Association of Low Reproductive Rates and High Contaminant Levels in Bald Eagles on Green Bay, Lake Michigan

Transcription

1 J. Great Lakes Res. 27(2): Internat. Assoc. Great Lakes Res., 2001 Association of Low Reproductive Rates and High Contaminant Levels in Bald Eagles on Green Bay, Lake Michigan Cheryl R. Dykstra 1, *, Michael W. Meyer 2, Kenneth L. Stromborg 1, D. Keith Warnke 3,, William W. Bowerman, IV 4,, and David A. Best 5 1U.S. Fish and Wildlife Service Green Bay, Wisconsin Bureau of Integrated Science Services Wisconsin Department of Natural Resources Rhinelander, Wisconsin Department of Fisheries and Wildlife Cooperative Research Unit University of Minnesota St. Paul, Minnesota Department of Fisheries and Wildlife Pesticide Research Center and Institute for Environmental Toxicology Michigan State University East Lansing, Michigan U.S. Fish and Wildlife Service East Lansing, Michigan ABSTRACT. Bald eagles (Haliaeetus leucocephalus) nesting on the shores of Green Bay, Lake Michigan, had reproductive rates significantly lower than those of neighboring eagles nesting in inland Wisconsin (0.55 vs. 1.1 young per occupied territory). This study investigated effects of two factors which have depressed eagle reproductive rates at other locations: exposure to organochlorine contaminants and low food availability. Levels of dichloro-diphenyl-dichloroethylene (DDE) and total polychlorinated biphenyls (PCBs) in addled eggs and eaglet blood samples from Lake Michigan and inland Wisconsin reference sites were measured. An index to food availability, the food delivery rates by parent eagles to nestlings, and other behavioral indices that have been associated with food availability, were measured and compared to inland reference data. Mean contaminant concentrations in addled eggs from Green Bay were 8.3 µg/g wet weight DDE and 31.3 µg/g wet weight total PCBs (1987 to 1996, n = 9). Mean concentrations in nestling blood plasma were 53 µg/kg wet weight DDE and 207 µg/kg wet weight total PCBs (1987 to 1995, n = 8). Indices of food availability were generally normal in comparison to inland reference data. Mean food delivery rate to the nestlings was 2.55 items per nestling per day (n = 7 nests). None of the behavioral measures differed significantly from inland reference values, suggesting that prey was adequate. It was concluded that organochlorine contaminants caused all or most of the depression in reproductive rates of Green Bay bald eagles. INDEX WORDS: Bald eagles, DDE, PCBs, Lake Michigan, food, reproductive rates, Green Bay. * Corresponding author. Present address: 7715 Mitchell Park Dr., Cleves, OH cheryldykstra@juno.com Present address: Wisconsin Department of Natural Resources, Madison, WI Present address: Department of Environmental Toxicology, Clemson University, Pendleton, SC INTRODUCTION Between 1947 and 1970, populations of bald eagles nesting in the continental United States declined due to reproductive failure caused mainly by DDE, a metabolite of the organochlorine insecticide 239

2 240 Dykstra et al. DDT (dichloro-diphenyl-trichloroethane; Wiemeyer et al. 1972, Colborn 1991). After the use of DDT and other organochlorines was banned, the North American eagle population rebounded quickly throughout most of its range. However, bald eagle populations in a few regions, including the Great Lakes shorelines (Colborn 1991, Best et al. 1994), have not increased as rapidly. The mean productivity of eagles near the shores of the Great Lakes (excluding Lake Ontario) from 1989 to 1993 was 0.79 young/occupied territory, compared to 1.0 young/occupied territory in inland Michigan and Minnesota (Bowerman 1993). Bald eagles were extirpated from historic Green Bay nest sites by Along the Lake Michigan shore in Wisconsin, the first post-ddt era nest was initiated in 1986 at the Peshtigo River. The number of nesting pairs on the Wisconsin shore of Green Bay increased slowly to a maximum of six pairs in Elsewhere along Lake Michigan, a similar slow increase in nesting pairs has been documented (Best et al. 1994). Despite population increases, the eagles nesting on the shores of Lake Michigan still exhibit reproductive rates lower than those of neighboring birds in inland Wisconsin and Michigan (Colborn 1991, Best et al. 1994). There is continued concern over the health of these eagles, not only because they are a threatened species, but also because they serve as a sensitive indicator of the overall health of the Green Bay ecosystem. In the Green Bay ecosystem, organochlorines in the sediments and water (Evans 1988) continue to contaminate the biota (Haseltine et al. 1981, Heinz et al. 1983). Double-crested cormorants (Phalacrocorax auritus) and Forster s terns (Sterna forsteri) continue to experience depressed reproduction, which was associated with the organochlorines in their prey (Kubiak et al. 1989, Custer et al. 1999). Bald eagles may be at higher risk from contaminants than either of these species because the fish that eagles consume are larger and higher on the food chain than the fish eaten by cormorants and terns. In some Great Lakes locations, eagles occasionally prey on cormorants and gulls, putting the eagles one step higher on the food chain (Dykstra 1995, Warnke 1996). Because of these dietary habits, eagles bioaccumulate lipophilic contaminants and thus suffer greater exposure to toxicants than species such as cormorants and terns. However, depressed productivity in eagles is not always the result of organochlorine contamination. In uncontaminated areas, eagle reproductive success (Hansen 1987) and population density (Gerrard et al. 1983, Whitfield and Gerrard 1985, Dzus and Gerrard 1993) have been associated with prey abundance. Inland eagle reproduction declined at small lakes in the Upper Peninsula of Michigan when rough fish were removed (Bowerman 1991). Even in some relatively contaminated regions, low reproductive rates have been found to be associated with low food availability (possibly in combination with environmental contaminants; Gill 1998). Within the Great Lakes, on Wisconsin s Lake Superior shore and Apostle Islands, low reproduction was likely caused by low food availability in Lake Superior (Dykstra et al. 1998). This study investigated contaminant loads, food availability, and reproductive rates for the Green Bay eagle population for the 10-year period from 1987 to Documented here are reproductive rates of Green Bay eagles which are compared to reference populations in inland Wisconsin. Total PCB and DDE concentrations are reported in addled eggs and eaglet blood samples, as well as the route of exposure to contaminants (species consumed). Food delivery rates by parent eagles to nestlings were used as an index to food availability in the environment, and finally, this study assesses whether current low productivity can be explained by organochlorine contaminants, low food availability, or neither. METHODS Study Sites Included in the study were all eagle nests located within 8 km of the shore of Green Bay (including the northern end of the bay, known as Big Bay de Noc and Little Bay de Noc, MI) during the 10 years from 1987 to One nest located on the eastern shore of Door County, WI, was also included because the waters of Lake Michigan in that area are near enough to Green Bay proper to be influenced by any contaminant conditions there (nest number DO-01, Fig. 1). Reproductive Rate Reproductive rate was assessed by the Wisconsin and Michigan Departments of Natural Resources (WDNR and MDNR) from 1987 to 1996 by inspecting nests from the air twice during the breeding season, once during incubation and again when

3 Low Productivity and High Contaminants in Green Bay Eagles 241 FIG. 1. Locations of bald eagle nesting territories on Green Bay and associated Lake Michigan shores. nestlings were 4 to 7 weeks old. In the first aerial survey, the eagle pairs that were occupying territories and/or incubating eggs were counted, and in the second flight, the resulting nestlings were counted. For a regional summary, the total number of young produced was divided by the total number of territories that birds occupied. An occupied territory was defined as one where eggs had been laid, or two eagles were present on the territory, or the nest had been visibly repaired (even if no adults or only one non-incubating adult was observed; Postupalsky 1974). From 1987 to 1993, WDNR tallied reproduction at inland nests in a slightly different way, where an occupied territory was defined simply as a territory where eggs had been laid (formerly called an active territory, described herein as Wisconsin-Occupied ). Contaminant Concentrations in Addled Eggs Between 1986 and 1996, 15 addled bald eagle eggs were collected from six Green Bay nests located in Michigan and Wisconsin, at the time of banding nestlings. Eggs were wrapped in aluminum foil and refrigerated until processing. Prior to opening, the eggs were weighed, and their lengths and widths were measured with calipers. After opening, the egg contents were frozen in chemically-clean jars until analysis. Most eggs collected in Wisconsin for WDNR were analyzed by the Wisconsin State Lab of Hygiene for organochlorine pesticides and total PCBs using gas chromatography, with confirmation by mass spectrometry (Wisconsin State Laboratory of Hygiene 1996). Total PCBs were determined by comparison to Aroclor mixtures. One Wisconsin egg was analyzed using simi-

4 242 Dykstra et al. lar methodology for the U.S. Fish and Wildlife Service at Texas A. & M. by the Geochemical and Environmental Research Group (Wade et al. 1988). Eggs collected in Michigan were analyzed at Patuxent Analytical Control Facility, National Fisheries Contaminant Research Center, or Hazleton Environmental Services Inc. using similar methodology (Cromartie et al. 1975, Keith et al. 1983, Hazleton Laboratories America, Inc., respectively). Residue concentrations were corrected for moisture loss as in Stickel et al. (1973). Values were reported as µg/g wet weight. Data from pairs of sibling eggs were averaged using geometric means. A fresh dead nestling approximately 1 to 3 days old was collected from one nest in Wisconsin. The neonate was homogenized by personnel at the WDNR and analyzed for the USFWS at Mississippi State Chemical Laboratory using the above methodology (Mississippi State University 1985). There was no adjustment for moisture loss since the nestling was collected shortly after it died. Percent moisture in the neonate was similar to that in eggs, but the percent of lipid in the nestling was somewhat lower than that in eggs (1.5% vs. 3.4 to 7%). Contaminant concentrations in this nestling were not included in statistical analyses but are shown in figures. Contaminant Concentrations in Nestling Blood In 1992 to 1995, six blood samples were collected from bald eagle nestlings at four Green Bay nests. Nestlings were sexed by footpad length (Bortolotti 1984), and aged by the length of the eighth primary (Bortolotti 1984) when age was not known from observations. Nestlings were age 5 to 8 weeks at the time of the blood collection. Syringes used were either sterile plastic or glass previously washed with hexanes and acetone. Approximately 10 ml of blood was drawn from the brachial vein. Blood was transferred to heparinized vacutainers, stored on wet ice until the end of the day, and separated by centrifuging in the evening. Plasma was drawn off, transferred to another vacutainer, and frozen upright at approximately 20 C. At the end of the 1994 field season, samples from 1992 to 1994 were shipped on dry ice to Michigan State University for analysis. Samples from 1995 were handled similarly, but shipped to and analyzed by Hazleton Environmental Services, Inc., Madison, WI. Organochlorine pesticides and total PCB concentrations in the nestling plasma were determined by gas chromatography with electron capture detection and were confirmed by mass spectrometry (Bowerman et al. 1994, Mora et al. 1993, U.S. EPA 1984). Detection limits were 2.5 µg/kg for DDE and 5.0 µg/kg for total PCBs. Contaminant concentrations below detection limits were assigned a value of half the detection limit for statistical analyses. Residue levels measured in the plasma of sibling nestlings in the same year were averaged (geometric mean) to produce one value, but residues measured at the same nesting territory in more than one year were reported separately. Food Delivery Rates and Other Behaviors Food delivery rates at seven Green Bay eagle nests were measured in 1995 to 1997 by two techniques, video cameras and direct observations. Remote time-lapse video cameras were mounted at two eagle nests in 1995, two nests in 1996, and one nest in At four of these nests, cameras were placed after chicks were 4 to 6 weeks old, to avoid disturbance; at one videotaped nest, the camera was placed in winter before breeding began. In 1995, dawn-to-dusk observations were made at two nests using 20 to 60 X spotting scopes. Observation blinds were placed 150 to 300 m from the nests. Observer fatigue was prevented by switching observers every 4 hours without disturbing the nesting eagles. Observers recorded all behaviors of nestlings and adults, with the time of occurrence and duration (nearest minute). Behaviors included feeding, brooding, preening, fighting, exercising, sleeping, resting, etc. Observers also monitored prey deliveries by adults to the nestlings, and attempted to identify the taxonomic Class and the size of each prey item. Prey were categorized into one of three Classes: fish, bird, or mammal. Recorded further was the identification to species or species-group if possible. Prey were categorized to one of four sizeclasses: 0 to 15 cm, 15 to 30.5 cm, 30.5 to 46 cm, or > 46 cm. Videotapes were reviewed after the end of the field season. Prey deliveries and selected behaviors were recorded, with their time of occurrence and duration. Prey items were classified as above. Behaviors selected for analysis were feeding and adult attendance at the nest, because these behaviors appeared to be indicators of the adequacy or inadequacy of the prey base (Warnke 1996). Video data and direct observer data were combined for analysis because simultaneous observations (validations)

5 Low Productivity and High Contaminants in Green Bay Eagles 243 with video cameras indicated that the two techniques were not significantly different for number of prey deliveries or other behaviors (Warnke unpubl. data, Dykstra et al. 1998). Analysis of Behavioral Data For both camera nests and direct observation nests, behavioral data were summarized based on the age of the nestling(s). The hatching day for the oldest nestling was defined as day 0. Week 0 included days 0 to 6; week 1 included days 7 to 13, etc. In most nests observed the hatching date was known to within 7 days. Time spent feeding was defined as all time in which a nestling fed itself or was fed by an adult. Time spent feeding (%) was summed over each day (nearest minute) and divided by the total observation time for that day (which was equal to the daylength), multiplied by 100. For nests with two chicks, the nestlings feeding-times were totaled separately and averaged. Adult attendance at the nest was defined as the time that at least one adult was present in the nest. Adult attendance (%) was summed over each day (nearest minute) and divided by the total observation time (daylength) as above. For feeding time, adult attendance, and prey delivery rates, only days which were recorded completely from dawn to dusk were included in the analysis. When a nest was observed more than 1 day in a week, the multiple observations were averaged to produce a single pooled datum for that week (Warnke 1996). For prey delivery rates, weekly data were averaged to produce a seasonlong prey delivery rate for each nest. For analysis of prey Classes, sizes, and species, all available data were used, including partiallyrecorded days. Multiple observations in 1 week were not pooled. Statistical Analyses All analyses were performed with Systat (Wilkinson 1988). Reproductive rates were compared with t-tests. Contaminant concentrations in blood samples and eggs are shown as geometric means. Contaminant concentrations in blood and eggs were log-transformed to account for non-normality before comparisons by t-test. Similarly, weekly prey delivery rates were log-transformed before comparison to nestling age by linear regression. Non-parametric Mann-Whitney U-tests were used for all other comparisons of behavioral data to inland reference data. A probability value < 0.05 was considered significant. RESULTS Reproductive Rate and Population The population of breeding pairs on Green Bay has increased slowly since 1987, and had reached 11 to 13 occupied territories from 1993 to However, the productivity of these birds was low. From 1987 to 1996, only 49 young were produced in 88 nest attempts (0.55 young per occupied territory). Productivity during the first 5 years of this period, 0.42 young per occupied territory (n = 31 territories, 1987 to 1991), did not differ from productivity during the last 5 years of this period (0.63 young per occupied territory, n = 57 territories, t 86 = 1.50, p = 0.14). Productivity for neighboring birds in Vilas and Oneida Counties, WI, was 1.10 young per occupied territory during 1987 to 1996 (1,370 chicks in 1,244 territories; 1987 to 1993 were estimated from the number of Wisconsin-occupied territories, based on the ratio of occupied to Wisconsin-occupied nests in the north-central district, WI, in 1994, and in Iron and Gogebic Counties, MI, 1991 to For Michigan data, detailed observer notes were used to determine the number of occupied nests that would have also been classified as Wisconsin-occupied by the WDNR definition, and these numbers were used to generate the above ratio; S. Postupalsky unpubl. data, see Dykstra 1995 for details). Contaminant Concentrations in Eggs Green Bay eggs had higher concentrations of DDE than did inland eggs (8.3 vs µg/g wet weight, n = 9 Green Bay eggs, n = 20 inland eggs, t 27 = 9.0, p < 0.001, Fig. 2). Concentrations of total PCBs in Green Bay eggs were also significantly higher than levels measured in inland Wisconsin eggs (31.3 vs. 2.4 µg/g wet weight PCBs, n = 9 Green Bay eggs, n = 20 inland eggs, t 27 = 12.5, p < 0.001, Fig. 3). Results were similar when contaminant concentrations in the dead neonate were included in the analyses. Contaminant Concentrations in Blood The geometric mean concentration of DDE in nestling plasma from Green Bay was 53 µg/kg wet weight (n = 8, Table 1), which was higher than that of nestlings in inland Wisconsin (4 µg/kg, from

6 244 Dykstra et al. FIG. 2. Concentrations of DDE in addled eggs from Green Bay nests One 1986 Green Bay egg shown here, but not included in statistical tests; nestling identified here not included in statistical tests. Inland Wisconsin egg data from WDNR, published in Dykstra et al. 1998, shown for comparison. Line at >15-16 g/ g is the concentration associated with near-total reproductive failure in a nationwide study (Wiemeyer et al. 1984, 1993). Line at <3-3.6 g/ g is the upper concentration associated with normal reproductive rates (Wiemeyer et al. 1984, 1993). Dykstra et al. 1998; p = 0.006). The geometric mean concentration of total PCBs in nestling plasma was 207 µg/kg wet weight (Table 1), compared to 34 µg/kg in inland Wisconsin nestlings (Dykstra et al. 1998; p < 0.001). Organochlorines other than DDE and PCBs were present only at low levels (unpubl. data). Food Delivery Rates and Other Behaviors Food delivery rates to nestlings averaged 2.55 prey items per nestling per day (Table 2), which was slightly lower than the average delivery rate in inland Wisconsin (2.99 prey items per nestling-day; reported in Warnke 1996), although the difference was not statistically significant (Mann-Whitney U- test, U = 53.0, p = 0.55, n = 7 Green Bay and 13 inland). Delivery rate did not increase or decrease with the age of the nestlings in either location (linear regression with log-transformed delivery rates, p = 0.31, n = 12 weeks, for inland data see Warnke 1996). Biomass and energy content of prey were FIG. 3. Total PCBs in addled eggs from Green Bay nests One 1986 Green Bay egg shown here, but not included in statistical tests; nestling identified here not included in statistical tests. Inland Wisconsin egg data from WDNR, published in Dykstra et al. 1998, shown for comparison. Line at >33 g/ g is the concentration associated with near-total reproductive failure in a nationwide study (Wiemeyer et al. 1984; see text for caveats). not determined because previous research showed that biomass and energy content did not differ between inland Wisconsin and Great Lakes locations (Dykstra 1995). A total of 447 prey items (78.9%) could be identified to taxonomic Class (Table 3). Of the identified prey, the most common were fish (97.1% of all identified prey items), followed by birds (2.5%). Only two mammals were identified (both at nest MT-17). The distribution of known prey types in the diet may have differed from that of inland eagles, whose diet contained 97.6% fish and only 0.7% birds (numbers of prey items in classes other than fish were too small for legitimate statistical comparison, n = 423 prey items in four taxonomic classes; Warnke unpubl. data ). Some prey items were identified to species or species-group. Of 434 fish identified, 161 were further categorized (Table 3). Suckers, northern pike, and yellow perch or perch/walleye made up the largest proportions of the identified fish (29.8%, 17.4%, and 16.8%), followed by carp (8.1%). However, these proportions may not exactly represent the species composition found in the entire diet;

7 Low Productivity and High Contaminants in Green Bay Eagles 245 TABLE 1. Contaminant loads in nestling blood plasma samples from Green Bay nests, 1987 to Nest DDE Total PCBs Year State Number Nest Name (µg/kg wet wt) (µg/kg wet wt) 1987 MI De-13 Granskog Lake a MI De-15 Boutlier Lake a WI MT-16 Peshtigo R. N WI MT-17 Blueberry Is WI DO-01 Toft Point WI OC-04 Oconto R. Thome WI DO-01 Toft Point b WI MT-17 Blueberry Is b GEOMETRIC MEANS a Data from Bowerman (1991) Data may be biased low (matrix spike recoveries low) TABLE 2. Prey delivery rates for Green Bay nests, 1995 to Nest and # Days Nestling Brood Prey Deliveries Prey Deliveries Year Observed Age (wks) Size (per day) (per nestling per day) OC DO De b MT b De b c Mm b OC b MEAN 2.55 a Nests in Wisconsin designated by two capitalized letters followed by a number; nests in Michigan designated by a capital letter and a lowercase letter followed by a number. b Only complete days (dawn-to-dusk) were included in the calculation of prey delivery rates, although many partial days were video-taped. c One nestling died at about 8 weeks of age, probably as a result of a fall from the nest. these species are relatively easy to identify and may be over-represented in the identified subset. Of the 11 birds identified, 8 were further classified to species or species-group (most were ring-billed or herring gulls, n = 5). Time spent feeding, which is likely an indicator of food availability (Dykstra 1995, Warnke 1996), did not vary with nestling age (Fig. 4a). The time that nestlings spent feeding was summarized in three nestlings phases, to make the data comparable to the reference inland data. Phases were defined as Early (nestlings age 1 to 3 weeks), Mid (4 to 7 weeks), and Late (8 to 11 weeks). Within each phase, data from each nest were averaged to produce one value per nest. Methods of tallying totals for Green Bay differed slightly from that for inland, but a comparison of the two methods on a subset of the data indicated that the methods were statistically indistinguishable and hence the two populations could be compared (paired t-test, p = 0.87, n = 20 d of Green Bay data). There was no difference in the time spent feeding at the two locations during any nestling phases (Early, U = 9.0, p = 0.73, n = 3 Green Bay and 7 inland; Mid, U = 20.0, p = 0.36, n=7 Green Bay and 8 inland; Late, U = 15.0, p = 0.081, n = 7 Green Bay and 9 inland; Warnke unpubl. data). Adult attendance at nests (Fig. 4b) is another probable indicator of food availability, because low attendance rates were associated with low food availability at Lake Superior nests (Warnke 1996). These data were also summarized in the three

8 246 Dykstra et al. TABLE 3. Prey items delivered to Green Bay bald eagle nests and identified to taxonomic class or species, 1995 to Number identified Number identified Prey Item to species (%) 1 to Class (%) 2 Class Osteichthyes 434 (97.1) Sucker (Catostomus spp.) 48 (28.4) Northern pike (Esox lucius) 28 (16.6) Yellow perch or perch/walleye (Perca favescens or Stizostedion vitreum) 27 (16.0) Bass (Micropterus spp.) 18 (10.7) Bullheads (Ictalurus spp.) 15 (8.9) Carp (Cyprinus carpio) 13 (7.7) Other Centrarchids 8 (4.7) Alewife (Alosa pseudoharengus) 1 (0.6) Gizzard shad (Dorosoma cepedianum) 1 (0.6) Coregonus (Coregonus spp.) 1 (0.6) Trout (unknown sp.) 1 (0.6) Class Aves 11 (2.5) Gulls (Larus spp.) 5 (3.0) Double-crested cormorant (Phalacrocorax auritus) 2 (1.2) Duck (unknown sp.) 1 (0.6) Class Mammalia 2 (0.4) TOTAL Percent of 169 total prey items that were identified to species 2 Percent of 447 total prey items that were identified to Class nestling phases for comparison to inland reference data. Methods of tallying totals also differed slightly, but, again, a test of the two methods on a subset of the data indicated that the methods were comparable (paired t-test, p = 0.14, n = 20 days of Green Bay data). There was no difference in adult attendance at the two locations (Early phase, U = 13.0, p = 0.57, n = 3 Green Bay and 7 inland; Mid, U = 36.0, p = 0.12, n = 6 Green Bay and 8 inland; Late, U = 18.0, p = 0.29, n = 6 Green Bay and 9 inland; Warnke unpubl. data). For adult attendance data, nest De-09 was not included in the analysis above because adults there behaved unusually after the introduction of the camera on 13 June Adults at De-09 mainly visited the nest to deliver prey, and rarely stood in the nest for more than a few seconds after a delivery. Two adults were never seen simultaneously. Such behavior occurred in only two other nests after the introduction of a camera (of a total of 33 camera introductions, Dykstra et al. unpubl. data). However, at De-09, prey delivery rates were normal, and the chick developed normally, fledging by 26 to 28 July DISCUSSION Productivity The productivity of Green Bay eagles was 50% below the normal rate of inland Wisconsin eagles. This reproductive rate was well below the rate which has been associated with a healthy population (1.0 young per occupied territory; Kubiak and Best 1991, Best et al. 1994), and also below the rate required to maintain population numbers (0.7 young per occupied territory; Sprunt et al. 1973). Such low rates were similar to those of the 1970s and suggest that some factor strongly depressed reproduction in this population during 1987 to There was, however, some indication that Green Bay production improved in recent years: productivity in 1992 to 1996 was marginally higher than that in 1987 to 1991, although the difference was not statistically significant. Additionally, productivity on Green Bay in 1997 to 1999 was greater than 1.0 young per occupied territory for the first time (unpubl. data). The data in this study indicated that successful nesting pairs on Green Bay raised an average of 1.36 young per pair (unpubl. data), while inland Wisconsin parents averaged 1.7 young per success-

9 Low Productivity and High Contaminants in Green Bay Eagles 247 breeding pair are young birds. It is not possible to determine the exact mechanism of low productivity without further intensive research. Regardless of the mechanism, the low reproductive rates indicate that this subpopulation was unable to maintain its numbers without immigration from outside the Green Bay area. It is likely that the highly-successful inland Wisconsin and inland Michigan subpopulations were the source of the birds that colonized the Green Bay shores in the past decade. These source subpopulations can likely continue to sustain the Green Bay subpopulation through repeated emigration, should the Green Bay population fail to increase its reproductive rate. FIG. 4a. Time spent feeding (mean and s.e.m.) at Green Bay nests. 4b. Adult attendance at Green Bay nests. Daily values were recorded as a percentage of the total available daylight. Time spent feeding and adult attendance defined as in text. Multiple observations for a single nest in one week were averaged to produce a single value for each nest (see Methods). Not all nests were observed every week. Only complete observation days (dawn-to-dusk) were included. Week 0 = days 0-6, etc. ful nest (Dykstra 1995). The cause of this difference was unknown. It may have resulted from (1) a small clutch size, (2) loss of eggs/young nestlings, or (3) nestling mortality at a later stage of the nestling period. For Green Bay nests studied in 1995 to 1997, only one case was documented of later-stage nestling mortality (after the first third of the nestling period). Thus, this late mortality appeared to be rare. It is possible that Green Bay pairs lost eggs or young nestlings due to infertility, embryo mortality, or nestling mortality; such phenomena could have resulted from high contaminant loads. However, it is also possible that small clutches were laid or that no eggs were laid; failure to lay or small clutch size may indicate that the Contaminants Concentrations of both DDE and total PCBs in addled eggs from Green Bay eagles were very high. Other organochlorine pesticides were also present in eggs. One of these, dieldrin, was present in some eggs at levels sometimes associated with reproductive impairment (Wiemeyer et al. 1984): dieldrin concentrations averaged 0.4 ± 0.1 µg/g in five Green Bay eggs for which data were available (unpubl. data). However, the concentration of dieldrin present in Green Bay eggs was lower than that found to cause reproductive impairment in several experimental studies in other species (Wiemeyer et al. 1984), and is not thought to be limiting to bald eagles in the Great Lakes (Giesy et al. 1995). The concentrations of DDE in Green Bay eggs were similar to or higher than those in contemporary eggs from the most contaminated areas where eagles breed in the U.S.: Maine, the upper Klamath Basin, the lower Columbia River, and Lakes Erie and Superior (Frenzel 1985, Henny and Anthony 1989, Bowerman et al. 1994, Welch 1994, J. Buck pers. comm.). Mean concentration of DDE (8.3 µg/g) was midway between the level associated with near-total reproductive failure ( µg/g), and the highest level associated with healthy reproduction ( µg/g; Wiemeyer et al. 1984, 1993). These data suggest that at least some reproductive depression was likely caused by DDE. The critical concentrations for total PCBs were less evident than those for DDE, because of the ubiquitous co-variance of PCBs and DDE (Nisbet 1989, Wiemeyer et al. 1993). The effect of PCBs on eagle reproduction was found to be less significant than that of DDE and, according to Wiemeyer and co-workers, the observed effect of PCBs was probably due to the correlation of PCBs with DDE in the

10 248 Dykstra et al. eggs studied (Wiemeyer et al. 1993). There have been few studies of effects of PCBs in the absence of DDE, due to a paucity of suitable samples, but one analysis of recent riverine samples indicated that relatively high levels of PCBs in nestling plasma were associated with normal-to-high rates of eagle reproduction, in the absence of DDE (or in the presence of negligible amounts of DDE; Karasov and Meyer 2000). However, given the above caveats, total PCB concentrations in Green Bay eggs (31.3 µg/g) were near the level associated with near-total reproductive failure in an early study (> 33 µg/g, Wiemeyer et al. 1984). Concentrations in Green Bay eggs were also higher than those in eggs from Maine, the upper Klamath Basin, the lower Columbia River and Lakes Erie and Superior (Frenzel 1985, Henny and Anthony 1989, Bowerman et al. 1994, Welch 1994, J. Buck pers. comm.). Only Lake Huron eggs had higher total PCB levels than those found in Green Bay (Bowerman et al. 1994). Such elevated concentrations suggest that some part of the reproductive impairment may have been caused by PCBs. Nestling blood samples also contained significantly elevated levels of both DDE and total PCBs. The average concentrations at Green Bay were 6 to 14 times higher than at inland sites. It is more difficult to assess the critical concentrations of contaminants associated with reproductive failure in nestling blood samples than in addled eggs, because few researchers have compared nestling blood samples to adult reproduction. However, using correlative techniques similar to those used for addled eggs, Elliott and Norstrom (1998) estimated that nestling blood levels of >170 µg/kg DDE or > 500 µg/kg total PCBs were associated with populationlevel effects on reproduction. Average nestling contaminant concentrations in Green Bay, 53 µg/kg DDE and 207 µg/kg total PCBs, were lower than suggested critical concentrations, although some individual nestlings concentrations exceeded these levels. It should be noted that no no-observable-adverse-effect-level (NOAEL) has been suggested for nestling blood contaminant concentrations; hence it is still unclear what level of contamination can be considered to be safe. Analysis of prey items delivered to the nest indicated a plausible route of exposure to bioaccumulative toxicants. Carp in the diet probably constituted the highest risk of exposure, since carp are highly contaminated (Giesy et al. 1995, WDNR unpublished data) and made up 8% of the fish in the diet by count and a greater proportion of the diet by biomass. Spawning carp were extremely abundant during the latter half of the nestling period near three of the four nests observed in The birds in the diet, mainly gulls, may have posed a significant risk of exposure for certain eagles and eaglets. The proportion of birds in the diet was as high as 15% (nest De-09), and the gulls and cormorants of Green Bay are highly contaminated (Dale and Stromborg 1993). The variation in prey types between nests located in differing ecological situations indicated that Green Bay bald eagles were opportunistic in selecting prey, and that some nests, such as those near carp spawning areas or gull colonies, may be at higher risk for contamination than others. This variation in prey types is likely reflected in the variation in egg and nestling plasma contaminant levels (Table 1, Fig. 2, Fig. 3); however, sample sizes were too small to attempt to differentiate the subtle variation in Green Bay habitats in this study. Food Delivery Rates and Behavioral Indices Food delivery rates per nestling to Green Bay nestlings were not significantly different from those measured in the reference inland population. Nor was there any difference in the distributions of prey Classes delivered to Green Bay nests, or in the distributions of prey sizes delivered to Green Bay nests (unpubl. data). Within the prey Classes, there were some differences between the two locations. For inland eagles, northern pike and bullheads made up the largest proportion of the identified fish prey (Warnke unpubl. data). However, at Green Bay nests, suckers made up the majority of the identified fish. No carp or gulls were identified at any inland nests. Quantitative behavioral indices of food availability, determined in a previous study (Warnke 1996), revealed that behaviors of Green Bay eagles and eaglets were normal. Time spent feeding was similar to inland nestlings, suggesting an adequate food supply (in contrast to results for Wisconsin s Lake Superior shore, where time spent feeding was significantly lower than at inland sites; Warnke 1996, Dykstra et al. 1998). Adult attendance at the nests was statistically indistinguishable from that of inland eagles. This contrasts with Lake Superior results, where adults spent more time away from the nest when nestlings were young; presumably these adults spent the additional time searching for food; Warnke 1996, Dykstra et al (As a caveat, note that these results at Green Bay were less con-

11 Low Productivity and High Contaminants in Green Bay Eagles 249 clusive for the early nestling stage than for the mid and late nestling stages due to small sample size in ENS, because most cameras were placed during the MNS.) Unquantified behaviors also indicated that foodstress did not occur at Green Bay nests. Green Bay chicks often left food uneaten in the nest for hours after receiving it and did not attempt to monopolize prey items. There was only minimal fighting among siblings, and it was not occasioned by the arrival of food. In fact, one Green Bay nestling was photographed feeding a morsel of food to its sibling, a rare event which also occurred in inland Wisconsin but not on Wisconsin s Lake Superior shore. On Green Bay, little nestling mortality occurred at age 4 to 7 weeks. In contrast, food-stressed nestlings near Lake Superior did not leave uneaten food and often fought over food or attempted to monopolize food. There was also considerable nestling mortality, especially at ages 4 to 7 weeks, when energy demands are likely highest (Dykstra et al. 1998, Dykstra et al. 2001). In summary, the behavioral data indicated that there was no evidence of any food-stress among Green Bay eagles. One caveat to the interpretation of the Green Bay behavioral data is that most nests studied had only one nestling. Only three Green Bay nests contained two nestlings, and one of those was reduced to a single chick at about 8 weeks of age. On Wisconsin s Lake Superior shore and Apostle Islands, the documented food-stress was evident only in nests with two nestlings; thus, it is possible that the Green Bay data set contained too few nests with two nestlings for adequate assessment of behaviors. CONCLUSIONS Although it seems likely that organochlorine contaminants caused most of the depression in reproduction, one additional factor may also have contributed to the problem. If the breeding eagles on Green Bay were younger than average, they may have had lower reproductive success simply because of their inexperience (Newton 1979). Anecdotal evidence indicates that some of the breeding eagles of Green Bay were young (brown-streaked heads; pers. obs.). If reproduction were depressed by the inexperience of the parents, it should improve over time, as the parents gain skills. However, overall reproduction would not improve over time if the breeding adults were constantly replaced by young parents. It has been suggested that Great Lakes adults have low survival rates and high replacement (turnover) rates (Kozie and Anderson 1991, Bowerman et al. 1995, Kubiak and Best 1991), but there are presently no data to test this hypothesis. Thus the available data indicated that it is very likely that contaminants and not low food availability caused most or all the reduction in reproduction of Green Bay eagles in the period from 1987 to Diet analysis indicated a plausible route of exposure to these contaminants. Intercorrelation prevented the separation of the effects of DDE and total PCBs. However, levels of both contaminants were very high and at least DDE and probably PCBs seemed likely to cause reproductive depression at the measured concentrations. ACKNOWLEDGMENTS We thank Dave Evans, Jack Holt, Joe Papp, and Bradley Richardson for retrieval of eaglets and eggs from the nests. For field assistance, we are grateful to James Anderson, Luke Cieslewicz, Jean Lundin, Dave Picard, Kiersten Rattray, Brian Wehausen, and Jim Woodford. For assistance with reporting of video data, we thank Kristie Sitar. The Wisconsin Department of Natural Resources, Michigan Department of Natural Resources, and Sergej Postupalsky provided productivity data. Special thanks to the Scott Paper Company, Dave Moody, Jim Johnson, and University of Wisconsin- Green Bay for access to private land. We thank C. J. Henny, S. Postupalsky, M. C. Shieldcastle, L. M. Valoppi, and S. N. Wiemeyer for providing helpful comments on earlier drafts of the manuscript. Funding for this project was provided in part by the U.S. Fish and Wildlife Service (project number ) and the Wisconsin Department of Natural Resources. REFERENCES Best, D.A., Bowerman, W.W., IV., Kubiak, T.J., Winterstein, S.R., Postupalsky, S., Shieldcastle, M.C., and Giesy, J.P., Jr Reproductive impairment of bald eagles Haliaeetus leucocephalus along the Great Lakes shorelines of Michigan and Ohio. In Raptor Conservation Today, eds. B.-U. Meyburg and R.D. Chancellor, pp World Working Group on Birds of Prey. East Sussex, Great Britain: Pica Press. Bortolotti, G.R Criteria for determining age and sex of nestling bald eagles. J. Field Ornithol. 55: Bowerman, W.W., IV Factors influencing breed-

12 250 Dykstra et al. ing success of bald eagles in upper Michigan. M.S. thesis, Northern Michigan University, Marquette, MI Regulation of bald eagle (Haliaeetus leucocephalus) productivity in the Great Lakes basin: an ecological and toxicological approach. Ph.D. dissertation, Michigan State University, East Lansing, MI, Best, D.A., Kubiak, T.J., Postupalsky, S., Shieldcastle, M.C., Giesy, J.P The influence of environmental contaminants on bald eagles populations in the Laurentian Great Lakes, North America. In Raptor Conservation Today, eds. B.-U. Meyburg and R.D. Chancellor, pp World Working Group on Birds of Prey. East Sussex, Great Britain: Pica Press., Giesy, J.P., Best, D.A., and Kramer, V.J A review of factors affecting productivity of bald eagles in the Great Lakes region: implications for recovery. Environ. Health Perspect. 103 (Suppl 4): Colborn, T Epidemiology of Great Lakes bald eagles. J. Toxicol. Environ. Health 33: Cromartie, E.W., Reichel, W.L., Locke, L.N., Belisle, A.A., Kaiser, T.E., Lamont, T.G., Mulhern, B.M., Prouty, R.M., and Swineford, D.M Residues of organochlorine pesticides and polychlorinated biphenyls and autopsy data for bald eagles, Pestic. Monit. J. 9: Custer, T.W., Custer, C.M., Hines, R.K., Gutreuter, S., Stromborg, K.L., Allen, P.D., and Melancon, M.J Organochlorine contaminants and reproductive success of double-crested cormorants from Green Bay, Wisconsin, USA. Environ. Toxicol. Chem. 18: Dale, T.B., and Stromborg, K.L Reconnaissance surveys of contaminants potentially affecting Green Bay and Gravel Island National Wildlife Refuges. U.S. Fish and Wildlife Service. Unpublished report. Dykstra, C.J.R Effects of contaminants, food availability and weather on the reproductive rate of Lake Superior bald eagles (Haliaeetus leucocephalus). Ph.D. dissertation, University of Wisconsin-Madison, Madison, WI., Meyer, M.W., Warnke, D.K., Karasov, W.H., Andersen, D.E., Bowerman, W.W., IV., and Giesy, J.P Low reproductive rates of Lake Superior Bald Eagles: low food delivery rates or environmental contaminants? J. Great Lakes Res. 24:32 44., Karasov, W.H., Meyer, M.W., and Warnke, D.K Daily energy expenditure of nestling bald eagles in northern Wisconsin. Condor 103: Dzus, E.H., and Gerrard, J.M Factors influencing bald eagle densities in north-central Saskatchewan. J. Wildl. Manage. 57: Elliott, J.E., and Norstrom, R Chlorinated hydrocarbon contaminants and productivity of bald eagle populations on the Pacific coast of Canada. Environ. Toxicol. Chem. 17: Evans, M.S Ed., Toxic contaminants and ecosystem health: a Great Lakes focus, Vol. 21. Wiley Series Advances in Environmental Science and Technology. New York: John Wiley and Sons. Frenzel, R.W Environmental contaminants and ecology of bald eagles in southcentral Oregon. Ph.D. dissertation, Oregon State University, Corvallis, OR. Gerrard, J.M., Gerrard, P.N., Bortolotti, G.R., and Whitfield, D.W. A A 14-year study of bald eagle reproduction on Besnard Lake, Saskatchewan. In Biology and management of bald eagles and ospreys, pp Ste. Anne de Bellevue, Quebec: Harpell Press. Giesy, J.P., Bowerman, W.W., Mora, M.A., Verbrugge, D.A., Othoudt, R.A., Newsted, J.L., Summer, C.L., Aulerich, R.J., Bursian, S.J., Ludwig, J.P., Dawson, G.A., Kubiak, T.J., Best, D.A., and Tillitt, D.E Contaminants in fishes from Great Lakes-influenced sections and above dams of three Michigan rivers: III. Implications for health of bald eagles. Arch. Environ. Contam. Toxicol. 29: Gill, C.E Environmental contaminants, food availability, and reproduction of bald eagles (Haliaeetus leucocephalus) on Vancouver Island, British Columbia. M.S. thesis, Simon Fraser University, British Columbia. Hansen, A.J Regulation of bald eagle reproductive rates in southeast Alaska. Ecol. 68: Haseltine, S.D., Heinz, G.H., Reichel, W.L., and Moore, J.F Organochlorine and metal residues in eggs of waterfowl nesting on islands in Lake Michigan off Door County, Wisconsin, Pestic. Monit. J. 15: Hazleton Laboratories America, Inc. Determination of organochlorine pesticides and polychlorinated biphenyls (PCBs) in biological tissues. Method MP- FWST-MA, Hazleton Laboratories America, Inc., Madison, WI. Heinz, G.H., Haseltine, S.D., Reichel, W.L., and Hensler, G.L Relationships of environmental contaminants to reproductive success in red-breasted mergansers Mergus serrator from Lake Michigan. Environ. Pollut. 32A: Henny, C.J., and Anthony, R.G Bald eagle and osprey. Western raptor management symposium and workshop. National Wildlife Federation, Scientific and Technical Series 12: Karasov, W.H., and Meyer, M.W Testing the role of contaminants in depressing avian numbers. Revista Chilena de Historia Natural 73: Keith, L.H., Crummet, W., Deegan, J., Libby, R.A., Taylor, J.K., and Wentler, G Principles of environmental analysis. Anal. Chem. 55: Kozie, K.D., and Anderson, R.K Productivity, diet, and environmental contaminants in bald eagles nesting near the Wisconsin shoreline of Lake Superior. Arch. Environ. Contam. and Toxicol. 20: Kubiak, T.J., and Best, D.A Wildlife risks associ-

13 Low Productivity and High Contaminants in Green Bay Eagles 251 ated with passage of contaminated, anadromous fish at Federal Energy Regulatory Commission licensed dams in Michigan. Unpubl. report, U.S. Fish and Wildlife Service, East Lansing, MI., Harris, H.J., Smith, L.M., Schwartz, T.R., Stalling, D.L., Trick, J.A., Sileo, L., Docherty, D.E., and Erdman, T.C Microcontaminants and reproductive impairment of the Forster s tern on Green Bay, Lake Michigan Arch. Environ. Contam. and Toxicol. 18: Mississippi State University Method 1: Analysis for chlorinated hydrocarbon pesticides and related compounds. Mississippi State Chemical Laboratory, Mississippi State, MS. Mora, M.A., Auman, H.J., Ludwig, J.P., Giesy, J.P., Verbrugge, D.A., and Ludwig, M.E Polychlorinated biphenyls and chlorinated insecticides in plasma of Caspian terns: relationships with age, productivity and colony site tenacity in the Great Lakes. Arch. Environ. Contam. Toxicol. 24: Newton, I Population Ecology of Raptors. Vermillion, South Dakota: Buteo Books. Nisbet, I.C.T Organochlorines, reproductive impairment and declines in bald eagles Haliaeetus leucocephalus populations: mechanisms and doseresponse relationships. In Raptors in the Modern World, eds. B.-U. Meyburg and R.D. Chancellor, pp World Working Group on Birds of Prey, Berlin. Postupalsky, S Raptor reproductive success: some problems with methods, criteria, and terminology. In Management of Raptors, eds. F.N. Hamerstrom, Jr., B.E. Harrell, and R.R. Ohlendorff, pp Proc. Conf. Raptor Conserv. Tech., Raptor Res. Report No. 2. Sprunt, A., IV., Robertson, W.B., Jr., Postupalsky, S., Hensel, R.J., Knoder, C.E., and Ligas, F.J Comparative productivity of six bald eagle populations. Trans. N. Amer. Wildl. Conf. 38: Stickel, L.F., Wiemeyer, S.N., and Blus, L.J Pesticide residues in eggs of wild birds: adjustment for loss of moisture and lipid. Bull. Env. Contam. Toxicol. 9: U.S. Environmental Protection Agency (EPA) Method 608. Federal Register 49 (209): Wade, T.L., Atlas, E.L., Brooks, J.M., Kennicutt, M.C., II, Fox, R.G., Sericano, J., Garcia, B., and DeFreitas, D NOAA Gulf of Mexico Status and Trends Program: trace organic contaminant distribution in sediments and oyster. Estuaries 11: Warnke, D.K Comparative behavioral ecology of two bald eagle populations in Northern Wisconsin. M.S. thesis, University of Minnesota, St. Paul, MN. Welch, L.J Contaminant burdens and reproductive rates of bald eagles breeding in Maine. M.S. thesis, University of Maine. Whitfield, D.W.A., and Gerrard, J.M Correlation of bald eagle density with commercial fish catch. In Bald Eagle in Canada: proceedings of Bald Eagle Days, 1983, eds. J.M. Gerrard and T.M. Ingram, pp Headingley, Manitoba: White Horse Plains Publishers. Wiemeyer, S.N., Mulhern, B.M., Ligas, F.J., Hensel, R.J., Mathisen, J.E., Robards, F.C., and Postupalsky, S Residues of organochlorine pesticides, polychlorinated biphenyls, and mercury in bald eagle eggs and changes in shell thickness, 1969 and Pestic. Monitor. J. 6:50 55., Lamont, T.G., Bunck, C.M., Sindelar, C.R., Gramlich, F.J., Fraser, J.D., and Byrd, M.A Organochlorine pesticide, polychlorinated biphenyls, and mercury residues in bald eagle eggs and their relationship to shell thinning and reproduction. Arch. Environ. Contam. Toxic. 13: , Bunck, C.M., and Stafford, C.J Environmental contaminants in bald eagle eggs and further interpretations of relationships to productivity and shell thickness. Arch. Environ. Contam. Toxic. 24: Wilkinson, L Systat: the system for statistics. Evanston, Illinois: Systat, Inc.. Wisconsin State Laboratory of Hygiene Methods of Organic Analysis Manual January 1, Wisconsin State Laboratory of Hygiene, Madison, WI. Submitted: 20 September 2000 Accepted: 11 February 2001 Editorial handling: R. Peter Richards