Archives of Environmental Contamination and Toxicology

, Volume 62, Issue 4, pp 681–695

Mercury and Other Mining-Related Contaminants in Ospreys Along the Upper Clark Fork River, Montana, USA

Authors

    • Geosciences DepartmentUniversity of Montana
  • Erick Greene
    • Division of Biological Sciences and The Wildlife Biology ProgramUniversity of Montana
  • Robert Domenech
    • Raptor View Research Institute
  • Molly F. Staats
    • Geosciences DepartmentUniversity of Montana
Article

DOI: 10.1007/s00244-011-9732-5

Cite this article as:
Langner, H.W., Greene, E., Domenech, R. et al. Arch Environ Contam Toxicol (2012) 62: 681. doi:10.1007/s00244-011-9732-5

Abstract

We investigated links between mining-related contaminants in river sediment and their occurrence in nestling ospreys (Pandion haliaetus) in the Clark Fork River Basin, Montana, USA. Blood and feather samples from 111 osprey chicks were collected during 4 years from nests along river sections with greatly different sediment concentrations of arsenic (As), cadmium (Cd), copper (Cu), lead (Pb), zinc (Zn), and mercury (Hg). No significant differences between river sections were found among Zn (3,150 ± 160 μg L−1) and Cd (<5 μg L−1) concentrations in blood. Cu, Pb, and As concentrations in blood were significantly increased in chicks from the most contaminated river sections (mean values of 298, 8.9, and 100 μg L−1, respectively). Cu, Zn, and Pb concentrations increased significantly during a year of above-average river runoff combined with high suspended sediment loads in rivers. Total Hg concentrations in blood and feathers were highly correlated and depended on the geographic locations of the nests. The lowest blood concentrations of Hg were observed in the most upstream river section (mean 151 μg L−1) where total sediment concentrations were increased (0.80 mg kg−1). River sections with intermediate blood concentrations (mean 206 and 303 μg L−1) were associated with low to intermediate sediment concentrations (0.058 and 0.46 mg kg−1). The highest concentrations of Hg in ospreys (mean 548 μg L−1) were observed downstream from a contaminated tributary (1–4 mg kg−1 in sediment). In river sections with lower Hg concentrations in sediment, there was a negative correlation between blood Hg concentration and chick mass, presumably due to high deposition rates into growing feathers. This relationship was absent in sections of high Hg exposure. Osprey blood and feathers are suitable for monitoring Hg in aquatic ecosystems; however, responses of As, Cd, Cu, Pb, and Zn are more subtle.

Ospreys (Pandion haliaetus) have been used extensively as sentinels for environmental contamination in aquatic environments (Henny et al. 1991; DesGranges et al. 1998; Elliott et al. 2000; Golden and Rattner 2003; Toschik et al. 2005; Hopkins et al. 2007; Rattner et al. 2008; Rivera-Rodríguez and Rodríguez-Estrella 2011). This is because ospreys eat almost exclusively fish, are top trophic-level predators in aquatic food chains and thus potentially at risk from environmental contamination, nest readily on man-made nest platforms, and habituate quickly to human activity and other disturbance (Grove et al. 2009; Henny et al. 2010).

Although many inorganic and organic contaminants have been quantified in nestling ospreys (reviewed in Grove et al. 2009), few data are available relating contaminant levels in osprey tissues to traditional measures of environmental contamination, such as concentrations in sediments, water, or other biota. We studied the patterns of mining-related contaminants in ospreys in the Upper Clark Fork River Basin (UCFRB [Fig. 1]) of Western Montana. This basin is one of the nation’s largest United States Environmental Protection Agency (USEPA) Superfund sites, with high levels of the priority contaminants arsenic (As), cadmium (Cd), copper (Cu), lead (Pb,) and zinc (Zn) (USEPA 2004, 2011). These contaminants originated from historic mining and smelting activities in the headwaters near Butte and Anaconda, Montana (Fig. 1), and have been transported, primarily in fine sediments of the river, several hundred kilometers downstream (Moore and Luoma 1990). The Clark Fork River (CFR) combines with two large rivers near Missoula—Blackfoot (BFR) and Bitterroot (BRR) Rivers—that are comparable in discharge but are minimally affected by historic mining. The BFR and BRR are also inhabited by ospreys and serve as excellent reference rivers for comparison.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-011-9732-5/MediaObjects/244_2011_9732_Fig1_HTML.gif
Fig. 1

UCFRB with osprey nest sites that were included in the study

Although mercury (Hg) is not a contaminant considered in the USEPA Superfund process in the UCFRB, it is recognized as a dangerous pollutant. Unlike the As, Cd, Cu, Pb, and Zn that were released into the CFR primarily from the Butte–Anaconda smelting operations, Hg was released from hundreds of historic precious metals mines throughout the CFR watershed, thus creating a spatially diffuse source. Local sources of Hg are overlaid by atmospheric inputs from regional and global sources (Schuster et al. 2002), affecting both the more pristine tributaries and the CFR. Effects of Hg on local biota may be greatly modified by different net rates of methylation and by the organisms’ position in the food web. No systematic studies have been conducted to characterize the methylmercury exposure of biota in the UCFRB across a comparable level in the food web; therefore, methodical testing of ospreys in the UCFRB may provide valuable baseline information on the current exposure status in local environments and help distinguish between local and external Hg sources.

We studied the relationships between mining-related contaminants in river sediments and their presence in nestling ospreys. We focused on osprey chicks because virtually all of their biomass grew from consumption of local fish, thus reflecting local environmental conditions. This is compared with adult ospreys, which can accumulate considerable concentrations of contaminants on their wintering grounds (Henny et al. 2010; Rivera-Rodríguez and Rodríguez-Estrella 2011). Our objective was to determine and compare source contaminant levels in river sediments and in osprey chicks as a representative predator at the top of the local aquatic food web. We targeted priority contaminants of the USEPA Superfund cleanup and Hg. The UCFRB is the subject of an extensive environmental remediation and restoration effort, and thus our results are important baseline information that will allow us to assess changes in levels of these contaminants in ospreys in the future.

Materials and Methods

Study Area and Sample Collection

The majority of osprey samples (87%) were obtained during the month of July, 2006–2009, and the remainder during June and August of these years. We sampled 36 osprey nests in the UCFRB in western Montana, USA (Table 1). Twenty-six nests were located along the main stem CFR within a 293 river-km reach between Warm Springs and Tarkio. The remaining nests were located at least 5 km away from the main stem CFR along the large tributaries Little Blackfoot River (LBF), BFR, and BRR. Active nests containing young between 4 and 7 weeks of age were accessed with a boom truck, and nestlings were brought to the ground for processing. Older nestlings were covered with a large net if there appeared to be any risk of them leaping from the nest. The birds were weighed and banded, and after treating the skin with an alcohol pad, approximately 2 mL of blood was obtained from the brachial vein using a sterile 25-gauge hypodermic needle. Whole-blood samples were split immediately between two preweighed 1.5-mL microcentrifuge tubes. Samples were immediately stored on ice, and whole-blood samples were frozen at −18°C on the same day. Two to four nape feather tips were sampled if fully developed. To avoid overheating, nestlings were processed in the shade and misted with water on hot days. Before returning the chicks to the nest, we removed baling twine and other potentially hazardous materials to the extent possible (Blem et al. 2002; Houston and Scott 2006). The field sampling procedure took between 15 and 40 min, depending on the number of young, ease of access, and the condition of the nest. Adult ospreys generally returned to the nest and resumed normal parenting behavior within several minutes after removing the boom truck. Almost all nests were accessed only once per season, but one nest (C03) was resampled after 5 days, and one chick from nest C24 was resampled after being found injured 1 month after fledging. These activities were covered by the following permits: United States Geological Survey (USGS) Federal Banding and Sampling Permit 23353, Montana Fish, Wildlife, and Parks Wild Bird Banding and Possession Permits 573 and 029, and University of Montana–Missoula Institutional Animal Care and Use Committee Animal Use Protocol No. 013-07EG-DBS-060807.
Table 1

Analytical results in whole blood and feathers

Nest

Year

Chick

Mass (g)

Blood

Feathers

Hg (μg L−1)

As (μg L−1)

Cu (μg L−1)

Pb (μg L−1)

Zn (μg L−1)

Hg (mg kg−1)

C01

2007

1

1116

139

 

346

22

3271

2.9

2

1170

111

 

365

7.8

3732

2.1

3

997

148

 

297

15

3148

2.5

2008

1

725

124

17

289

7.5

3031

 

C02

2008

1

1152

192

17

386

11

3683

 

2

860

205

21

387

13

3537

 

2009

1

1457

170

50

282

<5

3544

2.2

2

1417

172

76

282

<10

3503

2.0

3

1483

165

40

270

<4

3404

2.5

C03

2006

1

1400

100

 

243

7.2

2406

1.4

2007

1

1102

120

 

347

8.0

3717

4.4

2008

1

870

208

126

295

10

3307

 

2

1044

170

348

283

7.6

3406

 

2a

1209

156

51

344

15

3391

 

3

1165

194

150

340

24

3041

 

2009

1

1057

198

306

224

7.3

3002

2.7

C04

2006

1

1388

112

 

289

6.8

2938

2.7

2

1625

115

 

260

6.5

3191

2.4

2007

1

1270

97

 

357

7.9

4548

1.5

2008

1

1413

104

76

343

16

3791

 

2

1580

97

32

322

12

3332

 

2009

1

939

205

7.9

290

<5

2870

2.9

C05

2007

1

1200

240

 

254

24

2299

2.7

C06

2007

1

1305

523

 

209

3.6

3256

5.1

2

1570

527

 

234

6.6

2938

5.4

C07

2009

1

1109

467

28

247

9.8

3601

12

2

1372

482

24

227

13

2715

11

3

1391

504

18

233

12

3243

9.8

C08

2009

1

1679

690

10

294

12

3523

11

C09

2008

1

1530

725

8.1

354

12

3256

 

2

1810

460

8.2

396

12

3357

 

3

1886

532

10

359

11

3383

 

2009

1

1635

366

37

215

4.8

2879

4.0

C10

2009

1

1312

657

19

288

6.4

3801

15

2

1419

588

9.4

197

4.8

2450

8.8

C11

2007

1

1400

412

 

204

<3

2622

8.1

2008

1

1035

482

21

309

14

2777

 

2

1262

437

18

318

15

2668

 

2009

1

1252

279

5.1

179

5.7

2615

5.4

2

1342

317

<5

221

5.8

2905

5.0

3

1485

285

5.4

315

4.8

4819

3.7

C12

2007

1

1390

387

 

257

3.0

3431

3.8

2

1648

363

 

315

24

4110

3.4

2008

1

1412

100

17

341

8.6

3132

 

2

1395

100

18

346

8.3

2960

 

2009

1

1649

349

8.9

314

<5

2980

5.7

2

1664

357

7.4

224

<5

3049

4.6

C13

2006

1

1690

362

 

248

<10

2628

3.7

2

1480

347

 

256

<6

2781

3.3

2007

1

1546

259

 

201

<3

2397

3.4

2

1107

288

 

277

3.4

3194

3.8

C14

2008

1

1408

108

14

340

7.3

2842

 

2

1169

117

12

310

7.7

3412

 

2009

1

1273

305

6.2

237

<4

2548

2.2

C15

2007

1

1330

255

    

6.7

2

1140

260

 

350

3.8

3509

7.1

2009

1

957

272

6.6

236

6.4

2767

6.1

2

931

280

<6

216

<11

2519

5.2

C16

2007

1

1354

370

 

294

7.8

2892

4.9

2

1532

366

 

209

2.8

3075

6.3

C17

2008

1

1249

355

6.2

311

43

3506

 

C18

2006

1

1465

356

 

274

<5

2557

5.2

2

1890

320

 

264

<5

3077

6.6

C19

2006

1

1715

191

 

300

<6

2849

5.0

2

1630

200

 

265

<5

2389

8.6

C20

2006

1

1239

407

 

234

<7

2877

4.6

2

1299

371

 

242

<11

2998

7.8

2007

1

1408

369

 

238

<3

2754

6.1

2

1596

  

221

<8

2750

6.8

2008

1

1118

213

<5

294

5.0

3678

 

2009

1

1186

262

6.6

213

6.7

2504

5.2

2

922

235

7.6

217

5.7

2476

5.4

C21

2008

1

1473

613

<5

308

8.4

3497

 

2009

1

1381

408b

<4

232

<4

2901

5.4

2

1479

162

<4

276

5.0

3437

4.7

3

1386

144

<5

253

<5

3449

4.2

C22

2008

1

1735

241

<4

309

6.1

3182

 

2

1435

210

3.8

305

6.3

3014

 

C23

2006

1

1703

434

 

265

<6

3318

5.2

C24

2007

1

1407

369

 

303

25

2867

7.4

2009

1

1009

387

4.7

208

<4

2668

 

1c

1531

2786

5.8

427

<5

2099

6.7

2

1107

405

<5

251

<5

2744

 

C25

2008

1

1462

171

11

345

8.8

3832

 

C26

2007

1

935

239

 

257

3.4

3171

4.2

LB01

2008

1

1480

124

50

285

8.8

3201

 

2

1732

127

46

342

8.8

3404

 

3

1605

134

40

325

8.5

3547

 

BF01

2008

1

1407

138

5.2

264

4.4

3186

 

BF02

2007

1

1233

182

 

250

<4

2802

0.95

2

1642

182

 

203

<3

2786

3.0

3

1667

190

 

223

<2

2948

2.2

2008

1

1590

174

6.2

308

5.7

3506

 

2

1552

159

6.0

335

4.8

3753

 

BR07

2009

1

1371

166

<5

188

<5

2422

4.7

2

1549

163

7.5

206

<5

3199

4.5

BR06

2007

1

1346

221

 

201

<3

2788

3.3

2

1284

213

 

190

<4

3881

2.5

2008

1

1538

211

<6

430

6.7

3796

 

2

1307

197

<4

381

4.3

3618

 

BR05

2008

1

1710

249

6.3

247

<5

3182

 

2

1395

228

<5

264

<5

3122

 

2009

1

1172

219

7.3

267

<4

3352

3.4

2

999

260

7.9

271

<5

3483

4.8

3

1153

221

6.9

283

<5

3079

4.7

BR04

2006

1

1290

143

 

298

<9

3222

6.1

2

1586

146

 

271

<5

2594

5.9

3

1476

171

 

263

<8

2547

7.2

BR03

2009

1

1415

246

4.6

247

<4

2908

2.8

BR02

2008

1

1245

127

<10

379

18

3246

 

2

1610

452b

5.1

293

13

3823

 

BR01

2009

1

1079

294

15

229

<7

3188

3.6

2

805

307

8.8

218

<6

2770

4.5

Quantification limits vary between samples depending on the volume of blood analyzed. Arsenic results are only available for 2008 and 2009 as an improved inductively coupled plasma-mass spectrometry (ICP-MS) method was used. Feather Hg content is reported only where mature cover feathers were present

aResults from resampling the same bird after 5 days

bValues treated as outliers and omitted from statistical analysis

cResults from resampling the same bird one month after fledging

Fine-sediment samples were collected from the river bottom during base flow conditions in August 2009 using a common method to evaluate heavy-metal contamination (Shelton and Capel 1994). Briefly, we combined samples from the upper 1–3 cm of an area of approximately 30 m2 and passed them through a 63-μm plastic screen using river water. Duplicate sediment samples were obtained for each site. Samples were centrifuged, dried at 60°C, and homogenized with a glass mortar.

Sample Analysis

All analyses were conducted by the Environmental Biogeochemistry Laboratory at the University of Montana Geosciences Department. Microcentrifuge tubes were weighed before and after the addition of whole blood to determine the sample mass. For analysis, the complete frozen samples were transferred into 50 mL perfluoroalkoxy (PFA) digestion vials (Savillex 0202, precleaned as described by DeWild et al. 2002) by squeezing the conical bottom of the tube. Remaining sample was transferred by filling the tube with concentrated trace-metal grade HNO3 for 15 min followed by two water rinses. A total of 3.3 mL of concentrated HNO3 and 2 mL of deionized water were added to each blood sample. These mixtures were then digested at room temperature for 30 min before refluxing at 80–85°C for 4 h. After cooling, 1 mL of 30% concentrated ACS-grade H2O2 was added and heated again to 60–70°C for 30 min. Samples were then allowed to cool, and deionized water was added to a total volume of 50 mL. Digests were analyzed for total Hg using a Leeman Hydra AF mercury analyzer according to USEPA method 245.7 (Cold Vapor Atomic Fluorescence Spectrometry). As, Cd, Cu, Pb, and Zn concentrations were analyzed using a Perkin Elmer Elan DRCII inductively coupled plasma mass spectrometer (USEPA method 200.8). Practical quantitation limits were 0.2 μg L−1 for Hg and 5 μg L−1 for As, Cd, Pb, and Zn in whole blood. Recoveries for Hg, As, Cd, and Pb in whole-blood standard reference samples (Bio-Rad Lyphochek controls 1–3, in duplicate) were within the manufacturer-listed ranges, except when they were near or lower than detection limits, indicating the suitability of the digestion and analytical methods for blood samples (Table 2). Other quality-control samples were analyzed and checked as required by the respective USEPA methods used. To convert wet mass-based (μg kg−1) to volume-based (μg L−1) concentrations, we determined the density of whole blood on a subset of fresh osprey blood samples (1.06 kg L−1, n = 4).
Table 2

Measured concentrations (μg L−1) in whole-blood standard reference materials (Bio-Rad Lypochek controls 1–3)

 

Control 1

Control 2

Control 3

Mercury

8.6–8.7 (7.6–11)

42–46 (37–56)

84–89 (73–110)

Arsenic

<5–8.6 (2.5–7.0)

32–45 (32–47)

58–82 (57–85)

Cadmium

<5 (3.8–5.7)

9.4–9.8 (9.7–15)

27 (26–39)

Lead

79–84 (79–120)

230 (230–350)

430–440 (420–640)

Copper

390

380

380

Zinc

2500

2400

2300

n = 2 for each control. Manufacturer-listed concentrations are in parentheses if reported

Feather samples were soaked for 15 min in 1% residue-free detergent solution (Contrex, Decon Labs) and rinsed three times with deionized water before air drying in a laminar flow hood. Samples from 2006 and 2007 were weighed into 50-mL PFA vials (13–57 mg), then digested and analyzed according to the same protocol as for blood samples. Samples from 2008 and 2009 were washed and dried, and 3- to 10-mg subsamples were analyzed for Hg using a Milestone DMA-80 direct mercury analyzer (USEPA method 7473). Practical quantitation limits for Hg for both methods were >0.2 mg kg−1 in dry feathers depending on the sample mass. Standard reference materials (DORM-3, National Research Council Canada) were included with each of the two methods (n = 3 and n = 5, respectively) with measured Hg concentrations between 0.34 and 0.42 mg kg−1, which is within the certified range of 0.32–0.44 mg kg−1, justifying the change of methods.

Fine-sediment samples were homogenized, and 0.4-g subsamples were digested using a standard procedure (EPA method 3050B), which involves digestion with HNO3, HCl, and H2O2 at increased temperature. Digests were brought to 40 mL before filtering and analysis for Hg (Leeman Hydra AF mercury analyzer [USEPA method 245.7]) and for As, Cd, Cu, Pb, and Zn using a Perkin Elmer Optima 5300 inductively coupled plasma-optical emission spectrometer (EPA method 200.7). Practical quantitation limits for Hg, As, Cd, Cu, Pb, and Zn in dry sediment were 0.0006, 2, 0.5, 0.6, 6, and 0.1 mg kg−1, respectively. Average recoveries for these elements in a standard reference material (SRM 2710 “Montana Soil,” National Institute of Standards and Technology; n = 5, SD in parentheses) were 110% (8%), 93% (1%), 112% (3%), 99% (2%), 84% (2%), and 92% (5%), respectively.

Statistical Analyses

We constructed linear least squares analysis of variance (ANOVA) models to examine factors that influenced the concentrations of elements in the blood of osprey chicks. The concentration of each element was used as the response variable; sampling year, chick mass, and river sections were modeled as fixed effects. Because chicks from the same nest are not statistically independent sample units, measurements of individual chicks were nested within their nest identification number. The nests we sampled were treated as random effects in the model. We defined different river sections on the basis of natural hydrological features that would be potentially important in influencing levels of contaminants. These river sections were as follows:
  • Section A: Main stem CFR from Warm Springs to above the confluence of Flint Creek with the CFR at Drummond. This section of the CFR mainly had mining contaminants from the Butte–Anaconda smelting operation. Hg was not used in the smelting of these ores.

  • Section B: Main stem CFR from Drummond downstream to the site of the former Milltown Reservoir just above Missoula, upstream from the confluence with the large tributaries BFR and BRR. The Flint Creek drainage had extensive historic gold and silver mining and smelting operations, which included the use of large amounts of Hg. Much of the toxic mining sediments from the upper CFR were deposited in Milltown Reservoir, which forms the downstream end of the UCFRB Superfund Complex.

  • Section C: Main stem CFR below Milltown Reservoir from Missoula and downriver to Tarkio, Montana. Most nests in this section are located in the Missoula Valley with short distances to uncontaminated tributaries and artificial ponds.

  • Section D: Reference rivers. These included the LBF, BFR, and BRR.

In cases where there were significant effects in the models, we made post hoc comparisons of effect means with Tukey’s honestly significant difference (HSD) test with α = 0.05. This test adjusts for multiple statistical comparisons. Statistical analyses were conducted with JMP 9 (SAS, Chicago, IL).

Results

Sediments

Osprey habitats in the UCFR watershed bracket a wide range of mining-related contaminant levels. The highest sediment concentrations of As, Cd, Cu, Pb, and Zn were measured in the upper section of the CFR (section A) between Warm Springs and Drummond (Figs. 1, 2), where average concentrations and ranges for these elements in mg kg−1 were 108 (range 68–167), 4.4 (range 0.6–6.9), 971 (range 215–1,575), 127 (range 38–185), and 883 (range 241–1,173), respectively. Concentrations in mg kg−1 between Drummond and Missoula (section B) were 70 (range 41–102), 3.3 (range 2.0–5.8), 459 (range 322–548), 89 (range 61–112), and 749 (range 589–815), respectively. Concentrations decreased further in the Missoula Valley and downstream (section C) to 29 (range 18–39), 1.5 (range 0.95–2.4), 266 (range 164–391), 52 (range 30–63), and 536 (range 321–747) mg kg−1, respectively. The lowest concentrations of As, Cd, Cu, Pb, and Zn were measured in reference rivers LBF, BFR and BRR (section D): 13 (range 4.2–31), 0.32 (range 0.20–0.55), 20 (range 12–34), 20 (range 16–26), and 79 (range 42–150) mg kg−1, respectively.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-011-9732-5/MediaObjects/244_2011_9732_Fig2_HTML.gif
Fig. 2

Contaminant concentrations in fine sediment along the main stem CFR and tributaries trending downward in downstream direction for As, Cu, Zn, Cd, and Pb. The increase in Hg concentrations near Drummond is due to contributions from Flint Creek. The most upstream sample was collected in Warm Springs Ponds where concentrations were not representative of in-stream sites

Although not inhabited by ospreys, sediment concentrations in lower Flint Creek, a tributary to the CFR near Drummond, are shown in Fig. 2. This stream drains a basin of intensive historic mining, resulting in highly increased levels of As, Pb, and especially Hg. Our sediment samples from lower Flint Creek contained between 9.7 and 25 mg kg−1 total Hg, and sediment concentrations in the main stem CFR increased sixfold from approximately 0.7 mg kg−1 above Flint Creek to 4.4 mg kg−1 at a sampling site 1 km below the mouth of Flint Creek (Fig. 2). Average Hg concentrations in section A were 0.80 (range 0.15–1.2) mg kg−1, increasing in section B to 2.2 (0.96–4.4) mg kg−1, and decreasing in section C to 0.46 (0.32–0.71) mg kg−1. Hg concentrations in reference rivers LBR, BFR, and BRR (section D) were 0.12, 0.04, and 0.02 mg kg−1, respectively, which are all lower than conventional threshold effect levels of approximately 0.17 mg kg−1, below which no detrimental effects on the benthic community have been observed (MacDonald et al. 2000; Yao and Gao 2007). Total sediment Hg concentrations in the BFR and BRR may be considered background levels (Andersson 1979), although none of these watersheds are completely free of historic precious metals mines and atmospheric deposition (USEPA 2008).

Overview of Nests, Numbers of Chicks, and Sampling Years

From 2006 to 2009, osprey nestlings were sampled from 36 nests throughout the UCFRB (Fig. 1; Table 1), with 3 of the nests sampled during 4 years, 2 nests during 3 years, and 11 nests during 2 years, resulting in bird samples from 60 nest visits. Sampled nests contained 1 (19 nest visits), 2 (20 nest visits), 3 (18 nest visits), or 4 (3 nest visits) live chicks. Ten of the nests were located along tributaries at least 5 km from the contaminated main stem of the CFR. Nestlings from these sites were treated as references unlikely to have consumed fish from the CFR. The remaining osprey nests were located in the vicinity of the CFR main stem, with adults commonly observed fishing in the CFR; however, alternative water bodies, such as smaller tributaries or ponds, exist in the watershed where parent ospreys may have obtained fish to feed their young. Blood samples were obtained from 111 of the 127 osprey nestlings found in nests; sampling was sometimes decreased to 1 or 2 chicks/nest to shorten the time of disturbance and to minimize stress on the young birds.

Blood

Concentrations of contaminants in the blood and feathers of individual birds are listed in Table 1 and shown in Fig. 3. Statistical summaries of the ANOVAs for the different elements are listed in Table 3. Blood Pb concentrations were generally low, i.e., <20 μg L−1 in all 28 birds from reference nests and from the majority of birds (77 of 83) from the main stem of the CFR (Table 1; Fig. 3). Concentrations in the remaining six birds from the main stem CFR were between 20 and 50 μg L−1. There was a significant difference in Pb concentrations between years (Table 3), with concentrations in 2008 being significantly higher than in all other years (least square mean values for years in μg L−1 were 2006 = 4.8, 2007 = 7.6, 2008 = 11.1, and 2009 = 4.9 [2008 concentrations significantly higher than in all other years], p < 0.05 by Tukey HSD comparisons). Pb concentrations also varied significantly by river section (Table 3), with concentrations in the reference rivers being significantly lower than all sections of the CFR (least square mean values for river sections in μg L−1 were A = 8.89, B = 9.44, C = 6.45, and D = 3.62 [mean concentrations in D is significantly lower than all other sections], p < 0.05 by Tukey HSD).
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-011-9732-5/MediaObjects/244_2011_9732_Fig3_HTML.gif
Fig. 3

Contaminant concentrations in whole blood of osprey chicks along the main stem CFR (solid circles) and tributaries (open circles). Labels refer to sections A, B, and C used for statistical analysis. The combined tributaries (right panel) represent section D. Symbols show averages for each nest during all years based on average concentrations for each clutch. Error bars show the minimum and maximum concentrations measured in any chick from each site. The sudden increase of Hg concentration in section B coincides with large inputs from Flint Creek

Table 3

Summary statistics for fixed effects in least squares models based on whole-blood samples (see text)

Effect

Arsenic

Copper

Lead

Zinc

Mercury

 

F ratio

Prob >F

F ratio

Prob >F

F ratio

Prob >F

F ratio

Prob >F

F ratio

Prob >F

Year

0.0069

0.93

24

<0.01

7.8

<0.01

5.3

<0.01

0.27

0.84

Chick mass

0.0027

0.96

0.84

0.36

0.59

0.45

3.3

0.07

0.49

0.48

River section

9.0

<0.01

4.4

0.0065

5.1

<0.01

2.8

0.05

48

<0.01

Cd levels in blood were lower than the detection limit of 5 μg L−1 in all samples. For this reason no statistical analyses were performed for Cd.

Cu concentrations in blood varied significantly between sampling years and between different river sections (Table 3). Osprey chicks in section A had significantly higher blood Cu concentrations than all other sections (least square mean values in μg L−1 were A = 298, B = 257, C = 278, and D = 270 [section A significantly higher than sections B, C, and D, whereas B, C and D were not significantly different by Tukey HSD test]). In addition, Cu concentrations were significantly higher overall in 2008 than in other years (least square mean values in μg L−1 were 2006 = 263.5, 2007 = 264.8, 2008 = 328.3, and 2009 = 249.2 [2008 significantly higher than other years at p < 0.05 by Tukey HSD comparisons, but 2006, 2007 and 2009 were not significantly different at p > 0.05 by Tukey HSD comparisons]).

Zn concentrations in blood were similar in samples across all river sections (Table 3; least square mean values for river sections in μg L−1 were A = 3,310, B = 3,070, C = 2,990, and D = 3,099 [all comparisons p > 0.05 by Tukey HSD test]), and Zn concentrations did vary significantly between years (Table 3), but the difference was small. Mean Zn concentrations were lowest in 2006, increased in 2007 and 2008, and then decreased slightly in 2009 (least square mean values for years in μg L−1 were 2006 = 2,804, 2007 = 3,180, 2008 = 3,340, and 2009 = 3,138 [only 2006 and 2008 are significantly different from each other by Tukey HSD tests]).

As concentrations in blood varied widely, ranging between <4 and 348 μg L−1 for individual bird samples (Table 1; Fig. 3). As levels did not vary significantly between years (least square mean values in μg L−1 were 2008 = 33.9 and 2009 = 34.8 [p > 0.05]) or in relation to chick mass (Table 3 [F = 0.0027, p > 0.05]) but were highly affected by location. The mean level of As was an order of magnitude greater in the CFR section A than in all other river sections (least square mean values for river sections in μg L−1 were A = 100, B = 17, C = 7.6, and D = 12 [section A is significantly higher than all other sections, p < 0.05 by Tukey HSD comparisons]). The level of As in chick blood generally reflects the concentrations of As in the nearby river sediments. In section A, samples from eight chicks, all originating from the most upstream nests, exceeded 37 μg L−1 of As in blood. This is where As concentrations in the sediments are most increased, containing >125 mg kg−1 (Fig. 2). Chicks from the BFR and BRR contained low As levels, ranging from <4–15 μg L−1. These low levels in blood corresponded to the lowest sediment As levels in the watershed of ≤5 mg kg−1. Chicks in the nest located on the LBF had between 40 and 50 μg L−1 As, whereas sediment levels were somewhat increased at 31 mg kg−1.

Total Hg concentrations in blood varied more than an order of magnitude between 97 and 725 μg L−1 among all nestlings (Table 1; Fig. 3). Variation in Hg concentrations differed significantly between locations (p < 0.0001), with the concentrations spiking in the CFR below the confluence with Flint Creek (least square mean values for river sections in μg L−1 were A = 151, B = 548, C = 303, and D = 206 [concentration in A is significantly lower than all other sections; concentration in B is significantly higher than all other sections; and the intermediate concentrations in sections C and D do not differ significantly from each other; p < 0.05 by Tukey HSD comparisons]). There was no significant difference in Hg concentrations between sampling years. Concentrations exhibited impressive consistency among birds from the same nest, with blood samples from siblings of the same clutch differing by 9% (±8% SD [n = 39]) on average, only slightly more than the analytical method-related variability of 6% (±5% [n = 34]), which was obtained by comparing split samples from the same birds.

Although there was no significant effect of chick mass on Hg concentration in the blood, the large variation in Hg concentrations in different river sections may mask some important biological relationships. We examined these relationships separately for the four different river sections (Fig. 4). In the sections where Hg concentrations are generally high (B and C), there is no significant relationship between Hg concentration and chick mass (B: r2 = 0.000103 [p = 0.97] and C: r2 = 0.0008 [p = 0.85]). In the sections where Hg concentrations were the lowest (A and D), there were significant negative relationships between Hg concentration and chick mass (A: r2 = 0.191 [p = 0.041] and D: r2 = 0.333 [p = 0.0016]).
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-011-9732-5/MediaObjects/244_2011_9732_Fig4_HTML.gif
Fig. 4

Relationship between blood Hg concentration and mass of nestling osprey chicks. Significant negative relationships exist only in sections A and D where Hg levels are lowest

Feathers

Feather samples from birds that had fully developed nape feathers were analyzed for Hg during all 4 years of the study (Table 1). Concentrations varied widely between 0.95 and 15 mg kg−1 dry mass. Osprey chicks in section A had the lowest concentrations, with increasingly higher concentrations occurring in sections D, C, and B (respective means ± SDs in mg kg−1 were 2.6 ± 0.89 [n = 10]; 3.9 ± 1.4 [n = 13]; 5.5 ± 1.7 [n = 25]; and 7.6 ± 4.0 [n = 6], respectively). All sections were significantly (p ≤ 0.01) different from each other.

Discussion

Priority Contaminants

The designation of the UCFRB as a Superfund area by the USEPA (2004) was based primarily on increased levels of As, Cd, Cu, Pb, and Zn associated with mine wastes lining the CFR flood plain. The environmental impact of these priority elements is prominent throughout the drainage because they cause damage to riparian vegetation, fish disease, contamination of ground water, and other environmental problems. However, the concentrations of these elements and their biological effects have not been documented in high trophic-level species, such as osprey along the CFR.

We found that blood and feather samples from osprey nestlings reflect the level of contamination for some elements but not others. Cd was undetectable in blood and feathers throughout the study area, with concentrations <5 μg L−1 or 0.3 mg kg−1, respectively. This is similar to results reported for ospreys along the Coeur D’Alene River, Idaho, where most chicks had blood Cd levels <5 μg L−1, and those exceeding this value were not necessarily associated with contaminated reaches (Henny et al. 1991). In a study on other wild bird species, blood accounted for only 0.5% of the total body burden of Cd, whereas kidneys and liver contained >92% of total Cd. These organs are considered more suitable indicators of environmental Cd exposure in birds (Scheuhammer 1987; Garcá-Fernández et al. 1996). Cu in osprey blood was significantly increased in the most upstream reach of the CFR (Fig. 3; Table 3 [section A]) compared with all other reaches; however, the difference in osprey blood Cu between the reaches was relatively small compared with the order-of-magnitude gradient in sediment Cu (Fig. 2).This is consistent with other studies showing that Cu concentrations in mining-affected rivers approached normal levels with increasing trophic level (Farag et al. 1998; Woodward et al. 1994; Moore et al. 1991; Cain et al. 1992). The effect of river section on blood Zn concentration was statistically insignificant, reflecting the ability of aquatic birds to regulate Zn levels within certain exposure thresholds (Eisler 1993; Wiemeyer et al. 1987) (Table 3). Blood Pb concentrations were dependent on the river section (Table 3). Most birds throughout the study area had blood Pb levels <20 μg L−1, but six birds from the main stem CFR had blood Pb levels between 20 and 50 μg L−1. These Pb levels would be considered marginally toxic to humans (Gilbert and Weiss 2006) and are similar to concentrations found in ospreys in the Idaho’s Coeur d’Alene River watershed where no adverse effects on osprey reproduction were identified (Henny et al. 1991). Although Pb poisoning poses a major concern for raptors that feed on hunter-killed water fowl or carcasses (Church et al. 2006; Craighead and Bedrosian 2008), this contaminant is not appreciably transmitted up the food chain and is unlikely to directly affect the local osprey population. As levels in blood varied widely between individual birds. The highest blood levels were found exclusively in the most upstream nests of the main stem CFR (37–348 μg L−1 in C02, C03, C04), coinciding with the highest concentrations in sediments. Coincidentally, the bird with the highest blood As concentration (348 μg L−1) was resampled after 5 days, at which time its blood As level had decreased to only 51 μg L−1, which is consistent with rapid excretion of the element. We are unaware of attempts to determine toxicity levels for As in ospreys; however, results from feeding trials with domestic birds (Hermayer et al. 1977) and tissue analyses of other raptors (Erry et al. 1999) and seabirds (Savinov et al. 2003) suggest that toxicity thresholds for As are higher than the concentrations observed in this study.

ANOVA also showed interyear differences of several metal contaminants (Table 3). Cu, Zn, and Pb concentrations were significantly higher in 2008 than during other years. During that year, the transport of suspended heavy metals in the CFR was also highest due to relatively high spring runoff and breach of the Milltown Dam upstream of Missoula (Sando and Lambing 2011). This suggests that during times of high suspended contaminant load, the contamination of prey fish also increases, together with measurable effects on their predators. No temporal effects were showed for As, possibly due to lower sample numbers and a relatively high importance of other sources of variability. None of the priority contaminants exhibited a dependence on the mass of the birds as a proxy of their age (Table 3).

Spatial Distribution of Blood Hg

The spatial distribution of blood Hg concentrations in nestling ospreys deviates from other mining-related contaminants, reflecting the different environmental sources and biogeochemical behavior of this element. Osprey chicks growing up in the most upstream reach of the CFR (C01–C05 [section A]) were consistently among the lowest in blood Hg concentration, with values varying between 97 and 240 μg L−1 (average ± SD 151 ± 49 μg L−1 [n = 13], Fig. 5), although total Hg concentrations in river sediment were higher compared with the most downstream section of the CFR and tributaries (sections C and D [Fig. 2]). We hypothesize that net methylation rates and the associated accumulation in the upper trophic levels are lower in section A than elsewhere in the watershed. This is consistent with previous observations for mountainous river systems (Krabbenhoft et al. 1999) and may be a compounding product of low temperatures, short growth periods for vegetation, and fast-flowing streams with few anaerobic wetlands. We speculate that Hg methylation in section A may also decrease due to widespread metals contamination, resulting in diminished microbial activity and less riparian vegetation along the upper CFR.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-011-9732-5/MediaObjects/244_2011_9732_Fig5_HTML.gif
Fig. 5

Total Hg concentrations in whole blood and feathers of nestling ospreys. Bars correspond to nest sites along river sections of the main stem CFR (sections A, B, C) and tributaries (section D). All bars are statistically different from the others at the 95% CI level for both blood and feathers

Blood Hg concentrations increased dramatically in the section between Drummond and the upstream end of the Missoula Valley (C06–C10, section B, 548 ± 113 μg L−1, n = 6, p < 0.01), coinciding with a several-fold increase in total Hg in sediment downstream from Flint Creek (Fig. 2). This suggests that the gross rate of methylmercury formation and uptake into biota in the CFR is a function of total Hg supply to the system. Consequently, elimination of Hg sources in the Flint Creek drainage could significantly decrease the amount of Hg in the biota of the lower CFR ecosystem.

Compared with section B, blood Hg levels decreased significantly within the Missoula Valley and below (C11–C26, section C, 303 ± 83 μg L−1, n = 28, p < 0.01). Sediment Hg concentrations decreased as well, due in part to mixing with uncontaminated alluvium and the addition of sediment from two large tributaries, BFR and BRR.

Blood Hg levels from reference rivers (LBF, BFR, and BRR [section D]) were intermediate (206 ± 55 μg L−1 [n = 13]), even although total sediment Hg concentrations were one to two orders of magnitude lower than in the main stem CFR. Concentrations were lower than in sections B and C, and they were higher than in section A (Fig. 5). The highest blood Hg levels in section D were observed in nests of the BRR drainage (227 ± 50 μg L−1 [n = 9], Fig. 3), which may be due to their location within or downstream of the Lee Metcalf National Wildlife Refuge. This refuge includes >10 km2 of wetlands where individual sections are artificially flooded and dried periodically for water bird management, likely enhancing methylation rates and Hg biomagnification within the aquatic food webs. Much of the total Hg supply in the tributaries may be from atmospheric deposition of globally and regionally redistributed Hg, which is estimated to be 8–11 g km−2 annually throughout the watershed (model results from USEPA 2008). Hg from atmospheric deposition has been shown to support methylmercury levels in the biota comparable with local anthropogenic sources (Marvin-DiPasquale et al. 2009) and appears to be more prone to methylation than the Hg that has been in the aquatic system for longer durations (Harris et al. 2007).

Temporal Variability of Blood Hg

Our statistical analysis showed no significant differences between blood Hg concentrations during several years within the same nests (Table 3). However, there were differences in interyear variability among reaches. The largest range in blood Hg concentrations between years occurred along the central section of the CFR, from site C09 to the downstream end of the Missoula Valley (C23), where nest averages of consecutive years differed by 47 ± 42% (average ± SD [n = 11]). A common feature among these sites is their vicinity to less contaminated tributaries, thereby providing alternative sources of prey fish. Site C09 is located 3 km upstream from pristine Rock Creek, and ospreys in the Missoula Valley have access to several low-Hg water bodies, such as the BFR, BRR, and a number of artificial ponds. Anecdotal observations suggest that the ospreys use these alternatives, especially during periods when high turbidity impairs their ability to hunt in the main stem CFR. Throughout the remainder of the study area, the average difference between nests sampled during consecutive years was 17 ± 20% (n = 12). This range is surprisingly narrow considering that our 4-year study included a considerable range of flow events, including a 10-year flood and a substantial disturbance due to the removal of Milltown Dam (USEPA 2004). This suggests that methylmercury levels in the local prey fish communities of the UCFRB are relatively stable between years and that ospreys mirror these levels closely. In addition, we conclude that ospreys are most suitable as biomonitors for methylmercury exposure in local food webs where alternative water bodies are ≥5 km away.

Based on their extensive study, DesGranges et al. (1998) suggest that age of nestlings does not strongly affect Hg levels in blood. The blood Hg concentrations in osprey chicks in their study mirrored levels in prey fish, and the investigators speculated that a dynamic balance exists between methylmercury uptake with fish and rapid removal from the bloodstream through deposition into growing feathers. Our data agree in that no correlation between chick mass and blood Hg concentration is present when the complete data set is considered or in river sections where blood Hg levels in most chicks exceed 300 μg L−1 (sections B and C [Fig. 4]). However, in the river sections where Hg levels are generally <300 μg L−1 (sections A and D), highly significant negative relationships emerge between chick mass and blood Hg concentration. We suggest that at this lower, more natural Hg exposure, the rate at which growing feathers sequester Hg from the bloodstream is high enough to decrease the relatively low concentration of Hg in the blood and outstrip the rate of dietary addition of Hg into the blood. For osprey chicks exposed to relatively low Hg levels, this would explain the negative relationship between blood Hg concentration and body mass. In contrast, when blood Hg concentrations and the rate of addition of new Hg into the blood exceed some thresholds, the physiological rate of Hg removal from the blood into developing feathers is exceed by the rate of dietary addition of Hg into the blood. This would obscure any relationship between Hg levels in the blood and body mass. Several other observations are consistent with this physiological rate hypothesis. We obtained samples from one nestling 5 days apart: Its mass had increased from 1,044 to 1,209 g, and its feathers were still rapidly growing. Its blood Hg level had decreased slightly from 170 to 156 μg L−1 (Site C03 in 2008 [Table 1]). Another osprey chick was injured and was recaptured 1 month after fledging and after its feathers had stopped growing (Site C24, 2009 [Table 1]). Its blood Hg level had increased more than sevenfold, from 387 to 2,790 μg L−1, since being sampled as a nestling. These anecdotal observations are consistent with earlier findings for piscivorous water birds (DesGranges et al. 1998) and songbirds (Condon and Cristol 2009) noting a rapid increase in concentrations of blood Hg once body and feather growth has ceased.

Blood Hg concentrations in osprey nestlings could potentially be affected by Hg imported from the female parents’ wintering ranges and transferred through the eggs. We collected data opportunistically from six addled eggs collected from nests throughout the study area in 2008 and 2009, which contained 31, 34, 43, 52, 126, and 134 μg kg−1 Hg (wet weight, adjusted for loss of moisture). Considering a tenfold to 20-fold decrease of these levels due to dilution during growth, in addition to a disproportionally high Hg deposition in feathers, the amount of Hg transmitted through eggs is likely insignificant in blood samples of chicks, which is consistent with earlier findings (DesGranges et al. 1998; Elliott et al. 2000, 2007).

Hg in Feathers

Total Hg concentrations in feathers were generally higher than in blood (Table 1; Fig. 5) when compared on a live-weight basis, which was expected due to the sequestration of Hg in keratinized tissues (DesGranges et al. 1998; Rattner et al. 2008). Concentrations showed a similar geographic pattern as whole-blood samples, which is consistent with previous findings correlating feather and blood Hg concentrations (Rattner et al. 2008; Golden et al. 2003). The lowest concentrations were measured in section A and the highest in section B, with sections C and D having intermediate but significantly different levels (Fig. 5). Although geographic differences were significant for feathers, the analytical results were much more variable for feathers than for whole blood. Different feathers from the same bird differed on average by 16% (n = 4), which is similar to the difference between tips and bases of identical feathers (17% [n = 4]). This variability among duplicate feather samples was much higher than the difference among duplicate blood samples (6% [n = 34]). Therefore, blood samples appear to be better suited for monitoring Hg exposure in nestling ospreys. The use of feather samples for the monitoring of contaminants in ospreys is further complicated by the short time window available for collection of feathers from nestlings. The period between sufficient development of cover feathers and risk of chicks fledging prematurely on approach by human intruders may be <2 weeks (week 5 and 6), whereas the sampling of blood may start 2 weeks earlier.

Conclusion

The concentrations of As, Cu, Hg, Pb, and Zn varied in the blood of nestling osprey chicks in response to spatial contamination gradients in river sediments of the UCFRB. Levels of Cu, Pb, and Zn also increased during a year of high spring runoff, presumably in response to increased suspended contaminant loads. However, with the exception of Hg, these responses were small or highly variable, decreasing the suitability of osprey blood sampling for the monitoring of changes in the contaminant status of rivers. Hg concentrations in blood and feathers were primarily a function of the geographic location of the nests, with impressively low variability between chicks and during several years, especially where ospreys are presumed to obtain all of their prey fish from one river. The highest Hg levels in ospreys were associated with inputs of Hg-rich sediment from a contaminated tributary, emphasizing the overriding importance of total Hg inputs to the system. Consequently, the removal of sediment-bound Hg inputs from tributaries may significantly decrease Hg levels in aquatic food webs of the UCFRB. At low to intermediate sediment Hg concentrations, the Hg levels in osprey tissues appeared to be less dependent on levels in sediment than on the environmental methylation potential associated with wetland abundance and local climate. Hg concentrations in feathers are generally more variable than in blood samples, but both show similar geographic dependencies. We found the sampling of blood from nestling ospreys to be a simple and sensitive method of monitoring bioavailable Hg in food webs of a mining-affected river system. More research will be necessary to link blood Hg concentrations in osprey nestlings to actual environmental toxicity thresholds because blood concentrations are the steady state result of several sources and sinks and are not representative of levels in fledged ospreys or other predators. In the more contaminated river reaches, blood Hg concentrations were independent of chick mass, but concentrations in less contaminated reaches were found to decrease with chick mass, likely as a result of disproportionally high Hg deposition in growing feathers.

Acknowledgments

We thank the Montana State Natural Resource Damage Program and the Environmental Biogeochemistry Laboratory at the University of Montana Geosciences Department for partial funding of this research. Special thanks go to Dave Taylor Roofing of Missoula, MT, Sam Milodragovich and Roy Brunner of Northwestern Energy Corporation, and Bart Peterson of Missoula Electric Cooperative for making their boom trucks available. We also thank personnel from Grant-Kohrs Ranch National Historic Site, Atlantic Richfield Company, Tammy Talley, and Riverside Health Care Center, as well as the many private land owners who graciously gave us access to osprey nests. Kate Davis of Raptors of the Rockies, Karen Wagner, and an anonymous donor also helped facilitate this research in various ways. We thank three anonymous reviewers for their thorough and thoughtful comments, which greatly improved an earlier draft of this manuscript.

Copyright information

© Springer Science+Business Media, LLC 2011