P.F., Fox, G.A., 2000. Geographic variation in haematological variables in adult and prefledling herring Gulls (Larus argentatus) and possible associations with organochlorine exposure. Archives of Environmental Contamination and Toxicology 38, 244e253. Grue, C.E., Gardner, S.C., Gibert, P.L., 2002. On the significance of pollutant-induced alterations in the behaviour of fish and wildlife. In: Dell’Omo, G. (Ed.), Behavioural Ecotoxicology. John Wiley & Sons, West Sussex, pp. 1e90. Henriksen, E.O., Gabrielsen, G.W., Skaare, J.U., 1998. Validation of the use of blood samples to assess tissue concentrations of organochlorines in Glaucous gulls, Larus hyperboreus. Chemosphere 37, 2627e2643. Henriksen, E.O., Gabrielsen, G.W., Trudeau, S., Wolkers, J., Sagerup, K., Skaare, J.U., 2000. Organochlorines and possible biochemical effects in glaucous gulls (Larus hyberboreus) from Bjørnøya, the Barents Sea. Archives of Environmental Contamination and Toxicology 38, 234e243. Herbert, C.E., Norstrom, R.J., Weseloh, D.V.C., 1999. A quarter century of environmental surveillance: the Canadian Wildlife Service’s Great Lakes herring gull monitoring program. Environmental Reviews 7, 147e166. Hoffman, D.J., Rice, C.P., Kubiak, T.J., 1996. PCBs and dioxins in birds. In: Beyer, W.N., Heinz, G.H., RedmonNorwood, A.W. (Eds.), Environmental Contaminants in Wildlife. Interpreting Tissue Concentrations. CRC Press, Boca Raton, pp. 165e207. Hose, J.E., Guillette, L.J., 1995. Defining the role of pollutants in the disruption of reproduction in wildlife. Environmental Health Perspectives 103 (Suppl. 4), 87e91. Janssens, E., Dauwe, T., van Duyse, E., Beernaert, J., Pinxten, R., Eens, M., 2003. Effects of heavy metal exposure on aggressive behavior in a small territorial songbird. Archives of Environmental Contamination and Toxicology 45, 121e127. Jones, K.C., Voogt, de, P., 1999. Persistent organic pollutants (POPs): state of the science. Environmental Pollution 100, 209e221. Keith, J.O., Mitchell, C.A., 1993. Effects of DDE and food stress on reproduction and body condition of ringed turtle doves.

J.O. Bustnes et al. / Environmental Pollution 136 (2005) 175e185 Archives of Environmental Contamination and Toxicology 25, 192e203. Kelce, W.R., Stone, C.R., Laws, S.C., Gray, L.E., Kemppainen, J.A., Wilson, E.M., 1995. Persistent DDT metabolite p,p#-DDE is a potent androgen receptor antagonist. Nature 375, 581e585. Lebreton, J.-D., Clobert, J., 1991. Bird population dynamics, management, and conservation: the role of mathematical modelling. In: Perrins, C.M., Lebreton, J.D., Hirons, G.J.M. (Eds.), Bird Population Studies: Relevance to Conservation and Management. Oxford University Press, London, pp. 105e121. Love, O.P., Shutt, L.J., Silfies, J.S., Bortolottio, G.R., Smits, J.E.G., Bird, D.M., 2003. Effects of dietary PCB exposure on adrenocortical function in captive American Kestrels (Falco sparverius). Ecotoxicology 12, 199e208. Mackay, D., Wania, F., 1995. Transport of contaminants to the Arctic: partitioning, processes and models. The Science of the Total Environment 160/161, 25e38. McCarty, J.P., Secord, A.L., 1999. Reproductive ecology of tree swallows (Tachycineta bicolor) with high levels of polychlorinated biphenyl contamination. Environmental Toxicology and Chemistry 18, 1433e1439. McCullagh, P., Nelder, J.A., 1989. Generalized linear models, second ed. Chapman and Hall, London. McKinney, J.D., Fawkes, J., Jordan, S., Chae, K., Oatley, S., Coleman, R.E., Briner, W., 1985. 2,3,7,8-tetrachlorodibenzo-pdioxin (TCDD) as a potent and persistent thyroxine agonist: a mechanistic model for toxicity based on molecular reactivity. Environmental Health Perspective 61, 41e53. Mora, M.A., Auman, H.J., Ludwig, J.P., Giesy, J.P., Verbrugge, D.A., Ludwig, M.E., 1993. Polychlorinated-biphenyls and chlorinated insecticides in plasma of caspian terns - relationships with age, productivity, and colony site-tenacity in the Great-Lakes. Archives of Environmental Contamination and Toxicology 24, 320e331. Norstrom, R.J., 1988. Bioaccumulation of polychlorinated biphenyls in Canadian wildlife. In: Crine, J.P. (Ed.), Hazards, Decontamination and Replacements of PCBs. Plenum Publishing, New York, USA, pp. 85e100. Payne, J., Scholze, M., Kortenkamp, A., 2001. Mixtures of four organochlorines enhance human breast cancer cell proliferation. Environmental Health Perspectives 109, 391e397.

185

Peakall, D.B., 1996. Disrupted patterns of behavior in natural populations as an index of ecotoxicity. Environmental Health Perspectives 104 (Suppl. 2), 331e335. Petersen, R.E., Theobald, H.M., Kimmel, G.L., 1993. Development and reproductive toxicity of dioxins and related compounds: cross species comparisons. Critical Review of Toxicology 23, 283e335. Rattner, B.A., Franson, J.C., 1984. Methyl parathion and fevalerate toxicity in American kestrels: acute physiological responses and effects of cold. Canadian Journal of Physiology and Pharmacology 62, 787e792. Sagerup, K., Henriksen, E.O., Skorping, A., Skaare, J.U., Gabrielsen, G.W., 2000. Intensity of parasitic nematodes increases with organochlorine levels in Glaucous Gull. Journal of Applied Ecology 37, 532e539. SAS, 1993. SAS Technical Report P-243 SAS/STAT Software: The GENMOD Procedure, Release 6.09, SAS Institute Inc., Cary, NC. SAS, 1999. SAS OnlineDocÒ, Version 8, SAS Institute, Inc., Cary, NC. Verreault, J., Skaare, J.U., Jenssen, B.M., Gabrielsen, G.W., 2004. Effects of organcohlorine contaminants on thyroid hormone levels in Arctic breeding glaucous gulls Larus hyperboreus. Environmental Health Perspectives 112, 532e537. Vos, J.G., Luster, M.I., 1989. Immune alterations. In: Kimbrough, R.D., Jensen, A.A. (Eds.), Halogenated Biphenyls, Terphenyls, Naphthalenes, Dibenzodioxins, and Related Products. Elsevier, New York, pp. 295e322. Walker, C.H., 2003. Neurotoxic pesticides and behavioural effects upon birds. Ecotoxicology 12, 307e316. Walker, C.H., Hopkin, S.P., Sibly, R.M., Peakall, D.B., 2001. Principles of ecotoxiology, second ed. Taylor & Francis, London. Wiemeyer, S.N., 1996. Other organochlorine pesticides in birds. In: Beyer, W.N., Heinz, G.H., Redmon-Norwood, A.W. (Eds.), Environmental Contaminants in Wildlife. Interpreting Tissue Concentrations. CRC Press, Boca Raton, pp. 99e115. Wooller, R.D., Bradley, J.S., Croxall, J.P., 1992. Long-term population studies of seabirds. Trends in Ecology and Evolution 7, 111e114.