U.S. Army
Corps of Engineers
New England District
Concord, Massachusetts
^eDS%
i Q ' s
\¦
^ PRO-CS-0
U.S. Environmental
Protection Agency
New England Region
Boston, Massachusetts
ECOLOGICAL RISK ASSESSMENT FOR
GENERAL ELECTRIC (GE)/HOUSATONIC RIVER SITE,
REST OF RIVER
Volumes 1 and 2
Sections 1-12
DCN: GE-070703-ABRC
July 2003
Environmental Remediation Contract
GE/Housatonic River Project
Pittsfield, Massachusetts
Contract No. DACW33-00-D-0006
Task Order 0003
03P-0966-1
SOLUTIONS
-------
ECOLOGICAL RISK ASSESSMENT FOR
GENERAL ELECTRIC (GE)/HOUSATONIC RIVER SITE
REST OF RIVER
VOLUMES 1 AND 2
SECTIONS 1-12
ENVIRONMENTAL REMEDIATION CONTRACT
GENERAL ELECTRIC (GE)/HOUSATONIC RIVER PROJECT
PITTSFIELD, MASSACHUSETTS
Contract No. DACW33-00-D-0006
Task Order No. 0003
DCN: GE-070703-ABRC
Prepared for
U.S. ARMY CORPS OF ENGINEERS
New England District
Concord, Massachusetts
and
U.S. ENVIRONMENTAL PROTECTION AGENCY
New England Region
Boston, Massachusetts
Prepared by
WESTON SOLUTIONS, INC.
West Chester, Pennsylvania
July 2003
Work Order No. 20123.001.096.0733
MK01 |O:\20123001.096\ERA_PB\ERA_PB_TITLEPAGE_VOL 1-2.DOC
-------
AUTHORS/CONTRIBUTORS
Tod Delong, Florence Sevold
Avatar Environmental, LLC
West Chester, Pennsylvania
Scott Ferson, Troy Tucker
Applied Biomathematics, Inc.
Setauket, New York
Dwayne Moore, Roger Breton, Drew McDonald, Andrew Pawlicz, Scott Teed, Ryan Thompson
The Cadmus Group, Inc.
Ottawa, Ontario
Gary Lawrence, Chessy Langford
EVS Environmental Consultants
Vancouver, British Columbia
Rich DiNitto
Sleeman, Hanley & DiNitto
Boston, Massachusetts
Alice Shelly
TerraStat
Seattle, Washington
Susan Svirsky
U.S. Environmental Protection Agency
Boston, Massachusetts
Dick McGrath, Scott Campbell
Weston Solutions, Inc.
West Chester, PA; Pittsfield, Massachusetts
John Lortie, Bob Roy, Michael Thompson, Chris Werner
Woodlot Alternatives, Inc.
Topsham, ME
MK01 |0:\20123001.096\ERA_PB\ERA_PB_TITLEPAGE_VOL 1-2 (PDFONLY).DOC 7/11/2003
-------
1
TABLE OF CONTENTS
2 Section Page
3 ES. ECOLOGICAL RISK ASSESSMENT EXECUTIVE SUMMARY ES-1
4 ES.l OVERVIEW ES-1
5 ES.2 SITE DESCRIPTION AND HISTORY ES-2
6 ES.3 REGULATORY BACKGROUND ES-7
7 ES.4 OVERVIEW OF TECHNICAL APPROACH ES-8
8 ES.5 OVERVIEW OF THE ASSESSMENT ENDPOINT CONCLUSIONS ES-12
9 ES.5.1 Risks in the Primary Study Area ES-12
10 ES.5.2 Risks Downstream of the Primary Study Area ES-33
11 ES.6 BROADER IMPLICATIONS ES-43
12 ES.6.1 Implications for Other Species in the Primary Study Area ES-43
13 HS.7 SOURCES OF UNCERTAINTY ES-48
14 ES.8 SUMMARY AND CONCLUSIONS ES-50
15 1. INTRODUCTION 1-1
16 1.1 OVERVIEW 1-1
17 1.2 SITE HISTORY 1-5
18 1.3 REGULATORY BACKGROUND 1-9
19 1.4 SITE DESCRIPTION 1-11
20 1.5 OVERVIEW OF TECHNICAL APPROACH 1-20
21 1.5.1 Problem Formulation 1-23
22 1.5.2 Assessment of Representative Species 1-27
23 1.6 DATA SOURCES 1-28
24 1.7 QA/QC 1-29
25 1.8 REFERENCES 1-30
26 2. PROBLEM FORMULATION 2-1
27 2.1 OVERVIEW 2-1
28 2.2 PHYSICAL AND ECOLOGICAL CHARACTERIZATION OF THE
29 HOUSATONIC RIVER 2-5
30 2.2.1 Physical Characteristics of the Housatonic River Basin 2-5
31 2.2.2 Ecological Characterization of the Study Area 2-6
32 2.3 IDENTIFICATION AND SOURCES OF STRESSORS 2-10
33 2.3.1 Contaminant Stressors 2-10
34 2.3.2 Physical and Biological Stressors 2-11
O:\20123001.096\ERA_PB\ERA_PB_TOC.DOC • • 7/12/2003
-------
TABLE OF CONTENTS
(Continued)
Section Page
1 2.4 OVERVIEW OF PRE-ERA 2-14
2 2.4.1 Introduction 2-14
3 2.4.2 Data 2-14
4 2.4.3 Primary Study Area (PSA) Evaluation and Results 2-16
5 2.4.4 PCB Screening Evaluation Downstream of Woods Pond and Results 2-17
6 2.5 FATE AND TRANSPORT OF CONTAMINANT STRESSORS 2-23
7 2.5.1 Fate and Transport of PCB s 2-23
8 2.5.2 PCB Distribution by Media 2-25
9 2.5.3 Identification of Exposure Pathways 2-39
10 2.5.4 Changes in PCB Congener Patterns 2-41
11 2.5.5 Fate and Transport of Dioxins/Furans 2-44
12 2.6 EFFECTS ON RECEPTORS 2-45
13 2.6.1 Polychlorinated Biphenyls (PCBs) 2-46
14 2.6.2 Dioxins/Furans 2-49
15 2.6.3 2,3,7,8-TCDD Toxic Equivalence (TEQ) 2-51
16 2.7 CONCEPTUAL MODEL 2-54
17 2.7.1 Exposure Pathways 2-55
18 2.8 SELECTION OF ASSESSMENT AND MEASUREMENT ENDPOINTS 2-58
19 2.8.1 Assessment Endpoints 2-58
20 2.8.2 Measurement Endpoints 2-60
21 2.9 WEIGHT-OF-EVIDENCE APPROACH TO ANALYSIS 2-66
22 2.10 EXTRAPOLATION OF RISK ESTIMATES FOR SELECTED ENDPOINTS
23 DOWNSTREAM 01 WOODS POND 2-72
24 2.11 REFERENCES 2-73
25 3. ASSESSMENT ENDPOINT—COMMUNITY STRUCTURE, SURVIVAL,
26 GROWTH, AND REPRODUCTION OF BENTHIC INVERTEBRATES 3-1
27 3.1 INTRODUCTION 3-1
28 3.1.1 Conceptual Model 3-7
29 3.2 EXPOSURE ASSESSMENT 3-11
30 3.2.1 Selection of COCs for Benthic Invertebrates 3-11
31 3.2.2 Types of Exposure Data 3-13
32 3.2.3 Habitat Characterization 3-15
33 3.2.4 Assessment of Sediment Chemistry 3-17
34 3.2.5 Tissue Chemistry Assessment 3-22
35 3.2.6 Surface Water Chemistry Assessment 3-24
36 3.3 EFFECTS ASSESSMENT 3-26
O:\20123001.096\ERA_PB\ERA_PB_TOC.DOC • • • 7/12/2003
-------
TABLE OF CONTENTS
(Continued)
Section Page
1 3.3.1 Sediment Toxicity 3-26
2 3.3.2 Concentration-Response Analysis - Toxicity Test Endpoints 3-45
3 3.3.3 Toxicity Identification Evaluations 3-52
4 3.3.4 Tissue PCB Effects Thresholds 3-53
5 3.3.5 Sediment Quality Values (SQVs) 3-54
6 3.3.6 Benthic Macroinvertebrate Community Evaluation 3-54
7 3.3.7 Concentration-Response Analysis - Benthic Community Assemblages 3-58
8 3.4 RISK CHARACTERIZATION 3-61
9 3.4.1 Field Surveys 3-62
10 3.4.2 Comparison of Chemistry Data to Benchmarks 3-63
11 3.4.3 Site-Specific Toxicity Study Results 3-66
12 3.4.4 Integrated Station-by-Station Assessment 3-69
13 3.4.5 Weight-of-Evidence (WOE) Procedure for Assessing Risk from PCBs in
14 the Housatonic River PSA 3-69
15 3.4.6 Sources of Uncertainty 3-74
16 3.4.7 Extrapolation to Other Species 3-77
17 3.4.8 Downstream Assessment 3-77
18 3.4.9 Conclusions 3-78
19 3.5 REFERENCES 3-80
20 4. ASSESSMENT ENDPOINT—COMMUNITY CONDITION, SURVIVAL,
21 REPRODUCTION, DEVELOPMENT, AND MATURATION OF
22 AMPHIBIANS 4-1
23 4.1 INTRODUCTION 4-1
24 4.2 CONCEPTUAL MODEL 4-8
25 4.2.1 Amphibian Developmental Studies 4-13
26 4.2.2 Leopard Frog Study: EPA 4-14
27 4.2.3 Wood Frog Study Design (EPA Studies) 4-14
28 4.2.4 Context-Dependent Wood Frog Study: GE 4-17
29 4.3 EXPOSURE ASSESSMENT 4-19
30 4.3.1 Selection of COCs for Amphibians 4-19
31 4.3.2 Exposure Data 4-20
32 4.3.3 Habitat Characterization 4-21
33 4.3.4 Assessment of Sediment Chemistry 4-22
34 4.3.5 Surface Water Chemistry Assessment 4-25
35 4.3.6 Tissue Chemistry Assessment 4-27
36 4.4 EFFECTS ASSESSMENT 4-31
37 4.4.1 Sediment Toxicity 4-32
O:\20123001.096\ERA PB\ERA PB TOC.DOC • 7/12/2003
- - IV
-------
TABLE OF CONTENTS
(Continued)
Section Page
1 4.4.2 PCB Effect Thresholds 4-58
2 4.5 RISK CHARACTERIZATION 4-61
3 4.5.1 Concentration-Response Analysis - Toxicity Test Endpoints 4-62
4 4.5.2 Biological Community Endpoints 4-64
5 4.5.3 Comparison of Tissue Chemistry Data to Benchmarks 4-65
6 4.5.4 Integrated Station-by-Station Assessment 4-69
7 4.5.5 Weight-of-Evidence Procedure for Assessing Risk from PCBs in the
8 Housatonic River PSA 4-69
9 4.5.6 Sources of Uncertainty 4-76
10 4.5.7 Conclusions 4-81
11 4.6 REFERENCES 4-85
12 5. ASSESSMENT ENDPOINT - SURVIVAL, GROWTH, AND
13 REPRODUCTION OF FISH 5-1
14 5.1 INTRODUCTION 5-2
15 5.1.1 Conceptual Model 5-3
16 5.2 EXPOSURE ASSESSMENT 5-10
17 5.2.1 Refined Screening of COPCs for Fish 5-10
18 5.2.2 Tissue Chemistry Assessment (Exposure to PCBs and TEQ) 5-11
19 5.2.3 Sediment Chemistry Assessment (Exposure to PAHs) 5-17
20 5.3 EFFECTS ASSESSMENT 5-21
21 5.3.1 Derivation of Literature Tissue Effects Metrics 5-22
22 5.3.2 Site-Specific Toxicity Studies 5-28
23 5.3.3 Concentration-Response Analysis - Toxicity Endpoints 5-39
24 5.3.4 Derivation of Literature-Based Sediment Effects Metrics for PAHs 5-45
25 5.4 RISK CHARACTERIZATION 5-48
26 5.4.1 Introduction 5-48
27 5.4.2 Field Surveys 5-49
28 5.4.3 Comparison of Estimated Exposures to Derived Effects Metrics 5-54
29 5.4.4 Site-Specific Toxicity Studies 5-68
30 5.4.5 Weight-of-Evidence Analysis 5-68
31 5.4.6 Sources of Uncertainty 5-71
32 5.4.7 Downstream Extrapolation 5-73
33 5.4.8 Risk Assessment Conclusions 5-74
34 5.5 REFERENCES 5-75
35 6. WILDLIFE ASSESSMENT HIGHLIGHTS 6-1
36 6.1 OVERVIEW 6-1
O:\20123001.096\ERA PB\ERA PB TOC.DOC 7/12/2003
- - v
-------
TABLE OF CONTENTS
(Continued)
Section Page
1 6.2 WILDLIFE EXPOSURE MODEL 6-7
2 6.2.1 Food Intake Rate (FIR) 6-8
3 6.2.2 Body Weight (BW) 6-9
4 6.2.3 Proportions of Dietary Items 6-9
5 6.3 SPATIAL AND TEMPORAL AVERAGING 6-10
6 6.4 TOXIC EQUIVALENCE (TEQ) 6-11
7 6.4.1 Non-Detects 6-15
8 6.4.2 Congener Co-Elution 6-15
9 6.4.3 Summary of Decision Criteria for Estimating Exposure Point
10 Concentrations 6-16
11 6.5 PROBABILISTIC RISK ASSESSMENT 6-18
12 6.5.1 Distribution Selection 6-18
13 6.5.2 Monte Carlo and Probability Bounds Analysis 6-19
14 6.6 EFFECTS ASSESSMENT 6-21
15 6.6.1 Dose-Response Relationships Using the Generalized Linear Model
16 F ramework 6-22
17 6.6.2 Hypothesis Testing to Determine LOAEL and NOAEL 6-23
18 6.6.3 Field-Based Measures of Effect and Threshold Ranges 6-24
19 6.7 RISK CHARACTERIZATION 6-25
20 6.7.1 Risk Categorization 6-25
21 6.7.2 Weight-of-Evidence Assessment 6-28
22 6.8 REFERENCES 6-31
23 7. ASSESSMENT ENDPOINT - SURVIVAL, GROWTH, AND
24 REPRODUCTION OF INSECTIVOROUS BIRDS 7-1
25 7.1 INTRODUCTION 7-2
26 7.2 CONCEPTUAL MODEL 7-8
27 7.3 EXPOSURE ASSESSMENT 7-12
28 7.3.1 Exposure Models for Insectivorous Birds 7-15
29 7.3.2 TDI Model Results 7-30
30 7.3.3 Microexposure Model Results 7-49
31 7.3.4 Tree Swallow Tissue Data from Field Study 7-62
32 7.4 EFFECTS ASSESSMENT 7-65
33 7.4.1 Review of Effects of tPCBs 7-68
34 7.4.2 2,3,7,8-TCDD and Equivalence (TEQ) 7-69
35 7.4.3 Tree Swallow Field Study 7-70
O:\20123001.096\ERA PB\ERA PB TOC.DOC • 7/12/2003
- - VI
-------
TABLE OF CONTENTS
(Continued)
Section Page
1 7.4.4 American Robin Field Study (GE) 7-71
2 7.4.5 Selection of Effects Metrics for Characterizing Risk 7-72
3 7.5 RISK CHARACTERIZATION 7-75
4 7.5.1 Comparison of Modeled Exposures to Effects 7-75
5 7.5.2 Tree Swallow Field Study 7-78
6 7.5.3 American Robin Field Study (GE) 7-78
7 7.5.4 Weight-of-Evidence Analysis 7-79
8 7.5.5 Sources of Uncertainty 7-86
9 7.5.6 Conclusions and Extrapolation to Other Species 7-88
10 7.6 REFERENCES 7-92
11 8. ASSESSMENT ENDPOINT—SURVIVAL, GROWTH, AND
12 REPRODUCTION OF PISCIVOROUS BIRDS 8-1
13 8.1 INTRODUCTION 8-1
14 8.1.1 Overview of Approach 8-2
15 8.2 CONCEPTUAL MODEL 8-8
16 8.3 EXPOSURE ASSESSMENT 8-12
17 8.3.1 Exposure Model 8-12
18 8.3.2 Exposure Model Results 8-19
19 8.4 EFFECTS ASSESSMENT 8-30
20 8.4.1 Total PCBs 8-31
21 8.4.2 TEQ 8-34
22 8.4.3 Effects Metrics 8-34
23 8.4.4 Belted Kingfisher Field Study 8-36
24 8.5 RISK CHARACTERIZATION 8-37
25 8.5.1 Comparison of Estimated Exposures to Laboratory-Derived Effect
26 Doses 8-37
27 8.5.2 Belted Kingfisher Field Study 8-38
28 8.5.3 Weight-of-Evidence Analysis 8-41
29 8.5.4 Sources of Uncertainty 8-46
30 8.5.5 Extrapolation to Other Species 8-48
31 8.5.6 Summary and Conclusions 8-48
32 8.6 REFERENCES 8-49
33 9. ASSESSMENT ENDPOINT - SURVIVAL, GROWTH, AND
34 REPRODUCTION OF PISCIVOROUS MAMMALS 9-1
35 9.1 INTRODUCTION 9-1
O:\20123001.096\ERA PB\ERA PB TOC.DOC •• 7/12/2003
- - Vll
-------
TABLE OF CONTENTS
(Continued)
Section Page
1 9.2 CONCEPTUAL MODEL 9-8
2 9.3 EXPOSURE ASSESSMENT 9-11
3 9.3.1 Exposure Model 9-12
4 9.3.2 Results of Exposure Assessments 9-21
5 9.4 EFFECTS ASSESSMENT 9-36
6 9.4.1 Review of Toxicity from the Literature 9-36
7 9.4.2 Mink Feeding Study 9-40
8 9.4.3 Effects Metrics for Characterizing Risk 9-46
9 9.5 RISK CHARACTERIZATION 9-49
10 9.5.1 Field Surveys 9-49
11 9.5.2 Comparison of Estimated Exposures to Laboratory-Derived Effects
12 Doses 9-52
13 9.5.3 Mink Feeding Study 9-68
14 9.5.4 Weight-of-Evidence Analysis 9-69
15 9.5.5 Sources of Uncertainty 9-74
16 9.5.6 Comparison to Other Piscivorous Mammals 9-77
17 9.5.7 Risk Downstream of PSA 9-77
18 9.5.8 Conclusions 9-78
19 9.6 REFERENCES 9-80
20 10. ASSESSMENT ENDPOINT—SURVIVAL, GROWTH, AND
21 REPRODUCTION OF OMNIVOROUS AND CARNIVOROUS MAMMALS 10-1
22 10.1 INTRODUCTION 10-1
23 10.2 CONCEPTUAL MODEL 10-8
24 10.3 EXPOSURE ASSESSMENT 10-11
25 10.3.1 Exposure Model 10-13
26 10.3.2 Results of Exposure Assessments 10-22
27 10.4 EFFECTS ASSESSMENT 10-32
28 10.4.1 Review of Effects of tPCBs and TEQ 10-33
29 10.4.2 Effects Metrics for Characterizing Risk 10-36
30 10.5 RISK CHARACTERIZATION 10-41
31 10.5.1 Field Surveys (Performed by EPA) 10-41
32 10.5.2 Comparison of Estimated Exposures to Laboratory-Derived Effects
33 Doses 10-43
34 10.5.3 Population Demography Field Study (Performed by GE) 10-53
35 10.5.4 Weight-of-Evidence Analysis 10-55
36 10.5.5 Sources of Uncertainty 10-61
O:\20123001.096\ERA PB\ERA PB TOC.DOC ••• 7/12/2003
- - Vlll
-------
TABLE OF CONTENTS
(Continued)
Section Page
1 10.5.6 Conclusions 10-63
2 10.6 REFERENCES 10-68
3 11. ASSESSMENT ENDPOINT—SURVIVAL, GROWTH, AND
4 REPRODUCTION OF THREATENED AND ENDANGERED SPECIES 11-1
5 11.1 INTRODUCTION 11-1
6 11.1.1 Overview of Approach 11-2
7 11.1.2 Conceptual Model 11-7
8 11.1.3 Organization 11-11
9 11.2 EXPOSURE ASSESSMENT 11-12
10 11.2.1 Exposure Model 11-13
11 11.2.2 Results of Exposure Assessments 11-23
12 11.3 EFFECTS ASSESSMENT 11-35
13 11.3.1 Total PCBs 11-36
14 11.3.2 2,3,7,8-TCDD Toxic Equivalence (TEQ) 11-39
15 11.3.3 Effects Metrics for Characterizing Risk 11-40
16 11.4 RISK CHARACTERIZATION 11-45
17 11.4.1 Field Survey 11-45
18 11.4.2 Comparison of Estimated Exposures to Laboratory-Derived Effects
19 Doses 11-47
20 11.4.3 Weight-of-Evidence Analysis 11-58
21 11.4.4 Sources of Uncertainty 11-63
22 11.4.5 Conclusions 11-65
23 11.5 REFERENCES 11-69
24 12. RISK SUMMARY 12-1
25 12.1 OVERVIEW 12-1
26 12.2 SUMMARY OF THE ASSESSMENT ENDPOINT CONCLUSIONS 12-4
27 12.2.1 Results of Weight-of-Evidence Evaluation 12-4
28 12.2.2 Hazard Quotient Analyses 12-17
29 12.2.3 Risk Assessment Downstream of Woods Pond 12-26
30 12.3 SPECIES SENSITIVITY AND MECHANISMS OF TOXICITY 12-43
31 12.3.1 Mechanism of Action and Sensitivity of Species to tPCBs and TEQ 12-43
32 12.3.2 Congener Composition and Toxicity to Biota 12-46
33 12.4 BROADER IMPLICATIONS 12-47
34 12.4.1 Implications for Other Species in the Primary Study Area 12-47
35 12.4.2 Ecological Implications and Other Concerns 12-69
O:\20123001.096\ERA PB\ERA PB TOC.DOC • 7/12/2003
- - IX
-------
TABLE OF CONTENTS
(Continued)
Section Page
1 12.5 SOURCES OF UNCERTAINTY 12-74
2 12.5.1 Problem Formulation 12-74
3 12.5.2 Exposure Assessment 12-75
4 12.5.3 Effects Assessment 12-76
5 12.5.4 Risk Characterization 12-78
6 12.6 ERA CONCLUSIONS 12-79
7 12.7 REFERENCES 12-80
8
O:\20123001.096\ERA PB\ERA PB TOC.DOC 7/12/2003
" " X
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
LIST OF TABLES
Title Page
Table ES-1 Evidence of Harm and Magnitude of Effects for Measurement Endpoints
Related to Maintenance of a Healthy Benthic Community ES-14
Table ES-2 Evidence of Harm and Magnitude of Effects for Measurement Endpoints
Related to Maintenance of Amphibian Populations in the Housatonic
River PS A ES-15
Table ES-3 Evidence of Harm and Magnitude of Effects for Measurement Endpoints
Related to Maintenance of a Healthy Fish Community ES-17
Table ES-4 Evidence of Harm and Magnitude of Effects for Insectivorous Birds
Exposed to tPCBs in the Housatonic River PSA ES-18
Table ES-5 Evidence of Harm and Magnitude of Effects for Insectivorous Birds
Exposed to TEQ in the Housatonic River PSA ES-18
Table ES-6 Evidence of Harm and Magnitude of Effects for Piscivorous Birds
Exposed to tPCBs in the Housatonic River PSA ES-20
Table ES-7 Evidence of Harm and Magnitude of Effects for Piscivorous Birds
Exposed to TEQ in the Housatonic River PSA ES-20
Table ES-8 Evidence of Harm and Magnitude of Effects for Piscivorous Mammals
Exposed to tPCBs in the Housatonic River PSA ES-22
Table ES-9 Evidence of Harm and Magnitude of Effects for Piscivorous Mammals
Exposed to TEQ in the Housatonic River PSA ES-22
Table ES-10 Evidence of Harm and Magnitude of Effects for Omnivorous and
Carnivorous Mammals Exposed to tPCBs in the Housatonic River PSA ....ES-24
Table ES-11 Evidence of Harm and Magnitude of Effects for Omnivorous and
Carnivorous Mammals Exposed to TEQ in the Housatonic River PSA ES-24
Table ES-12 Evidence of Harm and Magnitude of Effects of T&E Species Exposed to
tPCBs in the Housatonic River PSA ES-26
Table ES-13 Evidence of Harm and Magnitude of Effects for T&E Species Exposed to
TEQ in the Housatonic River PSA ES-26
Table 2.1-1 Surveys Conducted for Ecosystem Characterization and Their Specific
Objective(s) 2-3
Table 2.4-1 Sediment and Surface Water Data Categories 2-16
Table 2.4-2 COPCs for Sediment Based on Tier III Evaluation 2-18
Table 2.4-3 COPCs for Surface Water Based on Tier III Evaluation 2-20
Table 2.4-4 COPCs for Soil Based on Tier III Evaluation 2-21
Table 2.4-5 COPCs for Fish Based on Tier III Evaluation 2-22
O:\20123001.096\ERA PB\ERA PB TOC.DOC
XI
-------
LIST OF TABLES
(Continued)
Title Page
1 Table 2.6-1 Common Effects of PCB Exposure Observed in Various Animals 2-47
2 Table 2.6-2 TEF Values for Mammals, Fish, and Birds as Predators 2-53
3 Table 2.8-1 Ecological Assessment and Measurement Endpoints 2-62
4 Table 2.8-2 Summary of GE Ecological Studies 2-65
5 Table 2.9-1 Attributes for Judging Measurement Endpoints 2-68
6 Table 3.3-1 Results of Pairwise Statistical Tests Comparing Exposed Stations to
7 Negative Control (T-Ctrl) and Reference (Al, A3) Sediment (Water-Only
8 Exposures Excluded) 3-30
9 Table 3.3-2 In Situ Evaluation of Toxicity in Housatonic River Sediment (Station-by-
10 Station Assessment) 3-43
11 Table 3.3-3 Laboratory Evaluation of Toxicity in Housatonic River Sediment (Station-
12 by-Station Assessment) 3-44
13 Table 3.3-4 Evaluation of Lines of Evidence for Housatonic River Sediment Toxicity,
14 Relative to Reference Responses 3-46
15 Table 3.4-1 Weighting of Measurement Endpoints for Weight-of-Evidence Evaluation.. 3-73
16 Table 3.4-2 Evidence of Harm and Magnitude of Effects for Measurement Endpoints
17 Related to Maintenance of a Healthy Benthic Community 3-75
18 Table 3.4-3 Weight-of-Evidence Risk Analysis Summary Indicating Concurrence
19 Among Endpoints for Coarse-Grained and Fine-Grained Sediment 3-76
20 Table 4.4-1 Summary of Male Adult Leopard Frog Reproductive Health 4-37
21 Table 4.4-2 Summary of Female Adult Leopard Frog Reproductive Health 4-38
22 Table 4.4-3 Summary of Leopard Frog Larval Development Endpoints Data at End-of-
23 Test 4-41
24 Table 4.4-4 Summary of Leopard Frog Larval Development Endpoints at Final Test
25 Duration 4-42
26 Table 4.4-5 Statistical Analysis Results: Wood Frog Reproduction and Development
27 Studies 4-46
28 Table 4.5-1 Hazard Quotients for Leopard Frog PCB Tissue Residues, Based on
29 Literature-Derived Effects Thresholds 4-67
30 Table 4.5-2 Integrated Assessment of Potential for Adverse Impacts to Amphibian
31 Populations (Leopard Frog Study) 4-70
32 Table 4.5-3 Integrated Assessment of Potential for Adverse Impacts to Amphibian
33 Populations (Wood Frog Study) 4-71
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - Xll
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF TABLES
(Continued)
Title Page
Table 4.5-4 Weighting of Measurement Endpoints for Amphibian Weight-of-Evidence
Evaluation 73
Table 4.5-5 Evidence of Harm and Magnitude of Effects for Measurement Endpoints
Related to Maintenance of Amphibian Populations in the Lower
Housatonic River 4-74
Table 4.5-6 Risk Analysis for Amphibians Exposed to tPCBs and Other COCs in the
Housatonic River PSA 4-75
Table 5.2-1 tPCB Concentrations in Representative Species Fish Tissue (mg/kg) in
the PSA; Data from EPA Tissue Collections (1998-2000) 5-13
Table 5.2-2 Total Lipid-Normalized PCB Concentrations (mg/kg lipid) for
Representative Species in the PSA; Data from EPA Tissue Collections
(1998-2000) 5-15
Table 5.2-3 Summary of tPCB Concentrations (mg/kg) from EPA Samples Collected
in Reach 8 5-18
Table 5.2-4 TEQ in Representative Species Fish Tissue in the PSA with DL
Substitution for NDs (ng/kg); Data from EPA Fish Collections (1998-
2000) 5-19
Table 5.2-5 Summary Statistics for Concentrations of PAH COCs in Main Channel
Sediment by Reach 5-20
Table 5.3-1 Criteria Used To Screen Available Studies for Determining Threshold
Body Burdens 5-23
Table 5.3-2 Calculated ED50 Values (tPCBs and TEQ) for Largemouth Bass, Medaka,
and Rainbow Trout Exposed in Ovo to Housatonic River Extracts and
PCB-126 and 2,3,7,8-TCDD Standards 5-41
Table 5.4-1 Probabilities of Exceedances in the PSA for tPCBs 5-62
Table 5.4-2 Probabilities of Exceedances in Reach 8 for tPCBs and TEQ, Based on
EPA Sampling 5-63
Table 5.4-3 Probabilities of Exceedances for TEQ 5-67
Table 5.4-4 Weighting of Measurement Endpoints for Fish Weight-of-Evidence
Evaluation 5-69
Table 5.4-5 Evidence of Harm and Magnitude of Effects for Measurement Endpoints
Related to Maintenance of a Healthy Fish Community 5-70
Table 5.4-6 Risk Analysis for Risk Exposed to tPCBs and TEQ in the Housatonic
River PSA 5-72
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - Xlll
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF TABLES
(Continued)
Table 6.4-1 World Health Organization Toxic Equivalency Factors (TEFs) for TCDD
and Equivalents 6-13
Table 6.7-1 Decision Criteria for Converting Risk Category and Range to Evidence of
Harm and Magnitude of Effect 6-30
Table 7.5-1 Summary of Qualitative Risk Statements for Insectivorous Birds from the
Housatonic River Study Area 7-77
Table 7.5-2 Weighting of Measurement Endpoints for Tree Swallow Weight-of-
Evidence Evaluation 7-81
Table 7.5-3 Weighting of Measurement Endpoints for American Robin Weight-of-
Evidence Evaluation 7-82
Table 7.5-4 Evidence of Harm and Magnitude of Effects for Insectivorous Birds
Exposed to tPCBs in the Housatonic River PSA 7-83
Table 7.5-5 Evidence of Harm and Magnitude of Effects for Insectivorous Birds
Exposed to TEQ in the Housatonic River PSA 7-83
Table 7.5-6 Risk Analysis Summary for Insectivorous Birds Exposed to tPCBs in the
Housatonic River PSA 7-84
Table 7.5-7 Risk Analysis Summary for Insectivorous Birds Exposed to TEQ in the
Housatonic River PSA 7-85
Table 8.5-1 Summary of Qualitative Risk Statements for Piscivorous Birds from the
Housatonic River PSA 8-39
Table 8.5-2 Weighting of Measurement Endpoints for Piscivorous Birds Weight-of-
Evidence Evaluation 8-42
Table 8.5-3 Evidence of Harm and Magnitude of Effects for Piscivorous Birds
Exposed to tPCBs in the Housatonic River PSA 8-43
Table 8.5-4 Evidence of Harm and Magnitude of Effects for Piscivorous Birds
Exposed to TEQ in the Housatonic River PSA 8-43
Table 8.5-5 Risk Analysis Summary for Piscivorous Birds Exposed to tPCBs in the
Housatonic River PSA 8-44
Table 8.5-6 Risk Analysis Summary for Piscivorous Birds Exposed to TEQ in the
Housatonic River PSA 8-45
Table 9.5-1 Results of Snow Tracking and Scent Post Station Surveys in the PSA and
Reference Areas 9-50
Table 9.5-2 Summary of Qualitative Risk Statements for Piscivorous Mammals from
the Housatonic River Study Area 9-54
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - XIV
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF TABLES
(Continued)
Title Page
Table 9.5-3 Weighting of Measurement Endpoints for Piscivorous Mammals Weight-
of-Evidence Evaluation 9-70
Table 9.5-4 Evidence of Harm and Magnitude of Effects for Piscivorous Mammals
Exposed to tPCBs in the Housatonic River PSA 9-71
Table 9.5-5 Evidence of Harm and Magnitude of Effects for Piscivorous Mammals
Exposed to TEQ in the Housatonic River PSA 9-71
Table 9.5-6 Risk Analysis Summary for Piscivorous Mammals Exposed to tPCBs in
the Housatonic River PSA 9-73
Table 9.5-7 Risk Analysis Summary for Piscivorous Mammals Exposed to TEQ in the
Housatonic River PSA 9-74
Table 10.5-1 Summary of Qualitative Risk Statements for Omnivorous and Carnivorous
Mammals from the Housatonic River Study Area 10-44
Table 10.5-2 Weighting of Measurement Endpoints for Omnivorous and Carnivorous
Mammals Weight-of-Evidence Evaluation 10-56
Table 10.5-3 Evidence of Harm and Magnitude of Effects for Omnivorous and
Carnivorous Mammals Exposed to tPCBs in the Housatonic River PSA.... 10-58
Table 10.5-4 Evidence of Harm and Magnitude of Effects for Omnivorous and
Carnivorous Mammals Exposed to TEQ in the Housatonic River PSA 10-58
Table 10.5-5 Risk Analysis Summary for Omnivorous and Carnivorous Mammals
Exposed to tPCBs in the Housatonic River PSA 10-60
Table 10.5-6 Risk Analysis Summary for Omnivorous and Carnivorous Mammals
Exposed to TEQ in the Housatonic River PSA 10-61
Table 11.4-1 Summary of Qualitative Risk Statements for T&E Species from the
Housatonic River Study Area 11-49
Table 11.4-2 Weighting of Measurement Endpoints for T&E Species Weight-of-
Evidence Evaluation 11-59
Table 11.4-3 Evidence of Harm and Magnitude of Effects for T&E Species Exposed to
tPCBs in Housatonic River PSA 11-61
Table 11.4-4 Evidence of Harm and Magnitude of Effects for T&E Species Exposed to
TEQ in the Housatonic River PSA 11-62
Table 11.4-5 Risk Analysis Summary for T&E Species Exposed to tPCBs in the
Housatonic River PSA 11-62
Table 11.4-6 Risk Analysis Summary for T&E Species Exposed to TEQ in the
Housatonic River PSA 11-63
O:\20123001.096\ERA PB\ERA PB TOC.DOC
XV
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF TABLES
(Continued)
Title
Table 12.2-1
Table 12.2-2
Table 12.2-3
Table 12.2-4
Table 12.2-5
Table 12.2-6
Table 12.2-7
Table 12.2-8
Table 12.2-9
Table 12.2-10
Table 12.2-11
Table 12.2-12
Table 12.2-13
Table 12.2-14
Table 12.2-15
Table 12.4-1
Page
Ecological Assessment Endpoints and Conclusions for the Primary Study
Area Portion of the Lower Housatonic River 12-5
Evidence of Harm and Magnitude of Effects for Measurement Endpoints
Related to Maintenance of a Healthy Benthic Community 12-7
Evidence of Harm and Magnitude of Effects for Measurement Endpoints
Related to Maintenance of Amphibian Populations in the Housatonic
River PSA 12-8
Evidence of Harm and Magnitude of Effects for Measurement Endpoints
Related to Maintenance of a Healthy Fish Community 12-10
Evidence of Harm and Magnitude of Effects for Insectivorous Birds
Exposed to tPCBs in the Housatonic River PSA 12-10
Evidence of Harm and Magnitude of Effects for Insectivorous Birds
Exposed to TEQ in the Housatonic River PSA 12-10
Evidence of Harm and Magnitude of Effects for Piscivorous Birds
Exposed to tPCBs in the Housatonic River PSA 12-12
Evidence of Harm and Magnitude of Effects for Piscivorous Birds
Exposed to TEQ in the Housatonic River PSA 12-12
Evidence of Harm and Magnitude of Effects for Piscivorous Mammals
Exposed to tPCBs in the Housatonic River PSA 12-14
Evidence of Harm and Magnitude of Effects for Piscivorous Mammals
Exposed to TEQ in the Housatonic River PSA 12-14
Evidence of Harm and Magnitude of Effects for Omnivorous and
Carnivorous Mammals Exposed to tPCBs in the Housatonic River PSA.... 12-15
Evidence of Harm and Magnitude of Effects for Omnivorous and
Carnivorous Mammals Exposed to TEQ in the Housatonic River PSA 12-15
Evidence of Harm and Magnitude of Effects for T&E Species Exposed to
tPCBs in Housatonic River PSA 12-17
Evidence of Harm and Magnitude of Effects for T&E Species Exposed to
TEQ in the Housatonic River PSA 12-17
Summary of the Assessment of Risks Conducted for Biota Exposed to
tPCBs in the Lower Housatonic River Below Woods Pond 12-28
Comparison of Risks of tPCBs and TEQ to Representative and Other
Species in the Housatonic River PSA 12-48
O:\20123001.096\ERA PB\ERA PB TOC.DOC
XVI
-------
1
LIST OF FIGURES
2 Title Page
3 Figure ES-1 Housatonic River ES-3
4 Figure ES-2 Primary Study Area (PSA) ES-5
5 Figure ES-3 Ecological Risk Assessment Roadmap ES-10
6 Figure ES-4 Hazard Quotients for Aquatic Biota Exposed to tPBCs in the Housatonic
7 River PSA ES-27
8 Figure ES-5 Hazard Quotients for Fish Exposed to 2,3,7,8-TCDD TEQ in the
9 Housatonic River PSA ES-28
10 Figure ES-6 Hazard Quotients for Wildlife Exposed to tPCBs in the Housatonic River
11 PSA ES-29
12 Figure ES-7 Hazard Quotients for Wildlife Exposed to 2,3,7,8-TCDD TEQ in the
13 Housatonic River PSA ES-30
14 Figure ES-8 Assessment of Risk to Benthic Invertebrates Exposed to tPCBs
15 Downstream of Woods Pond ES-34
16 Figure ES-9 Assessment of Risk to Amphibians in Floodplain Exposed to tPCBs
17 Downstream of Woods Pond in Massachusetts ES-36
18 Figure ES-10 Assessment of Risk to Amphibians in Sediment Exposed to tPCBs
19 Downstream of Woods Pond in Connecticut ES-37
20 Figure ES-11 Assessment of Risk to Warmwater Fish Exposed to tPCBs Downstream of
21 Woods Pond ES-3 8
22 Figure ES-12 Assessment of Risk to Trout Exposed to tPCBs Downstream of Woods
23 Pond ES-40
24 Figure ES-13 Assessment of Risk to Mink Exposed to tPCBs Downstream of Woods
25 Pond ES-41
26 Figure ES-14 Assessment of Risk to Otter Exposed to tPCBs Downstream of Woods
27 Pond ES-42
28 Figure ES-15 Assessment of Risk to Bald Eagle Exposed to tPCBs Downstream of
29 Woods Pond ES-44
30 Figure 1.1-1 Housatonic River 1-3
31 Figure 1.1-2 Primary Study Area (PSA) 1-4
32 Figure 1.1-3 Ecological Risk Assessment Roadmap 1-6
33 Figure 1.2-1 GE Plant Area, Removal Action Areas 1-7
34 Figure 1.4-1 Housatonic River, Reach 5 1-13
35 Figure 1.4-2 Housatonic River, Reaches 5A and 5B 1-14
O:\20123001.096\ERA PB\ERA PB TOC.DOC •• 7/12/2003
- - XVII
-------
LIST OF FIGURES
(Continued)
Title Page
1 Figure 1.4-3 Housatonic River, Reaches 5C and 5D 1-15
2 Figure 1.4-4 Housatonic River, Reach 6 1-16
3 Figure 1.4-5 Housatonic River, Reaches 7 to 9 1-18
4 Figure 1.4-6 Housatonic River, Reaches 10 to 13 1-19
5 Figure 1.4-7 Housatonic River, Reaches 14 to 17 1-21
6 Figure 1.5-1 Eight-Step Ecological Risk Assessment Process for Superfund 1-24
7 Figure 2.2-1 Housatonic River Ecological Characterization 2-9
8 Figure 2.5-1 Biphenyl and Representative PCB Congeners 2-24
9 Figure 2.5-2 Distribution of tPCB Concentrations Detected in Sediment Samples from
10 the GE Facility to Long Island Sound 2-26
11 Figure 2.5-3 Mean Total Sediment PCB Concentrations by Reach 2-27
12 Figure 2.5-4 Mean Total Surficial Soil PCB Concentrations at Floodplain Locations by
13 Reach 2-28
14 Figure 2.5-5 Spatially Weighted tPCB Concentrations in Floodplain Soil in the Primary
15 Study Area (Tile 1 of 7) 2-30
16 Figure 2.5-6 Spatially Weighted tPCB Concentrations in Floodplain Soil in the Primary
17 Study Area (Tile 2 of 7) 2-31
18 Figure 2.5-7 Spatially Weighted tPCB Concentrations in Floodplain Soil in the Primary
19 Study Area (Tile 3 of 7) 2-32
20 Figure 2.5-8 Spatially Weighted tPCB Concentrations in Floodplain Soil in the Primary
21 Study Area (Tile 4 of 7) 2-33
22 Figure 2.5-9 Spatially Weighted tPCB Concentrations in Floodplain Soil in the Primary
23 Study Area (Tile 5 of 7) 2-34
24 Figure 2.5-10 Spatially Weighted tPCB Concentrations in Floodplain Soil in the Primary
25 Study Area (Tile 6 of 7) 2-35
26 Figure 2.5-11 Spatially Weighted tPCB Concentrations in Floodplain Soil in the Primary
27 Study Area (Tile 7 of 7) 2-36
28 Figure 2.5-12 Total PCB Concentrations Measured in all Surface Water Samples
29 Collected from the Housatonic River Since 1980 2-37
30 Figure 2.5-13 Total Surface Water PCB Concentrations by River Mile (Data Collected
31 Since 1996) 2-37
32 Figure 2.5-14 Total PCB Concentration (mg/kg wet weight) in Selected Biota
33 (Excluding Fish) for Reaches 5 and 6 2-38
O:\20123001.096\ERA PB\ERA PB TOC.DOC ••• 7/12/2003
- - XV111
-------
LIST OF FIGURES
(Continued)
Title Page
1 Figure 2.5-15 Total PCB Concentration (mg/kg wet weight) in Reaches 5 and 6 Fish 2-39
2 Figure 2.5-16 EPA Fish Collections (1998-2000) - Median tPCB Concentrations - All
3 Ages by Subreach in the PSA 2-40
4 Figure 2.5-17 EPA Fish Collections (1998-2000) - Median Lipid Normalized PCB
5 Concentrations - All Ages by Subreach in the PSA 2-41
6 Figure 2.7-1 Housatonic River Ecological Risk Assessment Conceptual Model:
7 Principal Exposure Pathways for PCBs 2-55
8 Figure 2.9-1 Example Endpoint Weighting Sheet 2-70
9 Figure 2.9-2 Scoring Sheet for Evidence of Harm and Magnitude 2-71
10 Figure 2.9-3 Example of Qualitative Assessment 2-72
11 Figure 3.1-1 Conceptual Model Diagram: Exposure Pathways for Benthic Invertebrates
12 Exposed to Contaminants of Concern (COCs) in the Housatonic River 3-3
13 Figure 3.1-2 Overview of Approach Used to Assess Exposure of Benthic Invertebrates
14 to Contaminants of Concern (COCs) in the Housatonic River 3-4
15 Figure 3.1-3 Overview of Approach Used to Assess the Effects of Contaminants of
16 Concern (COCs) to Benthic Invertebrates in the Housatonic River 3-5
17 Figure 3.1-4 Overview of Approach Used to Characterize the Risks of Contaminants of
18 Concern (COCs) to Benthic Invertebrates in the Housatonic River 3-6
19 Figure 3.1-5 Summary of Studies Conducted in Conjunction with Ecological Risk
20 Assessment for Benthic Invertebrates, and Linkage to ERA 3-10
21 Figure 3.2-1 Benthic Invertebrate Sampling Locations and Simplified Station
22 Identifiers 3-12
23 Figure 3.2-2 Median Percent Fines and Percent TOC by Sampling Location, for
24 Benthic Community Grab Samples 3-16
25 Figure 3.2-3 Concentrations of tPCBs in Sediment by Sampling Location for Individual
26 Benthic Community Grab Samples, and Associated Measures of Central
27 Tendency 3-19
28 Figure 3.2-4 Comparison of tPCB Concentrations in Sediment Collected at Benthic
29 Toxicity Sampling Locations, from Various Sampling Efforts Conducted
30 in 1999 3-20
31 Figure 3.2-5 Medians and Quartiles of PCB and TOC in the Housatonic River PSA,
32 Subdivided by River Reach and 0.25 Mile Subreaches 3-21
33 Figure 3.2-6 Concentrations of tPCBs in Benthic Invertebrate Tissues by Location and
34 Functional Feeding Group 3-23
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - XIX
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
LIST OF FIGURES
(Continued)
Title Page
Figure 3.3-1 Survival of Hyalella azteca in Chronic Laboratory Toxicity Tests, at Three
Time Periods (28 days, 35 days, 42 days) 3-32
Figure 3.3-2 Growth of Hyalella azteca in Chronic Laboratory Toxicity Tests, at Two
Time Periods (28 days, 42 days) 3-33
Figure 3.3-3 Reproduction of Hyalella azteca in Chronic Laboratory Toxicity Tests,
Based on Mean Number of Young (Days 28-42) 3-34
Figure 3.3-4 Survival and Emergence of Chironomus tentans in Chronic Laboratory
Toxicity Tests (43 days) 3-35
Figure 3.3-5 Growth Endpoints for Chironomus tentans in Chronic Laboratory Toxicity
Test (20 days) 3-36
Figure 3.3-6 Survival of Hyalella azteca in 48-hour Low Flow In Situ Toxicity Tests
Conducted 14-16 June 1999 3-37
Figure 3.3-7 Survival of Hyalella azteca in 10-day Low Flow In Situ Toxicity Tests
Conducted 17-27 June 1999 3-38
Figure 3.3-8 Survival of Chironomus tentans in 48-hour Low Flow In Situ Toxicity
Tests Conducted 14-16 June 1999 3-39
Figure 3.3-9 Survival of Chironomus tentans in 10-day Low Flow In Situ Toxicity
Tests Conducted 17-27 June 1999 3-40
Figure 3.3-10 Survival of Daphnia magna in 48-hour Low Flow In Situ Toxicity Tests
Conducted 14-16 June 1999 3-41
Figure 3.3-11 Survival of Lumbriculus variegatus in 48-hour Low Flow In Situ Toxicity
Tests Conducted 14-16 June 1999 3-42
Figure 3.3-12 Statistical Endpoints for Toxicity Data, with Comparisons to Station A1
(sorted by LC5o/EC5o value) 3-48
Figure 3.3-13 Segmented Linear Regression Models Applied to Toxicity Data, Relating
Relative Performance Proportion (RPP) to Bulk Sediment tPCB
Concentrations (mg/kg) 3-51
Figure 3.3-14 Combined Effects and No-Effects Levels for PCB Concentrations (mg/kg
wet) in Benthic Invertebrate Tissue Samples - tPCBs and Aroclor 1254 3-55
Figure 3.3-15 Average Ranks Analysis for Six Benthic Community Metrics, with Equal
Weighting Assigned to Each Metric 3-59
Figure 3.3-16 Multidimensional Scaling for Benthic Community Health Metrics,
Showing Metric Medians on MDS Plot 3-60
O:\20123001.096\ERA PB\ERA PB TOC.DOC
" " XX
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
LIST OF FIGURES
(Continued)
Title Page
Figure 3.4-1 Hazard Quotients (Median and Range) Based on Median Sediment
Chemistry for tPCBs, for Samples Collected in 1999 Close to Sediment
Quality Triad Stations 3-64
Figure 3.4-2 Hazard Quotients (Median, Range) Based on Overlying Water PCB
Concentrations, Measured Synoptic with In Situ Toxicity Tests 3-67
Figure 3.4-3 Hazard Quotients for tPCB Tissue Residues in Benthic Invertebrates,
Relative to Two Effects Thresholds Derived from Literature Studies 3-68
Figure 3.4-4 Weight-of-Evidence Evaluation of Housatonic River Benthic Sampling
Locations, with Indications of Alteration/Risk Relative to Background 3-70
Figure 4.1-1 Conceptual Model Diagram: Exposure Pathways for Amphibians Exposed
to Contaminants of Concern (COCs) in the Housatonic River PSA 4-3
Figure 4.1-2 Overview of Approach Used to Assess Exposure of Amphibians to
Contaminants of Concern (COCs) in the Housatonic River PSA 4-4
Figure 4.1-3 Overview of Approach Used to Assess the Effects of Contaminants of
Concern (COCs) to Amphibians in the Housatonic River PSA 4-5
Figure 4.1-4 Overview of Approach Used to Characterize the Risks of Contaminants of
Concern (COCs) to Amphibians in the Housatonic River PSA 4-6
Figure 4.2-1 Leopard Frog Exposure Pathways 4-9
Figure 4.2-2 Wood Frog Exposure Pathways 4-10
Figure 4.2-3 General Model of Leopard Frog Vernal Pool (VP) Reproduction and
Development Study 4-15
Figure 4.2-4 General Model of Wood Frog Vernal Pool (VP) Reproduction and
Development Study 4-16
Figure 4.3-1 Frequency Distribution and Cumulative Percentage of Sediment tPCB
Exposure Point Concentrations for 66 PSA Temporary and Permanent
Pools (Based on EPA Spatially Weighted Data) 4-24
Figure 4.3-2 Total Sediment PCB Concentrations for Wood Frog Vernal Pool Study
(mean, n =2) and Leopard Frog Reproduction/Development Study 4-26
Figure 4.3-3 Comparison of Leopard Frog Tissue Samples to Sediment tPCB
Concentrations (Reproductive Study Data and Spatially Weighted Data) 4-28
Figure 4.3-4 Comparison of tPCB Concentrations in Tissue (in Various Phases of the
Wood Frog Developmental Study) with Mean Vernal Pool and Spatially
Weighted Mean tPCB Concentrations in Sediment 4-30
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - XXI
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF FIGURES
(Continued)
Title Page
Figure 4.4-1 Comparison of Percent Abnormal Sperm Heads (Mean) from Male Adult
Chemical Analysis Leopard Frogs, with Mean Sediment tPCB and
Spatially Weighted Mean tPCB 4-36
Figure 4.4-2 Comparison of Mean Percent of Oocytes at Stage VI (Mature) for Female
Leopard Frogs, with Mean Sediment tPCB and Spatially Weighted Mean
tPCB 4-36
Figure 4.4-3 Days to Gosner Developmental Stage 26 (±1) and Final Developmental
Stage Reached at End-of-Test, with Sediment tPCB and Spatially
Weighted Mean tPCB: 2000 Leopard Frog Study 4-43
Figure 4.4-4 Comparison of Phase I Larval Wood Frog Malformations as Gosner
Developmental Stage 20-24 to Mean Sediment tPCB and Spatially
Weighted Mean tPCB 4-51
Figure 4.4-5 Incidence of Malformation in Phase I Wood Frog Metamorphs, with Mean
Sediment tPCB and Spatially Weighted Mean tPCB 4-51
Figure 4.4-6 Incidence of Malformation in Phase I Wood Frog Metamorphs, Phase I
Metamorph Tissue tPCB 4-52
Figure 4.4-7 Ratio of Males to Females in Phase III Wood Frog Metamorphs, with
Sediment tPCB and Spatially Weighted Mean tPCBs 4-55
Figure 4.4-8 Ratio of Males to Females in Phase III Wood Frog Metamorphs, with
Tissue tPCBs 4-55
Figure 4.4-9 Percent Malformation in Phase III Wood Frog Metamorphs, with
Sediment tPCBs 4-56
Figure 4.4-10 Percent Malformation in Phase III Wood Frog Metamorphs, with Tissue
tPCBs 4-56
Figure 4.4-11 Summary of Available Literature Effects Data on PCB Tissue Residues in
Anuran Amphibians 4-60
Figure 5.1-1 Conceptual Model Diagram: Exposure Pathways for Fish Exposed to
COCs in the Housatonic River 5-4
Figure 5.1-2 Overview of Approach Used To Assess Exposure of Fish to COCs in the
Housatonic River 5-5
Figure 5.1-3 Overview of Approach Used To Assess the Effects of COCs to Fish in the
Housatonic River 5-6
Figure 5.1-4 Overview of Approach Used To Characterize the Risks of COCs to Fish in
the Housatonic River 5-7
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - XXII
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
LIST OF FIGURES
(Continued)
Title Page
Figure 5.2-1 Box-and-Whisker Plots of Lipid-Normalized tPCB Concentrations Plotted
by Sample Type for Species with Multiple Sample Types 5-14
Figure 5.2-2 Box-and-Whisker Plot of Largemouth Bass tPCB Concentrations (Lipid-
Normalized) Versus Age 5-16
Figure 5.3-1 Literature-Based PCB Fish Tissue Effects Concentrations 5-24
Figure 5.3-2 Literature-Based TCDD (TEQ) Fish Tissue Effects Concentrations 5-27
Figure 5.3-3 Effects of in Ovo Exposure to Increasing Doses of Reach 6 Extracts on
Largemouth Bass at Swim-Up 5-35
Figure 5.3-4 Effects of in Ovo Exposure to Increasing Doses of Reach 5BC Extracts on
Medaka at 5d Post Swim-Up 5-36
Figure 5.3-5 Effects of in Ovo Exposure to Increasing Doses of Reach 6 Extracts on
Medaka at 5d Post Swim-Up 5-36
Figure 5.3-6 Effects of in Ovo Exposure to Increasing Doses of Reach 5BC Extracts on
Rainbow Trout at 600 DTU 5-37
Figure 5.3-7 Effects of in Ovo Exposure to Increasing Doses of Reach 6 Extracts on
Rainbow Trout at 600 DTU 5-37
Figure 5.3-8 TEQ ED50 Estimates for Fish Exposed to Housatonic River Extracts and
PCB-126 and TCDD Standards (Logarithmic Scale) 5-42
Figure 5.3-9 tPCB ED50 Estimates for Fish Exposed to Housatonic River Extracts and
PCB-126 and TCDD Standards (Logarithmic Scale) 5-43
Figure 5.4-1 Hazard Quotients for tPCBs in PSA Fish Based on Comparison to the
Mean Site-Specific Fish Toxicity Threshold (49 mg/kg tPCB)
(Logarithmic Scale) 5-56
Figure 5.4-2 Complementary Cumulative Distribution Plot for tPCB Concentrations in
Whole Body Tissue Compared to Effects Concentrations - Brown
Bullhead 5-57
Figure 5.4-3 Complementary Cumulative Distribution Plot for tPCB Concentrations in
Whole Body Tissue Compared to Effects Concentrations - Largemouth
Bass 5-58
Figure 5.4-4 Complementary Cumulative Distribution Plot for tPCB Concentrations in
Whole Body Tissue Compared to Effects Concentrations - Pumpkinseed.... 5-59
Figure 5.4-5 Complementary Cumulative Distribution Plot for tPCB Concentrations in
Whole Body Tissue Compared to Effects Concentrations - White Sucker.... 5-60
Figure 5.4-6 Complementary Cumulative Distribution Plot for tPCB Concentrations in
Whole Body Tissue Compared to Effects Concentrations - Yellow Perch.... 5-61
O:\20123001.096\ERA PB\ERA PB TOC.DOC ••• 7/12/2003
- - XX111
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
LIST OF FIGURES
(Continued)
Title Page
Figure 5.4-7 Hazard Quotients for TEQ for Fish in Primary Study Area (PSA) Based on
Comparison to the Average Site-Specific Tissue Effects Threshold (42
ng/kg TEQ) (Logarithmic Scale) 5-65
Figure 5.4-8 Complementary Cumulative Distribution Plot for TEQ Concentrations in
Whole Body Tissue Compared to Effects Concentrations for All Species
Using DL Substitution for Non-Detects 5-66
Figure 6.1-1 Conceptual Model for the Assessment of Risks from tPCBs and TEQ to
Wildlife in the Housatonic River Primary Study Area 6-3
Figure 6.1-2 Framework Used to Model Exposure of Wildlife Species to Contaminants
of Concern (COCs) in the Housatonic River PSA 6-4
Figure 6.1-3 Approach Used to Model Effects of Contaminants of Concern (COCs) to
Representative Species in the Housatonic River PSA 6-5
Figure 6.1-4 Approach Used to Characterize the Risks from Contaminants of Concern
(COCs) to Representative Species in the Housatonic River PSA 6-6
Figure 6.4-1 Molecular Structure of the Planar Chlorinated Hydrocarbon, 2,3,7,8-
Tetrachlorodibenzo-p-dioxin 6-11
Figure 6.4-2 Decision Tree for Determining Appropriate Treatment of Data with Non-
Detects and Co-Elution 6-17
Figure 6.5-1 Example Exposure Distribution from Monte Carlo and Probability Bounds
Analyses (TDI = total daily intake of tPCBs) 6-21
Figure 6.6-1 Decision Criteria Used to Characterize Effects for Each Wildlife
Receptor-COC Combination 6-22
Figure 6.7-1 Example Risk Curves Indicating Low, Intermediate, and High Risk
Categories 6-26
Figure 7.1-1 Conceptual Model Diagram: Exposure Pathways for Insectivorous Birds
Exposed to Contaminants of Concern in the Housatonic River PSA 7-4
Figure 7.1-2 Overview of Approach Used to Assess Modeled Exposure of
Insectivorous Birds to Contaminants of Concern in the Housatonic River
PSA 7-5
Figure 7.1-3 Overview of Approach Used to Assess the Modeled Effects of
Contaminants of Concern to Insectivorous Birds in the Housatonic River
PSA 7-6
Figure 7.1-4 Overview of Approach Used to Characterize the Risks of Contaminants of
Concern to Insectivorous Birds in the Housatonic River PSA 7-7
Figure 7.2-1 Tree Swallow (Tachycineta bicolof) 7-10
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - XXIV
-------
LIST OF FIGURES
(Continued)
Title Page
1 Figure 7.2-2 American Robin (Turdus migratorius) 7-11
2 Figure 7.3-1 Tree Swallow Nest Box Locations and Soil Invertebrate Sampling
3 Locations in the Housatonic River PSA 7-14
4 Figure 7.3-2 TDI Exposure Model Input Distributions for Tree Swallows 7-19
5 Figure 7.3-3 Microexposure Model Input Distributions for Maternal Transfer for Tree
6 Swallows 7-20
7 Figure 7.3-4 TDI Exposure Model Input Distributions for American Robin 7-21
8 Figure 7.3-5 Concentration of tPCBs in Prey of Tree Swallows 7-27
9 Figure 7.3-6 Concentration of TEQ in Prey of Tree Swallows 7-28
10 Figure 7.3-7 Concentration of tPCBs in Prey of American Robins 7-29
11 Figure 7.3-8 Concentration of TEQ in Prey of American Robins 7-30
12 Figure 7.3-9 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte Carlo
13 Analysis and Probability Bounds Analysis at Holmes Road 7-31
14 Figure 7.3-10 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte Carlo
15 Analysis and Probability Bounds Analysis at New Lenox Road 7-32
16 Figure 7.3-11 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte Carlo
17 Analysis and Probability Bounds Analysis at Roaring Brook 7-33
18 Figure 7.3-12 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte Carlo
19 Analysis and Probability Bounds Analysis at Southwest Branch 7-34
20 Figure 7.3-13 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte Carlo
21 Analysis and Probability Bounds Analysis at Threemile Pond 7-35
22 Figure 7.3-14 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte Carlo
23 Analysis and Probability Bounds Analysis at Taconic Valley 7-36
24 Figure 7.3-15 Tree Swallow TDI Exposure Model for TEQ: Results of Monte Carlo
25 Analysis and Probability Bounds Analysis at Holmes Road 7-37
26 Figure 7.3-16 Tree Swallow TDI Exposure Model for TEQ: Results of Monte Carlo
27 Analysis and Probability Bounds Analysis at New Lenox Road 7-38
28 Figure 7.3-17 Tree Swallow TDI Exposure Model for TEQ: Results of Monte Carlo
29 Analysis and Probability Bounds Analysis at Roaring Brook 7-39
30 Figure 7.3-18 Tree Swallow TDI Exposure Model for TEQ: Results of Monte Carlo
31 Analysis and Probability Bounds Analysis at Southwest Branch 7-40
32 Figure 7.3-19 Tree Swallow TDI Exposure Model for TEQ: Results of Monte Carlo
33 Analysis and Probability Bounds Analysis at Threemile Pond 7-41
O:\20123001.096\ERA PB\ERA PB TOC.DOC 7/12/2003
- - XXV
-------
LIST OF FIGURES
(Continued)
Title Page
1 Figure 7.3-20 Tree Swallow TDI Exposure Model for TEQ: Results of Monte Carlo
2 Analysis and Probability Bounds Analysis at Taconic Valley 7-42
3 Figure 7.3-21 Exposure of American Robins to tPCBs at Site 13 of the Housatonic River
4 PSA 7-43
5 Figure 7.3-22 Exposure of American Robins to tPCBs at Site 14 of the Housatonic River
6 PSA 7-44
7 Figure 7.3-23 Exposure of American Robins to tPCBs at Site 15 of the Housatonic River
8 PSA 7-45
9 Figure 7.3-24 Exposure of American Robins to 2,3,7,8-TCDD TEQ at Site 13 of the
10 Housatonic River PSA 7-46
11 Figure 7.3-25 Exposure of American Robins to 2,3,7,8-TCDD TEQ at Site 14 of the
12 Housatonic River PSA 7-47
13 Figure 7.3-26 Exposure of American Robins to 2,3,7,8-TCDD TEQ at Site 15 of the
14 Housatonic River PSA 7-48
15 Figure 7.3-27 Tree Swallow Microexposure Model for tPCBs: Results of Monte Carlo
16 Analysis and Probability Bounds Analysis at Holmes Road 7-50
17 Figure 7.3-28 Tree Swallow Microexposure Model for tPCBs: Results of Monte Carlo
18 Analysis and Probability Bounds Analysis at New Lenox Road 7-51
19 Figure 7.3-29 Tree Swallow Microexposure Model for tPCBs: Results of Monte Carlo
20 Analysis and Probability Bounds Analysis at Roaring Brook 7-52
21 Figure 7.3-30 Tree Swallow Microexposure Model for tPCBs: Results of Monte Carlo
22 Analysis and Probability Bounds Analysis at Southwest Branch 7-53
23 Figure 7.3-31 Tree Swallow Microexposure Model for tPCBs: Results of Monte Carlo
24 Analysis and Probability Bounds Analysis at Threemile Pond 7-54
25 Figure 7.3-32 Tree Swallow Microexposure Model for tPCBs: Results of Monte Carlo
26 Analysis and Probability Bounds Analysis at Taconic Valley 7-55
27 Figure 7.3-33 Tree Swallow Microexposure Model for TEQ: Results of Monte Carlo
28 Analysis and Probability Bounds Analysis at Holmes Road 7-56
29 Figure 7.3-34 Tree Swallow Microexposure Model for TEQ: Results of Monte Carlo
30 Analysis and Probability Bounds Analysis at New Lenox Road 7-57
31 Figure 7.3-35 Tree Swallow Microexposure Model for TEQ: Results of Monte Carlo
32 Analysis and Probability Bounds Analysis at Roaring Brook 7-58
33 Figure 7.3-36 Tree Swallow Microexposure Model for TEQ: Results of Monte Carlo
34 Analysis and Probability Bounds Analysis at Southwest Branch 7-59
O:\20123001.096\ERA PB\ERA PB TOC.DOC • 7/12/2003
- - XXVI
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
LIST OF FIGURES
(Continued)
Figure 7.3-37 Tree Swallow Microexposure Model for TEQ: Results of Monte Carlo
Analysis and Probability Bounds Analysis at Threemile Pond 7-60
Figure 7.3-38 Tree Swallow Microexposure Model for TEQ: Results of Monte Carlo
Analysis and Probability Bounds Analysis at Taconic Valley 7-61
Figure 7.4-1 Effects of Aroclor 1254/1260 on Avian Species (mg/kg bw/d) 7-66
Figure 7.4-2 Effects of 2,3,7,8-TCDD TEQ on Avian Species (ng TEQ/kg bw/d) 7-67
Figure 8.1-1 Conceptual Model for Exposure of Piscivorous Birds to tPCBs and TEQ
in the Housatonic River PSA 8-4
Figure 8.1-2 Overview of Approach Used to Assess Modeled Exposure of Piscivorous
Birds to Contaminants of Concern (COCs) in the Housatonic River PSA 8-5
Figure 8.1-3 Overview of Approach Used to Assess the Modeled Effects of
Contaminants of Concern (COCs) to Piscivorous Birds in the Housatonic
River PSA 8-6
Figure 8.1-4 Overview of Approach Used to Characterize the Risks of Contaminants of
Concern (COCs) to Piscivorous Birds in the Housatonic River PSA 8-7
Figure 8.2-1 Belted Kingfisher (Ceryle alcyon) 8-10
Figure 8.2-2 Osprey (Pandion haliaetus) 8-11
Figure 8.3-1 Exposure Model Input Distributions for Belted Kingfisher 8-15
Figure 8.3-2 Exposure Model Input Distributions for Osprey 8-16
Figure 8.3-3 Concentrations of tPCBs in the Prey of Belted Kingfishers in the
Housatonic River PSA and Reference Areas 8-20
Figure 8.3-4 Concentrations of TEQ in the Prey of Belted Kingfishers in the
Housatonic River PSA and Reference Areas 8-20
Figure 8.3-5 Concentrations of tPCBs in the Prey of Ospreys in the Housatonic River
PSA and Reference Areas 8-21
Figure 8.3-6 Concentrations of TEQ in the Prey of Ospreys in the Housatonic River
PSA and Reference Areas 8-21
Figure 8.3-7 Exposure of Belted Kingfishers to tPCBs in Reach 5 of the Housatonic
River PSA 8-23
Figure 8.3-8 Exposure of Belted Kingfishers to tPCBs in Reach 6 of the Housatonic
River PSA 8-23
Figure 8.3-9 Exposure of Belted Kingfishers to tPCBs in the Housatonic River
Upstream Reference Area 8-24
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - XXV11
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF FIGURES
(Continued)
Title Page
Figure 8.3-10 Exposure of Belted Kingfishers to tPCBs in the Threemile Pond Reference
Area 8-24
Figure 8.3-11 Exposure of Belted Kingfishers to TEQ in Reach 5 of the Housatonic
River PSA 8-25
Figure 8.3-12 Exposure of Belted Kingfishers to TEQ in Reach 6 of the Housatonic
River PSA 8-25
Figure 8.3-13 Exposure of Belted Kingfishers to TEQ in the Housatonic River Upstream
Reference Area 8-26
Figure 8.3-14 Exposure of Belted Kingfishers to TEQ in the Threemile Pond Reference
Area 8-26
Figure 8.3-15 Exposure of Ospreys to tPCBs in Reaches 5 and 6 of the Housatonic River
PSA 8-27
Figure 8.3-16 Exposure of Ospreys to tPCBs in the Housatonic River Upstream
Reference Area 8-27
Figure 8.3-17 Exposure of Ospreys to tPCBs in the Threemile Pond Reference Area 8-28
Figure 8.3-18 Exposure of Ospreys to TEQ in Reaches 5 and 6 of the Housatonic River
PSA 8-28
Figure 8.3-19 Exposure of Ospreys to TEQ in the Upstream Reference Area 8-29
Figure 8.3-20 Exposure of Ospreys to TEQ in the Threemile Pond Reference Area 8-29
Figure 8.4-1 Effects of Aroclor 1254/1260 on Avian Species (mg/kg bw/d) 8-32
Figure 8.4-2 Effects of 2,3,7,8-TCDD TEQ on Avian Species (ng TEQ/kg bw/d) 8-33
Figure 9.1-1 Conceptual Model Diagram: Exposure Pathways for Piscivorous
Mammals Exposed to COCs in the Housatonic River PSA 9-4
Figure 9.1-2 Overview of Approach Used to Assess Modeled Exposure of Piscivorous
Mammals to Contaminants of Concern (COCs) in the Housatonic River
PSA 9-5
Figure 9.1-3 Overview of Approach Used to Assess the Modeled Effects of
Contaminants of Concern (COCs) to Piscivorous Mammals in the
Housatonic River PSA 9-6
Figure 9.1-4 Overview of Approach Used to Characterize the Risks of Contaminants of
Concern (COCs) to Piscivorous Mammals in the Housatonic River PSA 9-7
Figure 9.2-1 Mink {Mustela vison) 9-9
Figure 9.2-2 River Otter (Lutra canadensis) 9-10
Figure 9.3-1 Input Distributions Used in Exposure Modeling for Mink 9-15
O:\20123001.096\ERA PB\ERA PB TOC.DOC
- - XXV111
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
LIST OF FIGURES
(Continued)
Title
Page
Figure
9.3-2
Input Distributions Used in Exposure Modeling for River Otter
9-16
Figure
9.3-3
Concentrations of tPCBs in Prey of Mink
9-19
Figure
9.3-4
Concentrations of TEQ in Prey of Mink
9-20
Figure
9.3-5
Concentrations of tPCBs in Prey of River Otter
9-20
Figure
9.3-6
Concentrations of TEQ in Prey of River Otter
9-21
Figure
9.3-7
Exposure of Mink to tPCBs in Reach 5 of the Housatonic River
9-22
Figure
9.3-8
Exposure of Mink to tPCBs in Reach 6 of the Housatonic River
9-23
Figure
9.3-9
Exposure of Mink to tPCBs in the Housatonic River Upstream Reference
Area
9-24
Figure
9.3-10
Exposure of Mink to tPCBs in the Threemile Pond Reference Area
9-25
Figure
9.3-11
Exposure of Mink to 2,3,7,8-TCDD TEQ in Reach 5 of the Housatonic
River
9-26
Figure
9.3-12
Exposure of Mink to 2,3,7,8-TCDD TEQ in Reach 6 of the Housatonic
River
9-27
Figure
9.3-13
Exposure of Mink to 2,3,7,8-TCDD TEQ in the Housatonic River
Upstream Reference Area
9-28
Figure
9.3-14
Exposure of Mink to 2,3,7,8-TCDD TEQ in the Threemile Pond
Reference Area
9-29
Figure
9.3-15
Exposure of River Otter to tPCBs in Reaches 5 and 6 of the Housatonic
River
9-30
Figure
9.3-16
Exposure of River Otter to tPCBs in the Housatonic River Upstream
Reference Area
9-31
Figure
9.3-17
Exposure of River Otter to tPCBs in the Threemile Pond Reference Area....
9-32
Figure
9.3-18
Exposure of River Otter to 2,3,7,8-TCDD TEQ in Reaches 5 and 6 of the
Housatonic River
9-33
Figure
9.3-19
Exposure of River Otter to 2,3,7,8-TCDD TEQ in the Upstream Reference
Area
9-34
Figure
9.3-20
Exposure of River Otter to 2,3,7,8-TCDD TEQ in the Threemile Pond
Reference Area
9-35
Figure
9.4-1
Dose Response Curve for Effects of tPCBs on Fecundity of Mink
9-48
Figure
9.5-1
Total PCB Risk to Mink Exposed to tPCBs in Reach 5 of the Housatonic
River
9-55
Figure
9.5-2
Total PCB Risk to Mink (10% Foraging Time in Reach 5)
9-56
O:\20123001.096\ERA PB\ERA PB TOC.DOC • 7/12/2003
- - XXIX
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF FIGURES
(Continued)
Title Page
Figure 9.5-3 Total PCB Risk to Mink Exposed to tPCBs in Reach 6 of the Housatonic
River 9-57
Figure 9.5-4 Total PCB Risk to Mink (10% Foraging Time in Reach 6) 9-58
Figure 9.5-5 Total PCB Risk to Mink Foraging in the Upstream Reference Area 9-59
Figure 9.5-6 Total PCB Risk to Mink (10% Foraging Time in the Upstream Reference
Area) 9-60
Figure 9.5-7 Total PCB Risk to Mink Exposed to tPCBs in the Threemile Pond
Reference Area 9-61
Figure 9.5-8 Total PCB Risk to Mink (10% Foraging Time in the Threemile Pond
Reference Area) 9-62
Figure 9.5-9 Total PCB Risk to River Otter Exposed to tPCBs in Reaches 5 and 6 of
the Housatonic River 9-63
Figure 9.5-10 Total PCB Risk to River Otter (10% Foraging Time in Reaches 5 and 6) 9-64
Figure 9.5-11 Total PCB Risk to River Otter Foraging in the Upstream Reference Area.... 9-65
Figure 9.5-12 Total PCB Risk to River Otter (10% Foraging Time in the Upstream
Reference Area) 9-66
Figure 9.5-13 Total PCB Risk to River Otter Exposed to tPCBs in the Threemile Pond
Reference Area 9-67
Figure 9.5-14 Total PCB Risk to River Otter (10% Foraging Time in the Threemile Pond
Reference Area) 9-68
Figure 10.1-1 Conceptual Model Diagram: Exposure Pathways for Omnivorous and
Carnivorous Mammals Exposed to Contaminants of Concern (COCs) in
the Housatonic River PSA 10-4
Figure 10.1-2 Framework Used to Model Exposure of Wildlife Species to Contaminants
of Concern (COCs) in the Housatonic River PSA 10-5
Figure 10.1-3 Approach Used to Model Effects of Contaminants of Concern (COCs) to
Representative Species in the Housatonic River PSA 10-6
Figure 10.1-4 Overview of Approach Used to Characterize the Risks of Contaminants of
Concern (COCs) to Omnivorous and Carnivorous Mammals in the
Housatonic River PSA 10-7
Figure 10.2-1 Red Fox (Vulpes vulpes) 10-9
Figure 10.2-2 Northern Short-Tailed Shrew (Blarina brevicauda) 10-10
Figure 10.3-1 Input Distributions for the Exposure Modeling of the Red Fox 10-15
Figure 10.3-2 Input Distributions for the Exposure Modeling of Short-Tailed Shrew 10-16
O:\20123001.096\ERA PB\ERA PB TOC.DOC 7/12/2003
- - XXX
-------
LIST OF FIGURES
(Continued)
Title Page
1 Figure 10.3-3 Concentrations of tPCBs in Prey of Northern Short-Tailed Shrew (n=l for
2 invertebrates and earthworms) 10-21
3 Figure 10.3-4 Concentrations of TEQ in Prey of Northern Short-Tailed Shrew (n=l for
4 invertebrates and earthworms) 10-22
5 Figure 10.3-5 Exceedance Probability Distribution for Red Fox Exposed to tPCBs in
6 Reach 5 of the PSA 10-23
7 Figure 10.3-6 Exceedance Probability Distribution for Red Fox Exposed to TEQ in
8 Reach 5 of the PSA 10-24
9 Figure 10.3-7 Exceedance Probability Distribution for Short-Tailed Shrew Exposed to
10 tPCBs at Location 13 of the PSA 10-25
11 Figure 10.3-8 Exceedance Probability Distribution for Short-Tailed Shrew Exposed to
12 tPCBs at Location 14 of the PSA 10-26
13 Figure 10.3-9 Exceedance Probability Distribution for Short-Tailed Shrew Exposed to
14 tPCBs at Location 15 of the PSA 10-27
15 Figure 10.3-10 Exceedance Probability Distribution for Short-Tailed Shrew Exposed to
16 TEQ at Location 13 of the PSA 10-28
17 Figure 10.3-11 Exceedance Probability Distribution for Short-Tailed Shrew Exposed to
18 TEQ at Location 14 of the PSA 10-29
19 Figure 10.3-12 Exceedance Probability Distribution for Short-Tailed Shrew Exposed to
20 TEQ at Location 15 of the PSA 10-30
21 Figure 10.4-1 Dose-Response Curve for Effects of tPCBs on Mortality at Birth of Rats... 10-38
22 Figure 10.4-2 Dose-Response Curve for Effects of TEQ on Reproductive Fecundity of
23 Rat 10-39
24 Figure 10.4-3 Dose Response Curve for Effects of TEQ on Reproductive Fecundity of
25 Mouse 10-40
26 Figure 10.5-1 Risk Function for Red Fox Exposed to tPCBs in Reach 5 of the
27 Housatonic River 10-45
28 Figure 10.5-2 Risk Function for Red Fox Exposed to TEQ in Reach 5 of the Housatonic
29 River 10-46
30 Figure 10.5-3 Risk Function for Short-Tailed Shrew Exposed to tPCBs at Location 13 of
31 the PS A 10-47
32 Figure 10.5-4 Risk Function for Short-Tailed Shrew Exposed to tPCBs at Location 14 of
33 the PSA 10-48
O:\20123001.096\ERA PB\ERA PB TOC.DOC • 7/12/2003
- - XXXI
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
LIST OF FIGURES
(Continued)
Title
Figure 10.5-5
Figure 10.5-6
Figure 10.5-7
Figure 10.5-8
Figure 11.1-1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
Figure 1
.1-2
.1-3
.1-4
.2-1
.2-2
.2-3
.2-4
.2-5
.2-6
.2-7
.2-8
.2-9
.2-10
.2-11
.2-12
.2-13
Page
Risk Function for Short-Tailed Shrew Exposed to tPCBs at Location 15 of
the PS A 10-49
Risk Function for Short-Tailed Shrew Exposed to TEQ at Location 13 of
the PSA 10-50
Risk Function for Short-Tailed Shrew Exposed to TEQ at Location 14 of
the PSA 10-51
Risk Function for Short-Tailed Shrew Exposed to TEQ at Location 15 of
the PSA 10-52
Conceptual Model Diagram: Exposure Pathways for T&E Species
Exposed to COCs in the Housatonic PSA 11-3
Approach Used to Assess Modeled Exposure of T&E Species to COCs 11-4
Approach Used to Assess the Modeled Effects of COCs to T&E Species 11-5
Approach Used to Characterize Risks of COCs to T&E Species 11-6
Input Distributions for the Exposure Modeling of Bald Eagle 11-16
Input Distributions for the Exposure Modeling of American Bittern 11-17
Input Distributions for the Exposure Modeling of Small-Footed Myotis .... 11-17
Median Concentrations of tPCBs in Prey of Bald Eagles 11-21
Median Concentrations of TEQ in Prey of Bald Eagles 11-21
Median Concentrations of tPCBs in Prey of American Bittern 11-22
Median Concentrations of TEQ in Prey of American Bittern 11-22
Total Daily Intake (TDI) of tPCBs by Bald Eagles in the Housatonic River
Primary Study Area 11-25
Total Daily Intake (TDI) of TEQ by Bald Eagles in the Housatonic River
Primary Study Area 11-25
Bald Eagle Egg Exposure to PCBs in the Housatonic River Primary Study
Area 11-26
Bald Eagle Egg Exposure to TEQ in the Housatonic River Primary Study
Area 11-26
Total Daily Intake (TDI) of tPCBs by American Bittern in Reach 5A of
the Housatonic River Primary Study Area 11-28
Total Daily Intake (TDI) of tPCBs by American Bittern in Reach 5B of
the Housatonic River Primary Study Area 11-28
O:\20123001.096\ERA PB\ERA PB TOC.DOC
XXX11
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF FIGURES
(Continued)
Title Page
Figure 11.2-14 Total Daily Intake (TDI) of tPCBs by American Bittern in Reach 5C of
the Housatonic River Primary Study Area 11-29
Figure 11.2-15 Total Daily Intake (TDI) of tPCBs by American Bittern in Reaches 5D
and 6 of the Housatonic River Primary Study Area 11-29
Figure 11.2-16 Total Daily Intake (TDI) of TEQ by American Bittern in the Housatonic
River Primary Study Area 11-30
Figure 11.2-17 American Bittern Egg Exposure to tPCBs in Reach 5 A of the Housatonic
River Primary Study Area 11-30
Figure 11.2-18 American Bittern Egg Exposure to tPCBs in Reach 5B of the Housatonic
River Primary Study Area 11-31
Figure 11.2-19 American Bittern Egg Exposure to tPCBs in Reach 5C of the Housatonic
River Primary Study Area 11-31
Figure 11.2-20 American Bittern Egg Exposure to tPCBs in Reaches 5D and 6 of the
Housatonic River Primary Study Area 11-32
Figure 11.2-21 American Bittern Egg Exposure to TEQ in Reaches 5 and 6 of the
Housatonic River Primary Study Area 11-32
Figure 11.2-22 Total Daily Intake (TDI) of tPCBs by Small-Footed Myotis in Reach 5 of
the Housatonic River Primary Study Area 11-33
Figure 11.2-23 Total Daily Intake (TDI) of TEQ by Small-Footed Myotis in Reach 5 of
the Housatonic River Primary Study Area 11-34
Figure 11.3-1 Dose-Response Curve for Effects of tPCBs on Mortality at Birth of Rats... 11-43
Figure 11.3-2 Dose-Response Curve for Effects of TEQ on Mortality at Birth of Rats 11-44
Figure 11.4-1 Risk Curves for Bald Eagles Exposed to tPCBs in the Housatonic River
Primary Study Area 11-50
Figure 11.4-2 Risk Curves for Bald Eagles Exposed to TEQ in the Housatonic River
Primary Study Area 11-50
Figure 11.4-3 Risk for Bald Eagle Eggs Exposed to tPCBs in the Housatonic River
Primary Study Area 11-51
Figure 11.4-4 Risk for Bald Eagle Eggs Exposed to TEQ in the Housatonic River
Primary Study Area 11-51
Figure 11.4-5 Risk Curves for American Bittern Exposed to tPCBs in Reach 5 A of the
Housatonic River Primary Study Area 11-52
Figure 11.4-6 Risk Curves for American Bittern Exposed to tPCBs in Reach 5B of the
Housatonic River Primary Study Area 11-52
O:\20123001.096\ERA PB\ERA PB TOC.DOC ••• 7/12/2003
- - XXX111
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
LIST OF FIGURES
(Continued)
Title Page
Figure 11.4-7 Risk Curves for American Bittern Exposed to tPCBs in Reach 5C of the
Housatonic River Primary Study Area 11-53
Figure 11.4-8 Risk Curves for American Bittern Exposed to tPCBs in Reaches 5D and 6
of the Housatonic River Primary Study Area 11-53
Figure 11.4-9 Risk Curves for American Bittern Exposed to TEQ in the Housatonic
River Primary Study Area 11-54
Figure 11.4-10 Risk for American Bittern Eggs Exposed to tPCBs in Reach 5A of the
Housatonic River Primary Study Area 11-54
Figure 11.4-11 Risk for American Bittern Eggs Exposed to tPCBs in Reach 5B of the
Housatonic River Primary Study Area 11-55
Figure 11.4-12 Risk for American Bittern Eggs Exposed to tPCBs in Reach 5C of the
Housatonic River Primary Study Area 11-55
Figure 11.4-13 Risk for American Bittern Eggs Exposed to tPCBs in Reaches 5D and 6 of
the Housatonic River Primary Study Area 11-56
Figure 11.4-14 Risk for American Bittern Eggs Exposed to TEQ in the Housatonic River
Primary Study Area 11-56
Figure 11.4-15 Risk Curves for Small-Footed Myotis Exposed to tPCBs in Reach 5 of the
Housatonic River Primary Study Area 11-57
Figure 11.4-16 Risk Curves for Small-Footed Myotis Exposed to TEQ in Reach 5 of the
Housatonic River Primary Study Area 11-57
Figure 12.2-1 Hazard Quotients for Aquatic Biota Exposed to tPCBs in the Housatonic
River PSA 12-19
Figure 12.2-2 Hazard Quotients for Fish Exposed to 2,3,7,8-TCDD TEQ in the
Housatonic River PSA 12-22
Figure 12.2-3 Hazard Quotients for Wildlife Exposed to tPCBs in the Housatonic River
PSA 12-24
Figure 12.2-4 Hazard Quotients for Wildlife Exposed to 2,3,7,8-TCDD TEQ in the
Housatonic River PSA 12-25
Figure 12.2-5 Assessment of Risk to Benthic Invertebrates Exposed to tPCBs
Downstream of Woods Pond 12-30
Figure 12.2-6 Assessment of Risk to Amphibians in Floodplain Exposed to tPCBs
Downstream of Woods Pond in Massachusetts 12-31
Figure 12.2-7 Assessment of Risk to Amphibians in Sediment Exposed to tPCBs
Downstream of Woods Pond in Connecticut 12-32
O:\20123001.096\ERA PB\ERA PB TOC.DOC • 7/12/2003
- - XXXIV
-------
LIST OF FIGURES
(Continued)
Title Page
1 Figure 12.2-8 Assessment of Risk to Warmwater Fish Exposed to tPCBs Downstream of
2 Woods Pond 12-34
3 Figure 12.2-9 Assessment of Risk to Trout Exposed to tPCBs Downstream of Woods
4 Pond 12-36
5 Figure 12.2-10 Assessment of Risk to Mink Exposed to tPCBs Downstream of Woods
6 Pond 12-38
7 Figure 12.2-11 Assessment of Risk to Otter Exposed to tPCBs Downstream of Woods
8 Pond 12-40
9 Figure 12.2-12 Assessment of Risk to Bald Eagle Exposed to tPCBs Downstream of
10 Woods Pond 12-42
11
O:\20123001.096\ERA PB\ERA PB TOC.DOC 7/12/2003
- - XXXV
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Ecological Risk Assessment
EXECUTIVE SUMMARY
ES. ECOLOGICAL RISK ASSESSMENT
EXECUTIVE SUMMARY
Highlights of the ERA
¦ Total PCBs and other COCs in the PSA of the Housatonic River pose
unacceptable risks to some assessment endpoints.
¦ Risk is high for benthic invertebrates, amphibians, and piscivorous
mammals. Confidence in this conclusion is high because (1) multiple
lines of evidence with concordant results were available, (2) models used
to estimate risk were not conservative, and (3) consideration of
uncertainty indicates a high probability of effects.
¦ Risk is moderate to high for some piscivorous and carnivorous birds,
omnivorous and carnivorous mammals, and high for selected threatened
and endangered bird and mammal species. There is uncertainty
regarding these conclusions because corroborating lines of evidence
were generally not available.
¦ Risk is low to moderate for fish and confidence in this conclusion is
moderate.
¦ Risk is low for insectivorous birds, but confidence in this conclusion is not
high.
¦ Other species not included in the quantitative risk assessments may also
be at risk in the PSA.
¦ Assessment of risks to the most susceptible endpoints downstream of the
PSA indicates that benthic invertebrates, amphibians, warmwater and
coldwaterfish, mink, river otter, and bald eagles may also be at risk.
ES.1 OVERVIEW
The purpose of this ecological risk assessment (ERA) is to characterize and quantify the current
and potential risks to biota exposed to contaminants of potential concern (COPCs) in the
Housatonic River below the confluence of the East and West Branches (known as the "Rest of
River"), focusing on polychlorinated biphenyls (PCBs) and other hazardous substances
originating from the General Electric Company (GE) facility in Pittsfield, MA.
This information is synthesized, through a weight-of-evidence approach, into a discussion of the
nature and magnitude of the risks for the assessment endpoints, and the uncertainties associated
with the characterization of these risks. Multiple lines of evidence for each assessment endpoint
are evaluated, including where applicable or available:
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg j 7/11/2003
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1
2
3
4
¦ Field surveys/studies.
¦ Toxicity tests.
¦ Comparison of effects in the literature to a site-specific exposure model.
5 ES.2 SITE DESCRIPTION AND HISTORY
6 The Housatonic River flows from east of Pittsfield, MA, to Long Island Sound and drains an area
7 of approximately 1,950 square miles in Massachusetts, New York, and Connecticut (Figure
8 ES-1). The river is located in a predominantly rural area of western Massachusetts and
9 Connecticut, where farming was the main occupation from colonial settlement through the late
10 1800s. The entire site, known as the General Electric/Housatonic River Site, consists of the 254-
11 acre (103-hectare) GE manufacturing facility; the Housatonic River and associated riverbanks
12 and floodplain from Pittsfield, MA, to Long Island Sound; as well as other properties or areas
13 that have become contaminated as a result of GE facility operations.
14 Widespread contamination of the river downstream of the GE facility has resulted from the
15 transport of PCB-contaminated river sediment and floodplain soil by river flow, sediment
16 transport, and overbank flooding. Numerous studies conducted since 1988 have documented
17 PCB contamination of soil within the floodplain of the Housatonic River downstream of the GE
18 facility. PCBs have been detected in river sediment in Massachusetts as far downstream as the
19 border with Connecticut (BBL 1995), and in Connecticut as far as the Derby-Shelton Dam and
20 beyond into Long Island Sound (other sources have been identified downstream of this dam).
21 The PCBs detected in Housatonic River floodplain soil and sediment consist of predominantly
22 Aroclor 1260, with a minor contribution of Aroclor 1254.
23 The GE facility in Pittsfield is the only known source of PCBs found in the Housatonic River
24 sediment and floodplain soil in Massachusetts. GE began operations in its present location in
25 1903. Three manufacturing divisions have operated at the GE facility (Transformer, Ordnance,
26 and Plastics). Although GE performed many functions at the Pittsfield facility throughout the
27 years, the activities of the Transformer Division, including the construction and repair of
MK01\O:\20123001.096\ERA_PB\ERA_PB_ES.DOC
ES-2
7/11/2003
-------
Long Island Sound
LEGEND:
~
Housatonic River
Housatonic River Basin
A/ Primary Road
N
V
s
4
0 4
8
Miles
5
0 5 10
15
Kilometers
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-1
HOUSATONIC RIVER
| O:\gepitt\aprs\era_figures2.apr | layout - fig 1.1-11 o:\gepitttepsfiles\plots\in\fig_es1.eps 111:25 AM, 7/10/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Ecological Risk Assessment
EXECUTIVE SUMMARY
electrical transformers using dielectric fluids, some of which contained PCBs (primarily Aroclor
1260 and, to a lesser extent, 1254), were one likely significant source of PCB contamination.
Because of its size and complexity, the GE/Housatonic River Site has been divided into several
areas for investigation and cleanup. The Rest of River is the portion of the river from the
confluence of the East and West Branches of the Housatonic River (the confluence) to the
Massachusetts border with Connecticut, a distance of approximately 54 miles (87 km), and
beyond into Connecticut to Long Island Sound. In addition to the river itself, the Rest of River
includes the associated riverbank and floodplain. The lateral extent of the area under
investigation includes the floodplain extending to the 1-ppm total PCB (tPCB) isopleth, which is
approximately equivalent to the 10-year floodplain. The Rest of River portion of the Housatonic
River flows through one of the most biologically diverse regions of Massachusetts and
Connecticut. Dams play an integral role in the downstream migration of PCBs and other COPCs
from the GE facility.
The ERA focuses on the portion of the river from the confluence, 2 miles (3 km) below the GE
facility, to Woods Pond Dam, a distance of approximately 11 river miles (17.7 km). This area is
referred to as the Primary Study Area (PSA) (Figure ES-2), and is where much of the PCB
contamination was found in previous studies. The ERA also includes an evaluation of the river
and floodplain downstream of the PSA to the Derby-Shelton Dam in Connecticut, approximately
13 miles upstream from Long Island Sound.
The first 10.5 miles (16.9 km) from the confluence to the headwaters of Woods Pond is referred
to as Reach 5. Next to the initial 0.5-mile (0.8-km) reach bordering the GE facility, Reach 5 has
the highest concentrations and highest frequency of detections of PCBs in sediment. Reach 5 is
subdivided further into four segments: Reach 5 A, from the confluence to just above the Pittsfield
Wastewater Treatment Plant (WWTP); Reach 5B, from the WWTP to Roaring Brook; Reach 5C,
from Roaring Brook to the headwaters of Woods Pond; and Reach 5D, the backwaters above
Woods Pond. Reach 6 begins 10.5 miles (16.9 km) downstream of the confluence at Woods
Pond. The pond is approximately 0.2 mile (0.3 km) in length and has an area of 60 acres (24 ha).
This reach contains the first impoundment downstream from the GE facility and is a depositional
environment.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg ^
-------
-Jf
Silver
Lake [¦
1
GE Facility
^
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Ecological Risk Assessment
EXECUTIVE SUMMARY
In the PSA, the river channel ranges from 40 to 125 ft (12 to 38 m) in width, is bordered by
extensive floodplain (up to 3,000 ft [900 m] wide), and has a meandering pattern with numerous
oxbows and backwaters. Woods Pond is a shallow 54-acre (22-ha) impoundment that was
formed by the construction of a dam in 1864.
Reach 7, the first reach south (downstream) of the PSA, begins below Woods Pond Dam and
flows for 18.6 miles (29.9 km), ending at the headwaters of Rising Pond, which is Reach 8.
Reach 9 begins downstream of Rising Pond and extends for approximately 24.6 miles (39.6 km)
to the Massachusetts/Connecticut state line. This reach is wide with a flat floodplain and many
oxbows, and agriculture is a predominant land use.
Reach 10 begins at the Massachusetts/Connecticut border and extends 7.4 miles (12 km) to the
dam at Great Falls Village. Reach 11, which begins on the downstream side of the dam at Great
Falls and ends 11.5 miles downstream at Cornwall Bridge, is mostly shallow and fast flowing,
and much of the reach is designated as a Trout Management Area. Reach 12 extends from
Cornwall Bridge to the dams at Bulls Bridge, a length of 13.1 miles (21.1 km). Reach 13 starts
on the downstream side of the dams at Bulls Bridge and runs 10.9 miles (17.5 km) to the now-
submerged Bleachery Dam at New Milford, CT. Downstream of this point the river is largely
confined to a series of large lakes formed by power dams.
Reach 14, from the Bleachery Dam to Shepaug Dam, is known as Lake Lillinonah. Water
movement is slow through this reach and the river is deep. Reach 15 encompasses Lake Zoar,
from Shepaug Dam to Stevenson Dam. Some homes and boat launches are found on Lake Zoar.
Reach 16 is Lake Housatonic, which is formed by the Derby-Shelton Dam. The remaining 13
miles of the river, from Derby-Shelton Dam to Long Island Sound, is tidally influenced and has
other industrial sources ofPCBs.
The land uses of the floodplain properties in Massachusetts include residential,
commercial/industrial, agricultural, recreational (such as canoeing, fishing, and hunting), wildlife
management, and parks and a golf course. The Housatonic River floodplain is an attractive area
for recreation, including fishing and waterfowl hunting.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg g
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Ecological Risk Assessment
EXECUTIVE SUMMARY
The State of Connecticut posted a fish consumption advisory for most of the Connecticut section
of the river in 1977 as a result of the PCB contamination in the river sediment and fish tissue. In
1982, the Massachusetts Department of Public Health (MADPH) issued a consumption advisory
for fish, frogs, and turtles for the Housatonic River. In addition, in 1999, MADPH issued a
waterfowl consumption advisory from Pittsfield to Great Barrington due to PCB concentrations
in wood ducks and mallards collected from the river by the U.S. Environmental Protection
Agency (EPA).
ES.3 REGULATORY BACKGROUND
The GE/Housatonic River site has been subject to regulatory investigations dating back to the
early 1980s. In 1991, EPA issued a RCRA Corrective Action Permit to the GE Pittsfield facility.
Following appeals by GE and others and subsequent modification, the permit became effective in
1994. The permit included the 254-acre facility, some filled former oxbows, Silver Lake, the
Housatonic River and its floodplains and adjacent wetlands, and all sediment contaminated by
PCBs migrating from the GE facility.
EPA proposed the site to the Superfund National Priorities List (NPL) in September 1997.
Several federal and state government agencies and GE entered into negotiations late in 1997 with
the goal of reaching a comprehensive settlement, which included remediation, redevelopment,
and restoration components.
In September 1998, representatives of the federal and state government agencies, GE, the City of
Pittsfield, and the Pittsfield Economic Development Authority reached an agreement in
principle. This agreement was translated into a Consent Decree that was entered by the court on
27 October 2000. The agreement provides for, in general, the cleanup of the GE plant facility
and surrounding areas that have become contaminated as a result of facility operations, including
the Housatonic River.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg j
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1 The GE/Housatonic River site is made up of several separate response actions (as described in
2 the Consent Decree), including three actions in the river:
3 ¦ Upper V2-Mile Reach Housatonic River Removal Action ('/2-Mile Reach)
4 "1 /4-Mile Housatonic River Removal Action (1 /4-Mile Reach)
5 ¦ Rest of River
6
7 The primary COPCs are PCBs; other COPCs include volatile organics, dioxins/furans,
8 polycyclic aromatic hydrocarbons (PAHs), semivolatiles, and metals. Certain PCB congeners
9 (known as coplanar or dioxin-like congeners) exhibit a mechanism of toxicity similar to that of
10 the most toxic dioxin congener (2,3,7,8-TCDD). The combined toxicity of these coplanar
11 congeners can be expressed and evaluated as the equivalent toxicity of 2,3,7,8-TCDD using the
12 concept of toxic equivalence (TEQ).
13 EPA completed an investigation of the Rest of River below the l^-Mile Reach into Connecticut,
14 which focused on collecting information for and preparing the human health and ecological risk
15 assessments, and modeling PCB fate and transport in the river. The ecological risk assessment,
16 together with the human health risk assessment and the model of PCB fate, transport, and
17 bioaccumulation, will inform EPA's decision on what additional remedial actions, if any, may be
18 required in the future.
19 ES.4 OVERVIEW OF TECHNICAL APPROACH
20 This ERA follows the eight-step technical approach and guidelines detailed in EPA's Ecological
21 Risk Assessment Guidance for Superfund: Process for Designing and Conducting Ecological
22 Risk Assessments. The first two steps of the ERA process (Screening-Level Problem
23 Formulation and Screening-Level Exposure Estimate and Risk Calculation) were first addressed
24 in the Upper Reach-Housatonic River Ecological Risk Assessment and subsequently refined in
25 Appendix B of this document. Steps 3, 4, and 5 (Baseline Problem Formulation, Study Design
26 and DQO Process, and Verification of Field Data Analysis) are iterative components of the
27 eight-step ERA process. These three steps were initially presented in the Supplemental
28 Investigation Work Plan for the Lower Housatonic River and were modified as necessary during
29 the data collection phase of the project. Steps 6 and 7 (Site Investigation and Data Analysis and
30 Risk Characterization) are presented in detail in this Ecological Risk Assessment report. Step 8
MK01\O:\20123001.096\ERA_PB\ERA_PB_ES.DOC "C C Q 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ecological Risk Assessment
EXECUTIVE SUMMARY
(Risk Management) will be addressed after the ERA has undergone Peer Review. A roadmap for
the ERA is provided in Figure ES-3. A brief overview of each of the eight steps, particularly as
they relate to this document, is presented below.
Problem formulation is an important component of the ERA process that establishes the goals,
objectives, and scope for the ERA. The problem formulation portion of the ERA is discussed in
Section 2 and was outlined in the Supplemental Investigation Work Plan for the Lower
Housatonic River.
An extensive physical and ecological characterization of the Housatonic River is presented in
Section 2.2 and Appendix A (Ecological Characterization) of this document. These sections
detail the physical setting, habitats, and biotic communities of the river in both the aquatic and
terrestrial environments.
Investigations of the nature and extent of contaminants in the Housatonic River watershed have
previously been conducted by GE, EPA, and others. PCBs have been identified as the main
COPC, and other contaminants such as dioxins and furans and PAHs have also been identified at
the GE facility. In Section 2.3, the sources, amounts, and patterns of contaminant releases and
receiving bodies are presented.
The purpose of the Pre-ERA was to identify contaminants that warranted more detailed analyses
in the ERA, and those that could be removed from further consideration because they pose
minimal risk. A summary of the Pre-ERA is provided in Section 2.4. The complete Pre-ERA is
included as Appendix B to this document.
An overview of the environmental behavior of PCBs and other COPCs is presented in Section
2.5. This section includes discussions of the transport of the contaminants from their point(s) of
release, partitioning behavior in different media, and biotic and abiotic degradation in these
media.
The effects and mechanisms of toxicity to biota of the contaminants identified as COPCs within
the Housatonic River and floodplain are discussed, with an emphasis on PCBs in Section 2.6,
and in further detail in the effects assessment portion of each assessment endpoint section and
corresponding appendix.
MK01\O:\20123001.096\ERA_PB\ERA_PB_ES.DOC T? C C\ 7/11/2003
-------
Section 3
Benthic
Invertebrates
(Appendix D)
Section 4
Amphibians
(Appendix E)
Section 5
Fish
(Appendix F)
Section 7
Insectivorous
Birds
(Appendix G)
Section 8
Piscivorous
Birds
(Appendix H)
Section 9
Piscivorous
Mammals
(Appendix I)
Section 10
Omnivorous/
Carnivorous
Mammals
(Appendix J)
Section 11
Threatened
& Endangered
Species
(Appendix K)
Section 12
Risk Summary
Relevant to all endpoints + Relevant to specific assessment
endpoints
j j Wildlife assessment approach
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
Figure ES-3
MK01|O:\20123001.096\ERA_PB\ERA_PB_1_Fig ES-3.ppt
Ecological Risk Assessment Roadmap
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Ecological Risk Assessment
EXECUTIVE SUMMARY
Assessment and measurement endpoints are defined and described in Section 2.8. An
assessment endpoint is defined as the "explicit expression of the environmental value that is to
be protected." A measurement endpoint is defined as "a measurable ecological characteristic
that is related to the valued characteristic chosen as the assessment endpoint," and is a measure
of biological effects (e.g., mortality, reproduction, growth).
Assessment Endpoint Selected for the Rest of River ERA
¦ Community structure, survival, growth, and reproduction of benthic invertebrates.
¦ Community condition, survival, reproduction, development, and maturation of
amphibians.
¦ Survival, growth, and reproduction offish.
¦ Survival, growth, and reproduction of insectivorous birds.
¦ Survival, growth, and reproduction of piscivorous birds.
¦ Survival, growth, and reproduction of piscivorous mammals.
¦ Survival, growth, and reproduction of omnivorous and carnivorous mammals.
¦ Survival, growth, and reproduction of Special Status Species (Endangered,
Threatened).
The conceptual model outlined in Section 2.7 describes the relationship between the COPCs and
the assessment endpoints. Section 2.9 describes the analytical approach used to estimate risks
and the weight-of-evidence approach used to develop the conclusions.
Sections 3 through 11 (and their corresponding appendices) provide an overview of the exposure
assessment, the effects assessment, and the risk characterization for each representative species
or representative group of species. The exposure assessment sections include a description of the
data collection activities and the studies conducted to determine concentrations of COPCs in
water, soil, sediment, and biota samples. The list of COPCs was further narrowed with additional
screening for the specific endpoint, resulting in the list of contaminants of concern (COCs)
retained for that endpoint risk assessment.
The effects assessment sections begin with an overview of the toxicity of PCBs and the other
COCs. For each major representative species group and COC, the effects literature was
reviewed. The goal of this review was to identify studies that could be used to develop effects
metrics for use in risk characterization. The effects metrics developed ranged from
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Ecological Risk Assessment
EXECUTIVE SUMMARY
concentration- or dose-response curves to benchmarks depending on the quality and relevance of
the data available.
The risk characterization sections first provide an overview of site-specific studies, analyses of
the results, and conclusions, and then consider the three major lines of evidence (where
available) using a weight-of evidence (WOE) approach. There is also a discussion of the
uncertainties associated with the assessment for that endpoint, an evaluation of potential risks to
receptors other than those chosen as the representative species, and a discussion of potential risks
downstream of the PSA for receptors of concern.
Section 12 summarizes the conclusions of the ERA, and presents a discussion of these
conclusions in the context of the uncertainties and other factors that could not be expressly
quantified in the evaluation of the assessment endpoints.
ES.5 OVERVIEW OF THE ASSESSMENT ENDPOINT CONCLUSIONS
ES.5.1 Risks in the Primary Study Area
The assessments in the ERA were conducted using various lines of evidence including, in many
cases, different measurement endpoints and effects metrics. These lines of evidence—defined as
information derived from different sources that can be used to describe and interpret risk—were
integrated into a graphical representation of risk using the weight-of-evidence approach. The
WOE provides an objective process by which measurement endpoints are related to an
assessment endpoint to evaluate whether significant risk is posed to the environment. A formal
WOE can range from a simple qualitative assessment to a highly quantitative evaluation;
however, no matter what form the weight-of-evidence takes, it should provide documentation of
the thought process used when assessing potential ecological risk.
In the first step of the WOE approach, weights are assigned to measurement endpoints based on
10 attributes related to: (1) strength of association between assessment and measurement
endpoints; (2) data and study quality; and (3) study design and execution. The second step of the
approach is to evaluate the magnitude of response in the measurement endpoint, considering
whether the measurement endpoint indicates the presence or absence of harm, and if the
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg ^2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Ecological Risk Assessment
EXECUTIVE SUMMARY
magnitude of response is low, intermediate, or high. The WOE process concludes by plotting the
output of the two preceding steps in a matrix for all measurement endpoints associated with a
given assessment endpoint. The matrix allows easy visual examination of agreements or
divergences among measurement endpoints, facilitating interpretation with respect to the
assessment endpoint.
The results of the WOE process and the final WOE matrix are summarized below for each of the
eight assessment endpoints considered in this ERA. Following that discussion, to facilitate
comparison of risks among aquatic life and wildlife receptors and to give an overview of the
findings of the risk assessment, assessment results are converted to probabilistic hazard quotients
(HQs). A HQ is the expected environmental concentration or dose of a contaminant divided by
its estimated low or no toxic effect concentration or dose. Higher quotients indicate greater risk.
The methods used to calculate the probabilistic HQs and the results of these analyses for each
endpoint are summarized in this section.
ES.5.1.1 Aquatic Assessment Endpoints
Benthic Invertebrates—The weight-of-evidence results for the benthic invertebrate assessment
endpoint are shown in Table ES-1. In this table, the measurement endpoints for the three lines of
evidence: water, sediment, and tissue chemistry (C), toxicity tests (T), benthic community
measures (B), are listed, as are the weight of the measurement endpoint evidence of harm, and
magnitude of response. This table indicates that the majority of endpoints suggest some risk for
benthic communities in both coarse- and fine-grained sediment. The conclusion is that there is a
moderate to high risk to much of the benthic community indicated by the WOE evaluation.
Amphibians—The results of the weight-of-evidence assessment for amphibians are presented in
Table ES-2. In the amphibian weight-of-evidence matrix, the measurement endpoints for the
three lines of evidence: the tissue chemistry (C); wood frog toxicity tests (W) and leopard frog
toxicity tests (L); and field surveys (B) are listed. As shown on the table, many of the endpoints
indicated some degree of risk. The potential for two amphibian studies conducted for GE to
determine risk to amphibians was judged to be undetermined due to limitations in the study
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1 Table ES-1
2
3 Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related to Maintenance of a Healthy
4 Benthic Community
Measurement Endpoints
Weighting
Value (High,
Moderate,
Low)
Coarse-Grained Sediment
Fine-Grained Sediment
Evidence of
Harm (Yes, No,
Undetermined)
Magnitude (High,
Intermediate, Low)
Evidence of
Harm (Yes, No,
Undetermined)
Magnitude (High,
Intermediate, Low)
C. Chemical Measures
C-l. Concentration of PCB in overlying water in relation to
levels reported to be harmful to benthic invertebrates
Low/Moderate
Yes
Intermediate
Yes
Intermediate
C-2. Concentration of PCB in the sediment in relation to
levels reported to be harmful to benthic invertebrates
Low/Moderate
Yes
High
Yes
High
C-3. Concentration of PCB in invertebrate tissues in relation
to levels reported to be harmful to benthic invertebrates
Moderate
Yes
Intermediate
Yes
Intermediate
T. Toxicological Measures
T-l. Sediment toxicity to multiple invertebrate species, as
measured in laboratory toxicity tests
Moderate/
High
Yes
High
Yes
High
T-2. Sediment toxicity to multiple invertebrate species, as
measured in in situ toxicity tests
Moderate/
High
Yes
Intermediate
Yes
High
T-3. Indications of PCB as toxicity driver in toxicity
identification evaluations
Moderate
Undetermined
—
Yes
Intermediate
B. Benthic Community Measures
B-l. Abundance, richness, and biomass of invertebrates,
relative to reference stations of comparable substrate and
habitat (ANOVA)
Moderate
Yes
Intermediate
No
B-2. Benthic community structure, as assessed using a
multivariate assessment of key benthic metrics (rank analysis
and MDS)
Moderate
Yes
Intermediate
No
B-3. Water quality assessment using modified Hilsenhoff
Biotic Index (MHBI) indicator of organic pollution
Moderate
No
—
No
—
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
Table ES-2
Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related to Maintenance of Amphibian
Populations in the Housatonic River PSA
Measurement Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
C. Chemical Measures
C. Concentration of PCB in frog tissues in relation to concentrations
reported to be harmful to amphibians
Moderate
Yes
Low
W. Wood Frog Toxicological Measures
W-l. Sediment toxicity to hatchling/late embryo life stages
Mod/High
No
-
W-2. Sediment toxicity to larval life stages
Mod/High
Yes
Intermediate
W-3. Sediment toxicity to late larval/metamorph life stage
Mod/High
Yes
High
W-4. GE Study (juvenile wood frogs)
Low
Undetermined
-
L. Leopard Frog Toxicological Measures
L-l. Sediment toxicity to hatchling/late embryo life stages
Mod/High
Yes
Low
L-2. Sediment toxicity to larval life stages
Mod/High
Yes
High
L-3. Sediment toxicity to late larval/metamorph life stage
Mod/High
Yes
High
L-4. Sediment toxicity to adult leopard frogs (reproductive health)
Mod/High
Yes
High
B. Biology
B-l. Vernal pool community study
Mod/High
Yes
Low
B-2. GE leopard frog egg mass survey
Low/Mod
Undetermined
-
B-3. Anecdotal observations during collections for reproductive
study
Moderate
Yes
Low
MK01\O:\20123001.096\ERA_PB\ERA_PB_ES.DOC
ES-15
7/11/2003
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1 designs. The only endpoint that did not indicate potential risk was the earliest life stage wood
2 frog toxicity endpoint, for which there is mechanistic explanation for the lack of response. The
3 plots also indicate that four endpoints exhibited a high degree of risk combined with a moderate
4 to high confidence rating.
5 In addition, a population model was constructed for wood frogs to determine whether effects
6 from PCBs on individual wood frogs influence the populations within the PSA. A 10-year
7 simulation, both with and without the effects of PCBs, was conducted. The model demonstrated
8 that effects observed in the toxicity studies would result in population-level impacts.
9 The conclusion is that there is a significant risk to amphibians as indicated by the preponderance
10 of the evidence, the relative weights of the measurement endpoints, and the population modeling.
11 The "no harm" value of measurement endpoint W-l does not diminish the overall conclusion,
12 because the study demonstrated that the embryo/early larval life stages are fairly insensitive to
13 the effects of maternally transferred PCBs relative to later juvenile life stages exposed to
14 contaminated media.
15 Fish—The weight-of-evidence results for fish in the PSA are shown in Table ES-3. In the fish
16 WOE matrix, the measurement endpoints for the three lines of evidence: site-specific toxicity
17 tests (A); fish tissue chemistry (B); and field surveys (C) are listed. This table illustrates that
18 although a high probability of adverse impacts to fish from tPCBs and/or TEQ is predicted
19 throughout the PSA, the impacts predicted are for sensitive sublethal endpoints (reproduction
20 and development); mortality of adults is unlikely. Therefore, the magnitude of impact is not
21 predicted to be catastrophic in any reach; adverse effects, although high in probability, are
22 generally expected to be subtle. The field studies conducted in the PSA (fish community and
23 reproduction studies) support lack of catastrophic effects, but cannot be used to assess lesser
24 impacts.
25 ES.5.1.2 Wildlife Assessment Endpoints
26 Insectivorous Birds—The WOE results for exposure of insectivorous birds to tPCBs are
27 presented in Table ES-4, and for exposure to TEQ in Table ES-5. Two lines of evidence are
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg jg
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1
2 Table ES-3
3
4 Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related
5 to Maintenance of a Healthy Fish Community
Measurement Endpoints
Weighting
Value
(High,
Moderate, Low)
Evidence of Harm (Yes,
No, Undetermined)
Magnitude (High,
Intermediate, Low)
A. Site-Specific Toxicity
Al. Reproductive success relative to reference
Mod/High
Yes
Low
A2. Reproductive success dose-response
High
Yes
Intermediate
B. Fish Body Burden
Bl. Observed fish tissue/ Literature toxic levels
Mod
Yes
Low
B2. Observed fish tissue/ Phase I toxic levels
Mod/High
Yes
Low
B3. Observed fish tissue/ Phase II toxic levels
Mod/High
Yes
Low
C: Fish Community and Reproduction Studies
CI: EPA Study and GE Community Study
Low/Mod
Undetermined
-
C2: GE Reproduction Study
Low/Mod
Undetermined
-
6
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg j j
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1
2 Table ES-4
3
4 Evidence of Harm and Magnitude of Effects for Insectivorous Birds Exposed to
5 tPCBs in the Housatonic River PSA
Measurement Endpoints
Weighting Value
(High, Moderate, Low)
Evidence of Harm
(Yes, No,
Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Study
High (Tree swallow)
Moderate/High
(American robin)
No (Tree swallow)
No (American robin)
Low (Tree swallow)
Low (American robin)
Modeled Exposure and
Effects
Moderate
Yes
High
6
7
8
9 Table ES-5
10
11 Evidence of Harm and Magnitude of Effects for Insectivorous Birds Exposed to
12 TEQ in the Housatonic River PSA
Measurement Endpoints
Weighting Value
(High, Moderate, Low)
Evidence of Harm
(Yes, No,
Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Study
High (Tree swallow)
Moderate/High
(American robin)
No (Tree Swallow)
No (American robin)
Low (Tree Swallow)
Low (American robin)
Modeled Exposure and
Effects
Moderate
Yes
Intermediate
13
14
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg jg
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Ecological Risk Assessment
EXECUTIVE SUMMARY
presented in the table, the field studies, and modeled exposures and effects. The results from the
modeled exposure and effects line of evidence suggest that tPCBs and TEQ pose intermediate to
high risks to tree swallows living in the PSA. However, the field study line of evidence suggests
that, if effects are occurring, they are minor. The uncertainty concerning the field-based
threshold range for tPCBs likely means that risks of this COC are overestimated for the PSA.
Even the upper end of the tPCB range is associated with equivocal evidence for adverse effects
to tree swallows. For TEQ, the threshold range is quite broad. The available evidence from field
studies indicates that tree swallows are tolerant to exposure to persistent organochlorines such as
tPCBs and TEQ. If the tree swallow threshold is near the upper end of the threshold range, then
the current modeled exposure and effects line of evidence is overestimating risks of TEQ to tree
swallows. Thus, the WOE assessment supports a finding of low risk for tree swallows exposed
to tPCBs and TEQ in the PSA. This conclusion, however, is uncertain because of the conflicting
results in the WOE assessment.
The results from the modeled exposure and effects lines of evidence suggest that tPCBs and TEQ
pose an intermediate to high risk to American robins inhabiting the PSA of the Housatonic River.
The American robin field study, however, suggests that reproductive success is not being
impaired by the tPCBs and TEQ in the PSA. The uncertainty in the modeled exposure and
effects line of evidence, outlined below, likely means the approach overestimates the risks of
tPCBs and TEQ to American robins in the PSA. The WOE assessment, therefore, supports a
conclusion of low risk to American robins exposed to tPCBs and TEQ in the PSA. This
conclusion, however, is uncertain because of the conflicting results in the weight-of-evidence
assessment.
Piscivorous Birds—The WOE analysis indicated that exposure of piscivorous birds, such as the
belted kingfisher and osprey (Tables ES-6 and ES-7), to tPCBs and TEQ in the PSA, could lead
to adverse reproductive effects in some species. The two lines of evidence used to support this
conclusion were the field study of kingfisher productivity and the comparison of modeled
exposure with effects to piscivorous birds.
For the assessment of risks to kingfishers, both lines of evidence were employed. The modeled
exposure and effects line of evidence indicated that kingfishers in the PSA are likely to receive a
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg
-------
1
2
3
4
5
6
7
8
9
10
11
Ecological Risk Assessment
EXECUTIVE SUMMARY
Table ES-6
Evidence of Harm and Magnitude of Effects for Piscivorous Birds Exposed to
tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled Exposure and
Effects
M
Yes (Kingfisher)
Yes (Osprey)
High (Kingfisher)
High (Osprey)
Belted Kingfisher Field
Study (Henning 2002)
M/H
No (Kingfisher)
Low (Kingfisher)
Table ES-7
Evidence of Harm and Magnitude of Effects for Piscivorous Birds Exposed to
TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled Exposure and
Effects
M
Yes (Kingfisher)
Yes (Osprey)
Intermediate (Kingfisher)
Intermediate (Osprey)
Belted Kingfisher Field
Study (Henning 2002)
M/H
No (Kingfisher)
Low (Kingfisher)
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg 2Q 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Ecological Risk Assessment
EXECUTIVE SUMMARY
tPCB dose greater than what the most tolerant species known from the literature can be exposed
to without effects. For TEQ, the risk picture is less clear because the threshold range for this
COC is very wide and the exposure estimates for kingfishers fell within this range. Thus, without
effects data specific to kingfishers, it is difficult to make definitive conclusions about the risks of
TEQ to this species. The field study of kingfisher productivity, however, indicated that these
birds are able to reproduce in the PSA. This line of evidence was given a higher weighting than
the exposure and effects modeling, despite concerns about the field study. Therefore, kingfishers
are considered to be at low risk in the PSA as a result of exposure to tPCBs and TEQ. The
conclusion of low risk to kingfishers is uncertain because the two lines of evidence did not give
concordant results.
For ospreys, only the modeled exposure and effects line of evidence was available to assess risk
to these birds. As with kingfishers, this line of evidence indicated that ospreys in the PSA are
likely to receive a tPCB dose that is greater than what the most tolerant species known from the
literature can be exposed to without effects. The risks due to exposure to TEQ are unclear, as the
estimates for exposure also fell within toxicity threshold range. Ospreys, however, lack a site-
specific study that investigated the effects of COCs in the PSA. The PSA contains suitable
habitat for ospreys, with abundant prey, raising the possibility that they are not resident in the
area because of contaminants. Ospreys are therefore considered to be at risk in the PSA as a
result of exposure to tPCBs and TEQ.
Piscivorous Mammals—The results of the WOE assessment for piscivorous mammals are
presented for tPCB and TEQ, respectively, in Tables ES-8 and ES-9. All three lines of evidence
—field studies, feeding study, and modeled exposure and effects—indicated that the elevated
concentrations of tPCBs and TEQ in the PSA of the Housatonic River are causing adverse
effects of high magnitude to mink and river otter. The field surveys indicated that mink and river
otter are rarely present in the PSA, except during winter, and likely have not established home
territories close to the main channel despite suitable mink and otter habitat. The MSU feeding
study indicated that feeding adult female mink with a diet containing as little as 3.51% fish from
the PSA caused a statistically significant reduction (46% compared to controls) in kit survival to
6 weeks of age. Because mink in the wild typically consume between 20% or more fish in their
diet, the associated risk is correspondingly higher. In addition, other components of the mink
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg 2 J 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Ecological Risk Assessment
EXECUTIVE SUMMARY
Table ES-8
Evidence of Harm and Magnitude of Effects for Piscivorous Mammals Exposed to
tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No,
Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Surveys
EPA
Moderate/High
Yes
High
GE
Moderate
No
Low
Feeding Study
High
Yes
High
Modeled Exposure and
Effects
Moderate/High
Yes
High
Table ES-9
Evidence of Harm and Magnitude of Effects for Piscivorous Mammals Exposed to
TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Surveys
EPA
Moderate/High
Yes
High
GE
Moderate
No
Low
Feeding Study
High
Yes
High
Modeled Exposure and
Effects
Moderate/High
Yes
High
MK01\0:\20123001.096\ERA_PB\ERA_PB_ES.DOC -p-q ryry 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Ecological Risk Assessment
EXECUTIVE SUMMARY
diet in the PSA (e.g., crayfish) have high concentrations of tPCBs and TEQ. Further, the jaw
lesion study indicated that erosion of the jaw occurs at even lower doses and exhibits a dose-
response relationship. Such effects could eventually lead to starvation. The occurrence of jaw
lesions coincides with the induction of Ah-receptor-regulated enzymes (ECOD and EROD) also
in a dose-response manner.
The high risks evident from the feeding study are further supported by the modeled exposure and
effects line of evidence. The estimated potential for exposure is so high that even individual
mink and otter that only forage in the PSA for short periods of time (less than or equal to 10% of
foraging time) are at an intermediate or higher risk from tPCBs and TEQ.
Omnivorous and Carnivorous Mammals—The weight-of-evidence results for omnivorous and
carnivorous mammals are shown in Table ES-10 for tPCB and Table ES-11 for TEQ. Data from
three lines of evidence were available, including field surveys, a population demography field
study of short-tailed shrew, and exposure and effects modeling. The weight-of-evidence analysis
indicates an intermediate risk for short-tailed shrews exposed to tPCBs and TEQ in the PSA.
This conclusion, however, is uncertain because of the lack of definitive findings as to whether
effects are occurring in the field surveys and population demography field study, and the lack of
species-specific measures of effects in the literature. The weight-of-evidence also suggests,
based on one line of evidence for red fox, an intermediate risk to fox exposed to tPCBs and TEQ
in the PSA. This finding is also uncertain because a foraging rate of 50% in Reach 5 was used,
and species-specific measures of effects were not available.
The field surveys, and conclusions made in a study of short-tailed shrew populations at the site
conducted for GE, are not in agreement with the results from the modeling of exposure and
effects line of evidence. However, the results of the supplemental analyses of the data from the
GE study on survival of short-tailed shrews are in agreement with the modeling results,
suggesting that there are intermediate effects from exposure to COCs in the contaminated areas
of the PSA.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg 2^
-------
1
2
3
4
5
6
7
8
9
10
11
12
Ecological Risk Assessment
EXECUTIVE SUMMARY
Table ES-10
Evidence of Harm and Magnitude of Effects for Omnivorous and Carnivorous
Mammals Exposed to tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Field Surveys
Moderate/High
Undetermined
Low
Population
Demography Field
Study
Moderate/High
Undetermined (Shrew)
Intermediate
Modeled Exposure and
Effects
Moderate/High
Yes (Shrew)
Undetermined (Red Fox)
High
Intermediate
Table ES-11
Evidence of Harm and Magnitude of Effects for Omnivorous and Carnivorous
Mammals Exposed to TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Field Surveys
Moderate/High
Undetermined
Low
Population
Demography Field
Study
Moderate/High
Undetermined (Shrew)
Intermediate
Modeled Exposure and
Effects
Moderate/High
No (Shrew)
Undetermined (Red Fox)
Low
Intermediate
MK01\O:\20123001.096\ERA_PB\ERA_PB_ES.DOC tjc O A 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Ecological Risk Assessment
EXECUTIVE SUMMARY
Threatened and Endangered Species—The results of the weight-of-evidence evaluation for
threatened and endangered species using a single line of evidence, modeled exposures and
effects, are shown in Table ES-12 and Table ES-13 for tPCBs and TEQ, respectively. The
results of the risk characterization showed that the highest risk to T&E species is to bald eagles
and American bitterns from exposure to tPCBs. The risk for adult bald eagles exposed to TEQ
was low; however, risk to bald eagle eggs exposed to TEQ was high. These two risk estimates
were combined to yield intermediate risk to bald eagles.
The weight-of-evidence analysis indicates that T&E species such as bald eagle and American
bittern are at risk in the PSA as a result of exposure to tPCBs. Risks to bald eagles and
American bittern exposed to tPCBs are high. There are intermediate risks to bald eagles exposed
to TEQ, and risks to American bittern exposed to TEQ are undetermined. Risks to small-footed
myotis exposed to tPCBs and TEQ are undetermined.
ES.5.1.3 Hazard Quotients
To facilitate comparison of risks among aquatic life and wildlife receptors and to give an
overview of the findings of the risk assessment, assessment results were converted to
probabilistic hazard quotients (HQs). A HQ is the expected environmental concentration or dose
of a contaminant divided by its estimated low or no toxic effect concentration or dose, similar to
a hazard index that is used to describe noncancer risks to people. Higher quotients indicate
greater risk. Figures ES-4 through ES-7 summarize the ranges of hazard quotients for exposure
to tPCBs and TEQ determined for each the assessment endpoints in the PSA. The methods used
to calculate the probabilistic HQs and the results of these analyses for each endpoint are
summarized below.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg 2^
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1 Table ES-12
2
3 Evidence of Harm and Magnitude of Effects of T&E Species Exposed to tPCBs in
4 the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled exposure and
effects, Bald Eagle
Moderate/High
Yes
High
Modeled exposure and
effects, American
Bittern
Moderate/High
Yes
High
Modeled exposure and
effects, Small-Footed
Myotis
Moderate/High
Undetermined
High
5
6
7 Table ES-13
8
9 Evidence of Harm and Magnitude of Effects for T&E Species Exposed to TEQ in
10 the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled exposure and
effects, Bald Eagle
Moderate/High
Yes
Intermediate
Modeled exposure and
effects, American
Bittern
Moderate/High
Undetermined
High
Modeled exposure and
effects, Small-Footed
Myotis
Moderate/High
Undetermined
High
11
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg 2g
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1000
100
e
©
s
a
¦a
N
w
0.001
0.01
Max
MC 75th
l-1"!
Mean
)
Median
~
MC 25th
Min
T
— #
Fish HQs are based on tissue concentrations;
others based on sediment tPCB.
<
PL,
&
c3
O
3
P2
P2
<
PLh
P2
T3
c3
2
P2
P2
<
PLh
o
<
PLh
n
Ph
<
PLh
U
.a
H
PLh
<
PLh
Pi
o
PLh
<
PLh
u
o
PLh
<
PLh
I
&
pi
c3
P2
1
o
<
PLh
C3
P2
£
o
§
h-1
3
PLh
I
2
o
c3
P2
1
o
h-1
<
p2
o
in
in
in
'Sl
i
rn
£
§
§
§
£
'a
£
£
H
Oh
Oh
Oh
B
a
a
a
<
<
<
<
WB-R = Whole body,
reconstituted fish tissue. WB = Whole body fish tissue.
Figure ES-4 Hazard Quotients for Aquatic Biota Exposed to tPBCs in the Housatonic River PSA
MK01\O:\20123001,096\ERA PB\ERA PB ES.DOC
ES-27
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
10
e
©
s
a
¦a
•-
N
w
Max
MC 75th
Mean
Median
MC 25th
Min
X
T
o.i
-o
^ /-N
3 *
pq pq
c ^
£ o
o
!-h
PQ
T3
(L) /->
™ M
h-1 cS
pq
"3 pq
it
M
00 M
J ffl
§d
a B
S) oa
H cd
a3
pq
T3
S3
(L)
3
pq
c
£
o
!-h
pq
pq
WB-R = Whole body,
reconstituted fish tissue. WB = Whole body fish tissue.
2
3 Figure ES-5 Hazard Quotients for Fish Exposed to 2,3,7,8-TCDD TEQ in the Housatonic River PSA
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg 2g
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
10000
1000
100
s
•-P
o
s
a
¦s
•-
10
0.1
0.01
f ii
*
Upper P-Bound 90th Percentile
Monte Carlo 75th Percentile
Monte Carlo Mean
Monte Carlo Median
Monte Carlo 25th Percentile
Lower P-Bound 10th Percentile
¥ i
m
&
K
Pd
K
tZ5
3
£
"c3
£
tZ5
J
2
tZ5
CQ
CQ
s
&
s
&
•8
00
g
00
G
ss
Oh
e
o-
w
o
u
u
o
in
ss
Clh
ss
Clh
&
w
&
w
tZ5
tZ5
tZ5
Q
in
in
Ji
Q
Ja
Q
<
tZ5
<
tZ5
Oh
C/2
1
S-
-G
1
s-
-G
tZ5
1/2
1
S-
-G
*
O
tin
C/2
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1000
100
Upper P-Bound 90th Percentile
Monte Carlo 75th Percentile
Monte Carlo Mean
Monte Carlo Median
Monte Carlo 25th Percentile
10
"-C
o
¦o
N
X
o.i
o.oi
o.ooi
0.0001
04
04
in
J
£
72
of
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ecological Risk Assessment
EXECUTIVE SUMMARY
The thresholds used in HQ calculations represent levels at, or close to, levels demonstrated to
cause adverse responses to organisms in site-specific studies. Thus, HQ exceedances of 1 are
cause for concern.
Benthic Invertebrates—Hazard quotients were calculated by dividing concentrations of COCs
in site sediment by 3 mg/kg, the effects benchmark for benthic invertebrates exposed to tPCBs in
sediment. These results indicate that significant risk was observed in all reaches of the PSA.
Predicted risks were greatest in the upstream (Reach 5A) and Woods Pond (Reach 6) sediment.
Because of small-scale variation in sediment tPCB concentrations, although the majority of
benthic invertebrates in the PSA are at risk (i.e., HQ > 1), some individuals in less-contaminated
areas are not.
Amphibians—For amphibians, HQs were calculated by dividing summary statistics for vernal
pool sediment concentrations by the effects benchmark for amphibians exposed to tPCBs in
sediment (3 mg/kg tPCBs). These results indicate significant risk in all reaches of the PSA, with
HQs above 1. Predicted risks were greatest in the upstream (Reach 5A) vernal pool habitats.
These results indicate that the majority of amphibians are at risk (i.e., HQ > 1), with higher levels
of risk (i.e., HQ > 5) in a large percentage of vernal pools (about 50% of pools in Reaches 5A
and 5B).
Fish—For fish, HQs were calculated separately for the five representative warmwater species by
dividing summary statistics for exposure by the tissue effects benchmark protective of all species
of PSA fish (49 mg/kg tPCB). These results indicate that risk occurs in all reaches of the PSA.
Predicted risks were greatest for predator fish at the top of the food web (e.g., largemouth bass)
and when fish reach their maximum adult tPCB concentrations. The ERA indicates that these
HQs are indicative of sublethal (e.g., reproductive and developmental) responses to offspring;
the pathway for the manifestation of effects is through the maternal transfer of tPCBs to eggs.
Acute mortality to adults is not expected for most fish.
In addition to tPCBs, fish HQs were derived for TEQ using the average of the effects thresholds
relevant to warmwater fish species obtained from site-specific toxicity studies (42 ng/kg TEQ).
The magnitudes and probabilities of risk for TEQ are very similar to tPCB risks.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ecological Risk Assessment
EXECUTIVE SUMMARY
Wildlife Endpoints—For wildlife, the distributions from a Monte Carlo analyses for total daily
intake of COCs by representative species were each divided by the corresponding effects metrics
used to estimate risks. In the case of a dose-response curve effects metric (e.g., mink exposed to
tPCBs), the effects metric was specified as a uniform distribution of dose ranging from 10% to
20% effect. A similar approach was used for NOAEL-LOAEL ranges (e.g., bald eagles exposed
to tPCBs), field-based effects metrics (e.g., tree swallows exposed to tPCBs), and threshold
ranges (e.g., kingfishers exposed to TEQ).
In addition to plots developed for mink and otter exposed to tPCBs and TEQ using the results of
literature-based dose-response curve, plots were also developed using the results of the mink
feeding study conducted in support of this ERA. In this case, the denominator was the NOAEL
to LOAEL range, rather than the 10% and 20% effects doses from the literature-based dose-
response curve.
Unlike traditional HQs, the probabilistic HQs for wildlife do not include safety factors, i.e., are
not conservative. No safety factors were used to estimate the effects metrics (except in the case
of the bald eagle), and uncertainties regarding exposure model inputs were explicitly propagated
through the probability bounds and Monte Carlo analyses. Thus, HQ exceedances of 1 are cause
for concern.
Wildlife risks from tPCBs and TEQ are highest for mink and river otter, with HQs between 100
and 500 for tPCBs, and 5 and 10 for TEQ using the results of the literature-based dose-response
curve. The HQs for tPCBs were somewhat lower when the results of the site-specific mink
feeding study were used to derive the effects threshold. Wildlife risks from tPCBs are generally
higher than risks from TEQ by one to several orders of magnitude.
The risks from tPCBs and TEQ to many of the wildlife species are uncertain to the extent that the
range of HQs spans 1. The highest and lowest values depicted for the wildlife HQs are the
extreme representations of uncertainty because they are tail outputs from the probability bounds
analyses, a technique designed to propagate all forms of uncertainty (e.g., inability to precisely
specify distribution type or parameter values for a distribution). Thus, the range of HQs shown
in the boxes should be interpreted as representing a reasonable range within which the HQ
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ecological Risk Assessment
EXECUTIVE SUMMARY
estimate occurs for the receptor of interest, and the lines should be interpreted as representing the
extremes within which the HQ could occur.
ES.5.2 Risks Downstream of the Primary Study Area
Because of the reduced levels of contaminant concentrations downstream of the PSA and
significant shifts in aquatic habitat types associated both with river gradient and location of
dams, a different approach than that applied in the PSA was followed to assess potential
ecological risks of tPCBs to biota in areas downstream of Woods Pond. The assessment of
potential ecological risks was conducted using mapping (GIS) techniques and threshold
concentrations that indicate potential risk for six taxonomic groups selected based on the
outcome of the evaluations performed in the PSA and the habitat characteristics found
downstream. These groups or species are benthic invertebrates, amphibians, warmwater fish,
trout, mink, otter, and bald eagles.
For each of these groups, a maximum acceptable threshold concentration (MATC) for tPCBs
was developed based primarily on the detailed risk assessment performed for the PSA. Each
MATC was then compared to medium-specific data for areas downstream of Woods Pond to
Long Island Sound. Areas with MATC exceedances, indicating potential risk, were plotted on
maps of the river. The methods used for each of the six representative groups and the results of
the analyses are discussed in the following sections.
Benthic Invertebrates—For benthic invertebrates, the medium of interest was river sediment.
An MATC of 3 mg/kg tPCBs was used as a conservative measure of the potential for adverse
affects to benthic invertebrates downstream of Woods Pond. The MATC of 3 mg/kg tPCBs was
compared to recent surficial sediment data downstream of Woods Pond, and the results were
plotted to indicate samples above and below the MATC. Inverse distance weighting was used to
interpolate sediment concentrations between discrete sampling points, and the potential for risk
to benthic invertebrates was shown by shading the corresponding sections of the river channel
Figure ES-8).
In general, potential risks to benthic invertebrates occur in limited areas downstream of Woods
Pond to Rising Pond. These areas are depositional and tend to have higher concentrations of
MK01\O:\20123001.096\ERA_PB\ERA_PB_ES.DOC gg ^3 7/11/2003
-------
NEW
YORK
MASSACHUSETTS
NOTES:
Risk to benthic invertebrates is based on a maximum acceptable threshold
concentration (MATC) of 3.0 mg/kg total PCB (tPCB) (dry weight) in surficial
sediments (0-6 inches).
LEGEND:
~ Town
Reach Break
BENTHIC INVERTEBRATE RISK
Roads
O <3 mg/kg
• >=3 mg/kg
I Housatonic River Basin Hydrology
State Boundary
w^prE
s
0 2 4 6 8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-8
ASSESSMENT OF RISK TO
BENTHIC INVERTEBRATES
EXPOSED TO tPCBs
DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - benthos d4-71 o:\gepitt\epsfiles\plots\in\benthos_risk_es8.eps 111:33 AM, 7/8/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ecological Risk Assessment
EXECUTIVE SUMMARY
tPCBs. Below Rising Pond, sediment does not contain concentrations of tPCBs that represent a
potential risk to benthic invertebrates.
Amphibians—A floodplain soil MATC of 3.0 mg/kg tPCB (dry weight) was derived from the
amphibian risk assessment conducted for the PSA. Inverse distance weighting was used to
interpolate tPCB concentrations to the limit of the 100-year floodplain (10-year floodplain
contours are not available downstream of Woods Pond) using the 0 to 6 inch (0 to 15 cm) depth
data from the floodplain downstream of Woods Pond.
Several large areas of the floodplain may pose risk to amphibians between Woods Pond and
Rising Pond, with only small isolated areas of potential risk downstream of Rising Pond (Figure
ES-9). The floodplain risk mapping for amphibians was not conducted downstream of the
Massachusetts/Connecticut state line because the extent of the floodplain is more limited in
Connecticut and concentrations in Reach 9 floodplain soil were below the MATC.
Potential risks to amphibians exposed to sediments downstream from Woods Pond appear to be
limited to small areas between Woods Pond and Great Barrington (Figure ES-10).
Warmwater Fish—As was done for the PSA, risk to fish was evaluated based on concentrations
of tPCBs in fish tissue. An MATC of 49 mg/kg tPCB in tissue (whole body, wet weight)
developed for the PSA based on site-specific effects to warmwater fish was applied to areas
downstream of Woods Pond using the available (e.g., bass, perch, sunfish) tissue data for
warmwater species. Each downstream reach (Reaches 7 through 16) was evaluated as a unit, and
the mean adult fish tissue concentration in each reach was compared with the MATC to
determine potential risk (Figure ES-11). No risks were indicated in any of the reaches below the
PSA.
Trout—Trout were evaluated separately from warmwater fish species because of significant
differences in habitat requirements and differences in the sensitivity of some trout species to
tPCBs documented in the literature. Trout also tend to have higher tPCB concentrations because
of their higher lipid content. However, the site-specific studies did not indicate large differences
between the effects threshold for rainbow trout and warmwater species, but the strain of rainbow
trout used in the site-specific toxicity tests is less sensitive than other strains used widely in
MK01\O:\20123001.096\ERA_PB\ERA_PB_ES.DOC gg ^ ^
-------
NOTES:
Risk to amphibians is based on a maximum acceptable threshold concentration
(MATC) of 3.0 mg/kg total PCB (tPCB) concentrations (dry weight) in surficial
floodplain soils (0-6 inches). All available sample data were used to calculate
interpolated PCB concentrations in unsampled areas within the 100-year
floodplain, using an inverse distance weighting (IDW) approach.
LEGEND:
~
Town
AMPHIBIAN
RISK
Reach Breaks
O
•
< 3 mg/kg
>= 3 mg/kg
< 3 mg/kg
Roads
Housatonic River Basin Hydrology
>= 3 mg/kg
~
State Boundary
S
0.5 0 0.5 1 1.5 2 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-9
ASSESSMENT OF RISK
TO AMPHIBIANS IN
FLOODPLAIN EXPOSED TO
tPCBs DOWNSTREAM OF
WOODS POND IN
MASSACHUSETTS
| O:\gepitt\aprs\era_species.apr | layout - amphibs FLD e4-41 o:\gepitttepsfiles\plots\in\amphib_fld_risk_es9.eps 111:34 AM, 7/8/20031
-------
MASSACHUSETTS
NEW
YORK
NOTES:
Risk to amphibians is based on a maximum acceptable
threshold concentration (MATC) of 3.0 mg/kg total PCB (tPCB)
concentrations (dry weight) in surficial sediment samples (0-6 inches).
LEGEND:
~
Town
AMPHIBIAN
RISK
Reach Breaks
O
< 3 mg/kg
•
>= 3 mg/kg
/X/
Roads
< 3 mg/kg
LJ
Housatonic River Basin Hydrology
>= 3 mg/kg
~
State Boundary
w^prE
s
0 2 4
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-10
ASSESSMENT OF RISK
TO AMPHIBIANS IN SEDIMENT
EXPOSED TO tPCBs
DOWNSTREAM OF WOODS
POND IN CONNECTICUT
| O:\gepitt\aprs\era_species.apr | layout - amphibs SEDS e4-51 o:\gepitttepsfiles\plots\in\amphibs_sed_risk_es10.eps 110:41 AM, 7/10/20031
-------
NEW
YORK
MASSACHUSETTS
CONNECTICUT
NOTES:
Risk to warmwater fish is based on a maximum acceptable threshold
concentration (MATC) of 49 mg/kg total PCB (tPCB) concentrations
(wet weight) in whole body tissue.
* Only fish collected in 1998 to the present (2002) were included.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
* Brown bullhead filet samples were scaled by 1.5.
* Warm water filet samples were scaled by 2.3.
LEGEND:
~
Town
WARM WATER FISH RISK
A/
Reach Break
< 49 mg/kg
Roads
>= 49 mg/kg
r~^i
Housatonic River Basin Hydrology
~
State Boundary
w^prE
s
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-11
ASSESSMENT OF RISK
TO WARM WATER FISH
EXPOSED TO tPCBs
DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - fish warm f4-101 o:\gepitt\epsfiles\plots\in\warm_fish_risk_es11.eps 111:36 AM, 7/8/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ecological Risk Assessment
EXECUTIVE SUMMARY
toxicity testing. Furthermore, there are other trout species found downstream of the PSA (e.g.,
brown trout) for which sensitivity has not been assessed. Given that some trout species have
been documented to have greater sensitivity of PCBs and dioxins, relative to the warmwater
species considered in the development of the 49 mg/kg tPCB warmwater MATC, a factor of 4
was applied in recognition of these potential interspecies differences. Therefore, a tissue MATC
of 12 mg/kg tPCBs (whole body, wet weight) was derived for trout.
The results of this evaluation indicate that trout are potentially at risk in Reaches 7 and 9, but not
in reaches with suitable habitat further downstream (Figure ES-12). This assessment has high
uncertainty due to the number extrapolations required and the low magnitude of exceedance of
the MATC value. Potential risk to trout was not evaluated in Reach 8, Reach 10, and
downstream of Reach 12 due to lack of suitable trout habitat.
Mink—An MATC for mink of 2.65 mg/kg tPCBs in fish (whole body, wet weight) represents
the geometric mean of the NOAEL and LOAEL developed in a site-specific study of the toxicity
of a diet containing Housatonic River fish to mink. Mean fish concentrations were calculated for
each river reach downstream of Woods Pond using available whole body fish tissue data from
samples of fish with an overall body length between 7 and 20 cm, corresponding to the size
commonly preyed on by mink. Potential risk to mink due to consumption of contaminated fish
occurs from the Woods Pond Dam downstream to the Great Falls Dam, corresponding to
Reaches 7 through 10 (Figure ES-13).
River Otter—The mink MATC of 2.65 mg/kg tPCB in fish (whole body, wet weight) was also
used for river otter. Mean fish concentrations were calculated for such areas in river reaches
downstream of Woods Pond using available whole body fish tissue data from fish with an overall
body length between 5 and 50 cm, corresponding to the size commonly preyed on by otter.
Potential risk to otter due to consumption of contaminated fish occurs from the Woods Pond
Dam downstream to the Bulls Bridge Dam, corresponding to Reaches 7 through 12 (Figure
ES-14).
Bald Eagle—An MATC of 30.4 mg/kg tPCBs (whole body fish tissue, wet weight) was
developed for wintering bald eagles, assuming that the eagle diet was composed of 78% fish, and
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg 39
-------
Vest St<
BncT
£enox Dale
(41)
Glend
Btoekbridge
Sou
Berksare I
bremont Plain
ghts^
(23)
Hartsvilt
MASSACHUSETTS
Drough
\41\
NEW
YORK
Falls\VittaaS»
%r
CONNECTICUT
VesrOornwa^
202
(55)
ICandlewood
LakeJ
ki.
my
NOTES:
Risk to coldwater fish is based on a maximum acceptable threshold
concentration (MATC) of 12 mg/kg total PCB (tPCB) concentrations
(wet weight) in trout tissue (whole body).
* Only fish collected in 1998 to the present (2002) were included.
* Fish fillet samples were scaled by a factor of 2.3 to convert to whole
body.
* Where trout data were unavailable, averages by reach for warmwater
species were calculated and scaled by 2 for trout. In some reaches,
only warmwater fillets were available for conversions. The
fillets were first scaled up by a factor of 2.3, then 2 for coldwater fish.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
* Brown bullhead filet samples were scaled by 1.5.
* Warmwater filet samples were scaled by 2.3.
* Trout were not evaluated downstream of Bulls Bridge Dam based on
insufficient trout data and no suitable trout habitat in downstream reaches.
f V
LEGEND:
~
Town
COLD WATER FISH RISK
A/
Reach Break
<12 mg/kg
Roads
>=12 mg/kg
r~^i
Housatonlc River Basin Hydrology
~
State Boundary
w^prE
s
2 4 6
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-12
ASSESSMENT OF RISK
TO TROUT EXPOSED TO
tPCBs DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - fish cold f4-111 o:\gepitttepsfiles\plots\in\trouLrisk_es12.eps 111:37 AM, 7/8/20031
-------
Vest St<
BncT
£enox Dale
(41)
Glend
Btoekbridge
Sou
Berksare I
bremont Plain
(23)
Hartsvilt
MASSACHUSETTS
Drough
\41\
NEW
YORK
Saljsbun
Falls Village*
%r
CONNECTICUT
vestuomwaJI
202)
(55)
ICandlewood
Lake/
KMJ
f V
NOTES:
Risk to mink is based on a maximum acceptable threshold concentration
(MATC) of 2.65 mg/kg total PCB (tPCB) concentration (wet weight) in whole
body fish tissue 7-20 cm in length.
* Only fish collected in 1998 to the present (2000) were included.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
* Trout filet samples were scaled by 1.47.
* Brown bullhead filet samples were scaled by 1.5.
* Warmwater filet samples were scaled by 2.3.
LEGEND:
~
Town
MINK RISK
A/
Reach Break
< 2.65 mg/kg
Roads
>= 2.65 mg/kg
r~^i
Housatonlc River Basin Hydrology
~
State Boundary
w^prE
s
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-13
ASSESSMENT OF RISK
TO MINK EXPOSED TO
tPCBs DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - mink i4-151 o:\gepitttepsfiles\plots\in\mink_risk_es13.eps 111:38 AM, 7/8/20031
-------
Vest St<
BncT
£enox Dale
(41)
Glend
Btoekbridge
Sou
Berksare I
bremont Plain
(23)
Hartsvilt
MASSACHUSETTS
Drough
\41\
NEW
YORK
Saljsbun
Falls Village*
%r
CONNECTICUT
vestuomwaJI
202)
(55)
ICandlewood
Lake/
KMJ
f V
NOTES:
Risk to otter is based on a maximum acceptable threshold concentration
(MATC) of 2.65 mg/kg total PCB (tPCB) concentration (wet weight) in
whole body fish tissue 5-50 cm in length.
* Only fish collected in 1998 to the present (2000) were included.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
* Trout filet samples were scaled by 1.47.
* Brown bullhead filet samples were scaled by 1.5.
* Warmwater filet samples were scaled by 2.3.
LEGEND:
~
Town
OTTER RISK
A/
Reach Break
< 2.65 mg/kg
Roads
>= 2.65 mg/kg
r~^i
Housatonlc River Basin Hydrology
~
State Boundary
w^prE
s
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-14
ASSESSMENT OF RISK
TO OTTER EXPOSED TO
tPCBs DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - otter i4-16 | o:\gepitt\epsfiles\plots\in\otter_risk_es14.eps 111:39 AM, 7/8/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Ecological Risk Assessment
EXECUTIVE SUMMARY
that the remainder of the diet included non-aquatic species that, for the purpose of this analysis,
were not contaminated. Potential risk to nesting bald eagles was evaluated using methods
similar to those discussed above for mink (Figure ES-15).
A more in-depth analysis was performed for Reaches 14 and 15 where bald eagles have nested.
Bald eagles on average consume a summer diet consisting of 78.2% fish, 16.3% birds, and 5%
mammals.
The results of the evaluation indicate that potential risks to bald eagles from consuming
contaminated fish in areas downstream of Woods Pond are restricted to Reach 8, corresponding
to Rising Pond. However, Rising Pond is smaller that the typical eagle foraging area, so this
estimate of risk is conservative. In addition, the more in-depth analysis specific to Reaches 14
and 15 also did not show risk in the foraging area of the nesting bald eagles.
ES.6 BROADER IMPLICATIONS
The weight-of-evidence assessments indicate that COCs in the PSA of the Housatonic River,
particularly tPCBs, are causing risks to many of the species chosen to represent the assessment
endpoints. Risks from COCs, however, may potentially extend beyond adverse effects to
survival, growth, and reproduction of representative species. The Housatonic River ERA also
explores the broader implications of the risks of COCs to representative species, including
extension of the ecological risk assessment to species that occur in the Housatonic River
watershed, but that had not been considered explicitly in the quantitative ecological risk
assessments, and additional ecological implications.
ES.6.1 Implications for Other Species in the Primary Study Area
The major factors that influence exposure to tPCBs and TEQ and that were considered in the
analysis include:
¦ Dietary composition.
¦ Foraging behavior and home range.
¦ Size, metabolism, and life history characteristics.
¦ Sensitivity to COCs.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg ^
-------
MASSACHUSETTS
NEW
YORK
CONNECTICUT
NOTES:
Risk to eagles Is based on a maximum acceptable threshold concentration
(MATC) of 30.4 mg/kg total PCB (tPCB) concentration (wet weight) in whole
body fish tissue greater than or equal to 12 cm.
* Only fish collected in 1998 to the present (2000) were included.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
LEGEND:
~
Town
EAGLE RISK
/V
Reach Break
< 30.4 mg/kg
Roads
>= 30.4 mg/kg
r~^i
Housatonic River Basin Hydrology
~
State Boundary
w^prE
s
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE ES-15
ASSESSMENT OF RISK
TO BALD EAGLE EXPOSED
TO tPCBs DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - eagle k4-51 o:\gepitt\epsfiles\plots\in\eagle_risk_es15.eps 111:40 AM, 7/8/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Ecological Risk Assessment
EXECUTIVE SUMMARY
The ERA compares these factors between the representative species and other species in their
foraging groups. The comparison highlights similarities and differences, and their potential to
influence exposure and hence risks to tPCBs and TEQ.
Benthic Invertebrates—The benthic invertebrate ERA included the entire benthic community;
benthic community composition analysis was a measurement endpoint considered in the weight-
of-evidence assessment. Both the status of sensitive taxa and community composition are
considered indicators of overall health and productivity of the benthic community. Thus, there is
no need to extrapolate the findings of the benthic invertebrate assessment described previously to
other benthic invertebrate species in the PSA.
Amphibians—Certain amphibian species that were not studied may be more susceptible to the
effects of tPCBs because of their life history characteristics. For example, blue-spotted
(Ambystoma laterale) and spotted salamanders (Ambystoma maculatum) have a lifestage as
aquatic carnivorous, bottom-dwelling larvae. Thus, they could potentially bioaccumulate PCBs
more quickly than herbivorous amphibians. Salamanders appeared in lower numbers in vernal
pools with high sediment tPCB concentrations. Several salamander species occur in
contaminated habitat in the PSA, including the spotted salamander, the Jefferson salamander
(,Ambystoma jeffersonianum, formerly considered a variety of blue-spotted salamander), and the
four-toed salamander (Hemidactylium scutatum), the latter two of which are Species of Special
Concern.
Fish—There is evidence in the literature that salmonid species may have a higher sensitivity to
the effects of PCBs and other dioxin-like COPCs. The use of rainbow trout in the site-specific
toxicity testing program, combined with effects data from the literature, provides a high degree
of confidence that the ERA included an evaluation of fish species with equal or greater
sensitivities than the representative species listed above. However, the procedure used to
establish MATCs for fish in the PSA placed a low weight on studies conducted with fish species
known to be highly sensitive (e.g., lake trout), to avoid an overly conservative assessment. Risks
to coldwater fisheries (e.g., trout) downstream of the PSA were explicitly evaluated using
benchmarks developed for salmonids; the uncertainty in these downstream risk estimates is high
due to the number of extrapolations required. The risk of COCs to the occasional salmonid
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg ^
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1 occurring within the PSA is considered to be moderate. The PSA, however, is considered to be a
2 warmwater fishery, and thus salmonid abundance is expected to be low in this portion of the
3 river, even in the absence of chemical stressors.
4 Insectivorous Birds—The weight-of-evidence assessment indicated that exposure of
5 insectivorous birds, such as tree swallows and American robins, to tPCBs and TEQ is high but
6 unlikely to lead to adverse reproductive effects. Confidence in this conclusion, however, is not
7 high because the two available lines of evidence for both species did not produce concordant
8 results. There are a number of insectivorous birds with similar feeding habits as tree swallows in
9 the PSA, and these are generally believed to be at the same low to moderate risk as tree
10 swallows, although some species with higher food intake could be at higher risk.
11 Compared to American robins, eastern bluebirds and eastern towhees are expected to experience
12 lower to similar levels of risk from exposure to tPCBs and TEQ. The level of risk for the hermit
13 thrush, northern mockingbird, veery, and wood thrush is expected to be similar to American
14 robins. With the exception of earthworms in the robin diet, the dietary preferences of these birds
15 are similar to the American robin. The absence of earthworms, a major dietary source of
16 contaminants, will decrease their exposure to tPCBs and TEQ. However, their smaller body
17 sizes result in higher food intake rates and hence greater exposure to tPCBs and TEQ through
18 diet compared to American robins.
19 Piscivorous Birds—The weight-of-evidence assessment indicates that risks of tPCBs and TEQ
20 to belted kingfisher are low; however, risks of these COCs to osprey are high and could lead to
21 adverse reproductive effects.
22 The belted kingfisher and osprey were chosen to represent piscivorous birds inhabiting the
23 Housatonic River area. Belted kingfisher and osprey are common piscivorous birds in the PSA.
24 Great blue herons are also found in the PSA, and are discussed in Appendix K with other
25 piscivorous birds (e.g., American bittern).
26 Piscivorous Mammals—Mink and river otter, the representative species for piscivorous
27 mammals, are the only piscivorous mammals commonly found in the watershed of the
28 Housatonic River.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Ecological Risk Assessment
EXECUTIVE SUMMARY
Omnivorous and Carnivorous Mammals—The weight-of-evidence assessment indicates that
omnivorous and carnivorous mammals, such as red fox and short-tailed shrew, are at risk in the
PSA as a result of exposure to tPCBs and TEQ. Masked shrews are expected to experience a
level of risk similar to northern short-tailed shrews and smoky shrews are expected to be at
higher risk, based on their metabolic rates relative to short-tailed shrews. All three have similar
foraging behaviors and ranges.
Coyotes have a larger body size and foraging range that decreases their exposure to tPCBs and
TEQ. Considering these characteristics, coyotes are expected to experience lower risks from
exposure to tPCBs and TEQ than red fox. Gray and red foxes are expected to experience similar
risks from exposure to tPCBs and TEQ. Gray fox have a larger foraging range than red fox and
that may decrease their exposure to tPCBs and TEQ. Gray fox, however, have a greater reliance
on animal matter and therefore greater exposure to tPCBs and TEQ.
Fishers, long-tailed weasels, and short-tailed weasels are expected to experience similar to higher
levels of risk from exposure to tPCBs and TEQ compared to the red fox due to greater
consumption of animal matter and/or higher metabolic rate.
Threatened and Endangered Species—The bald eagle, American bittern, and small-footed
myotis were chosen to represent T&E species that are likely to be highly exposed to COCs in the
Housatonic River PSA. Other T&E species that occur in the area include one mussel (triangle
floater); three dragonflies (riffle snaketail, zebra clubtail, and arrow clubtail); a turtle (wood
turtle); three salamanders (Jefferson salamander, four-toed salamander, and northern spring
salamander); three hawks (northern harrier, sharp-shinned hawk, and Cooper's hawk); two
warblers (northern parula and blackpoll warbler); a wading bird (common moorhen); and a shrew
(northern water shrew). Some of these species were qualitatively assessed in other appendices
and compared to other, more appropriate, assessment endpoints (e.g., amphibians for
salamanders).
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Ecological Risk Assessment
EXECUTIVE SUMMARY
The level of risk for soras1 is expected to be lower than for American bitterns because of greater
consumption of vegetable matter. Great blue herons and king rails are expected to experience a
similar level of risk as American bitterns because of a combination of size differences and some
differences in dietary preferences. The least bittern, green heron, Virginia rail, and pied-billed
grebe are expected to experience higher levels of risk compared to the American bittern. The
foraging and life history characteristics of these birds are similar to the American bittern.
However, these birds are much smaller than the American bittern. Their smaller body sizes
result in a higher metabolism and greater exposure to tPCBs and TEQ.
The Indiana bat, northern myotis, and little brown bat are expected to have similar levels of risk
as the small-footed myotis. These species belong to the same genus (Myotis) and have similar
foraging behaviors and life histories.
ES.7 SOURCES OF UNCERTAINTY
The assessment of risks of COCs to aquatic and wildlife species in the Housatonic River contains
uncertainties. Each source of uncertainty can influence the estimates of risk; therefore, it is
important to describe and, when possible, specify the magnitude and direction of such
uncertainties. In this section, the most significant sources of uncertainty commonly encountered
throughout the ERA are described. The sources of uncertainty are grouped by phase of the ERA
(i.e., problem formulation, exposure assessment, effects assessment, risk assessment).
The problem formulation is intended to define the linkages between stressors, potential exposure,
and predicted effects on ecological receptors. As such, the conceptual model provides the
scientific basis for selecting assessment and measurement endpoints to support the risk
assessment process. Potential uncertainties arise from lack of knowledge regarding ecosystem
functions, failure to adequately address spatial and temporal variability in the evaluations of
sources, fate and effects, omission of stressors, and overlooking secondary effects (EPA 1998).
1 Several of the species included in this section (i.e., sora, great blue heron, green heron, Virginia rail, northern
myotis, little brown bat) are not threatened and endangered species either federally or in Massachusetts and
Connecticut (Appendix A). They are included in the discussion of T&E species because they are taxonomically
and ecologically similar to either American bittern or to small-footed myotis.
MK01\O:\20123001.096\ERA_PB\ERA_PB_ES.DOC t-.c AO 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Ecological Risk Assessment
EXECUTIVE SUMMARY
The types of uncertainties associated with the conceptual model that links contaminant sources to
effects include those associated with the identification of COCs, environmental fate and transport
of COCs, exposure pathways, receptors at risk, and ecological effects. Of these, the
identification of exposure pathways probably represents the primary source of uncertainty in the
conceptual model. The detailed ecological characterization performed at this site has greatly
reduced the uncertainties associated with problem formulation, yet some remain and are
described below.
The exposure assessment is intended to describe the actual or potential co-occurrence of stressors
with receptors. As such, the exposure assessment identifies the exposure pathways and the
intensity and extent of contact with stressors for each receptor or group of receptors at risk. The
exposure models for wildlife were energetics-based models requiring information on body
weight, free living metabolic rate, proportions of food items in the diets, and the concentrations
of COCs in these food items. Each of these variables has associated uncertainties, most of which
were propagated through the exposure models. The effects assessment is intended to describe the
effects caused by stressors, link them to the assessment endpoints, and evaluate how effects
change with fluctuations in the levels (i.e., concentrations or doses) of the various stressors. In
this assessment, the effects of tPCBs and other COCs to representative species were assessed.
There are several sources of uncertainty in the assessment of effects, including extrapolation
errors and a limited number of toxicity studies conducted with the representative species.
For benthos and amphibians, the effects benchmarks derived from the literature had a high
degree of uncertainty because of the need to extrapolate across sites and species. The site-
specific fish toxicity studies indicated variations in the concentration-response relationships
observed across species, reaches, and treatments, and introduced uncertainty into the
development of effects thresholds. The methodology used in site-specific fish studies was
developed recently, and there are potential uncertainties inherent to extrapolating these
laboratory-based results to Housatonic River fish. Similarly, the extrapolation of concentrations
of tPCBs in egg to whole body concentrations has a degree of associated uncertainty.
The greatest potential source of uncertainty for the fish and wildlife effects assessments,
however, was associated with the lack of toxicity studies involving the representative species.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg ^g
-------
Ecological Risk Assessment
EXECUTIVE SUMMARY
1 A weight-of-evidence procedure was used to assess risks of tPCBs and TEQ to the assessment
2 endpoints in the Housatonic River PSA. The analysis follows the methodology proposed by the
3 Massachusetts Weight-of-Evidence Workgroup (Menzie et al. 1996; see Section 2.9 for details).
4 In general, the weight-of-evidence approach is an inclusive process whereby multiple lines of
5 evidence are considered prior to determining risk. For the wildlife risk assessments, these lines
6 of evidence included the exposure and effects modeling results and, in some cases, field survey
7 results, and/or in situ or whole media toxicity test results. For the fish and benthic invertebrate
8 risk assessments, available lines of evidence included field survey results (e.g., community
9 evaluation for benthos), site-specific toxicity tests, and comparison of tissue and sediment
10 concentrations to benchmarks (both from the literature and site-specific benchmarks). The
11 largest source of uncertainty in the weight-of-evidence process was the development of
12 conclusions based upon only one or two lines of evidence.
13 ES.8 SUMMARY AND CONCLUSIONS
14 Weight-of-evidence assessments indicated that aquatic life and wildlife in the Primary Study
15 Area of the Housatonic River are experiencing unacceptable risks as a result of exposure to
16 tPCBs and other COCs. Confidence in this conclusion is high for benthic invertebrates,
17 amphibians, and piscivorous mammals because multiple lines of evidence gave concordant
18 results.
19 The risks of tPCBs and other COCs likely extend beyond the representative species considered in
20 the quantitative risk assessments described herein. Qualitative risk assessments indicated that
21 many other species in the PSA are potentially at risk. Further, there are likely indirect effects
22 (e.g., changes in predator-prey relationships, changes in metapopulation dynamics) occurring
23 inside and outside the PSA as a result of the direct impacts caused by tPCBs and other COCs.
24 An assessment of risk downstream of the PSA indicated that tPCBs could potentially be causing
25 adverse effects to benthic organisms in depositional areas as far as Reach 8, amphibians in
26 floodplain areas as far as Reach 8, trout in Reaches 7 and 9, mink as far as Reach 10, and river
27 otter as far as Reach 12.
MK01\OA20123001.096\ERA_PB\ERA_PB_ES.DOC gg
-------
l 1. INTRODUCTION
2
3
4
5
6
7
8
9
10
11
12 1.1 OVERVIEW
13 The purpose of this ecological risk assessment (ERA) is to characterize and quantify the current
14 and potential risks to biota exposed to contaminants of potential concern (COPCs) in the
15 Housatonic River below the confluence of the East and West Branches, focusing on
16 poly chlorinated biphenyls (PCBs) and other COPCs originating from the General Electric
17 Company (GE) facility in Pittsfield, MA. This ERA considers the fate and transport of PCBs
18 and other COPCs to ecological receptors in the river and associated floodplain, identifies
19 assessment endpoints that are representative of species potentially at risk, and identifies the
20 potential routes of exposure and toxicological effects of the COPCs for these receptors.
21 This information is synthesized, through a weight-of-evidence approach, into a discussion of the
22 nature and magnitude of the risks for the assessment endpoints, and the uncertainties associated
23 with the characterization of these risks.
24 Multiple lines of evidence for each assessment endpoint are evaluated, including where
25 applicable or available:
26 ¦ Field surveys/studies.
27 ¦ Toxicity tests.
28 ¦ Comparison of effects in the literature to a site-specific exposure model.
29
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1"1 7/11/03
Purposes of ERA
This ecological risk assessment (ERA) characterizes and quantifies the current and
potential risks to biota exposed to contaminants in the Housatonic River below the
confluence of the East and West Branches. Specific purposes of the ERA include:
1. To consider the fate and transport of PCBs and other contaminants of potential
concern (COPCs) to ecological receptors in the river and associated floodplain.
2. To identify the potential routes of exposure and toxicological effects of the
COPCs for these receptors.
3. To identify assessment endpoints representative of species potentially at risk.
4. To determine the risk to the assessment endpoints selected.
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
The Housatonic River flows from east of Pittsfield, MA, to Long Island Sound and drains an area
of approximately 1,950 square miles (500,000 hectares) in Massachusetts, New York, and
Connecticut (Figure 1.1-1). The Housatonic River, its sediment, and associated floodplain have
been contaminated with polychlorinated biphenyls (PCBs) and other hazardous substances
released from the General Electric Company (GE) facility located in Pittsfield, MA. The entire
site, known as the General Electric/Housatonic River Site, consists of the 254-acre (103-hectare)
GE manufacturing facility; the Housatonic River and associated riverbanks and floodplain from
Pittsfield, MA, to Long Island Sound; former river oxbows that have been filled; neighboring
commercial properties; Allendale School; Silver Lake; and other properties or areas that have
become contaminated as a result of GE's facility operations.
Because of its size and complexity, the GE/Housatonic River Site has been divided into several
areas for investigation and cleanup. The "Rest of River" is the portion of the river from the
confluence of the East and West Branches of the Housatonic River (the confluence) to the
Massachusetts border with Connecticut, a distance of approximately 54 miles (87 km), and
beyond into Connecticut to Long Island Sound. The total distance from the confluence to Long
Island Sound is approximately 139 miles (223 km). In addition to the river itself, the Rest of
River includes the associated riverbank and floodplain. The Rest of River is further defined in
the Consent Decree entered with the U.S. District Court, Western Region, Massachusetts, in
October 2000. The Rest of River includes areas of the Housatonic River and its sediment and
floodplain (except Actual/Potential Lawns), in which contaminants originating from the GE
facility are located. The lateral extent of the area under investigation includes the floodplain
extending to the 1-ppm total PCB (tPCB) isopleth, which is approximately equivalent to the 10-
year floodplain.
The ERA focuses on the portion of the river from the confluence of the East and West Branches
2 miles (3 km) below the GE facility, to Woods Pond Dam, a distance of approximately 11 river
miles (18 km). This area is referred to as the Primary Study Area (PSA) in the Supplemental
Investigation (Figure 1.1-2), and is where much of the PCB contamination was found in previous
studies. The river (which includes free-flowing and impounded sections) and the floodplain
downstream of Woods Pond to the Derby-Shelton Dam in Connecticut are also considered in the
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-2
-------
Long Island Sound
LEGEND:
~
Housatonic River
Housatonic River Basin
A/ Primary Road
N
V
s
4
0 4
8
Miles
5
0 5 10
15
Kilometers
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 1.1-1
HOUSATONIC RIVER
| O:\gepitt\aprs\era_figures2.apr | layout - fig 1.1-11 o:\gepitttepsfiles\plots\in\fig1-1-1.eps 111:25 AM, 7/10/20031
-------
LEGEND:
Roads
Housatonic River
GE Facility
N
V
s
0.1
0 0.1
0.2 Miles
0.2
0 0.2
Kilometers
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 1.1-2
PRIMARY STUDY AREA (PSA)
| O:\gepitt\aprs\era_figures2.apr I layout - fig1.1-21 o:\gepitttepsfiles\plots\in\fig1-1-2.eps 19:50 AM, 3/7/20031
-------
1 ERA. Beyond this dam, the river is subject to tidal influence, as well as COPCs (including
2 PCBs) from other hazardous waste sites.
3 This ERA is structured as an integrated report summarizing information included in the
4 supporting appendices and providing background common to all the assessment endpoints. The
5 potential risk for each assessment endpoint is discussed in detail in Appendices D through K and
6 is summarized in Sections 3 through 11. Other appendices provide additional information such
7 as a comprehensive Ecological Characterization, the identification of COPCs and the extent of
8 contamination for further consideration in the ERA, and other supporting information. Figure
9 1.1-3 provides a roadmap for the ERA and supporting appendices.
10 1.2 SITE HISTORY
11 The Housatonic River is located in a predominantly rural area of western Massachusetts and
12 Connecticut, where farming was the main occupation from colonial settlement through the late
13 1800s. As with most rivers, the onset of the industrial revolution in the late 1800s brought
14 manufacturing to the banks of the Housatonic River in Pittsfield, MA. GE began operations in
15 its present location in 1903. Three manufacturing divisions have operated at the GE facility
16 (Transformer, Ordnance, and Plastics).
17 The 254-acre (103-ha) GE facility in Pittsfield (Figure 1.2-1) has historically been the major
18 handler of PCBs in western Massachusetts, and is the only known source of PCBs found in the
19 Housatonic River sediment and floodplain soil in Massachusetts. Although GE performed many
20 functions at the Pittsfield facility throughout the years, the activities of the Transformer Division,
21 including the construction and repair of electrical transformers using dielectric fluids, some of
22 which contained PCBs (primarily Aroclor 1260 and, to a lesser extent, 1254), were one likely
23 significant source of PCB contamination. According to GE reports, from 1932 through 1977,
24 releases of PCBs reached the wastewater and stormwater systems associated with the facility and
25 were subsequently conveyed to the East Branch of the Housatonic River and to Silver Lake, a
26 25-acre (10-ha) lake adjacent to the GE facility.
27 During the 1940s, efforts to straighten the Pittsfield reach of the Housatonic River by the City of
28 Pittsfield and the U.S. Army Corps of Engineers (USACE) resulted in 11 former oxbows being
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1-5 7/11/03
-------
Section 3
Benthic
Invertebrates
(Appendix D)
Section 4
Amphibians
(Appendix E)
Section 5
Fish
(Appendix F)
Section 7
Insectivorous
Birds
(Appendix G)
Section 8
Piscivorous
Birds
(Appendix H)
Section 9
Piscivorous
Mammals
(Appendix I)
Section 10
Omnivorous/
Carnivorous
Mammals
(Appendix J)
Section 11
Threatened
& Endangered
Species
(Appendix K)
Section 12
Risk Summary
Relevant to all endpoints + Relevant to specific assessment
endpoints
j j Wildlife assessment approach
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
Figure 1.1-3
MK01|O:\20123001.096\ERA_PB\ERA_PB_1_Fig. 1.1-3.ppt
Ecological Risk Assessment Roadmap
-------
Unkamet Brook Area_
Building 71
Consolidation Area
20s Complex
40s Complex
30s Complex
LEGEND:
GE Plant Area:
Newell Street Area II
Silver Lake:
20s complex
30s complex
40s complex
East Street Area 1- North
East Street Area 1-South
(Groundwater only)
East Street Area 2 - North
East Street Area 2- South
Building 71 Consolidation Area
Hill 78 Consolidation Area
Hill 78 Area
Unkamet Brook Area
Silver Lake
Silver Lake Banks
Former Oxbow Areas:
Former Oxbow Areas A&C
Former Oxbow Areas J&K
Lyman Street Area
Newell Street Area I
Newell Street Area II
Allendale School Area
Former Oxbows
Notes:
1. Base features provived by General Electric Contractors.
2. Not all physical features are shown.
3. Site Boundaries are approximate.
4. Map produced by Roy F. Weston, Inc.
500
Scale in Feet
500 1000
1500 2000
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 1.2-1
GE PLANT AREA: REMOVAL ACTION AREAS
| O:\gepitt\aprs\raas_locs.apr | layout-fig 1.2-11 o:\gepitt\epsfiles\plots\in\figi-2-1.eps 110:31 AM, 7/10/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
isolated from the river channel. The oxbows were filled with material, much of which was later
discovered to contain PCBs and other hazardous substances.
The State of Connecticut posted a fish consumption advisory for most of the Connecticut section
of the river in 1977 as a result of the PCB contamination in the river sediment and fish tissue. In
1982, the Massachusetts Department of Public Health (MADPH) issued a consumption advisory
for fish, frogs, and turtles for the Housatonic River. In addition, in 1999, MADPH issued a
waterfowl consumption advisory from Pittsfield to Great Barrington due to PCB concentrations
in wood ducks and mallards collected from the river by the U.S. Environmental Protection
Agency (EPA).
Although the first 2 miles (3 km) downstream from the facility have been channelized, the
remainder of the river's course is relatively unaffected (with the exception of the numerous dams
downstream) in areas south of Pittsfield. The river, from the confluence of the East and West
Branches of the Housatonic to Woods Pond Dam in Lenox, is approximately 11 miles (18 km)
long. The channel ranges from 40 to 125 ft (12 to 38 m) in width, is bordered by an extensive
floodplain (up to 3,000 feet [900 m] wide), and has a meandering pattern with numerous oxbows
and backwaters. Woods Pond, the first impoundment below the GE facility, is a shallow 54-acre
(22-ha) impoundment that was formed by the construction of a dam in 1864 (Harza, 2001 as
cited in BBL and QEA, 2003).
The land uses of the floodplain properties in Massachusetts include residential,
commercial/industrial, agricultural, recreational (such as canoeing, fishing, and hunting), wildlife
management, and parks and a golf course. The Housatonic River floodplain is an attractive area
for recreation, including fishing and waterfowl hunting.
Numerous studies conducted since 1988 have documented PCB contamination of soil within the
floodplain of the Housatonic River downstream of the GE facility. PCBs have been detected in
river sediment in Massachusetts as far downstream as the border with Connecticut (BBL 1995),
and in Connecticut as far as the Derby-Shelton Dam and beyond into Long Island Sound (other
sources have been identified downstream of this dam). The PCBs detected in Housatonic River
floodplain soil and sediment consist of predominantly Aroclor 1260, with a minor contribution of
Aroclor 1254.
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1"8 7/11/03
-------
1 Numerous residential properties have been the focus of efforts by the Massachusetts Department
2 of Environmental Protection (MDEP) to coordinate cleanup of residential soil contaminated with
3 PCBs that was brought to the properties as fill from GE. GE has cleaned up approximately 170
4 properties to date under this program.
5 Other properties or areas in Pittsfield and the surrounding communities have been discovered
6 over the years to have received waste from the GE facility and/or are contaminated with PCBs,
7 including the Pittsfield Landfill, Rose Disposal Site (National Priorities List [NPL] site), and
8 Dorothy Amos Park located on the West Branch of the Housatonic River. Actions to address
9 these properties have been taken or investigation is underway.
10 The highest concentrations of Aroclors 1254 and 1260 have been detected in the vicinity of the
11 plant and downstream of Building 68 (WESTON 2000; BBL 1994, 1995; O'Brien & Gere
12 Engineers, Inc. 1995). Widespread contamination of the river downstream of the GE facility has
13 resulted from the transport of PCB-contaminated river sediment and floodplain soil by river
14 flow, sediment transport, and overbank flooding (WESTON 2000). Total PCBs have been
15 detected at concentrations of greater than 1 ppm in floodplain soil as far downstream as
16 Bartholomew's Cobble in Massachusetts, close to the Massachusetts-Connecticut state line.
17 1.3 REGULATORY BACKGROUND
18 The GE/Housatonic River site has been subject to regulatory investigations dating back to the
19 early 1980s. For several years, these investigations were consolidated under the following
20 regulatory mechanisms: two Administrative Consent Orders (ACOs) with MDEP and a
21 Corrective Action Permit with EPA under the Hazardous and Solid Waste Amendments to the
22 Resource Conservation and Recovery Act (RCRA).
23 In 1991, EPA issued a RCRA Corrective Action Permit to the GE Pittsfield facility. Following
24 appeals by GE and others, and subsequent modification, the permit became effective in 1994.
25 The permit included the 254-acre facility, some filled former oxbows, Silver Lake, the
26 Housatonic River and its floodplains and adjacent wetlands, and all sediment contaminated by
27 PCBs migrating from the GE facility.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-9
-------
1 In addition to the permit, the two ACOs between GE and MDEP became effective in 1990 and
2 included those areas defined in the permit, as well as additional filled former oxbows and
3 Allendale Elementary School. Under the ACOs, GE has performed several investigations and
4 short-term cleanups.
5 EPA proposed the site to the Superfund National Priorities List (NPL) in September 1997.
6 Several federal and state government agencies and GE entered into negotiations late in 1997 with
7 the goal of reaching a comprehensive settlement, which included remediation, redevelopment,
8 and restoration components.
9 In September 1998, representatives of the federal and state government agencies, GE, the City of
10 Pittsfield, and the Pittsfield Economic Development Authority reached an agreement in principle
11 relating to GE's Pittsfield facility, other contaminated areas in Pittsfield, and the Housatonic
12 River. This agreement was translated into a Consent Decree, lodged with the federal court on 7
13 October 1999, and entered by the court on 27 October 2000. The agreement provides for, among
14 other things, the cleanup of the GE plant facility, cleanup and restoration of the former oxbows,
15 cleanup and restoration of Silver Lake, cleanup of Allendale School, environmental restoration
16 projects related to the Housatonic River and floodplains, monetary compensation for natural
17 resource damages, and government recovery of past and future response costs. Entry of the
18 agreement also makes possible large-scale redevelopment of the GE facility.
19 The GE/Housatonic River site is made up of several separate response actions (as described in
20 the Consent Decree), including three actions in the river:
21 ¦ Upper V2-Mile Reach Housatonic River Removal Action ('/2-Mile Reach).
22 "1 /4-Mile Housatonic River Removal Action (1 /4-Mile Reach).
23 ¦ Rest of River.
24
25 and actions outside the river, including:
26
27
28
29
30
31
GE plant site soil remediation.
Unkamet Brook and floodplain remediation.
Hill 78/Building 71 consolidation areas.
Groundwater remediation.
Former oxbow areas.
Allendale School.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-10
7/11/03
-------
1 ¦ Floodplain current residential and nonresidential properties.
2 ¦ Silver Lake.
3
4 The primary COPCs are PCBs, specifically Aroclor 1260 and, to a lesser extent, 1254. Other
5 COPCs include volatile organics, dioxins/furans, polycyclic aromatic hydrocarbons (PAHs),
6 semivolatiles, and metals. These contaminants vary in their distribution in different areas.
7 EPA completed an investigation of the Rest of River below the l^-Mile Reach into Connecticut,
8 which focused on collecting information for and preparing the human health and ecological risk
9 assessments, and modeling PCB fate and transport in the river. Under the terms of the Consent
10 Decree, both of the risk assessments and three aspects of the modeling effort are to undergo
11 formal external Peer Review, with the review of the Modeling Framework Design having taken
12 place in April 2001. The ecological risk assessment, together with the human health risk
13 assessment and the model of PCB fate, transport, and bioaccumulation, will inform EPA's
14 decision on what additional remedial actions, if any, may be required in the river and floodplain
15 below the confluence.
16 Following the investigations, as required in the Revised RCRA Permit, GE has prepared a
17 Supplemental RCRA Facility Investigation Report (BBL and QEA 2003), will propose cleanup
18 levels (Interim Media Protection Goals), and will analyze cleanup alternatives (Corrective
19 Measures Study) for consideration by EPA. EPA will propose the draft Statement of Basis
20 (cleanup plan, scheduled for 2006) for the Corrective Measure(s) for the Rest of River and, after
21 public comment, will finalize the Statement of Basis. GE and other members of the public may
22 then appeal EPA's decision to the EPA Environmental Appeals Board (EAB) and the First
23 Circuit District Court. GE is then required, under the Consent Decree, to implement and pay for
24 the remedy selected after resolution of any appeals. The Rest of River response action, if any,
25 will be implemented through a modification to the Revised RCRA Permit and an amendment to
26 the CERCLA Consent Decree, and is estimated to begin in 2007.
27 1.4 SITE DESCRIPTION
28 The Rest of River portion of the Housatonic River flows through one of the most biologically
29 diverse regions of Massachusetts (Barbour et al. 1998) and Connecticut. Dams play an integral
30 role in the downstream migration of PCBs and other COPCs from the GE facility.
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1-11 7/11/03
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
The first 10.5 miles (16.9 km) from the confluence to the headwaters of Woods Pond is referred
to as Reach 5 (Figure 1.4-1). Other than the initial 0.5-mile (0.8-km) reach bordering the GE
facility, Reach 5 has the highest concentrations and highest frequency of detections of PCBs in
sediment. Reach 5 is subdivided further into four parts: Reach 5A, from the confluence to just
above the Pittsfield Wastewater Treatment Plant (WWTP); Reach 5B, from the WWTP to
Roaring Brook; Reach 5C, from Roaring Brook to the headwaters of Woods Pond; and Reach
5D, the backwaters above Woods Pond (Woods Pond is Reach 6) (Figures 1.4-2 and 1.4-3).
The Housatonic River meanders through Reach 5A, with widths between 50 and 120 ft (15 and
37 m) and depths up to 11 ft (3.4 m) (HEC 1996). Aquatic habitat includes snags (large woody
debris), undercut banks, and rocks. Land use in this section is predominantly forested and
cleared, with some residential areas. Reach 5B is similar to Reach 5A from the WWTP to New
Lenox Road. The land near New Lenox Road is predominantly agricultural and forested. Below
New Lenox Road, the river widens (60 to 160 ft [18 to 48 m]) and becomes shallower (4 to 8 ft
[1.2 to 2.4 m]). This portion of Reach 5B is dominated by a broad wetland floodplain, ranging
from 800 to 3,000 ft (240 to 910 m) wide (see Appendix A). Reach 5C is similar to Reach 5B,
although as the Housatonic River approaches Woods Pond, the velocity decreases and deep pools
occur (up to and exceeding 7 ft [2 m]), created by large snags that divert water flow, and the
effect of Woods Pond Dam becomes apparent. Dense vegetation lines the banks of the river in
the upper portion of this section, while extensive backwaters border the lower section. Reach 5D
consists of several upstream backwater areas associated with Woods Pond and covers more than
120 acres (49 ha).
Reach 6 begins 10.5 miles (16.9 km) downstream of the confluence at Woods Pond. The pond is
approximately 0.2 mile (0.3 km) in length and has an area of 60 acres (24 ha) (Figure 1.4-4).
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-12
-------
Woods Pond |
LEGEND:
O Town/City
N Roads
K t
f ^ Reach Division Line
Housatonic River
State Park
Municipal Boundary
10-Year Floodplain
N
V
s
0.2
0
0.2 0.4
0.6 Miles
0.4
0
0.4
0.8 Kilometers
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 1.4-1
HOUSATONIC RIVER,
REACH 5
| O:\gepitt\aprs\era_figures2.apr I layout - fig1.4-11 o:\gepittVepsfiles\plots\in\rig1-4-1.eps 110:13 AM, 3/7/20031
-------
LEGEND:
O Town/City
N Roads
K t
f ^ Reach Division Line
Housatonic River
State Park
Municipal Boundary
10-Year Floodplain
N
V
s
0.05
0
0.05
0.1 Miles
0.07
0
0.07
0.14 Kilometers
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 1.4-2
HOUSATONIC RIVER,
REACHES 5A AND 5B
| O:\gepitt\aprs\era_figures2.apr I layout - fig 1.4-21 o:\gepitttepsfiles\plots\in\fig1-4-2.eps 110:16AM, 3/7/20031
-------
LEGEND:
O Town/City
Dam
N Roads
~ #
' *
Reach Division Line
Housatonic River
State Park
Municipal Boundary
10-Year Floodplain
N
V
s
0.04
0
0.04
0.08 Miles
0.06
0
0.06
0.12 Kilometers
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 1.4-3
HOUSATONIC RIVER,
REACHES 5C AND 5D
| O:\gepitt\aprs\era_figures2.apr | layout - fig 1.4-31 o:\gepitttepsfiles\plots\in\fig1-4-3.eps 110:35 AM, 3/7/20031
-------
) JCNOX
Reach 6 Shallow
Lenox Station O
LEGEND:
O Town/City
(2) Dam
H Roads
^^ Reach Division Line
Housatonic River
State Park
Municipal Boundary
10-Year Floodplain
N
V
s
0.02
0
0.02 Miles
0.03
0
0.03 Kilometers
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 1.4-4
HOUSATONIC RIVER,
REACH 6
| O:\gepitt\aprs\era_figures2.apr I layout - fig1.4-41 o:\gepitttepsfiles\plots\in\fig1-4-4.eps 13:36 PM, 6/26/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
This reach contains the first impoundment downstream from the GE facility and is a depositional
environment (HEC 1996). The water in Woods Pond is relatively slow-moving and contains
aquatic habitat characteristic of a standing water environment. The maximum depth is 16 ft (4.9
m), but most of the pond is 1 to 3 ft (0.3 to 0.9 m) deep (HEC 1996; Stewart Laboratories, Inc.
1982; CR Environmental 1998). The banks of the pond provide extensive cover, such as
overhanging vegetation, woody debris, rock piles, and submerged macrophytes. The Woods
Pond Dam was built in 1864. In 1989, GE replaced the original dam with a concrete weir dam
located 180 ft (55 m) downstream of the original dam site.
Reach 7 extends 18.6 miles (29.9 km) from Woods Pond to the upstream end of Rising Pond in
Great Barrington (Figure 1.4-5). There are five dams in this reach, and the river has an average
gradient of 14.5 ft (4.42 m) per mile, and an average depth of 1 ft (0.3 m) (Stewart Laboratories,
Inc. 1982). Agricultural activity becomes more common in this area than in the upstream
reaches.
Reach 7 ends above Rising Pond, which is Reach 8 (Figure 1.4-5). This 45-acre (18-ha) pond
was created by the construction of a dam at the Rising Paper Company (WESTON 2000).
Rising Pond has depositional characteristics similar to Woods Pond.
Reach 9 begins downstream of Rising Pond and extends for approximately 24.6 miles (39.6 km)
to the Massachusetts/Connecticut state line (Figure 1.4-5). It contains low-gradient sections with
river habitat, as well as moderate gradient sections with riffle habitat. This reach is wide with
flat floodplains and several oxbows, and agriculture is a predominant land use.
Reach 10 begins at the Massachusetts/Connecticut border and extends 7.4 miles (12 km) to the
dam at Great Falls Village (Figure 1.4-6). The river characteristics are similar to those of Reach
9, with a meandering river course. Reach 11 begins on the downstream side of the dam at Great
Falls and ends at Cornwall Bridge, where Route 7 crosses the river (Figure 1.4-6). This reach is
11.5 miles (18.5 km) long. Reach 11 is mostly shallow and fast flowing, and much of the reach
is designated as a Trout Management Area. Reach 12 extends from Cornwall Bridge to the dams
at Bulls Bridge (Figure 1.4-6), a length of 13.1 miles (21.1 km). The river is relatively straight
through this reach and flows quickly for most of the run. Near the town of Kent, the river slows
and deepens as it enters the backwaters from the dams at Bulls Bridge. Reach 13 starts on the
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1 -17 7/11/03
-------
Richmon®Fun(iaci
JJne
\ M
St^pkbridge
P
Tanglewood
o
Statior
nO
Can© Mati-KeANac
'qdv 1 ^
' Beachwotf
^lahkfi
ena
LenVx Dale)o)
>20<
Old Woods Rbnd Dam
Wood^Pond Dam
)ctober Mountain SF
JiiagraJ\/lil^3?am
/Austerlitz
Camp Kingsmoi
o
at g
> Green River
Alforc
^est Stockbriilge Csater^
Rockdale Mill
Gler
Glendala*
Willia^vill^VyMonument IwHs dam #2
Hdusa^ic /^^Monument Mils Dam #3
LisingdaJ
kCH 8
Jsing Pbnd D'&m
\Van Deusenville
Columbia Mill Dam
3/Eagle Mills Dam
Eatofi-Bikeman OamVWV^Lee
lam
Beartown SF
North Bsremont
Great Barrington
Berkshire Heights
'S. Berkshire Power & Elep. Dam
elcher Square
Camp Meadofwlark
o
Egremont
Plain
Camp Jayson
\o.
East Mokntai:
SF
lartsville
o
Hephzjl
siglits
^J) New ]V|aritf<5rough
5 Mill River
oSouthfield X Sandisfield SF
BERKSHIRE
Mass.
Mo\mt
\o
¦¦¦mwOb \
Hammertown „,°nn' /
WeatogueJ
^racom^ ^
^shlpy Falls
^ast Sheffield;
JJ
° Clayton
° Sodom ^ana
Valley
n
LEGEND:
O Town/City
(2) Dam
Reach 5
Roads
Reach 6
Reach 7
Housatonic River
Housatonic Watershed
Reach 8
1 1
| | State Park
Reach 9
County Boundary
Reach 10
N
V
s
1
0
1
2 Miles
1.5
0
1.5
3 Kilometers
Ecological Risk Assessment
GE/Housatonlc River Site
Rest of River
FIGURE 1.4-5
HOUSATONIC RIVER,
REACHES 7 TO 9
| O:\gepitt\aprs\era_figures2.apr I layout - fig1.4-51 o:\gepitttepsfiles\plots\in\fig1-4-5.eps 13:43 PM, 6/26/20031
-------
ander —-JoyceV
Mass:
Conn.
^Whifehouse
Crossin
>» ^ VHammertown ... . .
MoVnt M Waconicr-x Weatogue/
©
OTwin
Lakes
Grov
SOUTHFIELD
OClaylon
CanaaiTValley
Sodom
aCanaan
^East Canaan
WestNorfoll
ck
Aain SP
lousatonic'
, Spencer Corners
IrondaV O S(ate
llerton ^Lin
GREAT FALLS£)AN
nesyille
Natthealt Center
The I
itemanStatio
O
Benni:
South 1
urkevCobble
irfolk0^
Sharon Statioa O
Sharon
CEDAR HILL [
{Kent;
Upper lylerryall
^rwinsville
ylordkvilli
ricksoaCorner
iweKMerryall
New Preston
bleDa
Mount Tb
Romford
fast Morris
LakesiS
pevuleP
i Sandy Beach
North fie
files' ShelterO
, Northville.
«\
$
orse Heave
ckRi
rdCp
Baker Cornetl
O
3MEACHERY DAI
ewojod Spring!
\R6^1
kissville
;bury Station
LEGEND:
o
Dam
Roads
Town/City
Housatonlc River
I I State Park
County Boundary
Housatonlc Watershed
Reach 9
Reach 10
Reach 11
Reach 12
Reach 13
Reach 14
N
4
V
s
1
0
1 2
3 4 Miles
2
0
2
4 Kilometers
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 1.4-6
HOUSATONIC RIVER,
REACHES 10 TO 13
| O:\gepitt\aprs\era_figures2.apr | layout - fig 1.4-61 o:\gepitttepsfiles\plots\in\fig1-4-6.eps 110:20 AM, 3/7/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
downstream side of the dams at Bulls Bridge and runs 10.9 miles (17.5 km) to the Bleachery
Dam at New Milford, CT (Figure 1.4-6). The Bleachery Dam is virtually submerged as a result
of the backwater created by the Shepaug Dam farther downstream. The river meanders more
than in the previous reach and, as in Reaches 11 and 12, flows quickly.
Reach 14, from the Bleachery Dam to Shepaug Dam, is known as Lake Lillinonah (Figure 1.4-
7). The reach is 11.5 miles (18.5 km) long. The Shepaug Dam is approximately 100 ft (30 m)
high. The backwater effect from the Shepaug Dam extends all the way to the upstream
Bleachery Dam. The Shepaug Dam is used for power generation and this may affect water
levels during the year. Water movement is slow through this reach and the river is deep. Reach
15 encompasses Lake Zoar, from Shepaug Dam to Stevenson Dam (Figure 1.4-7). This
predominantly slow-moving reach is 10.2 miles (16.4 km) long. The backwater effect of the
Stevenson Dam extends upstream for almost the entire reach. The Stevenson Dam is
approximately 100 ft (30 m) high and supports power generation. Some homes and boat
launches are found on Lake Zoar.
Reach 16 is Lake Housatonic and is bounded by the Stevenson Dam and the Derby-Shelton Dam
(Figure 1.4-7). The reach is 6.0 miles (9.7 km) long and, like the previous two upstream reaches,
is slow moving. The Derby-Shelton Dam (approximately 25 ft [7.6 m] high) is smaller than
either the Shepaug or Stevenson Dams. More homes and boat launches occur along this reach.
Reach 17, from Derby-Shelton Dam to Long Island Sound, is 13.7 miles (22.0 km) long (Figure
1.4-7). This reach is entirely tidally influenced. It is shallow in its upstream portions and
deepens downstream. The Naugatuck River enters the Housatonic River approximately 2 miles
(3.2 km) from the upstream end of this reach.
1.5 OVERVIEW OF TECHNICAL APPROACH
This ERA follows the technical approach and guidelines detailed in EPA's Ecological Risk
Assessment Guidance for Superfund: Process for Designing and Conducting Ecological Risk
Assessments (EPA, 1997a). Additional documents were also consulted, including, but not
limited to, the following:
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-20
-------
$
orse Heave
Well&il
I^AlCHERY DAfy
3 ad Spring^^l VRd^cbury St^tion^
kissville
Groves
Lane
3ridgbwater
*forifr\Woodbvi
food
Pome
, Oakville;
O
Faftnount"
BunkeriH
fest Side Hi*
erville I
0
LongHjll
WestVie
(lebt
Hiliprest„ „ ,
OCMnok
>o
; Town Rlot/1
Hi#
r Falls
iqlyffest \ | (^Candlewfjbd
3jub
) j^ajce jpa,
oil
;woW*fiores\
IrooktieK
Field
eOV
Gle
O
lillville/j
iug HaroprO
in Point) Ajua Visti
, , Pleasant Acres \
i \o
Lam
JO
lyestt
itowiyTS^
ntain SP
iethel'
ySHEPA
Jouthspritain/"
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
¦ Framework for Ecological Risk Assessment (EPA/630/R-92/001, 1992a).
¦ Guidelines for Ecological Risk Assessment (EPA/630/R-95-002F, April 1998).
¦ Ecological Risk Assessment Issue Papers (EPA/630/R-94/009, November 1994,
1994e).
¦ Wildlife Exposure Factors Handbook, Volumes I and II (EPA 600/R-93/187a and
187b, December 1993, 1993a).
¦ Guidance for Disposal Site Risk Characterization: Method 3 Environmental Risk
Characterization (MDEP 1996).
¦ The Role of BTAGs in Ecological Assessment, ECO Update, Volume 1, Number 1
(EPA 1991a).
¦ Ecological Assessment of Superfund Sites: An Overview, ECO Update, Volume 1,
Number 2 (EPA 1991b).
¦ The Role of Natural Resource Trustees in the Superfund Process, ECO Update,
Volume 1, Number 3 (EPA 1992b).
¦ Using Toxicity Tests in Ecological Risk Assessment, ECO Update, Volume 2, Number
1 (EPA 1994a).
¦ Catalogue of Standard Toxicity Tests for Ecological Risk Assessment, ECO Update,
Volume 2, Number 2 (EPA 1994b).
¦ Field Studies for Ecological Risk Assessment, ECO Update, Volume 2, Number 3
(EPA 1994c).
¦ Selecting and Using Reference Information in Superfund Ecological Risk
Assessments, ECO Update, Volume 2, Number 4 (EPA 1994d).
¦ Ecotox Thresholds, ECO Update, Volume 3, Number 2 (EPA 1996).
¦ RAGS, Volume 3, Part A: Process for Conducting Probabilistic Risk Assessment
(EPA 540-R-02-002, December 2001).
¦ Guiding Principles for Monte Carlo Analysis (EPA/63 C/R-97/001, 1997b).
¦ Ecological Risk Assessment and Risk Management Principles for Superfund Sites
(EPA 1999).
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
The Ecological Risk Assessment Guidance for Superfund (EPA 1997a) details an eight-step
process for conducting an ERA (Figure 1.5-1). This document provides the user with a basic
framework for the ERA process and ensures a consistent approach to conducting ERAs. In
addition to these steps, there are several scientific/management decision points (SMDPs). These
are opportunities for the risk manager and the risk assessment team to communicate ideas
concerning the scope, focus, and direction of the ERA. The first two steps of the ERA process
(Screening-Level Problem Formulation and Screening-Level Exposure Estimate and Risk
Calculation) were first addressed in the Upper Reach-Housatonic River Ecological Risk
Assessment (WESTON 1998) and subsequently refined in Appendix B of this document. Steps
3, 4, and 5 (Baseline Problem Formulation, Study Design and DQO Process, and Verification of
Field Data Analysis) are iterative components of the eight-step ERA process. Steps 3 through 5
were initially presented in the Supplemental Investigation Work Plan for the Lower Housatonic
River (SIWP) (WESTON 2000) and were modified as necessary during the data collection phase
of the project. Steps 6 and 7 (Site Investigation and Data Analysis and Risk Characterization)
are presented in detail in the following sections and appendices. Step 8 (Risk Management) will
be addressed after the ERA has undergone Peer Review.
1.5.1 Problem Formulation
Problem formulation is an important component of the ERA process that establishes the goals,
objectives, and scope for the ERA. Products of problem formulation include the identification of
assessment endpoints, illustration of exposure pathways (relating fate and transport to ecological
effects), a conceptual model depicting the relationships between COPCs and the assessment
endpoints, and risk hypotheses and questions that can be drawn from evident or suspected
effects. The problem formulation portion of the ERA is discussed in Section 2 and was outlined
in the Supplemental Investigation Work Plan for the Lower Housatonic River (WESTON 2000).
1.5.1.1 Physical and Ecological Characterization
An extensive physical and ecological characterization of the Housatonic River is presented in
Section 2.2 and Appendix A (Ecological Characterization) of this document. These sections
detail the physical setting, habitats, and biotic communities of the river in both the aquatic and
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1 "23 7/11/03
-------
Ecological Risk Assessment for the Housatonic River
CD
C
'-4—'
c
CO
o
'x
»
LD
03
0
E
Q.
o
4—
E
_c
o
O
c
o
»
o
0
CC
-t—»
03
Q
STEP 1: SCREENING-LEVEL:
•Site Visit
•Problem Formulation
•Toxicity Evaluation
STEP 2: SCREENING-LEVEL:
•Exposure Estimate
•Risk Calculation
STEP 3: PROBLEM FORMULATION
P
Toxicity Evaluation
1
Assessment
Endpoints
Conceptual Model
Exposure Pathways
tl
Questions/Hypotheses
Li
STEP 4: STUDY DESIGN AND DQO PROCESS
•Lines of Evidence
•Measurement Endpoints
•Work Plan and Sampling and Analysis Plan
STEP 5: VERIFICATION OF FIELD
SAMPLING DESIGN
STEP 6: SITE INVESTIGATION AND
DATA ANALYSIS
Risk Assessor
and Risk Manager
Agreement
SMDP
SMDP
SMDP
SMDP
~
[SMDP]
STEP 7: RISK CHARACTERIZATION
STEP 8: RISK MANAGEMENT
SMDP
Legend:
SMDP - Scientific/management decision point
[SMDP] - only if change to the sampling and analysis plan is necessary
Source: EPA (U.S. Environmental Protection Agency), Environmental
Response Team. 1997. Ecological Risk Assessment Guidance for
Superfund: Process for Designing and Conducting Ecological Risk
Assessments. Interim Final. EPA 540-R-97-006.
Housatonic River Project
Pittsfield, Massachusetts
Figure 1.5-1
EIGHT-STEP ECOLOGICAL RISK
ASSESSMENT PROCESS FOR SUPERFUND
MK01|0:\20123001.096\ERA_PB\ERA_PB_1_Fig. 1.5-1. ppt
1-24
-------
1 terrestrial environments. The physical and ecological characterizations aid in identifying
2 representative species and exposure pathways, provide information for the exposure analyses,
3 and can inform risk managers on the potential impacts of future remedial actions.
4 1.5.1.2 Stressors and Their Sources
5 Investigations of the nature and extent of contaminants in the Housatonic River watershed have
6 previously been conducted by GE, EPA, and others. PCBs have been identified as the main
7 COPC, and other contaminants such as dioxins and furans and PAHs have also been identified at
8 the GE facility. In addition, other COPCs, such as pesticides, were detected in the PSA,
9 although at lower concentrations and frequencies of detection. In Section 2.3, the sources,
10 amounts, and patterns of contaminant releases and receiving bodies are presented.
11 1.5.1
12
13
14
15
16
17
18
19 The purpose of the Pre-ERA (Appendix B) was to identify contaminants that warranted more
20 detailed analyses in the ERA, and those that could be removed from further consideration
21 because they pose minimal risk. For those contaminants that screened through to the ERA, the
22 primary media of concern as well as the sections of the study area that are potentially impacted
23 are identified. A summary of the Pre-ERA is provided in Section 2.4. The complete Pre-ERA is
24 included as Appendix B to this document.
25 1.5.1.4 Fate and Transport of Contaminant Stressors
26 An overview of the environmental behavior of PCBs and other COPCs is presented in Section
27 2.5. This section includes discussions of the transport of the contaminants from their point(s) of
.3 Pre-Ecological Risk Assessment
COPC vs COC
In the ERA, contaminants of potential concern (COPC) refer to contaminants
considered before, during, and immediately after the Pre-ERA process. A
contaminant is considered a contaminant of concern (COC) if it has passed through
all screening-level processes and is included as part of the exposure and effects
assessment conducted for a specific assessment endpoint.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-25
-------
1 release, partitioning behavior in different media, and biotic and abiotic degradation in these
2 media.
3 1.5.1.5 Effects on Representative Species
4 The effects and mechanisms of toxicity to biota of the contaminants identified as COPCs within
5 the Housatonic River and floodplain are discussed, with an emphasis on PCBs, in Section 2.6,
6 and in further detail in the effects assessment portion of each assessment endpoint section and
7 corresponding appendix.
8 1.5.1.6 Conceptual Model, Selection of Assessment and Measurement
9 Endpoints, and Analysis Plan
10 The conceptual model outlined in Section 2.7 describes the relationship between the COPCs and
11 the biota at the site. Development of a conceptual model includes review of sources of
12 contamination, evaluation of the spatial scale for the assessment, description of the exposure
13 pathways, and formulation of risk questions to be addressed.
14 Assessment and measurement endpoints are defined and described in Section 2.8. An
15 assessment endpoint is defined as the "explicit expression of the environmental value that is to
16 be protected" (EPA 1997a). Because it is often unrealistic to perform an assessment for all
17 species present at a contaminated site, species or populations are often grouped based on their
18 similarities (e.g., exposure pathway, contaminant sensitivity) or societal importance (e.g.,
19 threatened and endangered species), and assessment endpoints are established for these groups of
20 similar species. A measurement endpoint is defined as "a measurable ecological characteristic
21 that is related to the valued characteristic chosen as the assessment endpoint," and is a measure
22 of biological effects (e.g., mortality, reproduction, growth) (EPA 1997a). Measurement
23 endpoints are frequently numerical expressions of observations (e.g., toxicity test results,
24 community diversity measures) that can be compared statistically to a reference site to detect
25 adverse responses to a site contaminant (EPA 1997a).
26 Section 2.9 describes the analytical approach used to estimate risks and the weight-of-evidence
27 approach used to develop the conclusions.
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1-26 7/11/03
-------
1 1.5.2 Assessment of Representative Species
2 Sections 3 through 11 (and their corresponding appendices) provide an overview of the exposure
3 assessment, the effects assessment, and the risk characterization for each representative species
4 or representative group of species.
5 1.5.2.1 Exposure Assessment
6 The exposure assessment sections include a description of the data collection activities and the
7 studies conducted to determine concentrations of COPCs in water, soil, sediment, and biota
8 samples. Previous sampling and monitoring studies are also described in this section. Variation
9 in PCB and COPC concentrations over space and time in each environmental medium is briefly
10 characterized. For each assessment endpoint, one or more representative species were selected
11 and an appropriate exposure model identified.
12 1.5.2.2 Effects Assessment
13 The effects assessment sections begin with an overview of the toxicity of PCBs and other
14 COPCs. For each major representative species group and COPC, the effects literature was
15 reviewed. The goal of this review was to identify key studies that could be used to develop
16 effects metrics for use in risk characterization. The effects metrics developed ranged from
17 concentration- or dose-response curves to benchmarks depending on the quality and relevance of
18 the data available.
19 1.5.2.3 Risk Characterization
20 The risk characterization sections for each assessment endpoint consider three major lines of
21 evidence (where available): (1) comparison of estimated exposures to laboratory-based effects
22 metrics, (2) results of in situ or whole media toxicity tests, and (3) results of field surveys.
23 Probabilistic methods were used to integrate COPC exposure distributions in the study area with
24 laboratory-derived benchmarks and effects curves. The format of the discussion includes an
25 overview of the study, statistical analyses of the results and conclusions stating the observed
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1-27 7/11/03
-------
1 effects, and likely causal agents. The risk characterization concludes with a weight-of-evidence
2 assessment for each assessment endpoint. Primary sources of uncertainty are also identified.
3 1.6 DATA SOURCES
4 The Housatonic River ERA generally relies on data from studies and research specifically
5 designed for this assessment. Field surveys were conducted to support the ecological
6 characterization and ecological risk assessments for benthic invertebrates, amphibians, fish,
7 birds, and mammals in the Housatonic River floodplain. Prior to the surveys, literature reviews
8 were conducted to establish historic populations and habitats for species within the study area.
9 Surveys were also conducted at several reference sites (i.e., areas of relatively low contamination
10 within the Housatonic River watershed).
11 A variety of site-specific studies were conducted, including the following:
12 ¦ Survival, growth, and reproduction of benthic organisms as part of the Sediment
13 Quality Triad (SQT) approach.
14 ¦ Reproductive success of amphibians in the Housatonic River floodplain and the
15 effects of exposure to PCBs and other COPCs on these species.
16 ¦ Studies with largemouth bass {Micropterus salmoides) to determine if exposure of
17 adults to PCBs and COPCs in river water and sediment adversely affect the survival
18 and development of offspring.
19 ¦ Investigation of tree swallows (Tachycineta bicolof) to determine the extent to which
20 PCBs and other COPCs are impairing their reproduction.
21 "A reproductive toxicology study with farm-raised mink (Mustela vison) exposed to
22 PCBs and other COPCs in their diet from fish collected from the Housatonic River.
23 These and other studies are described in more detail in the Supplemental Investigation Work Plan
24 for the Lower Housatonic River (WESTON 2000) and its appendices. Information on study
25 design, methodology, and quality assurance (QA)/quality control (QC) procedures can also be
26 found in the Supplemental Investigation Work Plan for the Lower Housatonic River.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-28
-------
1 In addition, GE conducted the following studies in the PSA (unless otherwise noted):
2
¦ Productivity of robins.
3
¦ Productivity and density of belted kingfishers (Ceryle alcyon).
4
5
¦ Analysis of context-dependent effects on early life stages on wood frogs (Rana
sylvatica).
6
¦ Spatial and demographic effects on tree swallows (performed in Canada).
7
¦ Demographics of short-tailed shrews (Blarina brevicauda).
8
¦ Field observations of presence/absence of mink.
9
¦ Evaluation of largemouth bass habitat, population structure, and reproduction.
10 EPA project data are managed using a relational structure in Microsoft Access. The database
11 contains information on PCBs and other COPCs in soil, sediment, and tissue samples, and other
12 field study data collected by EPA and other parties, constituting more than 2 million records.
13 Arc View (geographic information system [GIS]) was used to illustrate spatial patterns. Data
14 originating from previous or concurrent studies conducted by GE and other sources were used if
15 data quality was acceptable. The procedure followed for evaluating data quality of historical
16 studies is described in Appendix C.
17 1.7 QA/QC
18 QA and QC procedures and techniques are established to guide data collection, analysis,
19 modeling, administration, and auditing. Three documents, the Quality Assurance Project Plan
20 (WESTON 2001a), the Supplemental Investigation Work Plan for the Lower Housatonic River
21 (WESTON 2000), and the Field Sampling Plan (WESTON 2001b), outline the QA/QC
22 procedures and techniques used in the studies conducted by EPA in support of this ecological
23 risk assessment. These documents also provide details on the methods used in the collection and
24 analyses of data.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-29
7/11/03
-------
1 1.8 REFERENCES
2 Barbour, H., T. Simmons, P. Swain, and H. Woolsey. 1998. Our Irreplaceable Heritage,
3 Protecting Biodiversity in Massachusetts. Massachusetts Heritage and Endangered Species
4 Program and The Massachusetts Chapter of the Nature Conservancy. 83 pp.
5 BBL (Blasland, Bouck & Lee, Inc.). 1994. MCP Interim Phase II Report and Current
6 Assessment Summary for East Street Area 2/U.S. EPA Area 4. Volumes I, II, III, IV, V, VI, VII,
7 IX, X, and XII.
8 BBL (Blasland, Bouck & Lee, Inc.). 1995. MCP Supplemental Phase II Scope of Work and
9 Proposal for RCRA Facility Investigation of Unkamet Brook Area/U. S. EPA Area 1.
10 BBL (Blasland, Bouck & Lee, Inc.) and QEA (Quantitive Environmental Analysis, LLC). 2003.
11 Housatonic River- Rest of River RCRA Facility Investigation Report Volume 1. Prepared for
12 General Electric Company. January 2003.
13 CR Environmental. 1998. Housatonic River Supplemental Investigation Sub-bottom Profiling,
14 Woods and Rising Ponds. Prepared for Roy F. Weston, Inc. December 1998.
15 EPA (U.S. Environmental Protection Agency). 1991a. The Role of BTAGs in Ecological
16 Assessment. Office of Solid Waste and Emergency Response, Office of Emergency and
17 Remedial Response, Hazardous Site Evaluation Division (OS-230). ECO Update, Intermittent
18 Bulletin, Volume 1, Number 1. Publication 9345.0-051. September 1991.
19 EPA (U.S. Environmental Protection Agency). 1991b. Ecological Assessment of Superfund Sites:
20 An Overview. Office of Solid Waste and Emergency Response, Office of Emergency and
21 Remedial Response, Hazardous Site Evaluation Division (OS-230). ECO Update, Intermittent
22 Bulletin, Volume 1, Number 2. Publication 9345.0-051. December 1991.
23 EPA (U.S. Environmental Protection Agency). 1992a. Framework for Ecological Risk
24 Assessment. Risk Assessment Forum, Washington, DC. EPA/630R-92/001.
25 EPA (U.S. Environmental Protection Agency). 1992b. The Role of Natural Resource Trustees in
26 the Superfund Process. ECO Update, Volume 1, Number 3.
27 EPA (U.S. Environmental Protection Agency). 1993a. Wildlife Exposure Factors Handbook.
28 Volumes I and II. EPA/600/R-93/187a, EPA/600/R-93/187b. U.S. Environmental Protection
29 Agency, Office of Research and Development.
30 EPA (U.S. Environmental Protection Agency). 1994a. Using Toxicity Tests in Ecological Risk
31 Assessment. Office of Solid Waste and Emergency Response, Office of Emergency and
32 Remedial Response, Hazardous Site Evaluation Division (OS-230). ECO Update, Intermittent
33 Bulletin, Volume 2, Number 1. Publication 9345.0-051, EPA 540-F-94-012, PB94-963303.
34 September 1994.
35 EPA (U.S. Environmental Protection Agency). 1994b. Catalogue of Standard Toxicity Tests for
36 Ecological Risk Assessment. Office of Solid Waste and Emergency Response, Office of
MK01|O:\20123001.096\ERA_PB\ERA_PB_1.DOC 1-30 7/11/03
-------
1 Emergency and Remedial Response, Hazardous Site Evaluation Division (5204G). ECO Update,
2 Intermittent Bulletin, Volume 2, Number 2. Publication 9345.0-051, EPA 540-F-94-013, PB94-
3 963304. September 1994.
4 EPA (U.S. Environmental Protection Agency). 1994c. Field Studies for Ecological Risk
5 Assessment. Office of Solid Waste and Emergency Response, Office of Emergency and
6 Remedial Response, Hazardous Site Evaluation Division (5204G). ECO Update, Intermittent
7 Bulletin, Volume 2, Number 3. Publication 9345.0-051. EPA 540-F-94-014, PB94-963305.
8 September 1994.
9 EPA (U.S. Environmental Protection Agency). 1994d. Selecting and Using Reference
10 Information in Superfund Ecological Risk Assessments. Office of Solid Waste and Emergency
11 Response, Office of Emergency and Remedial Response, Hazardous Site Evaluation Division
12 (5204G). ECO Update, Intermittent Bulletin, Volume 2, Number 4. Publication 9345.0-101, EPA
13 540-F-94-050, PB94-963319. September 1994.
14 EPA (U.S. Environmental Protection Agency). 1994e. Ecological Risk Assessment Issue Papers.
15 EPA/630/R-94/009. November 1994.
16 EPA (U.S. Environmental Protection Agency). 1996. Ecotox Thresholds. Office of Solid Waste
17 and Emergency Response, Office of Emergency and Remedial Response. ECO Update,
18 Intermittent Bulletin, Volume 3, Number 2. Publication 9345.0-12FSI, EPA 540/F-95/038,
19 PN95-963324. January 1996.
20 EPA (U.S. Environmental Protection Agency). 1997a. Ecological Risk Assessment Guidance for
21 Superfund: Process for Designing and Conducting Ecological Risk Assessments. Interim Final.
22 EPA 540-R-97-006. U.S. Environmental Protection Agency, Environmental Response Team.
23 EPA (U.S. Environmental Protection Agency). 1997b. Guiding Principles for Monte Carlo
24 Analysis. EPA/63C/R-97/001. March 1997.
25 EPA (U.S. Environmental Protection Agency). 1998. Guidelines for Ecological Risk Assessment.
26 Risk Assessment Forum, Washington, DC. EPA/630/R-95/002F.
27 EPA (U.S. Environmental Protection Agency). 1999. Ecological Risk Assessment and Risk
28 Management Principles for Superfund Sites.
29 EPA (U.S. Environmental Protection Agency). 2001. Risk Assessment Guidance for Superfund
30 (RAGS), Volume 3, Part A -Process for Conducting Probabilistic Risk Assessment. EPA 540-R-
31 02-002. December 2001.
32 Harza. 2001. Woods Pond Dam: Structural Integrity Assessment. Prepared for General Electric
33 Company, Pittsfield, MA.
34 HEC (Harrington Engineering and Construction, Inc.). 1996. Report on the Preliminary
35 Investigation of Corrective Measures for Housatonic River and Silver Lake Sediment. Prepared
36 for General Electric Company.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-31
-------
1 MDEP (Massachusetts Department of Environmental Protection). 1996. Guidance for Disposal
2 Site Characterization in Support of the Massachusetts Contingency Plan. Chapter 9, Method 3,
3 "Environmental Risk Characterization," Interim Final Policy. MDEP, Bureau of Waste Site
4 Cleanup, and Office of Research and Standards. April 1996.
5 O'Brien & Gere Engineers, Inc. 1995. Phase I Report (MCP)ZCurrent Assessment Summary, Hill
6 78 Area/Area 2.
7 Stewart Laboratories, Inc. 1982. Housatonic River Study 1980 and 1982. Volumes I and II.
8 TechLaw (TechLaw, Inc.). 1998. Preliminary Report: Wetland Characterization and Function-
9 Value Assessment, Housatonic River from Newell Street to Woods Pond. 4 May 1998.
10 United States of America, State of Connecticut, and Commonwealth of Massachusetts, Plaintiffs
11 vs. General Electric Company, Defendant. 1999. Consent Decree-Main Document and
12 Appendices A through W. October 1999.
13 WESTON (Roy F. Weston, Inc.). 1998. Upper Reach-Housatonic River Ecological Risk
14 Assessment. Prepared for U.S. Environmental Protection Agency.
15 WESTON (Roy F. Weston, Inc.). 2000. Supplemental Investigation Work Plan for the Lower
16 Housatonic River. Prepared for U.S. Army Corps of Engineers and U.S. Environmental
17 Protection Agency. 22 February 2000. DCN GEP2-020900-AAME.
18 WESTON (Roy F. Weston, Inc.). 2001a. Final Quality Assurance Project Plan, Vol. I - Text,
19 Vol. II - Appendix A, Vol. Ha - Appendix A cont'd., Vol. IV - Appendix E & F. Prepared for
20 U.S Army Corps of Engineers and U.S. Environmental Protection Agency. DCN GE-021601-
21 AAHM.
22 WESTON (Roy F. Weston, Inc.). 2001b. Final Field Sampling Plan. Prepared for U.S. Army
23 Corps of Engineers and U.S. Environmental Protection Agency. DCN GE-053001-AAMA.
24 WESTON (Weston Solutions, Inc.). 2002. Rest of River Site Investigation Data Report. Prepared
25 for U.S. Army Corps of Engineers and U.S. Environmental Protection Agency. DCN GE-
26 080202-ABDK.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_1 .DOC
1-32
-------
l 2. PROBLEM FORMULATION
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16 2.1 OVERVIEW
17 Problem formulation, the planning phase of the ecological risk assessment (ERA), plays an
18 important role in the development and direction of the risk assessment. It builds on and refines
19 the screening-level problem formulation, and with input from stakeholders and other parties,
20 shapes the analysis of ecological issues of concern at a site (EPA 1997). Problem formulation
21 results in three products:
22 ¦ Conceptual model(s).
23 ¦ Assessment and measurement endpoints.
24 ¦ Analysis plan.
25
26 This section describes the process that was followed in developing and refining the problem
27 formulation phase of this ERA.
28 The screening-level problem formulation provides initial guidance to the risk assessors and
29 managers by providing a preliminary look at potential issues of concern (see Figure 1.5-1). The
30 screening-level problem formulation describes: (1) the environmental setting and contaminants
31 known or suspected at the site; (2) contaminant fate and transport mechanisms; (3) ecotoxicity
Problem Formulation Highlights
The problem formulation establishes the goals, scope, and focus of the baseline
ecological risk assessment (ERA). It is a process for generating and evaluating
preliminary hypotheses about why ecological effects have occurred, or may occur, as
a result of human activity. The problem formulation includes discussions of the
following topics:
¦ Identification and sources of stressors.
¦ Determination of contaminants of potential concern (COPCs).
¦ Fate and transport of contaminant stressors.
¦ Contaminant effects on receptors.
¦ Site conceptual model.
¦ Assessment and measurement endpoints.
¦ Weight-of-evidence (WOE) approach.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-1
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
mechanisms and receptor categories of concern; (4) exposure pathways from contaminant
sources to receptors; and (5) the results of a screening of conservative ecotoxicity values.
Subsequently, the problem formulation is expanded and refined as data collection and analysis
proceed. The ERA performed for the Upper Reach of the Housatonic River (WESTON 1998),
with the ERA Work Plan developed by GE during the previous RCRA process (ChemRisk
1997), together fulfill the functional requirements of the screening problem formulation.
A detailed ecological characterization of the site was conducted (and subsequently refined) to
expand on the information used in the screening-level problem formulation and to refine the
initial conceptual model for the site. The final ecological characterization is summarized in
Section 2.2 and presented in its entirety as Appendix A. The objective of the ecological
characterization was to characterize the ecosystems within the Housatonic River watershed,
including both plant and animal communities, with a focus on the Primary Study Area (PSA).
Table 2.1-1 summarizes the specific ecological characterization surveys performed to
characterize the ecosystems potentially at risk, as well as the specific survey objective(s), and
references to the SI Work Plan appendix containing the detailed standard operating procedure
(SOP) for each survey. This information was then used as input to the problem formulation.
In Section 2.3, the sources, concentrations, and distribution of contaminants in the study area are
discussed. Contaminants of potential concern (COPCs) identified in the screening-level problem
formulation were re-examined to determine whether they should be retained in the Pre-ERA
screen for the ERA (Section 2.4). The availability of new data, information, or changes in
assumptions can alter the results of the preliminary screening. Lack of data was not reason alone
to eliminate a potential contaminant, rather best professional judgment was used, and discussion
regarding the uncertainty surrounding the decision is presented.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-2
-------
Table 2.1-1
Surveys Conducted for Ecosystem Characterization and Their Specific
Objective(s)
Survey
Specific Objective(s)
SI Work Plan
Appendix
Rare Plants and Natural
Communities
Determine the potential rare, threatened, or
endangered plants or animals occurring within the
study area.
Determine the presence and areal extent of habitats
capable of supporting special status species
potentially occurring within the study area.
Determine the presence and areal extent of
exemplary natural communities within the study
area.
A.6
Dragonfly
Determine species of dragonflies present in the study
area, with particular attention to rare species.
A.7
Mussel
Determine historical and current distribution within
and upstream of the study area.
Identify potential mussel hosts.
Identify wildlife species that prey upon mussels.
A. 8
Reptile and Amphibian
Use
Estimate amphibian and reptile species richness in
the study area by habitat type.
Sample larval amphibians in breeding habitats over a
range of PCB concentrations.
Determine chemical concentrations in herptiles
incidentally succumbing during trapping.
Note: The latter two objectives were intended for use
in ERA exposure and effects characterization (see
Section 7.3 of the SI Work Plan).
A. 9
Raptors and Waterfowl
Identify raptors and waterfowl breeding in study
area.
A. 10
Forest Bird and Marsh
and Wading Bird
Identify birds using the study area floodplain forests
and scrub-shrub habitats.
Identify birds using the study area wetland and
aquatic habitats.
A. 11
River Otter, Mink, and
Bat
Determine if mink and otter are present in the study
area and reference areas.
Identify bat species present in the study area.
Determine habitats bats use for feeding.
A. 12
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
In the problem formulation, the fate and transport of contaminants in the ecosystem potentially at
risk and the description of exposure pathways were expanded beyond those in the screening-
level problem formulation. This was generally accomplished through the collection of data or
other information (e.g., field studies, modeling results, observations) on the fate and transport of
contaminants, the ecological setting and flora and fauna of the site, and the extent of
contamination (Section 2.5). In addition, the potential effects and impacts associated with
contaminants of potential concern (COPCs) are described in the context of site-specific
environmental conditions (Section 2.6).
The next step in the problem formulation was to establish the assessment endpoints for the study.
Assessment endpoints are an "explicit expression of the environmental value that is to be
protected" (EPA 1998). Specific assessment endpoints focus the ERA on the issues that are
important at the site and identify the appropriate measurement endpoints required to address
these endpoints. Potential adverse effects on local populations and communities, such as
reproduction, growth, and survival, or changes in community structure or function, respectively,
were identified and described using measurement endpoints to quantify effects for the
assessment endpoints (Section 2.8). The identification of assessment and measurement
endpoints and the exposure pathway analysis were used to refine the conceptual model for the
site (Section 2.7). The intent of the conceptual model was, through the iterative process
described above, to develop a thorough understanding and description of the site in a systematic
and representative manner, and to identify important data or information gaps.
The problem formulation culminates in a scientific/management decision point (SMDP). A
SMDP is an agreement between the risk manager and risk assessor on the assessment endpoints
selected, exposure pathways, and questions presented in the conceptual model (EPA 1997).
The initial problem formulation for this ERA focused on the area within the 10-year floodplain
of the Housatonic River extending from the confluence of the East and West Branches of the
river to and including Woods Pond (Figures 1.4-2 and 1.4-3). This area, referred to as the
Primary Study Area (PSA), is downstream from the source of COPCs from the GE facility, as
well as the area where cleanup activity (including river sediment, bank soil, upland soil, and
groundwater) is currently in progress, and includes the river sediment and floodplain soil where a
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-4
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
majority of the PCBs are located, as indicated by the historical data and the evaluation of the
recent EPA data, and summarized in the RCRA Facility Investigation Report (RFI) (BBL and
QEA 2003). The RFI states that most of the PCB mass in the Housatonic River and floodplain
downstream of the GE facility is in the PSA.
In addition, risks are also evaluated for the portion of the river below Woods Pond, MA, to the
point of tidal influence below the Derby-Shelton Dam, approximately 13 miles (21 km) from
Long Island Sound and 128 miles downstream from the PSA, but using a less quantitative
approach than that used for the PSA and for a focused set of endpoints (see Section 2.4, and
Appendix A).
2.2 PHYSICAL AND ECOLOGICAL CHARACTERIZATION OF THE HOUSATONIC
RIVER
2.2.1 Physical Characteristics of the Housatonic River Basin
The Housatonic River is located in Berkshire County, MA, and western Litchfield, eastern
Fairfield, and western New Haven Counties, CT. The river flows approximately 166 miles (240
km) from the headwaters above Dalton, MA, to Long Island Sound, and drains an area of
approximately 1,950 square miles in Massachusetts, New York, and Connecticut (BBL and QEA
2003).
For much of its path through Berkshire County, MA, the river lies in a wide alluvial plain called
the Central Valley (Weatherbee 1996). The Central Valley is bounded to the east by the
Berkshire Plateau, a southern extension of Vermont's Green Mountains, and to the west by the
Taconic Range, extending from Vermont to New York. In Connecticut, this same alluvial valley
is called the Marble Valley. East of the valley, the Berkshire Plateau from Massachusetts
continues southward and is called the Litchfield Hills Plateau.
In general, the plateaus and mountains bounding the river valley are typified by rounded hills and
mountains draped with glacial deposits, and relatively narrow, steep-sided valleys cut into the
hills by streams and rivers. The principal bedrock underlying much of the river basin is marble
formed during the Devonian period, approximately 350 to 400 million years ago. Because of the
prevalence of marble, the Housatonic River basin exhibits characteristics different from most
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs c 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
other river systems in the northeastern United States. In particular, soil and water pH in the
valley are high (7.9 to 8.3) and the groundwater contains high concentrations of calcium and
magnesium (Harris 1997; Olcott 1995).
The area has a continental climate, similar to the rest of interior New England, characterized by
cold winters and hot summers. In Stockbridge, MA, near the northern end of the study area,
average annual temperature was 8 °C, and average daily July and January temperatures were 20
and -6 °C, respectively, for the period between 1951 and 1974. At Cornwall, CT, at
approximately the midpoint of the watershed, average annual temperature was 9 °C, and average
daily July and January temperatures were 21 and -4 °C, respectively. At Danbury, CT, nearer the
southern end of the study area, the average annual temperature was 10 °C, and average daily July
and January temperatures were 22 and -3 °C (SCS 1970, 1981, 1988). The number of frost-free
days (growing season) at those locations ranges from 103 to 183 days. Moisture supply usually
exceeds evaporation, except during periods of drought. Average total rainfall is 43 inches (110
cm) in Berkshire County, increases slightly southward to 45 inches (114 cm) in Litchfield
County, and 47 inches (119 cm) in Fairfield and New Haven Counties (SCS 1970, 1979, 1981,
1988), and is evenly distributed throughout the year. Conversely, average total snowfall for
these counties decreases markedly north to south and is 71, 61, 39, and 32 inches (180, 155, 99,
and 81 cm), respectively (SCS 1970, 1979, 1981, 1988).
2.2.2 Ecological Characterization of the Study Area
For the purposes of the EPA Supplemental Investigation, the Housatonic River was divided into
17 reaches from the headwaters to Long Island Sound, with some reaches further subdivided.
Reaches 1 to 17 were the focus of earlier ecological characterization studies (Chadwick &
Associates, Inc. 1994). Reaches 5 and 6, comprising the PSA, were further investigated in detail
by EPA from 1998 to 2000 (see Figure 1.1-2). As a result of that work, an ecological
characterization of the PSA was prepared (see Appendix A.l, Ecological Characterization of the
Housatonic River). Reaches 7 to 17 were also characterized, using aerial photograph
interpretation and data provided by regional references and state natural resource agencies (see
Appendix A.2, Ecological Characterization of the Housatonic River Downstream of Woods
Pond).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs ^ 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
2.2.2.1 Primary Study Area (PSA) Characteristics
Much of the PSA (approximately 770 acres) consists of state lands. Portions of the Housatonic
River Valley State Wildlife Management Area, which totals 818 acres including land ranging
from the confluence of the East and West Branches of the Housatonic River to Woods Pond
(Mass Wildlife 2002), fall within the PSA. Approximately two-thirds of the State Wildlife
Management Area is a continuous parcel from just north of New Lenox Road south to Woods
Pond. Additional large parcels occur near the confluence (approximately 80 acres) and north of
the Pittsfield wastewater treatment plant (WWTP) (approximately 120 acres). This area includes
most of the forested habitat within the PSA. October Mountain State Forest, which comprises
approximately 15,940 acres, occurs immediately adjacent to the eastern side of the lower PSA.
This large area consists mainly of mature hardwood, softwood, and mixed forests. The City of
Pittsfield owns a 45-acre parcel of land associated with the WWTP. Much of the land associated
with the WWTP has been developed and includes buildings, paved parking areas, access roads,
and maintained lawns. The remaining WWTP land near the river consists of transitional forests,
shrub swamps, and shallow emergent marsh. Canoe Meadows, a Massachusetts Audubon
Society property, is located just below Holmes Road. This area contains forests and fields, as
well as a large wetland complex.
A total of 18 natural communities occur within the PSA: 1 lacustrine community; 10 palustrine
communities primarily associated with the Housatonic River floodplain and shoreline; 3 riverine
communities either within the channel itself or draining into it; and 4 upland communities
included within the 10-year floodplain1 (Appendix A).
Portions of the PSA have been cleared for various purposes, primarily agriculture, residences,
and various rights-of-way (e.g., roads, railroads, power lines). Agricultural development was the
primary source of forest clearing within the floodplain. Several large wet meadows can be found
in the PSA in which the species composition is influenced by past farming practices. Shrub
swamps are common along pool and river channel borders, but are especially frequent as an
1 Natural communities have been identified and classified according to Swain and Kearsley (2000). Weatherbee
(1996) and Weatherbee and Crow (1992) were used to classify some river and lake systems.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-7
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
intermediate successional stage in areas where pasture is reverting to forested floodplain. Even
within some transitional floodplain forests, it is clear from the subcanopy species present that
these areas were farmed in the past century. For example, dotted hawthorn routinely colonizes
regenerating pastureland, and survives in the subcanopy of floodplain forests for some time after
the tree stratum has returned to the site.
Significant portions of the PSA are open wetlands and riverine systems dominated by submersed,
floating-leaved, and emergent herbaceous vegetation. Riverine point bars and beaches occur
occasionally along the river, primarily near bends in the river channel. Mud flats of limited size
begin to appear later in the season as the water level declines and exposes previously inundated
sediment. Deep emergent marshes, which are usually inundated through the season and
vegetated by robust herbs, are frequent along the river channel and backwater edges (Figure
2.2-1). These areas become much more abundant south of New Lenox Road, where backwater
sloughs, old oxbows, and cut-off channels are common due to the influence of Woods Pond.
Shallow emergent marshes, which are areas with saturated soil or shallow water and lower herbs,
are less common in the study area and most frequently observed within more permanent vernal
pools.
2.2.2.2 Housatonic River Below Woods Pond Dam
The Housatonic River below Woods Pond Dam in Lenoxdale, MA, extends downstream to Long
Island Sound in Connecticut, encompassing Reaches 7 through 17 (see Figures 1.4-5 to 1.4-7).
Reach 8 comprises the next significant impoundment below Woods Pond, formed by Rising
Pond Dam. Reach 17 is the tidal portion of the river downstream of the Derby-Shelton Dam and
was not included in the Rest of River investigations due to its tidal nature and a number of other
sources of COPCs. The river valley in Connecticut becomes narrower with steep uplands
flanking both sides, and the free-flowing reaches of the river flow over a harder, coarser
substrate of limestone, quartz, and granite (HVA 2001). Because of the constricting valley walls,
the floodplain becomes narrower than in much of Massachusetts. However, some localized areas
of broader floodplain exist. In these areas, agricultural activities are the dominant land use.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-8
-------
Northern Black Ash-Red Maple-
Hardwood- Tamarack Calcareous
Hemlock- Seepage Swamp
White Pine
Forest
Powerline Springfield
R.O.W. Terminal
Railroad
Black Ash-Red Maple-Tamarack
Calcareous Seepage Swamp
Deep Emergent
Marsh
I miit til
10 mi
HuoJp(itin
hlitcfc iish, red maple, bur oak
Transitional Low Gradient Red Maple
Floodplain Stream Swamp
Forest
Northern
Hardwood-
Hemlock-
White Pine
Forest
Housatonic Rivet-
Ecological Characterization
Representative Section of Primary Study Area
Lower Section
(Reach 5C south of Yokuro Brook confluence)
Not to scale September 2002
Figure 2.2-1
East —*¦
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-9
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
The Housatonic River in Connecticut is affected by six dams, all of which create impoundments;
five of the dams are used for electric power generation and all the impoundments are used for
recreational purposes. These impoundments are medium to large, deep reservoirs of lacustrine
habitat. Between these reservoirs, the free-flowing river is characterized as a medium-gradient
stream with moderate to fast currents and pool, riffle, and run habitats.
A total of 28 natural communities occur in the Lower Housatonic River study area. Aquatic
communities include moderately alkaline lakes and ponds in impounded reaches and low-,
medium-, and high-gradient stream communities in free-flowing riverine areas. Palustrine
communities include deep emergent marshes, shallow emergent marshes, wet meadows, mud
flats, riverside seeps, calcareous sloping fens, shrub swamps, red maple swamps, black ash-red
maple-tamarack calcareous seepage swamps, transitional floodplain forests, and high-terrace
floodplain forests. Within the terrestrial systems, there are riverine point bars and beaches, high-
energy riverbanks, riverside rock outcrops, calcareous rock cliff communities, northern
hardwoods-hemlock-white pine forests, red oak-sugar maple transition forests, spruce-fir-
northern hardwood forests, successional northern hardwoods, rich mesic forests, and cultural
grasslands. Developed land uses include agricultural, residential, commercial, and public
development, along with transportation.
2.3 IDENTIFICATION AND SOURCES OF STRESSORS
2.3.1 Contaminant Stressors
In this section, the sources, concentrations, and distribution of contaminants in the study area are
identified. More detailed discussions of these topics, including information regarding the
amounts, form, and conditions of release, are presented in the Modeling Framework Design
(MFD) (WESTON 2003, in preparation), the Supplemental Investigation Work Plan for the
Lower Housatonic River (WESTON 2000), and the GE RCRA Facility Investigation Report
(BBL and QEA 2003).
The GE facility in Pittsfield was the major handler of PCBs in western Massachusetts, and is the
only known point source of PCBs in the PSA and downstream to the Derby-Shelton Dam
(approximately 13 miles from Long Island Sound). According to previous GE reports, from
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs -i 7/10/2003
-------
1 1932 through 1977, releases of PCBs reached the wastewater and storm systems associated with
2 the facility and were subsequently conveyed to the East Branch of the Housatonic River and to
3 Silver Lake, or were released directly to these waters. In addition to the Housatonic River and
4 Silver Lake, areas of the 254-acre GE facility, filled former river oxbows, neighboring
5 commercial properties, Allendale School, and other properties or areas have become
6 contaminated as a result of the GE operations.
7 Based on historical data and facility operations, the contaminants listed below have been found
8 in the source areas and may have migrated to the Housatonic River:
9 ¦ PCBs.
10 ¦ Dioxins/furans.
11 ¦ Semivolatile organics (e.g., bis(2-ethylhexyl)phthalate, methylphenol, phenol, and
12 polycyclic aromatic hydrocarbon [PAHs]).
13 ¦ Volatile organics (e.g., acetone, benzene, chlorobenzene, tetrachloroethene,
14 trichloroethene, toluene, xylene, and other chlorinated hydrocarbons).
15 ¦ Inorganics (e.g., lead and zinc).
16 According to the Source Area Characterization Report, there are five general categories of
17 contaminant sources potentially impacting the river (WESTON 1998):
18 ¦ Nonaqueous phase liquid (NAPL) discharge.
19 ¦ Contaminated groundwater discharge.
20 ¦ Riverbank soil/river sediment transport.
21 ¦ Desorption/adsorption of residual riverbank and sediment contaminants.
22 ¦ Direct stormwater discharge and surface runoff to the river.
23
24 The major areas of contamination designated in the Consent Decree for purposes of investigation
25 and response are shown in Figure 1.2-1.
26 2.3.2 Physical and Biological Stressors
27 In addition to contaminant stressors (e.g., PCBs), physical and biological stressors can alter
28 processes within ecosystems, affect habitat types, and ultimately influence natural communities
29 by changing the diversity and abundance of species within habitat types. Physical stressors
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-11
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
include structures and events such as dams, ice scour, floods, and droughts; biological stressors
consist of changes in the biological components of a community, such as invasive plants out-
competing native plants within riparian areas. Examples of physical and biological stressors that
occur within the PSA and their subsequent effects on the natural communities are presented in
the following paragraphs.
As mentioned in previous sections, the Housatonic River is a low-gradient river with a complex
and diverse array of aquatic, riparian, and terrestrial habitats. These habitat types are largely
determined by local conditions such as geology, climate, and soil, but are also influenced by a
broad network of watershed processes such as hydrology and sediment transport. These
processes can alter habitats by changing river morphology (e.g., eroding banks and creating
pools) or by resetting high floodplain forest succession (e.g., uprooting overstory trees during a
windstorm). Thus, the variety of habitat types within the PSA varies both spatially and
temporally.
One of the natural watershed processes of the Housatonic River is for the river to meander
laterally within its valley by eroding riverbanks on the outside of bends and depositing sediment
on the inside of bends to create point bars. This typically occurs during or after high-flow events
and can affect various animals within localized natural communities. For example, high flows
undermine riverbanks and cause bank collapse, which may result in nest failures of bank-nesting
birds such as the belted kingfisher (Ceryle alcyori). High flows also can flood out animals that
den in riverbanks, such as muskrats (Ondatra zibethicus), causing mortality or forced relocation
of burrows, or flood nests in the floodplain for species such as waterfowl.
Floods also increase river velocities and shear stresses that can cause the riverbed to scour,
transport sediment, and then, as high flows subside, deposit sediment downstream and onto the
floodplain. Movement of bed sediment can affect aquatic organisms including benthic
macroinvertebrates and macrophytes that depend on specific substrate types. In addition,
changes in macroinvertebrate communities can stress localized populations of other aquatic
organisms that depend on these animals as food.
High winds, such as those that occur during hurricanes or nor'easters, are another physical
stressor that occurs along the Housatonic River. These winds can cause blowdown of localized
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs -i rs 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
areas of the overstory floodplain vegetation (e.g., silver maple [Acer saccharinum] and black
willow [Salix nigra}), which then resets forest succession to pioneer herbaceous species such as
goldenrod (Solidago sp.) that can tolerate increases in light and decreases in soil moisture.
Blown-down trees and those undermined by bank failure fall into the river and create complex
habitat (large woody debris) for a host of aquatic organisms. Large woody debris alters localized
channel processes and influences the development of natural communities. When a tree falls
into a riffle, it can cause local bed scour and pool formation that then provides areas of refuge
during high flows, hiding cover, rearing areas, and food sources for specific aquatic organisms.
For example, pools typically have slower velocities and deeper water depths that are used by fish
species such as northern pike (J is ox lucius), largemouth bass (Micropterus salmoides), and
common carp (Cyprinus carpio). Riffles and runs provide relatively faster moving water for
different fish species (e.g., longnose dace \Rhinichthys cataractae]).
Biological stressors within the PSA influence natural communities by changing the distribution
of species, which can then affect other components of the food chain. Such stressors include
insect infestations, diseases and pathogens, and exotic or invasive species. Invasive plant species
are common in the PSA and include species such as Asiatic bittersweet (Celastrus orbiculata),
garlic mustard (Alliaria petiolata), common buckthorn (Rhamnus cathartica), Japanese
knotweed (Polygonum cuspidatum), and purple loosestrife (Lythrum salicaria). These plants can
invade natural communities and out-compete native plants, create localized monocultures,
provide prolific seed source areas, and reduce species diversity. Such invasions can stress
species that depend on specific native plants (e.g., some butterflies require specific plants for
food).
Both physical and biological stressors within the PSA are influenced by anthropomorphic
changes that have occurred within the watershed. These changes include channelization,
riverbank armoring, dams, urban runoff, riparian area management, introduction of non-native
and exotic species, invasive plants, business and residential development, watershed restoration
projects, stormflow routing, bridge and railroad construction, wastewater transport and treatment
facilities, agricultural clearing and ditching, and power lines.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC 2 13
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
2.4 OVERVIEW OF PRE-ERA
2.4.1 Introduction
The purpose of the Pre-ERA was to narrow the scope of the ERA by refining the list of
contaminants to only those that pose potential risks to biota. The primary objectives of the Pre-
ERA are as follows:
1. Identify COPCs other than PCBs for the PSA. (Downstream of the PSA, numerous
potential sources of COPCs, other than the GE facility, exist along the river.)
2. Determine the downstream boundary beyond which PCBs from the GE facility pose a
negligible risk to aquatic biota and wildlife.
The following discussion provides a brief overview of the steps taken to identify COPCs for the
PSA and to determine the downstream extent of the ERA. The COPCs from the Pre-ERA are
further refined in the discussion of each individual assessment endpoint, resulting in endpoint-
specific COCs, as appropriate. A more detailed presentation of the Pre-ERA approach and
results is provided in Appendix B.
2.4.2 Data
Data sets were developed for the primary media of concern for the PSA, background areas, and
for the area below Woods Pond.
2.4.2.1 Primary Study Area
Data in the PSA were grouped by media (i.e., sediment, surface water, soil, and fish tissue),
subreach, and geomorphological type. The subreaches used for this evaluation were
hydraulically similar sections of the Housatonic River, identified by the project modeling team in
the Modeling Framework Design (WESTON 2003, in preparation). Geomorphological terrain
descriptions (geomorph codes) were assigned to sediment, soil, and surface water samples
collected by EPA. Each geomorph code represents a depositional or erosional feature or a terrain
type that was formed by a specific geologic process (e.g., main channel, vernal pools, and side
channels). The sediment and water data categories used for the Pre-ERA are shown in Table
2.4-1.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs -i a 7/10/2003
-------
1 Floodplain and riverbank soils were evaluated separately for Reaches 5A, 5B, 5C, 6A, and 6B.
2 Soil adjacent to Woods Pond (referred to as Reaches 6C and 6D in the Pre-ERA) was also
3 evaluated. A more detailed description of the reach designations is provided in Section 1.4.
4 Fish tissue samples were grouped based on reach (5 A, 5B, 5C, 6A, and 6B) and size class. Three
5 size classes were evaluated, small (< 3 inches [7.6 cm]), medium (> 3 inches [7.6 cm] but < 12
6 inches [30.5 cm]), and large (> 12 inches [30.5 cm]).
7 2.4.2.2 Background Data
8 Background data are media-specific (i.e., sediment, surface water, soil, and fish tissue) chemistry
9 data collected within the Housatonic River watershed that were not believed to be influenced by
10 contamination directly originating from the GE Pittsfield facility. The objective of the
11 determination of background concentrations was to identify what the media-specific chemical
12 concentrations would be in the absence of releases from the GE facility, and to use this
13 information in evaluating COPCs for the ERA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-15
-------
1
2
3
Table 2.4-1
Sediment and Surface Water Data Categories
Medium/Reach
Geomorphological Terrain Type
Main Channel/
Aggrading Bars
Side Channels and
Oxbows (SCOX)
Vernal Pools
Pond
Sediment
5A
V
V
V
5B
V
V
V
5C
V
V
V
6 A, 6B
V
V
V
V
6C, 6D
V
Surface Water
5A
V
V
V
5B
V
V
5C
V
V
V
6C, 6D
V
4
5 2.4.3 Primary Study Area (PSA) Evaluation and Results
6 The procedures used to screen potential COPCs were applied to the data groupings summarized
7 above. Three progressive evaluation tiers were used to determine COPCs for the PSA.
8 ¦ Tier I - A three-step process was used to establish the initial COPC list evaluating:
9 - Frequency of detection.
10 - Exceedance of benchmarks.
11 - Comparisons to background concentrations.
12 ¦ Tier II - A more detailed evaluation of frequency of exceedance of benchmarks was
13 performed for contaminants that were not eliminated from further consideration in the
14 Tier I evaluation.
15 ¦ Tier III - The spatial extent of contamination, magnitude of benchmark exceedance,
16 presence in other media, and mechanism of toxicity were evaluated subjectively for
17 contaminants not removed in Tier I or Tier II.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-16
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Using the three-tier approach, a final list of COPCs was developed for each medium within a
reach/terrain. The final generic COPC lists are presented in Tables 2.4-2 through 2.4-5. A
detailed description of this approach is provided in Appendix B (Pre-ERA). Although several
pesticides were retained as COPCs from the Pre-ERA process, a subsequent review of pesticide
concentrations indicated, in general, relatively few detects and low concentrations. Therefore, it
is believed that pesticides are generally not a site-related COPC. The COPCs were then further
evaluated for each assessment endpoint; this is discussed in detail in the specific assessment
endpoint appendices.
2.4.4 PCB Screening Evaluation Downstream of Woods Pond and Results
To determine the downstream limit for the ERA and potential areas of concern, the PCB
concentrations measured in sediment at locations downstream of Woods Pond to Derby-Shelton
Dam were compared with available benchmarks. PCB concentrations in sediment (rather than
another medium) were selected as an indicator of the spatial extent of potential ecological risk,
because:
¦ Sediment serves as a reservoir of PCBs released from the GE facility.
¦ Sediment concentrations generally reflect the relative concentrations that could be
expected in the floodplain.
¦ Exposure to PCBs in sediment is a major route of exposure for lower trophic levels
(and subsequently higher trophic levels) in the aquatic food chain.
¦ Relatively extensive data on PCB concentrations in sediment are available.
The conservative benchmark used for this analysis was a threshold effect concentration (TEC) of
0.0598 mg PCB/kg sediment (MacDonald et al. 2000).
Hazard quotients (HQs) were developed using detected PCB concentrations or sample
quantitation limits (SQLs), and the MacDonald TEC benchmark (0.0598 mg PCB/kg). After
evaluation of the magnitude by which the benchmark was exceeded and the consistency and
frequency of exceedances, the reaches from Woods Pond Dam to Derby-Shelton Dam were
retained for quantitative evaluation of risk from exposure to PCBs in the ERA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-17
-------
Table 2.4-2
COPCs for Sediment Based on Tier III Evaluation
Chemical
Reach/Geomorphological Type
5A
5B
5C
6AB
MC&
AB
scox
VP
MC&
AB
scox
VP
MC & AB
scox
VP
Pond
Semivolatiles
Dibenzofuran
X
...
...
...
...
...
...
...
...
...
PAHs
Acenaphthene
X
...
...
...
...
...
X
...
...
...
Acenaphthylene
X
...
...
...
...
...
X
...
...
...
Anthracene
X
...
...
...
...
...
X
...
...
...
Benzo(a)anthracene
X
...
X
X
X
X
X
...
X
...
Benzo(b)fluoranthene
X
X
X
X
X
X
X
...
X
X
Benzo(k)fluoranthene
X
...
X
...
...
...
X
...
X
...
Benzo(g,h,i)perylene
X
X
X
X
X
X
X
...
X
X
Benzo(a)pyrene
X
...
X
X
X
X
X
...
X
...
Chrysene
X
...
X
...
X
X
X
...
X
...
Dibenzo(a,h)anthracene
X
X
X
...
...
X
X
...
X
...
Fluoranthene
X
...
...
...
...
...
X
...
X
...
Fluorene
X
...
...
...
...
...
X
...
...
...
Indeno( 1,2,3 -cd)pyrene
X
X
X
X
X
X
X
...
X
X
Naphthalene
X
...
...
...
...
...
X
...
...
...
Phenanthrene
X
...
...
X
X
...
X
...
X
...
Pyrene
X
...
...
X
X
X
X
...
X
...
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC 2 18 7/10/03
-------
Table 2.4-2
COPCs for Sediment Based on Tier III Evaluation
(Continued)
Chemical
Reach/Geomorphological Type
5A
5B
5C
6AB
MC &
AB
SCOX
VP
MC &
AB
SCOX
VP
MC & AB
SCOX
VP
Pond
Dioxins/Furans
X
X
X
X
X
X
X
X
X
X
PCBs
X
X
X
X
X
X
X
X
X
X
Metals
Antimony
...
...
...
...
...
...
X
...
X
...
Barium
...
...
...
...
...
...
X
...
X
...
Beryllium
...
...
...
...
...
...
...
X
X
...
Cadmium
...
...
...
...
...
...
X
...
X
...
Chromium
...
...
...
...
...
...
X
...
X
X
Copper
...
...
X
...
...
X
X
...
X
X
Lead
...
...
X
...
...
X
X
...
X
X
Mercury
...
...
X
...
X
X
X
X
X
X
Selenium
...
...
...
...
...
...
...
...
X
X
Silver
...
...
X
...
X
X
X
...
X
X
Tin
...
...
...
...
...
X
X
...
X
X
1 MC - main channel
2 AB - aggrading bars
3 SCOX - side channels and oxbows
4 VP - vernal pools
5 Pond - Woods Pond
6 Note: Reach designations reflect previous reach boundaries; it is assumed that the revised reach designations do not impact reach-specific COPCs determined in
7 Appendix B.
MK01|O:\20123001.096\ERA_PB\ERA_PB_2.DOC ~ iq 7/10/03
-------
1 Table 2.4-3
2
3 COPCs for Surface Water Based on Tier III Evaluation
Chemical
Reach/Geomorphological Type
5A
5B
5C
6AB
MC & AB
VP
MC & AB
VP
MC& AB
VP
Pond
Dioxins/Furans
X
X
X
X
X
X
X
PCBs
X
X
X
X
X
X
X
4 MC - main channel
5 AB - aggrading bars
6 VP - vernal pools
7 Pond - Woods Pond
8 Note: Reach designations reflect previous reach boundaries; it is assumed that the revised reach designations do not impact reach-specific COPCs
9 determined in Appendix B.
10
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-20
-------
1
2
3
Table 2.4-4
COPCs for Soil Based on Tier III Evaluation
Chemical
Reach/Geomorphological Type
5A
5B
5C
6AB
Floodplain
Riverbank
Floodplain
Riverbank
Floodplain
Riverbank
Pond
Semivolatiles
Dibenzofuran
X
X
X
X
...
...
...
PAHs
Benzo(a)pyrene
...
X
...
...
...
...
...
Pyrene
...
X
...
...
...
...
...
Dioxins/Furans
X
X
X
X
X
X
X
PCBs
X
X
X
X
X
X
X
Metals
Chromium
X
X
X
X
X
X
X
Lead
...
...
...
...
X
...
X
Mercury
...
X
X
X
X
X
X
Selenium
...
...
...
...
X
...
X
4 Note: Reach designations reflect previous reach boundaries; it is assumed that the revised reach designations do not impact reach-specific COPCs determined in
5 Appendix B.
6
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-21
-------
1
2
3
Table 2.4-5
COPCs for Fish Based on Tier III Evaluation
Chemical
Reach/Fish Size
5A
5BC
6AB
Small
Medium
Large
Small
Medium
Large
Small
Medium
Large
Pesticides
4,4'-DDE
...
...
...
...
...
X
...
X
X
0,p'-DDT
X
X
X
X
X
X
X
X
X
4,4'-DDT
...
X
...
...
X
X
...
...
X
Heptachlor epoxide
X
X
X
...
...
X
X
X
X
Cis-Nonachlor
X
X
X
X
X
X
X
X
X
Trans-nonachlor
X
X
X
X
X
X
X
X
X
Oxychlordane
X
X
X
X
X
X
...
X
X
Dioxins/Furans
X
X
X
X
X
X
X
X
X
PCBs
X
X
X
X
X
X
X
X
X
4 Note: Reach designations reflect previous reach boundaries; it is assumed that the revised reach designations do not impact reach-specific COPCs determined
5 in Appendix B.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC 9 99 7/10/03
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
2.5
FATE AND TRANSPORT OF CONTAMINANT STRESSORS
Understanding the fate and transport characteristics of COPCs is a major component of the
problem formulation phase of an ERA. Although other COPCs are present in the study area, the
focus of the Rest of River evaluation is on PCBs, as well as dioxins/furans; therefore, the
objectives of this discussion are to:
¦ Provide a general description of PCB fate and transport mechanisms (Section 2.5.1).
¦ Present a summary of PCB distribution within the Housatonic River floodplain
(Section 2.5.2).
¦ Identify exposure pathways (Section 2.5.3).
¦ Discuss how PCB congener patterns change in environmental media (Section 2.5.4).
¦ Present a general overview of the fate and transport mechanisms of dioxins/furans
(Section 2.5.5).
2.5.1 Fate and Transport of PCBs
Polychlorinated biphenyls (PCBs) are formed when hydrogen atoms on a biphenyl molecule are
replaced by 1 to 10 chlorine atoms. First manufactured approximately 75 years ago, PCBs are
extremely persistent contaminants that are now ubiquitous in the global ecosystem (Eisler 1986).
There are 209 possible configurations of PCB molecules, based on the number and position of
chlorine substitutions on the biphenyl ring; these individual PCB configurations are known as
congeners. Although all possible congeners have been synthesized, only approximately 175 of
the 209 congeners were included in the various commercial formulations. Groups of PCB
congeners with similar numbers of substituted chlorine atoms are referred to as homologs,
including: mono-, di-, tri-, tetra-, penta-, hexa-, hepta-, octa-, nona-, and decachlorobiphenyl
(EPA 1996). Aroclors (Aroclor is a trade name of the Monsanto Company) are commercial
mixtures of PCB congeners that were formulated to have specific physical properties, which are
based, in general, on the overall amount of chlorine substitution (Figure 2.5-1).
The level of chlorination affects various physicochemical properties of the PCB molecule, such
as the octanol/water partition coefficient (Kow), solubility, vapor pressure, and Henry's Law
constant. These properties affect processes such as volatilization, and partitioning to water,
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC O O ^ 7/10/03
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
sediment, and floodplain soil. Similarly, the level of chlorination also controls (in part)
biologically mediated processes such as biotransformation, uptake, and bioaccumulation
(WESTON 2003, in preparation). In general, more chlorinated PCBs have greater stability and
environmental persistence (EPA 1996).
PCBs in the environment occur as mixtures of congeners that differ in composition from
commercial mixtures because of partitioning, contaminant transformation, and preferential
bioaccumulation over time (Aulerich et al. 1986; Hornshaw et al. 1983; EPA 1980). Some
congeners are retained in sediment, soil, and biological tissue. Bioaccumulated PCBs appear to
be more toxic than commercial PCBs (Aulerich et al. 1986; Hornshaw et al. 1983).
More detailed discussions on the fate and transport of PCBs can be found in the Modeling
Framework Design (WESTON 2003, in preparation).
meta ortho ortho meta
para (W A j 1> B j 4)Para
rnefa ortho ortho meta
biphenyl
monochlorobiphenyl
CI
CI/
2,2',4,4'-tetrachlorobiphenyl
Source: Adapted from Eisler 1986.
Figure 2.5-1 Biphenyl and Representative PCB Congeners
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-24
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
2.5.2 PCB Distribution by Media
This section provides an overview of the distribution of PCBs in sediment, soil, surface water,
and biota of the PSA. This section also presents a discussion of the sediment grain size analysis
and the concentrations of organic carbon in the sediment, soil, and water samples from the PSA
and their relationship with PCBs. Sediment and soil data used for this analysis included all
samples collected by any organization between 1998 and 2002, a span of 5 years. Earlier data
were not included to ensure that any potential temporal trends or the influence of different
analytical methods in the data would not potentially mask current spatial patterns. The analysis
of surface water included samples collected between 1996 and 2002. This slightly longer span of
time was used because of the more robust data set that was available. A full presentation of the
spatial and temporal trends is presented in the MFD (WESTON 2003, in preparation) and the
RFI report (BBL and QEA 2003).
The term sediment is defined for this study as solid material typically inundated under normal
hydrologic conditions. Soil samples are defined as those samples collected from areas not
typically inundated under normal hydrologic conditions. Sediment and soil samples were
collected from across the PSA and classified by the geomorphic terrains (i.e., main river channel,
floodplain, riverbanks, etc.) from which they were originally collected. The distribution of PCBs
in the PSA and in Reaches 7, 8, and 9 between the PSA and the Massachusetts/Connecticut state
line is illustrated in the stack bar figures (see Attachment 2.1).
2.5.2.1 Sediment
PCBs have been detected in sediment samples collected from all reaches of the Housatonic River
from just upstream of the GE facility through the PSA and downstream in Massachusetts and in
Connecticut. Figure 2.5-2 presents sediment PCB data for the entire Housatonic River (from the
vicinity of the GE facility to the point where the river empties into Long Island Sound).
Historically, over 7,500 sediment samples have been collected from the main channel of the river
in Massachusetts and Connecticut; almost 5,000 samples since 1998 alone. The highest
concentrations of PCBs have been detected in sediment adjacent to the GE facility (river mile
137; 9,411 mg/kg in a surficial sample) and continuing downstream to Woods Pond Dam (at
river mile 124.37). Within the PSA, the highest PCB concentrations detected by EPA were 614
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC O O £ 7/10/03
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
mg/kg in Reach 5A, 165 mg/kg in Reach 5B, 213 mg/kg in Reach 5C, and 668 mg/kg in Reach 6
(Woods Pond). Sediment samples collected prior to 1998 detected PCBs as high as 2,270 mg/kg
in Reach 5A. PCBs have also been detected as deep as 6 to 8 feet below the riverbed surface
throughout Reaches 5 and 6 (BBL and QEA 2003; WESTON 2003, in preparation).
PCB concentrations in sediment decrease downstream of Woods Pond Dam in Reaches 7, 8, and
9, and decrease further in concentration through most of Connecticut (BBL and QEA 2003). An
increase in PCBs was detected in the most downstream reach (Reach 17 - from the Derby-
Shelton Dam to Long Island Sound), attributable to other Superfund or designated hazardous
waste sites located within that portion of the river.
~
Massachusett
~
s Connecticut
. t
WP
RPD
G
F C
B BE
SD BE
) S
) s
t 1
DSD
f 1
ii«
4
i i
1
-r-#—
—+T-
400
350
300
250
O
O
200 ^
flj
150 o
o
o
100
50
0
150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0
River Miles from Long Island Sound
Notes:
1. All data are plotted at the approximate mid-point of each reach, and represent samples collected from within
the top 3 feet of the riverbed.
2. Total PCB concentrations above 400 mg/kg were not plotted.
3. Symbols represent significant features/names of reach boundaries: GE = General Electric facility; WPD =
Woods Pond Dam; RPD = Rising Pond Dam; GFD = Great Falls Dam; CB = Cornwall Bridge; BBD = Bulls
Bridge Dam; BD = Bleachery Dam; SD = Shepaug Dam; STD = Stevenson Dam; DSD = Derby-Shelton
Dam.
Figure 2.5-2 Distribution of tPCB Concentrations Detected in Sediment Samples
from the GE Facility to Long Island Sound
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-26
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
The mean tPCB concentrations in sediment samples are plotted by reach in Figure 2.5-3. The
data have been presented on a log scale to capture the mean tPCB concentration of 393 mg/kg in
Reach 3 (adjacent to the GE facility). For the purposes of calculating statistics, values for non-
detects were treated as half the reported detection limit. Likewise, the most commonly reported
detection limit of 0.5 mg/kg is shown on the figure for comparison.
1000
100
10
| ~Mean Total PCB Concentration
T I = Two standard errors of the
mean.
<
i
z
i 1
X , 1
Typical Detection Limit <
- 4 *
~
~
(-~H
1—~-)
1 •-
;; <
~
, T T •
O
CO
O
Q_
0.1
0.01
1/2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
Reach
Figure 2.5-3 Mean Total Sediment PCB Concentrations by Reach
As shown in Figure 2.5-2, the mean tPCB concentrations are highest adjacent to the GE facility
and on through the PSA to Woods Pond Dam (Reaches 3 to 6) and then generally decrease
through the remaining reaches in Massachusetts and Connecticut. Many samples from
Connecticut were non-detect, resulting in low (<0.5 mg/kg) mean concentrations of PCBs.
Reach 11, and approximately half of the length of Reach 10, is shallow and fast-flowing with
mostly a gravel to cobble stream bed where PCB-containing solids are not likely to be deposited,
resulting in very few samples having PCBs. Reach 16 represents the last impoundment along the
Housatonic River, and approximately 99% of the samples collected there were non-detect.
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-27
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
2.5.2.2
Soil
Soil samples were collected from the floodplain and riverbanks along the Housatonic River
within Massachusetts) (approximately 4,300 samples were collected by EPA, and 3,300 samples
were collected by GE) in the reaches upstream of Woods Pond Dam. Additional samples
(approximately 1,600 collected by EPA, and 200 collected by GE) were collected below the
PSA, and PCBs have been detected in floodplain soil in all reaches of the Housatonic River from
the GE facility to the Massachusetts/Connecticut state line. The highest concentration of tPCBs
detected in floodplain soil was 907 mg/kg from soil in Reach 5C above Woods Pond.
Conversely, the highest tPCBs detected in riverbank soil were adjacent to the GE facility and just
downstream, in Reaches 2 through 4 (between river miles 138 and 135).
Figure 2.5-4 presents the spatial distribution of the mean and median tPCB concentrations for all
surficial (0 to 6 inches [0 tol5 cm]) floodplain soil by reach for the portion of the river upstream
of the Massachusetts/Connecticut state line, at which point the average PCB concentration in
floodplain soil is less than 1 mg/kg. In addition, little floodplain exists within the Connecticut
portion of the river; therefore, no samples were collected from those reaches. Mean tPCB
concentrations are broadly similar within Reaches 4, 5, and 6, averaging slightly more than 15
mg/kg, and then on average decreasing by an order of magnitude in Reaches 7, 8, and 9.
However, localized areas of higher concentrations are found in Reach 7.
n =
85
Legend
q Modmn tPC8 Concentration
I Two standard $rron of the msan
n =
293
164 n = 806
J
O
n = 716
0
o
1
O
o
n = 700 n = 14 n = 104
1^1 rh
Reach 4 Reach 5A Reach 5B Reach 5C Reach 6 Reach 7 Reach 8 Reach 9
Reach
Figure 2.5-4 Mean Total Surficial Soil PCB Concentrations at Floodplain
Locations by Reach
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC 2 28 7/10/03
-------
1 Figure 2.5-5 Spatially Weighted tPCB Concentrations in Floodplain Soil in the
2 Primary Study Area (Tile 1 of 7)
Click Here to go to
l'"i«ures holder
\
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-30
-------
1 Figure 2.5-6 Spatially Weighted tPCB Concentrations in Floodplain Soil in the
2 Primary Study Area (Tile 2 of 7)
>
Click Here to go to
Figures Folder
\
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs ^ -i 7/12/03
-------
1 Figure 2.5-7 Spatially Weighted tPCB Concentrations in Floodplain Soil in the
2 Primary Study Area (Tile 3 of 7)
/
Click Here to go to
l'"i«ures holder
N
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-32
-------
1 Figure 2.5-8 Spatially Weighted tPCB Concentrations in Floodplain Soil in the
2 Primary Study Area (Tile 4 of 7)
Click Here to go to
l'"i«ures holder
S
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC O ^ 7/12/03
-------
1 Figure 2.5-9 Spatially Weighted tPCB Concentrations in Floodplain Soil in the
2 Primary Study Area (Tile 5 of 7)
Click Here to go to
l'"i«ures holder
S
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC O ^/I 7/12/03
-------
1 Figure 2.5-10 Spatially Weighted tPCB Concentrations in Floodplain Soil in the
2 Primary Study Area (Tile 6 of 7)
3
4
5
Click Here to go to
l'"i«ures holder
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-35
-------
1 Figure 2.5-11 Spatially Weighted tPCB Concentrations in Floodplain Soil in the
2 Primary Study Area (Tile 7 of 7)
3
4
Click Here to go to
Figures Folder
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-36
-------
1 Figures 2.5-5 through 2.5-11 display the spatially weighted floodplain tPCB concentrations in
2 the PSA using the inverse distance weighting procedures described in Appendix C.3.
3 Riverbank soil PCB concentrations are broadly similar in concentration ranges to the floodplain
4 soil and river sediment samples, being highest adjacent to the GE facility in Reach 3 and
5 immediately downstream in Reach 4, and decreasing in concentration within Reaches 6 and 7.
6 2.5.2.3 Surface Water
7 Sampling for PCBs in surface water was conducted during both low flow conditions and during
8 higher or storm flow conditions. During lower flow conditions, PCB-contaminated sediment act
9 as the primary source of PCBs in the water column through the processes of diffusion and
10 groundwater advection through the sediment and associated porewater. During higher flows, the
11 principal source of PCBs in the water column is from resuspended sediment, from both upstream
12 and within the PSA.
13 Results for all of the surface water samples collected and analyzed for tPCBs since 1980 are
14 presented in Figure 2.5-12 by river mile. In addition, Figures 2.5-12 and 2.5-13 identify the
15 locations of major impoundment structures found along the Housatonic River from the GE
16 facility to Long Island Sound.
17 ¦ GE = General Electric facility.
18 ¦ WPD = Woods Pond Dam.
19 ¦ RPD = Rising Pond Dam.
20 ¦ GF = Great Falls Dam.
21 ¦ CB = Cornwall Bridge.
22 ¦ BBD = Bulls Bridge Dam.
23 ¦ RDD = Bleachery Dam.
24 ¦ SD (River Mile 25) = Shepaug Dam.
25 ¦ SD (River Mile 15) = Stevenson Dam.
26 ¦ DSD = Derby-Shelton Dam.
27
28 While the analysis of spatial patterns discussed below only used data from 1996 to 2002, all
29 historical data were plotted to show the full set of results, because most of the data downstream
30 of Woods Pond, especially in Connecticut, were collected prior to 1996. Figure 2.5-13 presents
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-29
-------
Massachusetts
WPD
RPD
I
<
Connecticut
n = 1706
n in CT = 97
GFD CB
BBD BD SD StD DSD
2.5
2.25
1.75
1.5 o
1.25 8
c
o
a
i m
1 o
a
0.75 2
0.5
0.25
150 140 130 120 110 100 90 80 70 60 50 40 30 20 10 0
River Mile
Figure 2.5-12 Total PCB Concentrations Measured in all Surface Water Samples
Collected from the Housatonic River Since 1980
Massachusetts
GE
WPD
RPD
Connecticut
GFD CB
• Detects
o Non-Detects
BBD BD SD StD DSD
2.5
2.25
1.75 :
1.5 o
1.25
-i m
1 o
0.5
0.25
150 140 130 120 110 100 90
80 70
River Mile
60 50 40 30 20
10
Figure 2.5-13 Total Surface Water PCB Concentrations by River Mile (Data
Collected Since 1996)
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-37
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
only the surface water data collected since 1996. As indicated in this figure, tPCB
concentrations increase at the GE facility and then decrease downstream through to Rising Pond
Dam. Concentrations of tPCBs continue to decrease by an order of magnitude downstream of
Rising Pond Dam and into the Connecticut portion of the river. Most (approximately 80%) of
the samples collected in Connecticut, both before and since 1996, were non-detect. Within
Reaches 5 and 6, PCBs were detected at all surface water sampling locations and were fairly
constant in concentrations across the study area. More than half of the samples collected from
Reaches 5 and 6 contained tPCBs above the chronic ambient water quality criterion (cAWQC)
for protection of aquatic life of 0.014 (ig/L.
2.5.2.4 Biota
Biological tissue sampling was conducted to support both the human health and ecological risk
assessments and modeling study. In general, most tissue samples collected were analyzed for
tPCBs and PCB congeners, dioxins/furans, and organic carbon (OC) pesticides. Figures 2.5-14
and 2.5-15 present the distribution of tPCB concentrations for a majority of the biota used to
evaluate PCB exposure in the baseline ERA.
180
160
£ 140
O)
•
2 120
s
o>
% 100
E
GO
0 80
CL
1
h 60
40
20
~
~ IrxUvldual sample
• Arithmetic mean
Median
i
~
i
H
I
«
"T
*
i
I ! I ! *
w
-~
~
~
Figure 2.5-14 Total PCB Concentration (mg/kg wet weight) in Selected Biota
(Excluding Fish) for Reaches 5 and 6
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-38
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
|? 300
E
~
~
~ Individual sample
• Arithmetic mean
Median
15 150
~
I
n
~
a
t
~
~
~
*
~
~
r\
*
n
~
~
*
*
~
~
i i
Note: Includes young-of-year fish.
Figure 2.5-15 Total RGB Concentration (mg/kg wet weight) in Reaches 5 and 6
Fish
2.5.3 Identification of Exposure Pathways
Considerable variability was observed in PCB concentrations in various media in the Housatonic
River watershed. This variability is primarily attributable to the following factors:
¦ Differences in PCB concentrations among various abiotic exposure media (soil,
sediment, water), particularly the small-scale heterogeneity observed in PCB
concentrations in sediment.
¦ Analytical variability within a medium, which has been assessed and quantified (see
Appendix C.ll).
¦ Species-specific physiology, such as lipid content and metabolic requirements of the
animals.
¦ Differences in life history and foraging behavior that affect duration and magnitude of
PCB exposures.
¦ Position in aquatic or terrestrial food webs, thus affecting degree of biomagnification.
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-39
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Sediment-associated invertebrates have significant exposures to PCBs because they remain in
continuous contact with the sediment bed, which contains relatively high PCB concentrations.
Water column invertebrates also accumulate PCBs, either through respiration of PCBs in the
water column, or by ingestion of contaminated suspended particulate matter. Overall, food
ingestion is the dominant pathway of PCB uptake for aquatic organisms in the Housatonic River.
Fish species exhibit interspecies variation in PCB concentrations. This partly reflects the
differences in PCB concentrations in the abiotic media to which the fish are exposed. For
example, forage fish tend to have lower PCB body burdens compared to benthic fish, which are
in contact with contaminated sediment and porewater. However, the main reason for the
interspecies differences is not direct contact with PCB-contaminated media, but rather
differences in the dietary uptake patterns. Biomagnification in the food web also is a major
factor controlling the PCB concentrations in fish. Biomagnification represents trophic-level
differences in PCB concentrations and is measured as the increase in lipid-based contaminant
concentrations in predators over those in prey (Russell et al. 1999). The mean tPCB and lipid-
normalized tPCB concentrations in whole fish, by reach, are presented in Figures 2.5-16 and 2.5-
17.
200
— 150
o
U)
100
0Q
O
Q.
+¦»
50
0
Figure 2.5-16 EPA Fish Collections (1998-2000) - Median tPCB Concentrations -
All Ages by Subreach in the PSA
¦ Reach 5A
~ Reach 5B&C
~ Reach 6
i
Brown Bullhead Goldfish White Sucker Pumpkin Seed Largemouth Yellow Perch
Bass
Fish Species
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-40
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
4000
jg" 3000
|5.
o
J*
o> 2000
E
m
o
S: 1000
0
Figure 2.5-17 EPA Fish Collections (1998-2000) - Median Lipid Normalized PCB
Concentrations - All Ages by Subreach in the PSA
Organism foraging behavior plays a substantial role in the bioaccumulation of PCBs. Species
that remain in proximity to the areas of higher PCB concentrations (e.g., main channel sediment)
have increased exposure relative to those that use habitats such as distal floodplains or
woodlands. Some species (e.g., wood frogs) have high exposures during specific life history
stages but may migrate to less-contaminated habitats as adults. Other organisms (e.g., ducks,
large raptors) may have exposures to highly contaminated prey as both juveniles and adults, but
effectively "dilute" their exposures due to their large home ranges and/or seasonal residency in
the Housatonic River watershed.
2.5.4 Changes in PCB Congener Patterns
Because PCBs constitute a group of contaminants rather than a single contaminant, their fate in
the environment is complex. Some congeners are subject to degradation to a greater extent than
others, with the transformation of those congeners and the potential creation of, or enhancement
of, other congeners. In addition, congeners have different rates and extent of exchange in
different media, resulting in differential rates and patterns of transport.
-
1—1
¦ Reach 5A
~ Reach 5B&C
~ Reach 6
¦
-
1—1
—1
1—
1—1
i—i
i—i
i—i
-
—
—
Brown Bullhead Goldfish White Sucker FUrrpkin Seed Largerrouth Yellow Fterch
Bass
Rsh Species
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-41
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
In the Housatonic River, the predominant congeners are the highly chlorinated congeners
associated with the release of Aroclor 1260, and to a lesser degree, Aroclor 1254. The more
highly chlorinated congeners are more resistant to degradation. A number of studies have shown
that under laboratory conditions, PCB congeners in sediment samples from the Housatonic River
can degrade to varying degrees, with the losses of some congeners and increases in the
degradation product congeners (Bedard and May 1996). However, the congener data collected
from the river sediment and floodplain soil do not support degradation as a major removal
process.
During the release and transport of PCBs in the river, the level of chlorination of the congeners
controls, in part, the distribution and exchange of the congeners among the solid and liquid
phases. Increasing the degree of chlorination decreases the solubility of the congener and
increases its tendency to sorb to solid phases. As a result, surface water samples have congener
distributions that have a higher percentage of the lower-chlorinated congeners compared to the
congeners measured on the particulate matter in the same sample. Similarly, the less chlorinated
congeners are present at a higher percentage in porewater than those found in the sediment from
which the water is extracted. The effect of this partitioning phenomenon is that PCBs tend to
fractionate during transport and over time, with the loss of less-chlorinated congeners and the
retention of more highly chlorinated ones. In the Housatonic River, however, the PCBs
discharged from the facility were dominated by the more highly chlorinated congeners, primarily
those associated with Aroclor 1260. As a result, only limited changes in the congener
distribution are observed from differential congener transport.
In 2001, EPA and GE conducted a joint sampling effort to investigate site-specific PCB
partitioning behavior in the Housatonic River. The program entailed synoptic collection of
sediment and porewater, and in a complementary effort, synoptically in surface water and
suspended solids. The synoptic nature of the collections and analyses allowed a detailed
assessment of partitioning behavior and an assessment of shifts in congener distributions among
media. Findings from the study include:
¦ The analyses of approximately 50 paired bulk sediment/porewater samples indicate a
shift in the PCB homolog profiles between media. In bulk sediment, the homolog
profile averaged 5.9 chlorines per biphenyl (Cl/BP), with hepta-PCBs having the
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-42
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
largest contribution to tPCB mass. In contrast, synoptic porewater samples had an
average of 5.3 Cl/BP, with hexa-PCBs having the largest contribution to tPCB mass.
This pattern reflects the congener-dependent partitioning behavior described above.
¦ Spatial trends in chlorination level (which may be used as a surrogate for alterations
in congener distributions related to chemical properties) were evaluated in sediment
and porewater. No trend with distance downstream was observed in porewater. A
modest reduction in chlorination level was observed in bulk sediment, however.
Typically, the majority of Cl/BP ranged from 6 to 6.5 for samples collected within
Reach 5A of the PSA, whereas downstream samples (Reaches 5B and 5C) usually
had 5.5 to 6 Cl/BP. This confirms that changes in the congener distribution with
distance from the source are possible, but are limited because of the highly
chlorinated nature of PCBs in the source media.
¦ Surface water particulate matter in samples collected from four locations (Pomeroy
Avenue, West Branch, New Lenox Road, Woods Pond) exhibited congener/homolog
distributions comparable to bulk sediment. The particulate organic matter yielded an
average of 6.0 Cl/BP, compared to 5.9 Cl/BP for bulk sediment.
¦ Surface water samples from the same four locations yielded dissolved PCB profiles
(4.7 Cl/BP) that were slightly "lighter" than porewater samples (5.3 Cl/BP), primarily
because of an increased percentage of tri-PCBs.
These findings support the conceptual model that PCB congener distributions will differ in
aqueous and particulate media, primarily because of contaminant properties that favor
partitioning to solids (and reduced solubility) for higher chlorinated congeners. Some spatial
variation in the partitioning behavior for sediment is apparent, but does not dominate the
kinetics. Therefore, it appears that physical transport of PCBs (via bedload and suspended
particulate matter at higher flows, and diffusive flux at lower flows) is the dominant fate process,
with dechlorination of PCB mixtures a relatively minor process.
A more extensive evaluation of congener patterns in sediment, soil, and tissue, using multivariate
classification analysis (Euclidean distance) and principal components analysis (PCA), was
performed. These analyses included a broad range of media, including floodplain soil, bank soil,
bed sediment, suspended sediment, and tissue (e.g., bullfrogs, fish, tree swallows, crayfish, and
ducks). The analyses were conducted to investigate the similarity of congener patterns within
and among groups of samples for the purpose of measuring the differences between groups and
the level of consistency within groups. Overall, the congener evaluation (Appendix C.7)
indicated that differences in profiles are sometimes evident, but that most media exhibit congener
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-43
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
profiles similar to Aroclor 1260 across all reaches. Some changes in congener profiles were
observed both spatially and across media, with the latter differences larger than the former.
2.5.5 Fate and Transport of Dioxins/Furans
The following discussion presents an overview of the general fate and transport mechanisms
associated with dioxins/furans that were retained as COPCs in all media (see Section 2.4).
2.5.5.1 Transport and Partitioning
Dioxins and furans, similar to PCBs, are characterized by low solubility, low vapor pressure, and
high affinity for organic carbon (log Koc values as high as 7.39), which suggests that they will
strongly adsorb to sediment or soil and that their vertical movement in either medium will be
limited. The leaching of dioxins and furans is unlikely if water is the only transporting medium;
however, saturation of sorbed sites and the presence of organic solvents or petroleum may result
in vertical migration in sediment or soil.
Volatilization from soil during warm months may also be a major partitioning mechanism. In
general, the higher the degree of chlorination, the lower the relative degree of volatilization from
soil or water.
In the atmosphere, dioxins and furans are typically adsorbed to particulates with the vapor-phase
tending to be negligible (Paustenbach et al. 1991). Vapor pressure and ambient temperature are
the two environmental factors controlling the phase of congeners in the atmosphere. Congeners
having a vapor pressure greater than 10"4 mm Hg will exist primarily in the vapor phase. Dioxins
and furans have relatively long residence times in the atmosphere and are removed by wet, dry,
and gas-phase (vapor phase onto plant surface) deposition (ATSDR 1998). Contamination of
plant foliage via atmospheric deposition is the primary mechanism of accumulation in terrestrial
plants.
Dioxin and furan adsorption to particulates in the water column increases with increasing
chlorination. Dioxins and furans are removed from the water column primarily by binding with
particulates, sediment, or biota and to a lesser extent by volatilization (Paustenbach et al. 1992).
Resuspension of sediment-bound dioxins and furans can increase their transport and availability
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC O /I /I 7/10/03
-------
1 for uptake by aquatic biota. The primary route of exposure to dioxins and furans for lower
2 trophic-level organisms is uptake from water. Bioaccumulation appears to increase with
3 increasing chlorination up to T(tetra)CDDs and TCDFs. For higher trophic-level organisms, the
4 predominant route of exposure is via food chain transfer.
5 2.5.5.2 Transformation and Degradation
6 Photolysis of dioxins and furans in sediment or soil is a relatively slow process when compared
7 with aquatic photolysis rates. However, the addition of organic solvents to contaminated
8 sediment or soil can enhance photolytic transformation. Field and laboratory studies have shown
9 that several microorganisms (e.g., fungi and bacteria) are capable of degrading different
10 congeners. In general, the rate of biodegradation decreases with increasing chlorination.
11 In the atmosphere, dioxin and furan reactions with hydroxyl radicals appear to be the most
12 significant source of transformation. Vapor-phase dioxins and furans may also undergo
13 photolytic degradation. The estimated half-life for TCDD reactions with hydroxyl radicals is 2
14 to 8 days, and the estimated photolytic lifetime ranges from 1 to 7 days (ATSDR 1998). OCDDs
15 and OCDFs, with their low vapor pressure, partition to the particulate phase. Atmospheric
16 photodegradation of these highly chlorinated congeners is less likely.
17 Dioxins and furans in aquatic environments are primarily associated with particulate matter.
18 Photodegradation of bound dioxins and furans occurs near the water's surface and decreases with
19 water depth. Biodegradation in the water column does not appear to be a significant
20 transformation mechanism. Limited biodegradation of dioxins and furans has been observed in
21 sediment, with degradation rate decreasing with increasing chlorination.
22 2.6 EFFECTS ON RECEPTORS
23 There are a number of chemical stressors that may have an adverse impact upon organisms found
24 in the Housatonic River PSA. The preliminary ecological risk assessment (Pre-ERA) identified
25 24 COPCs that are of interest (Appendix B, Section 2.3). A short review of toxicity
26 mechanisms and the possible effects to aquatic and terrestrial organisms follows for PCBs and
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-45
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
dioxins/furans. A more detailed, receptor-specific review of COC toxicity is presented in each of
the assessment endpoint appendices (Appendices D through K).
2.6.1 Polychlorinated Biphenyls (PCBs)
The toxicology of PCBs varies considerably among congeners, depending on the number and
location of chlorines on the biphenyl molecule, and also between animal species due to
differences in absorption, metabolism, mechanism of action, and potential toxic effects (Eisler
and Belisle 1996).
PCB congeners vary in toxicity in many ways, including mode of action, potency, and potential
for interaction. PCB congeners may interact with each other and with other chemicals when
combined in a complex commercial PCB mixture. Lethal and sublethal effects of PCBs on
mammals, birds, and aquatic life, are discussed in detail in the appropriate assessment endpoint
appendices; a general summary of PCB-associated effects is presented in Table 2.6-1. The
following discussion of PCB toxicology focuses primarily on the general mechanisms of PCB
toxicity.
PCB congeners differ in their biological activities, and different animal species vary in their
sensitivity to the individual congeners. Multiple and diverse mechanisms are involved in the
toxicological responses of animals to PCB exposures. The mechanism of Ah-receptor binding is
an initial step in producing toxic effects, and is the basis for the World Health Organization's
(WHO) toxic equivalency factors (TEFs) approach for ranking the relative potency of PCBs,
PCDDs, and PCDFs (Van den Berg et al. 1998). The WHO TEFs only apply to Ah-receptor-
mediated biochemical responses and toxic effects. The relationship between PCB molecular
structure and the potential for toxic effects independent of Ah-receptor mediation is not clearly
understood. Research through the 1990s found increasing evidence for alternative mechanisms
for several PCB-induced effects such as neurotoxicity and disruption of neutrofil function
independent of Ah-receptor mediation (ATSDR 2000). In addition, there is a third category,
where PCB toxicity may be initiated by both Ah-receptor-dependent and independent
mechanisms.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-46
-------
1
2
3
Table 2.6-1
Common Effects of PCB Exposure Observed in Various Animals
System Affected
Specific Effect
Hepatic effects
¦ Hepatomegaly, bile duct hyperplasia
¦ Widespread (e.g., rabbit) or focal (e.g., mouse) necrosis
¦ Lipid accumulation, fatty degeneration
¦ Induction of microsomal monooxygenases and other enzymes
¦ Decreased activity of membrane ATPases
¦ Depletion of fat-soluble vitamins
¦ Porphyria
Gastrointestinal effects
¦ Hyperplasia and hypertrophy of gastric mucosa
¦ Gastric ulceration and necrosis
¦ Proliferation and invasion of intestinal mucosa (monkey)
¦ Hyperplasia, hemorrhage, necrosis (hamster, cow)
Respiratory system
¦ Chronic bronchitis, chronic cough
Nervous system
¦ Alterations in catecholamine levels
¦ Impaired behavioral responses
¦ Developmental deficits
¦ Depressed spontaneous motor activity
¦ Numbness in extremities
Skin
¦ Chloracne
¦ Edema, alopecia
Immunotoxicity
¦ Lymphoid involution (spleen, lymph nodes, especially thymus)
¦ Subsequent reduction of circulating lymphocytes
¦ Suppressed antibody responses
¦ Enhanced susceptibility to viruses
¦ Suppression of natural killer cells
Endocrine system
¦ Altered levels of circulating steroids
¦ Estrogenic, antiestrogenic, antiandrogenic effects
¦ Decreased levels of plasma progesterone
¦ Adrenocortical hyperplasia
¦ Thyroid pathology, changes in circulating thyroid hormones
Reproduction
¦ Increased length of estrus
¦ Decreased libido
¦ Embryo and fetal effects following in utero exposure
Carcinogenesis
¦ Promoter
¦ Attenuation of some carcinogens
Source: Hansen 1994
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-47
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
PCBs are able to induce hepatic Phase I enzymes (CYP oxygenases) and Phase II enzymes (e.g.,
UDP glucuronyltransferases, epoxide hydrolase, glutathione transferase). Most commercial PCB
mixtures induce both 3-methylcholanthrene type (CYP1A1 and 1A2) and phenobarbital-type
(CYP2B1, 2B2, and 3A) CYPs. Non-ortho and mono-ortho PCBs can assume a coplanar
molecular configuration and bind to the Ah receptor causing CYP1A1/1A2 induction in rodents
(Safe 1994). Effects from PCBs involving the Ah-receptor-initiated mechanisms include body
weight wasting, thymic atrophy, porphyria, and porphyria cutanea tardea (Safe 1994).
There are many examples of the complexity of the relationship between PCB molecular structure
and toxic effects independent of Ah-receptor initiation. For example, some PCBs with two ortho
chlorines and lateral chlorines induce both types of CYPs but demonstrate little Ah-receptor
affinity. Di-ortho PCBs with one or two para chlorines predominantly induce CYP2B1/2B2/3A
and have no affinity for the Ah receptor (Connor et al. 1995). The induction of phenobarbital-
type CYPs by PCBs is independent of the Ah receptor. PCBs with at least two ortho chlorines
and one or two para chlorines are the most potent CYP inducers.
Neurological and neurodevelopmental effects involving changes in brain dopamine levels are
PCB-induced effects that are Ah-receptor independent. Scientists have hypothesized that the
effect on dopamine levels is related to decreased dopamine synthesis by PCB inhibition of
certain enzymes or decreased dopamine uptake into vesicles (ATSDR 2000). It is also possible
that a connection exists between disruption of Ca2+ homeostatic mechanisms and neurological
and neurodevelopmental effects. It is clear that Ah-receptor-independent mechanisms are
important in the induction of neurotoxic effects by PCBs.
In vitro studies have indicated that PCBs can induce functional changes such as degranulation in
neutrofils (ATSDR 2000). These functional changes may be related to PCB toxicity such as
immunological effects and tissue damage. Immunological effects that involve neutrofils include
defenses against pathogens and inflammatory responses leading to tissue injury.
There are a number of effects that involve both Ah-receptor-dependent and -independent
mechanisms. These include liver hypertrophy, neurodevelopmental, or reproduction effects
involving changes in steroid hormone homeostasis and/or thyroid hormone disruption,
immunological effects, and cancer (Safe 1994; ATSDR 2000).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC O /IQ 7/10/03
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Safe (1994) reviewed numerous studies of PCB-induced hepatoxicity in mammals including rats,
mice, rabbits, guinea pigs, monkeys, and mink exposed to Aroclors including 1221, 1242, 1248,
1254, and 1260. From these studies, it appears that PCB-induced liver toxicity is mediated by
both Ah-receptor-dependent and -independent mechanisms.
Reproductive impairment following PCB exposure has been observed in mink, one of the most
sensitive mammals to PCB toxicity (Eisler and Belisle 1996; Moore et al. 1999). Although
congeners with high Ah-receptor affinity are more potent than congeners with low Ah-receptor
affinity, there is evidence that Ah-receptor-independent mechanisms may be involved.
Review of the scientific literature indicates that animals exposed to PCBs have an increased risk
of cancer. Lifetime oral exposures to a number of commercial PCB mixtures (Aroclors 1016,
1242, 1254, and 1260) have produced liver tumors in female rats and Aroclor 1260 has produced
liver tumors in male rats. Mixtures with high chlorine content such as Aroclor 1254 were
generally more potent than mixtures with low chlorine content such as Aroclor 1016 (Mayes et
al. 1998).
2.6.2 Dioxins/Furans
Many halogenated aromatic compounds have been described as exhibiting dioxin-like behavior,
such as poly chlorinated dibenzofurans (PCDFs) and some coplanar poly chlorinated biphenyls
(PCBs), based on similarities in toxicity and mechanism of action. The primary toxic
mechanism of action is binding of the PCDD, PCDF, or coplanar PCB compound to the Ah
receptor (described in the previous section). Because 2,3,7,8-TCDD binds to the Ah receptor
with a high affinity and has a high toxic potency, it has been the focus of experimental toxicity
studies. EPA, regulatory agencies in other countries, and international organizations such as
WHO use a TEF approach to reflect the varied toxicity of the different PCDDs, PCDFs, and
PCBs.
The impact of dioxins in the environment is directly related to their highly lipophilic and
hydrophobic nature as well as to the toxic effects of these compounds on plants and animals.
These toxic effects have been extensively studied in the laboratory and through evaluation of
animals exposed to dioxins in the environment. The toxicology of PCDD/PCDF varies
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-49
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
considerably between congeners and between animal species in absorption, metabolism,
mechanism of action, and potency of toxic and carcinogenic effects. The following discussion of
PCDD and PCDF toxicology focuses primarily on the general mechanisms of toxicity.
2,3,7,8-TCDD and equivalents share a mechanism of toxicity that initially involves binding of
the individual congener to the cystolic Ah receptor in all animal species. After initial binding,
the ligand-receptor complex is translocated to the nucleus of the cell. It then becomes associated
with the DNA and causes transcription of one or more contaminated genes (EPA 1993). The
physiological effects that follow are species-specific but there are many similarities, including
the induction of enzyme systems such as cytochrome P4501A1, "wasting syndrome," decreased
immunocompetence, reproductive effects, edema, and mortality. The Ah receptor is present in
all mammalian and bird species that have been tested, as well as in many species of fish. It is
unclear whether the Ah receptor is present in amphibians and reptiles.
A protein similar to the Ah receptor has been identified in terrestrial invertebrates, but there is no
evidence to support the existence of an Ah-receptor type protein in aquatic invertebrates (EPA
1993).
2,3,7,8-TCDD toxicity and the toxicity of the other 74 individual PCDD congeners is mediated
by the Ah receptor. Differences between species in sensitivity to 2,3,7,8-TCDD may be related
to the size and binding efficiency of the Ah receptor, pharmacokinetic differences between
species, and additional contributing factors. Ah-receptor affinity is determined by the chlorine
substitution pattern of the individual dioxin congener. 2,3,7,8-TCDD is substituted in all four
lateral positions and has the highest affinity for the Ah receptor. Less active congeners have an
additional one, two, or four nonlateral chlorine substituents or have lateral chlorine substituents
removed. 2,3,7,8-TCDD and structurally related halogenated aromatic compounds induce
microsomal hepatic enzymes such as hepatic aryl hydrocarbon hydroxylase (AHH) and
ethoxyresorufin-O-deethylase (EROD). Both AHH and EROD are markers of CYP1A1 activity.
Increased synthesis of cytochrome P4501A1 (CYP1A1) is induced by an individual dioxin
congener binding to the Ah receptor. CYP1A1 functions in the detoxification or activation of
endogenous and exogenous chemicals. Cytochrome P4501A2 (CYP1A2) is only induced in
hepatic tissue and has a similar function to CYP1A1.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-50
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Effects observed in the offspring of animals exposed to 2,3,7,8-TCDD include fetal/newborn
mortality, structural malformations, impaired development of the reproductive system,
neurodevelopmental effects, immunotoxicity, and thymic atrophy. Impaired development of the
reproductive system and neurobehavioral effects in the developing organism are the most
sensitive endpoints of 2,3,7,8-TCDD exposure (ATSDR 1998).
2,3,7,8-TCDD is a potent animal carcinogen and has tested positive for carcinogenicity in 19
different studies in four animal species: mice, rat, hamsters, and fish (Huff 1992; Johnson et al.
1992). EPA classifies 2,3,7,8-TCDD as a B2, probable human carcinogen (EPA 2002).
The exact mechanism of how 2,3,7,8-TCDD causes carcinogenicity is not well understood but
the evidence indicates that direct DNA damage through formation of DNA adducts is not the
mechanism. The carcinogenicity of 2,3,7,8-TCDD is thought to involve the Ah receptor.
2,3,7,8-TCDD is considered a nongenotoxic carcinogen and has tested as not mutagenic in the
Salmonella/Ames test. 2,3,7,8-TCDD is a potent tumor promoter and is either a weak initiator or
a non-initiator. 2,3,7,8-TCDD and the other carcinogenic dioxin congeners are whole and
complete carcinogens as tested in mice, rats, hamsters, and fish.
PCDD/PCDFs disrupt normal homeostatic processes that regulate cell growth and
differentiation. These disruptions produce a wide range of toxic effects and histopathological
changes. The PCDD/PCDF congeners vary in many ways including affinity for the Ah receptor,
potency, and potential for interaction.
2.6.3 2,3,7,8-TCDD Toxic Equivalence (TEQ)
The polychlorinated halogenated (PCH) congeners (including both PCBs and dioxins/furans)
have different toxicity potencies, and there may be synergistic and/or antagonistic effects among
the congeners. To estimate the relative toxicity of mixtures of PCH mixtures, a system of toxic
equivalency factors (TEFs) has been developed. This approach is based on in vivo and in vitro
toxicity of each of the PCH congeners in relation to 2,3,7,8-TCDD, which is considered to be the
most toxic of the PCH class of chemicals (Van den Berg et al. 1998; Birnbaum and DeVito 1995;
Safe 1994). There are a number of assumptions made when using the TEF approach. These
include: (1) PCH congeners are Ah-receptor antagonists and their toxicological potency is
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-51
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
mediated by their binding affinity; and (2) no interaction occurs between the congeners and thus,
the sum of the individual congener effects accounts for the potency of the PCH mixture. The
overall effect of these assumptions is a potency estimate or toxic equivalency (TEQ) value. To
generate a TEQ, the following equation (equation modified from Van den Berg et al. 1998) is
used:
7 10 12
TEQ = YJ[PCDDn x TEFn] + YJ[PCDFp x TEFp] + £[PCBq x TEFq]
n=1 p=1 q=1
where:
TEQ = Toxic equivalence
PCDDn = Polychlorinated dibenzo-p-dioxin congener concentration
PCDFp = Polychlorinated dibenzo-p-furan congener concentration
PCBq = Polychlorinated biphenyl congener concentration
TEFnp q = Toxic equivalency factor for appropriate individual PCDD/PCDF and PCB
congeners, respectively.
There are a number of TEF approaches available in the scientific literature for PCHs (e.g., Van
den Berg et al. 1998; Kennedy et al. 1996; Safe 1994; NATO 1988). For this ERA, the TEFs
presented by Van den Berg et al. (1998) were adopted. TEF values were developed for those
compounds that: (1) show a structural relationship to PCDDs and PCDFs; (2) bind to the Ah
receptor; (3) elicit an Ah-receptor-mediated biochemical and toxic response; and (4) are
persistent and accumulate in the food chain (Van den Berg et al. 1998; Birnbaum and DeVito
1995).
The Van den Berg et al. TEFs are the most recently proposed and are based on the best available
science in terms of identifying specific endpoints consistent with the mode of action of each of
the congeners. They have also been widely accepted and applied in the scientific literature since
1998 (Dyke and Stratford 2002; Lindstrom et al. 2002). Van den Berg et al. (1998) present TEF
values for use in deriving TEQ for mammals, fish, and birds as predators (Table 2.6-2).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-52
-------
1
2
3
Table 2.6-2
TEF Values for Mammals, Fish, and Birds as Predators
No.
Congener
TEF
Mammals
Fish
Birds
1
PCB-77
0.0001
0.0001
0.05
2
PCB-81
0.0001
0.0005
0.1
3
PCB-126
0.1
0.005
0.1
4
PCB-169
0.01
0.00005
0.001
5
PCB-105
0.0001
<0.000005*
0.0001
6
PCB-114
0.0005
<0.000005*
0.0001
7
PCB-118
0.0001
<0.000005*
0.00001
8
PCB-123
0.0001
<0.000005*
0.00001
9
PCB-156
0.0005
<0.000005*
0.0001
10
PCB-157
0.0005
<0.000005*
0.0001
11
PCB-167
0.00001
<0.000005*
0.00001
12
PCB-189
0.0001
<0.000005*
0.00001
13
1,2,3,4,6,7,8-HpCDD
0.01
0.001
<0.001*
14
1,2,3,4,6,7,8-HpCDF
0.01
0.01
0.01
15
1,2,3,4,7,8,9-HpCDF
0.01
0.01
0.01
16
1,2,3,4,7,8-HxCDD
0.1
0.5
0.05
17
1,2,3,4,7,8-HxCDF
0.1
0.1
0.1
18
1,2,3,6,7,8-HxCDD
0.1
0.01
0.01
19
1,2,3,6,7,8-HxCDF
0.1
0.1
0.1
20
1,2,3,7,8,9-HxCDD
0.1
0.01
0.1
21
1,2,3,7,8,9-HxCDF
0.1
0.1
0.1
22
1,2,3,7,8-PeCDD
1
1
1
23
1,2,3,7,8-PeCDF
0.05
0.05
0.1
24
2,3,4,6,7,8-HxCDF
0.1
0.1
0.1
25
2,3,4,7,8-PeCDF
0.5
0.5
1
26
2,3,7,8-TCDF
0.1
0.05
1
27
2,3,7,8-TCDD
1
1
1
28
OCDD
0.0001
<0.0001*
0.0001
29
OCDF
0.0001
<0.001*
0.0001
4 *Values that are less than should be considered to be the upper limit for use in any TEQ calculation.
5 Source: Van den Berg et al. 1998
6
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs c o 7/10/03
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
2.7 CONCEPTUAL MODEL
A conceptual model is a written description and visual representation of predicted relationships
between ecological entities and the stressors to which they may be exposed. In essence, the
conceptual model presents a series of working hypotheses regarding how the stressors might
affect ecological components at the site. Risk hypotheses are specific assumptions about
potential risk to assessment endpoints and may be based on theory and logic, empirical data, and
mathematical or probability models. The hypotheses are formulated using professional judgment
and available information of the ecosystem at risk, potential stressor sources and characteristics,
and observed or predicted effects on assessment endpoints. Conceptual models include
ecosystem processes that influence receptor responses, or exposure scenarios that qualitatively
link land use activities to stressors and describe primary, secondary, and tertiary exposure
pathways, ecological effects, and ecological receptors (EPA 1998).
The development of the conceptual model is a complex, non-linear process, with many parallel
activities that result in modifications to the conceptual model as additional information becomes
available. The objectives of the conceptual model presented here are to illustrate the important
relationships within the Housatonic study area, and to specify exposure scenarios evaluated in
the ERA, as a refinement of the conceptual model outlined in the SI Work Plan. The model was
refined based upon physical, chemical, and biological information collected during the
investigation, and on the body of scientific knowledge on the COPCs that has also evolved in the
interim. The following discussion presents an overview of the primary exposure pathways, risk
questions/testable hypotheses, and a visual representation of the predicted relationships between
ecological receptors and contaminant stressors (see Figure 2.7-1).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-54
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
d
C*
5
o
.2
¦3
Historic releases of PCBs from the Pittsfield GE facility
and surrounding disposal areas
' '
r '
r
Wetland and surface water
discharge
1 Floodplain runoff 1 1 Groundwater discharge 1
Legend
Direct uptake
Trophic transfer
Both
Potential Effects
Contaminated soil, sediment, water and biota
I
Terrestrial
invertebrates
Earthworm, slug
i
Herbivorous Birds
Goose, finch, dove
Aquatic vegetation, Algae
Phytoplankton, filamentous
algae, periphyton, macrophyte
Omnivorous Birds
Moorhen, duck, rail
I
Omnivorous Fish
Carp, white sucker
Pelagic invertebrates
Cladocerans, ostracods
Carnivorous Fish
Bass, bluegill, perch
Carnivorous Birds
Eagle, robin, heron,
kingfisher, tree swallow
Omnivorous Mammals
Raccoon, vole, mouse
Benthic organisms
Mussel, insect larva
i
Carnivorous Amphibian;
Frog, salamander
Carnivorous Mammals
Mink, otter, fox, shrew
Decreased Survival, Growth, or Reproduction
Figure 2.7-1 Housatonic River Ecological Risk Assessment Conceptual Model:
Principal Exposure Pathways for PCBs
2.7.1 Exposure Pathways
Exposure of receptors to COPCs is possible through various pathways including absorption
through gills, dermal contact, ingestion of sediment, ingestion of surface water, ingestion of soil,
ingestion of contaminated food, and inhalation of volatilized substances. Sediment may become
resuspended if hydrodynamics disturb the sediment bed and distribute suspended sediment
outside the river when floodplains are inundated. Organisms may also be exposed to chemical
contaminants through trophic transfer. Organisms lower in the food chain may ingest and
accumulate a contaminant, which is then passed on when they are consumed by higher food
chain predators.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-55
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Benthic and soil communities are at risk of direct exposure to PCBs and several other COPCs
(e.g., dioxins, furans, lead, mercury, PAHs). Species in these communities are exposed to
COPCs through direct contact with interstitial porewater, ingestion of sediment particles, and
ingestion of organisms that have also been exposed to contaminants.
Pelagic organisms in the Housatonic River system are exposed to COPCs through dermal and
gill contact with surface water; ingestion of water, suspended sediment, and organic matter;
ingestion of sediment for bottom-feeding fish; and ingestion of other benthic and pelagic
organisms. Uptake of PCBs by fish occurs mainly through the gills and the gastrointestinal tract
(Shaw and Connell 1984). Most PCB accumulation in top fish predators can be attributed to the
food pathway (Thomann 1989). Other species, such as amphibians, are also exposed to PCB-
contaminated surface water. The early life stages of these organisms are entirely aquatic, and
because the skin is a respiratory surface during this phase, dermal exposure may be important.
Insectivorous, carnivorous, and piscivorous birds and mammals that reside, or partially reside,
within the PSA are exposed to PCBs principally through diet and trophic transfer. PCBs are
highly bioaccumulative substances that increase in concentration as they are passed up the food
chain. For organisms inhabiting the Lake St. Clair ecosystem, Haffner et al. (1994) showed that
PCB concentrations increased from 935 (J,g/kg in sediment, to 1,360 (J,g/kg in bivalves, to 7,240
[j,g/kg in oligochaetes, and to 64,900 (J,g/kg in predatory gar pike. MacKay (1989) has also noted
the food chain biomagnification of PCBs for several piscivorous birds. The avian and
mammalian predators of the Housatonic River study area would similarly be expected to
accumulate PCBs from the prey they consume. Water, sediment, and soil consumption from
foraging activities likely contribute less to PCB exposure.
The exposure pathways for other COPCs depend largely on their chemical and physical
properties. Highly lipophilic substances, such as dioxins and furans, will behave similarly to
PCBs, partitioning to sediment and being upwardly mobile in the food chain.
Figure 2.7-1 characterizes the ecosystem in the Housatonic River PSA, as well as the major
exposure pathways for COPCs.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-56
-------
1 As a component of the development of the site conceptual model, testable hypotheses or "risk
2 questions" are developed to provide the basis for the study design and selection of measurement
3 endpoints. These hypotheses represent statements regarding anticipated ecological effects and
4 define the focus of the individual studies. In general, the primary question to be asked by the
5 risk hypothesis is "what probabilities are associated with effects of differing magnitudes as a
6 result of exposure of the assessment endpoint to the COPC?" The three major lines of evidence
7 used to answer this question are:
8 ¦ Comparison of an estimated or measured exposure concentration of a COPC to
9 concentrations known from the literature to be toxic to receptors associated with the
10 assessment endpoint.
11 ¦ Comparison of laboratory bioassay results using media from the site to the results
12 using media from a reference site, and/or comparing in situ toxicity test results at the
13 site to results at a reference location, or comparisons of results across a concentration
14 gradient.
15 ¦ Comparison of observed effects in the receptors in the field, with observations in
16 similar receptors at reference locations, or across a concentration gradient (e.g.,
17 exposure modeling).
18
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-57
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
2.8 SELECTION OF ASSESSMENT AND MEASUREMENT ENDPOINTS
The selection of endpoints for consideration in an ERA requires identification of ecological
characteristics that may be adversely affected by site contaminants. In an ERA, two types of
endpoints are required - assessment endpoints and measurement endpoints. Assessment
endpoints represent specific ecological values deemed important to protect; measurement
endpoints are the tools used to determine the outcome for the assessment endpoints.
2.8.1 Assessment Endpoints
Assessment endpoints are unambiguous statements or goals concerning specific ecological
characteristics (e.g., reproductive effects on aquatic organisms) that are to be evaluated and
protected (EPA 1994, 1998). Assessment endpoints determine the foundation for the ERA
because they:
¦ Provide guidance for evaluating the site and the extent of contamination.
¦ Establish a basis for assessing the potential risks to identified receptors.
¦ Assist in the identification of the ecological structure and function at the site.
Each site or area evaluated in an ERA has the potential to be biologically unique; therefore, there
is no universal list of assessment endpoints (Suter 1993). Because it is not practical or possible to
directly evaluate risks to all of the individual components of the ecosystem at a site, assessment
endpoints should focus the risk assessment on particular components of the ecosystem that could
be adversely affected by contaminants from the site (EPA 1997). According to EPA's Ecological
Risk Assessment Guidance for Superfund (EPA 1997):
"Assessment endpoints for the baseline ERA must be selected based on the ecosystems,
communities, and/or species potentially present at the site. The selection of assessment
endpoints depends on:
¦ The contaminants present and their concentration;
¦ Mechanisms of toxicity of the contaminants to different groups of organisms;
¦ Ecologically relevant receptor groups that are potentially sensitive or highly exposed
to the contaminant and attributes of their natural history; and
¦ Potentially complete exposure pathways."
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC 2 58 7/10/°3
-------
1 To guide this process, EPA (1998) provides further detail on the criteria that assessment
2 endpoints should satisfy:
3 Ecological relevance—Assessment endpoints must reflect biologically important
4 characteristics of the ecosystem and should be functionally related to other key
5 components of the system. Ecologically relevant assessment endpoints are particularly
6 valuable for identifying potential cascading adverse effects resulting from the loss or
7 reduction of a species or guild. For example, an alteration of the benthic community is of
8 concern not only to the benthos, but also to higher trophic levels because of disruption at
9 the base of the food web. Alternatively, an alteration at the higher trophic levels may
10 reflect an integration of effects throughout the ecosystem.
11 Susceptibility to known or potential stressors—There should be a cause-effect linkage
12 between the assessment endpoint and the magnitude of the contaminant stressor.
13 Relevance to management goals—The selection of endpoints that reflect societal values
14 and management goals, while not scientifically based, ensures that the risk assessment
15 will have utility for the risk management decisions that must be made. Management goals
16 are desired characteristics of the ecosystem deemed to have value to the public. For
17 example, fish abundance and biomass may be used as indicators of whether fisheries are
18 being adequately protected. The status of the benthic invertebrate community in the
19 study area is often a good indicator of the overall productivity of the aquatic ecosystem,
20 making it a relevant endpoint for maintaining a viable fishery in an area.
21 In addition, specific assessment endpoints should define the ecological value in sufficient detail
22 to identify measures needed to answer specific questions or to test specific hypotheses (EPA
23 1997). An assessment endpoint must be definable in a practical context, and requires both an
24 entity (that can be clearly defined) and an attribute (that can be measured or assessed). The
25 operational definition ensures that the assessment endpoint can be linked with a measured
26 response.
27 Ultimately, the value of an ERA depends on whether it can be used to determine if a baseline
28 risk is present and to support appropriate managerial decisions. Therefore, the selection of
29 assessment endpoints is fundamental in determining the utility of the risk assessment process.
30 Once assessment endpoints are selected and the conceptual model of exposure is developed,
31 testable hypotheses and measurement endpoints are developed to determine whether or not a
32 potential threat to the assessment endpoints exists (EPA 1997).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-59
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
2.8.2 Measurement Endpoints
A measurement endpoint is defined as "a measurable ecological characteristic that is related to
the valued characteristic chosen as the assessment endpoint." Measurement endpoints link the
conditions existing on-site to the goals established by the assessment endpoints through the
integration of modeled, literature, field, or laboratory data (Maughan 1993).
"Measurement endpoints are frequently numerical expressions of observations (e.g., toxicity test
results, community diversity measures) that can be compared statistically to a control or
reference site to detect adverse responses to a site contaminant" (EPA 1997). Measurement
endpoints can include measures of exposure (e.g., contaminant concentrations in water or
tissues) as well as measures of effect.
It is desirable to have more than one measurement endpoint for each assessment endpoint,
thereby providing multiple lines of evidence for the evaluation. However, the primary
consideration for selecting measurement endpoints should always be how many and which lines
of evidence are appropriate to support risk management decisions at the site. Once it has been
determined which lines of evidence are required to answer questions concerning the assessment
endpoint, the measurement endpoints by which the questions or test hypotheses will be examined
are selected (EPA 1997).
In selecting an appropriate measurement endpoint to represent an assessment endpoint, the
following criteria are considered (Suter 1991):
¦ Corresponds to or is predictive of an assessment endpoint.
¦ Readily measurable.
¦ Appropriate to site scale, exposure pathways, and temporal dynamics.
¦ Diagnostic.
¦ Broadly applicable.
¦ Standard.
In particular, measurement endpoints that address both sensitivity and likely exposure to
stressors are relevant to management concerns (EPA 1998).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-60
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
With the selection of measurement endpoints, the conceptual model development is essentially
completed. The conceptual model, which is discussed in Section 2.7, then is used to guide the
study design and development of data quality objectives (DQOs).
Over a period of several years preceding the Consent Decree, EPA, GE, and other stakeholders
discussed available information on contaminants and the Housatonic River ecosystem, and
determined the assessment endpoints appropriate for the ERA. Past discussions and written
comments between these parties demonstrated that while some parties expressed a preference for
measurement endpoints using controlled studies, others had a preference for field-based
observations and studies. The EPA SIWP addressed both of these preferences by including both
a field and a controlled study component for assessment endpoints, where possible and/or
appropriate. In addition, GE supplemented the measurement endpoints in the EPA SIWP with
studies they conducted independent from agency review, but subject to EPA oversight. GE
requested that these studies be incorporated into this ERA, and EPA has done so where the study
was determined to be relevant to the assessment endpoint. The assessment and associated
measurement endpoints that were used by EPA to evaluate potential ecological risks resulting
from PCBs, and possibly other contaminants in the Lower Housatonic River that were
established in the EPA SI Work Plan (WESTON 2000), are presented in Table 2.8-1. The
independent studies that GE conducted are summarized in Table 2.8-2.
The conceptual model for the site demonstrates the complexity of the ecosystem being evaluated.
It was necessary to develop assessment endpoints that were representative of the varying habitats
and exposure pathways that exist at the site, and for which there is the potential for differing
baseline risk to occur (i.e., a deepwater riverine reach versus a forested floodplain). In addition,
many studies conducted as part of this investigation included multiple measurement endpoints in
the design. Rather than list these individual measurement endpoints separately, the assessment
endpoint and principal measurement endpoints are presented in Table 2.8-1. A listing of all the
measurement endpoints included in the design is presented in the SOPs for the individual studies
(WESTON, 2000b). In some cases, the investigators added additional endpoints during the
conduct of the study. These are discussed in the individual investigator reports, and, where
relevant to the assessment endpoint, in the appropriate assessment endpoint appendix.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-61
-------
1
2
3
Table 2.8-1
Ecological Assessment and Measurement Endpoints
Receptor
Assessment
Endpoint
Measurement Endpoint
Benthic
Invertebrates
Community structure,
survival, growth, and
reproduction
Community composition; species richness, abundance, and biomass
and other metrics compared with similar metrics at reference
locations.
Sediment Quality Triad evaluation—Evaluation includes benthic
community composition, sediment toxicity testing, and sediment
chemistry.
Sediment macroinvertebrate chronic toxicity testing using Hyalella
azteca to determine survival, growth, and reproduction; and
Chironomus tentans to determine survival, growth, and emergence.
In situ toxicity studies using C. tentans, Daphnia magna, H. azteca,
and Lumbriculus variegatus to determine survival and growth.
(Growth evaluated only in C. tentans.)
Toxicity Identification Evaluation (TIE) laboratory 24-hour study
using Ceriodaphnia dubia to determine survival for different
porewater fractions of contaminant classes.
Comparison of sediment chemistry with sediment quality values
(SQVs) and tissue chemistry with tissue effects thresholds.
Amphibians
Community
condition, survival,
reproduction,
development, and
maturation
Semiquantitative sampling of larval amphibians in breeding habitats
with different sediment concentrations of stressors. Endpoints
include species richness per habitat type; species abundance; gross
pathology; and body, tail, and total length measurements.
Surveys of vernal pools to quantitate amphibians entering vernal
pools and determine breeding behavior and condition; egg laying,
hatching success, and larval growth and development;
metamorphosis and emigration.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-62
-------
Table 2.8-1
Ecological Assessment and Measurement Endpoints
(Continued)
Receptor
Assessment
Endpoint
Measurement Endpoint
Amphibians
(cont'd)
Amphibian toxicity tests designed with exposure over a gradient of
stressor concentrations in site sediment. Toxicity endpoints include
morphology of embryos and juveniles, limb development, skin
maturation, and tail resorption of Rana pipiens and R. sylvatica.
Gravidity of females; egg count; necrotic eggs; oocyte maturity;
sperm count, morphology, and viability; fertilization rate; embryo
viability; hatching success; mortality; and teratogenesis of Rana
pipiens collected from the study area over a contamination gradient
and compared with an external control.
In situ amphibian toxicity study evaluated how multiple stressors
(including population density and PCB exposure) affect survival and
growth of larval Rana sylvatica.
Fish
Survival, growth, and
reproduction
Determine the possible extent of adverse effects by comparing the
concentrations of COCs in sediment to the concentrations reported
in the literature to cause adverse effects on the survival, growth, or
reproduction of fish.
Compare the concentrations of COCs in fish tissues to the
concentrations in fish tissues that may result in adverse effects,
based on site-specific fish toxicity studies.
Compare the concentrations of COCs in fish tissues to
concentrations documented in the literature to result in adverse
effects.
Evaluate field survey information (fish biomass study, ecological
characterization study, and largemouth bass habitat and
reproduction study) to qualitatively assess potential effects.
Insectivorous Birds
Survival, growth and
reproduction
Reproductive performance of tree swallows (Tachycineta bicolor)
based on the nest box study conducted in areas of varying stressor
sediment concentrations. Parameters for evaluation include nest
building, egg presence/absence, number of eggs, and hatching
success.
Comparison of site-specific tissue concentrations in tree swallows
with reference area concentrations and with residue effects levels
from literature.
Quantitative comparison of daily intakes based on dietary intake of
stressors by tree swallows and American robins using site-specific
stressor levels in invertebrates and comparison with literature-based
effect values.
American robin productivity within the PSA and reference areas
was evaluated and compared to associated PCB exposure. Metrics
assessed included clutch size, hatching and fledgling success, and
PCB concentrations in robin eggs.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs ^ 7/10/03
-------
Table 2.8-1
Ecological Assessment and Measurement Endpoints
(Continued)
Receptor
Assessment
Endpoint
Measurement Endpoint
Piscivorous Birds
Survival, growth, and
reproduction
Quantitative comparison of daily intakes based on dietary intake of
stressors by belted kingfishers and osprey using site-specific fish
tissue concentrations and site-specific stressor levels in other
aquatic-related food items (e.g., crayfish and frogs), with literature-
based effect values.
Belted kingfisher nests within and adjacent to the PSA were
identified and monitored for productivity (i.e., number of eggs,
number of eggs hatched, and fledgling success). Habitat suitability
and modeled PCB exposure were also evaluated and related to nest
productivity.
Piscivorous
Mammals
Survival, growth, and
reproduction
Mink toxicity tests using Housatonic River fish in the diet. Toxicity
endpoints include body weight, food intake rate, length of gestation,
reproductive success (measured by number of females whelping,
newborns/female, litter weight, etc.), survival, histopathology,
presence/absence of jaw lesions, organ weights, and various
biochemical endpoints.
Quantitative evaluation of mink and otter presence using scent posts
and snow tracking, (two separate studies)
Quantitative comparison of daily intakes based on dietary intake of
stressors by mink and river otter using site-specific stressor levels in
fish and other aquatic prey with literature-based effect values.
Omnivorous and
Carnivorous
Mammals
Survival, growth and
reproduction
Reproductive evidence in trapped small mammals (e.g., examination
of placental scars to determine number of litters, and number/litter).
Quantitative comparison of daily intakes based on dietary intake of
stressors by northern short-tailed shrews and red fox using site-
specific stressor levels in soil invertebrates and small mammals with
literature-based effect values.
Demographic characteristics of short-tailed shrew populations were
assessed at six locations within the PSA that spanned a range of
PCB soil concentrations. Population characteristics measured at
each location included survival rate, sex ratio, reproduction and
growth rate, and body mass.
Special Status
Species
(Endangered,
Threatened)
Survival, growth, and
reproduction
Quantitative comparison of daily intakes based on dietary intake of
stressors using site-specific media concentrations and comparison
with literature-based effect values.
1
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC rs 7/10/03
-------
1
2
3
Table 2.8-2
Summary of GE Ecological Studies
Study
Objectives
Robin Productivity in the Housatonic River
Watershed (Henning, Robinson, and Jenkins 2002)
¦ Document reproductive output of robins in the PSA and
reference areas.
¦ Evaluate exposure of eggs and young to PCBs.
¦ Evaluate relationships between exposure and
reproductive output.
Productivity and Density of Belted Kingfishers on
the Housatonic River (Henning and Brooks 2002)
¦ Evaluate kingfisher productivity in situ in a system
with known PCB contamination.
¦ Determine whether estimated PCB dose, habitat
quality, phenology, and/or nest density were significant
predictors of reproductive success.
Experimental Analysis of the Context-Dependent
Effects of Early Life-Stage PCB Exposure on Rana
sylvatica (Resatarits 2002)
¦ Determine the effects and interactions of PCB exposure
and density-dependence on the growth and
development of amphibian offspring.
Spatial and Demographic Effects on Tree Swallow
Nest Quality and Reproductive Success (Robertson
and Jones 2002)
¦ Determine the effects of: (1) inter-nest spacing,
(2) proximity to edge, (3) settlement and nest-building
date, (4) availability of nesting material, (5) history of
the nest-box and nest-box grid, and (6) female and male
age, on both nest quality and reproductive success.
Demography of Short-Tailed Shrew Populations
Living on Contaminated Sites (Boonstra 2002)
¦ Assess whether PCBs adversely affect population
demography of short-tailed shrew living in a natural
environment.
Evaluation of Mink - Presence/Absence,
Distribution, and Abundance in the Housatonic
River Floodplain (BBL 2002)
¦ Qualitatively determine the presence/absence,
abundance, and distribution of free-ranging mink in the
PSA.
Evaluation of Largemouth Bass Habitat, Population
Structure, and Reproduction in the Housatonic
River, Massachusetts (R2 2002)
¦ Determine if largemouth bass (LMB) population in the
study reach is self-sustaining.
¦ Determine if the LMB population is dependent on
tributary recruitment.
¦ Identify which attributes of growth, size-class structure,
and reproduction of the LMB population are similar to
LMB populations in other systems.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-65
-------
1 Several field surveys were conducted to provide information specifically on species presence.
2 Although field surveys can also be used to assess community condition, the majority of the field
3 surveys (with limited exceptions) were designed for community characterization and were not
4 intended to be used as lines of evidence; therefore, they are not included in Table 2.8-1.
5 Tissue samples were collected for contaminant analyses for a number of species in support of the
6 ecological exposure assessment, human health risk assessment, and PCB fate and effects
7 modeling. Endpoints typically associated with residue effects range from general toxicity to
8 reproductive effects and lethality. Where comparable literature-based residue effects data were
9 identified through various literature and toxicity database searches, these were incorporated to
10 provide a comparison of site-specific tissue data with literature-based effects levels in the risk
11 assessment to provide additional lines of evidence.
12 Although many of the endpoints presented are linked to organism-level effects (e.g., survival and
13 reproduction), these endpoints are expected to be strong indicators of potential local population-
14 level effects (e.g., viability of the benthic community within the Housatonic River study area)
15 (EPA 1992, 1999). Extrapolation from organism-level to population-level effects may be
16 logically achieved based on the predictive nature of the endpoint and/or through the use of
17 process-based models.
18 2.9 WEIGHT-OF-EVIDENCE APPROACH TO ANALYSIS
19 Inferences in ERAs are often made by weight-of-evidence (WOE) rather than traditional
20 scientific standards of proof (EPA 1992). The WOE approach is a process by which
21 measurement endpoints are related to an assessment endpoint to evaluate whether significant risk
22 is posed to the environment (Menzie et al. 1996). A formal WOE can range from a simple
23 qualitative assessment to a highly quantitative evaluation; however, no matter what form the
24 WOE takes, it should provide documentation of the thought process used when assessing
25 potential ecological risk.
26 The term "line of evidence" as used in this discussion follows the definition of "Information
27 derived from different sources or by different techniques that can be used to describe and
28 interpret risk estimates" provided in the Guidelines for Ecological Risk Assessment (EPA 1998).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-66
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Unlike the term "weight-of-evidence," this definition does not imply assignment of qualitative or
quantitative weightings to information. The three general lines of evidence under which most
measurement endpoints fall are (Hull and Suter 1994; Suter et al. 1995):
¦ Biological survey data that indicate the state of the receiving environment.
¦ Media toxicity data that indicate whether the contaminated media are toxic (i.e.,
laboratory or in situ toxicity testing).
¦ Single contaminant toxicity data that indicate the toxic effects of the concentration
measured in site media (e.g., exposure modeling).
Two or three general lines of evidence were considered in evaluating potential risk for each
assessment endpoint. A more detailed presentation of the specific lines of evidence used in this
risk assessment is provided in the appendix for each assessment endpoint.
The WOE approach used in this ERA for each of the assessment endpoints follows the approach
originally described in the Massachusetts Weight-of-Evidence Special Report (Menzie et al.
1996).
According to Menzie et al. (1996), WOE is reflected in three characteristics of measurement
endpoints: (1) the weight assigned to each measurement endpoint; (2) the magnitude of response
observed in the measurement endpoint; and (3) the degree of concurrence among outcomes of
multiple measurement endpoints for a given assessment endpoint.
First, weights are assigned to measurement endpoints based on ten attributes (summarized in
Table 2.9-1) related to: (1) strength of association between assessment and measurement
endpoints; (2) data and study quality; and (3) study design and execution. The initial step in this
process involves assigning qualitative (low through high) weights to each attribute, which is a
subjective process involving professional judgment using criteria outlined in Menzie et al.
(1996). This process is described in the appendix for each assessment endpoint. Figure 2.9-1
provides a generic example of the measurement endpoint weighting process used to evaluate
each assessment endpoint.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-67
-------
1
2
3
Table 2.9-1
Attributes for Judging Measurement Endpoints
1. Strength of Association Between Assessment and Measurement Endpoints
Biological linkage between measurement endpoint and assessment endpoint—This attribute refers
to the extent to which the measurement endpoint is representative of, correlated with, or applicable to
the assessment endpoint. If there is no biological linkage between a measurement endpoint (e.g., a
study that may have been performed for some other purpose) and the assessment endpoint of interest,
then that study should not be used to evaluate the stated assessment endpoint. Biological linkage
pertains to similarity of effect, target organ, mechanism of action, and level of ecological organization.
Correlation of stressor to response—This attribute relates to the degree to which a correlation is
observed between levels of exposure to a stressor and levels of response and the strength of that
correlation.
Utility of measure—This attribute relates to the ability to judge results of the study against well-
accepted standards, criteria, or objective measures. As such, the attribute describes the applicability,
certainty, and scientific basis of the measure, as well as the sensitivity of a benchmark in detecting
environmental harm. Examples of objective standards or measures for judgment might include
ambient water quality criteria, sediment quality criteria, biological indices, and toxicity or exposure
thresholds recognized by the scientific or regulatory community as measures of environmental harm.
2. Data and Overall Study Quality
Quality of data and overall study—This attribute reflects the degree to which data quality objectives
and other recognized characteristics of high quality studies are met. The key factor affecting the
quality of the data is the appropriateness of data collection and analysis practices. The key factor of the
quality of the study is the appropriateness and implementation of the experimental design and the
minimization of confounding factors. If data are judged to be of poor or no quality, the study would be
rejected for use in the ERA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-68
-------
Table 2.9-1
Attributes for Judging Measurement Endpoints
(Continued)
3. Design and Execution
Site-specificity—This attribute relates to the extent to which media, species, environmental
conditions, and habitat types that are used in the study design reflect the site of interest.
Sensitivity of the measurement endpoint to detecting changes—This attribute relates to the ability
to detect a response in the measurement endpoint, expressed as a percentage of the total possible
variability that the endpoint is able to detect. Additionally, this attribute reflects the ability of the
measurement endpoint to discriminate between responses to a stressor and those resulting from natural
or design variability and uncertainty.
Spatial representativeness—This attribute relates to the degree of compatibility or overlap between
the study area, locations of measurements or samples, locations of stressors, and locations of
ecological receptors and their points of potential exposure.
Temporal representativeness—This attribute relates to the temporal compatibility or overlap
between the measurement endpoint (when data were collected or the period for which data are
representative) and the period during which effects of concern would be likely to be detected. Also
linked to this attribute is the number of measurement or sampling events over time and the expected
variability overtime.
Quantitativeness—This attribute relates to the degree to which numbers can be used to describe the
magnitude of response of the measurement endpoint to the stressor. Some measurement endpoints may
yield qualitative or hierarchical results, while others may be more quantitative.
Use of a standard method—The extent to which the study follows specific protocols recommended
by a recognized scientific authority for conducting the method correctly. Examples of standard
methods are study designs or chemical measures published in the Federal Register or the Code of
Federal Regulations, developed by ASTM, or repeatedly published in the peer-reviewed scientific
literature, including impact assessments, field surveys, toxicity tests, benchmark approaches, toxicity
quotients, and tissue residue analyses. This attribute also reflects the suitability and applicability of the
method to the endpoint and the site, as well as the need for modification of the method.
1 Source: Menzie et al. 1996.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-69
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Score each measurement endpoint from low to high
Assessment Endpoint:
Attribute
Measurement
Endpoint A
Measurement
Endpoint B
Measurement
Endpoint C
I. Relationship between Measurement and Assessment Endpoints
¦ Degree of Association
Moderate
High
High
¦ Stressor/Response
High
Moderate
High
¦ Utility of Measure
Moderate
High
High
II. Data Quality
¦ Quality of data
High
High
High
III. Study Design
¦ Site-specificity
High
High
High
¦ Stressor-specificity
Moderate
Moderate
Moderate
¦ Sensitivity
Moderate
Low
High
¦ Spatial representativeness
Moderate
High
Moderate
¦ Temporal representativeness
Low
Low
Moderate
¦ Quantitativeness
High
High
High
¦ Use of a standard method
Moderate
Moderate
Moderate
Total Value
Moderate
Moderate
Moderate-High
Figure 2.9-1 Example Endpoint Weighting Sheet
To ensure that the selected measurement endpoints would support the achievement of the study
objectives, a preliminary WOE was conducted in the SIWP. Therefore, it is expected that low
attribute values will not typically be assigned as total scores for a line of evidence in the final
WOE.
The second step of the Menzie et al. (1996) approach is to evaluate the magnitude of response in
the measurement endpoint, considering two questions:
¦ Does the measurement endpoint indicate the presence or absence of risk (yes, no, or
undetermined)?
¦ Is the response low or high?
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-70
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Figure 2.9-2 illustrates a matrix for an assessment endpoint that provides a simple
communication tool summarizing the conclusions of the WOE evaluation of the magnitude of
response.
Assessment Endpoint:
Measurement Endpoints
Weighting
Score (High,
Moderate,
Low)
Evidence of
Harm (Yes,
No,
Undetermined)
Magnitude
(High,
Intermediate,
Low)
Endpoint A
Endpoint B
Endpoint C
Adapted from: Menzie et al. 1996.
Figure 2.9-2 Scoring Sheet for Evidence of Harm and Magnitude
The third step of the WOE process evaluates the degree of concurrence among measurement
endpoints by plotting the output of the two preceding steps on a matrix for all measurement
endpoints associated with a given assessment endpoint (see Figure 2.9-3). The matrix allows
easy visual examination of agreements or divergences among measurement endpoints,
facilitating interpretation with respect to the assessment endpoint. Logical connections,
interdependence, and correlations among endpoints should also be considered when evaluating
concurrence. The generalized matrix shown in Figure 2.9-3 is used for each assessment endpoint
to illustrate the results of the WOE assessment of risks of PCBs and other COCs. Completed
matrices specific to each assessment endpoint are presented in the respective appendix for each
endpoint and each summary section of the report.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-71
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Assessment Endpoint: Maintenance of a henthic community that can serve as a prey base for local fish
~
Harm/Magnitude
Low Weight
Moderate Weight
High Weight
k
Yes/High
A
Y es/Low
B
Undetermined
No/Low
C
No/High
Notes:
Use letter designations to place measurement endpoints in boxes.
~ Indicates increasing confidence in weight.
Adapted from: Menzie et al., 1996.
Figure 2.9-3 Example of Qualitative Assessment
2.10 EXTRAPOLATION OF RISK ESTIMATES FOR SELECTED ENDPOINTS
DOWNSTREAM OF WOODS POND
Because of the decline in PCB mass and concentrations and the associated decrease in the
amount of data collected downstream of the PSA, the detailed approach followed in assessing
ecological risks in the PSA was not appropriate or possible. An estimate of potential ecological
risks was developed using mapping (GIS) techniques and threshold concentrations, that, if
exceeded, would indicate potential risk to six selected target groups: benthic invertebrates,
amphibians, warm-water fish, trout, mink, and bald eagles. These target groups were selected
based on the risks for these organisms observed in the PSA, and the occurrence of these
organisms in the reaches downstream.
For each of these groups, a maximum acceptable threshold concentration (MATC) for total PCBs
(tPCBs) in the appropriate medium was developed, based primarily on the detailed risk
assessment performed for the PSA. The MATC was then compared to available medium-
specific data for areas downstream of Woods Pond to Long Island Sound. Areas of exceedances
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC r* 7/10/03
-------
1 (HQ > 1), indicating potential risk, were plotted on maps of the river. The specific approaches
2 developed for each of the six target groups are discussed in the appropriate appendices and
3 summary sections of this report.
4 2.11 REFERENCES
5 ATSDR (Agency for Toxic Substances and Disease Registry). 1995. Toxicological Profile for
6 Polycyclic Aromatic Hydrocarbons. Prepared by Research Triangle Institute. August 1995.
7 ATSDR (Agency for Toxic Substances and Disease Registry). 1997. Poly chlorinated Biphenyls
8 (PCBs). www.atsdr.cdc.gov/tfactsl7.html. June 12, 2001.
9 ATSDR (Agency for Toxic Substances and Disease Registry). 1998. Toxicological Profile
10 ChlorinatedDibenzo-p-Dioxins. Prepared by Research Triangle Institute. December 1998.
11 ATSDR (Agency for Toxic Substances and Disease Registry). 2000. Toxicological Profile for
12 PolychlorinatedBiphenyls. November 2000.
13 Aulerich, R.J., R.K. Ringer, and J. Safronoff 1986. Assessment of primary versus secondary
14 toxicity of Aroclor 1254 to mink. Archives of Environmental Contamination and Toxicology
15 15:393-399.
16 BBL (Blasland, Bouck & Lee, Inc.). 2002. Evaluation of Mink - Presence/Absence, Distribution,
17 and Abundance in the Housatonic River Floodplain. Report prepared for the General Electric
18 Company.
19 BBL (Blasland, Bouck & Lee, Inc.) and QEA (Quantitative Environmental Analysis, LLC).
20 2003. Housatonic River - Rest of River RCRA Facility Investigation Report, Volume 1. Prepared
21 for General Electric Company. January 2003.
22 Bedard, D. and R. May. 1996. Characterization of the polychlorinated biphenyls in the sediments
23 of Woods Pond: Evidence for microbial dechlorination of Aroclor 1260 in situ. Environ. Sci.
24 Technol. 30:237-245.
25 Birnbaum, L.S. and M.J. DeVito. 1995. Use of toxic equivalency factors for risk assessment for
26 dioxins and related compounds. Toxicology 105:391-401.
27 Boonstra, R. 2002. Demography of Short-Tailed Shrew Populations Living on PCB
28 Contaminated Sites. Report prepared for General Electric Company.
29 Callahan, M.A., M.W. Slimak, N.W. Gabel, I.P. May, C.F. Fowler, J.R. Freed, P. Jennings, R.L.
30 Durfee, F.C. Whitmore, B. Maestri, W.R. Mabey, B.R. Holt, and C. Gould. 1979. Water-Related
31 Environmental Fate of 129 Priority Pollutants. Chapter 36. EPA-440/4-79-029a.
32 Chadwick & Associates, Inc. 1994. Aquatic Ecology Assessment of the Housatonic River,
33 Massachusetts 1993. Chadwick & Associates, Inc., Littleton, CO.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC 2 73
-------
1 Chem Risk. 1997. Work Plan for the Ecological Risk Assessment of the Housatonic River:
2 Volumes I and II. Chem Risk (24 May 1997).
3 Connor, K., S. Safe, and C.R. Jefcoate. 1995. Structure-dependent induction of CYP2B by
4 polychlorinated biphenyl congeners in female Sprague-Dawley rats. Biochem. Pharmacol.
5 50(11): 1913-1920.
6 Cornelissen, G., H. Rigterink, M.M.A. Ferdinandy, and P. van Noort. 1998. Rapidly desorbing
7 fractions of PAHs in contaminated sediments as a predictor of the extent of bioremediation.
8 Environmental Science and Technology 32:966-970.
9 Dyke, P.H. and J. Stratford. 2002. Changes to TEF schemes can have significant impacts on
10 regulation and management of PCDD/F and PCB. Chemosphere 47:103-116.
11 Eisler, R. 1986. Polychlorinated Biphenyl Hazards to Fish, Wildlife, and Invertebrates: A
12 Synoptic Review. Contaminant Hazard Reviews-7. 81 pp. PB86-170057.
13 Eisler, R. 1987. Polycyclic Aromatic Hydrocarbon Hazards to Fish, Wildlife, and Invertebrates:
14 A Synoptic Review. Contaminant Hazard Reviews, Biological Report 85(1.11). May 1987.
15 Eisler, R. and A. Belisle. 1996. Plannar PCB hazards to fish, wildlife, and invertebrates: A
16 synoptic review contaminant hazard reviews. Biological Report 31.
17 EPA (U.S. Environmental Protection Agency). 1980. Ambient Water Quality Criteria Document:
18 PolychlorinatedBiphenyls. p. B-2. EPA 440/5-80-068. As cited in ATSDR 2000.
19 EPA (U.S. Environmental Protection Agency). 1992. Framework for Ecological Risk
20 Assessment. Risk Assessment Forum, Washington, DC. EPA/630/R-92/001.
21 EPA (U.S. Environmental Protection Agency). 1993. Wildlife Exposure Factors Handbook.
22 EPA/600/R-93/187a. Office of Research and Development, Washington, DC.
23 EPA (U.S. Environmental Protection Agency). 1994. A Review of Ecological Assessment Case
24 Studies from a Risk Assessment Perspective, Volume II. Office of Research and Development,
25 Washington, DC. EPA/630/R-94-003.
26 EPA (U.S. Environmental Protection Agency). 1996. PCBs: Cancer Dose-Response Assessment
27 and Application to Environmental Mixtures. National Center for Environmental Assessment.
28 Office of Research and Development, Washington, DC. EPA/600/P-96/001F.
29 EPA (U.S. Environmental Protection Agency). 1997. Ecological Risk Assessment Guidance for
30 Superfund: Process for Designing and Conducting Ecological Risk Assessments. Interim Final.
31 Solid Waste and Emergency Response, U.S. Environmental Protection Agency. EPA-540-R-97-
32 006.
33 EPA (U.S. Environmental Protection Agency). 1998. Guidelines for Ecological Risk Assessment.
34 Risk Assessment Forum, Washington, DC. EPA/630/R-95/002F.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-74
-------
1 EPA (U.S. Environmental Protection Agency). 1999. Ecological Risk Assessment and Risk
2 Management Principles for Superfund Sites. OSWER Directive 9285.7-28P.
3 EPA (U.S. Environmental Protection Agency). 2002. Integrated Risk Information System.
4 Haffner, G.D., M. Tomczak, and R. Lazar. 1994. Organic contaminant exposure in the Lake St.
5 Clair food web. Hydrobiologia 281:19-27.
6 Hansen, L.G. 1994. Halogenated Aromatic Compounds. In Basic Environmental Toxicology.
1 L.G. Cockerham, and B.S. Shane eds. CRC Press, pp. 199-230. ISBN 0-8493-8851-1.
8 Harris, S.L. 1997. Inorganic and Organic Constituents and Grain-Size Distribution in Streambed
9 Sediment and Ancillary Data for the Connecticut, Housatonic, and Thames River Basins Study
10 Unit, 1992-1994. USGS Open-File Report 96-397.
11 Henning, M.H. and R.P. Brooks. 2002. Productivity and Density of Belted Kingfishers on the
12 Housatonic River. Report prepared for General Electric Company.
13 Henning, M.H., S. Robinson, and K. Jenkins. 2002. Robin Productivity in the Housatonic River
14 Watershed. Report prepared for General Electric Company.
15 Hornshaw, T.C., R.J. Aulerich, and H.E. Johnson. 1983. Feeding Great Lakes fish to mink:
16 Effects on mink and accumulation and elimination of PCBs by mink. Journal of Toxicology and
17 Environmental Health 11:93 3-946. As cited in EPA 1996.
18 Huff, J. 1992. 2,3,7,8-TCDD: A potent and complete carcinogen in experimental animals.
19 Chemosphere. 25:173-176.
20 Hull, R.N. and G.W. Suter II. 1994. Using the Weight-of-Evidence Approach for Ecological Risk
21 Assessment at a DOE Facility. SETAC Denver, CO. November 1, 1994.
22 HVA (Housatonic Valley Association). 2001. The Housatonic River Watershed. URL
23 http://www.hvathewatershedgroup.org/HousatonicFactSheet.htm.
24 Johnson, R., J. Tietge, and S. Botts. 1992. Carcinogenicity of 2,3,7,8-TCDD to medaka.
25 Toxicologist. 12(1), 476.
26 Kennedy, S.W., A. Lorenzen, and R.J. Norstrom. 1996. Chicken embryo hepatocyte bioassay for
27 measuring cytochrome P4501a-based 2,3,7,8-tetrachlorodibenzo-p-dioxin equivalent
28 concentrations in environmental samples. Environmental Science and Technology 30:706-715.
29 Lindstrom, G., L.S. Haug, T. Nicolaysen, and E. Dybing. 2002. Comparability of world-wide
30 analytical data on PCDDs, PCDFs and non-ortho PCBs in samples of chicken, butter and
31 salmon. Chemosphere 47:139-146.
32 MacDonald, D.D., C.G. Ingersoll, and T.A. Berger. 2000. Development and evaluation of
33 consensus-based sediment quality guidelines for freshwater ecosystems. Archives of
34 Environmental Contamination and Toxicology 39:20-31.
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-75
-------
1 Mackay, D. 1989. Modeling the long-term behavior of an organic contaminant in a large lake
2 application of PCBs in Lake Ontario. Journal of Great Lakes Research 15:283-297.
3 Mass Wildlife. 2002. http://www.state.ma.us/dfwele/dfw/dfw wd/htm.
4 Maughan, J.T. 1993. Ecological Assessments of Hazardous Waste Sites. Van Nostrand Reinhold,
5 New York, NY.
6 Mayes, B.A., E.E. McConnell, and B.H. Nell. 1998. Compenative carcinogenicity in Sprague
7 Dawley rets of polychlorinated biphenyl mixtures Arcoclor 1016, 1242, 1254, and 1260. Toxicol.
8 Sci. 41(l):62-76.
9 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S.
10 Petron, B. Potocki, S. Svirsky, and P. Tyler. 1996. Special Report of the Massachusetts Weight-
11 of-Evidence Workgroup: A Weight-of-Evidence Approach for Evaluating Ecological Risks.
12 Raman and Ecological Risk Assessment 2(2):277-304.
13 Moore, D.R.J., B.E. Sample, G.W. Suter, B.R. Parkhurts, R.S. Teed. 1999. A Probabilisitic Risk
14 Assessment of the Effects of Methylmercury and PCBs on Mink and Kingfishers along East Fork
15 Poplar Creek, Oak Ridge, Tennessee, USA. Environmental Toxicology and Chemistry, vol. 18,
16 No. 12, pp. 2941-2953.
17 NATO (North Atlantic Treaty Organization). 1988. Pilot study on international information
18 exchange on dioxins and related compounds. International toxicity equivalent factor (I-TEF)
19 method of risk assessment for complex mixtures of dioxins and related compounds. Committee
20 on the Challenges of Modern Society, Report 176. August 1988.
21 Olcott, P.G. 1995. Ground Water Atlas of the United States: Connecticut, Maine, Massachusetts,
22 New Hampshire, New York, Rhode Island, Vermont: Carbonate Rock Aquifers. U.S. Geological
23 Survey Publications HA 730-M.
24 Paustenbach, D.J., R.J. Wennig, V. Lau, N.W. Harrington, D.K Rennix, and A.H. Parson. 1992.
25 Recent developments on the hazards posed by 2,3,7,8-tetrachlorodibenzo-p-dioxin in soil:
26 Implication for setting risk-based cleanup levels at residential and industrial sites. Journal of
27 Toxicology and Environmental Health 36(2): 103-149. As cited in ATSDR 1998.
28 Resetarits, W.J. 2002. Experimental Analysis of the Context-Dependent Effects of Early Life-
29 Stage PCB Exposure on Rana sylvatica. Report prepared for General Electric Company.
30 Robertson, R.J and J. Jones. 2002. Spatial and Demographic Effects on Tree Swallow Nest
31 Quality and Reproductive Success. Report prepared for General Electric Company.
32 Russell, R.W., F.A.P.C. Gobas, and G.D. Haffner. 1999. Role of chemical and ecological factors
33 in trophic transfer of organic chemicals in aquatic food webs. Environmental Toxicology and
34 Chemistry 18:1250-1257.
MK0110:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-76
-------
1 R2 (R2 Resource Consultants Inc.). 2002. Evaluation of Largemouth Bass Habitat, Population
2 Structure, and Reproduction in the Upper Housatonic River, Massachusetts. Report prepared
3 General Electric Company.
4 Safe, S.H. 1994. Polychlorinated biphenyls (PCBs): Environmental impact, biochemical and
5 toxic responses, and implications for risk assessment. Critical Reviews in Toxicology 24(2):
6 87-149.
7 SCS (Soil Conservation Service, U.S. Department of Agriculture). 1970. Soil Survey of
8 Litchfield County, Connecticut. Storrs, CT, USA.
9 SCS (Soil Conservation Service, U.S. Department of Agriculture). 1979. Soil Survey of New
10 Haven County, Connecticut. Storrs, CT, USA.
11 SCS (Soil Conservation Service, U.S. Department of Agriculture). 1981. Soil Survey of Fairfield
12 County, Connecticut. Storrs, CT, USA.
13 SCS (Soil Conservation Service, U.S. Department of Agriculture). 1988. Soil Survey of
14 Berkshire County, Massachusetts, Amherst, MA, USA.
15 Shaw, G.R. and D.W. Connell. 1984. Factors controlling bioaccumulation of PCBs. In PCBs and
16 the Environment, Volume 1. National Technical Information Service, Springfield, VA. pp. 501-
17 516.
18 Suter, G.W., II. 1991. Ecological Endpoints. In Ecological Assessment of Hazardous Waste
19 Sites: A Field and Laboratory Reference. U.S. Environmental Protection Agency. EPA/540/R-
20 92/003.
21 Suter, G.W, II. 1993. Ecological Risk Assessment. Lewis Publishers, Boca Raton, FL. ISBN 0-
22 87371-875-5.
23 Suter, G.W., II, B.E. Sample, D.S. Jones, T.L. Ashwood, and J.M. Loar. 1995. Approach and
24 Strategy for Performing Ecological Risk Assessments for the U.S. Department of Energy's Oak
25 Ridge Reservation. Oak Ridge National Laboratory, Oak Ridge, TN. ES/ER/TM-33/R2.
26 Swain, P.C. and K.B. Kearsley. 2000. Classification of the Natural Communities of
27 Massachusetts. Massachusetts Natural Heritage and Endangered Species Program, Westborough,
28 MA, USA.
29 Thomann, R.V. 1989. Bioaccumulation model of organic chemical distribution in aquatic food
30 chains. Environmental Science and Technology 23:699-707.
31 Van den Berg, M., L. Birnbaum, A.T.C. Bosveld, B. Brunstrom, P. Cook, M. Freely, J.P. Giesy,
32 A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C. Larsen, F.X.R van Leeuwen, A.K.
33 Djien Liem, C. Nolt, R.E. Peterson, L. Poellinger, S. Safe, D. Schrenk, D. Tillitt, M. Tysklind,
34 M. Younes, F. Waern, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs) for PCBs,
35 PCDDs, PCDFs for humans and wildlife. Environmental Health Perspectives 106(12):775-792.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-77
-------
1 Weatherbee, P.B. 1996. Flora of Berkshire County Massachusetts. The Berkshire Museum,
2 Pittsfield, MA, USA.
3 Weatherbee, P.B. and G.E. Crow. 1992. Natural plant communities of Berkshire County,
4 Massachusetts. Rhodora 94:171 -209.
5 WESTON (Roy F. Weston, Inc.). 1998. Source Area Characterization Report. Prepared for U.S.
6 Army Corps of Engineers and U.S. Environmental Protection Agency. DCN GEPM-072198-
7 AABA.
8 WESTON (Roy F. Weston, Inc.). 2000. Supplemental Investigation Work Plan for the Lower
9 Housatonic River. Prepared for U.S. Army Corps of Engineers and U.S. Environmental
10 Protection Agency. 22 February 2000. DCN GEP2-020900-AAME.
11 WESTON (Weston Solutions, Inc.). 2003, In Preparation. Modeling Framework Design:
12 Modeling Study of PCB Contamination in the Housatonic River. Prepared for U.S. Army Corps
13 of Engineers and U.S. Environmental Protection Agency.
14
MK01 |O:\20123001.096\ERA_PB\ERA_PB_2.DOC
2-78
-------
1 3. ASSESSMENT ENDPOINT—COMMUNITY STRUCTURE, SURVIVAL,
2 GROWTH, AND REPRODUCTION OF BENTHIC INVERTEBRATES
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31 3.1 INTRODUCTION
32 The purpose of this section is to characterize and quantify the current and potential risks posed to
33 benthic invertebrates exposed to contaminants of concern (COCs) in the Housatonic River,
34 focusing on total PCBs (tPCBs) and other COCs originating from the General Electric (GE)
35 facility in Pittsfield, MA.
Highlights
Conceptual Model
¦ Sediment and biota tissue are the most relevant exposure media, with the water
column of lesser importance.
Exposure
¦ PCBs, PAHs, and metals retained in as COCs.
¦ COCs were measured in tissue, sediment, and water at up to 13 sediment quality
stations, synoptic with biological effects information. Chemistry data were also
collected at numerous other stations throughout the PSA to provide a broader
characterization of exposure.
Effects
¦ Site-specific toxicity tests (laboratory and in situ) indicate adverse responses,
relative to both reference stations and negative controls.
¦ Benthic community appears altered at multiple stations with elevated PCB
concentrations, relative to reference stations.
¦ Toxicological impacts are significantly correlated with PCB exposures.
Risk
¦ Comparison of exposure concentrations to literature effects benchmarks
(sediment, tissue, water) indicates high probability of harm, particularly due to
PCBs.
¦ Toxicity identification evaluation (TIE) implicates non-polar organics (e.g., PCBs)
as causal agent in toxicity tests.
¦ Contaminants other than PCBs and dioxins/furans do not exhibit concentration
gradients consistent with pattern of effects.
¦ Weight-of-evidence (WOE) approach used to characterize risks, suggesting
significant adverse impacts to benthos predicted throughout the PSA; low risks
predicted downstream of Woods Pond.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-1
-------
1 A Pre-ERA was conducted to narrow the scope of the ERA by identifying contaminants, other
2 than tPCBs, that pose potential risks to aquatic biota in the Primary Study Area (PSA) (Appendix
3 B). In the benthic invertebrate ERA, further screening was done to refine the list of
4 contaminants of potential concern (COPCs) to those that were specifically relevant to the
5 invertebrate community in the main channel of the Housatonic River. COCs that screened
6 through to the risk assessment for benthic invertebrates were tPCBs, several metals, several
7 polycyclic aromatic hydrocarbons (PAHs), and dibenzofuran.
8 A step-wise approach was used to assess the risks of these COCs to benthic invertebrates in the
9 Housatonic River watershed. The four main steps in this process included:
10 1. Development of a conceptual model (Figure 3.1-1).
11 2. Assessment of exposure of benthic invertebrates to COCs (Figure 3.1-2).
12 3. Assessment of the effects of COCs on benthic invertebrates (Figure 3.1-3).
13 4. Characterization of risks to benthic invertebrates (Figure 3.1-4).
14 This section is organized as follows.
15 ¦ Section 3.1 (Introduction and Conceptual Model)—Describes the conceptual
16 model for benthic invertebrates, including selection of representative taxa and
17 establishment of measurement and assessment endpoints.
18 ¦ Section 3.2 (Exposure Assessment)—Describes the quantification of exposures,
19 both specific to stations for which linked biological effects information was collected
20 (n=13) as well as for the broader study area.
21 ¦ Section 3.3 (Effects Assessment)—Describes the potential effects to benthic
22 invertebrates exposed to site COCs, as indicated by the toxicological and biological
23 investigations conducted in the PSA. This section also summarizes the ranges of
24 benchmarks (toxicity thresholds) derived from the literature.
25 ¦ Section 3.4 (Risk Characterization)—Integrates the exposure and effects
26 assessments, and makes conclusions regarding risk for benthic invertebrates in the
27 Housatonic River using three main lines of evidence. A discussion of the sources of
28 uncertainty regarding risk estimates follows. Section 3.4 also presents an
29 extrapolation of risks beyond the PSA to areas downstream of Woods Pond.
30
31
This section provides a summary of the ERA for benthic invertebrates, which is
presented in detail in Appendix D.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-2
-------
<0
o
<0
re
o
o
ac
o
in
.re
T3
O
S2
o
Q.
O
O
O
ac
w
O
£
| Survival, Growth, or Reproduction and/or Community Alteration |
Figure 3.1-1 Conceptual Model Diagram: Exposure Pathways
for Benthic Invertebrates Exposed to Contaminants of Concern (COCs)
in the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-3
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Exposure
Figure 3.1-2 Overview of Approach Used to Assess Exposure of Benthic
Invertebrates to Contaminants of Concern (COCs) in the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ a 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Effects
Figure 3.1-3 Overview of Approach Used to Assess the Effects of Contaminants
of Concern (COCs) to Benthic Invertebrates in the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ r 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Risk Characterization
Figure 3.1-4 Overview of Approach Used to Characterize the Risks of
Contaminants of Concern (COCs) to Benthic Invertebrates in the Housatonic
River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ s 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
3.1.1 Conceptual Model
Total PCBs, dioxins, and furans are persistent and hydrophobic and lipophilic. Therefore,
organic carbon pools (both living and non-living) are the primary uptake vectors for benthic
invertebrates. Less hydrophobic COCs, such as low molecular weight PAHs and metals, are not
as strongly associated with organic pools, and exhibit more complex partitioning behavior. The
COCs identified for benthic invertebrates exhibit both direct (i.e., contact with contaminated
source media) and indirect (i.e., food web bioaccumulation) pathways.
The conceptual model presented in Figure 3.1-1 illustrates the exposure pathways for benthic
invertebrates in the PSA. The benthic invertebrate ERA considered organisms that reside in, or
are in direct contact with, Housatonic River sediment. For sediment invertebrates, the dominant
abiotic exposure media were sediment (solid phase and/or porewater) and surface water.
Concentrations of COCs in tissues of benthic invertebrates were also considered. Tissue data
provide an organism-based measure of bioavailability, and provide an additional line of evidence
to consider along with the conventional Sediment Quality Triad approach (synoptic measurement
of sediment chemistry, toxicity, and invertebrate communities).
The problem formulation (Section 2) identified species used in toxicity tests as surrogates for the
Housatonic River freshwater benthic community (i.e., Chironomus tentans, Hyalella azteca,
Lumbriculus variegatus, Daphnia magna, Ceriodaphnia dubia). Both the status of sensitive
indicator taxa and the overall community composition are considered indicative of the condition
and productivity of the benthic community.
The assessment endpoints that are the subject of this section are benthic invertebrate community
structure, survival, growth, and reproduction. The measurement endpoints used to evaluate the
assessment endpoint are presented below.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-7
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 The approach used to characterize risks to benthic invertebrates was based upon evaluation of
19 numerous data sources, many of which support the Sediment Quality Triad approach. The
20 Sediment Quality Triad approach to assessment of sediment quality is based on synoptic
21 measurement of sediment chemistry, site-specific sediment toxicity, and benthic invertebrate
22 community structure (Long and Chapman 1985; Chapman 1996).
23
24
25
26
27
28
29
30
31
32
33
34
35
36 A summary of the studies conducted and their linkage to the ERA is provided in Figure 3.1-5. In
37 addition to the targeted Sediment Quality Triad studies summarized above, the ERA considered
Measurement Endpoints for Benthic Invertebrates
¦ Determine, based on field studies, the extent to which reductions in benthic
community abundance, biomass, species richness, and other community metrics
have occurred, including species-specific indications of adverse effects.
Determine if these changes can be related to exposure to PCBs or other COCs
in the sediment of the river.
¦ Determine, based on in situ and laboratory toxicity studies performed for this
ERA, the extent to which the exposure to PCBs and other COCs in the river
sediment may result in adverse impacts to survival, growth, and/or reproduction
of representative benthic taxa.
¦ Determine, based on effects information from the literature, the extent to which
the concentrations of PCBs and other COCs in Housatonic River sediment
and/or water may cause adverse impacts to the benthic community.
¦ Determine, based on a combination of in situ tissue measurements and literature
effects values, the extent to which the concentrations of PCBs bioaccumulated in
the tissues of the benthic organisms will cause effects to survival, growth, or
reproduction.
Sediment Quality Triad Components Investigated in this Study
Standard Triad Components:
¦ Site-specific toxicity studies (laboratory and in situ); multiple species (Hyalella
magna, Chironomus tentans, Daphnia magna, Lumbriculus variegatus), multiple
test durations (48-hour, 7-day, 10-day, 42-day).
¦ Benthic macroinvertebrate community structure.
¦ Abiotic media chemistry (sediment, overlying water, and porewater).
Additional Components:
¦ Bioaccumulation assessment (chemistry in resident invertebrates [predators and
shredders]; 7-day bioaccumulation assessment to deposit-feeding invertebrates
in laboratory [oligochaete]).
¦ Toxicity Identification Evaluations (TIEs).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-8
-------
1 broader site characterization information, such as PCB concentrations in surface sediment (0 to
2 6 inches [15 cm]), ecological characterization studies (Appendix A), and literature information
3 on the potential toxicity of COCs.
4
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-9
-------
2 Figure 3.1-5 Summary of Studies Conducted in Conjunction with Ecological Risk
3 Assessment for Benthic Invertebrates, and Linkage to ERA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-10
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
3.2 EXPOSURE ASSESSMENT
The exposure assessment estimates the exposure of benthic invertebrates to tPCBs and other
COCs in the Housatonic River PSA (Figure 3.1-2). The exposure assessment for benthic
invertebrates also considered the influences of habitat and sediment substrate, and assessed the
degree to which exposure data can be appropriately linked to biological effects studies. Unlike
higher trophic level receptors (Sections 6 - 11), a complex exposure model was not required.
Instead, exposures were assessed as either the COC concentrations in abiotic site media (i.e.,
sediment, water), or as the tissue body burdens that represent integrated exposure from all
sources.
To match exposure data with effects-based measures, many of the data considered were derived
from sampling conducted in association with the 13 benthic community sampling locations
and/or the 7 sediment toxicity locations (Figure 3.2-1). For the purposes of this report, the 13
benthic macroinvertebrate sampling stations are referenced using the "simplified IDs" presented
below, rather than the more complex field sampling IDs.
Summary of IDs for 13 Benthic Invertebrate Sampling Locations
¦ Upstream Reference Locations: A1, A2, A3 (arranged north to south).
¦ Exposed Locations on Housatonic River: 1 to 9 (arranged north to south).
¦ Downstream (watershed) Reference Location: R4 (Threemile Pond).
Exposure assessments were also undertaken for abiotic media at a broader scale than the stations
shown in Figure 3.2-1, such that findings from the Sediment Quality Triad study could be
extrapolated to the larger PSA and Rest of River areas. These extrapolations relied on the
development of exposure-response relationships from the Sediment Quality Triad stations.
3.2.1 Selection of COCs for Benthic Invertebrates
The contaminants initially considered in the benthic invertebrate exposure assessment (COPCs)
were identified in the Pre-ERA (Appendix B). The invertebrate Pre-ERA included screening on
a reach-by-reach basis and subdivision of COPCs by major hydrological/geomorphological
category.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ -> -> 7/10/2003
-------
oS j
A2
River Mile: N/A
HW
-SE000161
A
A3
River Mile: N/A
HW-SE000398
2
River Mile: 133.79
H3-SE000060
i
A
4
River Mile: 132.34 "
H3-SEEC0019
or
Reai
/20\
) / 0 \ —v n .
6
River Mile: 128.7
H3-SE000116
x/v/ >
o
(each 5E
(Reach 5C
8A
River Mile: 125.65
H3-SEEC0023
Reach 51
(backwater areas)\
8
River Mile: 125.65
H3-SEEC0031
Reach 6
/
Approximately 13 miles,
ririse'
jch 1
LEGEND. ^ Benthic Community Stations Without Toxicity Testing
A Benthic Community Stations With Toxicity Testing
Reach Breaks
H Roads
[^j Hydrography
I | Housatonic River Basin Boundary
Notes: - Code in left box of each ID represents "simplified" nomenclature used for benthic invertebrate ERA.
- Station 8A was positioned 12 meters from Station 8, and was tested for laboratory toxicity only.
N
V
s
2500
0
2500
5000 Feet
0.5
0
0.5
1 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 3.2-1
BENTHIC INVERTEBRATE
SAMPLING LOCATIONS
AND SIMPLIFIED STATION
IDENTIFIERS
| O:\gepitt\aprs\macro_int.apr I layout - mac Iocs 3.2-11 o:\gepitt\epsfiles\plots\in\mac_int_3-2-1 .eps 111:42 AM, 7/8/20031
-------
1 The sediment COCs identified in "main channel and aggrading bar" sediment are presented
2 below. PCBs were identified as sediment COCs in all PSA reaches. PAHs were retained
3 throughout the PSA, although the number of individual PAH compounds screened was greater
4 for Reaches 5A and 5C, relative to the other reaches. Dibenzofuran was retained only for Reach
5 5A. Metals were retained in Reaches 5C and 6 only.
6
7
8
9
10
11
12
13
14 Surface water COPCs identified in the pre-ERA (Appendix B) included dioxins/furans, PCBs,
15 and silver. Therefore, the water chemistry screening did not result in any additional
16 contaminants that were not already considered as sediment COCs for invertebrates.
17 Several additional contaminants (mainly pesticides) were determined in the pre-ERA to be below
18 detection limits in sediment, but had detection limits that exceeded screening benchmarks. An
19 examination of the detection limits for invertebrate tissues indicated that concentrations for most
20 of the pesticides of concern were below detection limits. Tissue effects concentration data were
21 available for many pesticides, and screening of maximum observed tissue concentrations showed
22 that even the contaminants detected in invertebrate tissues were below concentrations shown to
23 be of ecological concern. On this basis, and considering that some pesticide detections may be
24 attributable to laboratory interference artifacts, the entire suite of organochlorine pesticides was
25 eliminated from further consideration in the invertebrate portion of the ERA.
26 3.2.2 Types of Exposure Data
27 The approach used to characterize exposure to benthic invertebrates was based upon evaluation
28 of numerous data sources, including both sediment and water column COC concentrations and
Contaminants of Concern for Benthic Invertebrates
¦ Chlorinated organic compounds - tPCBs, dioxins/furans.
¦ Metals - antimony, barium, cadmium, chromium, copper, lead, mercury, silver,
and tin.
¦ Semivolatile organic compounds (SVOCs) - dibenzofuran.
¦ PAHs - numerous individual PAH compounds, including low- and high-molecular
weight PAHs.
MK01|O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
invertebrate tissue COC concentrations (Figure 3.1-2). Many of the data applicable to benthic
endpoints relate to the Sediment Quality Triad.
Use of the Sediment Quality Triad approach requires selection of exposure data to achieve
maximum correspondence between exposure and effects endpoints, while also recognizing the
need to address spatial and temporal variability in the data, and inherent limitations in field
sampling. Concentration-response relationships were investigated using both: (a) a combined
data set, screened using the criteria listed below; and (b) the single "most synoptic" chemistry
value paired with each toxicity endpoint. Section 3 and Appendix D present the results of the
first method, and Attachment D.5 presents an analysis using the latter method, for comparison.
In most cases, the two approaches yielded similar results, demonstrating that the interpretations
in the risk assessment were not an artifact of the data processing methods.
Criteria for Selection of Exposure Data Linked to Toxicity Endpoints
¦ Sediment data collected within a radius of 5 m from the benthic biota sampling
location.
¦ Exposure data collected between March and October 1999, the period over
which all site-specific effects measurements were performed.
¦ Sediment samples collected from within the top 6 inches (15 cm) of sediment,
and included at least the top 2 inches (5 cm).
¦ In merging data sets from multiple studies, exposure data were combined (using
a median of individual data) for all points collected on the same calendar day at
the same location. This was done to avoid bias resulting from a higher number
of samples or replicates on one day, which would have potentially obscured
temporal variability.
Because the sediment PCB concentration data were lognormally distributed, the median was
chosen as the measure of central tendency for use in concentration-response assessments. Unlike
wildlife ERA components, the benthos were assumed to be relatively sessile, and therefore,
integrating exposures over a home range (i.e., using arithmetic mean to integrate concentrations
as they vary over space) was not appropriate.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-14
-------
1 3.2.3 Habitat Characterization
2 To provide the foundation for the risk characterization, the preliminary results of physical and
3 ecological investigations were used to define appropriate "clustering" of benthic stations for the
4 exposure assessment. In this case, clustering refers to the grouping of stations of similar
5 properties (e.g., sediment types), to discriminate between substrate-related responses and those
6 attributable to contaminants.
7
8
9
10
11
12
13
14
15
16 Two approaches were used to evaluate the benthic sampling stations in terms of their gross
17 habitat characteristics:
18 ¦ Physical substrate variables were evaluated within and among benthic sampling
19 stations downstream of identified contaminant sources (Figure 3.2-2) to identify
20 significant break points in physical habitat features (e.g., substrate type, organic
21 carbon content, sediment particle size distribution).
22 ¦ Aquatic habitat was evaluated to identify broad biological regimes within the benthic
23 sampling locations. Factors such as surrounding vegetation, macrophyte coverage,
24 and cluster analysis of benthic invertebrate assemblages were used to identify
25 changes in macro-habitats.
26 Considering the above, a clear change in substrate and habitat type was observed between
27 Stations 5 and 6, which is coincident with the transition in river regime from Reach 5A to Reach
28 5B, and the location of the Pittsfield wastewater treatment plant (WWTP) outfall. This shift in
29 habitat (e.g., particle size distributions and organic carbon content of sediment) was used to
30 identify appropriate statistical contrasts in the ERA.
Rationale for "Clustering" of Benthic Sampling Stations in ERA
Clustering was relevant for the exposure assessment for three primary reasons:
¦ To determine the appropriateness of reference stations for making statistical
comparisons to exposed stations.
¦ To provide a means of separating physical and ecological "regimes" in a manner
consistent with both the exposure and effects assessment.
¦ To provide a tool to assess if there are other influences that were not well
characterized.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-15
-------
100
80 --
0)
¦- 60 +
0)
g
0)
Q.
"D
0)
40 --
20 --
o
¦A-A ,0,S,6,6
A1 A2 A3 1 2 3 4
o
o
o
S o
o
10%
8%
6% o
o
4%
2%
0%
¦o
0)
R4
• Percent Fines
1
2
3
4
5
6
O Total Organic Carbon
Note: Stations Al, A2, A3, and R4 represent less contaminated "reference" areas.
Figure 3.2-2 Median Percent Fines and Percent TOC by Sampling Location, for
Benthic Community Grab Samples
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-16
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Grouping of Benthic Sampling Stations Based on Habitat
Benthic sampling stations in Figure 3.2-1 were assigned to one of the following
categories:
¦ "Coarser" Reference Locations (C/R) - Low total organic carbon (TOC) (typically
less than 1%), sandy sediment found either upstream of influence from the GE
facility or on the West Branch. Three locations (A1, A2, A3).
¦ "Coarser" Contaminated Locations (C/C) - Low TOC (typically less than 1%),
sandy sediment found between the confluence and the Pittsfield VWVTP. Five
locations (1, 2, 3, 4, 5).
¦ "Finer" Reference Locations (F/R) - High TOC, silty sediment found outside the
PSA at Threemile Pond (Location R4).
¦ "Finer" Contaminated Locations (F/C) - High TOC (typically a few percent or
greater), silty sediment found downstream of the Pittsfield VWVTP. Four
locations (6, 7, 8, 9).
3.2.4 Assessment of Sediment Chemistry
3.2.4.1 Sources of Sediment Data
There are multiple sources of sediment data, each with a varying degree of correspondence to
various effects metrics. The use of specific data sets depended on the ERA goal; for example,
"discrete sampling" data not associated with benthic sampling stations were used to extrapolate
risk estimates, but were not used for development of concentration-response relationships.
Sediment Data Sources Used in Benthic Invertebrate ERA
¦ Benthic Community Grab Samples (1999) - 12 replicate samples taken at each
of 13 stations and analyzed for PCBs and other parameters; synoptic with
benthic community structure samples.
¦ Laboratory Toxicity Samples (1999) - Composite samples (mixture of five cores)
collected; samples were submitted for laboratory toxicity tests and analyses
including tPCBs and TOC.
¦ In situ Toxicity Samples - Composite samples taken in a similar fashion to the
laboratory samples; and at different time periods (i.e., end of 48-hour, 7-day, and
10-day exposure periods). All analyzed for tPCBs; 7-day samples also analyzed
for PCB congeners, PAHs, pesticides, metals, chlorinated benzenes, and TOC.
¦ Other Sediment Samples - Included discrete river sampling, screening samples
prior to the biological investigations, and collected with supplemental studies,
such as the Wright State University (WSU) TIE investigation. Other data not
considered synoptic were used only for screening, characterizing generic
exposures in the PSA, or for extrapolating concentration-response relationships.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-17
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
3.2.4.2 Distribution and Concentrations of PCBs
3.2.4.2.1 Benthic Community Grabs
Individual replicate concentrations of tPCBs for each benthic sampling station are presented on a
logarithmic scale in Figure 3.2-3. The data indicate highly elevated tPCB concentrations in the
C/C sites, with median values of approximately 5 to 25 mg/kg. In the F/C sites, median PCB
concentrations were significantly lower (pooled variance t-test; p <0.001). There was
considerable variability in tPCB concentrations between replicates at most stations (Figure
3.2-3), indicative of small-scale variability in PCB concentrations.
3.2.4.2.2 Toxicity Test Samples
Concentrations of tPCBs were measured in Housatonic River sediment in conjunction with
laboratory and in situ toxicity and bioaccumulation tests conducted between May and July 1999
(EVS 2003). As with the benthic community grab samples, tPCB concentrations were quite
variable within stations across the four toxicity sampling events (Figure 3.2-4).
These results suggest that PCB exposure data from the toxicity testing data sets should not be
extrapolated to benthic community composition endpoints, and also indicate that the variability
in the chemistry data associated with the toxicity program must be considered when deriving
concentration-response relationships. Because sediment samples were not replicated in
individual toxicity sampling events, data from all relevant sampling events (as defined in Section
3.2.2) were included in the development of concentration-response relationships for toxicity
endpoints.
3.2.4.2.3 Broad Scale Sediment Characterization
Figure 3.2-5 depicts the spatial distribution of tPCB concentrations within the PSA. The data
indicate that median PCB concentrations are highest in the upstream reaches of the PSA, and
decrease with distance from the GE facility. The median concentrations are lowest just
downstream of the WWTP, but increase moving farther downstream to Woods Pond. For areas
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-18
-------
1000
o>
O)
E
m
o
Q.
o
0)
O
C
o
O
100
10
0.1
A
I
&
t
A
s
t I
t
t
A
I
{ *
s
i
A1 A2 A3
3 4 5
Location ID
4
t
R4
A Median of 12 replicates
— Mean of 12 replicates
• Replicate concentration, assuming non-detected values equal to
half MDL
Figure 3.2-3 Concentrations of tPCBs in Sediment by Sampling Location for
Individual Benthic Community Grab Samples, and Associated Measures of
Central Tendency
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-19
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
1000
G)
O)
E
ca
o
o.
Is
o
100
10
0.1
0.01
X
o
A1
~
0
£
x
+
A
O
*
•_
A
~
X
A3
4 5 7
Station Location
X
o
8A
A48h in situ WSU toxicity (June 1999)
~ 7-day in situ WSU toxicity (June 1999)
• 10-day in situ WSU toxicity (June 1999)
O Laboratory Toxicity - WSU (May 1999)
XBenthic Macroinvertebrate (Median of 12; June 1999)
O Porewater TIE (AD) - WSU (Sept. 1999)
X Porewater TIE (AW) - WSU (Sept. 1999)
~ Mussel Study (May 1999)
+ Mussel Study (July 1999)
— Discrete River Sampling (October 1999)
APre-Toxicity Screening (March-May 1999)
Figure 3.2-4 Comparison of tPCB Concentrations in Sediment Collected at
Benthic Toxicity Sampling Locations, from Various Sampling Efforts Conducted
in 1999
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-20
-------
1000
100
g—i—i—i—|—i—i—i—r
E Reach 5A, N = 499
-i—r
-i—i—r
1—1—i—i—I—i—r-
Reach 5B, N = 218
—i—i—i—|—i—r
Reach 5C, N = 286
Reach 5D, N = 63
"I—
—I 1 |"T
Reach 6,
N = 225
O)
*
O)
E
CD
O
D.
10
0.1
= Total PCB
32 45 22 30 54 52 23 37 14 24
J I I J J I I J L
24 14 39 12 21 16 10
J L L I I—
5 36
-I L
31 12 19 34 ,38 39 20 25
1 I I I—I I I l_
15 23 14 9 28 18 21 42 23 19 42 21 11,
I I I I J I J I J I I I I
135
1000000
134
133
132
131
130
129
128
127
126
125
124
|—i—i—i—|—i—i—i—p
E Reach 5A, N = 472
T—I—I—I—I—T
1—1—i—i—I—i—r-
Reach 5B, N = 207
1—'—'—1—I—I—r-
Reach 5C, N = 288
Reach 5D, N = 69
"I—
—i—ttt
Reach B,
N = 174
100000
O)
*
O)
c 10000
•=- I i ; l
o
O
1000
100
1
: Total Organic Carbon
y 2:
73 40 21 28 53
26 36
_l L_
23 15 43
_1_
15 13 11
_l I L.
6 37
-I L
12 20 33 34 42 20
J I I I I I
|3 21 13 9^ 2^
18 20
—I I-
24 51 15
L
5 1^
135
100000
134
133
132
131
130
129
128
127
126
125
124
i—1—1—1—|—1—1—1—|—1—1—1—|—1—1—1—|—1—1—r
: Reach 5A, N = 440
1—1—1—1—I—1—r-
Reach 5B, N = 190
1—'—r-
N = 243
N = 63
1 <~
o
10000
O)
*
O) 1000
o
a.
100
10
Reach 5C,
Reach 5D,
: Carbon Normalized PCB
72 40 21 25 49 45 23 34 13 22 14 38 6 14 13 10
J I I I I I I I I I I I I I l_
4 34
_l L
29 11 17 31
1 I L I L
32 38 19 13
J I I
19 11
I l_
9 22 17 1
Reach S,
N = 154
17 40 11
I I L
135 134
LEGEND:
River Reach Median
River Reach Quartile
— Sub-reach Median and Quartile
I Reach 5D Median and Quartile
133
132
131
130
129
128
127
126
125
124
Mile Point (miles)
0.25 MILE SUB-REACH AVERAGING
Median and Quartiles of PCB and TOC in the Housatonic River (NOTE: Reach 5D is Backwaters)
(All Data, Half DL, Surface Layer: 0 - 6")
Figure 3.2-5 Medians and Quartiles of PCB and TOC in the Housatonic River PSA, Subdivided by River Reach
and 0.25 Mile Subreaches
MK0110:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
downstream of Woods Pond, the tPCB concentrations are lower than in the PSA (by
approximately an order of magnitude). Sediment tPCB concentrations downstream of the
Connecticut border were generally below 1 mg/kg, reflecting the general trend of decreasing
concentration with distance downstream.
3.2.4.3 Distribution and Concentration of Other COCs
Fewer sediment data are available for other COCs; however, there were sufficient data to
characterize patterns of concentrations throughout the PSA (Appendix D; Figures D.2-16
through D.2-28). A brief summary of spatial trends for these COCs is provided below.
Summary of Trends in Other COCs in Sediment Data
¦ Dioxin/furan concentrations are elevated at downstream fine-grained locations
relative to upstream and/or reference locations, and are positively correlated with
tPCB concentrations.
¦ Dibenzofuran concentrations do not occur in a pronounced spatial pattern
throughout the PSA; most concentrations were in the 0.1 to 1.0 mg/kg range.
This COC was eliminated from the ERA (rationale provided in Appendix D).
¦ Total PAH concentrations were highly variable, both spatially and between
sampling events. The median concentrations were greatest near the urbanized
areas of the Housatonic River watershed, and were lowest at Station A1 and at
the Woods Pond headwaters. The broad spatial pattern of PAH concentrations
in the toxicity locations was opposite to that observed for PCBs.
¦ Metals concentrations were typically lower at upstream sites relative to the fine-
grained sediment found downstream. Different metals had generally similar
concentration patterns. Metals concentrations were significantly correlated
(p <0.05) with TOC concentrations.
3.2.5 Tissue Chemistry Assessment
Benthic tissue data are less abundant than data for abiotic media, and generally did not include
replication, due to limited volumes of tissue available for chemical analysis. Nevertheless, the
available data provide a measure of the site-specific bioavailability of the COCs.
Figure 3.2-6 presents the distribution of tPCB concentrations by sampling location and tissue
type, for samples collected at the benthic sampling stations. Most reference samples had tPCB
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-22
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
50
40
30
o>
O)
E
8 20
Q.
10
0.30/0.22 0 35
0 L.
A1 A3 1
j ^ 0.05/0.50
8 9 R4
Predator
~ Shredder
Note: Text labels indicate detected values close to zero. Missing values with no text labels indicate
stations where no analysis was completed due to insufficient sample volume.
Figure 3.2-6 Concentrations of tPCBs in Benthic Invertebrate Tissues by Location
and Functional Feeding Group
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-23
-------
1 tissue concentrations below 1.0 mg/kg. In contrast, contaminated locations had elevated
2 concentrations ranging from 2 to 48 mg/kg.
3
4
5
6
7
8
9
10
11
12
13
14 Concentrations of other COCs, such as Appendix IX pesticides, are also available in the benthic
15 tissue chemistry data set. However, because pesticides were screened out of the benthic ERA
16 (based on detection limit considerations, conservative tissue concentration screening, and
17 potential for artificially high laboratory values due to PCB interference), these contaminants
18 were not considered further. Although PAHs and metals were retained as COCs, the tissue
19 analyses did not include these parameters due to the lack the sufficient sample volume for
20 analysis.
21 3.2.6 Surface Water Chemistry Assessment
22 Surface water chemistry data have limited application to the benthic ERA due to the uncertainty
23 in extrapolating from water chemistry to effects in sediment-dwelling biota. Because of this
24 uncertainty, the only data considered relevant were those collected synoptic with effects
25 measurements. Unfiltered overlying site water was collected in conjunction with the 7-day in
26 situ toxicity testing (EVS 2003) and evaluated for tPCBs, PCB congeners, PAHs, pesticides, and
27 metals. Only tPCBs were measured for the 48-hour and 10-day exposures.
Sources of Benthic Invertebrate Tissue Chemistry Data
¦ Analysis of composite samples of "predators" and "shredders," respectively,
conducted in 1999 by EPA. Each tissue sample was analyzed for lipids and a
number of organic contaminants, including PCBs (as congeners, as Aroclors,
and as tPCBs), dioxins/furans, and pesticides.
¦ Data from the 7-day in situ bioaccumulation study (EVS 2003) conducted with the
oligochaete worm Lumbriculus variegatus at 6 sampling stations.
¦ Academy of Natural Sciences of Philadelphia long-term historical tissue PCB
monitoring of dobsonfly, caddisfly, and stonefly nymphs/larvae collected
downstream of the PSA near Cornwall, CT (BBL & QEA 2003).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-24
-------
1
2
3
4
5
6
7
8
9
10
Summary of Water Chemistry COC Concentrations
¦ Concentrations of tPCBs at upstream reference stations were less than 10
nanograms per liter (ng/L); concentrations at contaminated stations ranged from
approximately 100 ng/L to 300 ng/L.
¦ Concentrations of furans matched the spatial patterns in tPCBs, with total
detected furans well below 10 picograms per liter (pg/L) at upstream reference
stations, and concentrations of 60 to 120 pg/L at contaminated stations.
¦ No dibenzo-p-dioxins or silver were detected in the samples collected in
conjunction with the 7-day in situ tests.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-25
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
3.3 EFFECTS ASSESSMENT
The effects assessment for benthic invertebrates (Figure 3.1-3) emphasizes the site-specific
biological investigations performed at the 13 benthic sampling stations because these studies
provided direct indications of the bioavailability, toxicity, and effects of site COCs. Both
toxicity assessments (i.e., laboratory toxicity, in situ toxicity, and TIE) and community
evaluations (i.e., benthic macroinvertebrate community composition) were compared to
appropriate field references to determine whether the exposed sites on the Housatonic River
exhibited biological impairment.
The effects assessment also provides an overview of the literature on the effects of tPCBs and
other COCs to survival, growth, and reproduction of benthic invertebrates. Studies were
screened and used to derive the most appropriate effects metrics for tissue, sediment, and water.
In recognition of the uncertainty inherent in threshold effects concentrations for these media,
ranges of benchmarks were derived instead of relying on single effects thresholds.
Detailed evaluations of concentration-response relationships are discussed immediately
following the broad discussion of inter-station differences (i.e., differences between exposure
location responses and control and/or reference location responses). This corresponds to
Sections 3.3.2 and 3.3.7 for sediment toxicity and benthic community structure, respectively.
The effects thresholds derived therein are carried forward into the Risk Characterization, and are
used as maximum acceptable threshold concentrations (MATCs) for extrapolation to areas
downstream of the PSA.
3.3.1 Sediment Toxicity
3.3.1.1 Methods
Wright State University (WSU) conducted site-specific toxicity testing of Housatonic River
sediment (EVS 2003). Test protocols, study methods, and other detailed documentation for the
Housatonic River sediment toxicity testing program are presented in EVS (2003) and in the
SIWP (WESTON 2000). Seven stations were sampled for in situ and laboratory toxicity
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-26
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
analyses. Of these, six were within the group of 13 locations where the benthic
macroinvertebrate sampling was conducted (Figure 3.2-1).
Toxicity Test Methods
Laboratory testing (EPA 2000 protocols')
¦ Chronic 42-day bulk sediment test using a freshwater amphipod (Hyalella
azteca). Duration and endpoints were 28 days, 35 days, and 42 days for
survival; 28 days and 42 days for growth; 35 days and 42 days for reproduction.
¦ Chronic 43-day sediment test using a freshwater midge (Chironomus tentans).
Duration and endpoints were 20 days for survival and growth, and 23 days to 43
days for mortality and emergence.
In situ testing (including both sediment and water-only exposures'):
¦ 48-hour toxicity test using a freshwater cladoceran (Daphnia magna, 48 hours
old). Endpoint: survival.
¦ 48-hour and 10-day toxicity test using a freshwater midge (Chironomus tentans,
8 to 12 days post-hatch). Endpoint: survival.
¦ 48-hour and 7-day toxicity test using a freshwater oligochaete worm (Lumbriculus
variegatus, multiple ages). Endpoint: survival and tissue bioaccumulation.
¦ 48-hour and 10-day toxicity test using a freshwater amphipod (Hyalella azteca, 1
to 14 days old). Endpoint: survival.
The organisms selected are both environmentally relevant to the Housatonic River (e.g.,
chironomid and oligochaete species were present in high numbers in PSA sediment) and have a
large toxicological database demonstrating their relative sensitivity to the COCs. The selected
species inhabit sediment during the life stages tested, remain relatively immobile, and have a
high potential for exposure. The test organisms selected are tolerant to a broad range of
sediment physicochemical characteristics (EPA 2000).
Sediment grain size is a significant variable that can affect benthic organisms, both in the field
and in some laboratory toxicity studies. Because the field reference sediment (Stations A1 and
A3) both represented coarse-grained sediment, it was important to assess the potential for
confounding effects of particle sizes in toxicity test treatments with fine-grained sediment. To
this end, a literature review was conducted to document the sensitivity of the test organisms to
changes in particle size distributions (Attachment D.4). The review indicated that the indicator
species chosen for toxicity testing are quite tolerant of a broad range of particle sizes (hence their
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC 1 01 7/10/2003
-------
1 selection as measurement endpoints). In addition, the in situ tests were conducted using
2 sediment exposure chambers placed immediately on top of the sediment. The organisms were
3 separated from the sediment by a fine-mesh screen and therefore not affected by the sediment
4 grain size distribution. Therefore, comparison of all contaminated sediment treatments to the
5 reference sediment A1 and A3 was appropriate.
6 3.3.1.2 Results
7 Table 3.3-1 presents the results of statistical tests of significance (comparisons to negative
8 controls and reference stations) for most toxicity test endpoints. Toxicological responses for
9 each test type and treatment are also presented graphically (Figure 3.3-1 to Figure 3.3-11). PCB
10 concentrations are presented in two ways, based on the two data processing approaches discussed
11 in the exposure assessment:
12 ¦ The values in bold represent the median of all spatially and temporally relevant PCB
13 concentrations at each station; and
14 ¦ The values in italics represent the single PCB concentration measurement taken
15 closest to the effects endpoint (i.e., most synoptic concentration).
16 The station-by-station summary assessment is presented in Tables 3.3-2 and 3.3-3 (in situ and
17 laboratory endpoints, respectively). Each endpoint was rated as exhibiting negligible, moderate,
18 strong, or very strong evidence for toxicological effects to freshwater organisms, based on the
19 effect magnitude observed relative to the negative control(s). The degree of confidence in the
20 assessment of potential for ecological risk is in large part a function of the degree of concordance
21 observed among endpoints.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-28
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Summary of Site-Specific Toxicity Outcomes
In situ exposures (acute mortality endpoints):
¦ Negligible toxicity at both reference locations (Stations A1, A3) and the most
upstream "contaminated" location with the lowest PCB concentration (Station 4).
¦ Toxicity was evident in multiple tests for the remaining three contaminated
locations (Stations 5, 7, 8), with the magnitude of response generally greatest at
the two locations with the highest PCB concentrations.
¦ Modest effects were observed in some water-column exposures, but most
pronounced effects were observed in the sediment-exposure treatments.
Laboratory exposures (chronic lethal and sublethal endpoints):
¦ Overall frequency of toxic responses greater than for acute in situ exposures.
¦ Minor indications of reduced endpoint performance (relative to negative control)
in both reference locations (Stations A1, A3). These responses consisted mainly
of marginal reductions in Hyalella reproduction and Chironomus endpoints.
¦ Large magnitude adverse responses were much greater at all four contaminated
locations (e.g., Chironomus toxicity, Hyalella mortality, and reproductive effects).
¦ Conclusion of a "high" toxicity rating for all four contaminated locations.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-29
-------
Table 3.3-1
Results of Pairwise Statistical Tests Comparing Exposed Stations to Negative Control (T-Ctrl) and Reference
(A1, A3) Sediment (Water-Only Exposures Excluded)
Station
4
5
7
8A
8
Pairwise Comparison
T-Ctri
A1
A3
T-Ctrl
A1
A3
T-Ctrl
A1
A3
T-Ctrl
A1
A3
T-Ctrl
A1
A3
28-d Hyalella survival
Yes
Yes
Yes
N/A
N/A
N/A
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
35-d Hyalella survival
Yes
Yes
Yes
N/A
N/A
N/A
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
42-d Hyalella survival
Yes
No
Yes
N/A
N/A
N/A
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
28-d Hyalella dry weight
No
No
No
N/A
N/A
N/A
NC
NC
NC
No
No
No
No
No
No
42-d Hyalella dry weight
No
No
No
N/A
N/A
N/A
NC
NC
NC
No
No
No
No
No
No
42-d Hyalella young per
female
Yes
No
No
N/A
N/A
N/A
NC
NC
NC
Yes
No
Yes
Yes
Yes
Yes
42-d Hyalella mean young
Yes
No
Yes
N/A
N/A
N/A
NC
NC
NC
Yes
Yes
Yes
Yes
Yes
Yes
20-d Chironomus survival
Yes
Yes
Yes
N/A
N/A
N/A
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
43-d Chironomus
emergence
Yes
Yes
Yes
N/A
N/A
N/A
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
20-d Chironomus dry
weight
Yes
Yes
Yes
N/A
N/A
N/A
NC
NC
NC
NC
NC
NC
Yes
Yes
Yes
20-d Chironomus ash-free
dry weight
Yes
Yes
Yes
N/A
N/A
N/A
NC
NC
NC
NC
NC
NC
Yes
Yes
Yes
48-h Hyalella survival
(sediment)
No
No
No
No
No
No
No
Yes
No
N/A
N/A
N/A
No
No
No
10-d Hyalella survival
(sediment)
No
No
No
No
No
No
No
Yes
Yes
N/A
N/A
N/A
No
Yes
Yes
48-h Chironomus survival
(sediment)
No
No
No
No
No
No
No
No
No
N/A
N/A
N/A
No
No
No
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ ^H 7/10/2003
-------
Table 3.3-1
Results of Pairwise Statistical Tests Comparing Exposed Stations to Negative Control (T-Ctrl) and Reference
(A1, A3) Sediment (Water-Only Exposures Excluded)
(Continued)
Station
4
5
7
8A
8
Pairwise Comparison
T-Ctri
A1
A3
T-Ctrl
A1
A3
T-Ctrl
A1
A3
T-Ctrl
A1
A3
T-Ctrl
A1
A3
10-d Chironomus survival
(sediment)
No
No
No
No
No
No
Yes
Yes
Yes
N/A
N/A
N/A
Yes
Yes
Yes
48-h Daphnia survival
(sediment)
No
Yes
No
No
No
No
Yes
Yes
Yes
N/A
N/A
N/A
Yes
Yes
Yes
48-h Lumbriculus survival
(sediment)
No
No
No
No
No
No
No
No
No
N/A
N/A
N/A
No
No
No
1 Yes = Statistically different at alpha = 0.05
2 No = Not statistically different at alpha = 0.05
3 NC = Sublethal endpoint not calculable (due to zero survival in treatment)
4 N/A = Not applicable; not tested for endpoint/station combination
5
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
0.28
100
Q
CO
^ 80
re
>
NA
3
(A
-4-i
C
a>
*2
a>
o.
c
re
a>
60
40
20
0.018
0.028
0.28
T-Control
A1
A3
54
213
4.6
31
8A
77
72
iml
Station Location
28-day (24 June 99) ~ 35-day (1 July 99) ~ 42-day (8 July 99)
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
Figure 3.3-1 Survival of Hyalella azteca in Chronic Laboratory Toxicity Tests,
at Three Time Periods (28 days, 35 days, 42 days)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-32
-------
1
T-Control A1 A3 4 7
Station Location
8A
28-day (24 June 99)
~ 42-day (8 July 99)
4
5
6
7
9
10
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
Figure 3.3-2 Growth of Hyalella azteca in Chronic Laboratory Toxicity Tests, at
Two Time Periods (28 days, 42 days)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-33
-------
1
2
3
4
5
6
7
8
9
10
11
80
o
CO
70
+
' '
(A
60
a>
13
c
50
o
a>
c
40
o
a>
30
.Q
E
3
20
C
C
re
10
a>
S
0
NA
UL
T-Control
0.018
0.028
A
A1
0.28
0.28
1
¦
5.9
8.7
54
213 46
(no reproduction 31
data due to zero
survival rate)
A3 4 7
Station Location
8A
77
72
28-42 day mean young per female ~ 28-42 day mean young (unstandardized)
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
Figure 3.3-3 Reproduction of Hyalella azteca in Chronic Laboratory Toxicity
Tests, Based on Mean Number of Young (Days 28-42)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-34
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Q 100
CO
+
a>
a
c
a>
E>
a>
E
LLJ
re
>
E
3
CO
80
60
40
20
T-Control C- F-
Control Control
A1
A3
54
273
4.6
31
8A
77
72
A
Station Location
120-day Survival ~ 20-day Emergence
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
T-Control, C-Control, and F-Control are negative laboratory controls ("Trout Farm", "Cellulose",
and "Florissant", respectively).
Figure 3.3-4 Survival and Emergence of Chironomus tentans in Chronic
Laboratory Toxicity Tests (43 days)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-35
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
5.0
CO
+ 4.0 -
T- C- F- A1 A3 4 7 8A 8
Control Control Control
Station Location
¦ Dry weight ~ Ash-free Dry weight
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
T-Control, C-Control, and F-Control are negative laboratory controls ("Trout Farm", "Cellulose",
and "Florissant", respectively).
Figure 3.3-5 Growth Endpoints for Chironomus tentans in Chronic Laboratory
Toxicity Test (20 days)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-36
-------
1
2
3
4
5
6
7
8
9
10
11
120
Q
100
CO
+
80
5?
re
>
60
E
D
40
CO
C
re
a>
20
S
0
NA
0.018
0.0001
Lab Control
A1
0.28
0.38
5.9
0.95
8.3
54
A3 4 5
Station Location
77
522
~ Water Column
I Against Sediment
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
Figure 3.3-6 Survival of Hyalella azteca in 48-hour Low Flow In Situ Toxicity
Tests Conducted 14-16 June 1999
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-37
-------
1
Lab Control A1 A3 4 5 7 8
Station Location
~Water Column BAgainst Sediment
2
3 Notes: Labels represent tPCB concentration (mg/kg) in sediment.
4 Value in bold represents median tPCB concentration (from all measurements made within 5 meters
5 of station in 1999; see Appendix D).
6 Value in italics represents "most synoptic" tPCB concentration; single concentration measured
7 closest to toxicity test in space/time.
8 Figure 3.3-7 Survival of Hyalella azteca in 10-day Low Flow In Situ Toxicity
9 Tests Conducted 17-27 June 1999
10
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-38
-------
1
2
3
4
5
6
7
8
9
10
11
120
q 100
CO
+
" 80
™ 60
W 40
c
re
a>
5 20
0.018 Q.28
5.9
NA
0.0001 0.38
0.95 83
54.1
77
522
Lab Control A1 A3 4 5
Station Location
~ Water Column
I Against Sediment
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
Figure 3.3-8 Survival of Chironomus tentans in 48-hour Low Flow In Situ
Toxicity Tests Conducted 14-16 June 1999
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-39
-------
1
2
3
4
5
6
7
8
9
10
11
120
q 100
CO
+
80
™ 60
W 40
c
re
-------
1
2
3
4
5
6
7
8
9
10
11
0.018 0.28 5.9 8.3 54 77
0.0001 0.38j 0.95 7.3 -j^q 522
mm
Lab Control A1 A3 4 5 7 8
Station Location
~ Water Column ¦ Against Sediment
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
Figure 3.3-10 Survival of Daphnia magna in 48-hour Low Flow In Situ Toxicity
Tests Conducted 14-16 June 1999
Q 100
CO
+
re
>
E
3
CO
C
re
a>
80
60
40
20
NA
I
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-41
-------
1
2
3
4
5
6
7
8
9
10
11
120
q 100
CO
+
" 80
re
>
E
3
CO
C
re
a>
60
40
20
AM
X.
0.028
0.0071
0-28
5.9
Lab Control A1
A3 4 5
Station Location
~ Water Column
I Against Sediment
Notes: Labels represent tPCB concentration (mg/kg) in sediment.
Value in bold represents median tPCB concentration (from all measurements made within 5 meters
of station in 1999; see Appendix D).
Value in italics represents "most synoptic" tPCB concentration; single concentration measured
closest to toxicity test in space/time.
Figure 3.3-11 Survival of Lumbriculus variegatus in 48-hour Low Flow In Situ
Toxicity Tests Conducted 14-16 June 1999
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-42
-------
1
2
3
Table 3.3-2
In Situ Evaluation of Toxicity in Housatonic River Sediment (Station-by-Station Assessment)
Sampling Station
(ID, Location, WESTON ID)
Median
Bulk
Sediment
[PCB]
(mg/kg)
H. azteca
48-hour
Survival
(Water-
Sediment)
C. tentans
48-hour
Survival
(Water -
Sediment)
L. Varie-
gatus
48-hour
Survival
(Water-
Sediment)
I), magna
48-hour
Survival
(Water-
Sediment)
H. azteca
10-day
Survival
(Water-
Sediment)
C. tentans
10-day
Survival
(Water-
Sediment)
L. varie-
gatus
Residue
(mg/kg
lipid)
Overall
Assessment
A1
Dalton Reference
011
0.018
O-O
O-O
O-O
O-O
O-O
O-O
5.3
O
A3
Lower West
Branch Reference
398
0.28
o-o
O-O
O-O
O-O
O-O
O-O
20.7
O
4
1.5 miles below
Holmes Road
019
5.9
O-O
o-o
o-o
o-o
o-o
o-o
232.6
o
5
Near WWTP
Discharge
428
7.3
o-o
o-o
o-o
o-o
o-o
o-o
380.6
o
7
2 miles below
New Lenox Road
389
54
o-m
o-o
o-o
O-0t
O -
O-0t
128.3
•
8
Vi mile above
Woods Pond
031
77
o-o
o-o
o-o
O-0t
O-0t
O-0t
314.5
•
4 O = Negligible to low toxicity: less than 20% effect size relative to negative control. Overall assessment - negligible indication of ecological risk.
5 O = Moderate toxicity: 20 to 50% effect size relative to negative control. Overall assessment - ecological effects possible, but not conclusive.
6 • = High toxicity; greater than 50% effect size relative to negative control. Overall assessment - strong indication of potential ecological effects.
7 • = Very strong toxic response for individual endpoint; greater than 90% effect size relative to negative control.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-43
-------
1 Table 3.3-3
2
3 Laboratory Evaluation of Toxicity in Housatonic River Sediment (Station-by-Station Assessment)
Sampling Station
(ID, Location, WESTON ID)
Median Bulk
Sediment
[PCB]
(mg/kg)
H. azteca
Survival
(28 day-
35 day -
42 day)
H. azteca
28-42 day
Reproduction
(young/female)
H. azteca
Dry weight
(28-day-
42 day)
C. tentans
Survival
C. tentans
Growth
(Total Wt-
Ash Free)
C. tentans
Emergence
Overall
Assessment
A1
Dalton Reference
011
0.018
o-o-o
O
O
I
o
O
O - •
O
O
A3
Lower West Branch
Reference
398
0.28
o-o-o
O
0
1
o
o
0
1
o
O
o
4
1.5 miles Below Holmes
Road
019
5.9
o-o-o
•
0
1
o
•
•
•
7
2 miles Below New Lenox
Road
389
54
•* - N/A -
N/A
N/A
N/A - N/A
•
N/A
•
•
8
Vi mile above Woods Pond
031
77
•
O
I
o
•
•
•
8A
Vi mile above Woods Pond
023
4.6
•
0
1
o
•
N/A
•
•
4 O = Negligible to low toxicity: less than 20% effect size relative to negative controls. Overall assessment - negligible indication of ecological risk.
5 O = Moderate toxicity: 20 to 50% effect size relative to negative controls. Overall assessment - ecological effects possible, but not conclusive.
6 • = High toxicity; greater than 50% effect size relative to negative controls. Overall assessment - strong indication of potential ecological effects.
7 • = Very strong toxic response for individual endpoint; greater than 90% effect size relative to negative control.
8 N/A = Sublethal endpoint not measured due to complete mortality in treatment.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-44
-------
1 A comparative approach (i.e., relative to reference) was also applied to help distinguish
2 background field reference responses from those observed at contaminated locations. In this
3 assessment, comparisons were made not to the negative control, but to the two upstream
4 reference locations (Stations Al, A3). Despite the modest toxicity observed in the laboratory
5 toxicity endpoints for Stations Al and A3, the comparative assessment (Table 3.3-4) still
6 indicated a moderate to strong incremental toxicity associated with contaminated PSA sediment.
7 The three most downstream stations (7, 8, and 8A) had "high" ratings due to the consistency and
8 severity of toxicity observed for numerous endpoints.
9 3.3.2 Concentration-Response Analysis - Toxicity Test Endpoints
10 A statistical assessment was conducted to quantify the observed relationship between toxicity
11 test endpoints and COC concentrations measured concurrent with the biological tests. The
12 assessment focused on the relationship between PCBs and toxicity endpoints because other lines
13 of evidence indicated a high probability that PCBs were a causal agent for toxicity to benthic
14 invertebrates within the Housatonic River PSA.
15 This section emphasizes concentration-response using the "median" sediment PCB exposure
16 concentration at each station. An alternative analysis, using only the "most synoptic" exposure
17 concentration is presented in Attachment D.5. Generally, the two approaches yield comparable
18 results (i.e., most endpoints within a factor of 2); the "median" analysis yielded effects
19 thresholds that were slightly lower than the other method.
20
21
22
23
24
25
26
Methods for Evaluating Concentration-Response for Toxicity Data
¦ Individual Endpoint Analysis - Each toxicity endpoint was investigated
individually using conventional descriptive statistics that related degree of effect
to PCB concentrations (e.g., LC50, IC2o, NOAEL, LOAEL).
¦ Combined Endpoint Analysis - The toxicological endpoints were integrated using
a general linear modeling approach to identify similarities and differences in
concentration-response relationships across species and endpoints.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-45
-------
1 Table 3.3-4
2
3 Evaluation of Lines of Evidence for Housatonic River Sediment Toxicity, Relative to Reference Responses
Sampling Station
(ID, Location, WESTON
ID)
Chronic Laboratory Endpoints
(20 day, 42 day)
Acute In Situ Endpoints (48 hours, 10 days)
Overall
Assessment
H. azteca
Laboratory
(Survival,
Growth,
Reproduction)
C. tentans
Laboratory
(Survival,
Emergence,
Growth)
H. azteca
In situ
Survival
(Water,
Sediment)
C. tentans
In situ
Survival
(Water,
Sediment)
I). magna
In situ
Survival
(Water,
Sediment)
L. variegatus
In situ
Survival
(Water,
Sediment)
. 1.5 miles below
4
Holmes Road
019
o-o-o
O-O
O-O
O-O
O-O
O
5 Near WWTP
Discharge
428
NA-NA-NA
NA - NA - NA
o-o
O-O
O-O
O-O
O
2 miles below
7 New Lenox
Road
389
• - NA - NA
o - •
o-m
o-m
O-O
•
„ . Vi mile above
Woods Pond
023
• - O - •
NA-NA
NA-NA
NA-NA
NA-NA
•
g Vi mile above
Woods Pond
031
• - O - •
o-m
O - •
o-m
O-O
•
4 O = Negligible to low toxicity: less than 20% effect size relative to upstream background (Al, A3). Overall assessment - negligible indication of
5 ecological risk.
6 O = Moderate toxicity: 20 to 50% effect size relative to upstream background (Al, A3). Overall assessment - ecological effects possible, but not
7 conclusive.
8 • = High toxicity; greater than 50% effect size relative to upstream background (Al, A3). Overall assessment - strong indication of ecological effects.
9 NA = Endpoint not measured due to complete mortality in treatment (Location 7), or samples not collected at station (Locations 5, 8A).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-46
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
3.3.2.1 Approach 1: Calculation of Individual Toxicity Test Endpoints
Point estimates were calculated for each toxicity test, including LC2o and LC50 values for
survival endpoints, and IC20 and IC50 values for sublethal response endpoints (e.g., growth,
reproductive success). For each data set, test endpoints were calculated based on comparison to
both negative controls and reference sediment (Stations Al, A3). Full statistics are presented in
Appendix D; a graphical presentation of toxicity thresholds relative to reference Station Al is
depicted in Figure 3.3-12.
Although there are small differences in the toxicity threshold values calculated using different
statistical methods (i.e., choice of extrapolation model or choice of reference sediment), the data
indicate an increase in the frequency and magnitude of adverse biological responses with
increasing sediment tPCB concentration. The following ranges of tPCB concentrations and
associated responses were developed, based on comparisons of contaminated station responses to
reference stations:
¦ <3 mg/kg - Some sensitive endpoints exhibited apparent responses, but the magnitude of
responses was not large. These subtle responses were difficult to evaluate precisely due
to statistical power limitations, caused in part by the high variability in some treatments.
¦ 3 to 10 mg/kg - Numerous endpoints indicated ecologically significant responses, with
many LC50/EC50 values falling in this range. Statistically significant responses were
observed in most Hyalella and Chironomus life-cycle endpoints at 4.56 mg/kg.
¦ 10 to 30 mg/kg - Nearly all toxicity endpoints indicated large (>50%) responses relative
to reference stations. The only endpoints that did not exhibit large responses in this
concentration range were either growth endpoints or were short-term (48-hour) tests
and/or with tolerant species.
¦ >30 mg/kg - The concentration-response analyses indicated that most survival and
reproduction endpoints exhibited very large reductions at these concentrations, with
complete mortality of some species.
Dose/response modeling was also conducted using individual chemistry values considered to be
"most synoptic" with the toxicity tests (Attachment D.5). These tests had results similar to those
presented above (i.e., most LC50/EC50 values were in the 3 to 30 mg/kg tPCBs range).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-47
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
o>
O)
E
m
o
Q.
100
>77 >77 73.2 >77
~ Graphed LC50/IC50
¦ LC20/IC20 Calculated Value
10
0.1
^
p'
p ^<9 ^ J?
< J?
# o
.>$• vsS1
i> i> rJ>
~ tSb'
j3> j3> j3>
& - & & $ ^ $ $ JP & & & & & &
o- o-
*•
f v# v# jr J?
o-
<&
*• *•
4X
-------
1 Threshold effects concentrations were calculated using the individual endpoint data, in order to
2 allow derivation of site-specific hazard quotients in the Risk Characterization, and to serve as
3 MATCs for downstream risk extrapolations. To calculate threshold effects concentrations, the
4 average of values from the six most sensitive endpoints was calculated for both 50% effects and
5 20% effects levels. This approach was based on the rationale that thresholds should consider
6 multiple sensitive endpoints, but should not be based on the single most sensitive endpoint. The
7 50% effects level corresponds to major impacts, for which there is a high degree of confidence in
8 a significant biological impact. The 20% effects levels correspond to lower but potentially
9 biologically significant effects.
10 Calculations were performed for comparisons to negative control sediment and also to field
11 reference sediment. In general, comparisons to field references were preferred for derivation of
12 sediment MATC values, since field references account for physicochemical factors that may
13 mediate sediment toxicity.
14
15
16
17
18
19
20
21
22
23
24
25
26 Using the "median" exposure data, the 50% effect level for sensitive toxicity endpoints is
27 approximately 3 mg/kg tPCB. The analysis conducted using "most synoptic" exposure data only
28 (Attachment D.5) indicated that the 50% effect level for sensitive toxicity endpoints is
29 approximately 6-7 mg/kg tPCB, and that the 20% effect level for sensitive toxicity endpoints is
30 approximately 3 mg/kg tPCB. Based on this information, 3 mg/kg tPCB was selected as the site-
31 specific threshold for sediment tPCB.
Summary of 50% and 20% Effects Levels
¦ Comparison to Negative Control - The mean of the lowest six 50% effects levels
was 1.3 mg/kg tPCB. The mean of the lowest six 20% effects levels was 0.1
mg/kg tPCB.
¦ Comparison to Reference A1 - The mean of the lowest six 50% effects levels
was 3.5 mg/kg tPCB. The mean of the lowest six 20% effects levels was 0.9
mg/kg tPCB.
¦ Comparison to Reference A3 - The mean of the lowest six 50% effects levels
was 3.3 mg/kg tPCB. The mean of the lowest six 20% effects levels was 0.9
mg/kg tPCB.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-49
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3.3.2.2
Approach 2: General Linear Model of Concentration-Response
The assessment of individual endpoints is sensitive to test variability, which can mask broader
trends in toxicity of PCBs. Therefore, a supplemental approach was applied that combined the
toxicity results from various endpoints to identify the overall trend(s) in concentration-response
observed. The endpoints for all toxicity tests were standardized so that the response variables
were equivalent (i.e., responses represented the proportion of their control mean response). This
transformation of all endpoints to the relative performance proportion (RPP) values standardized
results from different toxicological endpoints to similar ranges and facilitated the search for a
single unified model among all endpoints. The results of the general linear modeling are
depicted in Figure 3.3-13. Overall, the linear modeling indicated that seven of eight toxicity
endpoints evaluated were significantly correlated with log-transformed PCB concentration.
Differences between acute endpoints and chronic endpoints were observed; these are likely
related to the greater sensitivity of chronic endpoints in toxicity tests. The modeling procedure
enabled the identification of threshold tPCB concentrations in sediment. These results are in
agreement with the summary of individual test endpoints provided in Section 3.3.2.1. In
summary, sediment tPCB concentrations above 3 mg/kg indicate significant adverse effects for
sensitive (chronic) endpoints, and tPCB concentrations in the 10 to 30 mg/kg tPCBs range may
result in acute mortality to multiple organisms.
Attachment D.5 presents the results of the linear modeling using only the "most synoptic"
exposure data. Results were qualitatively similar to those described above; the analysis indicated
that the threshold for manifestation of tPCB effects is likely greater than 1 mg/kg and less than
10 mg/kg. There is some uncertainty within this concentration range due to variations in
exposure and effects data, with the frequency of adverse effects increasing toward the upper end
of this range.
3.3.2.3 Relationships of Effects with Other COCs
The data for other COCs were also evaluated qualitatively to assess whether the concentrations
of these contaminants were likely to have confounded the results of the PCB concentration-
response presented above. The spatial patterns in COC concentrations were compared against
the pronounced gradient in toxicity.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ r/-\ 7/10/2003
-------
All Acute
1.5
1.0
Q.
0.
a.
0.5
0.0
Q.
0.
a.
-1 o 1
log10(tPCB median concentration)
Remaining Chronic
-1 o 1
log10(tPCB median concentration)
o.
0.
a.
1.5
1.0
0.5
0.0
1.5
1.0
Q.
Q.
0.5
0.0
42-day Hyalella Survival
X
X
X
X
x—
^
X
X
^ ^ X
X
X
S ^ s XX
X
X
X
XX
X
XX ^
^ X
X
X X
-1
0 1
2
log10(tPCB median concentration)
42-day Hyalella Growth
X
K
M
X*
S
X
X
1
*
"If--
*
PR
X
~1
X
X
K
-1 0 1
log10(tPCB median concentration)
Figure 3.3-13 Segmented Linear Regression Models Applied to Toxicity Data,
Relating Relative Performance Proportion (RPP) to Bulk Sediment tPCB
Concentrations (mg/kg)
MK0110:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-51
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Comparison of Other COC Trends to Toxicity Trends
¦ PAHs - Most PAH data show a trend of reduced PAH concentrations with
distance downstream that is the reverse of the observed trend for toxic
responses. Therefore, with the possible exception of Station 7, there is no
evidence that PAHs were a major contributor to the observed pattern of
sediment toxicity.
¦ Dioxins/Furans - These analytes exhibited a spatial trend similar to the trend in
toxicity. This is likely due to co-occurrence between PCBs and dioxins/furans in
environmental samples.
¦ Metals - Metals generally exhibited a pattern of increasing concentration with
distance downstream, which matched the pattern of toxic responses. However,
the trends in metals concentrations also followed the sediment TOC and particle
size distributions. Once metals concentrations are normalized to the substrate
differences (thus accounting for lower bioavailability in downstream areas), there
was no indication that metals concentrations were responsible for observed
effects. This was confirmed by the low hazard quotients (HQs) for these metals,
and the results of the TIE.
On the basis of the information presented above, other COCs (other than tPCBs) do not explain
the patterns of toxicity observed in the in situ toxicity tests. One possible exception is for
dioxins and furans, which correlate strongly with tPCBs (i.e., co-contaminants in PCB mixtures).
3.3.3 Toxicity Identification Evaluations
EVS (2003) conducted TIEs to broadly define the physical/chemical characteristics of the
contaminants causing observed toxic responses. TIEs were conducted using porewater from
Housatonic River sediment to which the daphnid (Ceriodaphnia dubia) was exposed for 48
hours. Several TIE treatments were initiated in late summer 1999, including baseline tests,
oxidant reduction addition tests, EDTA chelation addition, pH-adjusted filtration, pH-adjusted
aeration, and pH-adjusted Ci8 solid phase extraction (SPE). None of the individual treatments
provided a definitive identification of toxic agent; however, integration of the results of various
treatments provides strong indications of the class of toxic agents.
Based on the TIE study, EVS (2003) concluded that non-polar organic compounds (most likely
PCBs) were responsible for the observed pattern of toxicity responses.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-52
-------
1 A more comprehensive Phase II or Phase III TIE would be required to make a definitive
2 conclusion. However, the indications of PCB toxicity in the Phase I TIE are consistent with the
3 large exceedances of sediment quality values (SQVs) and water quality guidelines for PCBs
4 observed in site media.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26 3.3.4 Tissue PCB Effects Thresholds
27 Data were compiled on PCB tissue concentrations associated with lethal or sublethal effects in
28 aquatic invertebrates. The purpose was to estimate threshold tissue concentrations beyond which
29 adverse effects might occur in Housatonic River benthos. The review focused on data for
30 Aroclors 1260 and 1254, in addition to tPCBs. No studies conducted specifically for Aroclor
31 1260 were identified. Because tissue effects data were limited for these PCB metrics, both
32 freshwater and marine/estuarine invertebrate species were considered in the review.
Rationale for Implication of Non-Polar Organic Compounds as Active
Toxicants in TIE Treatments
¦ Significant reduction in toxicity in the pH-adjusted/filtration treatments - Higher
survival in the filtration test was attributed to organic colloids in the samples
being filtered out and/or pH-mediated toxicity alteration of organic compounds
(EVS 2003). Filterable compounds can include non-polar organics, such as
PAHs, PCBs, and some metals.
¦ Significant reduction in toxicity in the pH-adjusted C18 SPE treatments - The
results of these manipulations indicated that the filtration reduced the toxicity of
the original samples. Therefore, this test implicated non-polar organics,
pesticides, and/or some metals.
¦ EDTA treatments - These treatments did not result in a reduction of toxicity.
This provides evidence against metals as the dominant causal agent.
¦ Sediment and porewater chemistry - PCB concentrations in TIE treatments were
observed to be well above upper-bound sediment quality guidelines and water
quality criteria applied to porewater. Conversely, PAH concentrations in these
TIE treatments (EVS 2003) were below applicable criteria (i.e., Swartz [1999]
sediment quality guidelines for total PAHs). Furthermore, the two samples
demonstrated to be the most toxic in the initial 24-hour screening toxicity test had
the highest sediment tPCB concentrations.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-53
-------
1 To estimate PCB effects thresholds, the tissue effects data were ranked in order of increasing
2 tissue concentration to illustrate studies where effects did and did not occur (Figure 3.3-14). The
3 figure includes only the subset of studies deemed appropriate for threshold derivation (screening
4 rationale provided in Appendix D). Ten studies of a total of 11 freshwater species and seven
5 estuarine/marine species were deemed applicable. The majority of data applied to effects on
6 mortality; however, there were also data for growth, development, behavior, physiological, and
7 cellular effects endpoints.
8 Figure 3.3-14 shows the distribution of no effect and effect tissue concentrations. Based on this
9 distribution, it appears that adverse effects are unlikely to occur at tissue concentrations at or
10 below 3 mg/kg, that they are likely to occur to sensitive organisms above 10 mg/kg, and that
11 there is some uncertainty about whether they will occur at concentrations between these points.
12 3.3.5 Sediment Quality Values (SQVs)
13 Numerous sediment quality benchmarks have been developed, using various derivation
14 procedures. The limitations associated with the derivation of the SQVs must be considered in
15 their application. In recognition of these limitations, SQVs were used in the benthic invertebrate
16 ERA as an additional line of evidence, rather than as a conclusive statement, regarding the
17 toxicity of COCs in Housatonic River sediment. A summary of the values used in the ERA is
18 provided in Appendix D (Table D.3-10).
19 3.3.6 Benthic Macroinvertebrate Community Evaluation
20 3.3.6.1 Methods
21 Multiple lines of evidence were considered in the evaluation of benthic community data.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-54
-------
1000
.5*
o
5
"5
:$
m
o
a.
O)
~a>
E
Open Bar = no significant effect (NOEL)
Solid Bar = significant effect (LOEL)
Figure 3.3-14 Combined Effects and No-Effects Levels for RGB Concentrations (mg/kg wet) in Benthic
Invertebrate Tissue Samples - tPCBs and Aroclor 1254
MK0110:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-55
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15 Because the regression/correlation approach represents an integration of exposure and effects
16 assessments, these analyses were deferred to the risk characterization (Section 3.4). The
17 comparisons of exposed locations to reference locations are discussed here.
18 Using the screening rationale provided in Appendix D (Attachment D.3), six benthic community
19 metrics were included in multivariate statistical analyses.
Statistical Approaches Applied in Benthic Macroinvertebrate
Community Assessment
Comparison of benthic assemblages between contaminated locations and reference
locations. Tools used to make these comparisons included:
¦ Average rank plots, combining relevant summary metrics in a non-parametric
multivariate approach.
¦ Multidimensional scaling (MDS) plots, using the same summary metrics in a
parametric multivariate approach.
¦ Univariate tests using key summary metrics.
¦ Analysis of the relationship between sediment COC concentrations and benthic
community structure indices, using a regression/correlation approach. This
required partitioning of the data set into broad habitat types to help reduce the
confounding effect of habitat type on benthic assemblages.
20
21
22
23
24
25
26
27
28
29
30
31
32
Benthic Community Metrics Evaluated
Multivariate Assessment:
¦ Organism abundance (number of animals per replicate or station).
¦ Taxonomic richness (number of unique taxa per replicate or station).
¦ "EPT" relative abundance (mayflies, stoneflies, caddisflies).
¦ Relative abundance of tolerant dipterans
¦ Relative abundance of tolerant oligochaetes.
¦ Relative abundance of tolerant gastropods.
Univariate Assessment:
¦ Organism abundance
¦ Taxonomic richness
¦ Modified Hilsenhoff Biotic Index (MHBI)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-56
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
3.3.6.2 Results
Detailed results are presented in Appendix D; a summary is provided below. Overall, the benthic
macroinvertebrate community evaluation indicated a high degree of variability, both within and
among locations. Despite within-station variability, some significant locational differences were
observed that were consistent across the metrics considered. Specifically, for most metrics, the
coarse-grained contaminated stations exhibited impaired benthic communities relative to the
three coarse-grained reference stations; impairment was most pronounced at Stations 3 through
5. No habitat differences were identified that would explain the differences in benthic
assemblages observed among coarse-grained locations. No strong or consistent differences in
benthic assemblages were observed among the fine-grained stations.
The results of the benthic community evaluation are summarized as follows:
¦ Average Rank Plots (Figure 3.3-15) - Median ranks at Stations 3, 4, and 5 were
significantly higher than all reference stations, indicating degraded conditions for the
six metrics evaluated. Although the median ranks for Stations 1 and 2 were higher
than at the coarse-grained reference sites, these differences were not statistically
significant. Fine-grained stations did not differ significantly from reference locations.
¦ Multidimensional Scaling (Figure 3.3-16) - In the MDS plot, all coarse-grained
contaminated (C/C) stations (Stations 1 through 5) were set apart from the remaining
stations, suggesting community alteration. Stations 1 and 2 indicated benthic
communities that were different but not consistently degraded relative to the coarse-
grained reference sites. The MDS analysis did not indicate benthic community
alteration at the fine-grained stations.
¦ ANOVA (Total Abundance) - The analysis indicated that all five coarse-grained
contaminated stations had significantly lower total abundances than coarse-grained
reference stations. There was no indication of impairment in fine-grained sediment
relative to reference, however.
¦ ANOVA (Taxa Richness) - In coarse-grained sediment, all five contaminated stations
yielded significantly lower taxa richness relative to references; differences were
somewhat more pronounced for Stations 3 through 5, compared to Stations 1 and 2.
No evidence of ecological disruption in the fine-grained sediment was seen.
¦ MHBI Metric - There was no compelling evidence of incremental habitat degradation
due to PCBs at any of the contaminated stations using the MHBI metric. However,
the appropriateness of this metric was questionable for the study area because the
MHBI was not developed to address effects of PCBs, and because the reference
locations indicated a high "background" proportion of pollution-tolerant taxa.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-57
-------
1 ¦ Biomass Assessment - Overall, the biomass assessment yielded similar findings to
2 the abundance assessment, in that lower biomass was evident in coarse-grained
3 contaminated stations compared to reference stations. No impairment was evident in
4 fine-grained sediment; however, the very high variability in taxonomic distributions
5 among fine-grained locations suggests that habitat variations may limit the ability to
6 detect perturbations.
7 3.3.7 Concentration-Response Analysis - Benthic Community Assemblages
8 The concentration-response assessment for benthic community assemblages was conducted
9 using only the PCB data collected synoptic with the benthic community grab sampling. The
10 replication at each station (i.e., characterization of micro-variation by using 12 replicates),
11 combined with the highly synoptic nature of these data, justified this approach.
12 Although the comparison-to-reference approach yielded significant differences for coarse-
13 grained contaminated stations, these differences were not supported by a linear relationship with
14 PCB concentrations over a wide range of PCB concentrations in coarse-grained sediment. It is
15 possible that the micro-scale variation in PCB sediment chemistry confounded the determination
16 of a relationship between PCB chemistry and benthic abundance and/or richness.
17
18
19
No pattern in other COC concentrations (or habitat variables) was observed that
would explain the impaired benthic communities observed in coarse-grained
contaminated sediment.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-58
-------
o
CM
O
O
C
(0
£
% o
O) oo
(0
a)
<
O
CD
O
"sf
r—|—i ;
—i—
1
p.
i
T
1
1
1
•
~
1—1—1
n
1
—!
l
]
•
1
_L
j
•
I
i_L
i
1
_L
•
1
~r
i
i
•
~r
i
•
•
I
i
•
1
4
»
i
i
i_L.
j
•
i
1
1
_l_
/
1
1
-
1
i_Li
1 2 3 4 56 7 8 9 A1A2A3R4
Station
2 Notes: The box for each station is the interquartile range (1st through 3rd quartiles), and the whiskers are
3 the entire range. The median of the average ranks is shown as a black dot within the box.
4 Statistical definitions of symbols are provided in Appendix D.
5 Stations with high average ranks indicate unfavorable conditions for the six benthic community
6 metrics assessed (e.g., low taxonomic richness, high percentage of tolerant taxa).
7
8 Figure 3.3-15 Average Ranks Analysis for Six Benthic Community Metrics, with
9 Equal Weighting Assigned to Each Metric
10
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-59
-------
1
MDS Axis
3 Note: Shaded ovals are for presentation purposes only, and have no statistical meaning.
4
5 Figure 3.3-16 Multidimensional Scaling for Benthic Community Health Metrics,
6 Showing Metric Medians on MDS Plot
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-60
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
3.4 RISK CHARACTERIZATION
Purpose of Benthic Invertebrate Risk Characterization
¦ Integrate exposure and effects assessments.
¦ Summarize three major lines of evidence and conduct WOE for adverse effects
to benthic invertebrates.
¦ Discuss sources of uncertainty.
¦ Extrapolate risk findings to other species and portions of the Housatonic River
downstream of the PSA.
The risk characterization evaluates the likelihood that adverse effects may occur as a result of
invertebrate exposure to COCs. Three lines of evidence were used in the risk characterization
for benthic invertebrates (Figure 3.1-4):
¦ Field surveys (i.e., benthic community structure) - For these endpoints, care was
exercised to discriminate, to the extent possible, between responses related to COCs and
those related to other factors such as substrate or habitat type.
¦ Comparison of field-measured exposures to effects levels or benchmarks - For these
endpoints, the risk characterization integrated exposure and effects by relating the two
terms quantitatively (e.g., hazard quotient [HQ] method for chemistry data compared to
SQVs from the literature and/or site-specific effects thresholds).
¦ Site-specific toxicity study results - These endpoints (e.g., in situ and laboratory
toxicity tests, TIEs) directly evaluated biological responses to COCs.
These three lines of evidence were independent, allowing for a robust weight-of-evidence
(WOE) assessment of the potential for risk using the approach of Menzie et al. (1996). All lines
of evidence suggested some degree of harm to benthic invertebrates in the Housatonic River. In
addition, for each category of measurement endpoint, there were indications that PCBs are
responsible for the observed patterns of responses.
A WOE assessment was conducted to combine the results from each line of evidence. This
included a station-by-station assessment of each benthic sampling location, as well as an overall
WOE assessment for the assessment endpoint. The section concludes with a discussion of
sources of uncertainty in the assessment of risks of COCs to invertebrates, and the conclusions
of the risk characterization.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ A1 7/10/2003
-------
1 3.4.1 Field Surveys
2 The benthic invertebrate community study (Section 3.3.6 and 3.3.7) directly assessed the
3 assemblages of organisms found throughout the PSA, and related these assemblages to
4 concentrations of COCs and other stressors. After controlling for broad habitat factors (sediment
5 particle size distributions and organic carbon content), significant differences between coarse-
6 grained contaminated sites and coarse-grained references were observed. These differences were
7 not observed in fine-grained sediment, however.
8 There are several possible explanations for the lack of community responses observed in the
9 downstream fine-grained sediment within the PSA, including:
10 ¦ Microhabitat variation - Unlike the coarse-grained sediment, the fine-grained portions of
11 the PSA exhibited considerable inter-station differences in invertebrate communities.
12 These variations may have masked any subtle impacts due to PCBs.
13 ¦ Lower sediment chemistry - The concentrations of tPCB in the benthic community
14 sampling program were lower than for other sampling efforts associated with effects
15 endpoints (e.g., toxicity studies). As shown in Figure 3.2.3, the median sediment tPCB
16 concentration was generally in the 1-10 mg/kg range in the fine-grained sediment
17 collected synoptic with the benthic community grabs. Because these concentrations are
18 close to the site-specific toxicity threshold of 3 mg/kg derived from sediment toxicity
19 endpoints, large alterations in community structure would not necessarily be observed at
20 these levels. Although some biological alteration may be occurring at this concentration
21 range, the statistical power for detecting these differences is very low given the other
22 sources of variability in the study.
23 ¦ Reduced bioavailability of tPCB - As shown in Figure 3.2-2, some of the downstream
24 stations exhibited high organic carbon content, which may act to sequester PCBs.
25 Although the fine-grained sediment were clearly toxic at the higher exposure
26 concentrations in the sediment toxicity tests, the high TOC may have been sufficient to
27 suppress effects in the benthic community grab samples that had tPCB concentrations
28 close to the 3 mg/kg threshold.
29 Overall, due to a relatively narrow range of exposure concentrations and high natural variability,
30 the benthic community study was not suited to the identification of low-level environmental
31 perturbations in fine-grained sediment. Responses in coarse-grained sediment were evident, and
32 consistent across a number of biologically relevant effects metrics (e.g., abundance, taxonomic
33 richness, multivariate community structure).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-62
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
3.4.2 Comparison of Chemistry Data to Benchmarks
For chemistry data (water, sediment, and invertebrate tissue), HQs were used to quantify the
degree to which chemistry measurements exceeded environmental benchmarks considered
protective of the assessment endpoint. To address the uncertainty in generic benchmarks, the
HQs assessment used in the benthic ERA considered multiple benchmarks from different
jurisdictions, and calculated a range of HQs. For each contaminant and medium, the full range
of HQs was considered. Furthermore, to depict the "central tendency" of the benchmarks, HQs
were also calculated using the median value of all applicable benchmarks. The extremes of the
HQ distribution are called "upper-bound" and "lower-bound" HQs in the following discussion.
SQVs derived from the literature are generally conservative and have high associated
uncertainty; hazard quotients greater than one based on literature SQVs must be interpreted in
this context. However, HQs based on site-specific effects thresholds are more reliable indicators
of potential effects. This section discusses both types of HQs; however, only literature-based
HQs were derived for COCs other than PCBs.
3.4.2.1 Sediment Chemistry
Figure 3.4-1 shows the ranges of HQs for the PCB measurements made at the seven toxicity
testing stations in 1999. Within the time period (March to October 1999) 11 sampling events
were conducted that were relevant to the effects data. The bars for each station indicate that the
range of benchmarks derived from the literature (and thus HQs) is more than two orders of
magnitude. The median SQV-based HQs for the contaminated stations are all greater than one,
usually by a large amount.
Figure 3.4-1 also depicts HQs derived using the site-specific effects threshold (MATC) of 3
mg/kg. From a comparison of the two types of HQs, it is apparent that site-specific thresholds
for toxicity observed in the Housatonic River fall within the range of values found in the
literature, but are toward the higher end of SQVs (and therefore the lower end of HQs). All
contaminated stations yielded HQ values greater than one.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-63
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
10000
- I
cu
a*.
0)
(C
O)
"O
0£ 0)
os
X
1000
100
10
0.1
0.01
0.001
A1
i
A3
5 7
Station
8A
Notes: Circle symbol and error bars represent the median and range of HQs calculated using literature-
derived sediment quality values.
Diamond symbols represent HQs calculated using site-specific sediment effects benchmark
(MATC) of 3 mg/kg.
Figure 3.4-1 Hazard Quotients (Median and Range) Based on Median Sediment
Chemistry for tPCBs, for Samples Collected in 1999 Close to Sediment Quality
Triad Stations
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-64
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
HQs were also determined for other COCs that have SQVs. HQs were calculated for Sediment
Quality Triad station concentration data, and also for broader reach-wide data.
Sediment Hazard Quotients for Other COCs
¦ Median antimony HQs were below 1, and maximum antimony HQs barely
exceeded 1 at downstream stations.
¦ Barium HQs barely exceeded 1, and only at downstream stations.
¦ Median cadmium HQs exceeded 1 only at Station 7, and maximum HQs were 10
or less even at the most contaminated sites.
¦ Median chromium concentrations barely exceeded 1 at downstream locations,
and maximum HQs were below 10.
¦ Maximum copper and lead HQs were 10 or less, even at the most contaminated
stations.
¦ Mercury and silver exhibited median HQs between 1 and 10 at most downstream
locations.
¦ The HQs for total PAHs also indicated low risk at stations from these
compounds, with median HQs below 3 at all stations. However, the wide range
of PAH SQVs resulted in higher HQs (i.e., greater than 10) if lower-bound SQVs
are applied.
The broader PSA data indicated HQs that were equal to or lower than those described above for
the sampling locations. For example, the PAH HQs were much lower using the broader PSA
data, with median HQs below 1 for all reaches and both substrate types.
Overall, the HQ assessment for sediment indicated that the chemical hazard for tPCBs was much
higher than for other COCs. The median HQ for tPCBs was often 100 to 1,000, compared to
other COCs that rarely exceeded an HQ of 10. This finding is in agreement with the TIE
conclusions, which implicated PCBs and/or other non-polar organics as the dominant causative
agents in toxicity tests. When HQs based on site-specific tPCB effects information are
considered, risks are moderate to high for most sediment found within the PSA.
3.4.2.2 Water Chemistry
HQs for PCBs (Figure 3.4-2) were calculated by comparing the PCB water column data derived
from the toxicity study (EVS 2003) to water quality criteria for PCBs. The median HQs for both
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC o s-r 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
reference stations (Al, A3) were less than 1.0 in all three sampling events. In contrast, the PCB
concentrations at contaminated locations exhibited median HQs that were elevated and fairly
consistent among stations and across monitoring events (i.e., median HQ of approximately 10).
The maximum HQs, using worst-case PCB benchmarks, were approximately 100. Overall, the
results indicate a moderately high hazard based on PCB chemistry in the water column, with
negligible risk from other water column contaminants.
3.4.2.3 Tissue Chemistry
HQs were derived for tissue PCB burdens in benthic invertebrates sampled near the Sediment
Quality Triad stations. Two sets of HQs were derived representing different levels of
conservatism (Figure 3.4-3). One set of HQs was based on comparison of observed tissue
concentrations to an effects benchmark of 3 mg/kg tPCBs, which represents the lowest
concentration at which significant adverse effects were found in the literature. Nearly all HQs
derived in this manner were greater than 1.0, and three HQs were greater than 10. The second
method compared observed tissue concentrations to 10 mg/kg tPCBs, a concentration that the
literature review suggested would cause impacts to numerous species. Even with this relaxed
benchmark, most HQs still exceeded 1.0.
3.4.3 Site-Specific Toxicity Study Results
Both the in situ and laboratory toxicity tests (Section 3.3.1) exhibited significant adverse effects
in both coarse- and fine-grained sediment, relative to both negative controls and field reference
sediment. The only toxicity test endpoints that did not yield significant adverse responses at the
highest tPCB concentrations were: (a) limited exposure pathways, such as water-only in situ
exposures; (b) short test durations; and/or (c) tolerant test species, such as freshwater
oligochaetes used for bioaccumulation. The large number of endpoints indicating significant
toxicity (even for some acute lethal endpoints), and the high magnitude of response at the highest
PCB concentrations (100% mortality in some treatments), indicates a significant potential for
environmental harm. The evaluation of concentration-response (Section 3.3.2) and the TIE study
(Section 3.3.3) both indicated that non-polar organics (principally PCBs) were likely the
dominant toxic agents in the toxicity tests.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-66
-------
(a) 48-hour in situ
¦o
1000
(U
100
n
(/)
(U
.Q
10
(/)
a>
1
c
CT)
m
c
¦o
0)
(U
0.1
E
a
0.01
X
A1
A3
Station
(b) 7-day in situ
¦o
1000
(U
ft
100
u
n
1
CO
(U
.Q
10
(/)
a>
1
c
o>
.5
c
S
a>
(U
0.1
E
a
0.01
X
i
:
\
> <
i
~
! <
> 1
1
A1
A3
4 5
Station
•a
c
(U
E "j/T
E m
> n
w '
—¦- a)
c o>
(V c
=5 5
0)
E
1000 w
100
uu r
10
1
0.1
0.01
(c) 10-day in situ
A1
A3
Station
Figure 3.4-2 Hazard Quotients (Median, Range) Based on Overlying Water PCB
Concentrations, Measured Synoptic with In Situ Toxicity Tests
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-67
-------
(a) Relative to lower-bound effects threshold of 3 mg/kg (HQ > 1 in black)
Predators Shredders
20
15
.92
4—)
o
O 10
ro
N
ro
x
A1 A3 1 2 3 4 5 6 7 8 9 R4
A1 A3 1
(b) Relative to upper-bound effects threshold of 10 mg/kg (HQ > 1 in black)
Predators
Shredders
5
4
+•»
c
0
o 3
D
o
~o
ro 2
N
to
x
1
0
-
- 1
TO
1 1
TO
¦ nsnl
O
Z
A1 A3 1 2 3 4 5 6 7 8 9 R4
A1 A3 1 2 3 4 5 6 7 8 9 R4
1
2 Figure 3.4-3 Hazard Quotients for tPCB Tissue Residues in Benthic Invertebrates,
3 Relative to Two Effects Thresholds Derived from Literature Studies
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-68
-------
1 3.4.4 Integrated Station-by-Station Assessment
2 Potential impacts of contaminated sediment to local ecological resources at each contaminated
3 station were assessed using a graphical approach that considered multiple lines of evidence
4 (Figure 3.4-4). Multiple measurement endpoints were used, and the results of each were
5 integrated into a single conclusion regarding potential ecological impacts. For the purposes of
6 evaluating each measurement endpoint, results were categorized/simplified based on ecologically
7 based decision criteria. The categorizations facilitated the interpretation of the results for each
8 leg of the Sediment Quality Triad, on a station-by-station basis.
9 Each measurement endpoint was assigned a rating of high, medium, or low impact. Where
10 applicable, indications of potential for harm were standardized to appropriate background
11 conditions (e.g., toxicity endpoints were compared to reference Stations A1 and A3 rather than
12 negative control sediment). The decision criteria used to make the evaluations are summarized
13 in Appendix D.
14 The ratings in Figure 3.4-4 indicate evidence for ecological disruption for all three components
15 of the Sediment Quality Triad. For each component, there are multiple indications of "major"
16 risk, and at multiple stations. The overall assessment yielded a rating of "high" overall risk for
17 all stations except Stations 6 and 9, for which no toxicity testing was conducted. Although there
18 was a high degree of overall concordance, one area of discrepancy was in the benthic community
19 endpoints for fine-grained stations. The strong toxicological responses at these stations were not
20 associated with strong indications of benthic community alterations. The differences in PCB
21 chemistry associated with these endpoints (i.e., higher tPCBs concentrations observed in the
22 toxicity samples relative to toxicity test samples) may explain this apparent difference.
23 3.4.5 Weight-of-Evidence (WOE) Procedure for Assessing Risk from PCBs in the
24 Housatonic River PSA
25 A formal WOE process was applied to determine whether PCBs pose a significant risk to the
26 Housatonic River benthos. The three-phase approach of Menzie et al. (1996) and the
27 Massachusetts Weight-of-Evidence Workgroup was applied for this purpose, in which WOE was
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-69
-------
1
Endpoint
Stations
1
2
3
4
5
6
7
8A
8
9
1. Sediment Toxicity
42-d Hyalella (lab)
Survival
-
-
-
o
-
-
•
•
•
-
42-d Hyalella (lab)
Growth
-
-
-
o
-
-
-
o
O
-
42-d Hyalella (lab)
Reproduction
-
-
-
o
-
-
-
•
•
-
43-d Chironomus (lab)
Survival
-
-
-
•
-
-
•
•
•
-
43-d Chironomus (lab)
Emergence
-
-
-
•
-
-
•
•
•
-
43-d Chironomus (lab)
Growth
-
-
-
•
-
-
•
•
•
-
10-d Hyalella {in situ)
Survival
-
-
-
o
o
-
•
-
•
-
10-d Chironomus {in situ)
Survival
-
-
-
o
o
-
•
-
•
-
48-h Daphnia {in situ)
Survival
-
-
-
o
o
-
•
-
•
-
48-h Lumbriculus {in situ)
Survival
-
-
-
o
o
-
o
-
o
-
TIE Treatments with Ceriodaphnia
Survival effect linked to PCBs
-
-
-
-
-
-
•
-
•
-
2. Benthic Community
Multivariate - Average Rank Plots
Equal Endpoint Weighting
O
o
•
•
•
o
o
-
o
o
Multivariate - MDS
Separation in 2-dimensional plot
O
o
o
o
•
o
o
-
o
o
Modified Hilsenhoff Bio tic Index
MUBI scores (ANOVA)
o
o
o
o
o
o
o
-
o
o
Taxa Richness
ANOVA versus references
o
o
•
•
•
o
o
-
o
o
Total Abundance
ANOVA versus references
•
•
•
•
•
o
o
-
o
o
2 Figure 3.4-4 Weight-of-Evidence Evaluation of Housatonic River Benthic Sampling Locations, with Indications of
3 Alteration/Risk Relative to Background
4
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ qrs 7/10/2003
-------
1
Endpoint
Stations
1
2
3
4
5
6
7
8A
8
9
3. Chemistry
Toxicity - Sediment tPCB
Synoptic with sediment toxicity tests
-
-
-
•
•
-
•
•
•
-
Benthos - Sediment tPCB
Synoptic with benthos collection
•
•
•
•
•
•
•
-
•
o
PSA Data - Sediment tPCB
Reach-wide sampling (median)
•
•
•
•
•
•
•
•
•
•
Water column tPCB
Synoptic with toxicity tests
-
-
-
O
o
-
•
-
O
-
Tissue tPCB in predators
Relative to literature benchmark
O
-
O
•
•
-
•
-
O
•
Tissue tPCB in shredders
Relative to literature benchmark
•
O
-
•
•
•
o
-
•
o
Tissue tPCB in oligochaetes (lab)
Relative to literature benchmark
-
-
-
o
o
-
o
-
o
-
4. Integrated Assessment
Toxicity Endpoints
Combined Assessment
-
-
-
o
o
-
•
•
•
-
Benthic Endpoints
Combined Assessment
o
o
•
•
•
o
o
-
o
o
Chemistry Endpoints
Combined Assessment
•
•
•
•
•
•
•
•
•
o
OVERALL
•
•
•
•
•
o
•
•
•
o
2 Notes:
3 • = major impact.
4 O = moderate impact.
5 O = negligible impact.
6
7
8 Figure 3.4-4 Weight-of-Evidence Evaluation of Housatonic River Benthic Sampling Locations, with Indications of
9 Alteration/Risk Relative to Background (Continued)
10
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC 1 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
reflected in the following three characteristics: (1) the weight assigned to each measurement
endpoint; (2) the magnitude of response observed in the measurement endpoint; and (3) the
concurrence among outcomes of the multiple measurement endpoints.
A discussion of attributes considered in the WOE is provided in Section 2, and the rationales for
weighting of measurement endpoints are provided in Appendix D. A summary of the derived
weightings for each attribute is provided in Table 3.4-1. The chemistry endpoints yielded the
lowest overall values because of lower site-specificity and some uncertainties in the biological
association between the measurement endpoints and the assessment endpoint(s). The toxicity
testing endpoints yielded the highest overall values, because of the high degree of biological
relevance of the tests. The benthic community structure endpoints had intermediate values.
Although these endpoints were site-specific, collected at a time when effects would be expected,
and were measures of the community structure component of the assessment endpoint, the
potential for the confounding effects of other factors in the direct attribution of the response to
the stressor reduced the utility of these endpoints to some degree.
The magnitude of the response in the measurement endpoint is considered together with the
measurement endpoint weight in judging the overall WOE (Menzie et al. 1996). This requires
assessing the strength of evidence that ecological harm has occurred, as well as an indication of
the magnitude of response, if present. The weighting scores, evidence of harm, and magnitudes
of responses were combined in a matrix format and are presented in Table 3.4-2.
A graphical method was used for displaying concurrence among measurement endpoints (Table
3.4-3). The method entailed plotting the nine symbols representing the toxicity (T), benthic
community (B), and chemistry (C) endpoints in a matrix, with the weight of the measurement
endpoint and the degree of response as axes. These graphics indicate that the majority of
endpoints suggest some risk for benthic communities in both coarse- and fine-grained sediment.
The plots also indicate that several of the endpoints suggest a high degree of risk with a
relatively high weight (e.g., toxicity endpoints). The conclusion from interpretation of Table
3.4-3 is that there is a moderate to high risk to much of the benthic community indicated by the
WOE evaluation.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-72
-------
1 Table 3.4-1
2
3 Weighting of Measurement Endpoints for Weight-of-Evidence Evaluation
Attributes
Endpoint Group C: Chemistry
Endpoint Group T: Toxicity
Endpoint Group B: Benthic
Community
C-l
(Water)
C-2
(Sediment)
C-3
(Tissue)
T-l
(Lab)
T-2
(in situ)
T-3 (TIE)
B-l
(Metrics)
B-2
(Multivar)
B-3
(MHBI)
I Relationship Between Measurement and Assessment Endpoints
1. Degree of Biological Association
Low
Low
Mod
Mod/High
Mod/High
Mod
Low/Mod
Low/Mod
Low/Mod
2. Stressor/Response
Low
Low/Mod
Mod
Mod/High
Mod/High
Mod/High
Low/Mod
Low/Mod
Low/Mod
3. Utility of Measure for Judging Risk
Low
Mod
Mod
High
High
Mod/High
Low/Mod
Low/Mod
Low/Mod
H. Data Quality
4. Data Quality
High
High
High
High
High
High
High
High
High
EX Study Design
5. Site Specificity
Low/Mod
Low/Mod
Low/Mod
Mod/High
Mod/High
Mod/High
High
High
High
6. Sensitivity to Detecting Changes
Low/Mod
Low/Mod
Low/Mod
High
Mod/High
High
Low/Mod
Low/Mod
Low/Mod
7. Spatial Representativeness
Mod/High
High
Mod
Mod
Mod/High
Low
Mod/High
Mod/High
Mod/High
8. Temporal Representativeness
High
High
Mod/High
Mod
Mod
Mod
Mod
Mod
Mod
9. Quantitativeness
Mod/High
Mod/High
Mod/High
High
High
Mod/High
Mod/High
Mod/High
Mod
10. Standard Method
Mod
High
High
High
Mod/High
Mod/High
High
High
High
Overall Endpoint Value
Low/Mod
Low/Mod
Mod
Mod/High
Mod/High
Mod
Mod
Mod
Mod
4 C. Chemical Measures
5 C-l. Concentration of PCB in overlying water in relation to concentrations reported to be harmful to benthic invertebrates.
6 C-2. Concentration of PCB in the sediment in relation to concentrations reported to be harmful to benthic invertebrates.
7 C-3. Concentration of PCB in invertebrate tissues in relation to concentrations reported to be harmful to benthic invertebrates.
8 T. Toxicological Measures
9 T-l. Sediment toxicity to multiple invertebrate species, as measured in laboratory toxicity tests.
10 T-2. Sediment toxicity to multiple invertebrate species, as measured in the in situ toxicity tests.
11 T-3. Indications of PCB as toxicity driver in TIE investigations.
12 B. Benthic Community Measures
13 B-l. Abundance, richness, and biomass of invertebrates, relative to reference stations of comparable substrate and habitat (ANOVA analysis).
14 B-2. Benthic community structure, as assessed using multivariate assessment of key benthic metrics (rank analysis and multidimensional scaling).
15 B-3. Water quality assessment using modified Hilsenhoff Bio tic Index (MHBI) indicator of organic pollution.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC ^ 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
3.4.6 Sources of Uncertainty
The assessment of risks to benthic invertebrates contains uncertainties. Each source of
uncertainty can influence the estimates of risk, therefore, it is important to describe and, when
possible, specify the magnitude and direction of such uncertainties. Appendix D contains a more
complete list of uncertainties; some of the most significant uncertainties are summarized below.
¦ Small-scale variability in COC exposure concentrations, which complicated the
development of concentration-response relationships - The variability in exposure
concentrations within and among studies, and the differences in spatial trends across
some studies required careful characterization of exposures that are appropriately
matched to effects data, particularly for sediment concentrations. Analytical variability
(Appendix C.l 1) and field variability both contribute to uncertainty in exposure data.
¦ Inconsistencies in exposure concentrations across studies - The patterns of PCB
concentrations observed in the benthic community study, sediment toxicity study, and
benthic invertebrate tissue sampling study were not always consistent. For example, the
PCB concentrations measured at Stations 7 and 8 during the benthic community sampling
(n=12), were lower than most other PCB measurements made at those locations. This
complicated the integrated station-by-station assessment presented in Section 3.4.4, since
the magnitudes of exposure at a given station were not always equal across all effects
endpoints.
¦ Calculations of site-specific effects thresholds (e.g., sediment MATC of 3 mg/kg) had the
following uncertainties: (a) uncertainty due to application of dose-response models
required for interpolation; (b) uncertainty regarding the choice of exposure data that is
considered synoptic to effects information; (c) uncertainty due to natural variability in
exposure data and effects data. These uncertainties were addressed in the ERA by
conducting multiple assessments (e.g., applying different statistical models and exposure
assumptions). The general concordance of the findings using many different data
processing assumptions provides confidence that the derived thresholds are not based on
spurious statistical outcomes.
¦ Occurrence of elevated PCB concentrations (tissue and sediment) in samples collected
from the West Branch of the Housatonic River near the confluence - These elevated
concentrations may be due to localized PCB inputs from contamination in the vicinity of
Dorothy Amos Park; this uncertainty cast some doubt on the appropriateness of Station
A3 as a reference.
¦ The effects benchmarks derived from the literature carry a high degree of uncertainty,
due to the need to extrapolate across sites, species, and PCB mixtures. The site-specific
Sediment Quality Triad studies indicated that the lower-bound (most conservative)
benchmarks are over-protective for PCB and other COCs, and that upper-bound
benchmarks are more indicative of Housatonic River effects thresholds.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-74
-------
1 Table 3.4-2
2
3 Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related to Maintenance of a Healthy
4 Benthic Community
Measurement Endpoints
Weighting
Value (High,
Moderate,
Low)
Coarse-Grained Sediment
Fine-Grained Sediment
Evidence of
Harm (Yes, No,
Undetermined)
Magnitude (High,
Intermediate, Low)
Evidence of
Harm (Yes, No,
Undetermined)
Magnitude (High,
Intermediate, Low)
C. Chemical Measures
C-l. Concentration of PCB in overlying water in relation to
levels reported to be harmful to benthic invertebrates
Low/Moderate
Yes
Intermediate
Yes
Intermediate
C-2. Concentration of PCB in the sediment in relation to
levels reported to be harmful to benthic invertebrates
Low/Moderate
Yes
High
Yes
High
C-3. Concentration of PCB in invertebrate tissues in relation
to levels reported to be harmful to benthic invertebrates
Moderate
Yes
Intermediate
Yes
Intermediate
T. Toxicological Measures
T-l. Sediment toxicity to multiple invertebrate species, as
measured in laboratory toxicity tests
Moderate/
High
Yes
High
Yes
High
T-2. Sediment toxicity to multiple invertebrate species, as
measured in in situ toxicity tests
Moderate/
High
Yes
Intermediate
Yes
High
T-3. Indications of PCB as toxicity driver in toxicity
identification evaluations
Moderate
Undetermined
—
Yes
Intermediate
B. Benthic Community Measures
B-l. Abundance, richness, and biomass of invertebrates,
relative to reference stations of comparable substrate and
habitat (ANOVA)
Moderate
Yes
Intermediate
No
B-2. Benthic community structure, as assessed using a
multivariate assessment of key benthic metrics (rank analysis
and MDS)
Moderate
Yes
Intermediate
No
B-3. Water quality assessment using modified Hilsenhoff
Biotic Index (MHBI) indicator of organic pollution
Moderate
No
—
No
—
5
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-75
-------
1
2
3
4
5
6
7
8
9
10
Table 3.4-3
Weight-of-Evidence Risk Analysis Summary Indicating Concurrence Among Endpoints for Coarse-Grained and
Fine-Grained Sediment
Assessment Endpoint:
Community, structure, survival, growth, and reproduction of benthic invertebrates
(a) Coarse-grained contaminated (C/C) sediment
Harm/Magnitude
Weighting Factors (increasing confidence or weight)
Low
Low/Moderate
Intermediate
Moderate/High
High
Yes/High
C-2
T-l
Yes/Intermediate
C-l
B-l, B-2, C-3
T-2
Yes/Low
Undetermined
No Harm
T-3
B-3
n
(b) Fine-grained contaminated sediment (F/C)
Harm/Magnitude
Weighting Factors (increasing confidence or weight)
Low
Low/Moderate
Intermediate
Moderate/High
High
Yes/High
C-2
T-l, T-2
Yes/Intermediate
C-l
T-3, C-3
Yes/Low
Undetermined
No Harm
B-l, B-2, B-3
n
Note: See Tables 3.4-1 and 3.4-2 for definitions of endpoint codes.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-76
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
¦ There is uncertainty with respect to the confounding role of micro-habitat for benthic
communities. Although the study design controlled for habitat (physical and biological)
to the extent possible, variations in micro-habitat factors may have obscured alterations
due to chemical stressors.
¦ Individual toxicity effects endpoints carry some uncertainty because individual taxa have
specific tolerances to both chemical and background environmental factors. The strength
of the Sediment Quality Triad approach comes from the multiple lines of evidence (lethal
and sublethal test endpoints with different exposure durations) from multiple test species.
The concurrence of findings from different taxa substantially reduced this uncertainty.
3.4.7 Extrapolation to Other Species
The benthic invertebrate ERA included the entire benthic community; benthic community
composition analysis was a measurement endpoint considered in the weight-of-evidence
assessment. Individual species were also used in toxicity tests as surrogates for the Housatonic
River freshwater benthic community. Both the status of sensitive taxa and community
composition are considered indicators of overall health and productivity of the benthic
community. As a result, no explicit extrapolation to other species was required. The toxicity test
species and endpoints encompass a range of toxicological sensitivities, ranging from sensitive
(e.g., Hyalella chronic reproduction) to tolerant (e.g., Lumbriculus survival); similar variation in
sensitivity can be expected in the field.
3.4.8 Downstream Assessment
Because of the more limited amount and spatial coverage of data on contaminant concentrations
downstream of the PSA, the more rigorous approach followed in assessing ecological risks in the
PSA was not appropriate or possible. Risk estimates for downstream of Woods Pond were
derived by comparing observed sediment and tissue concentrations with maximum acceptable
threshold concentrations (MATCs) for tPCBs developed from the Sediment Quality Triad.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-77
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18 Using the MATC values, potential risks to benthic invertebrates are predicted to occur in limited
19 areas downstream of Woods Pond to Rising Pond, where pockets of sediment contaminated with
20 higher concentrations of PCBs appear to have accumulated. Below Rising Pond through the
21 remainder of Massachusetts and Connecticut, sediment does not contain concentrations of PCBs
22 that are sufficiently elevated to represent a potential risk to benthic invertebrates. Tissue
23 concentration data for caddisflies, dobsonflies, and stoneflies (BBL and QEA 2003) collected
24 from Cornwall CT indicate tPCB concentrations at or below the toxicity threshold of 3 mg/kg,
25 and therefore support the conclusions based on sediment.
26 3.4.9 Conclusions
27 Overall, the benthic ERA indicates significant risk to aquatic invertebrates based on a WOE
28 evaluation of multiple Sediment Quality Triad endpoints. Furthermore, the available data
29 suggest that PCBs are the primary chemical stressor responsible for such impairment. The
30 confidence in the conclusion is moderate to high, based upon the concordance in predictions of
31 risk from multiple measurement endpoints.
32 Compelling evidence for ecological risk comes from the sediment toxicity tests, which not only
33 indicated significant toxicological effects in multiple appropriate indicator species and endpoints,
34 but also indicated a correlation between the level of effect and sediment PCB concentration.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
MATCs for PCBs Used to Assess Risks Below Woods Pond
¦ The sediment MATC of 3 mg/kg tPCB was used as a conservative measure of
the potential for adverse effects on benthic invertebrates downstream of Woods
Pond. This concentration was developed in the risk assessment for the PSA
using multiple lines of evidence (e.g., benthic community studies, in situ and
laboratory toxicity testing, bioaccumulation testing, Sediment Quality Triad) and
was selected as the concentration at which some sensitive endpoints exhibited
apparent responses, but the magnitude of responses was not large. Above a
concentration of 3 mg/kg tPCB, numerous endpoints indicated ecologically
significant responses, with many LC5o/EC5o values falling in this range.
¦ The tissue MATC of 3 mg/kg tPCB was used as a conservative measure of the
potential for adverse effects on benthic invertebrates downstream of Woods
Pond. This concentration was developed considering the frequency of adverse
effects observed in the literature studies; none of the available studies yielded
toxic responses below 3 mg/kg tPCB, but numerous studies above 3 mg/kg
yielded significant adverse responses.
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
This correlation was consistent with the TIE results, which implicated non-polar organics as the
dominant toxicants in Housatonic River sediment. The evidence of effects to benthic community
structure was not as compelling, because significant alteration relative to reference conditions
was not observed in the fine-grained sediment downstream of the WWTP.
The magnitude of risk to benthic invertebrates in the Housatonic River varies spatially, primarily
as a function of sediment PCB concentration and also in relation to sediment characteristics,
primarily organic carbon content. The WOE assessment of benthic invertebrate endpoints
indicates a high risk of ecologically significant effects at the PCB concentrations observed at the
Sediment Quality Triad stations. The toxicity studies within the PSA indicated that ecologically
significant effects were observed at sediment tPCB concentrations of 3 mg/kg or higher, and that
effects were large in magnitude (i.e., 50% responses in most test species) at 10 mg/kg tPCBs.
These concentrations are in general agreement with a threshold identified from benthic
community studies (5 mg/kg tPCB) and are in concordance with the higher end of the SQVs for
tPCBs identified in a literature review (i.e., 1 to 10 mg/kg). The spatial distribution of tPCB
concentrations in the PSA (Figure 3.2-5) indicates that most of the sediment in the PSA exceeds
these threshold effects levels. Unacceptable risks are predicted for the majority of sediment
sampled within Reach 5A. In the downstream reaches of the PSA, risks were lower; however,
the tPCB data indicate that a high percentage of samples still exceed the site-specific thresholds
described above. Downstream of Woods Pond, risks are reduced relative to the PSA, and are
negligible to low downstream of Rising Pond.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-79
-------
1 3.5 REFERENCES
2 Academy of Natural Sciences of Philadelphia, Patrick Center for Environmental Research. 1999.
3 PCB Concentrations in Fishes and Benthic Insects from the Housatonic River, Connecticut, in
4 1984-1998. November 15, 1999. Report No. 99-10F.
5 BBL (Blasland, Bouck & Lee, Inc.) and QEA (Quantitative Environmental Analysis, LLC).
6 2003. Housatonic River - Rest of River RCRA Facility Investigation Report. Prepared for
7 General Electric Company. January 2003.
8 Chapman, P.M. 1996. Presentation and interpretation of Sediment Quality Triad data.
9 Ecotoxicology 5:327-339.
10 EPA (U.S. Environmental Protection Agency). 2000. Methods for Measuring the Toxicity and
11 Bioaccumulation of Sediment-Associated Contaminants with Freshwater Invertebrates. EPA
12 600/R-99/064. 192 pp.
13 EVS Environment Consultants. 2003. Assessment of in situ Stressors and Sediment Toxicity in
14 the Lower Housatonic River. Final Report, adapted from a study by G.A. Burton, 2001, Institute
15 for Environmental Quality, Wright State University, Dayton Ohio.
16 Long, E.R., and P.M. Chapman. 1985. A Sediment Quality Triad: measures of sediment
17 contamination, toxicity and infaunal community composition in Puget Sound. Marine Pollution
18 Bulletin 16:405-415.
19 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S.
20 Petron, B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-
21 of-Evidence Workgroup: A weight of evidence approach for evaluating ecological risks. Human
22 and Ecological Risk Assessment 2:277-304.
23 Swartz, R.C. 1999. Consensus sediment quality guidelines for polycyclic aromatic hydrocarbon
24 mixtures. Environ. Tox. Chem. 18(4):780-787.
25 WESTON (Roy F. Weston, Inc.). 2000. Supplemental Investigation Work Plan for the Lower
26 Housatonic River. Prepared for U.S. Army Corps of Engineers and U.S. Environmental
27 Protection Agency. DCN: GEP2-020900-AAME.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_3.DOC
3-80
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
4. ASSESSMENT ENDPOINT—COMMUNITY CONDITION, SURVIVAL,
REPRODUCTION, DEVELOPMENT, AND MATURATION OF
AMPHIBIANS
Highlights of the Amphibian ERA
Conceptual Model
The assessment endpoint is the survival, development, and reproduction of
amphibians in the PSA. Amphibians, including leopard frogs and wood frogs,
selected as representative species for the ERA, are exposed to contaminants of
concern (COCs) via diet and possibly dermal absorption.
Exposure
Exposure of the representative species to tPCBs, dioxins and furans, metals, and
PAHs was determined through three site-specific studies that evaluated reproductive
performance and developmental effects. Routes of exposure and rates of
bioaccumulation were also assessed.
Effects
Reproductive performance and early developmental effects were assessed using a
number of measurement endpoints in frogs from contaminated areas in the PSA and
frogs from reference areas from the Housatonic River watershed and external
reference sources. These effects were compared to those reported in the literature
to identify similarity of responses for COCs, types of effects, and mechanisms of
effects.
Risk
There is a high probability of risk of ecologically significant effects at PCB
concentrations observed in the PSA. There were significant correlations between
adverse effects in late larval-stage wood frogs and PCB concentrations in sediment
and tissue. Leopard frogs appear more acutely sensitive than wood frogs, with
strong indications of toxicity observed through the range of tPCB concentrations
tested in the PSA. These findings suggest that amphibian populations are impacted
throughout much of the PSA. The indications of community responses from the
population studies (i.e., localized depressions of richness and abundance near high
tPCB vernal pools, and high incidence of malformations observed) substantiate these
findings.
4.1 INTRODUCTION
The purpose of this section of the ecological risk assessment (ERA) is to characterize and
quantify the current and potential risks posed to amphibians exposed to contaminants of potential
concern (COPCs) in the Housatonic River, focusing on total PCBs (tPCBs) and other COPCs
originating from the General Electric Company (GE) facility in Pittsfield, MA. The watershed is
located in western Massachusetts and Connecticut, discharging to Long Island Sound, with the
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC a -i 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
GE facility located near the headwaters of the watershed. The Primary Study Area (PSA)
includes the river and 10-year floodplain from the confluence of the East and West Branches of
the Housatonic River downstream of the GE facility to Woods Pond (Figure 1.1-2).
A Pre-ERA was conducted to narrow the scope of the ERA by identifying COPCs, other than
tPCBs, posing potential risks to aquatic biota in the PSA (Appendix B). The amphibian ERA
further screened COPCs for specific relevance to the amphibian community occupying the vernal
pool, floodplain, and backwater habitats of the Housatonic River. The contaminants of concern
(COCs) that were retained for the detailed risk assessment for amphibians were tPCBs, six
metals, several polycyclic aromatic hydrocarbons (PAHs), and dibenzofurans.
A step-wise approach was used to assess the risks of these COCs to amphibians in the
Housatonic River watershed. The four main steps in this process included:
1. Derivation of a conceptual model (Figure 4.1-1).
2. Assessment of exposure of amphibians to COCs (Figure 4.1-2).
3. Assessment of the effects of COCs on amphibians (Figure 4.1-3).
4. Characterization of risks to amphibians (Figure 4.1-4).
The detailed ecological risk assessment for amphibians is provided in
Appendix E.
This section is organized as follows:
¦ Section 4.2 (Conceptual Model) describes the conceptual model for amphibians,
including selection of representative taxa and establishment of measurement and
assessment endpoints.
¦ Section 4.3 (Exposure Assessment) describes the quantification of exposures, both
specific to the amphibian study's effects stations and for the broader study area.
¦ Section 4.4 (Effects Assessment) describes the potential effects to amphibians
exposed to site COCs, as indicated by the toxicological and field investigations
conducted in the PSA. Section 4.4 also summarizes the ranges of tissue benchmarks
(toxicity thresholds) derived from the literature.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-2
-------
(A
-------
Exposure
Figure 4.1-2 Overview of Approach Used to Assess Exposure of
Amphibians to Contaminants of Concern (COCs) in the
Housatonic River PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-4
-------
Effects
Figure 4.1-3 Overview of Approach Used to Assess the Effects of
Contaminants of Concern (COCs) to Amphibians in the
Housatonic River PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-5
-------
Risk Characterization
Figure 4.1-4 Overview of Approach Used to Characterize the Risks of
Contaminants of Concern (COCs) to Amphibians in the
Housatonic River PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-6
-------
1 ¦ Section 4.5 (Risk Characterization) integrates the exposure and effects assessments
2 and makes conclusions regarding risk for amphibians in the Housatonic River and
3 floodplain/backwater habitats using three lines of evidence. A discussion of the
4 sources of uncertainty regarding risk estimates follows. Section 4.5 also presents an
5 extrapolation of risks beyond the PSA to areas downstream of Woods Pond.
6
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-7
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
4.2 CONCEPTUAL MODEL
Total PCBs, dioxins, and furans are persistent and hydrophobic and lipophilic. Therefore,
organic carbon pools (both living and non-living) are the primary uptake vectors for juvenile
amphibians, with aquatic and terrestrial invertebrates an important uptake pathway for adults.
Less lipophilic COCs, such as low molecular weight PAHs and metals, are less associated with
organic pools, and exhibit more complex partitioning behavior. The COCs identified for
amphibians exhibit both direct (i.e., contact with contaminated source media) and indirect (i.e.,
food web bioaccumulation, maternal transfer) exposure pathways.
The conceptual models presented in Figures 4.2-1 and 4.2-2 illustrate the exposure pathways for
amphibians in the PSA. The amphibian assessment focused on life stages that are in direct
contact with Housatonic River sediment. For amphibian larvae, the dominant abiotic exposure
media were sediment (solid phase and/or porewater) and surface water. Concentrations of COCs
in tissues of amphibians were also considered. Tissue data provide an organism-based measure
of bioavailability, and provide an additional line of evidence to consider along with the effects
data gathered in the two frog developmental studies (FEL 2002a, 2002b).
Section 2, Problem Formulation, identified two indigenous species to be used in toxicity tests
representative of the Housatonic River amphibian community: leopard frogs (Rana pipiens) and
wood frogs (Rana sylvatica). Summary life history profiles for both species are found in the
following text boxes, and detailed profiles are provided in Appendices A and E.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-8
-------
Breeding Adult
Tadpoles
Summer
Metamorph
Summer Adult/
Juvenile
Winter
Adult/Juvenile
Rock with
Periphyton
Sediments
Uptake
y
Transfer —
—~
Flying Insect
(Pathway also includes
Terrestrial Insects)
Figure 4.2-1 Leopard Frog Exposure Pathways
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-9
-------
Flying Insect
S (Pathway also includes
Terrestrial Insects)
Breeding
Adult j
Breeding Adult ^
(7-10 Days)
Juvenile and
Metamorph
Late-larval/Pre-metamorph
Egg
Masses Tadpoles
Rock with Periphyton
Uptake
Floodplain Soils
Sediments
Transfer
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
Figure 4.2-2 Wood Frog Exposure Pathways
4-10
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Life History of Northern Leopard Frog
The northern leopard frog is a slender, medium-sized frog, sometimes referred to as the
"meadow frog" because of its preference for grassy habitats. It is not considered an
obligate vernal pool species in New England, primarily because it prefers lakes, ponds,
and slow-moving streams for breeding. Its life cycle includes an aquatic larval stage and
semi-terrestrial juvenile and adult stages.
¦ Habitat—Considered semi-terrestrial. Breed and overwinter in water bodies, adults
spend the entire post-breeding summer period in grassy meadows, open shrub
areas, or damp woods, often far from any water. In southern New England, appear
to be restricted to floodplains along large streams and rivers, wetlands along lake
margins, and meadows adjacent to freshwater and brackish tidal wetlands. Often
inhabit cattle pastures and hay fields, otherwise seem to avoid severely disturbed or
sites with poor water quality. In the spring, attracted to vegetated shorelines by a
greater abundance of food, moderated temperatures, and protective cover.
¦ Home Range and Territoriality—Adults show marked fidelity to home areas, with
individuals remaining in a relatively confined area for most of the summer, returning
to that area after nighttime excursions and the following year after hibernation and
breeding. Especially active during rainy nights, when they will often move to warm
road surfaces. Temperature (air and water) may play a major role in the timing of
their movements between wintering and breeding areas and between summering
and wintering areas. Have shown excellent homing ability when displaced moderate
distances (i.e., <1 km) from home area.
¦ Food Habits and Diet—Foods of adults and juveniles include insects, as well as
spiders, snails, and frogs. Availability rather than preference likely determines food
types; beetles are a staple in the diet of adults and juveniles. Moth and butterfly
larvae; grasshoppers and crickets; bees, wasps, and ants; and bugs are also
common. Vegetation can also make up a significant volume (10 to 20%) of adult and
juvenile food. Diet of adults more diverse than that of juveniles. Tadpoles are
primarily herbivorous, consuming algae, plankton, and small plant materials (detritus)
from the substrate and the undersides of aquatic vegetation within the natal pond.
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-11
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Life History of Wood Frog
The wood frog is one of the smaller frogs inhabiting the Northeast, Like the spotted and
Jefferson salamanders, it is considered an obligate vernal pool amphibian species
because it requires (or, more accurately, prefers) vernal pools for breeding. Its life cycle
includes an aquatic larval stage and terrestrial juvenile and adult stages.
¦ Habitat—Entirely terrestrial except during the brief breeding season, when they
move to vernal pools and other aquatic habitats to mate and lay eggs. Preferred
terrestrial habitats are cool, moist upland woods, often far from water, but also found
in wooded swamps and bogs. In summer, are active day and night. Use brush piles
and other terrestrial features for cover, rather than seeking aquatic escape like some
other frogs. During winter, hibernate in upland areas under rotting wood, moss,
stones, or decaying leaf litter, never in water. Preferred breeding habitat are vernal
pools, however, will also utilize ditches, cattail swamps, gravel pits, slow-moving
streams, and other ephemeral habitats that lack fish.
¦ Home Range and Territoriality—Summer home range estimated for adults was
77.2 square yards (695 sq ft), with a range of 3.5 to 440 sq yd, not significantly
different between males and females. Suggested that many remain in a "home
area," at least during the summer. Availability of food was likely one of the principal
factors affecting home range size. Adults exhibit high degree of fidelity to their
breeding ponds each year; some juveniles may disperse to breeding ponds other
than the ones in which they were born. No information was found in the literature
regarding the territoriality of adult in terrestrial habitats. Frogs in general may defend
their shelters against other amphibians. Males are only somewhat territorial in the
breeding pools during the brief mating period.
¦ Food Habits and Diet—Food includes insects, especially beetles, flies, slugs, snails,
spiders, bugs, moth larvae, and earthworms. Tadpoles thought to be mostly
herbivorous feeders, consuming algae, decaying plants (detritus), and various
microorganisms scraped from aquatic plants present in the breeding pools. Tadpoles
found to be extremely effective predators of American toad eggs and hatchlings
inhabiting the same pool, despite the fact that American toad eggs and larvae are
thought to be toxic or distasteful to other organisms.
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-12
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
The assessment endpoint that is the subject of this section is the maintenance of local populations
of amphibians by ensuring the survival, reproduction, and development of local species. The
measurement endpoints used to evaluate the assessment endpoint are presented below.
Measurement Endpoints for Amphibians
¦ Semiquantitative sampling of larval amphibians in breeding habitats with different
sediment concentrations of stressors. Endpoints include species richness per
habitat type; species abundance; gross pathology; and body, tail, and total length
measurements.
¦ Surveys of vernal pools to quantify amphibians entering vernal pools and
determine breeding behavior and condition; egg laying, hatching success, and
larval growth and development; metamorphosis and emigration.
¦ Amphibian toxicity tests designed with exposure over a gradient of stressor
concentrations in site sediment. Toxicity endpoints include morphology of
embryos and juveniles, limb development, skin maturation, and tail resorption of
Rana pipiens.
¦ Gravidity of females; egg count; necrotic eggs; oocyte maturity; sperm count,
morphology, and viability; fertilization rate; embryo viability; hatching success;
mortality; and teratogenesis of Rana pipiens collected from the study area
compared with a reference area.
4.2.1 Amphibian Developmental Studies
Three separate site-specific studies were conducted to evaluate reproductive performance and
developmental effects in frogs exposed to PCBs and other COCs (two studies were conducted by
EPA, and one was conducted by GE). The studies focused on reproduction, early development,
and maturation (metamorphosis) in northern leopard frogs, and development and maturation in
wood frogs. These represent critical stages in amphibian life cycles and provide information on
the capacity of PCB and other COCs to disrupt the life-cycle processes (Sparling et al. 2000).
Various reproductive and developmental endpoints were assessed, such as gravidity of female
frogs, egg mass fertilization and hatching success, larval and metamorph mortality, growth, and
incidence of larval and metamorph malformation. Bioaccumulation of COCs in amphibian
tissue was also evaluated. The selection of individual test endpoints was made a priori and based
on previous investigations into the sensitivity of various life stages to organic contaminants.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC 4 13
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
4.2.2 Leopard Frog Study: EPA
This study was designed to evaluate both the reproductive fitness of adult leopard frogs in the
PSA, as well as monitor development of hatchlings through the metamorphosis stage. Adult
frogs (male and female) were collected from nine sampling areas within the PSA; these animals
were to be fertilized in the laboratory, with the resulting larvae to serve as the test organisms in
the developmental portion of the study. The reproductive condition of the adults was examined,
including body weight, sperm count, sperm morphology, ovary weight, egg count, and egg
maturity. However, the field-collected females possessed virtually no mature oocytes, so
fertilization was unsuccessful. In addition, the males exhibited a high proportion of malformed
sperm heads. The study design was modified to include the field collection of egg
masses/hatchlings at five of the nine original sampling stations; these stations were the only ones
that contained live leopard frogs for sampling. The egg masses/hatchlings were returned to the
laboratory for developmental evaluation. Larval mortality, malformation, growth, and incidence
of metamorphosis were recorded. Figure 4.2-3 illustrates the general study design.
Cross-over treatments were also included, wherein control egg masses were cultured in
contaminated sediment; the resulting larvae remained in the test media and were observed
through metamorphosis. A sediment spiking treatment was conducted to further investigate the
relationship between vernal pool media and larval development (effectively removing any
influence of COC transfer via the maternal pathway). The treatment involved culture of control
egg masses in reference site sediment that had been spiked with 30 mg/kg Aroclor 1260. COC
tissue concentrations were measured in samples representing various leopard frog life stages
(adult whole body, egg mass/ovary, and larvae whole body).
4.2.3 Wood Frog Study Design (EPA Studies)
The wood frog study was initiated in April 2000 and evaluated the growth, development, and
maturation of wood frogs. The study design combined laboratory exposures to vernal pool water
and sediment and assessment of field-collected animals, and consisted of three separate phases.
Figure 4.2-4 shows a simplified model of the study design, showing the life stage, exposure
scenario, and endpoints evaluated for each phase of the study.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-14
-------
LEOPARD FROG
REPRODUCTION STUDY
Evaluation of Adult Leopard
Frog Reproductive Condition
LEOPARD FROG
DEVELOPMENTAL STUDY
Laboratory Culture of
Field-Collected Egg Masses
Field Collection of Adults
(Male. Female) from 9
Breeding Areas of PSA
End points:
MALE:
Sperm Count,
Sperm Head Abnormality.
Testes Weight,
Body Weight
FEMALE:
Ovary Weight,
Egg Count. Necrotic Eggs,
Oocyte Maturity,
Body Weight
Main Study
Culture of Target and
External Reference Site Egg
Masses in Natal Pool
Sediment/Water
Supplemental
Investigations
I
Endpoints:
LARVAE:
Survival,
Growth,
Malformation
METAMORPHS:
Incidence.
Malformation
Cross-over Study
Exposure:
Reference Site
Larvae to
Contaminated Site
Media
Aroclor 1260
Spiking Study
Exposure:
Reference Site
Larvae in Clean
Sediment Spiked
with 30 mg/kg
Aroclor 1260
Endpoints:
Mortality,
Malformation,
Metamorphosis,
Growth
I
Tissue:
Adult Whole Body
•
Female Offal
(Whole Body Minus Egg
Mass/Ovary)
•
Egg Mass/Ovary
Tissue:
Late Larval/Metamorph
Tissue:
Larvae
Figure 4.2-3 General Model of Leopard Frog Vernal Pool (VP) Reproduction and Development Study
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-15
-------
Laboratory Culture of
Field-Collected Egg Masses
Supplemental
Investigations
Cross-over Study
Exposure:
Reference Site
Larvae to Target
VP Media
Target Site Larvae
to Reference VP
Media
1
Endpoints:
Mortality,
Malformation,
Growth
(Length, Weight),
Metamorphosis
~
Tissue:
Individual Larvae
Aroclor 1260
Spiking Study
1
Exposure:
Reference Site
Larvae to
Sediment Spiked
with 30 mg/kg
Aroclor 1260
Reference Site
Larvae in
Reference Site VP
Media ("Clean")
Endpoints:
Malformation,
Mortality,
Metamorphosis,
Growth (Length)
i
Tissue:
Individual Larvae
Main Study
Culture of Target and
Reference Site Egg
Masses in Natal Pool
Sedimenl/Water
Endpoints:
EGG MASSES:
Count, Weight.
% Fertilized. % Necrotic.
Hatching Success
LARVAE:
Mortality,
Malformation,
Developmental Stage.
Growth (Length)
Metamorphosis
METAMORPHS:
€
Malformation,
Growth (Weight)
Tissue:
Individual Egg
Mass. Larvae.
Metamorph
Four Field-sampling
Events. Each Several
Weeks Apart
Endpoints:
Malformation.
Growth (Length) at
Each Sampling Event
Tissue:
Individual Larvae
(from Events 1 & 3)
One Sampling Event
Endpoints:
Malformation,
Sex Ratio.
Growth (Weight),
Necropsy of Internal
Organs
Tissue:
Metamorph
Composite
Figure 4.2-4 General Model of Wood Frog Vernal Pool (VP) Reproduction and Development Study
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-16
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
Design of Wood Frog Vernal Pool Study
¦ Phase I: Egg masses collected from 11 vernal pools - 8 contaminated vernal pools
with low, medium, and high PCB concentrations (the ninth pool, with the second-
highest tPCB concentration, did not have any egg masses); and 3 reference pools.
Egg masses were cultured in the laboratory in sediment and water collected from the
associated contaminated and reference pool. Egg mass fertilization, egg counts, egg
weight, hatching success, larval growth, percent metamorphosis, and malformations
were the endpoints for this phase. Egg mass tissue and metamorph samples were
analyzed fortPCBs.
¦ Phase II: Following natural hatching of egg masses in the pools, tadpoles were
collected during 4 sampling events, each about 2 weeks apart to assess in situ
development. Endpoints for this phase included larval growth and malformations.
Larval tissue samples from the first and third collection event were analyzed for
tPCBs.
¦ Phase III: 50 wood frog metamorphs were collected from the 11 pools, and
individual weight, gender, and malformations were recorded. Tissue samples (one
composite per station) were analyzed for tPCBs and other COCs (PAHs, OC-
pesticides, metals, and dioxins/furans).
4.2.4 Context-Dependent Wood Frog Study: GE
The objectives of this study were to address whether larval density in the natal pool and PCB
exposure affects the survival and growth of wood frog larvae and whether the two factors
interact to influence larval developmental success.
The study was initiated in April 2001 and began with the collection of egg masses from five
vernal pools within the PSA. The pools included 8-VP-4, 8-VP-5, 23b-VP-2, 40-VP-l, and 40-
VP-3 (see Figure E.2-1). Approximately 21 egg masses were collected from each pond; the eggs
were then transported to a building at the GE facility in Pittsfield, MA, and maintained until they
hatched. Composite tissue samples were collected from each pool (approximately 200
hatchlings per sample) and analyzed for whole body PCB content. Although there were no
associated egg mass tissue samples, the study assumed that these hatchling (i.e., larvae) samples
served as an indicator of the transfer of maternal contaminants from parent to offspring.
Hatchling tissue tPCB concentrations were the basis for the selection of experimental treatment
groups. The study design specified three concentrations for evaluation in the field exposures: a
"low" (3.3 mg/kg) hatchling tissue tPCB concentration bounded by a "very low" (0.89 mg/kg)
and a "high" (11.2 mg/kg) concentration.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-17
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
The experimental design utilized these three levels of hatchling tissue PCB concentration, (i.e.,
very low, low, and high), and three levels of initial larval density (i.e., 200, 400, and 800). Each
of these combinations was exposed in two vernal pools (23b-VP-l and 23b-VP-2). These two
vernal pools were chosen because they supported natural populations of wood frogs, had very
low concentrations of tPCBs, and were believed to be deep enough to hold water longer than
most of the other floodplains in the PSA. A total of 18 experimental treatments were
established:
Density
Hatchling
PCB = 0.89
mg/kg
(40-VP-3)
Hatchling
PCB = 3.3 mg/kg
(23b-VP-2)
Hatchling
PCB = 11.2 mg/kg
(8-VP-5)
200 larvae
N=2
N=2
N=2
400 larvae
N=2
N=2
N=2
800 larvae
N=2
N=2
N=2
Approximately 200, 400, or 800 larvae were selected at random from each of the three vernal
pool's hatchling crop and placed in the in situ experimental enclosures (depending on each
treatment's assigned initial larval density). Concurrent to this field selection, three additional
sets of larvae were selected at random for tissue PCB analysis (composited samples).
The survival and growth of both tadpoles and metamorphs were assessed in the study. An
analysis of variance (ANOVA) approach was used to test whether any combination of initial
larval density, hatchling tissue tPCB concentration, or vernal pool sediment tPCB concentration
affected the number and weight of juveniles of each life stage at the end of the test.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-18
-------
1 4.3 EXPOSURE ASSESSMENT
2 The exposure assessment estimates the exposure of amphibians to tPCBs and other COCs in the
3 Housatonic River PSA (Figure 4.1-1). Exposures were assessed as either the COC
4 concentrations in sediment or water, or as the tissue body burdens that represent integrated
5 exposure from all sources. Routes of exposure were assessed to determine the contribution of
6 maternal transfer and the extent of bioaccumulation during various stages of development.
7 To match exposure data with effects-based measures, many of the sediment tPCB data
8 considered were from sampling conducted as part of the EPA studies that evaluated reproductive
9 performance and developmental effects in frogs exposed to tPCBs and other COCs. These
10 synoptic sediment samples are referred to as discrete sample data. Additional exposure data
11 included spatially weighted sediment tPCB concentrations. These were calculated using all
12 available sediment data, to develop average concentrations based on habitat types preferred by
13 wood frogs and leopard frogs during the reproductive period.
14 Sediment data were included in the GE wood frog study for the five pools of original egg mass
15 selection and for the two experimental ponds used for the larval development portion of the
16 study. No source or methods of sample collection was given for the sediment data; however,
17 concentrations were fairly similar to the spatially weighted tPCB values for a given vernal pool.
18 4.3.1 Selection of COCs for Amphibians
19 The contaminants initially considered in the amphibian exposure assessment were identified in
20 the Pre-ERA (Appendix B). The Pre-ERA included screening on a reach-by-reach basis and
21 subdivision of COPCs by major hydrological/geomorphological category.
22 A refined screening was conducted on the sediment data collected in support of the two EPA
23 amphibian developmental studies. The intent of this exercise was to further identify the
24 sediment COCs that were most relevant to local amphibian populations. The sediment COCs
25 retained for the amphibian assessment are presented below. Total PCBs were identified as
26 sediment COCs in all reaches of the PSA. A number of PAHs were retained throughout Reach
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-19
-------
1 5, and three PAH compounds were retained for Reach 6 (Woods Pond). Dioxins/furans were
2 retained in the PSA, as well as some metals.
3
4
5
6
7
8
9 Several additional contaminants (mainly pesticides) were determined in the Pre-ERA to be below
10 detection limits in sediment, but had detection limits that exceeded screening benchmarks. An
11 examination of the amphibian tissue data indicated that concentrations for most of the pesticides
12 of concern were below detection limits or below background. On this basis, and considering that
13 some pesticide detections may be attributable to laboratory interference artifacts, the entire suite
14 of organochlorine pesticides listed above was eliminated from further consideration in the
15 amphibian portion of the ERA.
16 4.3.2 Exposure Data
17 The approach used to characterize exposure to amphibians was based upon evaluation of
18 numerous data sources, including both sediment and water column COC concentrations and
19 amphibian tissue COC concentrations (Figure 4.1-2). Exposure data were evaluated to achieve
20 acceptable synopticity between exposure and effects endpoints, while also recognizing the need
21 to address spatial and temporal variability in the data, and inherent limitations in field sampling.
22 Concentration-response relationships were investigated using both (1) the single "most synoptic"
23 chemistry value paired with each toxicity endpoint (i.e., the average of the April and May 2000
24 sediment sampling events); and (2) a combined data set, based on all available relevant sediment
25 data and used to generate a spatially weighted average exposure concentration for vernal pools
26 and backwaters within the PSA. In most cases, the two approaches yielded similar results and
27 helped to reduce the uncertainty associated with the use of either particular source of data.
28 A community-wide assessment of amphibians in the PSA was conducted from 1998 to 2000
29 using visual and audio surveys, dip-netting, funnel-trapping, and pit-trapping techniques. A total
Contaminants of Concern for Amphibians
¦ Chlorinated organic compounds - tPCBs, dioxins/furans.
¦ Metals - Cadmium, chromium, copper, lead, mercury, silver.
¦ PAHs - Some individual PAH compounds, including low and high molecular
weight PAHs.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-20
-------
1 of 13 species were observed, including 4 salamanders, 8 frogs, and 1 newt. Amphibian
2 observation and capture data were combined with detailed habitat-type maps to predict likely
3 occurrences per species and to identify breeding, post-breeding, and wintering habitat.
4 4.3.3 Habitat Characterization
5 To provide the foundation for the problem formulation, the preliminary results of physical and
6 ecological investigations were used to evaluate potential habitat influences on amphibian
7 reproduction and development. Detailed characterization of the sediment type provides an
8 indication of the potential contaminant bioavailability, and the characterization of habitat allows
9 for the identification of physical and ecological characteristics that could affect the amphibian
10 endpoints.
11 Sediment characterization was based on examination of gross physical parameters (such as total
12 organic carbon [TOC] and grain size distributions) known to affect contaminant partitioning, and
13 therefore, bioavailability. In addition, potential effects of the Pittsfield Wastewater Treatment
14 Plant (WWTP) discharge were also examined with respect to the backwater habitats (to evaluate
15 potential impacts of the WWTP). Evaluation of effects from the WWTP was based on the
16 assumption that backwater habitats receive a portion of their sediment load from the river main
17 stem; therefore, influence from the WWTP may be carried into the backwaters.
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Vernal Pool/Backwater Substrate Evaluation
¦ Vernal pool and backwater sediment is much richer in organic matter than main
channel sediment, particularly in the upper Reach 5 area. The range of TOC for the
amphibian sampling areas was 1.7 to 59.1%, with a median TOC concentration of
7% in the wood frog vernal pools, and 6% in the leopard frog sampling areas.
¦ Grain size distributions for the wood frog sampling areas in the PSA were fairly
homogeneous and similar to reference sites. However, comparison of backwater
sediment characteristics at locations in the leopard frog study indicated that the
reference station had more coarse-grained sediment than did the sampling areas in
the PSA.
¦ The relationship between tPCB and TOC appeared qualitatively similar upstream and
downstream of the WWTP (no excess organic enrichment attributed to the WWTP,
thus, no confounding influence).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-21
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
4.3.4 Assessment of Sediment Chemistry
4.3.4.1 Sources of Sediment Data
Sediment data were collected for different programs, each with varying degrees of synopticity to
amphibian effects metrics. The sediment data that were used included all floodplain samples
collected from depths of 0 to 6 inches; all floodplain vernal pool samples collected in 1998 and
1999 to characterize floodplain sediment tPCB contamination; and all leopard frog and wood
frog samples collected as part of the EPA amphibian developmental studies. Use of all relevant
data for the purposes of the risk assessment helped to reduce the uncertainty associated with any
of the individual data sets (i.e., small-scale spatial variability, analytical variability, limited
spatial coverage).
Sediment Data Sources Used in Amphibian ERA
m Vernal Pool Characterization Study: Sediment sampling in 66 temporary and
permanent pools was conducted in 1998 and 1999 to characterize PCB
contamination in floodplain habitats of the PSA and to select "representative" pools
for amphibian developmental studies. PCB concentrations were fairly high;
approximately 78% of the pools evaluated within the PSA had sediment tPCB
concentrations > 5 mg/kg.
¦ Spatially Weighted Sediment/Floodplain Soil Concentrations: All available
surficial sediment tPCB data were combined using spatial weighting to estimate
average concentrations and exposure in wood frog vernal pools and leopard frog
ponds and backwaters. The spatial weighting integrated wetland habitat types into
the inverse distance weighting (IDW) procedure (Appendix C.3).
¦ Amphibian Developmental Studies: The sediment chemistry data with greatest
synopticity with the effects endpoints in the amphibian ERA were data collected in
conjunction with collection of the amphibian tissues (egg masses and larvae) during
the developmental studies.
¦ GE Wood Frog Study. The source of the sediment chemistry data used in the study
is not known, but the tPCB concentrations in the two experimental ponds known to be
low: < 0.3 mg/kg were similar to EPA concentrations. PCB concentrations in the 5
egg mass source ponds ranged from 0.5 to 30.8 mg/kg (0.3 to 32 mg/kg).
4.3.4.2 Distribution and Concentrations of PCBs
4.3.4.2.1 Vernal Pool Characterization Data
The 1998 to 1999 ecological characterization sampling resulted in more than 500 samples
analyzed for tPCBs. Many of these samples were collected from vernal pools, with less than
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC /I OO 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
10% collected from pool perimeters. Perimeter samples were collected in areas that, in wetter
years, would be submerged during the breeding period. A few of these samples were analyzed
for other COCs: six for PCB congeners, seven for PAHs, eight for metals, and eight for
dioxins/furans.
4.3.4.2.2 Spatially Weighted Average Exposure Concentrations
Many floodplain soil and sediment samples were collected over a several-year period in the PSA
(a subset of which are described above). All surficial (0-15 cm) data collected in the PSA,
combined with detailed habitat type maps and an understanding of site-specific hydrodynamics,
were used to estimate spatially weighted surficial PCB concentrations. Data from similar habitat
types were used in conducting the spatial weighting exercise, and were grouped into six similar
habitat types; sampling area boundaries were then incorporated as an overlay. The approach is
summarized in Appendix C.3, and the results are shown on Figures 2.5.5 through 2.5.11 of the
ERA.
For application in the amphibian risk assessment, an exposure point concentration (EPC) was
computed as the spatially weighted average (arithmetic mean) of the cells contained within each
leopard or wood frog sampling area boundary. The EPC represents the estimated average
concentration for a juvenile frog during development or an adult during breeding and foraging.
Juvenile frogs were assumed to move at random within their natal ponds or pools, and to be
equally exposed to every point within these areas; thus, the spatially weighted average served as
an appropriate representation of exposure. The individual amphibian endpoints (from the EPA
studies) for a given sampling station were then evaluated with respect to both the discrete
developmental study data and the spatially weighted EPCs.
Spatially weighted EPCs for the leopard frog sampling areas ranged from 0.4 to 44 mg/kg, with
six of the nine areas greater than 20 mg/kg tPCB. Spatially weighted EPCs for the wood frog
vernal pools ranged from 0.2 to 99.5 mg/kg; approximately two-thirds were greater than 10
mg/kg tPCB (see Figure 4.3-1).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
- 100%
0.05 0.1 1 5 10 20 30 40 50 60 70 80 90 100
Spatially Weighted EPC (mg/kg)
0)
O)
re
£
0)
o
0)
CL
0)
>
'-5
re
3
E
3
u
i iFrequency Cumulative %
Figure 4.3-1 Frequency Distribution and Cumulative Percentage of
Sediment tPCB Exposure Point Concentrations for 66 PSA
Temporary and Permanent Pools (Based on EPA Spatially
Weighted Data)
4.3.4.2.3 Data Collected in Support of the EPA Amphibian Developmental Studies
The sediment chemistry data with greatest synopticity with the effects endpoints in the
amphibian ERA were data collected in conjunction with the collection of amphibian tissues (egg
masses and larvae) during the EPA developmental studies. These sampling efforts included:
¦ Nine samples from the PSA and one reference site sediment sample for the 2000
leopard frog study. These samples were analyzed for PCBs (total and Aroclors) and
all other COCs.
¦ A total of 23 sediment samples collected for the 2000 wood frog developmental study
(collected in April and May). Samples were analyzed for PCBs (total and Aroclors),
dioxins/furans, and inorganics (cyanide and sulfide). The May 2000 samples were
also analyzed for other COCs (metals, PAHs, herbicides, and organophosphate [OP]
pesticides). Samples collected during the April event were later analyzed for
congeners.
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-24
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Figure 4.3-2 shows the sediment tPCB concentrations for both amphibian developmental studies.
The pools used as experimental ponds in the GE study also are shown on Figure 4.3-2; the pools
are 23b-VP-l and 23b-VP-2.
Summary of Occurrence of Other COCs in Sediment Data
¦ Metals concentrations at leopard frog sites were similar to the Muddy Pond reference
station; only Station E-1 had elevated metals concentrations. Metals in the wood frog
vernal pools were similar to the reference stations.
¦ Total PAH concentrations were elevated at most leopard frog stations, relative to the
reference site. PAH concentrations in the wood frog vernal pools were elevated at
three stations. Elevated PAH compounds included acenaphthylene,
benzo(a)anthracene, benzo(a)pyrene, benzo(b)fluoranthene, benzo(k)fluoranthene,
and pyrene.
4.3.5 Surface Water Chemistry Assessment
Surface water chemistry data is applicable to the amphibian ERA because the early life stages of
both frog species are entirely aquatic (water and sediment exposure). However, there is
uncertainty associated with extrapolating water chemistry effects in biota when the animals also
have the potential for exposure to sediment. Variability in water column concentrations over
time adds additional uncertainty. For these reasons, bulk sediment and porewater contaminant
concentrations are more commonly used as exposure metrics for toxicity testing. Because of the
uncertainty in relating water chemistry to effects on amphibians, the most relevant water data
were those collected in conjunction with effects measurements (from EPA studies). These data
were collected in conjunction with sediment sampling conducted for the two EPA amphibian
developmental studies (10 water samples for the leopard frog study, 22 water samples for the
wood frog study).
Total PCBs in water samples collected from both amphibian studies were reported as either
Aroclors 1254 and 1260, or both, with tPCBs used as the single PCB metric in data evaluation.
Sediment and water tPCB concentrations were correlated in both amphibian studies.
Surfacwater tPCB concentrations were lowest in the reference stations and the target stations
with low sediment tPCB concentrations (0.01 to 0.03 |j,g/L for the wood frog study, 0.013 |j,g/L
for the leopard frog study). Elevated water tPCB concentrations corresponded to amphibian
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC /I O ^ 7/11/2003
-------
Pittsfield
Wood Frog Stations
Leopard Frog Stations
Dalton
Washington
Lenox
Legend
Town Line
Houaatonic Valley Stale Wildlife
Management Area
Approximate 10 Year Flood Line
River\Water Line
Location of Pool
a
Nnt«;
o
2)
Base Map Information provided by the USKPA.
Placement of Town linen la approximate. Source USCS
Quadrangles.
Pool® surveyed Include vernal poola, aa defined
by the Maaaachuaetta Natural Heritage and Endangered Speclea
Program, and other water bodlea tbat contained, or could
contain breading amphibians.
Housatonic River
Vernal Pool Study Area
Pools Surveyed for
2000 Harm sylvalica Study
with sediment total PCR concentration
Figure 4.3-2
Total Sediment PCB Concentrations for Wood Frog Vernal
Pool Study {mean, n =2) and Leopard Frog
Reproduction/Development Study
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-26
-------
1 sampling areas with elevated tPCB concentrations (0.1 to 0.47 |j,g/L for the wood frog study,
2 0.03 to 0.41 |j,g/L for the leopard frog study).
3 For the most part, all other COCs measured in the water samples were screened out of the
4 amphibian risk assessment and, therefore, out of further data analysis. No water data were
5 reported in the GE wood frog study report.
6
7
8
9
10
11
12
13
14 4.3.6 Tissue Chemistry Assessment
15 There was less tissue chemistry data from the two EPA studies than abiotic data, and these data
16 generally did not include replication due to limited volumes of tissue available for chemical
17 analysis. Nevertheless, the available data provide indications of the site-specific bioavailability
18 of the COCs.
19 Figure 4.3-3 presents the distribution of tPCB concentrations by sampling location and tissue
20 type for adult samples collected at the leopard frog stations. Adult whole body tissue
21 concentrations ranged from 0.15 to 5.4 mg/kg at PSA sites; Figure 4.3-3 shows whole body
22 samples only from sites with associated offal and egg mass/ovary samples. PSA site offal
23 (whole body minus ovary/egg mass) tPCB ranged from 0.02 to 2.6 mg/kg; egg mass/ovary
24 samples ranged from 0.24 to 45.1 mg/kg. Larval tissue concentrations ranged from 0.05 to 1.4
25 mg/kg tPCB. For all tissue sample types, there was a direct relationship between sediment tPCB
26 concentration and tissue tPCB concentration.
27 Tissue samples from the cross-over and Aroclor 1260-spiked treatments confirmed the
28 importance of the sediment uptake pathway; control animals raised in PSA site media had a
29 tissue concentration of 0.37 mg/kg, while the control animals raised in reference media had a
Elimination of Other COCs in Surface Water Data
¦ All metals screened in the sediment assessment were measured as non-detects in
the two water samples except for zinc. The measured value for zinc (17 |ag/L) is
below both the EPA federal and British Columbia provincial criteria for protection of
aquatic life. Therefore it was not retained as a COC.
¦ All PAHs were screened out of the water assessment because they were not
detected.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-27
-------
1000
O)
O)
E
m
o
o.
100
10
0.1
0.01
Pooled ref.
W-1
W-1
W-9a W-9a W-7a W-7a
Vernal Pool ID
^ Adult offal 0 Egg mass/ovary
4 Spatially weighted mean tPCB in sediment ¦ Adult whole body
EW-3 EW-3 W-6
_l Sediment [tPCB]
Note: Data include adult whole body and adult offal (with associated egg mass); male frogs were included in adult whole body tissue sample and data are
shown for informational purposes only. N=2 for offal/egg mass samples (except for station W-6; n=l). N=1 for adult whole body samples.
Figure 4.3-3
Comparison of Leopard Frog Tissue Samples to Sediment tPCB Concentrations (Reproductive
Study Data and Spatially Weighted Data)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-28
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
mean body burden of 0.056 mg/kg (based on two treatments). Animals in the spiking study had
a body burden of 0.55 mg/kg (spiked treatment) and 0.007 mg/kg (control treatment).
Sources of Amphibian Tissue Chemistry Data
m Leopard frog study (EPA): Nine composite whole body samples were analyzed for
tPCBs. Nine individual offal samples with associated egg mass/ovary tissue
removed were analyzed for tPCBs (Figure 4.3-3). Five offal samples were analyzed
for other COCs. Five composite larval samples were analyzed for tPCBs. Six
composite larval samples were analyzed for tPCBs from the cross-over and Aroclor
1260 spike treatments. All tissue samples confirmed contaminant bioavailability from
sediment.
¦ Wood frog study (EPA): 15 egg mass samples, 13 Phase I metamorph samples,
20 Phase II larval samples (from two discrete sampling events), 10 Phase III
metamorph samples (five of these samples analyzed for other COCs) were analyzed
for tPCBs. Four larval tissue samples from cross-over treatments and one larval
tissue sample from the Aroclor 1260 spike treatment were also analyzed for tPCBs.
Results of analysis showed that PCBs are bioavailable at all life stages, but that egg
mass tissue concentration is not related to sediment tPCB . However, all other tissue
samples showed a trend of increasing contaminant uptake with increasing exposure
concentration and duration in the vernal pools.
¦ Wood frog study (GE)\ Composite tissue samples were collected at two events
during the juvenile period: at the hatchling stage (1-2 days post-hatch) and at the
early larval stage (approximately 11-12 days post-hatch), one from each of the five
vernal pool stations during the first event, and 4 composites during the second event
(Pool 10.9 [EPA 40-VP-1] was not sampled).
Wood frog tissue concentrations (EPA study) across various life stages are included in Figure
4.3-4. PSA site egg mass tissue concentrations ranged from 0.01 to 2.1 mg/kg, but were
unrelated to sediment tPCB concentrations. This was not surprising, given that the egg mass
concentration would be more representative of the female's exposure prior to moving into the
pool to breed. Phase I metamorph (laboratory-cultured) tPCB concentrations ranged from 0.06
to 5.83 mg/kg in the PSA animals and were related to sediment tPCB concentration. Phase II
larval tissue samples (from collection event 1) ranged from 0.28 to 3.44 mg/kg and were not
related to sediment tPCB concentration. However, later-stage larval samples (from Phase II,
collection event 3) were related to sediment tPCB concentrations (tissue tPCBs ranged from 0.09
to 10.4 mg/kg). Phase III metamorph (field-collected) samples ranged from 0.13- to 15-mg/kg
tPCB and also were related to sediment tPCB concentrations.
In the cross-over and spiking studies, tissue tPCB concentrations were elevated in all samples
exposed to PCBs and were much lower in the reference exposures.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC /I OO 7/11/2003
-------
100
10
O)
£ 1
O)
E,
m
o
~ 0.1
0.01
0.001
H—Mean VP tPCB
Spatially weighted
mean tPCB
X Phase III
Metamorphs
~ Phase II, Event 3
+ Phase II, Event 1
A Phase I
metamorphs
~ Phase I egg mass
Vernal Pool ID
Note: No tissue samples for 39-VP-l; no Phase II, event 3 tissue data for WML-2; Phase I egg mass tissue values for WML-1, WML-2, and WML-3
were non-detect (ND) - values shown reflect detection limits (0.007, 0.004, and 0.008, respectively).
* WML-1 Phase III metamorph sample is anomalous.
Figure 4.3-4 Comparison of tPCB Concentrations in Tissue (in Various Phases of the Wood Frog
Developmental Study) with Mean Vernal Pool and Spatially Weighted Mean tPCB
Concentrations in Sediment
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-30
-------
Hatchling tissue concentrations in the GE study ranged from 0.26 mg/kg to 11.2 mg/kg; no other
COC tissue data were included in the report (n = 5 composites). The hatchlings from three of the
five pools were then selected for placement in the in situ vernal pool enclosures. Just prior to
placing the test organisms in the enclosures, larval tissue samples representing animals from the
three hatchling test concentrations were analyzed for PCB body burden. Larval tissue
concentrations ranged from 1.4 to 7.2 mg/kg tPCB. A fourth larval tissue sample was also
analyzed and contained 6.1 mg/kg tPCBs, although hatchlings from this location were not
evaluated in the experimental treatments.
4.4 EFFECTS ASSESSMENT
The effects assessment for amphibians (Figure 4.1-3) emphasizes the site-specific field
investigations; because these studies provided direct indications on the bioavailability, toxicity,
and effects of site-specific COCs. Both toxicity assessments (i.e., laboratory toxicity, in situ
toxicity) and the community evaluations (i.e., amphibian community composition) were
compared to appropriately matched field references to determine whether the exposed sites in the
Housatonic River vernal pool/backwater habitats exhibited biological impairment.
The effects assessment also provides an overview of the literature on the effects of tPCBs and
other COCs to survival, growth, and reproduction of amphibians. At the time of the literature
review, there were virtually no amphibian studies available that contained paired sediment or
water exposure data with effects data. However, there were sufficient studies available that
evaluated tissue PCB concentrations with response data. A total of five different frog species
were used in the studies, including both leopard frogs and wood frogs. There were a total of 18
"no effect" measurements (ranging from 0.02 to 11.2 mg/kg tPCB wet weight) and 11 "effect"
measurements (ranging from 0.96 to 128 mg/kg tPCB wet weight).
Detailed evaluation of concentration-response relationships, for both toxicity assessments and
amphibian community structure assessments, are not included in this section. These are
presented in the risk characterization section (Section 4.5). Accordingly, this section is limited
to a discussion of differences between effects at the exposure locations and control and/or
reference locations.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-31
-------
4.4.1 Sediment Toxicity
4.4.1.1 Data Evaluation
The two EPA amphibian studies contained over 200 individual metrics that were collected to
track reproduction and development through multiple life stages; however, because some of
these metrics were redundant, they were not used in the ERA. Metrics that were the most
biologically relevant and that provided independent measurements of an effect were selected to
assess the overall degree of effects to amphibian communities. The majority of endpoints
selected for detailed discussion in the ERA were determined on an a priori basis based on
hypotheses formulated during the study design; others were selected based on the patterns of
effects that emerged during initial exploration of the data. Approximately 50 discrete endpoints,
representing each major life stage of leopard and wood frogs, were initially evaluated for the
ERA data analysis. The approach used to evaluate these endpoints was based on two objectives:
¦ Determination of relative sensitivity of various life stages.
¦ Evaluation of COC concentration-response relationships.
The wood frog data were fairly well suited to the use of inferential statistics in the evaluation of
relationships. The leopard frog data, however, required a more qualitative approach.
4.4.1.1.1 Leopard Frog Data
Inferential statistics were not deemed appropriate for analysis of the leopard frog effects data, for
three reasons:
¦ The magnitude of the effects observed made the use of statistical analyses less
appropriate than for the wood frog data. The leopard frogs exhibited much more of a
threshold effect response for various endpoints; fairly low responses in the
reference/control treatments, and very high responses in the exposure treatments.
Such data distributions are better evaluated through visual interpretation (figures) and
examination of average response data (for a given endpoint) at each station.
¦ Sample sizes were small, primarily because of limited availability of test animals in
the field (often, n = 5 or n = 6), which makes the application of many standard
statistical tests inappropriate.
¦ Because fertilization of the field-collected females was unsuccessful due to the lack
of mature oocytes, there was no biological relationship between female PCB body
burdens or reproductive tissue endpoints (i.e., percent Stage VI oocytes, percent
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC /I ^O 7/11/2003
-------
malformed sperm cells) and larval/developmental effects endpoints (i.e., incidence of
metamorphosis).
In summary, the evaluation of leopard frog concentration and response data was limited to a
more qualitative presentation. Non-parametric Spearman's correlation tests were conducted on
adult tissue samples and sediment tPCB concentrations (sufficient paired sediment/tissue data),
to determine the potential influence of environmental exposure on contaminant uptake into
animal tissue. Data from the cross-over and spiked studies were evaluated via hypothesis testing
for differences among groups (always a two-sample comparison of a control treatment and a
target treatment of interest).
4.4.1.1.2 Wood Frog Data: EPA Study
Tests for correlation were used to determine relationships between variables of interest.
Evaluation of the distribution of much of the exposure/response data revealed that the
distributions were not normal, thus precluding the use of parametric statistics and reducing the
confidence that could be placed in such approaches as simple linear regression.
The non-parametric Spearman's rank-order correlation coefficient, rs (Spearman's correlation),
was selected. The Spearman's correlation is less sensitive to data "outliers," thereby reducing
their influence when evaluating a relationship. The choice of calculating a Spearman's
correlation provided a conservative and robust approach to the data analysis.
The overall consistent pattern of significant relationships between biological effects and PCB
concentration at sensitive life stages, combined with corroborating literature-based effects data,
provided a check against spurious correlations.
As with the leopard frog data, the cross-over and spiked study data were evaluated via hypothesis
testing for differences among groups (a two-sample comparison of a control treatment and a
target treatment of interest).
4.4.1.1.3 Wood Frog Data: GE Study
Multivariate and univariate analysis of variance (MANOVA and ANOVA) were used to evaluate
the interaction of vernal pool, hatchling tissue tPCB concentration, and initial larval density on
the survival and growth of the test organisms. In addition, correlation analyses were used to
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
-------
determine whether there was a relationship between hatchling and larval tissue concentration and
the sediment tPCB concentration of the vernal pool from which the egg masses were collected.
4.4.1.2 Results
4.4.1.2.1 Leopard Frog Study
Reproductive Fitness: Adult male and female leopard frogs (and some juveniles of both sexes)
were collected from the nine contaminated sampling areas in the PSA and transported to Fort
Environmental Laboratories, Inc. (FEL). No leopard frogs were collected at the three reference
areas; therefore, control animals purchased from a commercial supplier (Carolina Biological
Supply, CBS) were used. These frogs were collected in Vermont directly upon order, shipped to
CBS, and then forwarded to FEL (formerly part of The Stover Group).
It is not known why leopard frogs were not available in the reference areas or why no eggs were
found during the year this study took place. They had been observed in the prior year, and it was
assumed that there was a large enough population present for sampling. This assumption may
have been incorrect. Because of the timing of collection and the limited number of reference
areas with suitable potential habitat, and because soil and sediment samples confirmed only
background concentrations of tPCBs and COCs, it was necessary to obtain outside control frogs.
The timing of collection of adult leopard frogs from the target stations coincided with the normal
onset of reproductive receptiveness and initiation of breeding activity. Adult specimens were
collected between March 25 and April 22, 2000. Surface water temperatures in the PSA were
approximately 8 to 10° C at this time (WESTON 1998 - 1999). These temperatures represent
the ideal environmental "triggers" for the frogs to emerge in the early spring and gather in
breeding areas. Typically, males begin chorusing when water temperatures reach approximately
8° C, with oviposition peaking when water temperatures reach 10° C (Gilbert et al. 1994). Hine
et al. (1981) reported the occurrence of breeding when water temperatures reached or slightly
exceeded 10° C in Wisconsin ponds.
After collection and transport to the laboratory and acclimatization for 24 hours, female frog
gravidity was recorded, and mature (gravid) females were hormonally induced to super-ovulate
egg masses; fertilization was then attempted on these egg masses using sperm collected from
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-34
-------
male frogs from the same sampling area. The number of eggs produced per female, rates of
necrosis, and oocyte developmental stage distribution were determined. Sperm count,
morphology, and overall viability were also assessed. The eggs were monitored for fertilization,
morphology, and coloration.
Male body weight and sperm count did not appear to be related to exposure media tPCB
concentrations; however, there was a strong inverse relationship between incidence of sperm
head abnormalities and sediment tPCB concentrations (Table 4.4-1 and Figure 4.4-1). Sperm
head abnormalities may have contributed to the low fertilization success of the eggs from the
field-collected females. Tissue data were not collected on the male frogs, so a comparison of
body burden to sediment tPCB concentration or percent abnormal sperm heads could not be
conducted.
Data on female frog reproductive fitness were limited because of small sample sizes (see Table
4.4-2). Findings regarding reproductive fitness of the female leopard frogs include:
¦ Adult leopard frog specimens collected from contaminated sampling areas in the PSA
showed marked signs of reproductive stress.
¦ None of the females collected from Sites E-5, W-9a, W-8, and E-l (37.0, 4.3, 120.0,
and 160.0 mg/kg sediment tPCBs, respectively) were found to be gravid (eggs mature
enough for successful fertilization).
¦ Few of the PSA sites produced female specimens that possessed any biologically
significant quantity of Stage VI oocytes (mature eggs capable of fertilization), with
the exception of Station W-7a (Figure 4.4-2). Immature oocytes (< Stage III) were
observed in mature female specimens collected from all PSA sampling areas,
however developing oocytes were found in specimens from Sites W-7a, W-4, EW-3,
and W-l (18.0, 0.5, 30.0, and 0.2 mg/kg sediment tPCBs, respectively). Therefore,
the lack of success in artificially fertilizing oocytes from contaminated site specimens
was not surprising, and appeared to be the primary limiting factor in the reproductive
dysfunction observed in the contaminated site specimens evaluated from the PSA.
¦ Even though more advanced oocytes were found in specimens containing greater
concentrations of ovary tPCBs, only a few Stage VI oocytes were found, indicating
that the final stage of maturation that involved hormonal induction of the final
preparatory event known as germinal vesicle breakdown (GVBD) may have been
inhibited. Further, since oogenesis and, to a greater extent, egg maturation, were
inhibited in ovaries with tissue residues of as low as 0.3 mg/kg, the threshold for
inhibition appeared to be below this residue level.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-35
-------
R2, R3
I I mean % abnormal sperm heads
Sediment tPCB —A — Spatially weighted mean tPCB
Figure 4.4-1 Comparison of Percent Abnormal Sperm Heads (Mean) from
Male Adult Chemical Analysis Leopard Frogs, with Mean
Sediment tPCB and Spatially Weighted Mean tPCB
I i Stage VI oocytes (%)
Sediment PCB (mg/kg) — t —Spatially weighted mean tPCB
Figure 4.4-2 Comparison of Mean Percent of Oocytes at Stage VI (Mature)
for Female Leopard Frogs, with Mean Sediment tPCB and
Spatially Weighted Mean tPCB
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-36
-------
Table 4.4-1
Summary of Male Adult Leopard Frog Reproductive Health
Sampling
Area ID
Sediment tPCB
(mg/kg)*
Mean Total
Water PCB
(|J.g/L)
Mean Male
Body Weight (g)
(SD)
Mean Testes
Weight (% of
Total Body
Weight) (SD)
Mean %
Abnormal
Sperm Heads
(SD)
Mean Sperm
Count x 106/ g
Gonad Tissue
(SD)
tPCB
Sp. Wt.
PCB
R1
-
-
-
40.6 (6.29)
0.176 (0.016)
0.42 (0.32)
5.38 (1.25)
R2
-
-
-
40.8 (4.75)
0.126 (0.083)
0.89 (0.51)
7.88 (5.94)
R3
-
-
-
35.9 (1.48)
0.179 (0.039)
2.36 (0.49)
3.61 (0.56)
pooled Rl,
R2,R3
-
-
-
39.2(4.91)
0.162 (0.051)
1.14 (0.95)
5.60 (3.40)
W-l
0.15
0.4
0.013
36.5 (4.08)
0.163 (0.044)
4.33 (0.76)
5.06 (2.86)
W-4
0.46
0.4
0.013
37.2 (NA)
0.140 (NA)
3.15 (NA)
4.48 (NA)
W-9a
4.3
7.5
0.013
40.7 (4.82)
0.147 (0.023)
8.26 (2.39)
2.42 (0.87)
W-7a
18
27.6
0.03
34.2 (4.93)
0.106 (0.040)
12.0 (4.72)
3.52 (1.67)
EW-3
30
23.8
0.41
31.4 (4.28)
0.112(0.026)
49.5 (10.8)
7.87 (0.28)
W-6
42
21
0.22
39.6 (5.28)
0.138 (0.058)
37.3 (6.26)
2.01 (1.30)
W-8
120
43.5
0.14
34.8 (11.8)
0.089 (0.010)
42.7 (2.39)
6.08 (0.20)
E-l
160
26.6
0.24
41.4(7.39)
0.113 (0.075)
14.3 (4.43)
3.83 (3.26)
* tPCB = Value from amphibian developmental studies; Sp. Wt. PCB = mean tPCB for each sampling area based on spatial
weighting of sediment data.
SD = Standard deviation.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
-------
Table 4.4-2
Summary of Female Adult Leopard Frog Reproductive Health
Sampling Area
ID
Sediment tPCB
(mg/kg)*
Total Water
PCB (jig/L)
Mean Female
Tissue PCB
(mg/kg) (SD)
Mean Egg
Mass/Ovary
PCB (mg/kg)
(SD)
Mean Female
Body Weight
(g) (SD)
Proportion
Gravid
Mean Ovary
Weight (% of
Body Weight)
(SD)
Mean Total
Egg Count
(SD)
Mean % of
< Stage in
Oocytes (SD)
Mean % of
Stage VI
Oocytes (SD)
tPCB
Sp. Wt.
PCB
R1
-
-
-
0.012 (0.014)
0.036 (0.007)
77.41 (5.55)
4/4
30.78 (3.87)
1264 (908)
23.01 (32.75)
56.75 (25.20)
R2
-
-
-
0.011 (0.013)
0.012 (0.005)
79.32 (15.26)
4/4
23.32 (5.71)
2811 (1342)
27.82 (8.93)
65.25 (7.22)
R3
-
-
-
0.017(0.010)
0.024 (0.003)
76.10 (12.33)
5/5
32.14 (5.44)
119(49)
0
89.46 (6.16)
Pooled Rl,
R2,R3
-
-
-
0.015 (0.010)
0.024 (0.011)
77.61 (11.14)
13/13
29.01 (6.14)
1174 (1339)
14.62 (21.85)
72.50 (20.78)
W-l
0.15
0.4
0.013
0.022 (NA)
0.240 (NA)
48.83(18.28)
1/5
4.56 (4.59)
1008 (961)
84.86 (30.29)
0.44 (0.89)
W-4
0.46
0.4
0.013
-
43.26 (10.69)
2/2
6.23 (0.36)
1038 (110)
28.38 (7.90)
2.22 (0.60)
W-9a
4.3
7.5
0.013
1.260 (NA)
45.086 (NA)
51.59 (12.04)
0/3
3.88 (1.28)
1238 (799)
99.70 (0.52)
0
W-7a
18
27.6
0.03
1.407 (1.636)
14.219
(17.801)
52.43 (12.87)
5/5
21.35 (2.88)
2918 (1663)
20.00 (44.72)
4.88 (3.04)
EW-3
30
23.8
0.41
-
55.73 (4.97)
2/3
13.04 (5.67)
419 (436)
67.93 (29.50)
1.02 (1.37)
E-5
37
19.6
0.043
-
50.33 (NA)
0/1
1.25 (NA)
177 (NA)
100
0
W-6
42
21
0.22
0.386 (NA)
9.477 (NA)
57.35 (12.66)
2/5
4.65 (1.99)
2401 (841)
97.52 (3.39)
0.06 (0.13)
W-8
120
43.5
0.14
-
52.66(11.66)
0/1
5.03 (NA)
307 (NA)
100
0
E-l
160
26.6
0.24
-
37.91 (9.73)
0/4
3.17(0.97)
1168 (733)
99.26 (1.55)
0
* tPCB = Value from amphibian developmental studies; Sp. Wt. PCB = mean tPCB for each sampling area based on spatial weighting of sediment data.
SD = Standard deviation.
NA = Not applicable; only 1 replicate.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-38
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
¦ Only two PSA sampling areas had frogs in the same body size range as the control
frogs. Purchased frogs (collected in Vermont) were larger than most contaminated
site frogs, and had relatively larger ovaries. Vermont is approximately 45 miles north
of the PSA, and the local climatic conditions are similar to those found in western
Massachusetts. Because the control frogs were collected in a similar climate, region,
and timeframe, differences in body sizes were not expected.
¦ Juvenile frogs collected for this portion of the study were not included in the
assessment of overall female reproductive fitness.
In summary, the evidence supporting impairment of reproductive fitness related to PCB exposure
includes:
¦ Low rates of egg maturation.
¦ Poor egg mass fertilization from field-collected female frogs.
¦ High incidence of sperm head abnormalities in males from vernal pools with high
sediment PCB concentrations.
Few contaminated site female oocytes reached Stage VI (only 381 of 10,611 eggs from all
contaminated sites; these eggs came from only 4 frogs), whereas control eggs totaled 5,785, and
more than half reached Stage VI (3,653 eggs). Rosenshield et al. (1999) found a significant
negative correlation between sediment PCB concentration and hatching success of green frog
(Rana clamitans) and leopard frog embryos exposed along a PCB gradient.
Developmental Endpoints: Because of poor egg fertilization success, where most of the females
from the PSA sampling areas were reproductively unfit, the study design was modified to
include the field collection of egg masses from the leopard frog sampling areas, and to raise them
in the laboratory, as was done for wood frogs. Each contaminated area and reference area was
surveyed for egg masses and hatchlings, which were then collected at five of the nine
contaminated sampling areas (EW-3, W-6, W-l, W-7a, and W-4). No egg masses were found in
the two locations with the highest sediment PCB concentrations (Stations W-8 and E-l), or at the
reference areas. Therefore, control egg masses were obtained from the females leopard frogs
from CBS fertilized in the laboratory.
The four larval endpoints measured in the study (mortality, metamorphosis, malformation, and
growth) were not evaluated for the same amount of time; larval malformation and growth
endpoints had shorter test durations than larval mortality and metamorphosis endpoints.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC /I ^O 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Differences in test duration for the four endpoints were normalized to a common test duration to
make comparisons among the endpoints. To perform comparisons among the four endpoints,
data for the mortality and metamorphosis endpoints were taken from the last day that
malformation and growth were measured. Additional endpoints included:
¦ Mean percent metamorphosis at end-of-test (EOT).
¦ Mean percent larval mortality at EOT.
¦ Days to reach Gosner developmental Stage 26 (±1).
¦ Developmental stage reached at EOT.
Gosner Stage 26 is the point in development when tadpoles go from a relatively immobile
embryo to an active, feeding tadpole. EOT was used to designate the shorter test duration (the
last day that larval growth and malformation were measured). Final test duration refers to the
last day that larval mortality and metamorph data were recorded.
Table 4.4-3 and Table 4.4-4 show the responses of the four developmental endpoints from the
main study, as well as the results of the cross-over and Aroclor 1260-spiked treatments.
Findings with respect to the leopard frog developmental endpoints are presented in the following
paragraphs.
Mortality was high (85 to 100%) for larvae raised in contaminated sediment regardless of PCB
concentration, when compared to R3 control larvae (44%) raised in Muddy Pond sediment and
water. Effects, in the form of high larval mortality, occurred at all sampling sites in the PSA.
The incidence of larval malformations was low (0 to 3.4%) in sampling areas with tPCB
concentrations below 1 mg/kg, and higher (46 to 54%) in sampling areas with tPCBs greater than
20 mg/kg. Malformations were similar to those observed in studies of exposure of other frog
species, including other ranids and the South African clawed frog (Xenopus laevis) to PCBs and
similar contaminants (Birge et al. 1978; Eisler and Belisle 1996; Gutleb et al. 1999, 2000). Thus,
the effects observed in the leopard frog study appeared to be characteristic of exposure to these
PCB or PCB-like contaminants.
Larval developmental delay was observed in leopard frogs raised in contaminated sediment.
There was an obvious relationship between sediment tPCBs and the amount of time for the
larvae to reach Stage 26 (±1) (Figure 4.4-3). Control larvae reached this stage in 13 days,
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-40
-------
Table 4.4-3
Summary of Leopard Frog Larval Development Endpoints Data at End-of-Test
Sediment tPCB
(mg/kg)c
Living
Larvae
at End
of Test
(EOT)
Mean %
Larval
Mortality
at EOT
Mean % Larval
Malformed
(Based on
Surviving
Larvae)
Days to
Reach
Stage
25 (±2)
Mean
Larval
Growth:
Length at
EOT (cm)
Sampling Area IDa
Test
Durationb
Initial
Larval
Count
tPCB
Sp.
Wt.
PCB
Water
tPCB
(Mg/L)
Mean %
Metamorph
at EOT
Develop.
Stage at
EOT
EW-3
22
13
30
23.8
0.41
NA
100
NA
NA
NA
22
NA
W-6
91
98
42
21
0.22
11
88.8
0
54.5
91
25-27
4.48
W-l
105
105
0.15
0.4
0.013
14
86.7
0
0
49
37-40
3.88
W-7a
105
105
18
27.6
0.03
16
84.8
0
45.8
105
25-27
4.47
W-4
111
210
0.46
0.4
0.013
15
92.8
0.83
0
55
36-37
4.55
MP Ref.
69
160
0.04
-
0.013
125
21.8
0
3.4
13
38
4.30
Cross-
over
Study
R1 Target
91
40
120
-
0.14
19
52.5
2.5
25.9
91
26
4.06
R3 Targetd
69
80
120
-
0.14
70
12.5
0
26.1
69
32
4.39
Referenced
69
160
0.04
-
0.013
125
21.8
0
3.4
69
38
4.3
Aroclor
1260
Spike
Study
Spiked
23
80
30
-
-
57
28.7
NA
29.8
NA
NA
NA
Control
23
80
0.04
-
-
59
26.2
NA
0
NA
NA
NA
Sampling areas arranged in order of increasing test duration.
bDurations vary for endpoints; larval malformation and growth had shorter test durations than larval mortality/metamorphosis. Test durations shown here are for
the malformation/growth endpoints. Last day of test duration shown here is used as end-of-test (EOT) for a given sampling area.
ctPCB = Value from amphibian developmental studies; Sp. Wt. PCB = mean tPCB for each sampling area based on spatial weighting of sediment data,
treatments used in hypothesis testing.
NA = Not applicable.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-41
-------
Table 4.4-4
Summary of Leopard Frog Larval Development Endpoints at Final Test Duration
Sampling Area IDa
FEL
Site
ID
Final
Test
Duration6
Initial
Larval
Count
Sediment tPCB
(mg/kg)c
Living Larvae
and
Metamorphs
at End of Test
(EOT)
Final Mean %
Larval
Mortality
Living
Metamorphs
(EOT)
Mean % Metamorph
(at Final Test
Duration)
tPCB
Sp. Wt.
PCB
EW-3
37
28
13
30
23.8
NA
100
0
0
W-6
35
128
98
42
21
8
91.8
3
3.1
W-l
39
142
105
0.15
0.4
9
91.4
0
0
W-7a
34
142
105
18
27.6
13
87.6
1
0.95
W-4
36
148
210
0.46
0.4
9
95.7
2
0.95
MPRef.
40
106
160
0.04
-
90
43.8
9
5.6
Cross-
over
Study
R1 Target
-
128
40
120
-
19
60.0
2
5.0
R3 Targetd
-
106
80
120
-
70
32.5
7
8.8
Reference"1
-
106
160
0.04
-
125
43.8
6
5.6
Aroclor
1260
Spike
Study
Spiked
-
23
80
30
-
57
28.7
NA
NA
Control
-
23
80
0.04
-
59
26.2
NA
NA
Sampling areas arranged in order of increasing test duration.
bTest durations for the larval mortality/metamorphosis endpoints were longer than for the larval malformation/growth endpoints. Endpoint measures in this table
correspond to final test durations.
ctPCB = Value from amphibian developmental studies; Sp. Wt. PCB = mean tPCB for each sampling area based on spatial weighting of sediment data,
treatments used in hypothesis testing.
NA = Not applicable.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-42
-------
120 i
100
80
¥, 60
40
20
Control
Vl-4 W-7a EW-3
Leopard Frog Sampling Area
100%
Mortality
-- 40
35
30 3
O)
E
25 „
o
a.
20 1
15 ^
-- 10
5
i i Days to reach Stage 26 (+/-1) k\\\ Final Developmental Stage at End-of-Test
—¦—Mean sediment tPCB —± —Spatially weighted mean tPCB
Test durations:
Control 69 days
W-l ' 105 days
W-4 111 days
W-7a 105 days
EW-3 22 days
W-6 91 days
Figure 4.4-3 Days to Gosner Developmental Stage 26 (±1) and Final
Developmental Stage Reached at End-of-Test, with Sediment
tPCB and Spatially Weighted Mean tPCB (FEL 2002b): 2000
Leopard Frog Study
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-43
-------
1 whereas larvae from sampling area W-7a (18 mg/kg sediment tPCB) took 105 days, and larvae
2 from sampling area W-6 (42 mg/kg sediment tPCB) took 91 days. Even given the uncertainty
3 that comes with comparing the control animals to the contaminated larvae, this difference
4 appears too large to be attributable to genetics alone. In addition, larvae from the two most
5 contaminated stations (W-7a and W-6) never developed beyond the Stage 26 (±1) endpoint.
6 Extended time to metamorphosis, and a low incidence of metamorphosis, was observed in
7 juveniles from the PSA sites. Leopard frog larvae normally spend 63 to 90 days as tadpoles
8 (DeGraaf and Rudis 1983; Taylor and Kollros 1946; Gosner 1960) before metamorphosis. The
9 test durations for the four treatments with larvae surviving beyond day 28 exceeded the time
10 period of normal development, and metamorphosis was expected before the end of the test
11 durations. Few larvae reached metamorphosis.
12 The study results indicate that some endpoints demonstrate a very strong toxic response to PCBs,
13 even at low concentrations.
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Summary of Northern Leopard Frog Toxicity Study
Adult Reproductive Fitness Endpoints:
¦ Both male and female adult frogs showed signs of reproductive stress, with the
females showing more severe effects. Males exhibited a high incidence of
malformed sperm in the higher-sediment tPCB sites (up to 50%). Females had
virtually no mature eggs (Stage VI, which the eggs must reach in order for
fertilization to occur). Incidences of immature oocytes (Stage III or smaller) were
high in the sites with high concentrations of sediment tPCB (up to 99% Stage III).
Developmental Endpoints:
¦ High sensitivity to acute endpoints: larval mortality very high (88 to 100% in the
PSA treatments, 44% in the control treatment); low incidence of metamorphosis
(0 to 6% in the PSA treatments, 6% in the control treatment, but 63% in the
water-only control treatment).
¦ Minor indications of reduced endpoint performance for larval malformations.
¦ High incidence of larval developmental delay, such that subsequent
environmental changes (i.e., decreased water temperature) may prohibit animals
developing in situ from reaching metamorphosis.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-44
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
4.4.1.2.2 Wood Frog Study: EPA
Table 4.4-5 presents the results of statistical tests of significance (comparisons to reference
stations) for wood frog toxicity test endpoints. Toxicological responses for impacted endpoints
that were related to exposure media concentrations are also presented graphically in detail in
Appendix E. Sediment tPCB concentrations are presented in two ways based on the two data
processing approaches discussed in the exposure assessment: (1) the concentrations measured in
the sediment samples collected with the amphibian samples, which represent the single PCB
concentration measurement taken closest to the effects endpoint (i.e., most synoptic
concentration); and (2) spatially weighted exposure point concentrations (EPCs) for each vernal
pool as discussed above.
Phase I: Egg mass viability for each vernal pool was evaluated relative to PCB concentrations
in sediment, water, and tissue. In general, there were no significant relationships found between
egg mass tissue concentration and any of the egg mass endpoints (such as hatching success or
percent fertilization). Egg mass tissue concentrations or any egg mass endpoints were not
significantly related to sediment tPCB concentration.
The Phase I egg mass viability studies indicate that these early life stage endpoints do not exhibit
consistent adverse effects that can be linked to PCB concentrations at the concentrations
measured in this portion of the study. Egg mass tPCB concentrations were not related to any
relevant Phase I endpoints, such as larval or metamorph mortality or the incidence of
larval/metamorph malformations. These results (relative to other life stages) suggest that
maternal transfer in wood frogs is not the dominant exposure pathway through which toxicity
was manifested in this study.
Both the magnitude and duration of exposure were shown to be important factors in the
manifestation of adverse effects in developing wood frogs. Larval development, metamorphosis,
and mortality assessment endpoints included:
¦ Evaluating larval mortality and metamorphosis at day 95, the longest uniform
exposure duration that could be applied to all treatments.
¦ Evaluating larval mortality and incidence of malformation at the end of each test
treatment.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC a a r 7/11/2003
-------
Table 4.4-5
Statistical Analysis Results: Wood Frog Reproduction and Development
Studies
V ariable/Endpoint
(shading indicates significant relationship)
Sample
Size (n)
Statistical Test
Results
Phase I Egg Mass
tPCB sediment (discrete) and tPCB
egg mass tissue
11
Spearman's
r = 0.36, p> 0.05
tPCB sediment (SW) and egg mass
tissue
8
Spearman's
r = -0.19, p > 0.05
tPCB water and tPCB egg mass tissue
11
Spearman's
r = 0.31, p > 0.05
Egg Mass Viability
tPCB sediment (discrete) and mean
egg weight
12
Spearman's
r =-0.014, p> 0.05
tPCB sediment (SW) and mean egg
weight
9
Spearman's
r = 0.067, p> 0.05
tPCB sediment (discrete) and %
fertilized eggs
12
Spearman's
r = -0.29, p> 0.05
tPCB sediment (SW) and % fertilized
eggs
9
Spearman's
r = -0.28, p> 0.05
tPCB water and % fertilized eggs
12
Spearman's
r = -0.10, p > 0.05
tPCB egg mass tissue and % fertilized
eggs
11
Spearman's
r = -0.27, p> 0.05
tPCB sediment (discrete) and hatching
success
12
Spearman's
r = 0.18, p > 0.05
tPCB sediment (SW) and hatching
success
9
Spearman's
r = -0.067, p> 0.05
tPCB water and hatching success
12
Spearman's
r = -0.26, p > 0.05
tPCB egg mass tissue and hatching
success
11
Spearman's
r = -0.08, p> 0.05
Phase II Larvae, Event 1
tPCB sediment (discrete)and tPCB
Event 1 larvae tissue
10
Spearman's
r = 0.29, p> 0.05
tPCB sediment (SW) and tPCB Event
1 larvae tissue
8
Spearman's
r = -0.26, p > 0.05
tPCB water and tPCB Event 1 larvae
10
Spearman's
r = 0.098, p> 0.05
Phase II Larvae, Event 3
iPCB sediment (discrete)and iPCB
Event 3 larvae tissue
11
Spearman's
r = 0.89, p< 0.002
iPCB sediment (SW) and tPCB Event
3 larvae tissue
8
Spearman's
r = 0.74, p = 0.05
tPCB water and tPCB Event 3 larvae
tissue
10
Spearman's
r = 0.69, p< 0.05
Phase II Larval Growth and
Development (Field-
Collected Animals), Event 4
tPCB sediment (SW) and Event 4
larval malformations
8
Spearman's
r = 0.83, p = 0.02
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC a as- 7/11/2003
-------
Table 4.4-5
Statistical Analysis Results: Wood Frog Reproduction and Development
Studies
(Continued)
Variable/Endpoint
Sample
Statistical Test
Results
(shading indicates significant relationship)
Size (n)
Larval Development,
Metamorphosis, and
Mortality
tPCB sediment (discrete) and Phase I
larval mortality (day 95)
11
Spearman's
r = -0.41. p > 0.05
tPCB sediment (SW) and Phase I
larval mortality (day 95)
8
Spearman's
r = -0.88. p < 0.02
tPCB water and Phase I larval
mortality (day 95)
11
Spearman's
r = -0.65. p< 0.05
tPCB egg mass tissue and Phase I
larval mortality (day 95)
11
Spearman's
r = 0.19. p> 0.05
tPCB sediment (discrete) and Phase I
larval mortality (EOT)
11
Spearman's
r = -0.41. p > 0.05
tPCB sediment (SW) and Phase I
larval mortality (EOT)
8
Spearman's
r =-0.57. p> 0.05
tPCB water and Phase I larval
mortality (EOT)
11
Spearman's
r = -0.43. p > 0.05
tPCB egg mass tissue and Phase I
larval mortality (EOT)
11
Spearman's
r = 0.21. p> 0.05
tPCB sediment (discrete) and Phase I
larval metamorphosis (day 95)
11
Spearman's
r = 0.43. p> 0.05
tPCB sediment (SW) and Phase I
larval metamorphosis (day 95)
8
Spearman's
r = 0.57. p> 0.05
tPCB water and Phase I larval
metamorphosis (day 95)
11
Spearman's
r = 0.42. p> 0.05
tPCB egg mass tissue and Phase I
larval metamorphosis (day 95)
11
Spearman's
r = -0.13. p > 0.05
tPCB sediment (discrete) and Phase I
larval metamorphosis (EOT)
11
Spearman's
r = 0.41. p> 0.05
tPCB sediment (SW) and Phase I
larval metamorphosis (EOT)
8
Spearman's
r = 0.57. p> 0.05
tPCB water and Phase I larval
metamorphosis (EOT)
11
Spearman's
r = 0.43. p> 0.05
tPCB egg mass tissue and Phase I
larval metamorphosis (EOT)
11
Spearman's
r =-0.21. p> 0.05
tPCB sediment (discrete) and no. of
days to 50% mortality Phase I larvae
11
Spearman's
r = 0.53. p> 0.05
tPCB sediment (SW) and no. of days
to 50% mortality Phase I larvae
8
Spearman's
r = 0.81. p < 0.05
tPCB water and no. of days to 50%
mortality Phase I larvae
11
Spearman's
r = 0.71. p< 0.02
tPCB egg mass tissue and no. of days
to 50% mortality Phase I larvae
11
Spearman's
r = -0.30. p > 0.05
tPCB sediment (discrete) and % Phase
I larval malformation Gosner stage
20-24
11
Spearman's
r = 0.80. p = 0.005
tPCB sediment (SW) and % Phase I
larval malformation Gosner stage 20-
24
8
Spearman's
r = 0.74. p = 0.05
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC A A1 7/11/2003
-------
Table 4.4-5
Statistical Analysis Results: Wood Frog Reproduction and Development
Studies
(Continued)
Variable/Endpoint
Sample
Statistical Test
Results
(shading indicates significant relationship)
Size (n)
Larval Development,
Metamorphosis, and
Mortality (Cont'd)
tPCB water and % Phase I larval
malformation Gosner stage 20-24
11
Spearman's
r = 0.77. p< 0.01
tPCB tissue and % Phase I larval
malformation Gosner stage 20-24
11
Spearman's
r = 0.53 p> 0.05
tPCB sediment (discrete) and Phase 1
mean metamorph weight
11
Spearman's
r = 0.56. p> 0.05
tPCB sediment (SW) and Phase 1
mean metamorph weight
8
Spearman's
r = 0.26. p > 0.05
tPCB water and Phase 1 mean
metamorph weight
11
Spearman's
r = 0.67. p< 0.05
tPCB egg mass tissue and Phase 1
mean metamorph weight
11
Spearman's
r = 0.66. p < 0.05
tPCB sediment (discrete) and Phase 1
metamorph malformations
11
Spearman's
r = 0.84. p< 0.005
tPCB sediment (SW) and Phase 1
metamorph malformations
8
Spearman's
r = 0.81. p < 0.05
tPCB water and Phase 1 metamorph
malformations
11
Spearman's
r = 0.73. p< 0.02
tPCB egg mass tissue and Phase 1
metamorph malformations
11
Spearman's
r = 0.49. p> 0.05
tPCB metamorph tissue and Phase 1
metamorph malformations
11
Spearman's
r = 0.54. p> 0.05
Phase I Metamorphs
tPCB sediment and tPCB metamorph
tissue
11
Spearman's
r = 0.55. p> 0.05
tPCB sediment (SW) and tPCB
metamorph tissue
8
Spearman's
r = 0.76. p< 0.05
tPCB water and tPCB metamorph
tissue
11
Spearman's
r = 0.67. p> 0.05
tPCB egg mass tissue and tPCB
metamorph tissue
11
Spearman's
r = 0.16. p> 0.05
Phase III Metamorphs
tPCB sediment (discrete) and tPCB
metamorphs (all)
10
Spearman's
r = 0.43. p> 0.05
tPCB sediment (discrete) and tPCB
metamorphs (exclude WML-1)
9
Spearman's
r = 0.70. p = 0.05
tPCB sediment (SW) and tPCB
metamorphs (exclude WML-1)
8
Spearman's
r = 0.76. p< 0.05
tPCB water and tPCB metamorphs
(all)
10
Spearman's
r = 0.74. p< 0.05
tPCB water and tPCB metamorphs
(exclude WML-1)
9
Spearman's
r = 0.81. p< 0.02
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC A AQ 7/11/2003
-------
Table 4.4-5
Statistical Analysis Results: Wood Frog Reproduction and Development
Studies
(Continued)
V ariable/Endpoint
(shading indicates si^nificanl relationship)
Sample
Size (n)
Statistical Test
Results
Phase III Metamorph
Development (Field-
Collected Animals)
tPCB sediment and Phase III
metamorph mean weight
10
Spearman's
r = 0.25, p> 0.05
tPCB metamorph tissue (excluding
reference) and Phase III metamorph
mean weight
9
Spearman's
r = 0.37, p> 0.05
tPCB sediment (discrete) and sex
ratios
10
Spearman's
r =-0.77, p< 0.02
iPCB sediment (S\Y) and sex ratios
8
Spearman's
r =-0.91, p = 0.005
iPCB metamorph tissue (excluding
WMI.-1) and sex ratios
9
Spearman's
r =-0.91, p< 0.005
iPCB sediment (discrete) and °o
malformation Phase III melamorphs
10
Spearman's
r = 0.93, p< 0.001
tPCB sediment (S\Y) and "<>
malformation Phase III melamorphs
8
Spearman's
r = 0.93, p< 0.005
iPCB metamorph tissue (excluding
WMI.-l) and °<> malformation Phase
III melamorphs
9
Spearman's
r = 0.85, p< 0.01
iPCB sedimenl (discrete) and °o
female gonadal malformation.* Phase
III melamorphs
10
Spearman's
r = 0.95, p< 0.002
tPCB sediment (S\Y) and "o female
gonadal malformation.* Phase III
melamorphs
8
Spearman's
r = 0.95, p< 0.005
iPCB metamorph tissue (excluding
W'MI.-l) and "n female gonadal
malformation.* Phase III melamorphs
9
Spearman's
r = 0.88, p< 0.005
tPCB melamorph tissue and % female
gonadal malformation.* Phase III
melamorphs
10
Spearman's
r = 0.72, p< 0.05
Melamorph sex ratio and "« female
gonadal malformation*
10
Spearman's
r =-0.93, p< 0.002
1 * Although the relationship between female gonadal malformation and total incidence of malformation is arguably
2 correlated, these comparisons are still of interest. There are many types of malformations that a juvenile could
3 show; however, the malformed females had a high incidence of gonadal aberrations that increased in relation to
4 increasing sediment and tissue tPCB concentration. Gonadal malformations can lead to sterility of the females.
5
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC A AQ 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
¦ Evaluating larval malformations at the first observation period, which occurred at
approximately Gosner developmental Stages 20 to 24. These are pre-feeding stages
characterized by full development of external gills (Gosner 1960). The transition
from embryo to a free-swimming, feeding tadpole (Gosner Stage 25/26) occurs in
these stages.
The larval development and metamorphosis component of the Phase I wood frog study produced
mixed indications of toxicity to individuals. There were virtually no adverse effects related to
PCB concentrations for mortality, time to metamorphosis, incidence of metamorphosis, or
growth endpoints. The pattern of responses appears to be related to the exposure duration and/or
the organism life stage. Hatchling stages indicated no concentration-response relationships,
whereas larval malformations at Gosner developmental Stages 20 to 24, and metamorphs
exhibited indications of toxicity (Figure 4.4-4). The malformation endpoint appeared sensitive
and was significantly correlated with sediment, water, and tissue tPCB concentrations (see
Figure 4.4-5 and Figure 4.4-6). The highest sediment tPCB concentrations caused an order of
magnitude increase in the incidence of malformed metamorphs.
Cross-Over and Aroclor 1260-Spiked Treatments: These treatments confirmed the
importance of vernal pool media as an exposure pathway:
¦ Reference site larvae raised in sediment and water from their native reference site
locations (i.e., Sites WML-1 and WML-2) had low tissue PCB concentrations of
0.340 and 0.242 mg/kg, respectively, while larvae from the same reference sites
raised in PSA vernal pool media (i.e., Sites 38-VP-l and 38-VP-2) had tissue PCB
concentrations of 6.61 and 7.82 mg/kg, respectively.
¦ PSA pool larvae raised in native media had tissue PCB concentrations of 4.66 and
6.61 mg/kg, respectively, while larvae from the same pools raised in reference site
media had tissue PCB concentrations of 0.109 and 0.053 mg/kg, respectively. This
indicates that uptake from sediment is more important than maternal transfer.
The cross-over study confirmed the overall findings of the Phase I main study, and indicated that
mortality and metamorphosis endpoints were not significantly affected by PCB exposures, while
moderate toxicity was observed for the malformation endpoint. The study indicated that tissue
burdens in later larval and metamorph stages were more directly linked to contaminated
exposure media (sediment, water) than to maternal transfer to the eggs. This finding has
implications for the interpretation of other wood frog toxicity endpoints (particularly Phase III
metamorphs).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC a r /-\ 7/11/2003
-------
I 15
I 10
5
0
WML-1 WML-2 WML-3 23b-VP-2 23b-VP-1 46-VP-1 46-VP-5 18-VP-2 8-VP-1 38-VP-1 39-VP-1 38-VP-2
Vernal Pool ID
1 % Malformed
- Mean sediment tFCB - - ~ - - Spatially w eighted mean tRCB
Figure 4.4-4 Comparison of Phase I Larval Wood Frog Malformations as
Gosner Developmental Stage 20-24 to Mean Sediment tPCB
and Spatially Weighted Mean tPCB (FEL 2002a)
WML-1 23b-VP-2 WML-3 WML-2 23b-VP-1 46-VP-1 46-VP-5 18-VP-2 8-VP-1 38-VP-1 39-VP-1 38-VP-2
Vernal Pool ID
~ Average % malformed metamorphs
¦ Sediment tPCB — Jr —Spatially weighted mean tPCB
Figure 4.4-5
Incidence of Malformation in Phase I Wood Frog Metamorphs,
with Mean Sediment tPCB and Spatially Weighted Mean tPCB
(FEL 2002a)
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-51
-------
7.0
£ 16
TO
O
Q.
18 -
2
8
0
S:
0.0
46-VP-1 46-VP-5 23b-VP-2 23b-VP-1 WML-3 WML-2 WML-1 18-VP-2 38-VP-1 8-VP-1 38-VP-2
Vernal Pool ID
l l Average % malformed metamorphs ¦ Tissue tPCB
Note: No tissue sample for 39-VP-l; sediment tPCB analyses for WML-1, WML-2, and WML-3 were non-
detect (ND) - numbers shown represent detection limits (0.069, 0.13, and 0.11, respectively).
Figure 4.4-6 Incidence of Malformation in Phase I Wood Frog Metamorphs,
Phase I Metamorph Tissue tPCB
(FEL 2002a)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC a r r\ 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Tissue PCB concentrations in the spiked and unspiked reference site treatment were 0.526 and
0.138 mg/kg PCB, respectively. The percentage of malformed metamorphs was higher in the
spiked treatment than in the control treatment (11% vs. 0.5%), indicating that exposure to the
spiked sediment had an adverse effect on metamorph malformation; however, larval mortality
and the incidence of larval metamorphosis appeared unaffected.
Phase II: The Phase II larvae showed a similar pattern of responses to that of the Phase I
animals. While only growth and malformation were evaluated in Phase II, the effects for the two
endpoints were similar in the two study phases. Mean larval length at each station increased
between sampling events 1 and 3 (as the larvae grew); and mean length was similar between
PSA and reference pools, showing no apparent relationship to tissue PCB concentration. The
larval growth endpoint has not proven to be a strong indicator of adverse biological effects in
other studies; Berven (1990) found no significant relationship between juvenile size and adult
survival. However, as the exposure duration increased for the Phase II animals, the incidence of
malformation increased. The event 4 animals had the highest incidence of malformation, with
response showing a significant relationship to sediment tPCB concentration.
The findings of the Phase II study show that PCB accumulation and the incidence of
malformations increase with exposure duration.
Phase III: This phase of the study represented an in situ exposure, with the same pools visited
at larval metamorphosis as those where egg masses for Phase I were collected. Endpoints
included incidence of malformation, growth, and sex ratio (number of males to females). As
with the first two study phases, there was no significant relationship between metamorph weight
and sediment or tissue PCB concentration. However, the sex ratios changed with increasing
sediment concentration. The metamorph sex ratio (males:females) ranged from 0 (Site 8-VP-l,
24.6 mg/kg) to 1.0 (Site 23b-VP-l, 0.2 mg/kg) for the PSA pools, and from 0.62 to 1.0 for the
reference sites (0.07 and 0.11 mg/kg). In general, as the PCB concentration in sediment or tissue
increased, the proportion of males to females decreased. There was a significant correlation
between skewed sex ratios and sediment and tissue tPCB concentrations (Figure 4.4-7 and Figure
4.4-8). Berven (1990) found a juvenile sex ratio of 1:1 in a Maryland study of population
fluctuations in larval and adult wood frogs, suggesting that this ratio is biologically "normal."
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-53
-------
1 The incidence of juvenile malformations was highest in this phase (both internal and external
2 malformations were assessed). Incidence of malformation was correlated with sediment and
3 metamorph tissue tPCB concentration. The percentage of malformed metamorphs ranged from
4 4.9% to 66.7% for the contaminated sites, and from 0 to 2.9% for the two reference sites (0.07
5 and 0.11 mg/kg). There was a significant relationship between metamorph malformation and
6 sediment and tissue tPCB concentrations (Figure 4.4-9 and Figure 4.4-10). Except for site
7 WML-1 (no malformed metamorphs, but with an anomalous tissue tPCB concentration),
8 treatments with the highest sediment or tissue tPCB concentrations also had the highest
9 percentages of malformed metamorphs.
10 The incidence of metamorph malformation is expected to be significant at the population level,
11 as a high degree of malformations could lead to reduced population recruitment at local and
12 regional scales (Ouellet 2000). The types of malformations observed in the metamorphs may
13 affect survivorship by interfering with swimming, hopping, foraging, and predator avoidance
14 (see following photos).
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Summary of Wood Frog Toxicity
¦ No observed toxicity in egg mass viability; egg fertilization, hatching success, and
egg counts were unaffected by vernal pool tPCBs, or egg mass tissue tPCB
concentrations.
¦ Contaminant effects were not observed in early-stage juveniles, although high
mortality in the reference animals makes it difficult to assess the acute sensitivity of
the wood frogs. Incidence of metamorphosis appeared unaffected.
¦ Manifestation of effects increased with time spent in the vernal pools. Late-stage
larvae/metamorphs (laboratory-cultured and field-collected) had elevated levels of
both internal and external malformations, with magnitude of response related to
sediment and tissue tPCB concentrations.
¦ Metamorphs collected after in situ exposure in natal pools showed alteration in sex
ratio in relation to sediment and tissue tPCB concentrations.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-54
-------
WML-1 WML-3 23b-VP-2 23b-VP-1 46-VP-1 46-VP-5 18-VP-2 8-VP-1 38-VP-1 38-VP-2
Vernal Pool ID
i iSpy ratio fM R —¦—Sediment tPCB —h —Spatially Weighted mean tPCB
Figure 4.4-7
Ratio of Males to Females in Phase III Wood Frog Metamorphs,
with Sediment tPCB and Spatially Weighted Mean tPCBs (FEL
2002a)
46-VP-1 WML-3 23b-VP-1 46-VP-5 23b-VP-2 38-VP-1 18-VP-2 WML-1 38-VP-2 8-VP-1
Vernal Pool ID
I I Spy ratin fM FI —¦—Tissue tPCB
Figure 4.4-8 Ratio of Males to Females in Phase III Wood Frog Metamorphs,
with Tissue tPCBs (FEL 2002a)
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-55
-------
70
70
60
50 -
u>
.c
Q.
O
E
ro
ai
40 -
¦D
at
E
|
ns
30
20 -
10 -
Zero
malformed:
tissue value
is anomalous
60
50
40
at
E,
m
o
0.
30 =
ai
¦a
at
20 w
10
WML-1 WML-3 23b-VP-2 23b-VP-1 46-VP-1 46-VP-5 18-VP-2 8-VP-1 38-VP-1 38-VP-2
Vernal Pool ID
I % Abnormal
-Sediment tPCB —h —Spatially Weighted mean tPCB
Figure 4.4-9 Percent Malformation in Phase III Wood Frog Metamorphs,
with Sediment tPCBs (FEL 2002a)
70
60
rr 50
Q-
o
E
4 "
H a
38-VP-1 38-VP-2
Vernal Pool ID
] % Abnormal —m— Tissue tPCB
Figure 4.4-10 Percent Malformation in Phase III Wood Frog Metamorphs,
with Tissue tPCBs (FEL 2002a)
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-56
-------
Example of Axial Flexure and Notochord Lesions
Example of Normal Tail
4.4.1.2.3 Wood Frog Study: GE
The study reported that there was not a statistically significant relationship between vernal
sediment tPCB concentrations and juvenile tissue concentrations. This finding concurs with that
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC a cry 7/11/2003
-------
of the EPA study, wherein natal pool sediment tPCB was shown to be unrelated to egg mass
tissue concentration.
MANOVA conducted on the grouped response variables of interest (juvenile survival and
growth) showed a significant difference between the two ponds in which the experiment was
conducted. However, this conclusion is difficult to interpret, as the sediment tPCB
concentrations were so similar between ponds 23b-VP-l and 23b-VP-2 (both < 1 mg/kg).
Differences are more likely due to environmental conditions, as the pools varied both according
to habitat type and size. Overall, there was a high degree of uncertainty associated with this
study, due mostly to an inadequate evaluation of relevant exposure pathways and study duration.
4.4.2 PCB Effect Thresholds
Data were compiled on sediment and tissue PCB concentrations associated with lethal or
sublethal effects in representative amphibians from the literature. The purpose was to estimate
threshold tissue concentrations where adverse effects might occur in Housatonic River
amphibians. The review focused on data for Aroclor 1254 and PCB-126 (considered one of the
more toxic congeners) in addition to tPCBs. Studies evaluating Aroclor 1260 were not found in
the literature, and studies with soil and sediment effects data were limited.
A total of five different species of frogs were used in the studies, including the leopard frog and
wood frog. There were a total of 18 no-effect measurements (ranging from 0.02 to 11.2 mg/kg
ww) and 11 effect measurements (ranging from 0.96 to 128 mg/kg ww). The majority of data
applied to effects on growth, development, behavior, physiological, and cellular effects. Seven
studies also evaluated mortality.
Figure 4.4-11 shows the distribution of no effect and effect tissue concentrations. No adverse
effects were observed at tissue concentrations of 0.1 mg/kg ww, whereas above 1 mg/kg ww, the
frequency of occurrence of adverse effects was more than 40%. There were six instances of
adverse effects occurring between 1 and 10 mg/kg. Based on this distribution, it is unlikely that
adverse effects will occur at tissue concentrations below 1 mg/kg, and it is likely that they will
occur above 10 mg/kg.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-58
-------
Overall, the upper and lower bounds of the effect concentration ranges from the EPA wood frog
study closely match those derived from the literature. Three endpoints indicated a consistent
relationship between tissue tPCB concentrations and adverse effects: metamorph malformations
in Phase I and Phase III and skewed sex ratios in the Phase III metamorphs. Specifically, 1
mg/kg was the approximate tissue concentration where ecologically significant adverse effects
began to occur, and responses became frequent and more severe at approximately 10 mg/kg.
No tissue effects threshold could be established for the leopard frogs, due to the difficulty in
establishing relevant biological linkages between tissue data and effects endpoints. However, the
leopard frogs appeared more acutely sensitive than the wood frogs.
In addition, protection of urodels (salamanders) also was considered in the derivation of a site-
specific tissue effects threshold. These animals sometimes spend almost their entire lives in the
vernal pools if they are a facultative neotonic species (i.e., fails to complete metamorphosis).
Given the increased sensitivity of the leopard frogs relative to the wood frogs, and the possibility
of neotony in the salamanders (and thus a much longer exposure period than would be typical for
the ranids), some conservatism was applied in the derivation of the 1 mg/kg tPCB tissue effects
threshold concentration.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-59
-------
1000
O)
E.
m
0
o.
01
3
100
10
0.1
0.01
LIGHT GRAY BAR = no significant effect (NOAEL)
DARK GRAY BAR = significant effect (LOAEL/LC50/EC20)
Total PCB
¦;?; ?.?.
¦ ¦ ?.?.
¦ ¦ ?.?.
¦ ¦ ?.?.
¦ ¦ ?.?.
¦ ¦ ?.?.
¦ ¦ ?.?.
¦ ¦ ?.?.
¦ ¦ ?.?.
Q, o
P, o
o U,
F: Fontenot et al, 2000
H: Huang et al, 1998
G: Gutleb et al, 2000
S: Savage et al, 2002
Figure 4.4-11 Summary of Available Literature Effects Data on PCB Tissue Residues in Anuran Amphibians
MK0110:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-60
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
4.5 RISK CHARACTERIZATION
The risk characterization evaluates the likelihood that adverse effects may occur as a result of
amphibian exposure to tPCBs and/or other COCs. Three broad categories of measurement
endpoints in the Housatonic River amphibian risk assessment were used to develop the risk
characterization:
¦ Endpoints based on field surveys (i.e., amphibian community structure) - For these
endpoints, care was exercised to discriminate, to the extent possible, between
responses related to COCs and those related to other factors such as substrate or
habitat type.
¦ Endpoints based on site-specific toxicity study results - These endpoints (e.g.,
toxicity tests involving both in situ and laboratory exposures) directly evaluated
biological responses to COCs.
¦ Endpoints that compared field-measured exposures to effects levels or benchmarks -
For these endpoints, the risk characterization integrated exposure and effects by
relating the two terms quantitatively (e.g., hazard quotient [HQ] method for tissue
chemistry data and derivation of concentration-response relationships for toxicity
data).
These three categories of endpoints were independent, allowing for a robust weight-of-evidence
(WOE) assessment of the potential for risk using the approach of Menzie et al. (1996).
All three lines of evidence suggested some degree of harm to amphibians in the Housatonic
River. In addition, for each line of evidence, there were indications that PCBs are primarily
responsible for the observed patterns of responses.
A WOE assessment was conducted to combine the results from each line of evidence. This
included a station-by-station assessment of each amphibian sampling location, as well as an
overall WOE assessment for the assessment endpoint. The section concludes with a discussion
of sources of uncertainty in the assessment of risks of COCs to amphibians and the conclusions
of the risk characterization.
Much of the risk characterization that follows is devoted to quantifying the relationship between
tPCB exposure concentrations and corresponding effects to amphibians. The formal
concentration-response analyses (for toxicity endpoints) strengthen the findings of the exposure-
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-61
-------
1 response relationships identified in Section 4.5. Various statistical methods were applied, with a
2 resulting level of concordance supporting the risk conclusions.
3 4.5.1 Concentration-Response Analysis - Toxicity Test Endpoints
4 A statistical assessment was conducted to quantify the relationship between toxicity test
5 endpoints and COC concentrations measured concurrent with the wood frog study. The
6 assessment focused on the relationship between PCBs and toxicity endpoints, since other lines of
7 evidence indicated a high probability that PCBs (as opposed to other COCs) were a causal agent
8 for toxicity to amphibians within the Housatonic River PSA.
9
10
11
12
13
14
15
16
17
18
19
20 4.5.1.1 Calculation of Individual Toxicity Test Endpoints
21 Comparisons based on magnitude of effects for various endpoints deemed biologically relevant
22 were considered. Effects observed at frequencies of 20% and 50% were selected as indicators of
23 moderate and major toxic effects, respectively.
24 Three sets of exposure data (two sediment, one tissue) were used to evaluate tPCBs
25 concentration-response relationships. Summary metrics (e.g., EC20, EC50) were calculated for
26 each endpoint based on sediment tPCB concentrations measured concurrent with the tests, and
27 also with spatially weighted sediment tPCB concentrations. In addition, tissue tPCB
28 concentrations were compared to effects.
29 Calculation of EC50 and EC20 values (with their corresponding 95% confidence limits) was
30 conducted using a linear probit method. If the probit model was not appropriate for the data
Endpoints Selected for Concentration-Response Analysis
Regardless of study phase, the late larval/metamorph endpoints were consistently
correlated with contaminant media concentrations. Therefore, the following endpoints
were selected for the more detailed statistical assessment:
¦ Phase I metamorph percent malformed larvae (compared to sediment and Phase I
metamorph tissue tPCB concentrations).
¦ Phase III percent malformed metamorphs (compared to Phase III metamorph tissue
and sediment tPCB concentrations).
¦ Phase III metamorph sex ratio (proportion of females) (compared to sediment tPCBs
and Phase III metamorph tissue tPCB concentrations).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-62
-------
1 (based on a goodness-of-fit test), the EC50 value was estimated by the nonparametric statistical
2 procedure, the Trimmed Spearman-Karber method.
3 4.5.1.2 Results of Concentration-Response Analysis
4 4.5.1.2.1 Sediment
5 Ecologically significant adverse effects in late stage juvenile wood frogs occurred in the
6 sediment tPCB concentration range of 9.54 to 59.3 mg/kg, although responses of lesser
7 magnitude, yet statistically significant, were observed at 0.52 mg/kg tPCBs and lower. MATC
8 of 3 mg/kg was established for sediment.
9
10
11
12
13
14
15
16
17
18 4.5.1
19 The threshold concentration range for amphibian tissues was 0.60 mg/kg to 6.54 mg/kg tPCB,
20 and was based on the sex ratio endpoint (both an EC20 and EC50) and the Phase III metamorph
21 malformation endpoint (an EC50 point estimate). As there was not a 20% effect size for
22 malformations in the Phase III metamorphs, a tissue EC20 could not be calculated. Tissue
23 concentrations below 1 mg/kg are not expected to cause biologically significant adverse
24 responses in the wood frogs. The tissue concentration-response modeling predicted significant
25 risk in the range of 1 to 10 mg/kg. At tissue concentrations >10 mg/kg, adverse ecological
26 effects are expected to occur with certainty.
Estimated Toxicity Threshold Values
¦ Most endpoints followed a fairly smooth (typically sigmoidal) concentration-response,
which could be fit using the probit model.
¦ Concordance was observed among endpoints for sediment concentrations causing
significant effects (i.e., 50% responses occurred at sediment tPCB concentrations of
9.54 to 59.3 mg/kg).
¦ Concordance was observed among endpoints for tissue residues causing significant
effects (50% responses occurred at tissue tPCB concentrations of 3.09 to 6.54
mg/kg).
.2.2 Tissue
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-63
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
4.5.2 Biological Community Endpoints
4.5.2.1 Amphibian Community Evaluation: EPA
Population responses of amphibians were measured in field studies of amphibian communities
conducted in 1999 and 2000 (Woodlot Alternatives, Inc. 2003). Detailed data were collected for
wood frogs (e.g., numbers of frogs entering and leaving pools, numbers of metamorphs captured
leaving the pools). In addition, species abundance, richness, and presence of malformations
were assessed for multiple species in selected vernal pools. Data describing the dominant plant
communities, dominant plants per community, soil, and general site hydrology were collected for
the entire PSA, as well as each amphibian sampling area (Appendix A). Descriptive information
on more than 60 amphibian breeding sites (e.g., size of vernal pools, average depth, percent
shading, and amphibian species observed breeding), and species observed at each site, is
included in the ecological characterization. Although most of these data were not collected
directly in conjunction with effects data, they were used in design of the subsequent amphibian
developmental studies, and were used to characterize the relative abundance of physically
suitable breeding sites for both leopard and wood frogs.
The findings included:
¦ Species richness was lower in the vernal pools with higher average sediment tPCB
concentrations; 6 in 8-VP-2 (55 mg/kg tPCBs), 8 in 38-VP-2 (32.3 mg/kg tPCBs), 8
in 8-VP-l (24.6 mg/kg tPCBs), and 11 in 46-VP-5 (0.72 mg/kg tPCBs). Overall,
density and biomass (on a per m2 basis) were lower in the more contaminated vernal
pools; 0.5 g/m2 wood frogs in 8-VP-2 versus 10.7 g/m2 wood frogs in 46-VP-5.
¦ Salamanders appeared to be sensitive to tPCBs, appearing in lower numbers in vernal
pools with high sediment tPCB concentrations. Salamander species observed
included the Jefferson salamander (Ambystoma jeffersonianum) and the four-toed
salamander (Hemidactylium scutatum), both of which are Species of Special Concern
in Massachusetts.
Gross malformation rates in adults (wood frogs and spotted salamanders) and metamorphs (wood
frogs) were low. However, malformation rates in larval wood frogs were high in all pools. The
malformation rates in the pools were 46% in 8-VP-l (24.6 mg/kg tPCBs), 35% in 38-VP-2 (32.3
mg/kg tPCBs), and 30% in 46-VP-5 (0.72 mg/kg tPCBs).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-64
-------
1 4.5.2.2 Leopard Frog Egg Mass Survey: GE
2 In the spring of 2003, ARCADIS G&M, Inc. (ARCADIS) conducted a survey of leopard frog
3 egg masses occurring in the vernal pool and backwater habitats of the PSA. The primary
4 objective of the survey was to determine whether adult leopard frogs are reproducing
5 successfully in the PSA; the metric chosen for evaluation of leopard frog reproductive health was
6 the presence/absence of egg masses within breeding habitats.
7 The investigators examined 44 ponds within the PSA for the presence of leopard frog eggs and
8 found egg masses in 17 ponds (a total of 216 egg masses). The study concluded that there was
9 no relationship between vernal pool sediment tPCB concentration and the presence/absence of
10 egg masses. In addition, the investigators concluded that there was no evidence of reproductive
11 impairment in leopard frogs within the PSA.
12 4.5.2.3 Amphibian Community Measures Observed During Developmental Study
13 Field Sampling: EPA
14 Additional evidence for population responses of amphibians was derived from anecdotal
15 information from field studies collected in support of the FEL developmental studies. No egg
16 masses were found at three of the leopard frog sampling areas and one of the wood frog vernal
17 pools. Sediment tPCB concentrations at these areas were among the highest of the
18 concentrations measured for the two studies: between 50 and 160 mg/kg. In addition, female
19 leopard frogs were not found at three contaminated sampling areas. Sediment PCB
20 concentrations at two of these areas were over 100 mg/kg tPCBs.
21 4.5.3 Comparison of Tissue Chemistry Data to Benchmarks
22 As an additional line of evidence, hazard quotients (HQs) were used to quantify the degree to
23 which amphibian tissue COC concentrations exceeded the literature-based and site-specific
24 thresholds deemed protective of assessment endpoints. In theory, adverse ecological responses
25 are possible if any HQ exceeds 1.0 (i.e., if exposure exceeds the lower threshold level). Separate
26 HQs were calculated for each tissue type and species. Tissue HQs were based on comparison of
27 observed tissue residues to an effects threshold of 1 mg/kg tPCBs, which represents a
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-65
-------
1 conservative interpretation of the LOAELs at which significant adverse effects were found in the
2 literature.
3 4.5.3.1 Leopard Frog HQs
4 HQs for leopard frogs could not be derived using site-specific leopard frog effects data due to the
5 modification to the study design based on the field conditions (i.e., leopard frog tissue burdens
6 and effects data were not synoptic or biologically related). However, HQs were calculated based
7 on comparison to literature-derived effects thresholds, which in turn are supported by the wood
8 frog study effects threshold. For the purposes of this exercise, it was assumed that there was a
9 similar sensitivity of the two representative species (although it is likely that the leopard frog is
10 more sensitive, as discussed in Section 4.4.1.2). The LOAEL of 1 mg/kg in tissue was
11 compared to tissue concentrations. Table 4.5-1 presents the range of HQs for leopard frog
12 tissues, using the literature-derived LOAEL. The tissue types evaluated included adults,
13 metamorphs, and egg mass/ovaries.
14 Based on the comparison to the LOAEL and the 10 mg/kg effects threshold, the abundance of
15 tissue HQs between 1 and 10 indicate a strong likelihood for adverse effects. Site-specific
16 reproductive and developmental effects clearly support this LOAEL for tissue tPCBs.
17 4.5.3.2 Wood Frog HQs: EPA Study
18 Wood frog HQs based on concentrations of tPCBs measured in the egg mass were relatively low.
19 Only two stations (18-VP-2 and 23b-VP-2) had a HQ greater than 1.0, and not by a large
20 amount. These HQs reflect a PCB exposure attributable to maternal transfer of PCBs. The low-
21 to-marginal HQs indicate that the chemical hazard for this life stage is fairly low. This finding is
22 consistent with the lack of significant toxicity observed in the egg mass toxicity endpoints, such
23 as hatching success, percent fertilization, and percent necrotic eggs.
24 Tissue HQs based on Phase II wood frog tadpoles (event 3; approximately 9 to 12 weeks old)
25 were variable, ranging between <0.1 and 10. The three stations with the highest sediment tPCB
26 concentrations (14.5 to 62 mg/kg) had HQs greater than 1.0 (8-VP-l, 38-VP-l, and 38-VP-2).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-66
-------
Table 4.5-1
Hazard Quotients for Leopard Frog PCB Tissue Residues, Based on Literature-
Derived Effects Thresholds
Sampling Area ID
Life Stage or Tissue Type
HQ
Muddy Pond Reference
Adult chemical analysis (whole body)
0.03
Adult experimental (female whole body minus ovaries/egg
masses)
0.012
Ovary/egg mass (from adult experimental female)3
0.024
Larvae-to-metamorphsb
NA
W-l
Adult chemical analysis (whole body)
0.15
Adult experimental (female whole body minus ovaries/egg
masses)
0.023
Ovary/egg mass (from adult experimental female)3
0.26
Larvae-to-metamorphs
NA
W-4
Adult chemical analysis (whole body)
0.34
Adult experimental (female whole body minus ovaries/egg
masses)
NA
Ovary/egg mass (from adult experimental female)3
NA
Larvae-to-metamorphsb
1.4
W-9a
Adult chemical analysis (whole body)
3.59
Adult experimental (female whole body minus ovaries/egg
masses)
1.24
Ovary/egg mass (from adult experimental female)3
5.05
Larvae-to-metamorphsb
NA
W-7a
Adult chemical analysis (whole body)
2.11
W-7a
Adult experimental (female whole body minus ovaries/egg
masses)
1.4
Ovary/egg mass (from adult experimental female)3
6.61
Larvae-to-metamorphsb
1.11
EW-3
Adult chemical analysis (whole body)
4.26
Adult experimental (female whole body minus ovaries/egg
masses)
1.23
Ovary/egg mass (from adult experimental female)3
1.52
Larvae-to-metamorphsb
0.96
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC a 7/11/2003
-------
Table 4.5-1
Hazard Quotients for Leopard Frog PCB Tissue Residues, Based on Literature-
Derived Effects Thresholds
(Continued)
Sampling Area ID
Life Stage or Tissue Type
HQ
E-5
Adult chemical analysis (whole body)
1.31
Adult experimental (female whole body minus ovaries/egg
masses)
NA
Ovary/egg mass (from adult experimental female)3
NA
Larvae-to-metamorphsb
NA
W-6
Adult chemical analysis (whole body)
1.78
Adult experimental (female whole body minus ovaries/egg
masses)
0.386
Ovary/egg mass (from adult experimental female)3
9.45
Larvae-to-metamorphsb
0.67
W-8
Adult chemical analysis (whole body)
5.39
Adult experimental (female whole body minus ovaries/egg
masses)
NA
Ovary/egg mass (from adult experimental female)3
NA
Larvae-to-metamorphs
NA
E-l
Adult chemical analysis (whole body)
3.10
Adult experimental (female whole body minus ovaries/egg
masses)
NA
Ovary/egg mass (from adult experimental female)3
NA
Larvae-to-metamorphs
NA
Sampling areas arranged in order of increasing sediment PCB concentration.
NA = No sample available because specimens were not found.
aEgg mass/ovary HQs based on a geometric mean of the two tissue concentrations per station. This was
done because of the large difference between the two concentrations for a given station.
bHQs for larvae-to-metamorph samples cannot all be compared to one another, as the specimens were
not all the same age when the samples were collected. Animals from sampling areas W-6, W-4, and
EW-3 are comparable; animals from sampling areas W-7a and W-l are comparable.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-68
-------
1 Phase III wood frog metamorphs (12 to 15 weeks old) had tissue HQs exceeding 1.0 in several
2 samples, with maximum HQs above 10. This phase represented sediment exposure over the
3 entire juvenile period, and exhibited the most pronounced toxicological responses. The reference
4 tissue concentration for station WML-1 was not included in the HQ calculations.
5 4.5.4 Integrated Station-by-Station Assessment
6 Potential impacts of contaminated sediment to local amphibians at each location were assessed
7 using a graphical approach that considered multiple lines of evidence (Table 4.5-2, leopard frogs;
8 Table 4.5-3, wood frogs). Multiple measurement endpoints were included, and the results of
9 each were integrated into a single conclusion regarding potential ecological impacts. For the
10 purposes of evaluating each measurement endpoint, results were categorized and simplified
11 based on ecologically-based decision criteria. The decision criteria used to make the evaluations
12 are summarized in Appendix E.
13 In summary, there was evidence for ecological effects for both acute and chronic developmental
14 endpoints in the leopard frog study and for several important developmental endpoints in the
15 wood frog study. For both amphibian developmental studies, there are multiple indications of
16 significant risk at multiple stations. There was a high degree of overall concordance among the
17 late-larval/pre-metamorph stage endpoints.
18 4.5.5 Weight-of-Evidence Procedure for Assessing Risk from PCBs in the
19 Housatonic River PSA
20 A formal WOE process was applied to determine whether PCBs pose a significant risk to the
21 Housatonic River benthos. The three-phase approach of Menzie et al. (1996) and the
22 Massachusetts WOE Workgroup was applied for this purpose, in which WOE was developed
23 using the following three characteristics: (1) the weight assigned to each measurement endpoint;
24 (2) the magnitude of response observed in the measurement endpoint; and (3) the concurrence
25 among outcomes of the multiple measurement endpoints.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-69
-------
Table 4.5-2
Integrated Assessment of Potential for Adverse Impacts to Amphibian Populations
(Leopard Frog Study)
Sampling
Area
Sediment tPCB (mg/kg)
Adult
Reproductive
Health3
Larval
Development1"
Tissue
Concentration
Overall Rating
Sampling Area
tPCB
Spatially
Weighted
tPCB
W-l
0.15
0.4
O
•
O
O
W-4
0.46
0.4
•
•
o
•
W-9a
4.3
7.5
•
NA
•
•
W-6
42
21
•
•
o
•
EW-3
30
23.8
•
K
o
•
E-l
160
26.6
•
NA
o
•
W-7a
18
27.6
o
•
•
•
W-8
120
43.5
•
NA
o
•
Sampling areas sorted by spatially weighted tPCB concentration.
© = Negligible-to-low toxicity: negligible indication of ecological risk. No exceedances of tissue benchmark (1 mg/kg tPCB).
O = Moderate toxicity; ecological effects possible, but not conclusive. At least 1 exceedance of tissue benchmark.
• = High toxicity; strong indication of potential ecological effects. At least 1 tissue concentration is >10x the tissue benchmark.
It = Very strong toxic response.
includes 6 endpoints: Adult body weight (male and female), sperm count, % abnormal sperm heads, egg count, % mature oocytes,
includes 5 endpoints: Larval growth, % metamorphosis, % malformation, growth, and days to reach Gosner Stage 26+1.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-70
-------
Table 4.5-3
Integrated Assessment of Potential for Adverse Impacts to Amphibian Populations
(Wood Frog Study)
Sampling Area
ID
Sediment tPCB (mg/kg)
Egg
Mass3
Early larvae
(up to Gosner
Stage 20-24)b
Mid- to Late-stage
larvae (after
Gosner Stage 24)c
Metamorphd
Tissue tPCB
Overall
Rating
VP
PCB
Spatially
Weighted
tPCB
23b-VP-l
0.19
0.2
O
O
O
O
O
O
23b-VP-2
0.11
0.3
o
O
o
O
O
O
46-VP-5
2.18
0.7
o
o
o
o
o
o
46-VP-l
0.5
0.8
o
o
o
o
o
o
18-VP-2
6.05
4.9
o
o
o
•
o
o
8-VP-l
14.5
24.6
o
o
o
•
•
•
38-VP-l
28
28.5
o
o
o
•
•
•
38-VP-2
62
32.3
o
o
o
•
o
o
39-VP-le
52
43
NA
NA
NA
NA
NA
NA
Sampling area sorted by spatially weighted tPCB concentration.
O = Negligible-to-low toxicity: negligible indication of ecological risk. No exceedances of lower or upper tissue benchmarks (1 and 10 mg/kg tPCB).
O = Moderate toxicity; ecological effects possible, but not conclusive. At least 1 exceedance of lower tissue benchmark.
# = High toxicity; strong indication of potential ecological effects. At least 1 exceedance of upper tissue benchmark.
H = Very strong toxic response.
includes 4 endpoints: Phase I egg mass weight (total), % fertilized, %viable, and % hatching success.
includes 4 endpoints: Phase I early larval malformation (Gosner Stage 20-24); Phase II, Event 1 and 2 larval abundance, % malformed and growth.
Includes 4 endpoints: Phase I larval mortality at day 95; Phase II, Event 3 and 4 larval abundance, % malformed and growth,
includes 3 endpoints: Phase I malformed metamorphs, Phase III malformed metamorphs, Phase III sex ratio.
eNo frogs were found in this pool for collection and subsequent study.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-71
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
The rationale for weighting of measurement endpoints is provided in Appendix E, along with a
discussion of attributes considered. A summary of the weighting for each attribute is provided in
Table 4.5-4. The chemistry endpoints yielded the lower overall values due to low-to-moderate
site specificity and some uncertainty with the linkage between the measurement endpoints and
the assessment endpoint(s). There is a stronger biological linkage to effects expected when
exposures are considered at the organism tissue level (i.e., incorporation of bioavailability). The
toxicity testing endpoints yielded the highest overall weighting, due to the site specificity and
high degree of biological relevance in the reproductive endpoints. The three field studies of
biological community endpoints had intermediate values. Although these endpoints were highly
site-specific and were direct measures of the assessment endpoint(s), confounding effects of
other environmental factors and lack of a quantitative assessment method reduced the utility of
these endpoints.
The magnitude of the response in the measurement endpoint is considered together with the
measurement endpoint weight in judging the overall WOE (Menzie et al. 1996). This requires
assessing the strength of evidence that ecological harm has occurred, as well as an indication of
the magnitude of response, if present. The weighting values, evidence of harm, and magnitude
of response were combined in a matrix format and are presented in Table 4.5-5.
A graphical method was used for displaying the degree of concurrence among measurement
endpoints (Table 4.5-6). The 12 symbols representing the chemistry (C), wood frog toxicity
(W), leopard frog toxicity (L), field biology (B), and (P) population model endpoints were
displayed in a matrix, with the weight of the measurement endpoint and the degree of response
as the axes.
The resulting plots show that 9 out of the 12 endpoints indicated some degree of risk. The
potential for the two GE studies to determine risk to amphibians was judged to be undetermined
due to limitations in the study designs. The only endpoint that did not indicate potential risk was
the earliest life stage wood frog toxicity endpoint, for which there is mechanistic explanation for
the lack of response. The plots also indicate that four endpoints exhibited a high degree of risk
combined with a moderate to high confidence rating.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-72
-------
Table 4.5-4
Weighting of Measurement Endpoints for Amphibian Weight-of-Evidence Evaluation
Measurement Endpoints
Endpoint
Group:
Chemistry
Endpoint Group: Wood Frog Toxicology (W)
Endpoint Group: Leopard Frog Toxicology (L)
Endpoint Group: Biology
Attributes
C
(tissue)
W-l
(hatchling)
W-2
(larvae)
W-3
(metamorph)
W-4
(GE juveniles)
L-l
(hatchling)
L-2
(larvae)
L-3
(metamorph)
L-4
(adult)
B-l
(community)
B-2
(GE egg mass
survey)
B-3
(anecdotal)
I. Relationship between Measurement and Assessment Endpoints
1. Degree of Association
Mod
Mod/High
Mod/High
Mod/High
Low
Mod/High
Mod/High
Mod/High
Mod/High
Mod/High
Mod
Mod/High
2. Stressor/Response
Mod
Mod/High
Mod/High
Mod/High
Low
Mod/High
Mod/High
Mod/High
Mod/High
Mod
Low
Low/Mod
3. Utility of Measure
Mod
Mod
Mod/High
Mod/High
Mod
Mod
Mod/High
Mod/High
Mod
Low/Mod
Low
Low/Mod
II Data Quality
4. Data Quality
Mod/High
Mod/High
Mod/High
Mod/High
Low
Mod/High
Mod/High
Mod/High
Mod/High
Mod/High
Mod
Mod/High
IK Study Design
5. Site Specificity
Low/Mod
Mod/High
Mod/High
Mod/High
Low/Mod
Mod
Mod
Mod
Mod
High
Mod/High
Mod/High
6. Sensitivity
Mod
Mod
Mod/High
Mod/High
Low
Mod
Mod
Mod
Mod
Mod
Low
Low/Mod
7. Spatial Representativeness
Mod
High
High
High
Low
Mod/High
Mod/High
Mod/High
Mod/High
Low
Mod/High
Mod
8. Temporal Representativeness
Mod
Mod/High
Mod/High
Mod/High
Mod
Mod/High
Mod/High
Mod/High
Mod/High
Mod/High
Mod
Mod/High
9. Quantitative Measure
Mod
High
High
High
Low
Mod/High
Mod/High
Mod/High
Mod
High
Low
Low
10. Standard Method
Mod/High
Mod/High
Mod/High
Mod/High
Low
Mod/High
Mod/High
Mod/High
Mod/High
Mod/High
Mod/High
Mod/High
Overall Endpoint Value
Mod
Mod/High
Mod/High
Mod/High
Low
Mod/High
Mod/High
Mod/High
Mod/High
Mod/High
Low/Mod
Mod
C. Chemical Measures
C. Concentration of PCB in frog tissues in relation to levels reported to be harmful to amphibians
W. Wood Frog Toxicological Measures
W-l. Sediment toxicity to hatchling/late embryo life stages
W-2. Sediment toxicity to larval life stages
W-3. Sediment toxicity to late larval/metamorph life stage
W-4. GE Context-Dependent Wood Frog Study (hatchlings, tadpoles, and metamorphs evaluated)
L. Leopard Frog Toxicological Measures
L-l. Sediment toxicity to hatchling/late embryo life stages
L-2. Sediment toxicity to larval life stages
L-3. Sediment toxicity to late larval/metamorph life stage
L-4. Sediment toxicity to adult leopard frogs (reproductive health)
B. Biology
B-l. Vernal pool community study
B-2. GE Leopard frog egg mass survey
B-3. Anecdotal observations during collections for reproductive stud
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-73
-------
Table 4.5-5
Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related to Maintenance of Amphibian
Populations in the Lower Housatonic River
Measurement Endpoints
Weighting
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Moderate, Low)
C. Chemical Measures
C. Concentration of PCB in frog tissues in relation to concentrations reported
to be harmful to amphibians.
Moderate
Yes
Low
W. Wood Frog Toxicological Measures
W-l. Sediment toxicity to hatchling/late embryo life stages.
Mod/High
No
-
W-2. Sediment toxicity to larval life stages.
Mod/High
Yes
Moderate
W-3. Sediment toxicity to late larval/metamorph life stage.
Mod/High
Yes
High
W-4. GE Study (juvenile wood frogs)
Low
Undetermined
-
L. Leopard Frog Toxicological Measures
L-l. Sediment toxicity to hatchling/late embryo life stages.
Mod/High
Yes
Low
L-2. Sediment toxicity to larval life stages.
Mod/High
Yes
High
L-3. Sediment toxicity to late larval/metamorph life stage.
Mod/High
Yes
High
L-4. Sediment toxicity to adult leopard frogs (reproductive health).
Mod/High
Yes
High
B. Biology
B-l. Vernal pool community study.
Mod/High
Yes
Low
B-2. GE leopard frog egg mass survey
Low
Undetermined
-
B-3. Anecdotal observations during collections for reproductive study.
Moderate
Yes
Low
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC a *-ja 7/11/2003
-------
Table 4.5-6
Risk Analysis for Amphibians Exposed to tPCBs and Other COCs in the Housatonic River PSA
Assessment Endpoint: Community condition, survival, reproduction, development, and maturation of amphibians
Risk/Magnitude
Weighting Factors (increasing confidence or weight)
Low
Low/Moderate
Moderate
Moderate/High
High
Yes/High
L-2, L-3, L-4,
W-3
Yes/Moderate
W-2
Yes/Low
Undetermined
No Risk
W-4, B-2
_______
C, B-3
L-l, B-l
W-l
n
C = Chemistry (tissue).
W = Wood frog study (1 = hatchling, 2 = larvae, 3 = metamorphs, 4 = GE Study).
L = Leopard frog study (1 = hatchling, 2 = larvae, 3 = metamorphs, 4 = adult).
B = Field study (1 = community, 2 = GE egg mass survey, 3 = anecdotal).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-75
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
The conclusion from interpretation of Table 4.5-6 is that there is a significant risk to amphibians
as indicated by the preponderance of the evidence and the relative weights of the measurement
endpoints. The "no harm" value of measurement endpoint W-l does not diminish the overall
conclusion, because the study demonstrated that the embryo/early larval life stages are fairly
insensitive to the effects of maternally transferred PCBs relative to later juvenile life stages
exposed to contaminated media.
4.5.6 Sources of Uncertainty
4.5.6.1 EPA Studies
The assessment of risks to amphibians in the PSA contains uncertainties, which can influence
overall conclusions of risk. Uncertainty associated with the assessment of risk of PCBs and
other COCs to amphibian receptors are described below.
¦ The greatest uncertainties in the exposure assessment in the amphibian ERA are (1)
the mobility of the animals and their exposure to concentrations that are not known,
and (2) the potential for small-scale variability (as was observed in PCB
concentrations in the river main channel sediment) in exposure concentrations in the
sampling areas. These two factors can confound the extrapolation of the quantitative
concentration-response relationships to the spatial scale of the PSA and its associated
backwater habitats. To overcome these uncertainties, spatially-weighted data were
used to determine an overall average exposure concentration for the leopard frog and
wood frog sampling areas, and for many of the vernal pools within the PSA.
¦ There is some uncertainty associated with the range of effects thresholds in the
literature and, therefore, with the selection of the tissue effects threshold. However,
studies within the literature do not exist from which definitive generalizations
regarding amphibian sensitivity to chlorinated organic contaminants can be made,
either for the class itself or for an individual genera or species. Therefore, the
literature review served as a supplement to the site-specific developmental studies,
but not to supplant the study results.
The tissue effect threshold was calculated to provide a general indication of risk . The
possible increased sensitivity of the leopard frogs and salamanders (see Section
E.3.6.4) relative to the wood frogs outweighs the uncertainty associated with the
derivation of the literature effects threshold. Overall, added conservatism in
development of a tissue effects threshold for amphibians within the PSA is warranted.
¦ There were some unusual data observations that were evaluated in the exposure
assessment:
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-76
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
The observation of elevated PCB concentrations in a wood frog Phase III
metamorph sample (WML-1; 4.3 mg/kg) collected at a reference location is
anomalous. All other reference sample concentrations were <0.4 mg/kg, and
sediment PCB concentrations from this area were non-detect. Statistical analyses
were conducted both with and without this anomalous value to evaluate the
potential effect of this value.
The lipid content of the wood frog Phase I egg mass samples were variable. A
number of lipid measurements were 0.1%, while other samples collected from the
same location yielded lipid contents as high as 1.5% (i.e., a 15-fold difference).
Several of the higher biota-sediment accumulation factor (BSAF) values in the
frog studies appear to be driven by the low lipid contents (0.1%) in the samples,
resulting in higher BSAFs than would be predicted on the basis of theoretical
equilibrium partitioning or observed at other PCB-contaminated sites. When
central tendency values (i.e., 1.0%; comparable to the lipid contents observed in
many of the tissue samples) are substituted, the BSAFs are more plausible. This
suggests that the low lipid content values may be underestimates, although the
variability in sediment PCB concentrations must also be considered.
- Detection limits for PAH compounds in the wood frog sediment data were high,
particularly for reference locations.
- Detection limits for metals (vanadium and nickel) in the wood frog tissue data
were also high.
¦ There is some uncertainty associated with the wood frog mortality endpoint. The
poor performance of the reference animals could be masking a natural sensitivity in
the wood frogs with respect to mortality because mortality in PSA animals was high
(greater than 50% at five stations). Thus, the sediment threshold effect concentrations
may be overly conservative with respect to the acute sensitivity of wood frogs.
¦ In addition, there is some uncertainty associated with the lack of replication in the
leopard frog data. As the study progressed, there were simply not enough test
organisms available for adequate replication (due to failed fertilization of the field-
collected adult females and limited availability of field-collected egg masses for
laboratory culture). Lack of replication prohibited the use of quantitative inferential
statistics to determine relationships between contaminant exposure concentrations and
response variables of interest. However, the magnitude of impacts observed in the
target animals (high mortality, low incidence of metamorphosis) helps to reduce the
uncertainty associated with limited replication.
¦ Based on the above-noted uncertainty associated with the three site-specific
amphibian developmental studies, extrapolation to the level of population is
uncertain. The wood frog Phase I metamorph malformation endpoint is suspected of
contributing to reduced young-of-year recruitment. However, relationships between
juvenile malformations and mortality need to be better defined because the mortality
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-77
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
data revealed little difference between the reference and PSA animals. Still, the high
incidence of both internal and external malformations in the Phase III metamorphs
indicates that metamorph recruitment to the population may be affected by exposure
to COCs.
¦ Skewed sex ratios observed in the wood frog study could be affecting the frog
population. Berven (1990) conducted a large-scale study to examine the factors
affecting population fluctuations in larval and adult stages of the wood frog in
Maryland. The study determined the sex ratio of recently metamorphosed juveniles
to be approximately 1:1 (male:female). The ratio of breeding adults, however,
averaged 3.1:1 (male: female. This skewed sex ratio was attributed to the majority of
female frogs breeding at age two, while male frogs started breeding at age one. The
females breeding in their second year were exposed to an additional year of mortality
than males, resulting in 2.3 times as many males as females from a given clutch
surviving to breed.
The Phase III wood frog metamorph results for the Housatonic sampling
exhibited a range of sex ratios. Sediment in the two most contaminated vernal
pools (38-VP-l and 38-VP-2) contained PCB concentrations of 28.5 and 32.3
PCBs based on spatial weighting of sediment data. These two pools also
exhibited the lowest metamorph sex ratios at 0.25 and 0.24 (male:female).
Another pool with sediment PCB concentrations of 24.6 mg/kg (8-VP-l) had
entirely females, although there were only three individual metamorphs captured.
The two vernal pools (23b-VP-l and 23b-VP-2) with the lowest PCB
concentrations (0.2 and 0.3 mg/kg tPCB, respectively) had sex ratios of 1.00 and
0.65 (male:female). The two reference sites (WML-1 and WML-3) had
metamorph sex ratios of 0.62 and 1.00 (male:female), respectively.
Sex ratios were also determined for breeding adult wood frogs in four pools
within the Housatonic River study area (Woodlot Alternatives, Inc. 2003). The
two vernal pools with the highest contaminations in this study (38-VP-2 and 8-
VP-2) contained soil tPCB concentrations of 32.3 and 54.9 mg/kg, respectively.
The corresponding sex ratio of breeding wood frogs from these pools was 1.5:1
and 0.9:1 (male:female). Vernal pool 8-VP-l had tPCB concentrations of 24.6
mg/kg and a corresponding sex ratio of breeding wood frogs of 0.8:1
(male:female). In the pool with the lowest contamination (46-VP-5, 0.59 mg/kg
tPCB), breeding wood frogs had a sex ratio of 1.3:1 (male:female).
The Housatonic River vernal pool sex ratio data for wood frog metamorphs and
breeding adults exhibit strong differences from Berven's data at a non-
contaminated site. The general trend for the wood frogs examined near the
Housatonic River PSA is a marked decrease in the male to female ratio in both
metamorphs and breeding adults. This feminization of the wood frogs in this
study may be adversely impacting the local population. Hayes (2000) reports that
alterations of sex ratios in amphibians may result in decreased recruitment and
population declines in what otherwise appear to be normal healthy adults.
Studies have also found that breeding between normal and sex-reversed adults
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-78
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
can lead to even more skewed sex ratios (Mikamo and Witschi 1964; Richards
and Nace 1978, in Hayes 2000).
- Reeder et al. (1998) examined intersexuality and the effects of environmental
contaminants on the cricket frog in Illinois. Their study determined that sex
ratios in cricket frog metamorphs also varied significantly between PCB/PCDF
contaminated and control sites. In nature the sex ratio in cricket frog metamorphs
favors females (Burkett 1984, in Reeder et al. 1998).
4.5.6.2 GE Studies
4.5.6.2.1 Wood Frog Study
The GE study did not examine the exposure pathway of contaminated media (i.e., sediment and
water), as the vernal pool exposure scenario did not represent the range of sediment PCB
concentrations within the PSA; the two pools used in the development portion of the study had
concentrations near detection limits.
Because wood frog larvae were placed in relatively clean sediment throughout the experimental
period, exposures to developing larvae were underestimated, relative to exposure during in situ
development in much of the PSA. Vernal pool sediment tPCB concentration was shown to be a
significant factor in the tissue uptake of PCBs and the subsequent manifestation of effects as the
frogs matured in the pools (FEL 2002a, 2002b). Therefore, the GE study exposed the developing
larvae to an atypically low range of sediment tPCB concentrations that were not characteristic of
the floodplain/backwater habitats above Woods Pond. Exposure to the full range of sediment
tPCB concentrations is necessary throughout the developmental period to understand
contaminant fate and effects in wood frogs because it is the later larval stages that were shown to
be the most vulnerable to contaminant-induced effects. The failure to expose developing larvae
to representative sediment tPCB concentrations is the primary source of uncertainty associated
with the study.
While the explicit consideration of density dependence is a valid consideration in the
examination of COC effects on amphibians, the limitation of inadequate contaminant exposures
outweighs the explicit consideration of density-dependence (detailed in Section 2.8 of the ERA).
In addition, some wood frogs enclosures had predators, which were unevenly distributed, and not
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-79
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
documented in the GE study report. Differing levels of predation were not factored into the
study and could affect density-dependence.
The GE study report did not provide a rationale for the selection of the three levels of initial
larval density evaluated, and a work plan was not provided. The relevance of the three densities
to wood frog ecology was not demonstrated or discussed, nor did the report indicate what change
in density is required to affect populations. The potential confounding effects of larval density in
the vernal pools remains a source of uncertainty.
The GE report states that it has evaluated the contaminant/effects question at the population
level. However, given the two significant sources of uncertainty discussed above, the question
of the effect of PSA contaminants on amphibian populations remains outstanding based on the
findings from this study.
4.5.6.2.2 Leopard Frog Egg Mass Survey
The principal area of uncertainty associated with this study is the conclusion that the egg masses
found are indicative of the reproductive health of the leopard frogs. While 216 egg masses were
found, the study did not show how the total number of egg masses could be extrapolated to the
level of overall species reproductive health. While the study provided qualitative verification
that some leopard frog egg masses were fertilized, quantitative data on fertilization, embryo
development, hatching success, and larval growth and development were not collected. The egg
mass measurement endpoint alone is not adequate for an assessment of adult frog reproductive
health or fitness. There is much more information that is necessary to adequately determine the
reproductive health of the leopard frog population.
Finally, the study states that sediment tPCB concentration is not related to the presence/absence
of egg masses or to egg mass density. Given the mobility and life history of the leopard frog
adults, this statement is likely correct. The sediment tPCB concentration of a breeding area is
not likely to be representative of a female's exposure to PCBs throughout the PSA. A female's
selection of an egg deposition site will be dependent upon habitat suitability and the presence of
males; sediment tPCB concentration in the breeding area is not likely the most critical metric for
determining trends in the presence or abundance of leopard frog egg masses within the PSA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-80
-------
1 There is no known mechanism by which leopard frogs can discern the toxic concentrations of
2 chlorinated organic contaminants and thereby avoid exposure to the contaminant.
3 4.5.7 Conclusions
4 Overall, the amphibian ERA indicates significant risk to frog species in the Housatonic River
5 PSA based on a WOE evaluation of multiple effects endpoints and their associated contaminant
6 media. Furthermore, the available data implicate tPCBs as the stressor responsible for such
7 impairment. The confidence in the conclusion is moderate to high, based on the concordance in
8 predictions of risk from multiple measurement endpoints. The most compelling evidence for
9 ecological degradation comes from the two frog toxicity studies, which not only exhibited
10 significant toxicological effects in both frog species and endpoints, but which also indicated a
11 correlation between level of effect and sediment tPCB concentration.
12 4.5.7.1 Population Modeling
13 A stochastic population model was developed to determine whether effects from tPCBs on
14 individual wood frogs influences wood frog populations within the Housatonic River PSA
15 (Appendix E, Attachment E.3). The population model projected wood frog population trends 10
16 years into the future and computed the risk of population decline (Ginzburg et al. 1982) using
17 vital rate information from the literature (Berven 1990) and initial abundances derived from
18 studies of vernal pools in the PSA (Woodlot Alternatives, Inc. 2002, 2003).
19 The impact of tPCBs on the wood frog population was assessed by comparing population
20 projections from a base population model (i.e., a wood frog population in the absence of tPCBs),
21 with projections from population models that included the effect of tPCBs on population vital
22 rates (see FEL 2002b). Two projection comparisons were performed based on simulations of (1)
23 a non-declining base population, and (2) a declining base population. All models were
24 constructed using RAMAS Metapop (Ak9akaya 2002).
25 Findings from the population modeling were:
26 ¦ tPCBs have an impact on wood frog population growth and abundance.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-81
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
¦ tPCBs hasten population decline, reduce population numbers, and increase the
likelihood of extinction.
¦ Data collected in the PSA provide field evidence supporting the population-level
effects of tPCBs seen in the simulations.
¦ The relationship between sediment tPCB concentrations and adult male and female
density indicate that increased tPCB concentration leads to decreased density—
particularly for adult females.
4.5.7.2 Risks Within the PSA
The WOE assessment of amphibian endpoints indicated a high probability of risk of ecologically
significant effects at PCB concentrations observed in the PSA vernal pools included in the
studies. Extrapolation to other areas of the PSA required use of concentration-response
relationships derived from the site-specific studies. The ERA findings suggest that amphibian
populations are impacted throughout much of the PSA, with leopard frogs impacted at a wide
range of sediment concentrations (likely due to the life history of contact with sediment PCB
concentrations, which were not measured in the study), and with stronger responses from wood
frogs expected in the more highly contaminated vernal pools. The indications of community
responses from the population studies (i.e., localized depressions of richness and abundance near
high tPCB vernal pools, and high incidence of malformations observed) substantiate these
findings.
4.5.7.3 Extrapolation of Risk Estimates Downstream of Woods Pond
Amphibians are primarily exposed to PCBs in floodplain soil, particularly vernal pools and other
wet, low-lying areas. The risk assessment focused on vernal pools, but such detailed data were
not available below Woods Pond, so the parameter of interest was tPCBs in floodplain soil and
sediment in general. Sediment was included to account for more aquatic amphibians such as
bullfrogs, and to account for the aquatic life phases of leopard frogs. IDW was used to
interpolate PCB concentrations to the limit of the 100-year floodplain (10-year floodplain
contours are not available downstream of Woods Pond) using the 0- to 6-inch-depth data from
the floodplain downstream of Woods Pond.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-82
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
4.5.7.3.1 Sediment
Ecologically significant adverse effects in late stage juvenile wood frogs occurred in the
sediment tPCB concentration range of 9.54 to 59.3 mg/kg, although responses of lesser, yet
statistically significant, magnitude were observed at 0.52 mg/kg tPCBs and lower.
A sediment MATC of 3 mg/kg tPCB was determined based on the results of the point estimate
calculations presented in section E.4.3.3. This concentration was just below the EC20 values for
the Phase III metamorph malformation endpoint (based on both measured and spatially weighted
tPCB concentrations), and well within the range of the 95% confidence limit (based on the probit
statistical analysis) for the point estimate. The MATC was rounded to 3 mg/kg to account for
uncertainty. The EC20 for the sex ratio endpoint was 0.52 and 0.61 mg/kg tPCB (based on
measured and spatially weighted sediment tPCB concentrations, respectively). However, as
noted in Section E.4.3.3, the 20% effect size is likely not of biological relevance, and therefore a
sediment MATC based on the sex ratio EC20 may be overly conservative. Selecting an
MATC of 3 mg/kg for this endpoint, just outside the 95% confidence limits for the EC20, is
believed to provide adequate protection for other amphibian species.
4.5.7.3.2 Tissue
The threshold concentration range for amphibian tissues was 0.60 mg/kg to 6.54 mg/kg tPCB,
and was based on the sex ratio endpoint (both an EC20 and EC50) and the Phase III metamorph
malformation endpoint (an EC50 point estimate). As there was not a 20% effect response for
malformations in the Phase III metamorphs, a tissue EC2o for malformations could not be
calculated.
Again, the EC20 of approximately 0.60 mg/kg for the sex ratio endpoint was considered to be
more conservative than necessary from the standpoint of biological relevance for wood frogs.
However, use of the tissue EC50S for the two endpoints was considered under-protective,
particularly given the likely increased sensitivity of the leopard frogs and salamanders relative to
the wood frogs. A tissue MATC of 1 mg/kg was therefore believed to provide a suitable balance
between the protection of other amphibian receptors and the lower tissue MATC
of approximately 0.65 mg/kg tPCB.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-83
-------
1 In summary, tissue concentrations below 1 mg/kg are not expected to cause biologically
2 significant adverse responses in the wood frogs. The tissue concentration-response modeling
3 predicted significant risk in the range of 1 to 10 mg/kg. At tissue concentrations >10 mg/kg,
4 adverse ecological effects are expected to occur.
5
6
7
8
9
10
11
12
13
14
15
16
17
18
MATCs for PCBs Used to Assess Risks Below Woods Pond
¦ The soil and sediment MATC of 3 mg/kg tPCB was used as a measure of the
potential for adverse effects to amphibians downstream of Woods Pond (Figures F.4-
10 and F.4-11). This concentration was developed in the risk assessment for the
PSA using multiple lines of evidence (e.g., amphibian community studies, in situ and
laboratory-exposure toxicity testing) and was selected as the concentration at which
some sensitive endpoints exhibited apparent responses, but the magnitude of
responses was not large. Above a concentration of 3 mg/kg tPCB, numerous
endpoints indicated ecologically significant responses.
¦ The tissue MATC of 1 mg/kg tPCB was used as a conservative measure of the
potential for adverse effects to amphibians downstream of Woods Pond. This
concentration was developed considering the frequency of adverse effects observed
in the literature studies, in the site-specific studies, and in an effort to compensate for
the uncertainty associated with the sensitivity of salamander species.
19
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-84
-------
1 4.6 REFERENCES
2 Ak9akaya, H.R. 2002. RAMAS Metapop: Viability Analysis for Stage-structured
3 Metapopulations (Version 4.0). Applied Biomathematics, Setauket, New York.
4 Berven, K.A. 1990. Factors affecting population fluctuations in larval and adult stages of the
5 wood frog (Rana sylvatica). Ecology 71:1599-1608.
6 Birge, W.J., J.A. Black, and A.G. Westerman. 1978. Effects of polychlorinated biphenyl
7 compounds and proposed PCB-replacement products on embryo-larval stages of fish and
8 amphibians. University of Kentucky. Lexington, KY. pp. 1-33.
9 Burkett, R.D. 1984. An ecological study of the cricket frog, Acris crepitans. Vert. Ecol. Syst.
10 10:89-103.
11 DeGraaf, R.M. and D.D. Rudis. 1983. Amphibians and Reptiles of New England: Habitats and
12 Natural History. University of Massachusetts Press, MA.
13 Eisler, R. and A.A Belisle. 1996. Planar PCB Hazards to Fish, Wildlife, and Invertebrates: A
14 Synoptic Review. National Biological Service Report 31, Washington, DC.
15 FEL (Fort Environmental Laboratories, Inc.). 2002a. Final Report - Frog Reproduction and
16 Development Study. 2000 Rana sylvatica Vernal Pool Study. Study protocol no.: WESR01-
17 RSTS03-1. Prepared by Fort Environmental Laboratories Inc., Stillwater, OK.
18 FEL (Fort Environmental Laboratories, Inc.). 2002b. Final Report - Frog Reproduction and
19 Development Study. 2000 Rana pipiens Reproduction and Development Study. Study protocol
20 no.: WESR01-RSTS03-1. Prepared by Fort Environmental Laboratories Inc., Stillwater, OK.
21 Fontenot, L.W., G.P. Noblet, J.M. Akins, M.D. Stephens, and G.P. Cobb. 2000. Bioaccumulation
22 of polychlorinated biphenyls in ranid frogs and northern water snakes from a hazardous waste
23 site and a contaminated watershed. Chemosphere 40:803-809.
24 Gilbert, M., R. Leclair, Jr., and R. Fortin. 1994. Reproduction of the northern leopard frog {Rana
25 pipiens) in floodplain habitat in the Richelieu River, P. Quebec, Canada. Journal of Herpetology
26 28(4):465-470.
27 Ginzburg, L.R., L.B. Slobodkin, K. Johnson, and A.G. Bindman. 1982. Quasiextinction
28 probabilities as a measure of impact on population growth. Risk Analysis 2:171-181.
29 Gosner, K. 1960. A simplified table for staging anuran embryos and larvae with notes on
30 identification. Herpetologica 16(2): 183-190.
31 Gutleb, A.C., J. Appelman, M.C. Bronkhorst, J.H.J, van den Berg, A. Spenkelink, A. Brouwer,
32 and A.J. Murk. 1999. Delayed effects of pre- and early-life time exposure to polychlorinated
33 biphenyls on tadpoles of two amphibian species (Xenopus laevis and Rana temporaria).
34 Environ. Toxicol. Pharm. 8:1-14.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC a qc 7/11/2003
-------
1 Gutleb, A.C., J. Appelman, M. Bronkhorst, J.H.J, van den Berg, and A.J. Murk. 2000. Effects of
2 oral exposure to polychlorinated biphenyls (PCBs) on the development and metamorphosis of
3 two amphibian species (Xenopus laevis and Rana temporaria). Sci. Tot. Environ. 262:147-157.
4 Hayes, T.B. 2000. Endocrine Disruption in Amphibians. In Chapter 10A, Ecotoxicology of
5 Amphibians and Reptiles. D.W. Sparling et al., editors. SETAC Press, Pensacola, FL.
6 Hine, R.L., B.L. Les, and B.F. Hellmich. 1981. Leopard frog populations and mortality in
7 Wisconsin, 1974-76. Wis. Dep. Nat. Resour., Tech. Bull. No. 122.
8 Huang, Y.W., M.J. Melancon, R.E. Jung, and W.H. Karasov. 1998. Induction of cytrochrome
9 P450-associated monooxygenases in northern leopard frogs, Rana pipiens, by 3,3'4,4'5-
10 pentachlorobiphenyl. Environ. Tox. Chem. 17(8): 1564-1569.
11 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S.
12 Petron, B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-
13 of-Evidence Workgroup: A weight of evidence approach for evaluating ecological risks. Human
14 and Ecological Risk Assessment 2:277-3 04.
15 Mikamo, K. and Witschi, E. 1964. Masculinization and breeding of the WW Xenopus.
16 Experentia 20:622-623.
17 Ouellet, M. 2000. Chapter 11, Amphibian Deformities: Current State of Knowledge. In
18 Ecotoxicology of Amphibians and Reptiles. Society of Environmental Toxicology and Chemistry
19 (SETAC), Pensacola, FL. 904 p.
20 Reeder, A.L., G.L. Foley, D.K. Nichols, L.G. Hansen, B. Wikoff, S. Faeh, J. Eisold, M.B.
21 Wheeler, R. Warner, J.E. Murphy, and V.R. Beasley. 1998. Forms and prevalence of
22 intersexuality and effects of environmental contaminants on sexuality in cricket frogs (Acris
23 crepitans). Environ. Health Persp. 106:261-266.
24 Richards, C.M. and Nace, G.W. 1978. Gynogenetic and hormonal sex reversal used in tests of
25 the XX-XY hypothesis of sex determination in Rana pipiens. Growth 42:319-331.
26 Rosenshield, M.L., M.B. Jofre, and W.H. Karasoy. 1999. Effects of polychlorinated biphenyl
27 126 on green frog (Rana clamitans) and leopard frog (Rana pipiens) hatching success,
28 development, and metamorphosis. Environmental Toxicology and Chemistry 18(11):2478-2486.
29 Savage, W.K., F.W. Quimby, and A.P. DeCaprio. 2002. Lethal and sublethal effects of
30 polychlorinated biphenyls on Rana sylvatica tadpoles. Environmental Toxicology and Chemistry
31 21(1): 168-174.
32 Sparling, D.W., G. Linder, and C.A. Bishop. Editors. 2000. Ecotoxicology of Amphibians and
33 Reptiles. Society of Environmental Toxicology and Chemistry (SETAC), Pensacola, FL. 904 p.
34 Taylor A.C. and J.J. Kollros. 1946. Stages in the normal development of Rana pipiens larvae.
3 5 The Anatomical Record 94:7-23.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-86
-------
1 Woodlot Alternatives, Inc. 2002. Ecological Characterization of the Housatonic River. Prepared
2 for U.S. Environmental Protection Agency.
3 Woodlot Alternatives, Inc. 2003. Amphibian Reproductive Success within Vernal Pools
4 Associated with the Housatonic River, Pittsfield to Lenoxdale, Massachusetts. Report prepared
5 for Weston Solutions, Inc. DCN: 99-1275.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_4.DOC
4-87
-------
1 5. ASSESSMENT ENDPOINT - SURVIVAL, GROWTH, AND
2 REPRODUCTION OF FISH
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Highlights
Conceptual Model
¦ Conceptual model for fish indicates that the most important exposure pathways
are diet and contaminated sediment.
Exposure
¦ COCs evaluated were total PCBs (tPCBs), 2,3,7,8-TCDD TEQ, and polycyclic
aromatic hydrocarbons (PAHs).
¦ Whole body fish concentrations in five representative fish species were used to
evaluate exposures to both tPCBs and TEQ.
¦ Sediment concentrations were used to evaluate risks to fish from PAHs.
Effects
¦ Site-specific toxicity tests (Phase I and Phase II) indicate adverse effects at
contaminated locations relative to reference areas, and although variable,
general PCB dose-dependency. Effects observed are indicative of a dioxin-like
etiology.
¦ Literature review indicates that PCB and TEQ threshold tissue concentrations
identified in the literature are in the same range as those from site-specific
toxicity tests.
Risk Characterization
¦ A weight-of-evidence approach was used to characterize risks. A high probability
of adverse impacts to fish from tPCBs and/or TEQ is predicted throughout the
PSA. Impacts are likely for sensitive sublethal endpoints (reproduction and
development), but mortality of adults is unlikely.
¦ Risks attributable to PAHs are negligible to low.
¦ Impacts downstream of Woods Pond are limited to marginal risks for coldwater
fish (trout).
¦ Magnitude of impact is not predicted to be catastrophic in any reach; adverse
effects, although high in probability, are generally expected to be subtle.
¦ Field surveys (fish community and reproduction studies) support lack of
catastrophic effects, but cannot be used to assess lesser impacts.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-1
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
5.1 INTRODUCTION
The purpose of this section is to characterize and quantify the current and potential risks posed to
fish exposed to contaminants of concern (COCs) in the Housatonic River, focusing on total
PCBs (tPCBs) and other COCs originating from the General Electric (GE) facility in Pittsfield,
MA.
A Pre-ERA was conducted to narrow the scope of the ecological risk assessment by identifying
contaminants, other than tPCBs, that posed potential risks to aquatic biota in the PSA
(Appendix B). The ERA further screened the above contaminants of potential concern (COPCs)
for specific relevance to fish inhabiting the main channel of the Housatonic River, and identified
COCs to be retained in the detailed assessment. COC groups that were retained were tPCBs,
dioxin-like TEQ, and polycyclic aromatic hydrocarbons (PAHs).
A stepwise approach was used to assess the risks of COCs to fish in the Housatonic River. The
four main steps in this process included the following:
1. Development of a conceptual model (Figure 5.1-1).
2. Assessment of exposure of fish to COCs (Figure 5.1-2).
3. Assessment of the effects of COCs to fish (Figure 5.1-3).
4. Characterization of risks to fish (Figure 5.1-4).
This section is organized as follows:
¦ Section 5.1 (Introduction and Conceptual Model)—Describes the conceptual
model for fish, including selection of representative species and establishment of
measurement and assessment endpoints.
¦ Section 5.2 (Exposure Assessment)—Describes the quantification of exposures,
both within the Primary Study Area (PSA) and downstream of Woods Pond.
¦ Section 5.3 (Effects Assessment)—Describes the potential effects to fish exposed to
site COCs, as indicated by the toxicological and field investigations conducted in the
PSA. Assesses the concentration-response relationships from site-specific studies and
identifies corresponding effects thresholds. This section also summarizes the ranges
of relevant tissue and sediment effects thresholds (toxicity thresholds) derived from
the literature.
¦ Section 5.4 (Risk Characterization)—Integrates the exposure and effects
assessments, summarizes field surveys, and makes conclusions regarding risk for fish
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-2
-------
1
2
3
4
in the Housatonic River using three main lines of evidence. A discussion of the
sources of uncertainty regarding risk estimates follows. Section 5.4 also presents an
extrapolation of risks beyond the PSA to areas downstream of Woods Pond, for both
coldwater and warmwater species.
5
6
This section provides a summary of the ERA for fish, which is presented in detail in
Appendix F.
7
8 5.1.1 Conceptual Model
9 Total PCBs, dioxins, and furans are persistent and hydrophobic and lipophilic. Food web
10 bioaccumulation and biomagnification represent the most important pathways for PCBs and
11 TEQ (Oliver and Niimi 1985). Direct uptake pathways, through respiration of dissolved PCBs or
12 through incidental ingestion of sediment, are less important pathways for most fish species.
13 PAHs are metabolically transformed in most teleosts, and because of their poor water solubility
14 are more strongly associated with sediment. In summary, the COCs identified for fish exhibit
15 both direct (i.e., contact with contaminated media) and indirect (i.e., food web bioaccumulation)
16 pathways, with emphasis on the latter pathway for PCBs.
17 The conceptual model presented in Figure 5.1-1 illustrates the exposure pathways for fish in the
18 PSA. For strongly hydrophobic COCs (PCB, dioxins/furans), the dominant exposure media
19 were COC tissue concentrations, with uptake into tissues occurring mainly via ingestion of
20 contaminated prey. Tissue concentrations reflect the net COC uptake from food, sediment,
21 overlying water, and porewater, and therefore integrate all primary exposure pathways of
22 interest. For PAHs, sediment were considered the most relevant exposure media. Because
23 PAHs are rapidly degraded to daughter products (some toxic), tissue PAH concentrations cannot
24 be used as exposure metrics for linkage to effects (Johnson et al. 2002, Malins et al. 1985).
25 Five fish species were selected as the representative species for the ERA. The selected fish
26 species include representatives of the different trophic levels and exposure routes for fish in the
27 PSA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-3
7/11/2003
-------
(A
0)
(A
re
0)
0)
a:
(A
0)
£
3
o
(/)
ra
T3
0)
(A
Q.
0)
O
0)
o£
(A
o
it
LU
1
2 Figure 5.1-1 Conceptual Model Diagram: Exposure Pathways for Fish Exposed
3 to COCs in the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-4
-------
Exposure
2
3 Figure 5.1-2 Overview of Approach Used To Assess Exposure of Fish to COCs
4 in the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r r 7/11/2003
-------
Effects
Figure 5.1-3 Overview of Approach Used To Assess the Effects of COCs to
Fish in the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r s 7/1
-------
Risk Characterization
2 Figure 5.1-4 Overview of Approach Used To Characterize the Risks of COCs to
3 Fish in the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r q 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
Representative Species
¦ Largemouth bass (Micropterus salmoides) - predator fish.
¦ Yellow perch (Perca flavescens) - predator fish.
¦ Brown bullhead (Ameiurus nebulosis) - bottom feeder.
¦ White sucker (Catostomus commersoni) - bottom feeder.
¦ Pumpkinseed (Lepomis gibbosus) - forage fish.
Criteria considered in selecting representative fish species for the ERA included trophic level
and feeding preferences, abundance and biomass in the study area, availability of site-specific
data, and availability and appropriateness of toxicological data. Because trout have greater
importance downstream of the PSA (due to the presence of suitable habitat for a coldwater
fishery), a separate analysis using information on trout toxicity from general literature and site-
specific studies was also conducted as a part of the downstream risk prediction section of the
ERA.
The assessment endpoint that is the subject of this section is the survival, growth, and
reproduction of fish. Measurement endpoints were selected to assess risks of PCBs and other
COCs to the representative fish species:
Measurement Endpoints for Fish
¦ Determine the possible extent of adverse effects by comparing the
concentrations of COCs in sediment to the concentrations reported in the
literature to cause adverse effects on the survival, growth, or reproduction offish.
¦ Compare the concentrations of COCs in fish tissues to the concentrations in fish
tissues that may result in adverse effects, based on site-specific fish toxicity
studies.
¦ Compare the concentrations of COCs in fish tissues to concentrations
documented in the literature to result in adverse effects.
¦ Evaluate field survey information (fish biomass study, ecological characterization
study, and largemouth bass habitat and reproduction study) to qualitatively
assess potential effects.
The approach used to characterize risks to fish was based upon evaluation of numerous data
sources. These included site-specific toxicity investigations, chemical measurements of fish
tissue and sediment, biological/community assessments, and literature reviews.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC ^ g 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Summary of Studies Used To Characterize Risks to Fish
¦ Phase I site-specific studies - Investigated contaminant accumulation and effects
in largemouth bass from the PSA, Rising Pond, and a reference area and in their
offspring. Adults were evaluated for contaminant-related biochemical, cellular,
and organism level effects. Offspring were monitored for survival, development,
gross abnormalities, and biochemical alterations.
¦ Phase II site-specific studies - Simulated the maternal transfer of contaminants
to developing oocytes. Investigated the effects of injected organic contaminants
(extracted from largemouth bass tissue collected in the Phase I studies) on the
development of largemouth bass, medaka (Orizias latipes), and rainbow trout
eggs.
¦ Sampling and analysis offish tissue.
¦ Sampling and analysis of sediment.
¦ Field studies - EPA Fish Community and Ecological Characterization Study
(Appendix A) and GE Largemouth Bass Community, Reproduction, and Habitat
Study.
¦ Literature review - Evaluated the range of PCB, TEQ, and PAH concentrations
observed to elicit adverse effects to fish.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-9
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
5.2 EXPOSURE ASSESSMENT
The exposure assessment estimates the exposure of fish to tPCBs and other COCs in the
Housatonic River PSA (Figure 5.1-2). The COPCs that were retained in the Pre-ERA (Appendix
B) were screened specifically for relevance to fish, resulting in the COCs that were retained in
the exposure assessment.
The vast majority of relevant exposure data were those collected within the PSA. More limited
tissue data were also available for areas downstream of the PSA. Extrapolations of risk to most
downstream areas relied on the development of exposure-response relationships developed from
the site-specific studies and the literature. In some cases, upstream data were used to standardize
downstream data for use in the ERA (e.g., lipid-based conversions of fillet PCB concentrations
to whole-body concentrations).
5.2.1 Refined Screening of COPCs for Fish
The Pre-ERA (Appendix B) developed separate lists of COPCs for fish tissue, water, and
sediment. Water and sediment screening included comparisons to thresholds considered
protective of aquatic life, including invertebrates and fish. Because fish tissue COPC
concentrations in the Pre-ERA were not screened against fish tissue effects thresholds, a step was
conducted to ensure that no bioaccumulative COPCs were eliminated prematurely that could be
of concern to fish.
PCBs - PCBs were identified as COPCs for sediment and water in all PSA reaches. The ERA
for fish considered tPCB risks as well as dioxin-like toxicity (2,3,7,8-TCDD TEQ) from coplanar
PCBs, dioxins, and furans.
Dioxins and furans - The ERA considered dioxin and furan risks in the context of their
contribution to TEQ. The semivolatile contaminant dibenzofuran was eliminated on the basis of
the very few isolated exceedances of screening criteria (Reach 5A only).
PAHs - Of the principal PAH compounds detected in sediment, only pyrene and fluoranthene
had significantly elevated concentrations relative to those observed in reference areas. Although
the conservative screening using sediment quality values suggested a potential for aquatic risk,
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r -> rx 7/11/2003
-------
1 concentrations of most PAHs appear to be similar across contaminated and reference sites in the
2 PSA. Using a conservative approach, total PAHs plus eight individual PAHs were retained in
3 the fish risk assessment.
4 Pesticides - 4,4'-DDE and 4,4'-DDT were identified as COPCs for sediment in the Pre-ERA,
5 but were identified only in isolated vernal pools and side channels. Furthermore, some pesticide
6 detections may be attributable to laboratory interference artifacts. Pesticides were therefore not
7 considered further in the fish assessment.
8 Metals - Eleven metals were identified as COPCs in sediment based on thresholds developed for
9 the protection of benthic invertebrates. However, all metals were eliminated from the fish ERA,
10 based on an ecological relevance screen (Appendix F; Attachment F. 1).
11 Summary - The list of COCs retained in the risk assessment for fish is provided below.
12
13
14
15
16
17
18 5.2.2 Tissue Chemistry Assessment (Exposure to PCBs and TEQ)
19 The most robust data set for fish tissue concentrations of tPCBs and TEQ was collected by EPA
20 from September 1998 through October 2000. Other fish tissue tPCB data collected within the
21 PSA by GE and others from 1977 to 2002 are also available. These additional data sets were
22 evaluated, and either the inclusion or exclusion of these data would result in very similar risk
23 conclusions. For example, BBL and QEA (2003) state that "PCB concentrations in largemouth
24 bass collected in 2002 from Reach 5B and Reach 6 were similar to each other and, on a whole-
25 body basis, to the 1998 EPA data." Therefore, the ERA conclusions are not sensitive to the
26 inclusion or exclusion of data from this recent GE sampling effort.
27
COCs for Fish
¦ Chlorinated organic compounds - PCBs as tPCBs and TEQ, dioxins/furans
expressed as TEQ equivalents.
¦ PAHs - Total PAH, benzo(g,h,i)perylene, indeno(1,2,3-cd)pyrene, phenanthrene,
anthracene, benzo(a)anthracene, pyrene, fluorene, fluoranthene.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-11
-------
1
2
3
4
5
6
7
8
9
10 In general, game fish of edible and/or legal size were analyzed separately as skin-off fillet and
11 offal for use in the human health risk assessment and the ERA. The whole-body tissue
12 concentrations for these fish (WB-R) were calculated for this assessment as the weighted average
13 concentration of the individual tissues, using calculation methods described in Appendix F.
14 5.2.2.1 Total PCBs
15 Table 5.2-1 presents summary statistics for fish tissue tPCB concentrations in the PSA by sample
16 type and species, for all representative species. The table combines samples across all PSA
17 reaches. The reach-by-reach distributions are evaluated in Appendix F; in general there are
18 relatively consistent fish tissue concentrations across the entire PSA. The composite fish tissue
19 samples had lower PCB concentrations than other tissue types (Figure 5.2-1); the small fish in
20 the composite samples were typically representative of younger fish or of species that are small
21 even as adults (i.e., dace).
22 Table 5.2-2 presents summary statistics for lipid-normalized tPCB concentrations by sample
23 type, species, and reach. The mean and median concentrations in Table 5.2-2 indicate that the
24 highest PCB concentrations (i.e., median >2,000 mg/kg lipid) are observed in adult (WB-R)
25 predator fish, due to biomagnification in the food web. Age can be a factor even after lipid
26 normalization because older fish have accumulated PCBs over a longer period of time. This is
27 evident in Table 5.2-2, which shows that the lowest mean and median fish concentrations were
28 observed for composites (CM) of smaller younger fish. Figure 5.2-2 shows the relationship
29 between age and PCB body burden in largemouth bass.
Fish Chemistry Types Considered in the ERA
¦ CM - Composite samples - represent the combination of multiple fish, typically
young-of-year or other small fish.
¦ WB - Whole body samples - represent the analysis of single larger fish, often for
species that were not considered in the human health risk assessment (e.g.,
white sucker).
¦ WB-R - Whole body reconstituted samples - represent individual fish
concentrations that were calculated/estimated using separate fillet and offal
measurements.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-12
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Table 5.2-1
tPCB Concentrations in Representative Species Fish Tissue (mg/kg)
in the PSA; Data from EPA Tissue Collections (1998-2000)
Sample
Type
Species
Sample
count
Min
25th
Percentile
Median
Mean
75th
Percentile
Max
WB-R
BB
43
7.19
25.3
32.3
37.6
45.6
103
LB
38
10.9
42.3
67.8
97.1
125
424
PS
51
7.82
23.2
34.6
36.7
44.7
82.1
YP
75
6.11
61.3
76.1
87.3
104
329
WB
BB
2
20.9
21.3
21.7
21.7
22.0
22.4
GF*
42
10.8
95.5
143
163.6
215
447
LB
26
3.03
22.7
36.5
57.1
78.4
220
WS
57
7.96
36.2
56.6
70.6
86.5
216
CM
LB
12
9.03
19.9
26.1
27.9
36.3
51.2
PS
9
00
00
26.4
27.5
26.2
27.9
35.1
YP
15
16.5
27.4
31.0
31.4
35.7
46.9
*Goldfish (GF) were not selected as a representative species, but were included because of large sample size and
high tPCB concentrations.
BB
Brown Bullhead
LB
Largemouth Bass
PS
Pumpkinseed
WS
White Sucker
YP
Yellow Perch
CM
Composite
WB
Whole Body
WB-R
Reconstituted Whole Body
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC ^ ^3
-------
Largemouth Bass
Pumpkinseed
CO
O
0_
CM WB WB-R
Sample Type
m
O
0_
CM
WB-R
Sample Type
Yellow Perch
o
o
o - .
o
Q.
O -
CM WB-R
Sample Type
1
2 The shaded box represents the interquartile range, the white bar represents the median, and the whiskers extend to
3 the full range of the data.
4 CM = composite samples
5 WB = whole body samples
6 WB-R = whole body reconstituted samples
7
8 Figure 5.2-1 Box-and-Whisker Plots of Lipid-Normalized tPCB Concentrations
9 Plotted by Sample Type for Species with Multiple Sample Types
10
MK0110:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-14
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Table 5.2-2
Total Lipid-Normalized PCB Concentrations (mg/kg lipid) for
Representative Species in the PSA; Data from EPA Tissue Collections (1998-2000)
Sample
Type
Species
Sample
count
Min
25th Percentile
Median
Mean
75th Percentile
Max
WB-R
BB
43
224
1010
1520
2160
1870
14700
LB
38
591
1600
2490
2960
3720
8280
PS
51
210
692
1270
1370
1960
3600
YP
75
154
1410
2060
2510
2920
9990
WB
BB
2
2030
NA
2060
2060
NA
2090
GF*
42
578
1050
1480
1550
1770
4710
LB
26
168
837
1420
1980
2270
8150
WS
57
252
927
1480
2780
2900
44700
CM
LB
12
636
936
1080
1440
1490
3580
PS
9
664
758
854
998
1070
1760
YP
15
681
990
1210
1340
1410
3350
*Goldfish (GF) were not selected as a representative species, but were included because of large sample size and
high tPCB concentrations.
BB
Brown Bullhead
LB
Largemouth Bass
PS
Pumpkinseed
WS
White Sucker
YP
Yellow Perch
CM
Composite
WB
Whole Body
WB-R
Reconstituted Whole Body
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r i r 7/11/2003
-------
2 The shaded box represents the interquartile range, the white bar represents the median, and the whiskers extend to
3 the full range of the data.
4
5 Figure 5.2-2 Box-and-Whisker Plot of Largemouth Bass tPCB Concentrations
6 (Lipid-Normalized) Versus Age
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-16
-------
1 Fish tissue PCB data for samples collected downstream of Woods Pond were also evaluated to
2 determine trends in concentrations for the Rest of River area. Overall, fish tissue tPCB
3 concentrations were significantly lower below Woods Pond (Table 5.2-3) relative to those
4 measured in the PSA. Details on the various sources of downstream fish tissue data, and on the
5 non-EPA data sets considered within the PSA, are provided in Appendix F.
6 5.2.2.2 2,3,7,8-TCDD TEQ
7 Fish tissue TEQ were calculated using the approach outlined in Appendix C.10 (Van den Berg et
8 al. 1998). Some individual congener and dioxin/furan concentrations were below the method
9 detection limit (DL). Following the standard approach to non-detects adopted for this ERA (see
10 Appendix C.2), results were compared using 0 and DL substitution for non-detects. A bounding
11 analysis with the two methods indicated median TEQ concentrations that were within a factor of
12 two for nearly all cases. Table 5.2-4 presents the summary statistics for TEQ concentrations
13 (representative species only) in PSA fish tissue by sample type and species, with DL substitution
14 for non-detects. As with tPCBs, there are trends of increasing TEQ concentrations with age and
15 fish size.
16 5.2.3 Sediment Chemistry Assessment (Exposure to PAHs)
17 There were no data on fish tissue concentrations in the PSA for the eight individual PAHs
18 retained as COCs for fish, or for total PAHs because PAHs are readily metabolized by most
19 aquatic biota, including fish (Johnson 2000). Exposure for these contaminants was therefore
20 assessed based on sediment concentrations only. Table 5.2-5 displays the mean, minimum,
21 median, and maximum main channel sediment concentrations for the eight individual PAH
22 COCs and total PAHs. These sediment concentrations are compared to threshold sediment
23 concentrations in the risk characterization section.
24
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-17
-------
1 Table 5.2-3
2
3 Summary of tPCB Concentrations (mg/kg) from EPA Samples
4 Collected in Reach 8
Sample
Type
Species
Sample
Count
Min
25th
Percentile
Median
Mean
75th
Percentile
Max
WB-R
BB
7
3.46
3.58
3.83
4.97
3.93
12.5
LB
17
1.30
23.0
28.8
38.2
40.6
145
PS
13
5.87
12.7
13.7
14.6
14.9
26.0
YP
6
13.3
23.5
31.8
50.0
41.5
158
WB
LB
14
12.8
18.1
22.4
24.2
29.3
41.5
CM
LB
5
9.98
10.5
10.6
11.9
13.0
15.3
PS
5
9.74
9.98
10.4
10.5
10.7
11.8
YP
5
8.08
8.70
8.91
9.62
11.2
11.2
5
BB
Brown Bullhead
6
LB
Largemouth Bass
7
PS
Pumpkinseed
8
YP
Yellow Perch
9
CM
Composite
10
WB
Whole Body
11
WB-R
Reconstituted Whole Body
12
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-18
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Table 5.2-4
TEQ in Representative Species Fish Tissue in the PSA with DL Substitution
for NDs (ng/kg); Data from EPA Fish Collections (1998-2000)
Sample Type
Species
Sample
Count
Min
25th Percentile
Median
Mean
75th
Percentile
Max
WB-R
BB
31
22.9
48.8
62.2
70.3
86.2
152
LB
29
20.3
48.0
58.5
78.2
100
196
PS
31
24.3
33.6
41.3
44.5
51.3
108
YP
45
17.8
74.2
91.7
102
122
246
WB
BB
1
43.3
NA
NA
NA
NA
43.3
GF*
29
37.8
69.0
104
118
142
288
LB
15
12.6
23.6
37.5
43.0
54.1
86.9
CM
LB
12
29.2
36.8
41.9
46.1
57.2
63.1
PS
9
27.7
31.7
32.7
36.1
42.0
47.1
YP
15
34.2
41.3
42.9
45.8
52.3
63.1
DL=detection limit
ND=non-detect
*Goldfish (GF) were not selected as a representative species, but were included because of large sample size and
high tPCB concentrations.
BB Brown Bullhead
LB Largemouth Bass
PS Pumpkinseed
YP Yellow Perch
CM Composite
WB Whole Body
WB-R Reconstituted Whole Body
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-19
-------
1
2
3
Table 5.2-5
Summary Statistics for Concentrations of PAH COCs in Main Channel Sediment by Reach
5A
5B
5C
6
o
II
a
n = 6
n = 11
n = 3
Mean
Median
Range
Mean
Median
Range
Mean
Median
Range
Mean
Median
Range
Contaminant
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
(mg/kg)
Total PAHs
21.8
7.72
0.37-159
3.13
3.65
0.056 - 5.48
26.8
3.62
0.289 - 255
3.13
3.28
1.9-4.20
Anthracene
0.982
0.18
0.023 - 11
0.220
0.12
0.076 - 0.48
1.59
0.46
0.024 - 14
0.435
0.078
0.026 - 1.2
Benzo(a)anthracene
1.96
0.635
0.03 - 15
0.337
0.375
0.072 - 0.48
2.07
0.26
0.025 - 20
0.213
0.19
0.15-0.3
Benzo(g,h,i)perylene
0.706
0.34
0.022 - 3.8
0.247
0.21
0.076 - 0.48
0.628
0.22
0.03-4.9
0.217
0.25
0.12-0.28
Fluoranthene
3.15
1.1
0.058 - 20
0.445
0.475
0.027 - 0.84
4.13
0.42
0.037 -40
0.413
0.46
0.26-0.52
Fluorene
0.469
0.125
0.031 -4
0.203
0.112
0.032 - 0.48
1.32
0.57
0.048 - 10
0.601
0.56
0.043 - 1.2
Indeno( 1,2,3 -cd)pyrene
0.748
0.345
0.021-4.4
0.232
0.215
0.066 - 0.48
0.623
0.2
0.026 - 5
0.197
0.22
0.1-0.27
Phenanthrene
3.16
0.73
0.037 -29
0.423
0.405
0.08-0.84
5.20
0.24
0.096 - 54
0.256
0.33
0.079 -0.36
Pyrene
3.26
1.1
0.058 - 22
0.440
0.495
0.029-0.75
3.98
0.61
0.054 -36
0.490
0.54
0.34-0.59
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-20
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
5.3 EFFECTS ASSESSMENT
This section describes the literature and site-specific studies used to characterize the effects of
PCBs and other COCs to fish in the Housatonic River PSA. Results of site-specific fish toxicity
studies and literature effects levels were synthesized to develop tissue concentration ranges at
which adverse developmental effects can be expected in the representative fish species in the
Housatonic River.
Results of field surveys are described in Section 5.4 (Risk Characterization) because they are
largely qualitative in nature and could not be used to derive effects metrics in the fish ERA.
Three sources of data were considered in the development of tissue effects thresholds for PCBs
and TEQ (Figure 5.1-3). These include the following:
¦ General Literature—The literature review evaluated the range of PCB and TEQ
concentrations observed to cause adverse effects to ecologically relevant endpoints in
fish, such as reproduction and development. The types of responses observed in PCB
and dioxin spiking studies were also evaluated, to assess whether the effects observed
in the Phase I and Phase II studies were consistent with a PCB and/or dioxin-based
mechanism of action. The literature review emphasized studies conducted in the
laboratory using freshwater fish species, and for which egg or whole body tissue
concentration data were reported. In addition, a review was conducted to assess field
studies, marine fish species, and thresholds based on liver concentrations (Attachment
F.6).
¦ Phase I Site-Specific Studies—These studies investigated contaminant accumulation
and effects in largemouth bass from the PSA and in their offspring. Adult largemouth
bass from the PSA, Rising Pond, and a reference area (Threemile Pond) were
evaluated for contaminant-related biochemical, cellular, and organism-level effects.
Adult fish were spawned and the development of their offspring was monitored for
survival, development, gross abnormalities, and biochemical alterations.
¦ Phase II Site-Specific Studies—These studies investigated the effects of organic
contaminants extracted from largemouth bass tissue collected in the Phase I studies
on the development of uncontaminated largemouth bass, medaka, and rainbow trout
eggs exposed via injection. This study design simulated the maternal transfer of
contaminants to developing oocytes.
In this ERA, separate effects thresholds were derived for each of the sources of data described
above. Concordance between these separate effect levels was observed not only in the
magnitude of the thresholds derived, but also in the type of effects observed.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-21
-------
1 Effects of PCBs that have been documented on fish include mortality, growth-related effects,
2 behavioral responses, biochemical alterations, and adverse reproductive effects. Of particular
3 concern are the effects of dioxin-like PCB congeners that have the same toxic mechanism as
4 2,3,7,8-TCDD (Walker and Peterson 1991; Zabel et al. 1995a). Reproductive and developmental
5 effects have also been observed in fish exposed to 2,3,7,8-TCDD or other dioxin-like substances.
6 The early life stages of offspring from exposed adults are more sensitive to TCDD toxicity than
7 are adults (Walker et al. 1994). The effects of TCDD in sac fry include yolk sac edema,
8 pericardial edema, multifocal hemorrhages, craniofacial malformations, and mortality (Zabel et
9 al. 1995b).
10 5.3.1 Derivation of Literature Tissue Effects Metrics
11 A literature review was conducted to develop threshold effects concentrations for species that
12 occur in the Housatonic River.
13 5.3.1.1 Total PCBs
14 A total of 39 scientific papers were reviewed to identify the range of tPCB concentrations
15 associated with adverse effects on survival, growth, and reproductive success in fish. The papers
16 were screened using the criteria summarized in Table 5.3-1. Of the papers reviewed for the
17 effects of PCBs on fish, 6 met the screening criteria outlined above. The effects and no-effects
18 levels from these studies are shown in Figure 5.3-1. None of these articles described controlled
19 toxicity studies performed directly on the representative fish species selected for the Housatonic
20 River. Reported tissue concentration LOAELs ranged from 1.53 mg/kg for increased mortality
21 in lake trout (Salvelinus namaycush) sac fry (Berlin et al. 1981) to 125 mg/kg for fry mortality in
22 brook trout {Salvelinus fontinalis) that were exposed to Aroclor 1254 in water pre- and post-
23 hatch (Mauck et al. 1978).
24 A number of different methods have been used to select tissue effects concentrations, based on
25 approaches used in deriving sediment and water quality criteria and guidelines. Given the
26 uncertainties in relying on one method only, the following potential effects thresholds were
27 calculated, following the lines of evidence approach by EPA (1999):
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-22
-------
Table 5.3-1
Criteria Used To Screen Available Studies for Determining Threshold Body
Burdens
Criterion
Decision
Accept
Reject
PCBs and TCDD
Body burden data
Reported (whole body preferred)
Not reported or reported fillet
concentrations only
Endpoints
Population-level reproductive,
development and survival effects
Chemical level (i.e., enzyme induction)
effects
Exposure route
Studies mimicking maternal transfer,
exposure of eggs, juveniles and adults
via diet, water and/or sediment
Intraperitoneal injection of adults
Statistics
Study included a control
No control
PCBs only
PCB type
Aroclor 1254, Aroclor 1260, Clophen
A50, tPCBs
Other PCB mixtures or individual
congeners, or when co-occurring
contaminants present
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
-------
100
90
80
70
60
50
0)
Q- 40
30
20
10
— 92.9 [jg/g ww - paired NOAEL/LOAEL geomeans
— 71 [jg/g ww - highest NOAEL
— 61 [jg/g ww - average of all effects
— 16.5 [jg/g ww - 50th percentile of juvenile tissue effects
- 6.3 [jg/g ww - 50th percentile of all tissue effects
- 2.98 [jg/g ww - 10th percentile of juvenile tissue effects
1.53 [jg/g ww - lowest LOAEL
— 1.57 [jg/g ww - 10th percentile of all LOAELs
1.64 [jg/g ww - 10th percentile of all effects
10 100
PCB tissue concentration (|jg/g ww)
Developmental
Endpoints
~ NOAEL
• LOAEL
A Effects
Survival
Endpoints
¦ Effects
1000
Figure 5.3-1 Literature-Based PCB Fish Tissue Effects Concentrations
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-24
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
¦ Highest no observed adverse effect levels (NOAEL) and lowest observed adverse
effect levels (LOAEL) reported in the literature for fish.
¦ Selected percentiles for various effects concentrations and tissue types - 10th and 50th
percentiles calculated for all effects data and for groups of effects data (i.e., all
LOAELs; all effects concentrations; egg tissue effects concentrations; juvenile tissue
effects concentrations; and adult tissue effects concentrations where possible), where
sufficient data-points exited.
¦ Geometric means of paired NOAELs and LOAELs.
¦ The effects data and the potential effects thresholds for PCBs are shown in Figure
5.3-1. There are two groups of results, one in the range of 1-11 mg/kg ww and the
other in the range of 61-90 mg/kg ww. The lower results (i.e., 1-11 mg/kg)
correspond to the lowest LOAEL selected from the literature, the 10th and 50th
percentiles for sac-fry/juvenile effects, and, the 10th and 50th percentiles of the
concentrations in all life stages for which data were available (i.e., egg, sac-fry,
juvenile). The higher group of results (i.e., 61-90 mg/kg) corresponds to the highest
NOAEL reported, the paired NOAEL/LOAEL geometric means, and the mean
concentration for effects observed in all tissues (i.e., egg, sac-fry, juvenile).
Based on the lines of evidence approach, a threshold effects concentration of 61 mg/kg ww tPCB
for egg/sac-fry tissue was chosen. This value corresponds to the following:
¦ The average concentration for all effects reported in the studies used (61 mg/kg).
¦ The highest NOAEL reported in the studies used (71 mg/kg ww).
¦ The geometric mean of the paired NOAEL/LOAELs reported in the studies used
(92.9 mg/kg ww).
To scale the selected egg/sac-fry tissue concentration to a whole body concentration for
warmwater fish, a factor of 0.5 was applied based on site-specific and literature information
indicating that egg PCB concentrations are higher than the maternal whole body PCB
concentrations in PSA fish species (Section F.3.4.2.2). As a result, a whole body tissue
concentration of 31 mg/kg ww is recommended and is expected to be protective of reproductive
and developmental endpoints for PSA fish species.
In summary, adult fish with tissue concentrations greater than 31 mg/kg ww may have reduced
reproductive success and/or their offspring may experience adverse early life stage
developmental effects. Attachment F.4 provides additional details for the derivation of this
threshold. This effects threshold is consistent with a recent review of PCB toxicity conducted by
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r rs c 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
the National Oceanographic and Atmospheric Administration (Monosson 1999). The NOAA
study concludes that Aroclor 1254 at concentrations ranging from 25 mg/kg to 70 mg/kg in the
liver of adult fish "interferes with the proper functioning of the reproductive system." Because
the report also concludes that the "liver is estimated to have similar concentrations as the whole
body or eggs," the NOAA effects range is consistent with the 31 mg/kg threshold derived above.
Additional information on liver effects thresholds and field-based assessments of PCB toxicity
are provided in Attachment F.6.
5.3.1.2 2,3,7,8-TCDD TEQ
Nineteen papers were reviewed to identify the range of TEQ concentrations associated with
adverse effects on survival, growth, and reproductive success in fish. The same screening
criteria applied to tPCBs were used to select the relevant papers (Table 5.3-1). The 11 studies
that met the screening criteria involved maternal transfer to eggs from adults fed TEQ-
contaminated diets, waterborne exposure of juveniles, or injection of eggs. Most studies
provided egg concentrations, with only one of the papers reporting adult female whole body
contaminant concentrations.
Nine of the 11 studies included effects for mortality of sac fry, a reproductive effect. The species
included in these studies were primarily trout (lake trout, rainbow trout, and brook trout), but the
El on en et al. (1998) study included assessment of several warmwater species, including white
sucker (a representative species in the PSA). Warmwater species were less sensitive than most
trout species, with increases in early-lifestage mortality observed in the 100 to 1,000 ng/kg
TCDD range (Elonen et al. 1998).
These data are summarized in Figure 5.3-2. There appear to be two general groups of results—a
lower group at approximately 50 ng/kg ww, and a higher group between 400 and 1200 ng/kg
ww. Basing a threshold concentration for adult whole bodies on the lower group (which
corresponds to the lowest LOAEL observed, and the 10th percentile of egg concentrations at
which effects were observed) may be overly conservative due to the known sensitivity of the
trout species (e.g., lake trout) used in those studies. Conversely, basing the threshold
concentration on the higher group may not be protective against adverse effects for all the
species of concern in the Housatonic River.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC £ O A
-------
100
90
80
70
60
S
- 50
40
30
20
10
10
A A
- 40 ng/kg ww - lowest LOAEL
A A
A
A
51 ng/kg ww - 10th percentile of LOAELs
54 ng/kg ww - 10th percentile of all egg effects
~
~
••
A A
A
- 437 ng/kg ww - 50th of LOAELs
_ 560 ng/kg ww - geomean of paired LOAEL/NOAELs
577 ng/kg ww - 50th percentile of all egg effects
¦ 1,220 ng/kg ww - highest NOAEL
A A
100 1,000
TCDD tissue concentration (ng/kg ww)
Developmental
Endpoints
~ NOAEL
• LOAEL
A Effects
10,000
Figure 5.3-2 Literature-Based TCDD (TEQ) Fish Tissue Effects Concentrations
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-27
-------
1 An egg effects threshold of 100 ng/kg TEQ, which is intermediate between the high and low
2 groups discussed above, was selected for the Housatonic River PSA fish. This concentration
3 represents the level at which early lifestage mortality starts to increase in several species,
4 including warmwater fish species. Although the mortality rates at 100 ng/kg are not statistically
5 significant (i.e., NOAELs in Elonen et al. 1998 are 270 ng/kg and higher), the gradient in
6 toxicity is fairly steep above 100 ng/kg, such that approximately 50% mortality to several species
7 is observed at 1,000 ng/kg TEQ (see Attachment F.4).
8 As with the tPCB threshold derivation, the threshold egg concentration was scaled to an adult
9 whole body concentration based on the site-specific data and literature information (Niimi 1983;
10 Monosson 1999) indicating that the concentration of PCBs in eggs will be higher than the
11 maternal whole body. Using a conversion factor of 0.5 (Section F.3.4.2.2), a whole body tissue
12 concentration of 50 ng/kg ww TEQ was derived and is expected to be protective of sensitive
13 reproductive and developmental endpoints in PSA fish. Adult fish with tissue concentrations
14 greater than 50 ng/kg ww may have reduced reproductive success and/or their offspring may
15 experience adverse early life stage developmental effects.
16 5.3.2 Site-Specific Toxicity Studies
17 5.3.2.1 Housatonic River Fish Toxicity Study - Phase I
18 5.3.2.1.1 Methods
19 The Phase I study for the Housatonic River fish reproductive health assessment was conducted
20 by the USGS Columbia Environmental Research Center (CERC) (Tillitt et al. 2003a). The study
21 investigated contaminant-associated effects in fish collected from the PSA and spawned under
22 controlled conditions. The main objective of the study was to characterize differences that may
23 be related to PCB toxicity between Housatonic River fish and those at a reference area
24 (Threemile Pond).
25 Adult largemouth bass were collected from two Housatonic River locations within the PSA
26 (Reach 5C and Reach 6), from Rising Pond (Reach 8) and from the Threemile Pond reference
27 location. Adult fish (both pre- and post-spawning) were analyzed for elevated liver enzymes
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC ^ 9 S 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
(EROD), abnormal gonadal and liver histology, elevated occurrences and rates of macrophage
aggregates, body and organ sizes and weights, and steroid hormone concentrations.
Adult largemouth bass from the PSA and the reference area were spawned, and their lab-reared
offspring were monitored for survival, developmental delays and deformities, growth, and
cytochrome P450 induction.
5.3.2.1.2 Results
Adult largemouth bass from the Housatonic River sites exhibited multiple sublethal effects at
frequencies of occurrence that were statistically different from those observed in bass from the
reference location.
Phase I Adult Largemouth Bass - Effects Observed
¦ Elevated EROD levels In livers.
¦ Thickened lobule wall in testes.
¦ Elevated occurrence of macrophage aggregates.
¦ Reduced growth in females.
¦ Reduced estrogen levels.
The enzymatic responses, histopathologies, macrophage aggregates, and estrogen responses
listed above are biomarkers for the induction of organic contaminant-derived responses.
Although not necessarily indicative of ecologically significant impairment to the adults
themselves, they are indications of biological and chemical alterations that may lead to
reproductive effects. These effects are also consistent with the summary of Monosson (1999) in
which Aroclor 1254 was found to interfere with the reproductive system (altered steroid
hormone metabolism, altered testes and ovarian development, and altered concentrations of
neurotransmitters and gonadotrophins) at concentrations of approximately 25 to 70 mg/kg tPCB
in liver tissue.
Effects were also observed in offspring from the contaminated areas that were statistically
different from those observed in offspring from the reference site. At 15 days post swim-up,
deformities were observed in fry from all three reaches on the Housatonic River, while none
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC £ O Q 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
were observed in fry from the reference location. Partially external swim bladders, an extremely
unusual deformity that has also been observed in dioxin dosing studies, were observed at all
three Housatonic River reaches. In addition, shortened opercula were observed in 22% of fry
from Reach 5BC and tail deformities were observed in 2% of fry from Reach 6 at 15 days post
swim-up; no opercular or tail deformities were observed in fry from the other Housatonic River
locations or the reference location.
Phase I Largemouth Bass Offspring - Effects Observed
¦ Survival - Reduced survival from hatch to swim-up, or reduced survival post
swim-up.
¦ Development - Developmental delays (increased days to swim-up).
¦ Growth - Reduced growth from swim-up to 15 days post swim-up.
¦ Deformities - Increase in eye deformities from hatch to swim-up; shortened
opercula; tail deformities; external swim bladders.
In summary, the effects observed in the Phase I study were suggestive of PCB-related toxicosis.
Although responses were not all consistent across all exposed reaches, the adults and offspring
both exhibited a suite of symptoms that was consistent with PCB-related toxicity.
5.3.2.2 Housatonic River Fish Toxicity Studies - Phase II
5.3.2.2.1 Methods
In Phase II of the fish toxicity studies conducted by USGS CERC (Tillitt et al. 2003b), organic
contaminants present in the largemouth bass from the Phase I studies were extracted and injected
into cultured eggs of largemouth bass, medaka, and rainbow trout. This study provided a
simulation of maternal transfer of PCB contamination to offspring.
Egg Production and Injection
Largemouth bass and medaka eggs were produced by brood stocks held in experimental ponds
and rainbow trout eggs were produced in the laboratory using unfertilized rainbow trout eggs and
milt obtained from a hatchery. Eggs were randomly distributed to each treatment. The number
of trials for each species/life stage/treatment combination varied with egg availability and quality
(i.e., control survival). Each trial had three replicates, with 10 eggs per replicate for largemouth
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC £ ^ n 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
bass and medaka, and approximately 24 eggs per replicate for rainbow trout. Trials for each
species/life stage/treatment combination were conducted at different times.
Extracts were produced from tissue homogenates of largemouth bass collected from Reaches
5BC (i.e., Reaches 5B and 5C combined), Reach 6, and Reach 8 and the Three-mile Pond
reference location during the Phase I study. Extracts for each site were diluted into five different
concentrations and injected into clean largemouth bass, rainbow trout, and medaka eggs;
concentrations of tPCBs (mg tPCB/kg egg) and TEQ (ng TEQ/kg egg) for each dose group were
calculated. For comparative purposes, PCB-126 and 2,3,7,8-TCDD (TCDD) standards diluted
into six different concentrations were injected into the eggs. Uninjected eggs and eggs injected
with triolein were used as negative controls.
Assessment of Lethal and Sublethal Effects
The assessment of lethal and sublethal effects in fish was focused on the later (i.e., swim-up and
post swim-up) stages of development. At early stages of development (i.e., hatch), effects are
less pronounced or likely to occur because chemicals present in the egg yolk have not yet been
fully absorbed by the developing fish (Papoulias, personal communication 2003a). At later
stages of development, after yolk absorption is complete, effects are more apparent.
Accordingly, data collection, analyses, and interpretation were focused on effects on swim-up
and post swim-up fish.
The percent survival was determined for each treatment group and compared to controls. To
ensure data quality, trials in which largemouth bass survival was <50% or medaka or rainbow
trout survival were <70% were excluded from statistical analyses, in accordance with standard
toxicity testing data quality objectives (ASTM 2002). LD50 (or lethal dose 50) values were also
calculated to determine the concentration at which mortality is observed in 50% of the
population, relative to the negative controls.
Lengths and weights of largemouth bass and medaka were measured at the end of the experiment
(i.e., 15 days post swim-up) and compared to controls. Lengths and weights of rainbow trout
were not measured.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-31
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Each fish was examined for deformities at important life stages. For each deformity, all trials for
each species/life stage/treatment combination were combined and normalized for sample size
(per 1,000 fish) to allow for comparisons across treatment groups with different sample sizes.
ED50 values (or 50% effective doses for sublethal and lethal effects), which represent the
concentration at which mortality or one or more abnormalities were observed in 50% of the
population, were calculated relative to the negative controls. Trials were included in statistical
evaluations where effects (mortality or one or more pathologies were observed) were <50% in
negative controls and >50% in the high dose treatment groups (ASTM 2002).
5.3.2.2.2 Results
The following effects were observed in offspring of largemouth bass, medaka, and rainbow trout
exposed to extracts from the Housatonic River and PCB-126 and TCDD standards.
Survival and Growth
Survival of largemouth bass, medaka, and rainbow trout was assessed at swim-up, swim-up and
3 and 15 days post swim-up, and 600 daily temperature units (DTU; approximately 3 days post
swim-up), respectively.
Statistically significant reductions in survival were most evident in fish exposed to PCB and
TCDD standards. Reduced survival was also observed in medaka exposed to Housatonic River
extracts. Between 3 and 15 days post swim-up, medaka exposed to extracts from Reach 5BC
and Reach 6 exhibited statistically significant reductions in survival relative to control fry.
Survival was not affected in largemouth bass and rainbow trout exposed to Housatonic River
extracts.
High mortality was observed in largemouth bass control fish between 3 and 15 days post swim-
up. Largemouth bass, which are not typically used in toxicity testing, encountered difficulties
during the transition from endogenous to exogenous feeding and starved to death. Other sources
of variability in survival data (for all fish) were the occurrence of fungal infections, as well as
differences in egg quality and unknown sources of variation associated with temporally distinct
trials.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-32
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
LD5os were calculated for select trials for largemouth bass (swim-up), medaka (3 and 15 days
post swim-up), and rainbow trout (600 DTU) for Reaches 5BC and 6, PCB-126, and/or TCDD.
LD50s estimates could not be calculated for trials for all species and life stage combinations
because the data did not pass the acceptance criterion and/or the magnitude of the response
observed was not sufficient. LD50s (confidence intervals) for Reach 5BC extracts ranged from
16 (-7 to 40) to 95 (75 to 116) TEQ/kg egg. LD50s for Reach 6 ranged from 7 (-18 to 31) ng
TEQ/kg egg to 178 (68-287) ng TEQ/kg egg. LD50s for PCB-126 ranged from 580 (379 to 782)
ng TEQ/kg egg to 5,217 (3,610 to 6,824) ng TEQ/kg egg. TCDD LD50s ranged from 889 (673-
760) ng TEQ/kg egg to 5,481 (3,950 to 7,013) ng TEQ/kg egg.
Overall, medaka at 15 days post swim-up exhibited the lowest LD50s, relative to other species,
for all extracts and standards, with the exception of TCDD. The overall results (i.e., order of
magnitude difference in TEQ-based LD50s between site extracts and standards) appears to
indicate that the Housatonic River extracts are more toxic than would be predicted on the basis of
an additive model of dioxin-like toxicity alone. The increased toxicity observed with the
Housatonic River extracts could be attributed to synergistic effects of PCB mixtures and effects
of other PCBs in the mixture that are not considered using the TEQ approach (Tillitt, personal
communication 2003).
Largemouth bass and medaka length and weight (growth endpoints) were not affected by
exposure to extracts or standards.
Individual Deformities
At certain stages of development, largemouth bass, medaka, and rainbow trout were examined
for a variety of abnormalities. Several abnormalities exhibited an apparent dose-related or
threshold response to high doses of Housatonic River extracts or standards, relative to control
fish.
Overall, increased rates of deformities were most evident in fish at swim-up and post swim-up
following in ovo exposure to PCB and TCDD standards. Fish exposed in ovo to high doses of
these standards exhibited a variety of gross pathologies that are characteristic of PCB and dioxin
exposure (craniofacial deformities, spinal deformities, swim bladder deformities, hemorrhage,
pericardial edema, peritoneal edema, yolk sac edemas, and delayed development). Generally,
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC Z 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
rainbow trout appeared to be the most sensitive of the three species evaluated and medaka
appeared to be the least sensitive.
Fish exposed in ovo to high doses of Housatonic River extracts exhibited similar types of gross
pathologies as the dioxin-like standards, including craniofacial deformities, fin deformities,
spinal deformities, swim bladder deformities, hemorrhage, pericardial edema, peritoneal edema,
yolk sac edemas, and larval weakness and delayed development, along with a group of "other"
abnormalities. Rainbow trout exhibited the largest magnitude dose-related response to
Housatonic River extracts; delayed development was observed in 62% of trout exposed to Reach
6 extract containing 83 mg tPCBs/kg egg. Some of the pathologies were only observed in one of
the three species. For example, rainbow trout were the only species that exhibited opercular
deformities.
Some of the deformities observed in fish were only weakly related to tPCB or TEQ
concentrations for one species/life stage/treatment combination. The lack of a dose-response in
fish injected with Housatonic River extracts and/or PCB and TCDD standards and the
occurrence of these deformities in fish injected with control and reference site extracts indicates
that these abnormalities are not the most reliable markers of PCB exposure.
Total Abnormalities
To provide an overall picture of relationship between PCB exposure and the occurrence of
abnormalities, the proportion of fish exhibiting one or more abnormalities was determined and
compared to the control fish (at swim-up and post swim-up). In some cases, the highest
frequency of abnormalities was observed in the second highest dose group; the reason for this
was not apparent. Fish exposed to PCB and TCDD standards exhibited significantly higher
percentages of abnormalities, relative to control fish. Rainbow trout were the most sensitive of
the three species. Similar, but more variable, dose-response relationships were observed in fish
exposed to Housatonic River extracts. Figures 5.3-3 to 5.3-7 summarize the concentration-
response relationships for the PSA reaches. Again, rainbow trout appeared to be the most
sensitive of the three species.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-34
-------
40
30
20
10
Uninjected Triolein 1 2 3 4 5
Dose Identification Number
2 Notes: Bar height indicates percentage of fish affected with one or more pathologies.
3 Asterisks indicate significant differences from negative controls (uninjected and triolein).
4 Doses are 1, 2, 19, 93, and 185 mg/kg tPCB and 0.6, 1, 12, 59, and 118 ng/kg TEQ for Dose IDs 1-5.
5 Source: Adapted from Tillitt et al. 2003b.
6 Figure 5.3-3 Effects of in Ovo Exposure to Increasing Doses of Reach 6
7 Extracts on Largemouth Bass at Swim-Up
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-35
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Uninjected Triolein 1 2 3 4 5
Dose Identification Number
Notes: Bar height indicates percentage of fish affected with one or more pathologies.
Asterisks indicate significant differences from negative controls (uninjected and triolein).
Doses are 2, 3, 31, 155, and 310 mg/kg tPCB and 0.6, 1, 13, 64, and 128 ng/kg TEQ for Dose IDs 1-5.
Source: Adapted from Tillitt et al. 2003b.
Figure 5.3-4 Effects of in Ovo Exposure to Increasing Doses of Reach 5BC
Extracts on Medaka at 5d Post Swim-Up
251
20 -
¦s
0)
2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
60
50
40
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Cytochrome P450
Cytochrome P450 induction was evaluated qualitatively in largemouth bass, medaka, and
rainbow trout tissues using immunochemical histological techniques. Cytochrome P450
induction was observed in fish exposed to both standards and Housatonic River extracts.
Rainbow trout was the most sensitive test species, exhibiting apparent dose-related increases in
cytochrome P450 induction. The strongest response (i.e., highest induction) was observed in
trout exposed to Reach 5BC extracts. Low and moderate level cytochrome P450 induction was
observed in bass exposed to 6 |ig TCDD/kg egg and medaka exposed to 2 to 6 |ig TCDD/kg egg.
Medaka also exhibited moderate dose-related cytochrome P450 induction following exposure to
reference site extracts containing 0.15 mg tPCBs/kg egg. Largemouth bass did not appear to
exhibit a dose-related induction of cytochrome P450 following exposure to Housatonic River
extracts.
5.3.2.2.3 Study Conclusions
A high degree of variability was observed in many of the parameters evaluated in the Phase II
study. However, despite this variability, an overall pattern of PCB-related toxicity was apparent.
The types of abnormalities observed in fish exposed to Housatonic River extracts in the Phase II
study corresponded with abnormalities reported in the Phase I study, as well as with dioxin-like
effects documented in the literature. As expected, effects were most pronounced at later stages
of development (i.e., swim-up and post swim-up), after maximum contaminant exposure (i.e.,
completion of yolk absorption) occurred. Rainbow trout and medaka appeared to be more
sensitive to the Housatonic River extracts and PCB-126 and TCDD standards than largemouth
bass.
Given the range of tPCB and TEQ concentrations used in the Phase II study, it was not
unexpected that the magnitude of effects observed with the standards (2,3,7,8-TCDD and PCB-
126) would be greater than those observed with the Housatonic River extracts. However, when
ED50 concentrations were normalized using TEQ, the Housatonic River extracts were more toxic
than the standards. The increased toxicity associated with the Housatonic River extracts could
be attributed to synergistic toxicity of the PCB mixtures, as well as the effects of PCBs that are
not incorporated into the TEQ model (Tillitt, personal communication 2003).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC 5 38 7/11/2003
-------
1 5.3.3 Concentration-Response Analysis - Toxicity Endpoints
2 5.3.3.1 Phase I Study Threshold Effects Concentration Derivation
3 The Phase I fish toxicity study identified reproductive effects including reduced survival and
4 growth, as well as developmental delays and deformities, in Housatonic River offspring.
5 Specific abnormalities were observed in Housatonic River fish that were not observed in the fish
6 from the Threemile Pond reference location.
7 Both tPCB and TEQ tissue effects thresholds were derived from the Phase I study results. These
8 thresholds are not bounded because adverse effects were observed in spawn of bass from all
9 three Housatonic River sites, at the lowest tissue concentration measured, 45 mg/kg tPCBs (or 38
10 ng/kg TEQ); consequently, adverse effects may also occur at lower tissue concentrations. These
11 tissue effects thresholds are similar to those derived from a literature review of PCB and dioxin
12 effects (Section 5.3.1) and the Phase II studies (Section 5.3.2.2).
13 5.3.3.2 Phase II Study Threshold Effects Concentration Derivation
14 The primary objective of the Phase II study was to determine the toxic effects of in ovo exposure
15 of fish to extracts containing organic contaminants from adult Housatonic River fish. Although
16 the Phase I study identified a suite of effects that were consistent with PCB-related toxicity, the
17 Phase II study evaluated the cause-effect linkage more directly. The results of the Phase II study
18 indicated that fish exposed to Housatonic River extracts exhibited decreased survival and
19 increased abnormalities and biochemical alterations, in response to high doses of tPCBs and
20 TEQ. The patterns of responses observed were not always consistent across species and
21 treatments; however, gross pathologies observed were characteristic of PCB-related effects
22 reported in the literature and corresponded with a number of the effects observed in the Phase I
23 study.
24 5.3.3.2.1 ED50 Estimates
25 ED50s derived from the Phase II study data were used to develop thresholds for Housatonic River
26 extracts and PCB-126 and TCDD standards. An ED50 value represents the concentration at
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-39
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
which sublethal or lethal effects (i.e., either mortality or one or more abnormalities) was
observed in 50% of the population, relative to the negative controls. This combined toxicity
endpoint provides an indication of the concentration threshold for sublethal and lethal effects in
early life stages of fish. The ED50 endpoint represents a large effect size and indicates an
unacceptable level of biological harm. However, given the extent to which early lifestage fish
may be impacted by multiple natural and chemical stressors, it was considered desirable to select
a large effect size, with high statistical power for detecting a response. Because of the large
effect size, the use of an ED50 endpoint requires some conservatism in the processing of ED50
values from multiple trials and treatments; this process is described below.
ED50 values were calculated using raw data from the Phase II Studies (Attachment F.7), using
methods described in Appendix F (Section F.3.4.2.1). ED50 values calculated for largemouth
bass (swim-up), medaka (swim-up and 15 days post swim-up), and rainbow trout (600 DTU or
approximately 3 days post swim-up) exposed to Housatonic River extracts and PCB-126 and
TCDD standards are presented in Table 5.3.2 and Figures 5.3-8 and 5.3-9. Because Tillitt et al
(2003b) emphasize TEQ as exposure measures, TEQ were converted to tPCB concentrations
using linear regression (i.e., TEQ versus tPCB doses); regression equations used for the
conversions are provided in Appendix F (Table F.3-9). For many of the trials, an ED50 value
could not be calculated because the data did not meet the criteria specified above, the magnitude
of the effect was too small to calculate a toxicity threshold, and/or there was no dose-related
response.
5.3.3.2.2 Threshold Derivation
Rather than selecting an individual ED50 concentration as a threshold value, the entire
distribution of ED50 values was considered. The following procedures were applied in the
derivation of an egg-based maximum acceptable tissue concentration (MATC):
¦ Selection of Controls—Where ED50 values were calculated separately for the two
controls (due to statistically significant differences between controls), the arithmetic
mean of the two ED50 values was used. This prevented bias in the MATC derivation
from double counting the results of a single trial.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-40
-------
1 Table 5.3-2
2
3 Calculated ED50 Values (tPCBs and TEQ) for Largemouth Bass, Medaka, and
4 Rainbow Trout Exposed in Ovo to Housatonic River Extracts and PCB-126 and
5 2,3,7,8-TCDD Standards
Endpoint"
Species
Life Stage
Extract
Concentration1"
TEQ (pg/g)
Largemouth bass
Swim-up
Reach 6 (Trial 1)
136.58 (xgtPCBs/g egg
87
PCB-126 (Trial 2)
657.2 ng PCB-126/g egg
3286
Reach 5BC (Trial 1)
82.49 |xg tPCBs/g egg
(uninjected) 34.11 |xg tPCBs/g
egg (triolein)
34 (uninjected)
14 (triolein)
Reach 5BC (Trial 3)
43.78 |xg tPCBs/g egg
18
Swim-up
PCB-126 (Trial 2)
46.6 ng PCB-126/g egg
(uninjected) 64.4 ng PCB-126/g
egg (triolein)
23 3 (uninjected)
322 (triolein)
Medaka
PCB-126 (Trial 3)
44.8 ng PCB-126/g egg
224
2,3,7,8-TCDD (Trial 3)
2667 ng TCDD/g egg
2667
ed50
Reach 5BC (Trial 1)
48.62 |xg tPCBs/g egg
20
15 days post
swim-up
Reach 5BC (Trial 3)
9.91 |xg tPCBs/g egg (uninjected)
22.25 |xg tPCBs/g egg (triolein)
4.0 (uninjected)
9.1 (triolein)
PCB 126 (Trial 2)
29 ng PCB-126/g egg
145
PCB 126 (Trial 3)
31 ng PCB-126/g egg
155
Reach 6 (Trial 1)
11.85 |xg tPCBs/g egg
7.6
Reach 5BC (Trial 4)
107.21|xg tPCBs/g egg
44
Rainbow trout
600 DTU
PCB-126 (Trial 1)
24.2 ng PCB-126/g egg
121
PCB-126 (Trial 2)
84.6 ng PCB-126/g egg
423
2,3,7,8-TCDD (Trial 1)
294 pg TCDD/g egg
294
2,3,7,8-TCDD (Trial 2)
152 pg TCDD/g egg
152
6 aED50 endpoints were based on the combined effects observed per fish (pathology and mortality combined) for each
7 species/life stage/treatment/dose combination.
8 bTotal PCB concentrations were interpolated from non-standardized concentrations through linear regression. PCB-
9 126 and 2,3,7,8-TCDD concentrations were converted using toxicity equivalent factors for fish provided in Van
10 den Berg et al. (1998)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-41
-------
10000
1000
O)
O)
0)
<3
LU
O)
Q.
100 ¦:
Abbreviations
Species: MED = medaka, LMB = largemouth bass, RBT = rainbow trout; Life Stages: SU = swim up, 15 PSU = 15 d post swim-up; Extracts: TMP = Three Mile Pond
Figure 5.3-8 TEQ ED50 Estimates for Fish Exposed to Housatonic River Extracts and PCB-126 and TCDD
Standards (Logarithmic Scale)
MK0110:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-42
-------
O)
O)
D
t/>
c
O
c
d)
o
c
0
o
Q
Q
O
I-
"O
c
CO
CO
CM
1
CD
O
D.
CD
O
D.
1000
100 -!
~
ED50(s+l) estimate w as greater than the highest dose tested ( >
50% effects in negative controls and < 50% effects in high dose
group).
|ED50(s+l) was calculated ( < 50% effects in negative controls
and > 50% effects in high dose group) and a dose response was
observed.
1 ED50(s+l) was calculated ( < 50% effects in negative controls
and > 50% effects in high dose group) but the dose response
0.1 -S
0.01 -|
0.001 -i
0.0001
LO LO LO
0
0
0
CN
CN
CN
CN
CN
CN
CN
0
0
0
i-
T—
T—
T—
T—
I-
I-
co
m
CQ
CQ
CD
CQ
CQ
I
I
O
0
O
O
O
O
O
D
D
D
Q_
Q_
CL
CL
CL
CL
CL
I-
I-
CO
1
1
1
1
1
1
1
O
O
Q
D
D
D
D
D
D
D
O
O
LU
1-
CO
CO
CO
CO
CO
1-
O
CD
O
CD
O
O
CL-
IO
CL
LO
Q
LU
Q
LU
Q
LU
O
O
I—
CO
I—
CO
O
CD
O
O
O
CD
01
01
1—
CD
cn
LU
LU
1—
CQ
cn
D
C0
CL
LO
o
LU
D
C0
CL
LO
Q
LU
D
C0
CL
LO
o
LU
co
¦o1
CD
co
00
-C
CQ
-C
-C
-C
0
LO
O
0
0
ra
ro
ro
ro
0
0
0
0
(Z
ro
01
01
01
1
0
1
1
1
D
01
D
D
D
1-
>
1-
CO
CO
Q
D
Q
Q_
Q_
O
CO
O
LO
LO
0
Q
0
CD
LU
CD
O
Q
1—
1—
LU
LU
CQ
01
CQ
01
Abbreviations
Species: MED = medaka, LMB = largemouth bass, RBT = rainbow trout; Life Stages: SU = swim up, 15 PSU :
~
CD
O
O
CD
O
CQ
CQ
CD
CQ
LO
-C
Q
LO
LO
O
LO
-C
TO
-C
-C
ro
-C
O
0
O
0
0
0
ro
01
ro
ro
01
ro
0
0
0
0
01
D
01
01
D
01
D
CO
D
D
CO
D
1-
CQ
1-
1-
0
CO
Q
Q
Q
LU
Q_
O
_l
O
O
LO
0
0
0
CD
CD
CD
Q
LU
1—
1—
1—
CQ
CQ
CQ
01
01
01
D
C0
o
LU
15 d post swim-up; Extracts: TMP = Three Mile Pond
Figure 5.3-9
tPCB ED50 Estimates for Fish Exposed to Housatonic River Extracts and PCB-126 and TCDD
Standards (Logarithmic Scale)
MK0110:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-43
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
¦ Exclusion of Reference Station—Threemile pond extracts were excluded from the
MATC derivation; the maximum PCB concentrations tested for these reference fish
(0.15 mg/kg tPCB and 6 pg/g TEQ) were insufficient to yield large toxic responses or
provide meaningful information on the magnitude of the ED50 value. The
concentrations in these extracts were well below the levels causing effects in the
contaminated site extracts.
¦ Treatment of Unbounded ED50 —Where no ED50 value could be determined, a
value was set equal to the highest concentration tested. This approach is
conservative, since effects in these trials would theoretically only have occurred at
concentrations higher than the highest concentration tested. However, it was
considered necessary to include results from all trials (including those that did not
yield a 50% response), and some conservatism was warranted due to the large effect
size under consideration.
¦ Statistical Measure —The arithmetic mean of the relevant ED50 values was chosen
as the MATC. The analysis was also conducted using median values, and the results
were very similar. Use of a central tendency value ensures that the thresholds derived
are not based on results from a single trial, and balances sensitive trials with those
that did not yield large effect sizes.
¦ Species Included —As indicated above, the ED50 values were variable but did not
differ substantially among the three test species. The mean rainbow trout ED50 values
were only slightly lower than the combined medaka and largemouth bass values (i.e.,
within a factor of two). Therefore, the ED50 values from all three species were
combined to yield an integrated ED50 value for all species that is deemed protective of
all PSA fish species.
Based on the above criteria, ED50 values in eggs were calculated as 104 mg/kg tPCB and 89 pg/g
TEQ.
Before the threshold values could be applied in risk characterization, the egg concentrations first
required conversion to whole body concentrations. For this purpose, three independent
evaluations of egg versus whole weight concentrations were considered:
¦ Site-Specific Conversion—PCB chemistry data were available in whole bodies and
ovaries of largemouth bass. These data have the advantage of site-specificity, but are
limited to a single species, and require the assumption that ovaries and eggs are in
chemical equilibrium during the reproductive phase.
¦ Direct Comparison of Eggs to Whole Bodies (from Literature)—Niimi (1983)
measured the concentrations in whole body and egg tissue concurrently for several
freshwater species. These data have the advantage of multiple species, some of
which are found in the PSA, but are not site-specific.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-44
-------
1 ¦ Direct Comparison of Eggs to Muscle (from Literature)—Monosson (1999)
2 compiled information on ratios of ovary or egg concentrations relative to muscle
3 tissue (fillet). Results of eleven studies are presented; however the approach requires
4 estimation of the relationship between muscle and whole body concentrations.
5 The greatest weight was placed on the first method, due to site-specificity, with moderate weight
6 placed on the Niimi (1983) information. The lowest weight was placed on the Monosson (1999)
7 data due to the extrapolations required and lack of site- and species-specificity. Details on each
8 method are provided in Appendix F (Section F.3.4.2.2).
9 Considering the three methods for extrapolation of concentrations from egg to adult whole body,
10 a value of 0.5 was selected, and was applied for both tPCB and TEQ. This value is slightly
11 higher than the site-specific ratios based on regression of site-specific ovary and whole body
12 data, but is lower than the values indicated by the literature reviews.
13
14
15
16
17
18
19
20
21
22
23
24
25 5.3.4 Derivation of Literature-Based Sediment Effects Metrics for PAHs
26 PAHs tend to be associated with bottom sediment because of the chemical and physical
27 properties that affect their environmental fate. As a result, fish that are closely associated with
28 sediment (i.e., benthic species) are most affected through direct contact or incidental ingestion of
29 PAH-contaminated sediment. Exposure to PAHs in sediment can result in reproductive,
30 developmental, and carcinogenic effects in fish.
Conversion of tPCB Egg Threshold to Whole Body Threshold
¦ The mean ED50 value fortPCB Phase II warmwater fish toxicity (largemouth
bass, medaka, and rainbow trout) was 104 mg tPCBs/kg egg.
¦ The mean ED50 value for TEQ Phase II fish toxicity (largemouth bass, medaka,
and rainbow trout) was 89 ng TEQ/kg egg.
¦ A review of egg to whole body conversion factors for PCBs and TEQ was
conducted using site-specific and literature information, yielding an estimate of
0.5. This value was used to extrapolate tPCB egg concentrations to whole body
concentrations.
¦ The threshold egg concentrations were converted to whole body adult tissue
concentrations of 52 mg tPCBs/kg and 45 ng TEQ/kg warmwater.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-45
-------
1
2
3
4
5
6
7
8
9
10
11 The assessment of fish toxicity of PAHs is complicated by the fact that PAHs are readily
12 metabolized by most aquatic animals, including teleost fish (Johnson et al. 2002). Because tissue
13 concentrations are not a reliable predictor of adverse effects in fish, the relationship between
14 exposure to contaminated sediment and effects can be used to derive an effects threshold
15 (described below).
16 Development of the sediment PAH effects metrics were based on two major groups of studies of
17 field exposures of benthic fish, conducted in Black River, Ohio, and in the Puget Sound,
18 Washington (detailed in Appendix F; Attachment F.5). Because these studies investigated
19 effects in benthic fish with high sediment exposures, the thresholds derived from these studies
20 represent conservative estimates, which are expected to be protective of fish species present in
21 the Housatonic River. Lack of biomagnification of PAHs (due to metabolism) means that fish
22 feeding directly on bottom sediment and associated prey are at the greatest risk from PAH
23 contamination.
24
25
26
27
28
29 The 10 mg/kg threshold was considered the most appropriate value for application to the
30 Housatonic ERA because of common environments (i.e., freshwater), species (i.e., brown
31 bullhead), and endpoints of interest. The worst-case threshold of 1 mg/kg provides a level at
Potential Effects of PAHs to Fish
¦ Reproductive and Developmental Effects - PAHs have been shown to be
immunotoxic and to have adverse effects on reproduction (reduced egg fertility,
increased fry mortality) and development, with egg and larval stages the most
vulnerable.
¦ Carcinogenic Effects - A number of studies of bottom-dwelling fish, including
tomcod, English sole, Pacific staghorn sculpin, rock sole, brown bullhead, and
winter flounder indicate a link between sediment exposures and hepatic
neoplasms.
Sediment Threshold for Total PAH
¦ Most relevant threshold (10 mg/kg PAH) - Based on brown bullhead effects
observed over changing sediment PAH levels at an industrial site on the Black
River, Ohio.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-46
-------
1 which there is very high confidence in a conclusion of no effect, but it is likely overprotective,
2 especially for non-benthic Housatonic River species.
3 In addition to the threshold for total PAHs, thresholds of 0.92, 0.68, and 0.64 mg/kg were
4 identified for individual PAHs, phenanthrene, benzo(a)anthracene, and indeno(l,2,3-cd)pyrene,
5 respectively; these values were based on the no observed effect concentrations (NOECs) from
6 EPA (2000). Refer to Attachment F.5 for additional information on the derivation procedure.
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-47
-------
1 5.4 RISK CHARACTERIZATION
2 5.4.1 Introduction
3 The risk characterization for fish integrates the exposure assessment (Section 5.2) and effects
4 assessment (Section 5.3) to evaluate the assessment endpoint of survival, growth, and
5 reproduction of fish in the Housatonic River.
6 The following three lines of evidence were used to develop the risk characterization in the
7 Housatonic River fish risk assessment (Figure 5.1-4):
8 ¦ Field surveys - Two field surveys were conducted in the study area. EPA evaluated
9 fish abundance/biomass and conducted an ecological characterization of the site. GE
10 independently evaluated largemouth bass reproduction, community, and habitat data.
11 Interpretation of these studies was constrained by the data limitations; therefore, only
12 a qualitative analysis was performed.
13 ¦ Comparison of field-measured exposures to effects levels or thresholds - For
14 these endpoints, the risk characterization integrated exposure and effects by relating
15 the two terms quantitatively. This method consisted of a comparison of tissue
16 chemistry (PCBs and TEQ) to tissue effects thresholds, and sediment chemistry
17 (PAHs) to literature-based sediment effects thresholds. Hazard quotients were
18 calculated for PCBs by comparing observed tissue concentration data to site-specific
19 MATCs. Probabilities of exceeding various effects threshold levels were also
20 calculated.
21 ¦ Site-specific toxicity study results - These endpoints (e.g., Phase I and Phase II
22 toxicity tests) directly evaluated biological responses to COCs.
23 These lines of evidence allowed for a robust weight-of evidence assessment of the potential for
24 risk using the approach of Menzie et al. (1996).
25 Two of the three lines of evidence listed above suggest some degree of harm to fish in the
26 Housatonic River. Although the field surveys suggest that PCBs and/or other COCs are not
27 causing catastrophic effects to fish reproduction and community structure, they were
28 inconclusive with respect to evaluating more subtle potential alterations to fish community health
29 and reproductive capacity.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-48
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
5.4.2 Field Surveys
5.4.2.1 EPA Fish Community Study
A survey of fish biomass in the mainstem and Woods Pond (Reaches 5 and 6) was conducted by
EPA in fall 2001 to generate information for use in the modeling study and risk assessments
(Woodlot 2002).
Biomass estimates were developed for each species for both prey-sized fish (fish <10 cm) and all
other size classes pooled (fish >10 cm). Separate estimates were developed for each species/size
class group in each of five reaches (i.e., Reaches 5A, 5B, 5C, 5D, and 6). Separate age-class
biomass estimates, however, were generated for largemouth bass in each reach. Survey methods
were based on existing population estimation protocols (Zippin 1958, Ricker 1975, Mitro and
Zale 2000).
Electrofishing sampling was conducted during 21 to 25 August 2000 and 23 to 25 October 2000.
Eleven sample locations were randomly chosen and surveyed within each reach. One location
per reach was surveyed using multi-pass, or depletion, sampling; and the remaining 10 sites were
surveyed using single-pass sampling (i.e., one pass of the electrofishing boat). Survey areas
included the entire channel width and were generally about 200 m in length. Based on previous
electrofishing surveys in each reach, this length was expected to provide a suitable number of
fish for the study (Woodlot 2002).
An ecological characterization of the Housatonic River PSA (Appendix A.l) was also conducted
to describe the aquatic and ecological habitats throughout the study area. Four separate fish
collection events occurred within the PSA during 1998-2000. The principal method employed to
collect fish was electro-shocking fish (electrofishing) from one or two boats operated by the
United States Fish and Wildlife Service (USFWS 1999). In September and October 1998,
electrofishing was conducted to collect fish community characterization data and fish tissue.
Timed (30-minute) surveys to collect community composition data were conducted between
river miles 3 and 4 and between river miles 8 and 11. During each timed event, the total number
of all fish per species observed was estimated and recorded. In addition, target species within
different taxonomic fish groups (e.g., largemouth bass, yellow perch, brown bullhead, common
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r a q
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
carp) were collected for tissue analysis. Collections occurred along river miles 3 and 7 to 11,
and at Woods Pond. Each fish was weighed and measured prior to processing for analysis. A
sample of otoliths and scales was collected from largemouth bass to estimate ages of specimens
(USFWS 1999).
Results of the EPA biomass study are presented in Woodlot (2002) and summarized in Table
F.4-1. Chadwick (1993, 1994) also conducted an earlier biomass study that generally yielded
lower biomass estimates (Table F.4-2).
Fish captures in the Woodlot (2002) study totaled 7,064 individuals and included 17 species, 2
hybrids (bluegill-pumpkinseed and chain-redfin pickerel), and 1 group of taxa (cyprinids). The
most common predators were largemouth bass, yellow perch, and northern pike. Common
forage fish were bluegill, rock bass, pumpkinseed, and cyprinids, whereas the most abundant
bottom feeders were white sucker, common carp, brown bullhead, and goldfish. Other species
captured included smallmouth bass, chain pickerel, redfin pickerel, brown trout, rainbow trout,
black crappie, yellow bullhead (a single individual), and the hybrids mentioned above. Woodlot
(2002; Attachment C) provides the capture data in a standardized biomass (g/m2) format.
Results of the fish collection events associated with the Ecological Characterization Study are
provided in Appendix A.l. White suckers were clearly the dominant fish species in Reaches 5A
and 5B. They still represented the greatest component of the sample biomass in Reach 5C, but
they declined to a smaller component of the fish community in the backwaters and Woods Pond.
In Reach 5C, as well as in the backwaters and Woods Pond, carp become a common member of
the bottom-feeding guild. Goldfish and brown bullhead also represent significant portions of the
bottom-feeding guild in the backwaters and Woods Pond. Bluegills, pumpkinseed, cyprinids,
and rock bass share dominance of the forage fish group, which comprised 11 to 24% of the
overall fish community (based on biomass), in Reaches 5A-5C and the backwaters. Bluegills,
however, were abundant in Woods Pond, where they represented 30% of the total biomass
sample and forage fish, as a group, comprised 40% of the overall fish community. Largemouth
bass and yellow perch were the predominant predators in all reaches.
Based on these and other biological surveys, it is clear that the five representative fish species
chosen for this assessment (i.e., largemouth bass, pumpkinseed, yellow perch, white sucker,
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r rrx
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
brown bullhead) are found in suitable habitats of the PSA. There is also evidence that these
species are self-sustaining; therefore, total reproductive failure or catastrophic effects to the fish
populations are not apparent. A pattern of fish assemblages was apparent from the biological
surveys; this pattern appears related to major changes in habitat across the PSA. Specifically, the
downstream areas of the PSA (Reach 5C, 5D, and Woods Pond) contained a greater proportion
of fish that (a) have a soft-sediment-based feeding strategy, or (b) exploit epifauna on the
macrophyte beds, which are more abundant downstream. Within the forage fish category, the
upstream coarser-grained areas had a higher biomass of cyprinids (e.g., fallfish, shiners),
whereas the downstream areas had a higher biomass of centrarchid sunfish (e.g., bluegill,
pumpkinseed). These differences in species assemblages appear to be mainly related to habitat
differences among reaches.
Chadwick (1993, 1994) also conducted an earlier biomass study that generally yielded lower
biomass estimates. Based on these and other field surveys, it is clear that the five representative
fish species chosen for this assessment (i.e., largemouth bass, pumpkinseed, yellow perch, white
sucker, and brown bullhead) are found within suitable habitats of the PSA. There is also
evidence that these species are self-sustaining; therefore, total reproductive failure or
catastrophic effects to the fish populations are not apparent.
Because of the variability in habitat across the PSA, small-scale variability in PCB
concentrations, and the small gradient in PCB tissue concentrations across the PSA, it is difficult
to discriminate habitat influences from potential contaminant influences. Therefore, meaningful
statistical assessment of PCB or TEQ relationship to fish community parameters was not
feasible. In summary, no quantitative conclusions can be drawn regarding the health of the
Housatonic River fish community based on the field surveys conducted to date.
5.4.2.2 GE Largemouth Bass Community and Reproduction Study
A field study was conducted in the summer and early fall of 2000 and 2001 to assess largemouth
bass habitat, community structure, and reproduction in the Housatonic River (R2 Resource
Consultants Inc. 2002). The assessment of these metrics was made independent of contaminant
concentrations in the environment and fish tissue; therefore no assessment of concentration-
response was attempted. The study area extended from the confluence of the East and West
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r r ->
-------
1 branches downstream to the Woods Pond Dam, and included major stream branches and
2 tributaries.
3 The methods applied in the GE study included the following:
4 ¦ Nest surveys
5 ¦ Habitat surveys
6 ¦ Community estimates
7
8 The study confirmed that there is a population of largemouth bass in the Housatonic River.
9 However, the extent of tributary recruitment remains unknown, and more important, the question
10 of whether PCBs are causing or may cause effects to population strength or viability is largely
11 unaddressed.
12 Reproduction was confirmed to occur in the study area, although the data suggest that
13 reproductive success is lower than in other systems. Growth also appears lower than in other
14 systems. The current population (biomass), which is not exploited due to a consumption
15 advisory that in effect creates a catch-and-release fishery, is dominated by older, larger fish in
16 good condition. Although the population appears stable, this is likely due to the lack of typical
17 harvesting pressure, which places very little demand on recruitment in order to sustain overall
18 numbers.
19 Summaries of the three major components of the GE study are provided below. Additional
20 details are provided in Appendix F (Section F.4.1.2) and in R2 Resource Consultants Inc. (2002).
21 5.4.2.2.1 Nest Surveys
22 In the spring and summer of 2000 and 2001, nest surveys were conducted to determine if
23 largemouth bass in the Housatonic River were successfully reproducing and to assess the
24 condition of young-of-year (YOY) bass. Data collected included nest condition, egg presence
25 and condition, and identification/enumeration of fry.
26 The nest survey data do not provide strong evidence of a high degree of reproductive success. Of
27 the 77 nests observed, only 13 (16.9%) contained eggs described as being in "good" condition;
28 furthermore, only 17 (22%) produced sac fry or swim-up fry. Comparisons to literature
29 information suggest relatively poor spawning success at the sites examined in the Housatonic.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r rrs
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Length data for YOY suggest that YOY from the Housatonic are somewhat smaller than YOY
largemouth bass sampled at other locations at similar latitudes. Overall, although the data
provide an indication that reproduction is occurring, the data do not provide a measure of
reproductive rates. Comparisons of nesting success and growth for YOY suggest that
reproduction may not be similar to other systems. Ultimately, the apparent self-sustaining nature
of this population may be more a function of the low mortality rate of the adults rather than high
reproductive output.
5.4.2.2.2 Habitat Surveys
In the spring of 2000, habitat surveys were conducted at a total of 13 locations in the main river
channel, backwater areas, three major branches, and six tributaries. In 2001, surveys were
conducted at 15 locations in the backwater areas. During the surveys, habitat characteristics
(e.g., stream gradient, substrate), water chemistry (pH, DO, conductivity, temperature), stream
velocity, and other physical attributes were measured.
The results indicate that in the main channel largemouth bass habitat is good and that the
tributaries generally have poor habitat, with the exception of Moorewood and Yokun Brooks.
The study concludes that the fish community is not controlled by fluctuations in temperatures
and water levels, although nesting and spawning success may be sensitive to these factors. The
incidence of dead or fungus-affected eggs in the nests was also attributed to conventional water
quality variables. The investigators did not speculate about the influence of other potential
factors, including contaminants.
5.4.2.2.3 Community Estimates
In June or late July/early August 2000, electrofishing surveys were conducted in the main
channel, backwater areas, and the East and West Branches. In 2001, sampling was conducted
only in main channel and backwater sites located between the confluence of the East and West
Branches. Data collected included lengths/weights, fish aging, and select community metrics
including the proportional stock density (PSD; a measure of the dominance of large fish in the
population) and relative weight (a measure of condition of the fish).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-53
-------
1 Generally, largemouth bass were found throughout the sites sampled, except for selected
2 tributaries. No sampling was done on Moorehead and Yokun Brooks, even though they had been
3 identified as the two tributaries that offered suitable habitat. Weight and length data were used
4 to assess the condition of largemouth bass; these data demonstrated that bass were robust and in
5 good condition, relative to other systems.
6 The age analyses indicated that the bass population consists primarily of large older fish. These
7 analyses also indicated that largemouth bass in the Housatonic River grow at a slower rate as
8 they age. The high proportion of older fish is a function of the lack of harvesting pressure on the
9 community. Comparison of the age structure of the largemouth bass community and/or
10 condition of individual fish to other sites is not a reliable measure of effects.
11 5.4.3 Comparison of Estimated Exposures to Derived Effects Metrics
12 For the representative fish species (largemouth bass, pumpkinseed, yellow perch, white sucker,
13 brown bullhead), risks were assessed separately for three different river segments (PSA, Reaches
14 7 and 8, and Reach 9 and below).
15 5.4.3.1 Total PCBs
16
17
18
19
20
21
22
23 5.4.3.1.1 PSA (Reaches 5 and 6)
24 Overall, the independent lines of evidence exhibit strong concordance in the concentrations of
25 PCBs expected to cause adverse responses to PSA fish. Although there is some variation in the
26 responses across reaches and across species, there is sufficient concordance for the development
27 of threshold levels applicable to all species and reaches. In the risk characterization, exposure
28 concentrations were displayed in relation to multiple effects thresholds (i.e., literature-based and
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r r/\
Summary of Tissue Effects Thresholds for tPCB
¦ 31 mg/kg tPCB - based on literature review (all species).
¦ <45 mg/kg tPCB - based on Phase I study (largemouth bass).
¦ 52 mg/kg tPCB - based on Phase II study (warmwater species and rainbow
trout).
¦ 12 mg/kg tPCB - based on Phase I/Phase II studies, and literature information
on trout sensitivity (coldwater species downstream of the PSA).
-------
1 site-specific reproductive study based). These multiple thresholds also depict some of the
2 uncertainty inherent in the effects thresholds.
3 Figure 5.4-1 depicts hazard quotients for PSA fish tissue concentrations compared to the average
4 of the site-specific (Phase I and Phase II) fish effects thresholds derived for the PSA (i.e., 49
5 mg/kg tPCB). All mean HQs are below 3 and median HQs are below 2, indicative of an
6 ecologically significant but low magnitude of risk.
7 Figures 5.4-2 through 5.4-6 show the cumulative distribution plots for observed whole body
8 tPCB concentrations for each species and reach within the PSA. The vertical lines represent
9 effects concentrations from the literature (31 mg/kg tPCB), Phase II toxicity (52 mg/kg tPCB),
10 and site-specific toxicity studies (<45 mg/kg tPCB). Table 5.4-1 displays the 95th percentile and
11 exceedance probabilities for the threshold effects concentrations by species and PSA reach for
12 observed fish tissue concentrations.
13 Within the PSA, Table 5.4-1 indicates a moderate to high probability of exceedance for most
14 representative species and reaches. However, Figures 5.4-2 through 5.4-6 indicate that the
15 magnitude of exceedance is low to moderate, depending on the effects threshold adopted.
16 5.4.3.1.2 Downstream of PSA (Warmwater Fish)
17 Table 5.4-2 displays the exceedance probabilities for the literature-based and site-specific effects
18 concentrations, as well as the 95th percentile tPCB whole body tissue concentration for each
19 species collected in Reach 8 during the EPA 1998-99 sampling. Although there were limited
20 data for this portion of the river, significant risk to representative fish species (e.g., largemouth
21 bass and yellow perch) is predicted from tPCBs downstream of Woods Pond (Reach 6) to Rising
22 Pond (Reach 8). However, predicted risks are low for pumpkinseed and brown bullhead.
23 The analysis of the Stewart (1982) data, which required extrapolation from fillet concentrations
24 to whole body concentrations, generally supports the evaluation of the EPA data. Specifically,
25 there was a low probability of exceedance of the site-specific thresholds of 45 and 52 mg/kg
26 tPCB, and a moderate probability of exceedance for the literature-based threshold of 31 mg/kg.
27 The data also indicate a reduction in risk with distance downstream from the PSA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-55
-------
100
10
73
•~
es
N
es
X
0.1
0.01
T3
C3
O
jg
*3
m
m
&
c3
2
m
m
2
C/2
T3
o
o
!
&
U
&
43
a
o
Oh
U
Qh
c3
m
"3 i
o
a 1
o
BP
c3
1
2
MC75th and MC25th represent quartiles (i.e., 25th and 75th percentiles)
WB = whole body individuals; WB-R = whole body reconstituted; CM = multiple whole body fish composites
03
m
'S
"3 I
o
§1
i1
c3
m
£
"3
o
§
i1
h-1
o
4
5
Figure 5.4-1
Hazard Quotients for tPCBs in PSA Fish Based on Comparison to the Mean Site-Specific Fish
Toxicity Threshold (49 mg/kg tPCB) (Logarithmic Scale)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-56
-------
Reach 5BC
Reach 6
1
2
3
4
5
6
9
10
.Q
CO
.a
o
8
C
CD
"O
o
o
LU
GO
o
CO
o
o
d
.a
cc
.o
o
c
as
T3
0)
a)
>3
LU
oo
o
CO
o
o
CM
o
o
o
0.1
1
10
100 1000
Total PCBs (mg/kg)
0.1 1 10 100 1000
Total PCBs (mg/kg)
Fish effects thresholds (vertical bars):
31 mg/kg tPCB - Literature based threshold protective of PSA species
<45 mg/kg tPCB - Phase I toxicity study threshold (unbounded)
52 mg/kg tPCB - Phase II toxicity study effects threshold for warmwater species and rainbow trout
Figure 5.4-2
Complementary Cumulative Distribution Plot for tPCB
Concentrations in Whole Body Tissue Compared to Effects
Concentrations - Brown Bullhead
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-57
-------
Reach 5A
Reach 5BC
1
2
3
4
5
6
7
8
9
10
_Q
CO
.Q
O
1
LU
00
o
CO
o
d
CN
o
o
o
0.1 1 10 100 1000
Total PCBs (mg/kg)
0.1 1 10 100 1000
Total PCBs (mg/kg)
Fish effects thresholds (vertical bars):
¦ 31 mg/kg tPCB - Literature based threshold protective of PSA species
¦ <45 mg/kg tPCB - Phase I toxicity study threshold (unbounded)
¦ 52 mg/kg tPCB - Phase II toxicity study effects threshold for warmwater species and rainbow trout
Figure 5.4-3
Complementary Cumulative Distribution Plot for tPCB
Concentrations in Whole Body Tissue Compared to Effects
Concentrations - Largemouth Bass
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-58
-------
Reach 5BC
Reach 6
.o
CO
.Q
O
CD
O
c
C0
~o
o
0
o
X
LU
GO
o
CO
o
o
CM
o
o
o
0.1 1 10 100 1000
Total PCBs (mg/kg)
_Q
CO
_Q
O
CD
O
c
CO
"O
(D
-------
Reach 5A
Reach 5BC
.o
CO
.Q
O
CD
O
c
CO
~o
o
0
o
X
LU
GO
o
CO
o
¦<3-
o
CM
o
o
o
0.1
10
100 1000
_Q
CO
_q
o
CD
O
c
CO
"O
(D
-------
Reach 5A
Reach 5BC
O
o
T_
O
Q
T"
00
8
00
!5 d
Q
CO
A
CO
.Q
X
_Q
O CO
Q
O CO
CL O
Q
ol O
CD
Q
CD
O
Q
o
C ¦'t
Q
C ¦'t
¦§ o
g
TO d
CD
CD
O
X
LU
00
O
CD
O
o
CM
o
o
o
0.1 1 10 100 1000
Total PCBs (mg/kg)
3
4
5
6
7
Fish effects thresholds (vertical bars):
¦ 31 mg/kg tPCB - Literature based threshold protective of PSA species
¦ <45 mg/kg tPCB - Phase I toxicity study threshold (unbounded)
¦ 52 mg/kg tPCB - Phase II toxicity study effects threshold for warmwater species and rainbow trout
Figure 5.4-6
Complementary Cumulative Distribution Plot for tPCB
Concentrations in Whole Body Tissue Compared to Effects
Concentrations - Yellow Perch
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-61
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 5.4-1
Probabilities of Exceedances in the PSA for tPCBs
Species
Reach 5A
Reach 5BC
Reach 6
95th
Percentile
(mg/kg)
Probability of Exceeding
Thresholds
95th
Percentile
(mg/kg)
Probability of Exceeding
Thresholds
95th
Percentile
(mg/kg)
Probability of Exceeding
Thresholds
31 mg/kg
45 mg/kg
52 mg/kg
31 mg/kg
45 mg/kg
52 mg/kg
31 mg/kg
45 mg/kg
52 mg/kg
BB
NAa
NA
NA
NA
73
47%
21%
21%
75
60%
28%
24%
LB
213
50%
50%
50%
209
80%
56%
56%
193
81%
58%
52%
PS
NA
NA
NA
NA
73
68%
36%
28%
59
56%
12%
12%
WS
170
88%
81%
75%
170
77%
62%
58%
169
87%
60%
53%
YP
228
96%
96%
96%
114
100%
100%
80%
123
84%
84%
80%
NA = not available
BB
Brown Bullhead
LB
Largemouth Bass
PS
Pumpkinseed
WS
White Sucker
YP
Yellow Perch
Fish effects thresholds (vertical bars):
¦ 31 mg/kg tPCB - Literature based threshold protective of PSA species
¦ <45 mg/kg tPCB - Phase I toxicity study threshold (unbounded)
¦ 52 mg/kg tPCB - Phase II toxicity study effects threshold for warmwater species and rainbow trout
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Table 5.4-2
Probabilities of Exceedances in Reach 8 for tPCBs and TEQ,
Based on EPA Sampling
tPCBs
Species
95th
Percentile
(mg/kg)
Probability of Exceeding
Thresholds
31 mg/kg
45 mg/kg
52 mg/kg
Observed
BB
9.94
0%
0%
0%
LB
604
32%
10%
10%
PS
23.7
0%
0%
0%
YP
130
67%
17%
17%
TEQ
Species
95th
Percentile
(ng/kg)
Probability of Exceeding
Thresholds
38 ng/kg
45 ng/kg
50 ng/kg
0 Substitution
BB
16.4
0%
0%
0%
LB
123
48%
43%
33%
PS
29
0%
0%
0%
YP
213
100%
67%
67%
DL Substitution
BB
28.2
0%
0%
0%
LB
131
71%
57%
43%
PS
44.5
15%
8%
0%
YP
222
100%
83%
67%
BB Brown Bullhead; LB Largemouth Bass; PS Pumpkinseed; YP Yellow Perch
Fish effects thresholds:
¦ 45 ng/kg TEQ - Phase II toxicity study threshold for warmwater species and rainbow trout
¦ <38 ng/kg TEQ - Phase I toxicity study threshold (unbounded)
¦ 50 ng/kg TEQ - Literature based threshold protective of PSA species
¦ 31 mg/kg tPCB - Literature based threshold protective of PS A species
¦ 52 mg/kg tPCB - Phase II toxicity study threshold for warmwater species and rainbow trout
¦ <45 mg/kg tPCB - Phase I toxicity study threshold (unbounded)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r 7/11/2003
-------
1 There were limited data below Reach 8 (i.e., below Rising Pond), but the data suggest negligible
2 to low risk from tPCBs to warmwater fish species.
3 5.4.3.2 2,3,7,8-TCDD TEQ
4
5
6
7
8 5.4.3.2.1 PSA
9 Figure 5.4-7 depicts hazard quotients for PSA fish tissue concentrations compared to the average
10 site-specific effects threshold (i.e., 42 ng/kg TEQ). All 75th percentile-based HQs exceed 1, but
11 mean and median HQs for adult fish are below 3 for all species. These HQs are indicative of
12 ecologically significant but low magnitude risk.
13 Figure 5.4-8 shows the cumulative distribution plots for observed whole body TEQ tissue
14 concentrations for each species (using DL substitution for ND congeners). The vertical lines
15 represent the effects concentrations from the literature (50 ng/kg) and site-specific toxicity
16 studies (45, 38 ng/kg). Table 5.4-3 displays the exceedance probabilities for the effects
17 concentrations, as well as the 95th percentile for exposure for each species, substituting
18 concentrations equal to zero or to the DL, respectively, for non-detected congeners. There was a
19 moderate to high probability of exceedance for most fish species for all three effects thresholds.
20 5.4.3.3 PAHs
21 Table F.2-18 displays summary statistics for sediment chemistry concentrations for the eight
22 individual PAHs identified as COCs. For the three individual PAHs for which thresholds could
23 be determined, the median sediment concentration in each river reach was below the toxicity
24 threshold, indicating negligible to low risk to fish from these contaminants. The median
25 sediment concentrations for total PAH ranged from 3 to 8 mg/kg in the PSA, below the "most
26 relevant" effects concentration of 10 mg/kg tPAH for fish.
Summary of Effects Thresholds for TEQ
¦ 50 ng/kg TEQ - based on literature review (relevant to PSA fish species)
¦ <38 ng/kg TEQ - based on Phase I study (largemouth bass)
¦ 45 ng/kg TEQ - based on Phase II study (warmwater species and rainbow trout)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-64
-------
10
Max
MC 75th
i-L-i
Mean
X
Median
—
MC 25th
Min
T
1 -¦
¦o
~-
N
a
41
0.1
-o
cd
.3
I
fS
oi
pi)
-o
VI
• S Q
f\
o, O
S w
s
Plh
-a
Plh
£
_o
"S
>H
Pi
m
-a
H
c3
PQ
o
§
I?
J
P^
PQ
c3
PQ
o
§
I?
J
PQ
c3
PQ
*s
=3
o
§
I?
J
o
-o
cd
a
m ®
PQ
WB-R = Whole body,
reconstituted fish tissue. WB = Whole bodyfish tissue.
2 MC75th/MC25th are quartiles (i.e., 25th/75th percentiles); WB = whole individuals; WB-R = whole body reconstituted; CM = multiple whole fish composites
3 Figure 5.4-7 Hazard Quotients for TEQ for Fish in Primary Study Area (PSA) Based on Comparison to the
4 Average Site-Specific Tissue Effects Threshold (42 ng/kg TEQ) (Logarithmic Scale)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-65
-------
1
2
3
4
5
6
7
8
9
10
11
38 45 50
oo
d
-O
as
_Q
o
q
a
£
(0
"O
0
0
o
X
LU
CO
d
CM
d
o
d
0
50 100 150
TEQ Concentration (ng/kg)
200
Key: BB = Brown Bullhead; LB = Largemouth Bass; PS = Pumpkinseed; YP = Yellow Perch
Fish effects thresholds (vertical bars):
¦ 45 ng/kg TEQ - Phase II toxicity study threshold for wannwater species.
¦ <38 ng/kg TEQ - Phase I toxicity study threshold (unbounded).
¦ 50 ng/kg TEQ - Literature based threshold protective of coldwater and wannwater species.
Figure 5.4-8
Complementary Cumulative Distribution Plot for TEQ
Concentrations in Whole Body Tissue Compared to Effects
Concentrations for All Species Using DL Substitution for Non-
Detects
MK0110:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-66
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Table 5.4-3
Probabilities of Exceedances for TEQ
Non-Detect
Method
Species
95th
Percentile
(ng/kg)
Probability of Exceeding
Thresholds
38 ng/kg
45 ng/kg
50 ng/kg
0 Substitution
BB
125
75%
66%
56%
LB
117
61%
55%
41%
PS
58.9
26%
13%
10%
YP
151
96%
87%
87%
DL Substitution
BB
130
88%
78%
72%
LB
126
77%
64%
57%
PS
66
58%
32%
29%
YP
163
98%
98%
96%
BB Brown Bullhead
LB Largemouth Bass
PS Pumpkinseed
YP Yellow Perch
0 substitution = non-detectable concentrations were substituted with zeroes
DL substitution = non-detectable concentrations were substituted with method detection limit
Fish effects thresholds:
¦ 45 ng/kg TEQ - Phase II toxicity study threshold for warmwater species
¦ <38 ng/kg TEQ - Phase I toxicity study threshold (unbounded)
¦ 50 ng/kg TEQ - Literature based threshold protective of coldwater and warmwater species
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-67
-------
1 5.4.4 Site-Specific Toxicity Studies
2 Results and interpretation of the site-specific toxicity studies are summarized in Sections 5.3.2
3 and 5.3.3 and detailed in Appendix F; therefore only a synopsis is provided here.
4 The fish toxicity studies indicate that PCBs are acting upon early life stages of fish, and causing
5 various reproductive and developmental responses. The types of malformations and other effects
6 observed are suggestive of an Ah-receptor (i.e., dioxin-like) etiology. However, the threshold
7 effect levels (as identified in both the literature review and the site-specific studies) each have a
8 moderately high uncertainty. For example, individual ED50 values from Phase II site extracts
9 data span the range of PCB concentrations found in PSA fish.
10 5.4.5 Weight-of-Evidence Analysis
11 A weight-of-evidence evaluation was conducted for the multiple measurement endpoints in the
12 fish ERA to determine whether significant risk is posed to fish from tPCBs. The three-phase
13 approach of Menzie et al. (1996) and the Massachusetts Weight-of-Evidence Workgroup was
14 used. The weight-of-evidence approach involves: (a) assigning weights to each measurement
15 endpoint, (b) determining the magnitude of response observed in the measurement endpoint, and
16 (c) determining the concurrence among measurement endpoints.
17 The attributes considered in the weight of evidence are described in Section 2, and the rationale
18 for weighting of the measurement endpoints are provided in Appendix F. A summary of the
19 derived weightings for each attribute is provided in Table 5.4-4. The Phase II study yielded the
20 highest overall rating because of the site specificity of the study and the connection to the
21 exposure pathway of greatest interest (maternal transfer).
22 The magnitude of the response in the measurement endpoint is considered together with the
23 measurement endpoint weight in judging the overall weight-of-evidence (Menzie et al 1996).
24 This requires assessing the strength of evidence of ecological risk, as well as an indication of the
25 magnitude of the response, if present. The weighting values, evidence of risk, and magnitude of
26 responses were combined in a matrix format and are presented in Table 5.4-5.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-68
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Table 5.4-4
Weighting of Measurement Endpoints for Fish Weight-of-Evidence Evaluation
Measurement Endpoints:
Endpoint A: Site-Specific Toxicity
Endpoint B: Fish Tissue Chemistry
Endpoint C: Field Surveys
Attributes
Al. Phase I Study
A2. Phase II
Study
Bl. Observed /
Literature Effects
B2. Observed /
Phase I Effects
B3. Observed /
Phase II Effects
CI.
Community
Studies
C2.
Reproduction
Study
I. Relationship Between Measurement and Assessment Endpoints
1. Degree of Association
High
High
Mod
High
High
Low/Mod
Low/Mod
2. Stressor/Response
Mod
Mod
Low/Mod
Mod
Mod
Low
Low
3. Utility of Measure
High
High
Mod/High
Mod/High
Mod/High
Low/Mod
Low/Mod
II. Data Quality
4. Data Quality
High
High
High
High
High
High
High
III. Study Design
5. Site Specificity
High
High
Low/Mod
High
High
High
High
6. Sensitivity
Low/Mod
Mod/High
Low/Mod
Mod
Mod/High
Low
Low
7. Spatial Representativeness
Mod
Mod
Mod/High
Mod/High
Mod/High
Mod
Mod
8. Temporal Representativeness
High
High
High
High
High
Mod/High
Mod/High
9. Quantitative Measure
High
High
Mod
Mod
Mod
Mod
Low
10. Standard Method
High
High
Mod
Mod
Mod
Mod
Mod
Overall Endpoint Value
Mod/High
High
Mod
Mod/High
Mod/High
Low/Mod
Low/Mod
A: Site-Specific Toxicity
A1 - Reproductive success in site-specific toxicity tests, relative to reference condition
A2 - Reproductive success in site-specific toxicity tests, using dose-response analysis
B: Fish Tissue Chemistry
B1 - Observed fish tissue concentrations relative to
B2 - Observed fish tissue concentrations relative to
B3 - Observed fish tissue concentrations relative to
C: Fish Community and Reproduction Studies
CI - EPA and GE Fish Community Studies
C2 - GE Fish Reproduction Study
MK01|O:\20123001.096\ERA_PB\ERA_PB_5.DOC r rq 7/11/2003
literature toxicity threshold
Phase I study toxicity threshold
Phase II study toxicity threshold
-------
Table 5.4-5
Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related
to Maintenance of a Healthy Fish Community
Measurement Endpoints
Weighting
Value
(High,
Moderate, Low)
Evidence of Harm (Yes,
No, Undetermined)
Magnitude (High,
Intermediate, Low)
A. Site-Specific Toxicity
Al. Reproductive success relative to reference
Mod/High
Yes
Low
A2. Reproductive success dose-response
High
Yes
Intermediate
B. Fish Tissue Chemistry
Bl. Observed fish tissue/ Literature toxic levels
Mod
Yes
Low
B2. Observed fish tissue/ Phase I toxic levels
Mod/High
Yes
Low
B3. Observed fish tissue/ Phase II toxic levels
Mod/High
Yes
Low
C: Fish Community and Reproduction Studies
CI: EPA Study and GE Community Study
Low/Mod
Undetermined
-
C2: GE Reproduction Study
Low/Mod
Undetermined
-
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r qrs 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
A graphical method was used for displaying concurrence among measurement endpoints (Table
5.4-6); the method involved plotting the five symbols representing site-specific toxicity (A) and
fish tissue chemistry (B) endpoints in a matrix, with the weight of the measurement endpoint and
the degree of response as axes. This table illustrates that the majority of endpoints indicate, with
a moderately high degree of confidence, that there are low magnitude risks to fish in the PSA.
5.4.6 Sources of Uncertainty
The assessment of risks to fish contains uncertainties. Each source of uncertainty can influence
the estimates of risk; therefore, it is important to describe and, when possible, specify the
magnitude and direction of such uncertainties. Appendix F contains a more complete list of
uncertainties; some of the most significant uncertainties are described below.
¦ Potential for seasonal variation in fish tissue concentrations - The vast majority of the
PCB data for this project were collected in the late summer and early fall (September
- October). Because maternal transfer to juveniles is expected to be the most
important risk pathway for PCBs, changes in lipid content during the spawning period
may affect the concentrations of PCBs delivered to the eggs.
¦ Fillet data - There is some uncertainty with respect to risks for fish for which only
fillet concentration data are available, since this required conversions to estimated
whole body concentrations.
¦ Biological relevance of the injected doses used in Phase II toxicity study - To
develop a concentration-response relationship, it was assumed that the dose delivered
to the egg was proportional to ovary concentration in the parent fish, and that these
ovary concentrations would be related to whole body tissue concentrations on the
basis of equilibrium partitioning to lipids. These assumptions required extrapolations
and carried associated uncertainties.
¦ Lack of synopticity in Phase I toxicity study - Because different fish were evaluated
for tissue chemistry than were assessed for pathologies, the exposure and effects
information for Phase I were not directly synoptic. However, the approach used to
estimate the exposure concentrations was unbiased.
¦ Literature threshold uncertainty - The literature effects thresholds for tPCBs and for
TEQ have a number of uncertainties associated with them, including lack of
information for Housatonic River representative species; limited number of studies
that met the screening criteria used in the literature review; limited number of studies
that replicated the maternal transfer of contaminants to offspring; and variability of
effects thresholds identified.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-71
-------
1 Table 5.4-6
2
3 Risk Analysis for Risk Exposed to tPCBs and TEQ in the Housatonic River PSA
Assessment Endpoint:
Survival, growth, and reproduction of fish
4 ~
Harm/Magnitude
Weighting Factors (increasing confidence of weight)
~
1
1
1
J
Low
Low-Moderate
Moderate
Moderate-High
High
Yes/High
Yes/Intermediate
A2
Yes/Low
B1
A1,B2,B3
Undetermined
C1,C2
1
1
1
t
No/Low
No/High
5
6 A: Site-Specific Toxicity
7 A1 - Reproductive success in site-specific toxicity tests, relative to reference condition
8 A2 - Reproductive success in site-specific toxicity tests, using dose-response analysis
9 B: Fish Tissue Chemistry
10 B1 - Observed fish tissue concentrations relative to literature toxicity threshold
11 B2 - Observed fish tissue concentrations relative to Phase I study toxicity threshold
12 B3 - Observed fish tissue concentrations relative to Phase II study toxicity threshold
13 C: Fish Community and Reproduction Studies
14 CI - EPA and GE Community Studies
15 C2 - GE Reproduction Study
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-72
-------
1 ¦ Toxicity study uncertainties - The concentration-response relationships in the Phase I
2 and Phase II studies were variable, and other organic contaminants (other than PCBs)
3 may have been present that could confound the concentration-response curves for
4 PCBs. In the Phase II studies, effects observed were variable for a given standard or
5 extract among the different species and life stage combinations.
6 ¦ Extrapolation of egg thresholds to whole body burdens - The extrapolation of
7 concentrations of PCBs in egg to whole body concentrations has a degree of
8 associated uncertainty. The lipid content and type were not measured in eggs, which
9 can vary between species and between adults and eggs. PCB accumulation and
10 distribution in fish tissues is strongly influenced by the lipid content and lipid type
11 (polar, neutral, nonpolar); thus accumulation and deposition of PCBs into tissues
12 including developing oocytes can be influenced by the percent and type of these
13 lipids (Monosson 1999).
14 ¦ The exposure component for the field surveys is highly uncertain. The largemouth
15 bass study did not quantify contaminant exposures in either sediment or fish tissue.
16 Where chemistry data are available for relation to other field survey data (e.g., EPA
17 biomass and abundance estimates), the exposure assessment is confounded by
18 significant habitat variation across the PSA.
19 ¦ The assessment of effects in the field survey studies was uncertain because these
20 studies could not detect anything less than very large responses in the local
21 population.
22 5.4.7 Downstream Extrapolation
23 5.4.7.1 Risks to Warmwater Fish Downstream of the PSA
24 As was done for the PSA, risks to warmwater fish were evaluated based on concentrations of
25 tPCBs in fish tissue. A maximum acceptable threshold concentration (MATC) of 49 mg/kg
26 tPCB in tissue (whole-body, wet weight) was developed for the PSA using the average of site-
27 specific (Phase I and Phase II) toxicity effects thresholds, and was also applied to areas
28 downstream of Woods Pond using the available warmwater fish (e.g., bass, perch, sunfish) tissue
29 data.
30 In the case of the fish, each downstream reach (Reaches 7 through 16) was evaluated as a unit,
31 and the mean fish tissue concentration in the reach was compared with the threshold
32 concentration to determine potential risk. Only data collected since 1998 were used in this
33 analysis.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC ^
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Results are provided in Appendix F (Figure F.4-10). The moderate risks observed in the PSA
decline to below levels of concern in Reaches 7 through 9, in the section of the river between
Woods Pond and the Massachusetts/Connecticut state line. Potential risks were also not
indicated in the Connecticut portion of the river.
5.4.7.2 Risks to Coldwater Fish (Trout) Downstream of the PSA
Trout were evaluated separately from PSA fish species because of apparent differences in the
sensitivity of trout to PCBs as indicated by the results of the site-specific fish toxicity studies
(Tillitt et al. 2003b) and the generally higher PCB concentrations in trout due to their increased
lipid content. Although the ED50 values for trout were within a factor of 2 of warmwater species
in the Phase II trials, other indications of toxicity (Tillitt et al. 2003b) suggest that rainbow trout
were slightly more toxic than the warmwater species. Furthermore, the rainbow trout strain
applied in the Phase II testing (Tillitt, personal communication 2003) is less sensitive than other
test strains, and the sensitivity of other downstream trout species (e.g., brown trout) has not been
assessed. Therefore, the PSA effects threshold of 49 mg/kg tPCB was divided by a factor of 4 to
account for potential increased sensitivity of downstream coldwater species (i.e., coldwater
MATC of 12 mg/kg tPCB whole body, wet weight). Because of the more limited database for
trout, a number of extrapolations were necessary to convert available warmwater fish data and/or
trout fillet data to estimated whole body concentrations for trout. These extrapolations are
summarized in Appendix F (Section F.4.6.2).
Results are provided in Appendix F (Figure F.4-11). In general, some potential risk to trout from
PCBs was found in river reaches from Woods Pond Dam down to and including Reach 9. These
risks were marginal, and are uncertain due to incertitude about the sensitivity differences for
various trout species. Potential risk to trout was not evaluated downstream of Reach 12 due to
lack of suitable trout habitat.
5.4.8 Risk Assessment Conclusions
Overall, evaluation of the fish assessment endpoint suggests ecologically significant but low
magnitude risk to fish in the Housatonic River from both tPCBs and PCB TEQ, based on a
weight-of-evidence evaluation of multiple endpoints. Other COCs, such as PAHs and metals,
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC r
-------
1 were not present in the PSA at concentrations expected to cause pronounced effects, although
2 marginal PAH toxicity could not be conclusively ruled out.
3 The confidence in the numerical effects thresholds that support this conclusion is moderate.
4 Strength in the conclusions was derived from the concordance in predictions of risk from
5 multiple measurement endpoints; however, there is some uncertainty associated with several of
6 the endpoints. Because the effects thresholds derived in this study span a similar range as the
7 observed fish tissue concentrations in the PSA, the uncertainty inherent in the threshold
8 derivation has large implications for the prediction of actual risks to the local fish population.
9 Use of a lower-bound threshold results in a prediction of significant adverse effects for the vast
10 majority of species in all PSA reaches. However, use of higher-end thresholds would lead to a
11 conclusion of low risks for the same fish. Catastrophic risks, such as total reproductive failure or
12 widespread direct mortality of adults, are not predicted for any species because the magnitude of
13 exceedance of conservative effects thresholds is generally marginal to moderate (i.e., within a
14 factor of 5).
15 5.5 REFERENCES
16 ASTM (American Society for Testing and Materials). 2002. Standard guide for conducting early
17 life-stage toxicity tests with fishes. In: Annual Book of ASTM Standards, E 1241-98, West
18 Conshohocken, PA.
19 BBL (Blasland, Bouck & Lee, Inc.) and QEA (Quantitative Environmental Analysis, LLC).
20 2003. Housatonic River - Rest of River RCRA Facility Investigation Report. Prepared for
21 General Electric Company. January 2003.
22 Berlin, W.H., R.J. Hesselberg, and M.J. Mac. 1981. Growth and mortality of fry of Lake
23 Michigan lake trout during chronic exposure to PCBs and DDE. Tech. Pap. U.S. Fish Wildl. Ser.
24 105:11-22.
25 Chadwick (Chadwick & Associates). 1993. Fisheries Investigation of the Housatonic River,
26 Massachusetts. Prepared for General Electric Company, Pittsfield, MA.
27 Chadwick (Chadwick & Associates). 1994. Aquatic Ecology Assessment of the Housatonic
28 River, Massachusetts. Prepared for General Electric Company, Pittsfield, MA.
29 Coles, J.F. 1996. Organochlorine Compounds and Trace Elements in Fish Tissue and Ancillary
30 Data for the Connecticut, Housatonic, and Thames River Basin Study Unit, 1992-1994. U.S.
31 Geological Survey Open-File Report 96-358, 26 pp.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
-------
1 EPA (United States Environmental Protection Agency). 1999. Development of a tissue trigger
2 level from bioaccumulated tributyltin in marine benthic organisms: West Waterway, Harbor
3 Island Superfund Site, Seattle, WA. U.S. Environmental Protection Agency, Region 10, Seattle,
4 WA.
5 EPA (United States Environmental Protection Agency). 2000. Health of bullhead
6 in an urban fishery after remedial dredging. United States Environmental
7 Protection Agency, Great Lakes National Program Office. Chicago, Illinois. URL:
8 http://www.epa. gov/ grtlakes/sediments/Bullhead/report. html.
9 Johnson, L, T. Collier, and J.E. Stein. 2002. An analysis in support of sediment quality
10 thresholds for poly cyclic aromatic hydrocarbons (PAHs) to protect estuarine fish. Aquatic
11 Conservation: Marine and Freshwater Ecosystems 12:518-538.
12 Johnson, R.D., J.E. Tietge, K.M. Jensen, J.D. Fernandez, A.L. Linnum, D.B. Lothenbach, G.W.
13 Holcombe, P.M. Cook, S.A. Christ, D.L. Lattier, and D.A. Gordon. 1998. Toxicity of 2,3,7,8-
14 tetrachlorodibenzo-p-dioxin to early life stage brook trout (Salvelinus fontinalis) following
15 parental dietary exposure. Environ. Toxicol. Chem. 17(12):2408-2421.
16 Malins, D.C., B.B. McCain, D.W. Brown, S. Chan, M.S. Myers, J.T. Landahl, P.G. Prohaska,
17 A.J. Friedman, L.D. Rhodes, D.G. Burrows, W.D. Gronlund, and H.O. Hodgins. 1984. Chemical
18 pollutants in sediments and diseases of bottom-dwelling fish in Puget Sound, Washington.
19 Environmental Science and Technology 18(9):705-713.
20 Malins, D.C., M.M. Krahn, M.S. Myers, L.D. Rhodes, D.W. Brown, C.A. Krone, B.B. McCain
21 and Sin-Lam Chan. 1985. Toxic chemicals in sediments and biota from a creosote-polluted
22 harbor: relationships with hepatic neoplasms and other hepatic lesions in English sole
23 (Parophrys vetidus). Carcinogenesis 6(10):1463-1469.
24 Mauck, W.L., P.M. Mehrle, and F.L. Mayer. 1978. Effects of the polychlorinated biphenyl
25 Aroclor® 1254 on growth, survival, and bone development in brook trout (Salvelinus fontinalis)
26 J. Fish. Res. Board Can. 35:1084-1088.
27 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S.
28 Petron, B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-
29 of-Evidence Workgroup: A weight-of-evidence approach for evaluating ecological risks. Human
30 and Ecological Risk Assessment 2:277-304.
31 Mitro, M.G. and A.V. Zale. 2000. Predicting fish abundance using single pass removal
32 sampling. Canadian Journal of Fisheries and Aquatic Sciences. 57: 951-961.
33 Monosson, E. 1999. Reproductive, developmental and immunotoxic effects of PCBs in fish: a
34 summary of laboratory and field studies. Prepared for Damage Assessment Center, National
35 Oceanographic and Atmospheric Administration, Silver Spring, Maryland. NOAA Contract 50-
36 DSNC-7-90032. March 1999.
MK0110:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-76
-------
1 Oliver, B.G. and A.J. Niimi. 1985. Bioconcentration factors of some halogenated organics for
2 rainbow trout: limitations in their use for prediction of environmental residues. Environ. Sci.
3 Technol. 19:842-849.
4 Papoulias, D. 2003a. Personal Communication (e-mail to M. Ptashynski, EVS Consultants,
5 North Vancouver, BC, regarding interpretation of survival and pathology data). Fisheries
6 Biologist, U.S. Geological Survey. Columbia Environmental Research Center. Columbia
7 Missouri. June 16, 2003.
8 Papoulias, D. 2003b. Personal Communication (e-mail to M. Ptashynski, EVS Consultants,
9 North Vancouver, BC, regarding excerpt from final report describing statistical analyses for
10 Phase II Studies). Fisheries Biologist, U.S. Geological Survey. Columbia Environmental
11 Research Center. Columbia Missouri. June 18, 2003.
12 Papoulias, D. 2003c. Personal Communication (e-mail to M. Ptashynski, EVS Consultants,
13 North Vancouver, BC, regarding mortality in largemouth bass control fish). Fisheries Biologist,
14 U.S. Geological Survey. Columbia Environmental Research Center. Columbia Missouri. June
15 16,2003.
16 PSWQAT (Puget Sound Water Quality Action Team). 2000. 2000 Puget Sound Update: Seventh
17 Report of the Puget Sound Ambient Monitoring Program. Puget Sound Water Quality Action
18 Team. Olympia, Washington.
19 Reichert, W.L., B. French, T. Horn, H.R. Sanborn, J.E. Stein. 1996. Chronic exposure to
20 sediment-associated poly cyclic aromatic compounds: toxicokinetics of hepatic DNA adducts and
21 biliary fluorescent aromatic compounds in English Sole. Marine Environmental Research
22 42(1):279.
23 Ricker, W.E. 1975. Computation and Interpretation of Biological Statistics of Fish Populations.
24 Bulletin of the Fisheries Research Board of Canada. 191. 3 82 pp.
25 R2 (R2 Resource Consultants Inc.). 2002. Evaluation of Largemouth Bass Habitat, Population
26 Structure, and Reproduction in the Upper Housatonic River, Massachusetts. Report prepared for
27 General Electric Company.
28 Smith, S.B. and J.F. Coles. 1997. Endocrine Biomarkers, Organochlorine Pesticides, and
29 Congener Specific Polychlorinated Biphenyls (PCBs) in Largemouth Bass (Micropterus
30 salmoides) from Woods Pond, Housatonic River, Massachusetts, September 1994 and May
31 1995. U.S. Geological Survey Administrative Report. Prepared in cooperation with the U.S.
32 Environmental Protection Agency. Reston, VA.
33 Spies, R.B., D.W. Rice, Jr. and J.W. Felton. 1988. The effects of organic contaminants on
34 reproduction of starry flounder, Platichthys stellatus (Pallas) in San Francisco Bay. Part I.
35 Hepatic contamination and mixed-function oxidase (MFO) activity during the reproductive
36 season. Marine Biology 98:181-189.
37 Spies, R.B. and D.W. Rice, Jr. 1988. The effects of organic contaminants on reproduction of
38 starry flounder, Platichthys stellatus (Pallas) in San Francisco Bay. Part II. Reproductive success
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-77
-------
1 of fish captured in San Francisco Bay and spawned in the laboratory. Marine Biology 98:191-
2 202.
3 Stewart Laboratories, Inc. 1982. Housatonic River Study 1980 and 1982, Volumes I and II.
4 Tillitt, D. 2002. Personal communication (e-mail to Alice Shelly, TerraStat Consulting Group,
5 Everett, Washington regarding extrapolation of contaminant concentration from egg to ovary and
6 ovary to whole body). Columbia Environmental Research Center, USGS, Columbia, Missouri
7 July 31.
8 Tillitt, D. 2003. Personal communication (conference call with EVS Environmental Consultants,
9 USGS study team members, Weston Solutions, and US EPA regarding the Phase II studies).
10 USGS, Columbia Environmental Research Center, Columbia, Missouri. June 11, 2003.
11 Tillett, D., D. Papoulias, and D.R. Buckler. 2001. Fish Reproductive Health Assessment in PCB
12 Contaminated Regions of the Housatonic River, Massachusetts, USA: Investigations of Causal
13 Linkages Between PCBs and Fish Health. Prepared for U.S. Fish and Wildlife Service, Concord,
14 New Hampshire and U.S. Environmental Protection Agency, Boston, Massachusetts.
15 Tillett, D., D. Papoulias, and D.R. Buckler. 2002. Fish Reproductive Health Assessment in PCB
16 Contaminated Regions of the Housatonic River, Massachusetts, USA: Investigations of Causal
17 Linkages Between PCBs and Fish Health. Interim Report of Phase II Studies. Prepared for U.S.
18 Fish and Wildlife Service, Concord, New Hampshire and U.S. Environmental Protection
19 Agency, Boston, Massachusetts.
20 U.S. Army Corps of Engineers (USACE). 1988. Environmental effects of dredging technical
21 notes: relationship between PCB tissue residues and reproductive success of fathead minnows.
22 U.S. Army Corps of Engineers, U.S. Army Engineer Waterways Experiment Station, EEDP-01-
23 13.9 pp.
24 Van den Berg, M., et al. 1998. Toxic equivalency factors (TEFs) for PCBs, PCDDs, PCDFs for
25 humans and wildlife. Environmental Health Perspectives. 106(12): 775-792.
26 Walker, M.K. and R.E. Peterson. 1991. Potencies of polychlorinated dibenzo-p-dioxin,
27 dibenzofuran, and biphenyl congeners, relative to 2,3,7,8-tetrachlorodibenzo-p-dioxin, for
28 producing early life stage mortality in rainbow trout (Oncorhynchus mykiss). Aquat. Toxicol.
29 21:219-238.
30 Walker, M.K., P.M. Cook, A.R. Batterman, B.C. Butterworth, C. Berini, J.J. Libal, L.C.
31 Hufnagle and R.E. Peterson. 1994. Translocation of 2,3,7,8-tetrachlorodibenzo-p-dioxin from
32 adult female lake trout (Salvelinus namaycush) to oocytes: effects on early life stage
33 development and sac fry survival. Can. J. Fish. Aquat. Sci. 51:1410-1419.
34 WESTON (Roy F. Weston, Inc). 2000a. Supplemental Investigation Work Plan. Prepared for
35 U.S. Army Corps of Engineers and U.S. Environmental Protection Agency.
36 WESTON (Roy F. Weston, Inc). 2000b. Quality Assurance Project Plan (QAPP). Prepared for
37 U.S. Army Corps of Engineers and U.S. Environmental Protection Agency.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-78
-------
1 Woodlot Alternatives. 2002. Fish Biomass Estimate for Housatonic River Primary Study Area.
2 Prepared for U.S. Army Corps of Engineers. DCN: GE-061202-ABBF.
3 Zabel, E.W., P.M. Cook and R.E. Peterson. 1995a. Toxic equivalency factors of polychlorinated
4 dibenzo-/>dioxin, dibenzofuran and biphenyl congeners based on early life stage mortality in
5 rainbow trout (iOncorhynchus mykiss). Aquat. Toxicol. 31:315-328.
6 Zabel, E.W., P.M. Cook and R.E. Peterson. 1995b. Potency of 3,3',4,4',5-pentachlorobiphenyl
7 (PCB-126), alone and in combination with 2,3, 7,8-tetrachl orodi benzo-/>dioxin (TCDD), to
8 produce lake trout early life-stage mortality. Environ. Toxicol. Chem. 14(12): 2175-2179.
9 Zippin, C. 1958. The removal method of population estimation. Journal of Wildlife Management
10 22:82-90.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_5.DOC
5-79
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
6. WILDLIFE ASSESSMENT HIGHLIGHTS
Highlights
¦ Conceptual model for wildlife indicates diet is the major route of exposure to
tPCBs and TEQ. Therefore, wildlife exposure modeling approach focuses only
on dietary exposure.
¦ Probabilistic methods used to propagate uncertainty through wildlife exposure
model.
¦ Selection of dietary concentration variables required consideration of spatial
and temporal averaging of exposure.
¦ Options available to characterize exposure-effects relationships for tPCBs and
TEQ include developing dose-response curves, NOAELs or LOAELs, field-
based thresholds, or threshold ranges.
¦ Weight-of-evidence approach was used to characterize risks.
6.1 OVERVIEW
The purpose of this section is to describe the general approach and methods used to estimate
risks of contaminants of concern (COCs) to wildlife. Unlike the aquatic endpoints, the same
general approach was appropriate for the majority of the wildlife endpoints. The endpoints for
which this discussion applies include:
¦ Insectivorous Birds (Section 7)
¦ Piscivorous Birds (Section 8)
¦ Piscivorous Mammals (Section 9)
¦ Omnivorous and Carnivorous Mammals (Section 10)
¦ Threatened and Endangered Species (Section 11)
An overview of the following topics is included:
¦ Selection of COCs for wildlife
¦ Development of wildlife conceptual model
¦ Approach to wildlife exposure modeling
¦ Spatial and temporal averaging of exposure
¦ Probabilistic methods for propagating uncertainty
¦ Approach for effects assessment
¦ Approach for risk characterization
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-1
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
This section provides a comprehensive overview of the approaches used in modeling risk for
wildlife to reduce repetition in each of the subsequent sections, but does not provide the technical
details specific to each endpoint. The specific details are discussed in Appendix C and the
wildlife risk assessment appendices (Appendices Gto K).
Representative Wildlife Species
Insectivorous Birds - Tree Swallow and American Robin
Piscivorous Birds - Belted Kingfisher and Osprey
Piscivorous Mammals - Mink and River Otter
Omnivorous and Carnivorous Mammals - Red Fox and Short-Tailed Shrew
Threatened and Endangered Species - Bald Eagle, American Bittern, and Small-
Footed Myotis
Based on the results of the conservative, deterministic Pre-ERA (Section 2.4, Appendix B), the
contaminants of potential concern (COPCs) for wildlife include total polychlorinated biphenyls
(tPCBs), dioxins/furans, and several organochlorine pesticides. These pesticides were
subsequently screened out of the ERA because, in general, they were detected at low frequencies
and low concentrations, and are not considered to be from site-related sources. Further, the
actual concentrations of organochlorine pesticides are likely much lower than the measured
values due to laboratory interference. Dioxin, furan, and coplanar PCB congeners were
considered in the risk assessment by calculating 2,3,7,8-TCDD toxic equivalence (TEQ).
Methods for calculating TEQ concentrations are described in Section 6.4.
The conceptual model for the wildlife assessments is shown in Figure 6.1-1. The conceptual
model outlines the ecosystem processes that qualitatively link stressor releases and primary,
secondary, and tertiary exposure pathways to ecological receptors. Thus, the conceptual model
provides a visual representation of the potential risk pathways to wildlife from COCs. Each
representative species has a species-specific conceptual model. These are presented in Sections
7 to 11 and Appendices G to K.
The wildlife risk assessments have three main components: the exposure assessment, the effects
assessment, and risk characterization. The process used in each of these components is described
below.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-2
-------
1
2
3
4
5
6
7
8
9
0
1
0)
Contaminated soil, sediment, water and biota
£
in
Figure 6.1-1 Conceptual Model for the Assessment of Risks from tPCBs and TEQ
to Wildlife in the Housatonic River Primary Study Area
Figure 6.1-2 depicts the framework for the exposure assessment.
During the exposure assessment, exposure of wildlife to tPCBs and TEQ in the Housatonic River
Primary Study Area (PSA) was determined, beginning with a description of the exposure model.
Input variables for the exposure model were established using life history information for the
representative species, and concentrations of tPCBs and TEQ in prey collected in the PSA. For
those input variables that were uncertain, variable, or both, distributions were used rather than
point estimates. Monte Carlo and probability bounds analyses were then performed to propagate
input variable uncertainties through the exposure model for each COC.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-3
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
EXPOSURE
Figure 6.1-2 Framework Used to Model Exposure of Wildlife Species to
Contaminants of Concern (COCs) in the Housatonic River PSA
Figure 6.1-3 shows the approach used in conducting the effects assessment. The effects
assessment includes a comprehensive review of the literature on the effects of tPCBs and TEQ
on survival, growth, and reproduction of representative wildlife species or reasonable surrogate
species. Each of the studies was evaluated using acceptability criteria established for this ERA
(see box below). Appropriate studies were then selected and used to derive the effects metric.
Acceptability Criteria for Wildlife Studies
Were appropriate controls used?
Were appropriate statistics applied?
Were acceptable methods (e.g., laboratory methods) used?
Was there an appropriate range of exposure doses?
Was the experimental effect attributable to the COC?
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC s a 7/10/2003
-------
EFFECTS
1
2 Figure 6.1-3 Approach Used to Model Effects of Contaminants of Concern (COCs)
3 to Representative Species in the Housatonic River PSA
4 The final component of the wildlife risk assessments is the characterization of risk combining the
5 results of the exposure and effects assessments, and other available lines of evidence (e.g., whole
6 media or in situ studies, field surveys). Figure 6.1-4 presents the general approach used in the
7 risk characterization process.
8
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC r- r 7/10/2003
-------
RISK CHARACTERIZATION
1
2 Figure 6.1-4 Approach Used to Characterize the Risks from Contaminants of
3 Concern (COCs) to Representative Species in the Housatonic River
4 PSA
5 In the risk characterization, the likelihood and magnitude of adverse effects occurring as a result
6 of exposure of the representative wildlife species to tPCBs and TEQ was evaluated. A weight-
7 of-evidence approach (WOE) was used to make a risk determination for each representative
8 species. Several lines of evidence were available to characterize risks to wildlife from exposure
9 to tPCBs and TEQ, however, not all lines of evidence were available for each species:
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC s s 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
¦ Modeled Exposure and Effects - This line of evidence determines the extent to which
the concentrations of tPCBs and TEQ ingested in the diet will cause adverse effects to
the survival, reproduction, or growth of wildlife. Estimated exposures were compared
to results of toxicological studies reported in the literature to determine if the
representative wildlife species are exposed to tPCBs and TEQ at levels likely to
induce adverse effects.
¦ Field Surveys - When available, this line of evidence was used to determine the
relationship between the concentrations of tPCBs and TEQ and the abundance of
wildlife in the Housatonic River floodplain.
¦ Whole Media or In Situ Studies - When available, this line of evidence was used to
examine the relationship between tPCB and TEQ concentrations at specific sites or in
whole media from the PSA and effects observed in wildlife species. This line of
evidence is considered analogous to the site-specific toxicity testing line of evidence
used in the assessments for aquatic receptors.
Each wildlife risk characterization includes a discussion of sources of uncertainty in the
assessment of risks of COCs to wildlife, and the conclusions of the risk characterization.
6.2 WILDLIFE EXPOSURE MODEL
The approach for conducting the modeled exposure assessment for wildlife relies on the use of
total daily intake models. The primary focus of the model is on ingestion of prey. The dietary
exposure pathway is by far the most important exposure pathway for bioaccumulative substances
such as tPCBs and TCDD and equivalentce (TEQ) (Moore et al. 1997, 1999). Thus, the wildlife
exposure assessments do not include environmental media in the exposure model calculations.
The wildlife exposure model follows the general form:
n
TDI=FIR*FT*Y C.•P.
(Eq. 1)
where
FIR
TDI
Total daily intake (mg/kg bw/d tPCBs, ng/kg bw/d TEQ)
Proportion of z'th food item in the diet (unitless)
Normalized food intake rate (kg/kg bw/d)
Concentration in z'th food item (mg/kg tPCBs, ng/kg TEQ)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-7
7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
FT = Fraction of time in the contaminated area (unitless)
This general exposure model was customized accordingly for each representative species to
reflect feeding habits, foraging range, habitat preferences, and life history. Extensive literature
searches were conducted and data collected to determine the appropriate model inputs. Each of
these inputs is briefly discussed below.
6.2.1 Food Intake Rate (FIR)
Data on food intake rate (FIR) are only available for a few species, primarily due to the
difficulties in measuring intake for free-ranging wildlife. This assessment does not use measured
food intake rates determined using captive animals, because such animals do not expend energy
foraging for food and water, avoiding predators, defending territories, etc. (EPA 1993). Thus,
food intake rates estimated for captive animals considerably underestimate expected food intake
rates for free-ranging animals. In this assessment, allometric equations developed from
measurements of free metabolic rate (FMR) in free-ranging animals (see text box below) were
used to estimate food intake rate for each representative wildlife species. Food intake rate is
derived from FMR using the following equation:
FIR(g / day) =__™? (Eq. 2)
^ (AEi x GEi)
2=1
where Ali, is the assimilation efficiency of the z'th food item (unitless) and GEi is the gross energy
of the 7th food item (kcal/g). Where measured food intake rates are available for free-ranging
animals for a representative species, the measured food intake rates are compared to the
corresponding food intake rate derived using the allometric modeling approach. Such
comparisons can be found in the wildlife sections and appendices.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-8
-------
1
Example Wildlife Free Metabolic Rate Equations
2
Birds
3
FMRikJ!day) = 10.5 • BW(g)0 681 (Coraciiformes)
4
Mammals
5
FMR(kJ/day) = log0.221 • BW(g)0'869 (Carnivores)
6
where
BW =
(Nagy et al. 1999)
7
body weight (g).
8
9 6.2.2 Body Weight (BW)
10 Body weights for each of the representative wildlife species were determined through data from
11 the literature or data collected in the PSA of the Housatonic River. Data were combined from
12 each relevant, acceptable study, and the mean and standard deviations calculated. Body weight
13 is assumed to be a normally distributed parameter. The uncertainty associated with the variable
14 is generally due to natural variability, rather than a lack of knowledge or data (i.e., body weight
15 is easy to measure and data are available for each of the representative species).
16 6.2.3 Proportions of Dietary Items
17 Extensive literature searches were conducted to locate data and information on the dietary
18 preferences of the wildlife species assessed. The information in the literature on dietary
19 preferences was evaluated to determine relevance to representative species in the PSA and the
20 timing of their exposures to COCs. Some wildlife species have dietary preferences that can
21 include a large number of prey items. Therefore, only dietary items that comprise at least 10%
22 of the total diet of each species were included in the exposure model. In these cases, dietary
23 items for prey items comprising >10% of the diet were adjusted resulting in the sum of all
24 dietary components equaling 100%. Because diets vary between locations and individuals and
25 are also uncertain because of the limited data available for some species, distributions were used
26 for dietary variables.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-9
7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
6.3 SPATIAL AND TEMPORAL AVERAGING
Concentrations of COCs vary spatially and temporally in prey. The representative wildlife
species forage over distances ranging from tens of meters to greater than 10 km. Thus,
individuals tend to integrate spatial variation in the tissue concentrations of their prey over time.
Therefore, estimates of the central tendency (i.e., arithmetic means) are used in the exposure
model as an expression of the spatial and temporal averaging of concentrations of COCs in prey
tissues (EPA 1999). In the Monte Carlo analysis, it was assumed that the spatially and
temporally averaged exposure estimate did not vary between individuals foraging in the same
area. Thus, the point estimate of centrality was the minimum of:
1. The 95% upper confidence limit calculated using the Land H-statistic (assuming data
are lognormally distributed), or
2. The maximum concentration measured.
In the probability bounds analyses, however, the uncertainty regarding the arithmetic mean was
accounted for with a different procedure.
The procedure for the probability bounds analysis generally involved using the Land H-statistic
to estimate the lower and upper 95% confidence limits on the mean (Gilbert 1987), and then
using these lower and upper confidence limits to derive bounds on all possible distributions that
exist within this range. This approach results in an expression of the uncertainty about the true
value of the arithmetic mean that arises due to the small sample size. In cases where the 95%
upper confidence limit could not be estimated, or exceeded the maximum measured
concentration, other techniques were used to derive the bounds on the mean (see Appendices G
to K). Appendix C.5 describes the procedures for parameterizing prey concentration variables in
more detail.
EPA (1992) states that because of the uncertainty associated with estimating the true average
concentration for a site, the 95% upper confidence limit (UCL) of the arithmetic mean should be
used for this variable." For lognormal data, EPA (1992) recommends the Land method using the
H-statistic. Several authors (e.g., Ott 1995; Seiler and Alvarez 1996; Hattis and Burmaster 1994)
have argued that concentrations of contaminants in environmental media tend to be lognormally
distributed and that this may be expected because of mechanistic reasons. Current EPA guidance
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC r- -> rx 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
(EPA 1997; also see Haimes et al. 1994) states that distributions should be chosen for input
variables on the basis of mechanistic or theoretical reasons, if possible, because such
distributions have the highest degree of confidence. As a result, concentrations of contaminants
in prey were assumed to be lognormally distributed in this ERA, and hence the Land H-statistic
was used to estimate the 95% UCL. To determine the reasonableness of this assumption, the
Shapiro-Wilk test was used to test for lognormality. Over two-thirds of the data sets used in the
wildlife assessments passed the test for lognormality (i.e. , P > 0 .05), which supports the
assumption of lognormality for concentrations of contaminants. That said, it is recognized that
the Land method can produce high values for the UCL, particularly when data are not
lognormally distributed, sample size is small, or variation is high (Singh et al. 1997; Schultz and
Griffin 1999). EPA's (1992) guidance recognized this problem and recommended that the
maximum detected concentration be used when the calculated UCL exceeds this value. This
guidance was followed in this assessment.
6.4 TOXIC EQUIVALENCE (TEQ)
Some PCB congeners belong to a large class of chemicals called planar chlorinated
hydrocarbons (PCH) that are regularly detected in the environment. The PCHs also include
poly chlorinated dibenzo-p-dioxins (PCDDs), and poly chlorinated dibenzo-furans (PCDFs).
PCHs have a common structural relationship that includes lateral halogenation (i.e., the addition
of a halogen such as fluorine or chlorine to a compound), and the ability to assume a planar
conformation (Figure 6.4-1).
Figure 6.4-1 Molecular Structure of the Planar Chlorinated Hydrocarbon, 2,3,7,8-
Tetrachlorodibenzo-p-dioxin
MK0110:\20123001.096\ERA_PB\ERA_PB_6.DOC ^ i i 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
This structure is important because it leads to a common mechanism of action in many animal
species that involves binding to the aryl hydrocarbon (Ah) receptor and elicitation of an Ah-
receptor-mediated biochemical and toxic response (Van den Berg et al. 1998; Newsted et al.
1995; Safe 1994). Toxic responses include:
¦ Lethality.
¦ Hepatic lesions.
¦ Immunotoxicity.
¦ Tumor promotion.
¦ Adverse effects on reproduction.
¦ Induction of drug-metabolizing enzymes (Van den Berg et al. 1998; Newsted et al.
1995).
The planar structure determines the ability of the chemical to bind with the Ah receptor
(Birnbaum and Devito 1995; Newsted et al. 1995). The Ah receptor facilitates the translocation
of PCHs into the nucleus of affected cells and the binding of the PCH-Ah receptor complex to
sites on the DNA (Newsted et al. 1995). Environmental degradation of PCH congeners varies
due to their unique physical/chemical properties (Cogliano 1998) and thus there can be
substantial differences between the congeners detected in environmental samples and the
congener makeup of the original product (Cogliano 1998; Van den Berg et al. 1998). The
congeners also have different toxic potencies. To address these issues and effectively estimate
the relative toxicity of these mixtures, various systems have been created involving the
development and use of toxic equivalency factors (TEFs) to derive toxic equivalence (TEQ)
(Van den Berg et al. 1998; Safe 1990, 1994; EPA 1987, 1989, 1991; Kennedy 1996; NATO
1988a, 1988b; Ahlborg et al. 1994). The TEQ approach is based on the in vivo and in vitro
toxicity of each of the PCH congeners in relation to 2,3,7,8-tetrachlorodibenzo-p-dioxin
(TCDD). TCDD is considered to be the most toxic member of the PCH class of chemicals (Van
den Berg et al. 1998; Birnbaum and DeVito 1995; Safe 1994). For this ERA, the TEFs proposed
by Van den Berg et al. (1998) (also referred to as the World Health Organization or WHO TEFs)
have been adopted (Table 6.4-1). These TEF values were developed for compounds that:
¦ Show a structural relationship to PCDDs and PCDFs.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC ^ -> rs 7/10/2003
-------
Table 6.4-1
World Health Organization Toxic Equivalency Factors (TEFs) for TCDD and
Equivalents (Van den Berg et al. 1998)
Congener
Mammals
Fish
Birds
TEF
PCB-77
0.0001
0.0001
0.05
PCB-81
0.0001
0.0005
0.1
PCB-126
0.1
0.005
0.1
PCB-169
0.01
0.00005
0.001
PCB-105
0.0001
<0.000005*
0.0001
PCB-114
0.0005
<0.000005*
0.0001
PCB-118
0.0001
<0.000005*
0.00001
PCB-123
0.0001
<0.000005*
0.00001
PCB-156
0.0005
<0.000005*
0.0001
PCB-157
0.0005
<0.000005*
0.0001
PCB-167
0.00001
<0.000005*
0.00001
PCB-189
0.0001
<0.000005*
0.00001
1,2,3,4,6,7,8-HpCDD
0.01
0.001
<0.001*
1,2,3,4,6,7,8-HpCDF
0.01
0.01
0.01
1,2,3,4,7,8,9-HpCDF
0.01
0.01
0.01
1,2,3,4,7,8-HxCDD
0.1
0.5
0.05
1,2,3,4,7,8-HxCDF
0.1
0.1
0.1
1,2,3,6,7,8-HxCDD
0.1
0.01
0.01
1,2,3,6,7,8-HxCDF
0.1
0.1
0.1
1,2,3,7,8,9-HxCDD
0.1
0.01
0.1
1,2,3,7,8,9-HxCDF
0.1
0.1
0.1
1,2,3,7,8-PeCDD
1
1
1
1,2,3,7,8-PeCDF
0.05
0.05
0.1
2,3,4,6,7,8-HxCDF
0.1
0.1
0.1
2,3,4,7,8-PeCDF
0.5
0.5
1
2,3,7,8-TCDD
1
1
1
2,3,7,8-TCDF
0.1
0.05
1
OCDD
0.0001
<0.0001*
0.0001
OCDF
0.0001
<0.001*
0.0001
* Values that are "less than" should be considered to be the upper limit for use in any TEQ
calculation.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC r 10 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
¦ Bind to the Ah receptor.
¦ Elicit an Ah-receptor-mediated biochemical and toxic response.
¦ Are persistent and accumulate in the food chain (Van den Berg et al. 1998;
Birnbaum and DeVito 1995).
The WHO TEFs are the most recent estimates of 2,3,7,8-TCDD equivalence and are based on
current scientific research (Dyke and Stratford 2002). They have been accepted and applied in
numerous jurisdictions worldwide (Dyke and Stratford 2002). Assumptions are made when
using the TEF approach, including:
¦ PCH congeners are Ah-receptor antagonists and their toxicological potency is
mediated by their binding affinity.
¦ No interaction occurs between the congeners and thus, the sum of the
individual congener effects accounts for the potency of the PCH mixture.
The overall effect of these assumptions is a potency estimate or toxic equivalence (TEQ) value.
To generate a TEQ the following equation (Equation 1- modified from Van den Berg et al. 1998)
is used:
6 10 12
TEQ = X \PCDDn x TEFn ] + £ \PCDFp x TEFp ] + £ \PCBq x TEFq ] (Eq. 3)
n=1 p=1 q=1
where
TEQ = Toxic equivalence
PCDDn = Polychlorinated dibenzo-p-dioxin congener concentration
PCDFp = Polychlorinated dibenzo-furan congener concentration
PCBq = Polychlorinated biphenyl congener concentration
TEF„:P:q = Toxic equivalency factor for appropriate individual PCDD/PCDF and PCB
congeners, respectively
Two circumstances often arise when calculating a TEQ value:
¦ Congener concentrations may be below the detection limit (i.e., non-detects), and
¦ Some congeners may not be resolved due to co-elution during analysis.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-14
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
The approach used to address each of these circumstances in the ERA is discussed in the
following sections.
6.4.1 Non-Detects
Congeners detected at or below the detection limit (DL) were included in the TEQ calculations
by investigating three options: first, setting the value for the congener equal to zero (0), setting it
to half the DL, and, finally, setting it equal to the DL (Appendix C.2). A comparison of the
results of this bounding analysis provides a description of the uncertainty surrounding the TEQ
value due to concentrations of one or more congeners being below the detection limit. This
approach is also useful for determining the relative influence of individual non-detected
congeners on the estimated TEQ value. Concentrations of tPCBs in prey in the PSA were all
above the detection limit; therefore, there is no non-detect issue for tPCBs. However, treatment
of non-detects remains a concern for the TEQ congeners.
6.4.2 Congener Co-Elution
The development of a TEQ using the WHO approach requires the concentrations for each of 29
unique congeners (12 PCB and 17 PCDD/PCDF congeners). During analysis of many of the
tissue samples collected for the risk assessment, 2 of the 29 TEQ congeners (i.e., PCB-123 and
PCB-157) co-eluted with other congeners. PCB-123 co-eluted with PCB-149 (PCB-123/149)
and PCB-157 co-eluted with PCB-201 and PCB-173 (PCB-157/201/173). Assuming that the
concentration of the congener PCB-123 is equal to the doublet concentration and that the
concentration of PCB-157 is equal to the triplet concentration would likely overestimate the TEQ
concentration. Conversely, assuming that concentrations of the two congeners (i.e., PCB-123
and PCB-157) were equal to zero would likely underestimate the TEQ concentration. These two
approaches are useful to estimate TEQ bounds, but say little about the relative probabilities of
values between the bounds.
Where possible, independent data sets were located for tissue types where analytical results were
available for the co-eluted congeners in the Housatonic River database. Priority was given to
data sets with tissue samples taken from the Housatonic River to minimize uncertainty associated
with congener metabolism and environmental degradation. Only one appropriate data set was
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC r- -> r 7/10/2003
-------
1 located, for fish tissue, in the Housatonic River that had unique results for each of the congeners
2 in the doublet and triplet. Ratios of the congeners found in the independent data sets were
3 generated and applied to the co-eluted congener data. The co-elution ratio was then multiplied
4 by the reported result for the doublet and triplet concentrations to estimate the PCB-123 and
5 PCB-157 concentrations for fish tissue samples. Uncertainty associated with the method for
6 treating the co-eluted congeners includes interlaboratory variance due to different analytical
7 methods, laboratory conditions, and analyst experience and expertise. The calculated ratios also
8 do not account for differences between species found in the tissue database. A full description of
9 the approach to developing the co-elution ratios from the independent data sets is provided in
10 Appendix C. 10.
11
12
13
14
15 6.4.3 Summary of Decision Criteria for Estimating Exposure Point
16 Concentrations
Co-Elution Ratios
PCB-123/149 - 0.003/0.997
PCB-157/201/173-0.195/0.632/0.174
17 To deal with the uncertainty arising from co-elution or non-detect congeners when estimating
18 exposure point concentrations (EPCs) for use in the exposure analyses, the following decision
19 criteria (Figure 6.4-2) were developed (also see Appendix C.2):
20 ¦ Concentrations of COCs in samples where the concentration was below the
21 detection limit - To determine whether this source of uncertainty was important,
22 arithmetic means were calculated for tissue concentrations assuming a concentration
23 of zero for non-detected COCs (ND = 0), and assuming a concentration equal to the
24 detection limit (ND = DL). If the ratio of the ND = DL mean to the ND = 0 mean
25 was less than 1.3, this source of uncertainty was deemed unimportant. In these cases,
26 exposure calculations were done assuming that concentrations of non-detected COCs
27 were equal to half the detection limit (ND = V2 DL). 1 In cases where the ratio
28
1 This decision criterion supplements the procedures described in Appendix C.2.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-16
-------
UCL = Lower of the 95% UCL from the Land H-statistic or the dataset max
LCL = Higher of the 95% LCL from the Land H-statistic or the dataset min
Figure 6.4-2 Decision Tree for Determining Appropriate Treatment of Data with
Non-Detects and Co-Elution
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-17
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
exceeded 1.3, the source of uncertainty was considered sufficiently important to
incorporate in the exposure analysis. In the Monte Carlo analysis, for samples with
COC concentrations below the detection limit three estimates of the EPC (i.e.,
estimates assuming ND = 0, ND = V2 DL, ND = DL) were used as parameters in a
triangular distribution (i.e., minimum, best estimate, maximum). In the probability
bounds analysis, the distribution-free range was the range spanning the LCL
calculated assuming ND = 0 for the lower limit, and the UCL calculated assuming
ND = DL for the upper limit.
¦ Concentrations of TEQ in tissue samples (other than fish) with co-eluted
congeners - In some tissue samples, two PCB congeners required in the TEQ
calculation (PCB-157 and PCB-123) co-eluted with other congeners. As a result, the
concentrations of the triplet PCB-201/157/173 and the doublet PCB-149/123 are
known, but not the concentrations of PCB-157 and PCB-123. This source of
uncertainty was accounted for in the exposure calculations using an approach similar
to that used to account for uncertainty stemming from non-detected COCs. For each
tissue concentration variable, a ratio was calculated for mean TEQ concentration
assuming that the concentration of PCB-157 and PCB-123 was zero, and the mean
TEQ concentration assuming that the concentrations of these congeners were equal to
the triplet and doublet concentrations, respectively. If the ratio was less than 1.3, this
source of uncertainty was deemed unimportant. In these cases, exposure calculations
were done assuming that concentrations of PCB-157 and PCB-123 were equal to the
triplet and doublet concentrations, respectively. In cases where the ratio exceeded
1.3, the source of uncertainty was considered sufficiently important to incorporate in
the exposure analysis. The procedures followed to accomplish this task were the
same as used to deal with uncertainty due to non-detected COCs.
6.5 PROBABILISTIC RISK ASSESSMENT
6.5.1 Distribution Selection
Input distributions for the exposure analyses were generally assigned as follows: lognormal
distributions for variables that were right skewed with a lower bound of zero and no upper bound
(e.g., amount of COC transferred from mother to offspring via egg tissue), beta distributions for
variables bounded by zero and one (e.g., proportion of a prey item in the diet), normal
distributions for variables that were symmetric and not bounded by one (e.g., body weight), and
point estimates for minor variables or variables with low coefficients of variation. In certain
situations (e.g., poor fit of data), other distributions were fit to the data or other approaches were
used. To quantify uncertainty, two approaches were used as described in Section 6.5.2, below.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-18
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
6.5.2 Monte Carlo and Probability Bounds Analysis
General Risk Assessment Approaches
Deterministic Methods - Methods in which all biological, chemical, physical, and
environmental parameters are assumed to be constant and can be accurately
specified.
Probabilistic Methods - Methods in which important biological, chemical, physical,
and environmental parameters are assumed to vary or are uncertain and therefore,
are specified using distribution of possible values.
Monte Carlo and probability bounds analysis are two uncertainty propagation techniques used in
the Housatonic River wildlife risk assessments. The use of probabilistic methods in risk analysis
is growing rapidly and EPA has produced guidance on how to conduct such analyses in
Superfund and other programs (EPA 1997, 1999). The benefit of using probabilistic methods in
risk assessment is that they give the risk assessor the ability to fully characterize risk, rather than
providing a best estimate or a conservatively biased estimate of risk. For example, calculating a
mean risk (i.e., deterministic method) may exclude the potential for relatively rare, but serious,
extreme events (e.g., species extinction). This is generally undesirable, because although rare,
these events can occur and have significant impacts on individuals, communities, and
populations of species. By including the entire distribution for risk, all events are considered and
all of the data and information collected to characterize a situation are included. The remainder
of this section provides a short overview of Monte Carlo and probability bounds analysis as
applied in the wildlife risk assessments. Further technical detail on these methods can be found
in Appendix C.4.
Probabilistic Methods
Monte Carlo Analysis - A technique where parameter values are drawn at random
from defined input probability distributions, combined according to a model equation,
and the process repeated iteratively until a stable distribution of solutions results. It
is most useful when input distributions are known precisely.
Probability Bounds Analysis - Separates uncertainty and variability to obtain
bounds on the result that explicitly account for the uncertainty about the input
distributions.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-19
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
The primary goal of a Monte Carlo analysis in the risk assessment is to characterize
quantitatively, the uncertainty and variability in estimates of exposure and risk (EPA 1997). A
secondary goal is to identify key sources of variability and uncertainty and to quantify the
relative contribution of these sources to the overall range of wildlife exposure model results.
While Monte Carlo methods are appropriate for the determination of exposure risks when input
distributions are known precisely, they may not adequately represent the effects of uncertainty
about how to parameterize variability in the input distributions (Ferson 1996). In many
ecological risk assessments, the available data are limited and consequently the input
distributions used to calculate risks are uncertain. Probability bounds analysis is a tool for
separating variability and uncertainty to obtain bounds on the result that explicitly account for
uncertainty about the input distributions. As in Monte Carlo analysis, the overall slopes of the
bounds indicate how much variability exists in the system. The distance between the bounds, on
the other hand, is an indication of the uncertainty that exists due to lack of knowledge. An
example of exposure model outputs from Monte Carlo and probability bounds analyses is
presented in Figure 6.5-1.
The wildlife exposure models contain multiple variables, some of which may be correlated. The
assumption of independence can be inappropriate, because dependencies can affect the estimates
of exposure. If correlations are not accounted for, the variance and the tails of the exposure
distribution may be poorly estimated. The wildlife assessments use several approaches to
address correlations between variables. These approaches include simulation of observed
correlations, assumption of perfect covariance (e.g., when the diet consists of two prey items, the
proportion of one item in the diet is equal to one minus the other item), or no assumptions at all
about dependencies (e.g., all possible relationships between two variables can occur). The
specific approach used depends on the type of data and the application. In cases where
independence of variables seemed intuitively obvious (e.g., COC concentration in the prey item
and proportion of that item in the diet), independence was assumed.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-20
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
100
•Monte Carlo
- Upper PBound
Lower P Bound
0
0
20
40
60
TDI (nig/kg bw/d)
Figure 6.5-1 Example Exposure Distribution from Monte Carlo and Probability
Bounds Analyses (TDI = total daily intake of tPCBs)
6.6 EFFECTS ASSESSMENT
Effects data can be characterized and summarized in a variety of ways ranging from benchmarks
designed to be protective of most or all species to dose-response curves for the representative
species of interest. In this ERA, effects characterization preferentially relied on concentration-
or dose-response curves, but defaulted to benchmarks or other estimates of effect (e.g., no
observed adverse effect level [NOAEL], or lowest observed adverse effect level [LOAEL]) when
insufficient data were available to derive dose-response curves. Effects associated with growth,
survival, and reproduction were generally the preferred measures of effect as they most closely
relate to the assessment endpoints for wildlife. This section provides an overview of the
procedures used for characterizing effects information and describes the decision criteria for
choosing among them for each receptor-COC combination.
Figure 6.6-1 displays the hierarchy of decision criteria used to characterize effects for each
receptor-COC combination. In all cases, the units of the effects metrics were consistent with the
MK0110:\20123001.096\ERA_PB\ERA_PB_6.DOC r 7/10/2003
-------
1 units of the exposure metrics. To the extent possible, effects metrics were based on long-term
2 studies to match expected exposure durations.
3 The remainder of this section provides details on how the effects metrics were derived from the
4 decision tree.
5. Derive a range wherein the threshold for the receptor of interest is expected to occur. Because
information on the sensitivity of the receptor of interest is lacking, it is difficult to derive a
threshold that is not biased high or low. If data are available for several other species, however,
calculate thresholds for the most sensitive and tolerant species to determine a threshold range
that is likely to include the threshold for the representative species.
5
6 Figure 6.6-1 Decision Criteria Used to Characterize Effects for Each
7 Wildlife Receptor-COC Combination
8 6.6.1 Dose-Response Relationships Using the Generalized Linear Model
9 Framework
10 Most probabilistic risk assessments previously conducted estimated the probability that exposure
11 exceeded a specified no-observed-effects or lowest-observed-effects dose. An alternative
12 approach is to estimate the probabilities of effects of varying magnitude. To do this, a
13 concentration- or dose-response model is required. Generally, five or more treatments are
14 required to develop concentration- or dose-response relationships, either from a single study or
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC A O O 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
from several studies that used a similar methodology. The Generalized Linear Model (GLiM)
framework described by Kerr and Meador (1996) and Bailer and Oris (1997) is a useful
framework for deriving these relationships. The framework involves using link functions to
transform effects metrics (e.g., probit or logit link functions for quantal responses) and assigning
appropriate error distributions (e.g., binomial distribution for quantal responses). Linear
regression can then be conducted on the transformed data to derive the dose-response
relationship. Thus, the framework can be used for all available types of response variables
(Moore et al. 2000). By adding a quadratic term to the linear model, the framework can be
adapted to incorporate simulation at low doses. The GLiM framework was used to derive dose-
response relationships in this ERA when five or more treatments were available from a single
study for the receptor of interest or a reasonable surrogate. In some cases, it was necessary to
convert concentration-response relationships to dose-response relationships by multiplying the
former by the food intake rate of the species (Moore et al. 1999).
Dose-response relationships are combined with the corresponding exposure distribution in risk
characterization to derive risk curves that characterize the relationship between probability and
magnitude of effect.
6.6.2 Hypothesis Testing to Determine LOAEL and NOAEL
Analysis of variance (ANOVA) is the most common method of estimating low-level toxic
effects from chronic tests. There are several reasons for this, including the wide availability of
software capable of performing ANOVA and related nonparametric tests, and the familiarity of
regulators with the technique. Until recently, most toxicity-testing protocols specified
experimental designs more suited to hypothesis-testing methods such as ANOVA than to
regression-based approaches. However, hypothesis testing as an approach for estimating low-
level toxic effects has some limitations, including:
1. NOAELs and LOAELs are test doses that do not correspond with specified effects
levels from one test to the next.
2. Poor experimental design may mistakenly indicate that a contaminant is less toxic
than it really is.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
3. Most information available from the toxicity test is not used (Stephan and Rogers
1985; Pack 1993; Suter 1996).
As a result, hypothesis testing was not the preferred method for analysis of toxicity data in this
ERA.
However, in many cases, toxicity studies with five or more treatment levels are not available for
the representative species of interest or for a reasonable surrogate for tPCBs and TEQ. In those
cases, the use of hypothesis testing was necessary to estimate the NOAEL and LOAEL. In many
toxicological studies, these endpoints were previously determined and reported. Such studies
were evaluated to determine that proper statistical procedures were followed. Where the data
could be obtained from the reports or directly from the authors, the data were re-analyzed. In
cases where a re-analysis was conducted, information regarding the minimal difference required
to give a significant result was reported (e.g., number of replicates, test variance, a, P, test dose
intervals). The percent effect associated with the LOAEL, relative to the control, was also
reported.
6.6.3 Field-Based Measures of Effect and Threshold Ranges
Field-based measures of effect were derived from monitoring or in situ toxicity tests conducted
on the representative species or a reasonable surrogate. There are several methods available for
deriving field-based measures of effect. For benthic invertebrates, chemistry and effect data
from surveys of sediment and biota in various locations have been combined to develop sediment
concentrations that are generally protective or, conversely, likely associated with adverse effects
(Long et al. 1995; MacDonald et al. 1996). Similar approaches can be used with wildlife
species. With in situ studies, if sufficient data were available, regression-based approaches (e.g.,
GLiM models) could be used to link concentrations or doses with effects observed in the field.
When data are lacking on the toxicity of a particular COC to the representative species or a
reasonable surrogate, threshold ranges were developed. In these cases, it is not known whether
the representative species is sensitive or tolerant. Therefore, a threshold range was developed
that spanned the concentrations (or doses) that would be protective of sensitive species to those
that would be protective only for tolerant species. The assumption is that the threshold for the
representative species lies between these two extremes. To derive a threshold range, the toxicity
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC A O /I 7/10/2003
-------
1 literature was reviewed to determine the most sensitive and the most tolerant species for which
2 studies have been conducted. Thresholds were derived for both the most sensitive and the most
3 tolerant species using methods similar to those used in the Pre-ERA (see Section 2.4 and
4 Appendix B). The two resulting thresholds become the threshold range, which was then
5 compared to the exposure assessment results in the risk characterization.
6 6.7 RISK CHARACTERIZATION
7 6.7.1 Risk Categorization
8 Whenever possible, risk should be expressed quantitatively (Wentsel et al. 1997). For example,
9 a risk could be expressed as a 10% probability of >25% mortality for a particular species. In this
10 ERA, quantitative expressions of risk were derived for each of the wildlife assessment endpoints
11 (Appendices G to K) to facilitate discussion and to simplify comparisons of risk between species,
12 COCs, and locations. The following criteria were used to categorize risks to wildlife as high,
13 intermediate, or low:
14 ¦ Scenarios with effects data for the representative species (or a reasonable surrogate):
15 - If the probability of 10% or greater effect (or of exceeding the NOAEL) was less
16 than 20%, then the risk was categorized as low (Figure 6.7-1).
17 - If the probability of 20% or greater effect (or of exceeding the LOAEL) was
18 greater than 50%, then the risk was categorized as high (Figure 6.7-1).
19 - All other outcomes were categorized as intermediate risk (Figure 6.7-1).
20 ¦ Scenarios with effects data for the representative threatened and endangered species
21 (or a reasonable surrogate):
22 - If the probability of 10% or greater effect (or of exceeding the NOAEL) was less
23 than 20%, then the risk was categorized as low.
24 - If the probability of 10% or greater effect (or of exceeding the NOAEL) was
25 greater than 50%, then the risk was categorized as high.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-25
-------
loo r?
Intermediate Risk
-Q
-Q
O
•-
-
u
=
¦a
&
&
u
M
80
60
40
20
¦ Low Risk
-High Risk
20
40 60
% Effect
80
100
1
2 Note: Risk curves passing below the filled circle symbol indicate low risk, while those passing above the open circle
3 symbol indicate high risk.
4
5 Figure 6.7-1 Example Risk Curves Indicating Low, Intermediate, and High Risk
6 Categories
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-26
-------
1
All other outcomes were categorized as intermediate risk.
2
3
¦ Scenarios with threshold concentrations for the representative species (or a reasonable
surrogate):
4
5
If the probability of exceeding the threshold for the most sensitive species was
less than 20%, the risk was categorized as low.
6
7
If the probability of exceeding the threshold for the most tolerant species was
greater than 20%, the risk was categorized as high.
8
All other outcomes were categorized as intermediate risk.
9 Each categorization of risk was derived from the results of the Monte Carlo exposure analyses
10 (Figure 6.7-1). To capture the uncertainty about a risk categorization, the results from the
11 corresponding probability bounds analysis were compared to the above criteria to determine a
12 risk range (risk category using lower probability bound to risk category using upper probability
13 bound).
14 These risk categorization criteria were based on several considerations including:
15 ¦ Efroymson and Suter (1999) and others (e.g., Pack 1993) suggested that reductions in
16 survival, growth, or reproduction of 20% or greater is indicative of significant effects
17 to wildlife. Thus, a better than even chance (i.e., >50%) of exceeding this effect level
18 was deemed to represent a high risk situation. However, because effects at or above
19 the 20% level possibly may not be ecologically significant, these categorizations
20 should be considered further in each situation. For example, a stressor causing a 20%
21 decline in reproductive fecundity of brook trout was shown to lead to a general
22 lowering of risks of population decline compared to unexposed conditions because
23 the negative consequences of overcrowding were diminished (Ferson et al. 1996).
24 Similar effects on other fish species, however, have led to population collapses
25 (Myers et al. 1995).
26 ¦ Although there are exceptions (such as threatened and endangered species), an effect
27 level of 10% is unlikely to be ecologically significant. Thus, if the probability of
28 exceeding this effect level is relatively low (<20%), risk is deemed to be negligible to
29 low.
30 ¦ Several studies have shown that NOAELs are generally associated with effects of
31 10%) or greater (85% of studies examined by Moore et al. 1997), and LOAELs are
32 generally associated with effects of 20% or greater (79% of studies examined by
33 Moore et al. 1997) (also see Hoeckstra and Van Ewijk 1993; Pack 1993). Therefore,
34 the decision criteria above equated NOAELs with the 10% effect level, and LOAELs
35 with the 20% effect level.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC A O 7 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
¦ When toxicity data are lacking for representative species or reasonable surrogates, the
toxicity threshold for representative species is assumed to be between the thresholds
of the most sensitive and tolerant species tested. Thus, if the probability of exceeding
the lowest threshold is low (<20%), risk is deemed to be negligible to low. Tolerant
species may have thresholds one to several orders of magnitude higher than sensitive
species (see effects assessment sections in Appendices G to K). Thus, at the highest
threshold, it is likely that some representative species would be adversely affected,
possibly quite seriously. Thus, even a relatively low probability (20% or greater) of
exceeding the upper threshold may be cause for concern.
¦ Any effect to threatened and endangered species is cause for concern (Massachusetts
Office of Environmental Affairs 1999; Massachusetts Division of Fisheries and
Wildlife 2003; United States Congress 1973, Endangered Species Act). Because a
LOAEL generally represents >20% effect, the criterion separating intermediate and
high risk was adjusted for threatened and endangered species. Thus, a better than
even (>50%) chance of exceeding a NOAEL was deemed to represent a high risk
situation for threatened and endangered species.
The risk categories should not be used alone to determine whether risk management actions are
necessary. Risk categories are based on the results of the Monte Carlo exposure analysis only.
Risk categories are uncertain in cases where the risk range is wide (e.g., low to high). Risk
categories are also uncertain for assessment endpoints without corroborating lines of evidence.
Thus, the risk categories should be considered as a qualitative ranking of risk to facilitate
comparisons between COCs, locations, and assessment endpoints. They are intended to
contribute to weight-of-evidence assessments, not to replace them.
6.7.2 Weight-of-Evidence Assessment
A WOE approach was used in the risk assessments for wildlife. The WOE approach is a process
by which measurement endpoints are related to an assessment endpoint to evaluate whether
significant harm is posed to the environment (Menzie et al. 1996). The WOE approach used in
this ERA follows the approach originally described in the Massachusetts Weight of Evidence
Special Report (Menzie et al. 1996). A detailed review of the WOE approach used in the
Housatonic River ERA is provided in Section 2.9. In general, the WOE approach is an inclusive
process whereby multiple lines of evidence are considered prior to determining risk. For the
wildlife risk assessments, these lines of evidence included the exposure and effects modeling
results, field survey results, and/or in situ or whole media toxicity test results.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-28
-------
1 For the wildlife assessment endpoints, risk categories and risk ranges were developed for the
2 modeling of exposure and effects line of evidence. The Massachusetts Weight-of-Evidence
3 (WOE) approach requires a determination for evidence of harm and magnitude of effect for each
4 assessment endpoint-COC scenario. For this assessment, criteria were developed for converting
5 risk category and risk range to evidence of harm and magnitude of effect on the Massachusetts
6 WOE scoring sheets (Table 6.7-1).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-29
-------
1
2
3
Table 6.7-1
Decision Criteria for Converting Risk Category and Range to Evidence of Harm and Magnitude of Effect
Risk Category
Risk Range
Low
Low/Intermediate
Intermediate
Intermediate/High
High
Low/High
Low
Evidence=No
Magnitude=Low
Evidence=No
Magnitude=Low
...
...
...
Evidence=Undetermined
Magnitude=Low
Intermediate
...
Evidence=Undetermined
Magnitude=Intermediate
Evidence=Yes
Magnitude=Intermediate
Evidence=Yes
Magnitude=Intermediate
...
Evidence=Undetermined
Magnitude=Intermediate
High
...
...
...
Evidence=Yes
Magnitude=High
Evidence=Yes
Magnitude=High
Evidence=Undetermined
Magnitude=High
4 Evidence=Evidence of Harm (Yes, No, Undetermined), Magnitude=Magnitude (High, Intermediate, Low), — Indicates outcome is not possible
5
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC s- o f\ 7/10/2003
-------
1 6.8 REFERENCES
2 Alhborg, U.S., G.C. Becking, L.S. Birnbaum, A. Brouwer, H.J.G.M. Derks, M. Feeley, G. Golor,
3 A. Hanberg, J.C. Larsen, A.K.D. Liem, S.H. Safe, C. Schlatter, F. Wern, M. Younes, and E.
4 Yrjanheikki. 1994. Toxic equivalency factors for dioxin-like PCBs: Report on a WHO-ECEH
5 and IPCS consultation, December 1993. Chemosphere 28:1049-1067.
6 Bailer, A.J. and J.T. Oris. 1997. Estimating inhibition concentrations for different response
7 scales using generalized linear models. Environmental Toxicology and Chemistry 16:1554-1559.
8 Birnbaum, L.S. and M.J. DeVito. 1995. Use of toxic equivalency factors for risk assessment for
9 dioxins and related compounds. Toxicology 105:391-401.
10 Cogliano, V.J. 1998. Assessing the cancer risk from environmental PCBs. Environmental Health
11 Perspectives 106(6) :317-323
12 Dyke, P.H. and J. Stratford. 2002. Changes to the TEF schemes can have significant impacts on
13 regulation and management of PCDD/F and PCB. Chemosphere 47:103-116.
14 Efroymson, R. and G.W. Suter II. 1999. Finding a niche for soil microbial toxicity tests in
15 ecological risk assessment. Human and Ecological Risk Assessment 5(4):635-868.
16 EPA (U.S. Environmental Protection Agency). 1987. Interim Procedures for Estimating Risk
17 Associated with Exposure to Mixtures of Chlorinated Dibenzo-p-dioxins and Dibenzofurans
18 (CDD:CDFs). U.S. Environmental Protection Agency. EPA 625-3-87012.
19 EPA (U.S. Environmental Protection Agency). 1989. Interim Procedures for Estimating Risks
20 Associated with Exposure to Mixtures of Chlorinated Dibenzo-p-dioxins and Dibenzofurans and
21 1989 Update. Risk Assessment Forum, EPA/625/3-89/016.
22 EPA (U.S. Environmental Protection Agency). 1991. Workshop Report on Toxicity Equivalence
23 Factors for Poly chlorinated Biphenyl Congeners Risk Assessment Forum. U.S. Environmental
24 Protection Agency. EPA/625/3-91/020.
25 EPA (U.S. Environmental Protection Agency). 1992. A Supplemental Guidance to RAGS:
26 Calculating the Concentration Term. Publication 9285.7-081. Office of Research and
27 Development, Washington, DC. http://www.deq.state.ms.us/newweb/opchome.nsf/pages/
28 HWDivisionFiles/ $file/uclmean.pdf.
29 EPA (United States Environmental Protection Agency). 1993. Wildlife Exposure Factors
30 Handbook. Volumes I and II. Office of Research and Development. Washington, DC. EPA
31 EPA/600/R-93/187a, EPA/600/R-93/187b.
32 EPA (U.S. Environmental Protection Agency). 1997. Guiding Principles for Monte Carlo
33 Analysis. U.S. Environmental Protection Agency, Office of Research and Development,
34 Washington, DC. EPA/630/R-97/001.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC ^ ^ i 7/10/2003
-------
1 EPA (U.S. Environmental Protection Agency). 1999. Risk Assessment Guidance for Superfund,
2 Volume 3 - Part A, Process for Conducting Probabilistic Risk Assessment. Draft. Office of Solid
3 Waste and Emergency Response, U.S. Environmental Protection Agency, Washington, DC.
4 Ferson, S. 1996. What Monte Carlo methods cannot do. Human and Ecological Risk Assessment
5 2(4):990-1007.
6 Ferson, S., L.R. Ginzburg, and R.A. Goldstein. 1996. Inferring ecological risk from toxicity
7 bioassays. Journal of Water, Air and Soil Pollution 90:71-82.
8 Gilbert, R.O. 1987. Statistical Methods for Environmental Pollution Monitoring. Van Nostrand
9 Reinhold Company, New York.
10 Haimes, Y.Y., T. Barry, and J.H. Lambert, Editors. 1994. When and how you can specify a
11 probability distribution when you don't know much? Risk Analysis 14:661-706.
12 Hattis, D. and D.E. Burmaster. 1994. Assessment of variability and uncertainty distributions for
13 practical risk analyses. Risk Analysis 14:713-730.
14 Hoeckstra, J.A. and P.H. Van Ewijk. 1993. Alternatives for the no-observed-effect-level.
15 Environmental Toxicology and Chemistry 12:187-194.
16 Kennedy, S.W., A. Lorenzen, and R.J. Norstrom. 1996. Chicken embryo hepatocyte bioassay for
17 measuring cytochrome P4501A-based 2,3,7,8-tetrachlorodibenzo-p-dioxin equivalent
18 concentrations in environmental samples. Environmental Science and Technology 30:706-715.
19 Kerr, D.R. and J.P. Meador. 1996. Modeling dose response using generalized linear models.
20 Environmental Toxicology and Chemistry 15:395-401.
21 Long, E.R., D.D. MacDonald, S.L. Smith, and F.D. Calder. 1995. Incidence of adverse
22 biological effects within ranges of chemical concentrations in marine and estuarine sediments.
23 Environmental Management 19(l):81-97.
24 MacDonald, D.D., R.S. Carr, F.D. Calder, E.R. Long, and C.G. Ingersoll. 1996. Development
25 and evaluation of sediment quality guidelines for Florida coastal waters. Ecotoxicology 5:253-
26 278.
27 Massachusetts Division of Fisheries and Wildlife. 2003. 321 CMR 10.00 and 11.00.
28 Massachusetts Endangered Species Act Regulations. Boston, MA.
29 Massachusetts Office of Environmental Affairs. 1999.310 CMR 40.00. Massachusetts
30 Contingency Plan (MCP). Boston, MA.
31 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S.
32 Petron, B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-
33 of-Evidence Workgroup: A weight-of-evidence approach for evaluating ecological risks. Human
3 4 and Ecological Risk Assessment 2(2):277-3 04.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-32
-------
1 Moore, D.R.J., W. Warren-Hicks, B.R. Parkhurst, R.S. Teed, R.B. Baird, R. Berger,
2 D.L. Denton, and J.J. Pletl. 2000. Intra- and intertreatment variability in reference toxicant tests:
3 Implications for whole effluent toxicity testing programs. Environmental Toxicology and
4 Chemistry 19(1): 105-112.
5 Moore, D.R.J., B.E. Sample, G.W. Suter, B.R. Parkhurst, and R.S. Teed. 1999. A probabilistic
6 risk assessment of the effects of methylmercury and PCBs on mink and kingfishers along East
7 Fork Poplar Creek, Oak Ridge, Tennessee, USA. Environmental Toxicology and Chemistry
8 18(12):2941-2953.
9 Moore, D.R.J., R.L. Breton, and K. Loyd. 1997. The effects of hexachlorobenzene on mink in
10 the Canadian environment: An ecological risk assessment. Environmental Toxicology and
11 Chemistry 16(5): 1042-1050.
12 Moore, D.R.J, and P.-Y. Caux. 1997. Estimating low toxic effects. Environmental Toxicology
13 and Chemistry 16(4):794-801.
14 Myers, R.A., N.J. Barrowman, J.A. Hutchings, and A.A. Rosenberg. 1995. Population dynamics
15 of exploited fish stocks at low population levels. Science 269:1106-1108.
16 Nagy, K.A., I. A. Girard, and T.K. Brown. 1999. Energetics of free-ranging mammals, reptiles,
17 and birds. Annual Reviews of Nutrition 19:247-77
18 NATO (North Atlantic Treaty Organization). 1988a. International Toxicity Equivalence Factors
19 (I-TEF) Method of Risk Assessment for Complex Mixtures of Dioxins and Related Compounds.
20 Pilot Study on International Information Exchange on Dioxins and Related Compounds. Report
21 Number 176, August 1988. North Atlantic Treaty Organization, Committee on Challenges of
22 Modern Society.
23 NATO (North Atlantic Treaty Organization). 1988b. Scientific Basis for the Development of
24 International Toxicity Equivalency Factor (I-TEF), Method of Risk Assessment for Assessment of
25 Complex Mixtures of Dioxins and Related Compounds. Report No. 178.
26 Newsted, J.L., J.P. Giesy, G.T. Ankley, D.E. Tillitt, R.A. Crawford, J.W. Gooch, P.D. Jones, and
27 M.S. Denison. 1995. Development of toxic equivalency factors for PCB congeners and the
28 assessment of TCDD and PCB mixtures in rainbow trout. Environmental Toxicology and
29 Chemistry 14(5): 861 -871
30 Ott, W.R. 1995. Environmental Statistics and Data Analysis. Lewis Publishers, Boca Raton, FL.
31 Pack, S. 1993. A Review of Statistical Data Analysis and Experimental Design in OECD Aquatic
32 Toxicology Test Guidelines. Organization of Economic Cooperation and Development. Paris,
33 France.
34 Safe, S. 1990. Polychlorinated biphenyls (PCBs) dibenzo-p-dioxins (PCDDs), dibenzofurans
35 (PCDFs), and related compounds: environmental and mechanistic considerations which support
36 the development of toxic equivalence factors (TEFs). Critical Reviews of Toxicology 21:51-88.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
-------
1 Safe, S.H. 1994. Polychlorinated biphenyls (PCBs): Environmental impact, biochemical and
2 toxic responses, and implications for risk assessment. Critical Reviews in Toxicology
3 24(2):87-149.
4 Schulz, T.W. and S. Griffin. 1999. Estimating risk assessment exposure point concentrations
5 when the data are not normal or lognormal. Risk Analysis 19:577-584.
6 Seiler, F.A. and J.L. Alvarez. 1996. On the selection of distributions for stochastic variables. Risk
1 Analysis 16:5-18.
8 Singh, A.K., A. Singh, and M. Engelhardt. 1997. The lognormal distribution in environmental
9 applications. EPA/600/R-97/006. U.S. Environmental Protection Agency, Washington, DC.
10 Stephan, C.E. and J.W. Rogers. 1985. Advantages of using regression analysis to calculate
11 results of chronic toxicity tests. In Aquatic Toxicology and Hazard Assessment. R.C. Bahner and
12 D.J. Hansen, Editors. STP 737. American Society for Testing and Materials, Philadelphia, PA.
13 pp. 377-387.
14 Suter, G.W., II. 1996. Toxicological benchmarks for screening contaminants of potential concern
15 for effects on freshwater biota. Environmental Toxicology and Chemistry 15(7): 1232-1241.
16 United States Congress. 1973. Endangered Species Act of 1973 (87 Statute 884). Washington,
17 DC.
18 Van den Berg, M., L. Birnbaum, A.T.C. Bosveld, B. Brunstrom, P. Cook, M. Freely, J.P. Giesy,
19 A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C. Larsen, F.X.R van Leeuwen,
20 A.K. Djien Liem, C. Nolt, R.E. Peterson, L. Poellinger, S. Safe, D. Schrenk, D. Tillitt,
21 M. Tysklind, M. Younes, F. Waern, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs)
22 for PCBs, PCDDs, PCDFs for humans and wildlife. Environmental Health Perspectives
23 106(12):775-792.
24 Wentsel, R.S., M. Simini, R.T. Checkai, D. Ludwig, and T.W. La Point. 1997. Ecological Risk
25 Assessment Procedures for U.S. Army sites. Environmental Toxicology and Risk Assessment:
26 Modelling and Risk Assessment (Sixth Volume), ASTM STP 1317, F.J. Dwyer, T.R. Doane and
27 M.L. Hinman, Eds, American Society for Testing and Materials, West Conshohocken, PA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_6.DOC
6-34
-------
1 7. ASSESSMENT ENDPOINT - SURVIVAL, GROWTH, AND
2 REPRODUCTION OF INSECTIVOROUS BIRDS
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
Highlights
Conceptual Model
The assessment endpoint is the survival, growth, and reproduction of insectivorous
birds in the Housatonic River PSA. Insectivorous birds, including tree swallows and
American robins, are exposed to tPCBs and TEQ via trophic transfer. These two
species were selected as the representative species for the ecological risk
assessment (ERA).
Exposure
Exposure of the representative species to tPCBs and TEQ was determined from: (1)
concentrations found in prey items, (2) an estimation of the daily intake of these
contaminants of concern (COCs) from consumption of prey, and (3) for tree swallows
exposed to tPCBs only, tissue concentrations in 14-day-old nestlings. Site-specific
nesting and reproduction studies were conducted to evaluate adverse effects to tree
swallows and American robins from Housatonic River COCs.
Effects
Field data on the toxicity of tPCBs were available for tree swallows, but not American
robins. No data were available on the toxicity of TEQ to tree swallows or American
robins. Thus, surrogate species were used to estimate effects to American robins
exposed to tPCBs and for both representative species exposed to TEQ. A threshold
range spanning sensitive and tolerant surrogate species was used in these cases.
Risk
Modeled exposure and effects for tree swallows and American robins suggests that
they are at intermediate to high risk as a result of exposure to tPCBs and TEQ in the
Housatonic River PSA. However, the more highly weighted field study line of
evidence suggests that if effects are occurring, they are minor for both species.
Therefore, the weight-of-evidence (WOE) assessment favors a finding of low risk for
insectivorous birds exposed to tPCBs and TEQ in the PSA. This conclusion,
however, is uncertain because of the conflicting results in the WOE assessment.
Other insectivorous bird species common to the PSA are expected to have either a
similar to lower level of risk (e.g., cliff swallow, eastern kingbird, eastern bluebird,
eastern towhee); a similar level of risk (e.g., barn swallow, common nighthawk,
eastern phoebe, hermit thrush, northern mockingbird, veery, wood thrush); or a
similar to higher level of risk (e.g., bank swallow, chimney swift, northern rough-
winged swallow, gray catbird) compared to the representative species.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-1
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
7.1 INTRODUCTION
The purpose of this section is to characterize and quantify the current and potential risks posed to
insectivorous birds exposed to contaminants of potential concern (COPCs) in the Housatonic
River and floodplain, focusing on total PCBs (tPCBs) and other COPCs originating from the
General Electric Company (GE) facility in Pittsfield, MA. The watershed is located in western
Massachusetts and Connecticut, discharging to Long Island Sound, with the GE facility located
near the headwaters of the watershed. The Primary Study Area (PSA) includes the river and 10-
year floodplain from the confluence of the East and West Branches of the Housatonic River
downstream of the GE facility, to Woods Pond Dam.
A pre-ERA was conducted to narrow the scope of the ERA by identifying contaminants, other
than tPCBs, that pose potential risks to aquatic biota and wildlife in the PSA (Appendix B). A
three-tiered deterministic approach was used to screen COPCs. The deterministic assessments
compared potential conservative estimates of exposure with conservative adverse effects
benchmarks to identify COPCs for insectivorous birds in the Housatonic River. A hazard
quotient (total daily intake/effect benchmark) greater than 1 for insectivorous birds in the
Housatonic River area resulted in the COPC being screened through to the next tier of the
assessment, and to the probabilistic ERA if necessary. Subsequent to the pre-ERA, several other
COPCs (primarily organochlorine pesticides) were screened out because their actual
concentrations in the PSA were likely much lower than the measured values due to laboratory
interference (see Section 2.4).
In summary, the COPCs that screened through to the probabilistic risk assessment for
insectivorous birds were tPCBs and 2,3,7,8-TCDD toxic equivalence (TEQ). Total PCBs and
TEQ were retained as contaminants of concern (COCs) for this assessment endpoint. Total
PCBs detected in Housatonic River samples closely resemble the commercial PCB mixtures
Aroclor 1260 and Aroclor 1254, which are similar in congener makeup. TEQ is calculated from
coplanar PCB and dioxin and furan congeners using the toxic equivalency factor (TEF) approach
developed by Van den Berg et al. (1998) (see Section 6.4).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
A step-wise approach was used to assess the risks of tPCBs and TEQ to insectivorous birds in
the Housatonic River watershed. The four main steps in this process include:
1. Derivation of a conceptual model (Figure 7.1-1).
2. Assessment of exposure of birds to COCs (Figure 7.1-2).
3. Assessment of the effects of COCs on birds (Figure 7.1-3).
4. Characterization of risks to the insectivorous bird species (Figure 7.1-4).
This section is organized as follows.
¦ Section 7.2 presents the conceptual model for assessing the ecological risk to
insectivorous birds.
¦ Section 7.3 describes the exposure models, input parameters, and techniques to
propagate uncertainty. Also presented in this section are the exposure modeling
results for tree swallows and American robins.
¦ Section 7.4 provides an overview of the literature on the effects of tPCBs and TEQ to
survival, growth, and reproduction of tree swallows, American robins, and other bird
species. Field studies on the reproduction of tree swallows (Custer 2002) and
American robins (Henning 2002) are discussed. Key studies are then selected and
used to derive the most appropriate effects metrics.
¦ The two lines of evidence for each representative species are discussed in the risk
characterization, Section 7.5, followed by a discussion of the sources of uncertainty in
this assessment, and the conclusions regarding risks of tPCBs and TEQ to
insectivorous birds in the Housatonic River PSA.
This section provides a summary of the ecological risk assessment for insectivorous
birds, which is presented in detail in Appendix G.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-3
-------
0)
G
L_
s
o
<0
¦o
0)
Wetland and
surface water
discharge
Floodplain
runoff
Groundwater
discharge
Contaminated soil, sediment, water and biota
Terrestrial Vegetation
Blackberries,
raspberries
£
o
Q.
0)
G
0)
~ i
Terrestrial
Invertebrates
Earthworms, beetles,
grasshoppers
Insects
Mosquitoes, midges,
gnats, mayflies
Insectivorous Birds
Tree swallow and American robin
o
a)
!fc
LLI
Decreased Survival, Growth, or Reproduction
Figure 7.1-1 Conceptual Model Diagram: Exposure Pathways for Insectivorous
Birds Exposed to Contaminants of Concern in the Housatonic
River PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q a 7/10/2003
-------
EXPOSURE
2
3 Figure 7.1-2 Overview of Approach Used to Assess Modeled Exposure of
4 Insectivorous Birds to Contaminants of Concern in the
5 Housatonic River PSA
6
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q r 7/10/2003
-------
1
2
EFFECTS
4
5 Figure 7.1-3 Overview of Approach Used to Assess the Modeled Effects of
6 Contaminants of Concern to Insectivorous Birds in the
7 Housatonic River PSA
8
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q s 7/10/2003
-------
RISK CHARACTERIZATION
Consider Two Lines of Evidence
Compare Exposure
Distributions to
Effects Metrics
Results of Field
Studies
Generate Risk Curves
or Estimates of Risk
Classify Risks as High,
Intermediate, or Low
Weight-of-Evidence Analysis
Discuss Sources of
Uncertainty
2 Figure 7.1-4 Overview of Approach Used to Characterize the Risks of
3 Contaminants of Concern to Insectivorous Birds in the
4 Housatonic River PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-7
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
7.2 CONCEPTUAL MODEL
The conceptual model presented in Figure 7.1-1 illustrates the exposure pathways for
insectivorous birds exposed to tPCBs and TEQ in the PSA. Total PCBs and TEQ are persistent,
lipophilic, and hydrophobic. Therefore, they are highly bioaccumulated by aquatic and
terrestrial biota directly through the consumption of contaminated prey as part of the food chain
(Haffner et al. 1994; Senthilkumar et al. 2001; Borga et al. 2001). Emergent aquatic insects and
terrestrial invertebrates are major dietary items for insectivorous birds. Insectivorous birds that
reside, or partially reside, within the study area are exposed to tPCBs and TEQ principally
through diet and trophic transfer. Other routes of exposure, considered to be less important to
overall exposure, include inhalation, water consumption, and soil ingestion (Moore et al. 1999).
The problem formulation (see Section 2) identified the tree swallow (Tachycineta bicolor; Figure
7.2-1) and American robin (Turdus migratorius; Figure 7.2-2) as the representative species for
insectivorous birds potentially exposed to tPCBs and TEQ from consumption of contaminated
prey. Life history profiles for these bird species are summarized in the following text boxes.
Additional life history information for both species can be found in Sections G.2.1.3 and G.2.1.4.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
-------
Life History of the Tree Swallow
The tree swallow is distributed widely
throughout northern and central North
America and is one of the most widespread
members of its genus. Tree swallows breed
on both coasts and northward to the tree
line.
Habitat - Prefer open habitat near water,
including fields, marshes, shorelines, and
wooded swamps. Tree swallows are hole
nesters and depend on woodpeckers and
other excavators to furnish nesting cavities.
When defending a nest site, and especially
when feeding nestlings, tree swallows show
the greatest home range tenacity and
typically remain within 100 to 200 m of their
nest site. Large colonies can reach
densities of 150 pairs per 0.7 acre (0.28 ha).
Diet - Tree swallows actively pursue flying
insects, and occasionally glean insects from
the water surface or vegetation. Food items
include mosquitoes, midges, gnats,
mayflies, and beetles.
Life History of the American Robin
The American robin is a common, wide-
ranging North American bird, present
throughout the continental United States
and Canada, excluding the extreme north
and high altitudes.
Habitat - Occupies a wide range of habitat
types, from closed canopy forests to
residential areas. In summer, commonly
observed in cleared areas with short herbs,
such as natural forest openings, lawns, and
recently cleared or burned stands. Territory
is small, ranging from 0.3 to 2.0 acres (0.1
to 0.8 ha).
Diet - Forages on the ground and in
vegetation for invertebrates and fruits;
proportions vary by season. In spring, diet
comprises less than 10% fruits. In fall and
early winter, fruits account for 80% to 99%
of the diet, with the remainder of the diet
invertebrates.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-9
-------
MK0110:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-10
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Copyright0 1999, Don Baccus (dhogaza@pacifier.com)
Figure 7.2-2 American Robin (Turdus migratorius)
The assessment endpoint that is the subject of this section is the survival, growth, and
reproduction of insectivorous birds in the Housatonic River PSA. The measurement endpoints
used to evaluate the assessment endpoint include: (1) determining, by comparisons of modeled
exposure to doses reported in the literature to cause adverse effects, the extent to which the
concentrations of tPCBs and TEQ ingested in the diet will cause adverse effects to the survival,
growth, and reproduction of insectivorous birds; and (2) determining, by conducting field
studies, the relationship between concentrations of tPCBs and TEQ in prey and the reproductive
performance of insectivorous birds in the Housatonic River floodplain. The 3-year tree swallow
field study measured clutch size, hatching success, and tissue concentrations of tPCBs and TEQ
and other variables at three locations in the PSA, and three reference locations. The American
robin field study evaluated the relationship between reproductive success and tissue
concentrations of tPCBs in eggs and young.
MK0110:\20123001.096\ERA_PB\ERA_PB_7.DOC q i i 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
7.3 EXPOSURE ASSESSMENT
The exposure assessment for insectivorous birds focuses on both the PSA and several reference
locations. Trophic transfer and exposure through ingestion of contaminated food items are the
major exposure pathways for insectivorous birds exposed to tPCBs and 2,3,7,8-TCDD TEQ
(TEQ). Other routes of exposure, considered to be negligible contributors to overall exposure,
include inhalation, water consumption, and soil ingestion (Moore et al. 1999). PCBs and TEQ
tend to bioaccumulate in the food chain because:
¦ PCBs and TEQ are persistent, and highly lipophilic and hydrophobic substances.
¦ When released to aquatic systems, these COCs form associations with dissolved
and/or particulate matter in the water column and remain in the sediment;
biodegradation is considered to be a relatively minor fate process in water (NRCC
1981; Howard et al. 1991).
¦ Aquatic sediment provides a sink for these compounds and may represent long-term
sources to the aquatic food web (Kuehl et al. 1987; Muir 1988; Corbet et al. 1983;
Tsushimoto et al. 1982). Both PCBs and TEQ are bioaccumulated by aquatic and
terrestrial biota through the consumption of contaminated prey as part of the food
chain (Haffner et al. 1994; Senthilkumar et al. 2001; Borga et al. 2001).
In summary, insectivorous birds that reside, or partially reside, within the PSA are exposed to
tPCBs and TEQ principally through diet and trophic transfer.
Exposure of tree swallows and American robins to tPCBs and TEQ was estimated using the
standard total daily intake model. Tree swallow exposure was also estimated using an explicit
microexposure model. The total daily intake model was adapted from the Wildlife Exposure
Factors Handbook (EPA 1993) and was also used in the other wildlife assessments. For the
microexposure model, accumulation of tPCBs and TEQ over the first 15 days of the birds'
development was estimated. The results of the two modeling approaches are compared, and the
tissue concentration data collected from tree swallow nestlings in the PSA and reference areas
compared to predicted tissue concentrations from the microexposure model.
The tree swallow field study was performed with nest boxes placed at six locations, three of
which were located downstream of the GE facility within the PSA (Holmes Road, New Lenox
Road, and Roaring Brook). Three other locations (Threemile Pond, Southwest Branch, and
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-12
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
Taconic Valley) were expected to serve as reference locations (Custer 2002). In this field study,
nest boxes were placed adjacent to the river and on the nearby floodplain. A map of the tree
swallow nest box locations is shown in Figure 7.3-1. The reference location at Threemile Pond
(not depicted on the map) is within the Housatonic River drainage, but is upgradient of the river.
This location was expected to be representative of background contaminant concentrations. Two
other reference locations (Southwest Branch and Taconic Valley sites) lie upstream of the major
source of PCB contamination. The exposure assessment focused on the six locations used in the
field study for tree swallows by Custer (2002).
Exposure of American robins was estimated for three areas (Locations 13, 14, and 15; see Figure
7.3-1 and text box) in the PSA.
An exposure assessment was not conducted for American robins in reference areas because
concentrations of COCs in robin prey items were not available from these locations. In addition,
concentrations of COCs in prey in the floodplain reference areas were not estimated because
nearly all sediment samples from this area did not have detectable concentrations of tPCBs,
therefore, it is not likely that the floodplain soil have detectable concentrations of PCBs.
Description of Sites 13, 14, and 15
Location 13 is a relatively flat area on the west shore of the river, in the floodplain
adjacent to river mile 133.2, situated at an elevation of 965 ft (294 m). The
community type is transitional floodplain forest that is flooded seasonally and is
moderately well drained, with extensive vegetation cover (80%) and alluvial silt-loam
soil. The PCB concentrations in floodplain soil averaged 55.2 mg/kg dw.
Location 14 is a relatively flat low-lying area located on the west shore of the river, in
the floodplain adjacent to river mile 129.9, situated at an elevation of 965 ft (294 m).
The community type is transitional floodplain forest that is flooded seasonally with
extensive vegetation cover (70%) and fluvial silt soil. The PCB concentrations in
floodplain soil averaged 26.1 mg/kg dw.
Location 15 is a flat area located on the west shore of the river, in the floodplain
adjacent to river mile 126.7, situated at an elevation of 965 ft (294 m). Community
types are circumneutral hardwood swamp and transitional floodplain forest that are
flooded seasonally. This site has 60% vegetation cover, 40% leaf litter cover, and a
primarily mineral soil. The PCB concentrations in floodplain soil averaged 0.484
mg/kg dw.
MK01|O:\20123001.096\ERA_PB\ERA_PB_7.DOC 7 13
-------
/20\
Reacft^5A
teach 5E
ss:
I
JL
TS (Reach 5C
Reach 5E\<
(backwater areas)!
Reach 6
LEGEND:
<•> Tree Swallow Nest Box
• Soil Invertebrate Sample Location
>v Reach Breaks
H Roads
[^j Hydrography
I | Housatonic River Basin Boundary
N
V
s
2500
0
2500
5000 Feet
0.5
0
0.5
1 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 7.3-1
TREE SWALLOW NEST BOX AND
SOIL INVERTEBRATE SAMPLE
LOCATIONS IN THE
HOUSATONIC RIVER PSA
| o:\gepitl\aprs\swallows.apr | Layout - swallow/soilinvert Iocs 7.3-11 o:\gepitt\plots\in\tsnb_sis_locs_7-3-1.eps 110:21 AM, 7/3/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
This section begins with a description of the standard total daily intake (TDI) model used for tree
swallows and American robins. For tree swallows, the TDI model was used as the basis for
developing a microexposure model to estimate accumulation of tPCBs in tissues of nestlings
from hatch to 14 days post-hatch. Subsequent sections describe the inputs used in the exposure
analyses. The section concludes with a description of the Monte Carlo and probability bounds
analyses conducted to estimate exposure of tree swallows and American robins to tPCBs and
TEQ in the Housatonic River PSA and reference areas.
7.3.1 Exposure Models for Insectivorous Birds
7.3.1.1 Total Daily Intake Model
Exposure of the representative species, tree swallows and American robins, to tPCBs and TEQ
was estimated using a total daily intake model adapted from the Wildlife Exposure Factors
Handbook (EPA 1993) and related publications. The model used in the exposure analysis was:
n
TDI = FTx FIRYj C, x I] (Eq. 1)
z = l
where
TDI = Total daily intake (mg/kg bw/d tPCBs, ng/kg bw/d TEQ).
FT = Foraging time in the area of interest (unitless and set equal to one for
insectivorous birds).
FIR = Normalized food intake rate (kg/kg bw/d).
Pi = Proportion of the z'th dietary item (unitless).
Ci = Concentration in food item (mg/kg for tPCBs, ng/kg for TEQ).
The model considers the food intake rates of representative species (FIR), the concentrations of
COCs in each food item (Ci), and the proportion of diet accounted for by that food item (Pi). For
those input variables that are uncertain, variable, or both, distributions are used rather than point
estimates. Monte Carlo and probability bounds analyses are used to propagate uncertainties
about input variables through the exposure model for each COC. A description of these
techniques and the methods used to parameterize input variables is presented in Section 6.5 and
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-15
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Appendix C.4. The results of the Monte Carlo analysis are used to estimate the probability of
exposure exceeding an effects threshold or doses that cause adverse effects of differing
magnitudes. The probability bounds analysis is conducted to determine how uncertainty
regarding the distributions of the input variables influences the estimated exposure distribution.
The results of these analyses are discussed in detail in Appendix G.
7.3.1.2 Microexposure Model
The exposure model is a simple model driven by the concentration of COCs in the diet and by
the food intake rate. In this section, a more complex, dynamic microexposure model is derived
for the tree swallow exposure assessment. This model determines the whole body contaminant
content in the swallows as a function of time over the first 15 days of their development.
A simulation length of 15 days was chosen because swallows reach adult body weight at
approximately this time. In addition, nestling swallow tissue samples were collected between 12
and 14 days in the Custer (2002) study.
The microexposure model equation for whole body tPCB and TEQ tissue concentrations at 15
days {TOT BURDEN) is given below:
TOT BURDEN = (TOT MT + Zt [(FIR , x Cdiet) x 1 d/j /BW (Eq. 2)
where:
TOT BURDEN = Total amount of COCs accumulated by tree swallows from maternal
transfer plus their first 15 days of food consumption per gram of
body weight (mg/kg for tPCBs, ng/kg for TEQ).
TOT MT = COC tissue concentration (|Lxg) from maternal transfer.
i = Refers to the time steps in the simulation (range from 0 to 14 days1).
FIR, = Food intake rate (g/d) calculated using a variable body weight
specific to each day of the simulation.
1 Note that the first day of the simulation is day zero, and the last day of the simulation is day 14; therefore, the total
simulation length is 15 days.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-16
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Cdiet = Concentration in food item (mg/kg for tPCBs, ng/kg for TEQ).
BW = Body weight at day 14 (g).
All of the subscripts i refer to the time steps in the simulation, which range from 0 to 14 days.
Qiet (mg/kg) is the same as C, in the TDI model. Intake rate of food (FIRi, measured in g/d) was
calculated in the same manner as in the TDI model (see Section G.2.1.3.2), except the body
weight variable was made specific to each day of the simulation. Because the value of food
intake rate is a function of body weight in the microexposure model, it is subscripted by day to
account for the effect of changing body weight over the simulation. The intake rate was
converted to the amount consumed on each day i by multiplying through by one day (1 d). The
term TOT MT represents the COC tissue concentration from maternal transfer (see Section
G.2.1.3.2). The sum of maternal transfer and the expression in the summation was divided at the
end of the simulation by body weight at day 14 (BW). Thus, TOT BURDEN estimates the total
amount of COCs accumulated by tree swallows from maternal transfer plus their first 15 days of
food consumption per kilogram of body weight.
Two issues often arise when calculating a TEQ concentration in prey:
¦ Congener concentrations may be below the method detection limit (i.e., non-detects).
¦ Some congeners may not be resolved due to co-elution during analysis.
An approach was developed to address these issues and is presented in Section 6.4 and Appendix
C.2). Briefly, congeners detected at or below the detection limit (DL) were included in the TEQ
calculations by investigating three options:
¦ Setting the value for the congener equal to zero (0).
¦ Setting it to half the DL.
¦ Setting it equal to the DL.
A comparison of the results of this bounding analysis provides a description of the uncertainty
surrounding the TEQ value due to concentrations of one or more congeners being below the
detection limit.
To resolve the co-elution issue, the concentrations of coplanar (TEQ) congeners that co-eluted
with other congeners were assumed to be equal to the total concentration of the co-elutes (likely
overestimate of TEQ concentration) or zero (likely underestimate of TEQ concentration). The
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q -> q 7/10/2003
-------
1
2
3
decision criteria in Section 6.4 were followed to address the uncertainty arising from co-elution
or non-detection of congeners when estimating exposure point concentrations (EPCs) for use in
the exposure analyses.
4 Input distributions to the exposure analyses were generally assigned as follows:
5
6
7
¦ Lognormal distributions for variables that were right skewed with a lower bound of
zero and no upper bound (e.g., amount of COC transferred from mother to offspring
via egg tissue for tree swallow).
9
8
¦ Beta distributions for variables bounded by zero and one (e.g., proportion of a prey
item in the diet).
10
11
¦ Normal distributions for variables that were symmetric and not bounded by one (e.g.,
body weight).
12
¦ Point estimates for minor variables or variables with low coefficients of variation.
13 In certain situations (e.g., poor fit of data), other distributions were fit to the data or other
14 approaches were used. To quantify uncertainty, two approaches were used as described in
15 Section 6.5.2. The distributions used in the exposure analyses for tree swallows and American
16 robins are shown in Figures 7.3-2, 7.3-3, and 7.3-4 and are described in the following sections.
17 These distributions are also presented in greater detail in Appendix G.
18 Foraging Time
19 The foraging ranges of the two representative species are within the area of the PSA. Prey
20 availability and an abundance of suitable foraging habitat suggest that the birds that forage in the
21 PSA are able to meet their needs exclusively within this section of the river and floodplain. The
22 assessment of risk to insectivorous birds inhabiting the PSA of the Housatonic River will,
23 therefore, focus on those birds that spend 100% of their time foraging within the PSA. Foraging
24 time is a point estimate; therefore, it is not shown in Figures 7.3-2, 7.3-3, and 7.3-4.
25
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-18
7/10/2003
-------
1
Body We ight
FMR Slope Term (a)
1
0.8
0.6
0.4
0.2
0
10
20 25
Weight (g)
30
35
0.2 0.4 0.6
0.8 1 1.2 1.4 1.6
Log (a)
1
0.8
0.6
FMR PowerTerm (b)
0.2
0.4
0.6
b
3 Figure 7.3-2 TDI Exposure Model Input Distributions for Tree Swallows
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-19
-------
Maternal Transfer at Holmes Road
Maternal Transfer at Southwest Branch
100 150 200
Weight (fig)
250 300
100 150 200
Weight Qig)
250 300
Maternal TransferatNewLenoxRoad
Maternal Transferat Three mile Pond
£
Is 1
100 150 200
Weight (jig)
250 300
100 150 200
Weight (jag)
250 300
*
1 1
Maternal Transferat Roaring Brook
100 150 200 250 300
Weight (jig)
Maternal Transferat Taconic Valley
50 100 150 200 250 300
Weight Qig)
Figure 7.3-3 Microexposure Model Input Distributions for Maternal Transfer for
Tree Swallows
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-20
-------
Bodyweight distribution
Body Weight (g)
FMR slope term (a)
FMR power term {b)
Proportion earthworms in diet
2 Figure 7.3-4 TDI Exposure Model Input Distributions for American Robin
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-21
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Body Weight (BW)
TDI Model
Tree swallows are small birds with an average adult body weight of about 20 g. Body weight can
be as low as 16.5 g when food availability is low, and as high as 25.5 g for females during the
mating season (Robertson et al. 1992). Newly hatched nestlings weigh about 1.5 g and achieve
adult weight in about 14 days. Based on data cited in Dunning (1984), the mean adult body
weight was estimated to be 20.1 g and standard deviation of 1.58 (Figure 7.3-2).
American robins are a sexually monomorphic species with similar male and female body weights
(Clench and Leberman 1978; Wheelwright 1986; Marcum et al. 1998). Robins monitored in
Delta Marsh, Manitoba, Canada, ranged from 72 to 86 g, with females gaining slightly more
weight during the incubation period (Bierman and Sealy 1985). Clench and Leberman (1978)
found an average mass of 77 g when data from both sexes were pooled. The distribution of the
body weights of American robins is depicted in Figure 7.3-4.
Microexposure Model
The body weight of juvenile birds was modeled as a function of time, leveling off at 12 to 14
days, using a logistic model from Teather (1996). In the Monte Carlo version of the
microexposure model, point values were used for body weight. In the probability bounds
analysis, the uncertainty surrounding body weight as a function of time was taken into account.
Tables G.2-4 and G.2-5 show the body weight distributions used in the Monte Carlo and
probability bounds analyses, respectively.
Food Intake Rate (FIR)
TDI Model
The food intake rate of tree swallows and American robins has not been well characterized. The
food intake rate for American robins has been measured in captive animals (Hazelton et al.
1984). Because the animals were captive, the measured food intake rates likely underestimated
food intake rates of free-living robins (EPA 1993). Free-living robins, unlike captive robins,
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 7 O O 7/10/2003
-------
1
2
3
expend energy foraging for prey, avoiding predators, defending territories, etc. As a result, an
allometric modeling approach, described below, was used to estimate food intake rate for
American robins rather than the rates measured in the controlled study.
4 Nagy (1987) and Nagy et al. (1999) derived allometric equations for estimating the metabolic
5 rate of passerine birds using the following general equation:
7 The slope (a) and power (b) distributions were based on the error statistics reported in Nagy et al.
8 (1999) and our own analyses of the raw data, assuming an underlying normal distribution for
9 each. For passerine birds, the mean slope term for log a is equal to 1.02 with a standard error of
10 0.0883 in logio units. The power term b had a reported mean of 0.680 and a standard error of
11 0.0682. The body weight (BW) distribution was described above. The results of the calculation
12 were then converted to kcal/kg bw/d.
13 Food intake rate is derived from FMR using the following equation:
15 where AE is the assimilation efficiency of invertebrates (unitless) by birds and GE is the gross
16 energy of invertebrates (kcal/g). The gross energies of various wildlife food sources are
17 summarized in the Wildlife Exposure Factors Handbook (EPA 1993). For tree swallows, the
18 mean assimilation efficiency was 77% (SD = 8.4%)(Karasov 1990; EPA 1993). The mean gross
19 energy for grasshoppers and crickets is 17 kcal/g (SD = 260) (Cummins and Wuycheck 1971;
20 Collopy 1975; Bell 1990), and for adult beetles, the mean is 1.5 kcal/g (Cummins and Wuycheck
21 1971; Collopy 1975; Bell 1990). Grasshoppers, crickets, and beetles were used as
22 representatives of emergent aquatic insects; their mean gross energy is 1.6 kcal/g. For American
23 robins, mean assimilation efficiencies were 72% for both earthworms and litter invertebrates, and
24 the mean gross energies were 0.81 kcal/g wet weight for earthworms and 1.6 kcal/g wet weight
25 for litter invertebrates. Point estimates were used for these variables in the Monte Carlo and
26 probability bounds analyses because of their relatively small coefficients of variation (i.e., CV
27 <10%>). As a result, these input variables are not included in Figures 7.3-2, 7.3-3, and 7.3-4.
6
FMR (kJ/day) = axBW(g)b
(Eq. 3)
14
(Eq. 4)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-23
7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Microexposure Model
The allometric relationship for food intake rate was modeled as a function of body weight. This
was recalculated as the microexposure model was stepped forward in time over the life of the
swallow. For each day, the corresponding point estimate of body weight (in the case of the
Monte Carlo analysis) or interval estimate of body weight (in the case of the probability bounds
analysis) was substituted into the allometric equation for food intake rate. As with the intake
rates used in the TDI model, the allometric model equation variables (the slope and power terms)
were modeled as normally distributed.
Proportion of Dietary Items (P,)
An analysis of the diet delivered to swallow nestlings indicated that it consisted of Diptera
(45.9%), mayflies (15.6%), and other insects (8.7%) by number, and Diptera (41.8%), mayflies
(21.3%>), and moths and butterflies (9.2%) by total dry mass (Blancher and McNicol 1991). A
separate study also showed that mayflies and Diptera were common prey for swallows
(Robertson et al. 1992). Consumption of contaminated aquatic insects is presumed to be the
primary route of exposure of swallows to tPCBs. Direct stomach content samples were taken
from birds at nest sites within the Housatonic River PSA and at reference sites (Custer 2002).
These samples were used as the primary source of contaminant input concentrations for the
exposure models developed in this assessment.
American robin diets for the spring and summer were used in this assessment, because the focus
was on estimating reproductive effects to robins. Proportions of each prey item in the diet were
assumed to follow a beta distribution in the Monte Carlo analysis. For robins, the available
literature indicates that earthworms comprise about 15% of the diet on average during the spring
and summer, but may range from 10 to 20%. Litter invertebrates generally comprise about 60%
of the diet during spring and summer, with an approximate range of 45 to 75%. The proportion
of fruit in the robin diet during spring and summer was calculated as one minus the total of
earthworms and litter invertebrates.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-24
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Maternal Transfer (TOT_MT)
Maternal transfer refers to the total amount of COCs (in |Lxg) transferred from the mother to the
offspring via egg tissue. Maternal transfer was estimated at each site using the concentrations of
COCs in the tissues of the newborn swallows (pippers) and egg samples. Table G.2-6
summarizes the ratios of the means of tPCB concentrations in eggs and pippers (due to maternal
transfer only) to the tPCB concentrations in nestlings (aged 12 to 14 days). Low ratios indicate
that the majority of tPCB tissue concentrations originated from feeding activity at the site.
Ratios approaching or greater than one indicate that the tPCB content was primarily due to
maternal transfer, and not from feeding locally over the period from birth to 14 days.1 High
ratios of tPCB concentrations in pippers and eggs to total concentrations in nestlings occurred at
sites expected to be relatively uncontaminated locations. Those locations included Threemile
Pond, Taconic Valley, and Southwest Branch. At these locations, the ratios ranged from 0.57 at
the Taconic Valley in 1999 to 1.89 at Threemile Pond in 2000. In contrast, the ratios at the more
contaminated locations ranged from 0.13 at Roaring Brook in 1998 to 0.60 at the same location
in 1999 (see Table G.2-6).
For the Monte Carlo simulation, maternal transfer was assumed to have a lognormal distribution,
with location-specific mean and standard deviation. For the probability bounds analysis,
probability bounds were derived using the site-specific lower 95% confidence limit (LCL) and
the upper 95% confidence limit (UCL) around the means calculated using the Land H-statistic.
Concentrations of COCs in Food Items (Cf)
The concentrations of tPCBs and TEQ in the dietary items of insectivorous birds are illustrated
in Figures 7.3-5 to 7.3-8. In these figures, the box and star symbols depict the median and
arithmetic mean concentration of each COC in each dietary item in each of the areas for the risk
assessment. The vertical line depicts the interquartile range for the concentration. The
concentrations of COCs used in the exposure analyses are shown in Tables G.2-7, G.2-8, G.2-27,
and G.2-28 of Appendix G. Total PCB concentrations in the prey of tree swallows are similar at
1 Ratios greater than one, observed at Southwest Branch in 1998, Threemile Pond in 1999 and 2000, and Taconic
Valley in 2000, would seem to indicate depuration or growth dilution over the 14 days of growth. These ratios are
considered indicative of very high proportions of maternal transfer relative to intake of local contaminants.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC ry ~ r 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Holmes Road, New Lenox Road, and Roaring Brook. All of the reference locations have lower
tPCB concentrations than sites in the PSA. TEQ concentrations in prey items of tree swallows
are highest at Holmes Road and lowest at the Taconic Valley reference area. Both tPCB and
TEQ concentrations in the prey items of American robins are highest at Location 13 and lowest
at Location 15.
Consumption of contaminated emergent aquatic insects is presumed to be the primary route of
exposure of swallows to tPCBs. Soil, water, and sediment exposure was not considered because
tree swallow exposure via these pathways was determined to be extremely limited. Both the TDI
model and the microexposure model used stomach content data collected from juvenile swallows
during the Custer (2002) study. The microexposure model also used tissue data to estimate the
proportion of tPCB and TEQ concentrations in offspring tissue due to maternal transfer.
Available data on contaminant concentrations in benthic invertebrates were not used because of
the availability of stomach content samples, which provide a more direct measure of
concentrations of COCs in tree swallow food.
Direct stomach content samples were taken from birds at nest sites within the Housatonic River
PSA and at reference sites. Stomach content samples were collected in 1998, 1999, and 2000:
analytical results are included in Custer (2002). The median concentration of tPCBs in stomach
contents measured at Holmes Road was 3.24 mg/kg. The 25th and 75th percentiles were 2.56 and
10.7 mg/kg, respectively. The median, 25th, and 75th percentile concentrations of TEQ were 996,
669, and 1,324 ng/kg, respectively. Median tPCB concentrations in stomach contents at Holmes
Road, New Lenox Road, and Roaring Brook were approximately one order of magnitude higher
than at reference locations, except for Taconic Valley, which had comparable concentrations to
the PSA locations. Median TEQ concentrations in stomach contents at all locations, including
reference locations, were in the same order of magnitude. Statistics for concentrations of tPCBs
and TEQ in stomach contents at the six locations are shown in Tables G.2-7 and G.2-8.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-26
-------
30
25
St
20
15
a
o
U
09
O
Ph
10
~ Median
X Arithmetic Mean
I 75 Percentile
25th Percentile
Holmes Road New Lenox Road Roaring Brook
Southwest Threemile Pond Taconic Valley
Branch
1
2
Location
Figure 7.3-5 Concentration of tPCBs in Prey of Tree Swallows
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-27
-------
1400
1
2
1200
1000
If
"3d
£
S 800
g 600
u
o
W
H
400
200
~ Median
X Arithmetic Mean
75 Percentile
25th Percentile
Holmes Road New Lenox Roaring Brook Southwest Threemile Pond Taconic Valley
Road Branch
Location
Figure 7.3-6 Concentration of TEQ in Prey of Tree Swallows
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-28
-------
30
St
=
o
o
09
O
Ph
25
20
15
10
~ Median
X Arithmetic Mean
75 Percentile
25th Percentile
1
2
N#
J?
<
-------
1
2
3
4
5
6
7
8
9
10
11
12
1400
1200
St
~Bjd
£
=
o
1000
800
600
o
W
H 400
200
N#
3*
~ Median
X Arithmetic Mean
175 Percentile
125th Percentile
-------
Holmes Road
100
80
a
«
a
©
a.
4>
w
=
¦a
0)
0)
w
H
H
60
40
20
Monte Carlo
LPB
UPB
10 20
Dose (mg/kg bw/d)
30
40
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-9 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte
6 Carlo Analysis and Probability Bounds Analysis at Holmes Road
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-31
-------
New Lenox Road
0 10 20 30 40
Dose (mg/kg bw/d)
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5
6 Figure 7.3-10 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte
7 Carlo Analysis and Probability Bounds Analysis at New Lenox
8 Road
9
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-32
-------
Roaring Brook
0 10 20 30 40
Dose (mg/kg bw/d)
1
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-11 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte
6 Carlo Analysis and Probability Bounds Analysis at Roaring Brook
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
-------
Southwest Branch
©
a.
4>
w
=
¦a
0)
0)
w
H
H
100
80
60
40
20
¦Monte Carlo
-UPB
-LPB
10 20
Dose (mg/kg bw/d)
30
40
2
3
4
5
6
7
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-12
Tree Swallow TDI Exposure Model for tPCBs: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Southwest
Branch
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-34
-------
Three mile Pond
0 10 20 30 40
Dose (mg/kg bw/d)
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-13 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte
6 Carlo Analysis and Probability Bounds Analysis at Threemile
7 Pond
8
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-35
-------
Taconic Valley
0 10 20 30 40
Dose (mg/kg bw/d)
1
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-14 Tree Swallow TDI Exposure Model for tPCBs: Results of Monte
6 Carlo Analysis and Probability Bounds Analysis at Taconic Valley
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-36
-------
Holmes Road
o
Monte Carlo
LPB
— UPB
• Low-interm. criterion
O Interm. - high criterion
O
0
5,000
10,000 15,000 20,000 25,000
Dose (ng/kg bw/d)
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-15 Tree Swallow TDI Exposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Holmes Road
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 7^7 7/10/2003
-------
New Lenox Road
Monte Carlo
LPB
UPB
• Low- interm. criterion
O Interm. - high criterion
5,000
10,000 15,000
Dose (ng/kg bw/d)
20,000
25,000
2
3
4
5
6
7
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-16
Tree Swallow TDI Exposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at New Lenox
Road
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-38
-------
100 in
80
©
a.
4>
w
=
¦a
0)
0)
w
H
H
60
40
20
Roaring Brook
— Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
5,000
10,000 15,000
Dose (ng/kg bw/d)
20,000
25,000
2
3
4
5
6
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-17
Tree Swallow TDI Exposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Roaring Brook
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-39
-------
Southwest Branch
o
• Low- interm. criterion
O Interm.-high criterion
Monte Carlo
LPB
UPB
O
0
5,000
10,000 15,000 20,000 25,000
Dose (ng/kg bw/d)
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-18 Tree Swallow TDI Exposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Southwest
Branch
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 1 A f\ 7/10/2003
-------
Three mile Pond
100
— Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm. - high criterion
5,000
10,000 15,000
Dose (ng/kg bw/d)
20,000
25,000
2
3
4
5
6
7
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-19
Tree Swallow TDI Exposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Threemile
Pond
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-41
-------
Taconic Valley
100
Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
_i i
5,000
10,000 15,000
Dose (ng/kg bw/d)
20,000
25,000
2
3
4
5
6
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-20
Tree Swallow TDI Exposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Taconic Valley
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-42
-------
Site 13
Dose (mg/kg bw/d)
1
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-21 Exposure of American Robins to tPCBs at Site 13 of the
6 Housatonic River PSA
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-43
-------
Site 14
Dose (mg/kg bw/d)
1
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-22 Exposure of American Robins to tPCBs at Site 14 of the
6 Housatonic River PSA
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-44
-------
Site 15
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-23 Exposure of American Robins to tPCBs at Site 15 of the
6 Housatonic River PSA
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-45
-------
Site 13
1
2
3
4
5
6
-O
-O
o
e
¦a
100
80
60
40
8 20
x
W
1 10
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
^"Monte Carlo
LPB
UPB
# Low-intermed.
crtierion
O Interned.-high
criterion
100 1000
Dose (ng/kg bw/d)
10000
100000
Figure 7.3-24
Exposure of American Robins to 2,3,7,8-TCDD TEQ at Site 13 of
the Housatonic River PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-46
-------
Monte Carlo
LPB
UPB
# Low-intermed.
criterion
O Interned.-high
criterion
O
1 10 100 1000 10000 100000
Dose (ng/kg bw/d)
1
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-25 Exposure of American Robins to 2,3,7,8-TCDD TEQ at Site 14 of
6 the Housatonic River PSA
7
Site 14
100
80
.o
C3
-C
O
¦-
a.
CJ
a
es
¦a
o
o
cj
x
w
60
40
20
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-47
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Site 15
100
'Monte Carlo
LPB
UPB
9 Low-intermed.
criterion
O Interned.-high
criterion
O
0
' ' '
10
100
1000
10000
100000
Dose (ng/kg bw/d)
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-26 Exposure of American Robins to 2,3,7,8-TCDD TEQ at Site 15 of
the Housatonic River PSA
The probability bounds estimated for tree swallows foraging at Holmes Road are depicted in
Figure 7.3-9. The 10th percentile of the probability envelope formed by the lower and upper
bounds ranged between 1.04 and 18.2 mg/kg bw/d. The 50th percentile ranged between 1.27 and
22.1 mg/kg bw/d, and the 90th percentile ranged between 1.81 and 32.3 mg/kg bw/d. In
comparison, the 10th percentile of the Monte Carlo output was 15.3, the 50th percentile was 22.0,
and the 90th percentile was 32.1 mg/kg bw/d (Table G.2-9).
Exposures of tree swallows to tPCBs at two other PSA locations, New Lenox Road and Roaring
Brook, were lower than at Holmes Road, having mean total daily intakes of 11.0 and 13.3 mg/kg
bw/d, respectively (Table G.2-9). Exposures at the reference locations, Southwest Branch and
Threemile Pond, were very low with mean total daily intakes of 0.73 and 1.15 mg/kg bw/d. The
third reference location, Taconic Valley, had relatively high exposure with a mean of 14.2 mg/kg
bw/d. The highest concentrations of TEQ were at Holmes Road with a mean of 1,227 ng/kg
bw/d. New Lenox Road and Roaring Brook had mean total daily intakes of 701 and 681 ng/kg
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 1 AQ 7/10/2003
-------
1 bw/d, respectively (Table G.2-10). Mean exposure concentrations at the three reference
2 locations ranged from 396 to 866 ng/kg bw/d.
3 American robins had the highest exposure to tPCBs and TEQ at Location 13, with mean
4 concentrations of 5.66 mg/kg bw/d and 82.3 ng/kg bw/d, respectively. Exposure at Locations 14
5 and 15 was somewhat lower for both tPCBs and TEQ. Mean tPCB concentrations at Locations
6 14 and 15 were 4.69 and 1.22 mg/kg bw/d, respectively. Mean TEQ concentrations at Locations
7 14 and 15 were 57.5 and 33.0 ng/kg bw/d, respectively.
8 7.3.3 Microexposure Model Results
9 Exposure distributions for exposure of tree swallows to tPCBs and TEQ at Holmes Road, New
10 Lenox Road, Roaring Brook, Southwest Branch, Threemile Pond, and Taconic Valley are
11 presented in Figures 7.3-27 to 7.3-38.
12 Figure 7.3-27 depicts the cumulative distribution of tPCB intake rates for tree swallows at
13 Holmes Road. The Monte Carlo analysis indicated that accumulation of tPCBs by tree swallows
14 at Holmes Road could range from a minimum of 75.8 mg/kg in tissues of 14 d nestlings (5.06
15 mg/kg bw/d)1 to a maximum of 595 mg/kg (39.7 mg/kg bw/d). The mean exposure was 222
16 mg/kg (14.8 mg/kg bw/d), and the median exposure was 215 mg/kg (14.3 mg/kg bw/d); 80% of
17 the exposure estimates were between 154 (10.3) and 253 (15.6) mg/kg.
1 Estimated by dividing tissue residues by 15, which was the number of days included in the microexposure model.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-49
-------
Holmes Road
©
a.
4>
w
=
¦a
0)
0)
w
H
H
1
2
3
4
100
80
60
40
20
Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
0 50
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
100 150 200
Tissue Residue (mg/kg)
250
300
5 Figure 7.3-27 Tree Swallow Microexposure Model for tPCBs: Results of Monte
6 Carlo Analysis and Probability Bounds Analysis at Holmes Road
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-50
-------
New Lenox Road
0 50 100 150 200 250 300
Tissue Residue (mg/kg)
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-28 Tree Swallow Microexposure Model for tPCBs: Results of Monte
6 Carlo Analysis and Probability Bounds Analysis at New Lenox
7 Road
8
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-51
-------
Roaring Brook
100
80
a
«
a
©
a.
-------
Southwest Branch
©
a.
4>
w
=
¦a
0)
0)
w
H
H
100
80
60
40
20
Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
1
2
3
4
5
6
7
50
100 150 200
Tissue Residue (mg/kg)
250
300
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-30
Tree Swallow Microexposure Model for tPCBs: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Southwest
Branch
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-53
-------
Three mile Pond
100
80
a
«
a
©
a.
i/
w
=
¦a
0)
0)
S 20
H
60
40
— Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
>o
J I L.
-I I
50
100 150 200
Tissue Residue (mg/kg)
250
300
2
3
4
5
6
7
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-31
Tree Swallow Microexposure Model for tPCBs: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Threemile
Pond
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-54
-------
Taconic Valley
0 50 100 150 200 250 300
Tissue Residue (mg/kg)
1
2 Notes:
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 7.3-32 Tree Swallow Microexposure Model for tPCBs: Results of Monte
6 Carlo Analysis and Probability Bounds Analysis at Taconic Valley
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-55
-------
Holmes Road
©
a.
4>
w
=
¦a
0)
0)
w
H
H
100
80
60
40
20
Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
O
LLLil
10 100 1,000 10,000
Tissue Residue (ng/kg)
100,000 1,000,000
2
3
4
5
6
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-33
Tree Swallow Microexposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Holmes Road
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-56
-------
New Lenox Road
1
2
3
4
5
6
7
a
«
a
©
a.
4>
w
=
¦a
0)
0)
w
H
UJ
100
80
60
40
20
Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
uiu i
o
iiLl
10
100 1,000 10,000
Tissue Residue (ng/kg)
100,000 1,000,000
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-34
Tree Swallow Microexposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at New Lenox
Road
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-57
-------
Roaring Brook
1
2
3
4
5
6
a
«
a
©
a.
4>
w
=
¦a
0)
0)
w
H
H
100
80
60
40
20
—Monte C arlo
" LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
•
O
100 1,000 10,000
Tissue Residue (ng/kg)
100,000 1,000,000
1 10
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-35 Tree Swallow Microexposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Roaring Brook
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-58
-------
Southwest Branch
1
2
3
4
5
6
7
100
& 80
a
«
a
©
a.
4>
w
=
¦a
0)
0)
w
H
H
60
40 -
20
10
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Monte Carlo
LPB
UPB
# Low-interm. criterion
O Interm.-high criterion
100 1,000 10,000
Tissue Residue (ng/kg)
100,000 1,000,000
Figure 7.3-36
Tree Swallow Microexposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Southwest
Branch
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-59
-------
Three mile Pond
1
2
3
4
5
6
7
a
«
a
©
a.
4>
w
=
¦a
0)
0)
w
H
UJ
100
80
60
40
20
Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm.-high criterion
LU|_
LUL.
o
LUl_
LUl
100 1,000 10,000
Tissue Residue (ng/kg)
1 10 100 1,000 10,000 100,000 1,000,000
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-37 Tree Swallow Microexposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Threemile
Pond
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-60
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
a
«
a
o
•~
a.
4>
w
=
¦a
0)
0)
w
H
H
100
80
60
40
20
Taconic Valley
— Monte Carlo
LPB
UPB
• Low-interm. criterion
O Interm. - high criterion
lllL.
10
o
Lil|_
LLLLl
100 1,000 10,000
Tissue Residue (ng/kg)
100,000 1,000,000
Notes:
LPB = Lower probability bound
UPB = Upper probability bound
Figure 7.3-38
Tree Swallow Microexposure Model for TEQ: Results of Monte
Carlo Analysis and Probability Bounds Analysis at Taconic Valley
The probability bounds estimated for tree swallows foraging at Holmes Road are depicted in
Figure 7.3-27. The 10th percentile of the probability envelope formed by the lower and upper
bounds ranged between 11.5 mg/kg (0.77 mg/kg bw/d) and 215 mg/kg (14.3 mg/kg bw/d). The
50th percentile ranged between 12.6 mg/kg (0.84 mg/kg bw/d) and 227 mg/kg (15.2 mg/kg
bw/d), and the 90th percentile ranged between 14.4 mg/kg (0.96 mg/kg bw/d) and 273 mg/kg
(18.2 mg/kg bw/d). In comparison, the 10th percentile of the Monte Carlo output was 154 mg/kg
(10.3 mg/kg bw/d), the 50th percentile was 215 mg/kg (14.3 mg/kg bw/d), and the 90th percentile
was 253 mg/kg (15.6 mg/kg bw/d).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-61
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Accumulation of tPCBs by tree swallows at the other two PSA sites, New Lenox Road and
Roaring Brook, was lower than at Holmes Road, having mean tissue residues of 109 mg/kg (7.26
mg/kg bw/d) and 133 mg/kg (8.85 mg/kg bw/d), respectively (Tables G.2-13 and G.2-14).
Accumulation at the reference locations at Southwest Branch and Threemile Pond was very low
with mean tissue residues of 8.44 mg/kg (0.56 mg/kg bw/d) and 11.7 mg/kg (0.78 mg/kg bw/d).
The third reference location, Taconic Valley, had a relatively high accumulation of tPCBs with a
mean of 137 mg/kg (9.10 mg/kg bw/d). The highest tissue residues of TEQ were at Holmes
Road with a mean of 11,580 ng/kg (772 ng/kg bw/d). New Lenox Road and Roaring Brook had
mean tissue residues of 6,644 ng/kg (443 ng/kg bw/d) and 6,466 ng/kg (431 ng/kg bw/d),
respectively (Tables G.2-15 and G.2-16). Mean tissue residues at the three reference locations
ranged from 3,783 ng/kg (252 ng/kg bw/d) to 8,224 ng/kg (548 ng/kg bw/d). The results of the
microexposure model, when converted to daily intake units of tPCBs, correspond fairly closely
with the results of the TDI models.
7.3.4 Tree Swallow Tissue Data from Field Study
7.3.4.1 Tissue Residue Data
Analyses of the COC concentrations in tree swallow tissues provided direct measures of tPCB
and TEQ exposure. Tissue samples were collected from eggs or just hatched young (herein
referred to as pippers) and 12- to 14-day-old nestlings in 1998 at four sites and in 1999 and 2000
at six sites (Custer 2002). These samples provided tPCBs concentrations data for pippers and
12- to 14- day old nestlings (i.e., pre-migratory full-grown birds) and TEQ concentrations for
pippers.
7.3.4.2 12- to 14-day-old Bird Tissue Concentrations
The microexposure model estimated the whole body concentrations of tPCBs in 14-day-old
nestlings. These estimates can be compared directly to the tPCB concentrations measured in 12-
to 14-day-old swallow tissue samples in 1998, 1999, and 2000 in the Custer (2002) study.
Summary statistics by location are in Table G.2-20 (Custer 2002). The highest median
concentrations, observed at Holmes Road, New Lenox Road, and Roaring Brook, ranged from
21.5 to 36.0 mg/kg ww during the 3-year period. Median concentrations at the upstream and the
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q ^ 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
other reference locations (Threemile Pond) ranged from 0.77 to 3.30 mg/kg ww. Total PCB
concentrations typically found in the nestling tissue data at Holmes Road, New Lenox Road, and
Roaring Brook indicated that there was an accumulation above the concentrations found in the
eggs and pippers at these locations.
The nestling tissue data were compared to the results of the microexposure model directly
because the microexposure model estimated whole body concentrations of tPCBs (mg/kg) for
nestlings aged 14 days and the tissue samples were taken from nestlings aged 12 to 14 days.
Because of the maternal transfer term in the microexposure model, the tissue data and the
modeled estimates for exposure are not independent. The Kolmogorov-Smirnov test for
goodness of fit (Miller 1956) was used to compare the cumulative frequency distributions of
tPCBs estimated by the microexposure model and those observed in the tissue sample data at
each location. For all locations, the measured tissue concentration distributions were
significantly lower than the distributions estimated by the Monte Carlo analysis of the
microexposure model. This is likely partially due to the dissimilar exposure durations between
the location-specific data and modeled data. However, all measured distributions were within
the microexposure model probability bounds, as expected.
7.3.4.3 Pipper Tissue Concentrations
Total PCBs
Total PCB concentrations were generally higher in tree swallow pippers from locations in the
Housatonic River PSA than at the reference locations (Table G.2-22). Median concentrations in
pippers ranged from 44.9 to 80.2 mg/kg ww. In comparison, median concentrations ranged from
7.5 to 14.4 mg/kg ww in pippers at the reference sites. Concentrations of tPCBs were generally
consistent among years at each site. However, pippers at Roaring Brook had higher
concentrations in 1999 than in 1998. Also, pippers from Southwest Branch had lower tPCB
concentrations in 2000 than in 1998 and 1999 (Custer 2002).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-63
-------
1 TEQ
2 Concentrations of TEQ were highest at sites in the Housatonic River PSA. The median
3 concentrations in pipper samples ranged from 1,390 ng/kg ww at Holmes Road to 2,890 ng/kg at
4 Roaring Brook. In comparison, median concentrations ranged from 562 to 730 ng/kg ww in
5 pippers at the reference sites (Table G.2-23) (Custer 2002).
6
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-64
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
7.4 EFFECTS ASSESSMENT
The effects assessment has two objectives. The first is to review the scientific literature for
effects of tPCBs (mainly Aroclor 1254 and 1260 mixtures) and TEQ to insectivorous birds. Of
primary interest are documented effects to the representative species in this assessment: tree
swallows and American robins. In the absence of data for these species, other avian species are
considered. The other objective is to derive the effects metrics that will be used, in conjunction
with the exposure assessment results, to estimate risks to insectivorous birds from exposure to
COCs in the Housatonic River PSA.
This section focuses on effects that have an influence on the long-term maintenance of bird
populations (i.e., mortality or impairment of reproduction or growth). Studies involving multiple
exposure treatments and where reported results were evaluated statistically to identify significant
differences from controls are preferred.
The COCs in this assessment are tPCBs and 2,3,7,8-TCDD TEQ. The congeners used to estimate
TEQ concentration share the ability to bind to the Ah receptor protein (Bosveld and Van den
Berg 1994) and elicit an Ah-mediated biochemical and toxic response. The toxic response to this
group of chemicals is related to the three-dimensional structure of the molecule, including the
degree of chlorination and positions of the chlorine on the aromatic frame (Van den Berg et al.
1998; Newsted et al. 1995; Safe 1994). Figures 7.4-1 and 7.4-2 illustrate the ranges of effects of
tPCBs and TEQ, respectively, to various avian species.
Sensitivities of avian species to tPCBs and TEQ have been shown to vary greatly in literature
studies. Wild turkey embryos were found to be 20 to 100 times less sensitive than chicken
embryos to egg yolk injection of PCB-77. This difference in toxicity may be attributed to
differences in the availability of Ah receptors. Ah receptors were found in hepatocytes of 7-day-
old chicken embryos, but not in liver cells of 9-day-old turkey embryos (Brunstrom and Lund
1988). Presented below is a brief review of the scientific literature for both field-based studies
on tree swallow exposure to tPCBs and the laboratory studies used to derive the threshold range
for TEQ for insectivorous birds, and tPCBs for American robin.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-65
-------
Figure 7.4-1 Effects of Aroclor 1254/1260 on Avian Species (mg/kg bw/d)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q 7/10/2003
-------
Figure 7.4-2 Effects of 2,3,7,8-TCDD TEQ on Avian Species (ng TEQ/kg bw/d)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-67
-------
1
Toxicity of tPCBs and TEQ to Avian Species
2
Mode of Action
3
Bind to the aryl-hydrocarbon (Ah) receptor, eliciting an Ah-receptor-mediated
4
biochemical and toxic response.
5
Types of Toxicity
Specific Effects
6
hepatotoxicity
mortality
7
immunotoxicity
decreased growth
8
neurotoxicity
weight loss
9
embryotoxicity
porphyria
10
teratogenicity
reduced hatching
11
embryo mortality
12 7.4.1 Review of Effects of tPCBs
13 Heath et al. (1972) studied the effects of Aroclor 1254 on mortality of four avian species. The
14 most sensitive (after oral dosing for 5 days with Aroclor 1254) was the bobwhite quail, with a
15 median lethality response occurring at a dietary PCB concentration of 604 mg/kg. Other species
16 tested, such as the Japanese quail, mallard duck, and ring-necked pheasant, were less sensitive,
17 with oral LC50s of 2,898, 2,699, and 1,091 mg/kg diet, respectively. Prestt et al. (1970)
18 estimated the median lethal dietary dose rate of Aroclor 1254 to adult Bengalese finches to be
19 256 mg/kg/d.
20 Reproductive impairment of birds caused by tPCBs has been investigated in several species, in
21 dietary and egg injection studies, as well as field studies examining egg and hatchling
22 concentrations and hatching success. The most commonly noted effects to the reproduction of
23 avian species are reduced egg productivity, egg hatchability, and chick growth rates (CCME
24 1999). Of the species studied, chickens appear to be the most sensitive, followed by pheasants,
25 turkeys, ducks, and herring gulls (Bosveld and Van den Berg 1994). Total PCBs appear to have
26 no adverse effects on total egg weight, eggshell weight, or eggshell thickness (Lillie et al. 1974;
27 Britton and Huston 1973; Scott 1977). Lillie et al. (1974) exposed hens to 2 mg/kg Aroclor 1254
28 for 63 days in feed, giving birds a daily PCB dose of approximately 0.12 mg/kg bw/d. At this
29 treatment level, no significant effects were observed on fertility, egg production, shell thickness,
30 or hatchability, but the growth rate of chicks was slightly reduced. American kestrels were given
31 an approximate tPCB dose of 7 mg/kg bw/d for 100 days (Fernie et al. 2001) and birds
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-68
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
experienced a slight, but statistically insignificant, decrease in clutch size and the numbers of
fertile eggs, hatchlings, and fledglings per breeding pair.
Large clutch sizes have been reported in other populations of bird species exposed to chemical
contamination (McCarty and Secord 1999a). The long-term impact of this effect is unknown,
but is considered an indicator of disturbed reproductive biology. McCarty and Secord (1999a)
indicated that the overall reproductive success of tree swallows along the Hudson River, which is
contaminated by PCBs, was generally low. Hatchability of eggs was among the lowest reported
for this species, nest abandonment was higher than expected, and the number of large clutches
was high. Additionally, nest quality based on nest weight and number of feathers lining the nest
was low relative to control sites (McCarty and Secord 1999b). Concentrations of tPCBs in
nestlings were as high as 62.2 mg/kg. Overall nest quality can have an important impact on
length of incubation, length of nestling period, nestling mass, and number of fledglings produced
(McCarty and Secord 1999b). The reference data for this study were collected from an area
>230 km away from the Hudson River study area and 3 years previous (1990-91) (Custer 2002).
Comparing reproductive results from different years can be misleading (Custer 2002).
A field study conducted for GE examined the effects of ecological stressors (i.e., nest box
density) on tree swallow reproduction (Robertson and Jones 2002). The study took place 50 km
north of Kingston, Ontario, Canada. Ecological factors considered in the study included, but
were not limited to, nest spacing and proximity to forest or shrub edge. The study was not
considered in this risk assessment because the study did not evaluate the effects of tPCBs or TEQ
on tree swallow reproduction; therefore, it was unrelated to the assessment endpoint.
7.4.2 2,3,7,8-TCDD and Equivalence (TEQ)
In single, oral doses of TCDD, bobwhite quail, mallards, and ringed turtledoves were found to
have 37-day LD50s of 15,000, 108,000, and 810,000 ng/kg bw (Hudson et al. 1984). Ring-
necked pheasants given intraperitoneal TCDD doses weekly of 10, 100, or 1,000 ng/kg bw/week
for 10 weeks (Nosek et al. 1992) experienced no mortality or body weight effects in the two
lowest treatment groups, but the 1,000 ng/kg bw/week treatment group experienced 60%
mortality by the 23rd week (13 weeks after the dosing period).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-69
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Nosek et al. (1992) also investigated reproductive effects of TCDD to ring-necked pheasant
hens. The two lower doses (10 and 100 ng/kg bw/week) caused no significant impairment to egg
production. The highest dose, 1,000 ng/kg bw/week, caused a decrease in cumulative egg
production of approximately 70% over 7 weeks. The geometric mean of the LOAEL and
NOAEL from this study was 44 ng/kg bw/d. Hoffman et al. (1996) investigated the
developmental effects of TEQ on American kestrels and observed that skeletal growth was
significantly reduced at treatment levels of 25,000 ng TEQ/kg bw/d. This dose did not translate
into significant effects on hatchling success or weight gain.
7.4.3 Tree Swallow Field Study
The tree swallow reproduction field study was conducted between 1998 and 2000 in the PSA and
reference areas. The study focused on nest box occupancy rate, clutch size, nesting success, and
determination of tissue concentrations of tPCBs and TEQ.
Methods
Nest boxes were deployed in 1998 at four locations along the Housatonic River and its tributaries
and in 1999 at two additional reference locations. Nest boxes were placed on poles in early
spring and monitored thereafter for reproductive success. Samples of eggs or just hatched young
(pippers) were collected for organochlorine contaminant analyses. Twelve-day-old nestlings
were collected from selected sites to quantify concentration and accumulation rates in nestlings
and to assay EROD activity. Stomach content samples were taken from birds in 1998, 1999, and
2000 (Custer 2002). EROD activity analysis of tPCBs and PCB congeners was conducted using
standard methods. 2,3,7,8-TCDD equivalents (TEQ) were calculated using the toxic equivalency
factors (TEFs) of Van den Berg et al. (1998).
Results
Tree swallows nested at all study locations in the Housatonic River watershed. Nesting attempts
increased from the first to second year and stabilized in the third year. Average clutch sizes in
1998, 1999, and 2000 were 5.43, 5.37, and 5.46, respectively. There was a significant negative
relationship between tPCB concentrations in eggs and hatching success in 1999 (p = 0.044);
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q qrs 7/10/2003
-------
1 however, the fit of the model was poor. In 1998 and 1999, clutches that contained dead embryos
2 had significantly higher concentrations of tPCBs than those that hatched normally (p < 0.001).
3 The geometric mean concentrations in clutches that had reduced hatching were 62.8 mg/kg ww
4 in eggs in 1998 and 69.1 mg/kg ww in eggs in 1999.
5 Differences in the geometric mean concentrations of dioxins and furans between sites were
6 similar to the pattern for tPCBs. EROD activity was significantly induced at sites in the
7 Housatonic River PSA compared to Threemile Pond.
8 Conclusions
9 The fecundity of tree swallows in the PSA was not severely impacted by contaminants. In 1998
10 and 1999, clutches with dead embryos had geometric means of 62.8 and 69.1 mg/kg ww tPCB in
11 eggs. These concentrations exceeded the field-based threshold of 62.2 mg/kg ww tPCBs in eggs
12 established from the studies by McCarty and Secord (1999a, 1999b) and Secord et al. (1999).
13 Multivariate regression models indicated that dioxins and furans in the PSA could be
14 contributing to the observed reduced hatching success.
15 7.4.4 American Robin Field Study (GE)
16 The reproductive output of American robins was studied in the PSA and reference areas during
17 the 2001 breeding season (Henning 2002). The objective of the study was to evaluate the
18 relationship between reproductive success and tissue concentrations of tPCBs in eggs and young.
19 Methods
20 The study was conducted in the Housatonic River 10-year floodplain and in nearby public and
21 private lands. Active robin nests (having eggs or a fresh, wet mud lining) were monitored every
22 three days to record the number of eggs and hatchlings, as well as depredation, abandonment,
23 parental behavior, and development of young. Eggs and nestlings were analyzed for tPCB
24 concentrations. A nest was deemed "successful" if at least one young fledged. Clutch size,
25 hatching success, and fledging success were determined for successful nests.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-71
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Results
Clutch size averaged 2.91 eggs per nest in the PSA and 2.87 eggs per nest in the reference areas.
Clutch sizes in the PSA and reference areas were not significantly different according to Tukey's
studentized range test. Mean hatching success rates were 0.81 hatchlings per nest in the PSA and
0.89 in the reference areas, and mean fledging success rates were 0.61 in the PSA and 0.62 in the
reference areas. The differences were not significant according to Tukey's studentized range
test.
Total PCB concentrations in robin eggs averaged 83.6 mg/kg ww in the PSA (n=9) and 0.153
mg/kg ww in the reference areas (n=2). Only nests from which egg samples were taken for tPCB
analysis were included in the test. The test indicated no relationship between tPCB
concentrations in eggs and clutch size.
Conclusions
Concentrations of tPCBs in American robin eggs and nestlings were significantly higher in the
PSA compared to reference areas, with tPCB concentrations in robin eggs averaging 83.6 mg/kg
ww in the PSA (n=9) and 0.153 mg/kg ww in the reference areas (n=2). There were, however,
no significant differences in any of the measures of effects on reproduction included in this
study. Further, clutch size, hatching success, and fledging success in the target and reference
areas were within ranges reported for American robins (Brehmer and Anderson 1992; Kemper
and Taylor 1981; Fluetsch and Sparling 1994).
7.4.5 Selection of Effects Metrics for Characterizing Risk
Effects data can be characterized and summarized in a variety of ways ranging from benchmarks
designed to be protective of most or all species to concentration- or dose-response curves. A
summary of the decision criteria used to derive effects metrics is provided in the text box.
Further details on the decision criteria used in selecting effects metrics is provided in Section 6.6
of the ERA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-72
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Decision Criteria for Derivation of Effects Metric
The following is the hierarchy of decision criteria used to characterize effects for each
receptor-COC combination:
1. Have single-study bioassays with five or more treatments been conducted on the
receptor of interest or a reasonable surrogate? If yes, estimate the
concentration- or dose-response. If not, go to 2.
2. Are multiple bioassays with similar protocols, exposure scenarios, and effects
metrics available that, when combined, have five or more treatments for the
receptor of interest or a reasonable surrogate? If yes, estimate the dose-
response relationship as in 1. If not, go to 3.
3. Have bioassays with less than five treatments been conducted on the receptor of
interest or a reasonable surrogate? If yes, conduct or report results of
hypothesis testing to determine the NOAEL and LOAEL. If not, go to 4.
4. Are sufficient data available from field studies and monitoring programs to
estimate concentrations or doses of the COC that are consistently protective or
associated with adverse effects? If yes, develop field-based effects metrics. If
not, go to 5.
5. Derive a range where the threshold for the receptor of interest is expected to
occur. Because information on the sensitivity of the receptor of interest is
lacking, it is difficult to derive a threshold that is biased neither high nor low. If
bioassay data are available for several other species, however, calculate a
threshold for each to determine a threshold range that spans sensitive and
tolerant species. That range is likely to include the threshold for the receptor of
interest.
Although much work has been done to investigate the effects of tPCBs and TEQ to various bird
species in controlled laboratory settings, no suitable studies were available for tree swallows,
American robins, or for bird species that could be considered reasonable surrogates. Therefore,
no dose-response relationship could be established between exposure of tree swallows or
American robins to tPCBs (either Aroclor 1254 or 1260) or TEQ and adverse effects on
mortality, growth, or reproduction. It was also not possible to establish a NOAEL or LOAEL for
adverse effects from available laboratory studies.
7.4.5.1 Effects of tPCBs to Tree Swallows
Based on the results of the field-based effects studies reviewed, a field-based threshold range
was derived for tree swallows exposed to tPCBs. The low threshold was based on the McCarty
MK01|O:\20123001.096\ERA_PB\ERA_PB_7.DOC j
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
and Secord (1999b) study where tPCB concentrations in whole body nestlings were 62.2 mg/kg
ww. At this concentration, reproductive effects were observed (e.g., abnormal nest abandonment
behavior, larger clutch sizes), but the ecological significance of these effects could not be
determined. The high effects threshold was based on the field study conducted by Custer (2002)
where tree swallow pipper tissue concentrations of 69 mg/kg ww were associated with hatching
problems. This value also has uncertainty as to whether it caused adverse effects to tree swallow
hatching success in the PSA (Custer 2002). Thus, despite the narrowness of the tPCB threshold
range, there is some uncertainty about this effects metric. In addition, field studies are subject to
a number of factors that are impossible to control, including weather, predation, disease, and
presence of other COPCs. Therefore, the effects threshold range for insectivorous birds for
tPCBs is 62.2 to 69 mg/kg ww.
7.4.5.2 Effects of tPCBs to American Robins
Based on a review of avian toxicity literature, white leghorn chickens were the most sensitive
avian species to the reproductive effects of tPCBs and the most reproductively tolerant avian
species to tPCBs was the American kestrel. The threshold range for the reproductive success of
American robins exposed to tPCBs selected for this assessment is 0.12 to 7.0 mg/kg bw/d based
on reproductive studies conducted on white leghorn chickens (Lillie et al. 1974) and American
kestrels (Fernie et al. 2001), respectively.
7.4.5.3 Effects of TEQ to Tree Swallows and American Robins
The low toxicological threshold for effects of TEQ to sensitive birds is based on the study by
Nosek et al. (1992) using ring-necked pheasants. A dose of 44 ng/kg bw/d World Health
Organization (WHO) TEQ (geometric mean of LOAEL and NOAEL) was assumed to be the
threshold for sensitive avian species exposed to TEQ. The threshold for tolerant avian species
was derived from Hoffman et al. (1996) where 25,000 ng/kg bw/d TEQ was determined to be the
reproductive threshold for American kestrels exposed to TEQ. Thus, the effects threshold range
is 44 to 25,000 ng/kg bw/d1 for TEQ.
1 A tissue residue threshold range was derived for TEQ by multiplying this threshold range by 15 days, which is the
number of days included in the microexposure model. However, this threshold range was not used in the final
assessment of risk to tree swallows because the TDI model results were used to estimate risk of TEQ to tree
swallows.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC ry ~ , 7/10/2003
-------
1 7.5 RISK CHARACTERIZATION
2 This section characterizes risk to insectivorous birds exposed to tPCBs and 2,3,7,8-TCDD TEQ
3 in the PSA of the Housatonic River. The risk characterization uses two lines of evidence to
4 determine potential ecological risks to tree swallows and American robins. The major lines of
5 evidence are considered to be independent and will be combined in a weight-of-evidence
6 assessment. The key risk questions and the lines of evidence are summarized in the text box.
7
8
9
10
11
12
13
14
15
16 Section 7.5.1 compares the modeled exposure of tree swallows and American robins to the
17 effects metrics. Sections 7.5.2 and 7.5.3 present brief overviews of the field studies conducted
18 for tree swallows and American robins, respectively. A more detailed presentation of this
19 information is presented in Appendix G. A weight-of-evidence assessment is presented in
20 Section 7.5.4 along with the sources of uncertainty (Section 7.5.5), and the overall findings of the
21 risk characterization and extrapolation to other species (Section 7.5.6).
22 7.5.1 Comparison of Modeled Exposures to Effects
23 For tree swallows, exposure was assessed separately in the three PSA locations of Holmes Road,
24 New Lenox Road, and Roaring Brook, and in the three reference locations of Southwest Branch,
25 Threemile Pond, and Taconic Valley. The probabilistic exposure analysis for tree swallows used
26 a total daily intake rate modeling approach for calculating exposure to tPCBs and TEQ. In
27 addition to this model, a microexposure model was also employed to estimate tissue
28 concentrations of tPCBs and TEQ in juvenile birds after the first 2 weeks of their development.
29 The results from the microexposure model were used to estimate risk for tPCBs and the TDI
Key Risk Questions
¦ Are the concentrations of tPCBs and TEQ present in the prey of insectivorous
birds sufficient to cause adverse effects to individuals inhabiting the PSA of the
Housatonic River?
¦ If so, how severe are the risks and what are their potential consequences?
Lines of Evidence
¦ Field-based tree swallow and American robin reproduction studies.
¦ Probabilistic exposure and effects modeling.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-75
-------
1 model was used to estimate the risk for TEQ, to match the type of effect metric developed for
2 each COC.
3 Exposure of American robins was estimated in three areas (Locations 13, 14, and 15, see Figure
4 7.3-1) in the PSA. The probabilistic exposure analysis for American robins used a total daily
5 intake model to estimate exposure to tPCBs and TEQ.
6 For each COC-location combination, a category of low, intermediate, or high risk was assigned,
7 using the guidance below, when integrating the exposure and effects distributions.
8
9
10
11
12
13
14
15 This exercise was done separately for the results of the Monte Carlo analyses and the lower and
16 upper bounds from the probability bounds analyses. The "risk category" refers to the level of
17 risk based on the results of the Monte Carlo analyses. The "risk range" refers to the levels of risk
18 based on the results of the probability bounds analyses.
19 The results of the risk characterization are summarized in Table 7.5-1. The highest risk to tree
20 swallows is from exposure to tPCBs in the PSA sites and in the Taconic Valley reference site,
21 with low risk in the other two reference areas. As shown in Figure 7.3-27, the tPCB exposure
22 distribution from the Monte Carlo analysis in the Holmes Road location is above the upper
23 toxicity threshold. This means that the estimated exposure of tree swallows to tPCBs is greater
24 than the upper bound threshold for adverse effects to this species. The risk category for tree
25 swallows is high at the three PSA locations, and the risk range is low-high, indicating high
26 uncertainty. The risk category and the risk range are low for the Southwest Branch and
27 Threemile Pond reference locations; for the Taconic Valley reference location, the risk category
28 is high and the risk range is low-high (Table 7.5-1). The highest risk to American robins is from
29 exposure to tPCBs at Location 13. The risk category at Locations 14 and 15 is intermediate.
Guidance for Integrating the Exposure and Effects Distributions
¦ If the probability of exceeding the lower toxicity threshold is less than 20%, the
risk is considered to be low.
¦ If the probability of exceeding the upper toxicity threshold is greater than 20%,
the risk is considered to be high.
¦ All other outcomes are considered to have intermediate risk.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-76
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
The risk range from exposure to tPCBs at Locations 13 and 14 is low-intermediate, and at
Location 15 the risk range is intermediate.
Tree swallows are at intermediate risk from exposure to TEQ in the PSA and in the reference
locations, with risk ranges of intermediate for all sites except Southwest Branch, which had a
risk range of low-intermediate (Table 7.5-1). Similarly, American robins are at intermediate risk
from exposure to TEQ at Locations 13 and 14, and are at low risk at Location 15. The risk range
from exposure to TEQ at Location 13 is intermediate, and at Locations 14 and 15 the risk range
is low/intermediate.
The complete characterization of risks of tPCBs and TEQ to insectivorous birds is presented in
Appendix G.
Table 7.5-1
Summary of Qualitative Risk Statements for Insectivorous Birds from the
Housatonic River Study Area
Bird/Location
Qualitative Risk Statements
tPCBs
TEQ
Risk Category
Risk Range
Risk Category
Risk Range
Tree Swallow
Holmes Road
High
Low/High
Intermediate
Intermediate
New Lenox Road
High
Low/High
Intermediate
Intermediate
Roaring Brook
High
Low/High
Intermediate
Intermediate
Southwest Branch
Low
Low
Intermediate
Low/Intermediate
Threemile Pond
Low
Low
Intermediate
Intermediate
Taconic Valley
High
Low/High
Intermediate
Intermediate
American Robin
Location 13
High
Intermediate/High
Intermediate
Intermediate
Location 14
Intermediate
Intermediate/High
Intermediate
Low/Intermediate
Location 15
Intermediate
Intermediate
Low
Low/Intermediate
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q qq 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
7.5.2 Tree Swallow Field Study
Custer (2002) conducted a tree swallow reproduction study in the years 1998 through 2000 in the
PSA and reference areas. Nest boxes were placed at three locations in the PSA (Holmes Road,
New Lenox Road, and Roaring Brook) and at three reference locations (Southwest Branch,
Taconic Valley, and Threemile Pond in 1999 and 2000 only). Several indicators of reproductive
performance were monitored, including clutch size, nesting success, and determination of tissue
concentrations of tPCBs and TEQ in eggs. Pippers and 12- to 14-day-old nestlings were also
measured. The field study indicated that tree swallows did not experience serious adverse effects,
despite high tissue concentrations of tPCBs and TEQ in nestlings in the PSA locations.
In central Massachusetts, the mean clutch size was 4.8 to 5.3 eggs/clutch over a 22-year period
(Chapman 1955). The mean clutch sizes for tree swallows in the PSA in 1998, 1999, and 2000
were 5.43, 5.37, and 5.46 eggs/clutch, respectively. Thus, the fecundity of tree swallows in the
PSA was unaffected by tPCBs and TEQ. McCarty and Secord (1999a) similarly reported large
clutch sizes in tree swallows exposed to PCBs in the Upper Hudson River watershed, NY.
The geometric means of tPCB concentrations in tree swallow pippers and nestlings collected
from the Housatonic River ranged from 32 to 101 mg/kg ww whole body. These are the highest
concentrations reported in the literature (Custer 2002) and are substantially higher than
concentrations in samples obtained from reference sites (6 to 19 mg/kg ww whole body). Total
PCBs, dioxins, and furans were negatively correlated with hatching success in 1998 and 1999,
but the correlations were weak. Hatching success was not correlated with these COCs in 2000,
probably because concentrations were reduced in 2000 and because cold weather contributed to
poor hatching at all locations.
7.5.3 American Robin Field Study (GE)
A study of American robin reproduction was performed during the 2001 breeding season in the
PSA and reference areas. The study evaluated the relationship between tissue concentrations and
reproductive output. The study was conducted in the Housatonic River 10-year floodplain and in
nearby public and private lands within the watershed. American robin nests were located by
direct observation and identification, as well as by tracking robins in the area to their nests.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 7 78 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Eggs and nestlings were collected for analysis of tPCB residues. One egg that was viable and
had been incubated for at least 10 days, was randomly chosen from each nest having at least four
eggs; no eggs were taken from nests with fewer eggs. Some nonviable eggs were also collected
for chemical analysis. The largest 7-day old nestling was selected for analysis of tPCB residues
from successful nests with three or more nestlings; no nestlings were collected from nests with
fewer young. A successful nest was defined as a nest that fledged at least one young and the
percent of successful nests was determined by dividing the number of successful nests by the
number of active nests. Clutch size, and hatching and fledging success were also determined for
active nests.
Concentrations of tPCBs in American robin eggs and nestlings were significantly higher in the
PSA compared to reference areas, yet there were no significant differences in the measures of
effects included in this study. Clutch size, hatching success, and fledging success all exhibited
no difference in target and reference areas and were within ranges typical for American robins
(Brehmer and Anderson 1992; Kemper and Taylor 1981; Fleutsch and Sparling 1994). The
number of nonviable eggs per successful nest was higher in the PSA compared to the reference
areas, although this difference was not significant (p = 0.89). The difference in egg tPCB
concentrations at study and reference sites indicated that robins in the PSA were receiving a local
exposure. The principal concerns with the study are the inclusion of only active nests in the
measurement of reproductive success, the small number of eggs and nestlings that were collected
in the reference areas and analyzed for tPCBs, the non-random selection of eggs and nestlings,
the methods used to determine viability of eggs, and the use of some questionable statistical
methods to analyze the results. A reanalysis of the clutch size data was performed, including all
nests with these data, as opposed to only active nests. This relationship was examined with a
Chi-squared test for trend, but results showed no significant relationship between PCB
concentration in eggs and clutch size.
7.5.4 Weight-of-Evidence Analysis
A weight-of-evidence analysis procedure was used to assess risks of tPCBs and TEQ to
insectivorous birds. The goal of this analysis was to determine whether significant risk is posed
to insectivorous birds in the Housatonic River PSA as a result of exposure to tPCBs and TEQ.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-79
-------
1 The three-phase approach of Menzie et al. (1996) and the Massachusetts Weight-of-Evidence
2 Workgroup was applied for this purpose, in which weight-of-evidence was reflected in the
3 following three characteristics: (1) the weight assigned to each measurement endpoint, (2) the
4 magnitude of response observed in the measurement endpoint, and (3) the concurrence among
5 outcomes of the multiple measurement endpoints (see Section 2.9 for details).
6 A discussion of attributes considered in the WOE is provided in Section 2, and the rationale for
7 weighting of measurement endpoints are provided in Appendix G. A summary of how attributes
8 were weighted for the tree swallow and American robin lines of evidence is provided in Tables
9 7.5-2 and 7.5-3, respectively. The attribute values, evidence of harm, and magnitudes of
10 responses for both tree swallows and American robins are presented in Table 7.5-4 (tPCBs) and
11 Table 7.5-5 (TEQ).
12
13
14
15
16
17
18
19
20
For tree swallows exposed to tPCBs and TEQ, the field-based reproductive study
line of evidence was given a high weighting, and the modeled exposure and effects
line of evidence was given a moderate weighting.
For American robins, the field study was given a moderate/high weighting for both
tPCBs and TEQ. The modeled exposure and effects line of evidence was given
moderate value for tPCBs and TEQ.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-80
-------
Table 7.5-2
Weighting of Measurement Endpoints for Tree Swallow Weight-of-Evidence
Evaluation
Attributes
Field Study
Custer (2002)
Modeled Exposure and
Effects for tPCBs and TEQ
I. Relationship Between Measurement and Assessment Endpoints
1. Degree of Association
High
Moderate
2. Stressor/Response
Moderate/High
Moderate
3. Utility of Measure
High
Moderate
II. Data Quality
4. Data Quality
High
Moderate
III. Study Design
5. Site Specificity
High
Low/Moderate
6. Sensitivity
Moderate/High
Low/Moderate
7. Spatial Representativeness
High
Moderate
8. Temporal Representativeness
High
Moderate
9. Quantitative Measure
Moderate/High
Moderate/High
10. Standard Method
High
Moderate
Overall Endpoint Value
High
Moderate
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC q q -i 7/10/2003
-------
1 Table 7.5-3
2
3 Weighting of Measurement Endpoints for American Robin Weight-of-Evidence
4 Evaluation
Attributes
Field Study
Henning (2002)
Modeled Exposure and
Effects for tPCBs and TEQ
I. Relationship Between Measurement and Assessment Endpoints
1. Degree of Association
High
Moderate
2. Stressor/Response
Moderate
Moderate
3. Utility of Measure
Moderate/High
Moderate
II. Data Quality
4. Data Quality
Moderate/High
Moderate
III. Study Design
5. Site Specificity
High
Low/Moderate
6. Sensitivity
Moderate/High
Low/Moderate
7. Spatial Representativeness
High
Moderate
8. Temporal Representativeness
Moderate
Moderate
9. Quantitative Measure
Moderate/High
Moderate/High
10. Standard Method
Moderate/High
Moderate
Overall Endpoint Value
Moderate/High
Moderate
5
6 The magnitude of the response in the measurement endpoint is considered together with the
7 measurement endpoint weight in judging the overall weight-of-evidence (Menzie et al. 1996).
8 This requires assessing the strength of evidence that ecological harm has occurred, as well as an
9 indication of the magnitude of response, if present. For both tree swallows and American robins
10 exposed to tPCBs (Table 7.5-4) and TEQ (Table 7.5-5), the modeled exposure and effects line of
11 evidence indicated that there was evidence of harm, and that the magnitude was high risk for
12 tPCBs and intermediate risk for TEQ. For both tPCBs and TEQ, the tree swallow field study line
13 of evidence and the American robin field study line of evidence indicated that there was little
14 evidence of harm, and that the magnitude was low.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-82
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Table 7.5-4
Evidence of Harm and Magnitude of Effects for Insectivorous Birds Exposed to
tPCBs in the Housatonic River PSA
Measurement Endpoints
Weighting Value
(High, Moderate, Low)
Evidence of Harm
(Yes, No,
Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Study
High (Tree swallow)
Moderate/High
(American robin)
No (Tree swallow)
No (American robin)
Low (Tree swallow)
Low (American robin)
Modeled Exposure and
Effects
Moderate
Yes
High
Table 7.5-5
Evidence of Harm and Magnitude of Effects for Insectivorous Birds Exposed to
TEQ in the Housatonic River PSA
Measurement Endpoints
Weighting Value
(High, Moderate, Low)
Evidence of Harm
(Yes, No,
Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Study
High (Tree swallow)
Moderate/High
(American robin)
No (Tree Swallow)
No (American robin)
Low (Tree Swallow)
Low (American robin)
Modeled Exposure and
Effects
Moderate
Yes
Intermediate
The final component in the weight-of-evidence approach addresses the concurrence among the
measurement endpoints as they relate to each assessment endpoint. The methodology for
detecting concurrence involves the use of a graphical method where measurement endpoints are
plotted on a matrix that also includes the weight of each endpoint and degree of response. Tables
7.5-6 and 7.5-7 depict the outcomes for tree swallows and American robins exposed to tPCBs
and TEQ, respectively. The analyses were conducted separately for tPCBs and TEQ.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 7 83 7/10/2003
-------
Table 7.5-6
Risk Analysis Summary for Insectivorous Birds Exposed to tPCBs in the
Housatonic River PSA
Assessment Endpoint: Survival, growth, and reproduction of insectivorous birds
~
Harm/Magnitude
Weighting Factors (increasing confidence of weight)
Low
Low/Moderate
Moderate
Moderate/High
High
Yes/High
MEE
Yes/Intermediate
Yes/Low
Undetermined/High
Undetermined/Intermediate
Undetermined/Low
No/Low
FS (American
robin)
FS (Tree
swallow)
No/Intermediate
No/High
FS=Field study
MEE=Modeled exposure and effects
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC -y Q/| 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Table 7.5-7
Risk Analysis Summary for Insectivorous Birds Exposed to TEQ in the
Housatonic River PSA
Assessment Endpoint: Survival, growth, and reproduction of insectivorous birds
~
Harm/Magnitude
Weighting Factors (increasing confidence of weight)
Low
Low/Moderate
Moderate
Moderate/High
High
Yes/High
Yes/Intermediate
MEE
Yes/Low
Undetermined/High
Undetermined/Intermediate
Undetermined/Low
No/Low
FS (American robin)
FS (Tree
swallow)
No/Intermediate
No/High
FS = Field Study
MEE = Modeled exposure and effects
The results from the modeled exposure and effects line of evidence suggest that tPCBs and TEQ
pose a intermediate to high risk to tree swallows living in the PSA. However, the field study
line of evidence suggests that, if effects are occurring, they are minor. The uncertainty
concerning the field-based threshold range for tPCBs likely means that risks of this COC are
overestimated for the PSA. Even the upper end of the tPCB range is associated with equivocal
evidence for adverse effects to tree swallows. For TEQ, the threshold range is quite broad. The
available evidence from field studies indicates that tree swallows are tolerant to exposure to
persistent organochlorines such as tPCBs and TEQ (Section G.3.2). If the tree swallow threshold
is near the upper end of the threshold range, then the current modeled exposure and effects line
of evidence is overestimating risks of TEQ to tree swallows. Thus, the WOE assessment
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 7 85 7/10/2003
-------
1 supports a finding of low risk for tree swallows exposed to tPCBs and TEQ in the PSA. This
2 conclusion, however, is uncertain because of the conflicting results in the WOE assessment.
3 The results from the modeled exposure and effects lines of evidence suggest that tPCBs and TEQ
4 pose an intermediate to high risk to American robins inhabiting the PSA of the Housatonic River.
5 The American robin field study, however, suggests that reproductive success is not being
6 impaired by the tPCBs and TEQ in the PSA. The uncertainty in the modeled exposure and
7 effects line of evidence, outlined below, likely means the approach overestimates the risks of
8 tPCBs and TEQ to American robins in the PSA. The WOE assessment, therefore, supports a
9 conclusion of low risk to American robins exposed to tPCBs and TEQ in the PSA. This
10 conclusion, however, is uncertain because of the conflicting results in the WOE assessment.
11 7.5.5 Sources of Uncertainty
12 The assessment of risk to insectivorous birds contains uncertainties. Each source of uncertainty
13 can influence the estimates of risk; therefore, it is important to describe and, when possible,
14 specify the magnitude and direction of such uncertainties. The sources of uncertainty associated
15 with the assessment of risks of tPCBs and TEQ to insectivorous birds are described below.
16 "In this assessment, it was assumed that dietary exposure represented the most
17 important pathway for exposure of insectivorous birds to COCs. Although unlikely
18 to provide a major contribution to the risk, other pathways could increase exposure
19 and perhaps increase risk slightly (Moore et al. 1999). Deterministic calculations
20 were conducted in which estimates of exposure to COCs via drinking water and soil
21 ingestion were included in the exposure model. Inclusion of these routes did not
22 substantially increase overall exposure of insectivorous birds to the COCs.
23 ¦ The microexposure model used the ratio of tissue concentrations in pippers and
24 nestlings to indirectly estimate the amount of tPCBs and TEQ transferred from
25 mothers to offspring via egg tissue. The ratio was calculated using a sample of
26 limited size. Variability in the ratio was high. Each of these parameters has
27 associated uncertainties, but the potential magnitude and direction of uncertainty is
28 unknown.
29 ¦ The Monte Carlo sensitivity analyses suggested that the free metabolic rate (FMR)
30 slope and power terms were generally the most influential variables on predicted total
31 daily intakes of COCs. However, no direct measurements of free metabolic rate are
32 available for tree swallows. Similarly, measured food intake rates are not available
33 for free-living tree swallows and American swallows. Therefore, free metabolic rates
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 7 Q A 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
were estimated using allometric equations. The use of allometric equations
introduced some degree of uncertainty into the exposure estimates because they have
model-fitting error. The uncertainty due to model-fitting error was propagated in the
uncertainty analyses by using distributions as input for the allometric slope and power
terms.
¦ Sample sizes were limited for the analyses of COC concentrations in tree swallow
stomach contents. Only four samples from each of 3 years were available for each
location (2 years for Taconic Valley and Threemile Pond). In addition, these samples
were pooled from collections from individual tree swallows to augment sample mass,
but the exact number of individuals pooled was variable and reported only as a range.
¦ Sample sizes were limited for the analyses of COC concentrations in some prey
items, including earthworms, litter invertebrates, and benthic invertebrates. In the
case of TEQ analyses in earthworms, only one composite sample was available at
each location, composed of between 20 and 45 worm samples. To address this
uncertainty in the Monte Carlo analysis, the UCL or data set maximum (see Section
6.4 and Appendix C.5) was used as an estimate of COC concentrations in prey items.
The potential magnitude of the uncertainty associated with small sample sizes for
COC concentrations is unknown, but this approach likely overestimates exposure.
The probability bounds analysis used an unbiased approach (e.g., distribution free
range from lower confidence limit [LCL] to upper confidence limit [UCL]) to deal
with sample size uncertainty.
¦ PCB congeners 123 and 157 co-eluted with other congeners (PCB-123 with PCB-
149; PCB-157 with PCB-173 and PCB-201). As a result, decision criteria were
developed (see Section 6.4 of the ERA) for co-eluted congeners to determine TEQ
concentrations used as distribution parameters in the Monte Carlo and probability
bounds analyses. These criteria were designed to explicitly incorporate this source of
uncertainty in the probabilistic analyses. Thus, this source of uncertainty has been
incorporated in this risk assessment.
¦ Many TEQ congeners were detected at or below the method detection limit (see
Appendix G and Section 6.4 of the ERA), particularly in the reference locations. In
the Monte Carlo analyses, for prey concentration distributions affected by non-
detects, a triangular distribution was used with the minimum TEQ assuming non-
detects equal to zero, the best estimate TEQ assuming non-detected equal to one half
the detection limit, and the maximum TEQ assuming non-detects equal to the
detection limit. In the probability bounds analyses, a distribution-free range was used
with the lower limit assuming non-detects equal to zero and the upper limit assuming
non-detects equal to the detection limit. Thus, this source of uncertainty has been
incorporated in this risk assessment. Concentrations of tPCBs in prey tissues in the
PSA were all above the detection limit. Thus, there was no non-detect issue for
tPCBs.
¦ A source of uncertainty in the effects assessment for American robins exposed to
tPCBs and both representative species exposed to TEQ was the lack of controlled
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-87
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
laboratory studies involving tree swallows and American robins. As a result,
laboratory studies on surrogate species were used to estimate the effects range of
tPCBs and TEQ to insectivorous birds. This extrapolation introduced uncertainty in
the tPCB and TEQ effects assessment because of the variations in physiological and
biochemical factors such as uptake, metabolism, and disposition that can alter the
potential toxicity of a contaminant. The sensitivity of birds to an environmental
contaminant may differ from that of a laboratory or domestic species due to
behavioral and ecological parameters, including stress factors (e.g., competition,
seasonal changes in temperature or food availability), disease, and exposure to other
contaminants. Inbred laboratory animal strains may also have an unusual sensitivity
or resistance to a tested substance.
¦ A source of uncertainty in the effects assessment for tPCBs was due to the lack of
controlled laboratory studies involving tree swallows exposed to tPCBs. As a result,
a field-based threshold range was used as the effects metric for effects of tPCBs to
tree swallows. This extrapolation introduced uncertainty in the tPCB effects
assessment for two reasons. First, the value selected for the lower end of the range
(62.2 mg/kg) has uncertainty as to whether it caused adverse effects to tree swallow
reproduction, including abnormal nest abandonment behavior and larger clutch sizes,
in the upper Hudson River area (McCarty and Secord 1999a). The ecological
significance of these effects could not be determined. In addition, the value selected
for the upper end of the range (69 mg/kg) has uncertainty as to whether it caused
adverse effects to tree swallow hatching success in the PSA (Custer 2002). Thus,
despite the narrowness of the tPCB threshold range, there is some uncertainty about
this effects metric. Second, field studies are subject to a number of factors that are
impossible to control, including weather, predation, disease, and other COPCs.
¦ Concentrations of tPCBs in tree swallow pippers and nestlings collected within the
PSA, although some of the highest ever reported and well above previously reported
thresholds, showed a weak negative correlation to hatching success in 1998 and 1999,
and no correlation in 2000. Poor hatching success in 2000 was attributed to cold
rainy weather in the nesting season. While reproductive effects may have been
observed in tree swallows nesting at other PCB-contaminated sites, evidence
indicating that these effects are occurring or are solely the result of tPCB exposures is
lacking for the PSA and other PCB-contaminated sites.
7.5.6 Conclusions and Extrapolation to Other Species
The WOE analysis indicated that exposure of insectivorous birds, such as tree swallows and
American robins, to tPCBs and TEQ in the PSA is unlikely to lead to adverse reproductive
effects. This conclusion, however, is uncertain because the lines of evidence did not produce
concordant results. The lines of evidence used in this conclusion were the field-based tree
swallow and American robin reproductive studies and the comparison of modeled exposure with
effects.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC 7 QQ 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Custer (2002) conducted a field-based tree swallow reproduction study in the years 1998 through
2000 in the PSA and reference areas. The mean clutch sizes for tree swallows in the PSA in
1998, 1999, and 2000 were 5.43, 5.37, and 5.46 eggs/clutch, respectively, compared to central
Massachusetts, where the mean clutch size was 4.8 to 5.3 eggs/clutch over a 22-year period
(Chapman 1955). Thus, the fecundity of tree swallows in the PSA was unaffected by tPCBs and
TEQ. The geometric means of tPCB concentrations in tree swallow pippers and nestlings
collected from the Housatonic River ranged from 31.5 to 101 mg/kg ww whole body. These are
the highest concentrations reported in the literature (Custer 2002) and are substantially higher
than concentrations in samples obtained from reference sites (6 to 19 mg/kg ww whole body).
Total PCBs, dioxins, and furans were negatively correlated with hatching success in 1998 and
1999, but the correlations were weak.
The field-based tree swallow reproduction study did not detect obvious adverse effects to tree
swallow reproduction, despite high tissue concentrations in nestlings. This study supports the
conclusion that tree swallows are not being adversely affected due to exposure to tPCBs and
TEQ in the PSA.
The American robin field study (Henning 2002) was conducted during the 2001 breeding season
in the PSA and reference areas. The study evaluated the relationship between tissue
concentrations and reproductive output. Concentrations of tPCBs in American robin eggs and
nestlings were significantly higher in the PSA compared to reference areas, with tPCB
concentrations in robin eggs averaging 83.6 mg/kg ww in the PSA (n=9) and 0.153 mg/kg ww in
the reference areas (n=2). There were, however, no significant differences in any of the
measures of effect in this study. Clutch size, hatching success, and fledging success all exhibited
no difference in target and reference areas and were within ranges typical for American robins
(Brehmer and Anderson 1992; Kemper and Taylor 1981; Fleutsch and Sparling 1994).
The modeled exposure and effects line of evidence for tree swallows compared estimated body
burdens of nestlings (for tPCBs) or estimated daily intake by adult females (for TEQ) with COC
levels found in the literature. Field studies investigating the reproductive effects of tPCBs on
tree swallows were available in the literature, so the effects characterization employed a field-
based toxicity threshold range to describe the potential effects of tPCBs to tree swallows. The
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-89
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
most tolerant bird species to tPCBs found in the literature was the tree swallow, which had a
threshold tPCB tissue concentration range of 62.2 to 69 mg/kg ww, based on field-based
reproductive studies conducted on tree swallows by Custer (2002) and McCarty and Secord
(1999a). The upper threshold is the geometric mean tPCB concentration in tree swallows
associated with reduced hatching success (Custer 2002). The lower threshold represents the
whole body tPCB concentrations in tree swallow nestlings at which no adverse reproductive
effects were observed (McCarty and Secord 1999a). The microexposure model results indicated
that tree swallow nestlings are likely to attain body burdens greater than 69 mg/kg ww at the
three PSA study locations in this assessment.
Insufficient field-based data were available for American robins exposed to tPCBs. Thus, for
American robins a threshold range for surrogate species was used. White leghorn chickens were
the most sensitive avian species to the reproductive effects of tPCBs and the most reproductively
tolerant avian species to tPCBs was the American kestrel. The threshold range for the
reproductive success of insectivorous birds exposed to tPCBs selected for this assessment was
0.12 to 7.0 mg/kg bw/d based on reproductive studies conducted on white leghorn chickens
(Lillie et al. 1974) and American kestrels (Fernie et al. 2001b), respectively. The total daily
intake model results indicated that American robins are at high risk in the PSA.
A laboratory-based toxicity threshold range was used to describe the potential effects of TEQ to
insectivorous birds. The most sensitive and most tolerant bird species were used to develop the
TEQ toxicity threshold range, with the assumption that tree swallows and American robins
would begin to experience adverse effects in this range. The toxicity threshold range is very
wide (44 to 25,000 ng/kg bw/d). Most exposure estimates for tree swallows and American robins
fell within this range, placing the birds at an intermediate risk from exposure to TEQ. While the
modeled exposure and effects line of evidence suggested that tree swallows and American robins
are at risk in the PSA, confidence in this line of evidence is reduced compared to the field study.
The results from the modeled exposure and effects line of evidence suggest that tPCBs and TEQ
pose intermediate to high risks to tree swallows and American robins living in the PSA.
However, the more highly weighted field study line of evidence suggests that if effects are
occurring, they are minor for both species. Thus, the WOE assessment favors a finding of low
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-90
-------
1 risk for insectivorous birds exposed to tPCBs and TEQ in the PSA. This conclusion, however, is
2 uncertain because of the conflicting results in the WOE assessment.
3 Other insectivorous bird species common to the PSA include the bank swallow, northern rough-
4 winged swallow, barn swallow, cliff swallow, chimney swift, common nighthawk, eastern
5 kingbird, eastern phoebe, eastern bluebird, eastern towhee, gray catbird, hermit thrush, northern
6 mockingbird, veery, and wood thrush. Effects data are not available for other insectivorous bird
7 species living in the Housatonic River area. As a result, the same effects data used to estimate
8 effects to tree swallows were used for other insectivorous species. A qualitative analysis was
9 conducted to compare exposure of tree swallows, American robins, and other insectivorous birds
10 to tPCBs and TEQ. The major factors that influence avian exposure to tPCBs and TEQ include:
11 ¦ Foraging behavior and dietary composition.
12 ¦ Foraging and home range size.
13 ¦ Species body weight and other life history characteristics.
14
15 Tree swallows and other insectivorous bird species were compared using these factors and the
16 results are provided in Appendix G.4.6. A qualitative analysis of risk to these species indicates
17 that the cliff swallow, eastern kingbird, eastern bluebird, and eastern towhee have a similar to
18 lower level of risk compared to the representative species; barn swallow, common nighthawk,
19 eastern phoebe, hermit thrush, northern mockingbird, veery, and wood thrush have a similar
20 level of risk; and bank swallow, chimney swift, northern rough-winged swallow, and gray
21 catbird have a similar to higher level of risk compared to the tree swallow.
22
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-91
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
ERA Summary
The weight-of-evidence analysis indicates that insectivorous birds are likely at low
risk in the PSA as a result of exposure to tPCBs and TEQ. This conclusion, however,
is uncertain. Risks to tree swallows and American robins in the PSA are intermediate
to high based on modeled exposure and effects, but field studies detected no
obvious adverse reproductive effects in the PSA.
Other insectivorous bird species common to the PSA include the bank swallow,
northern rough-winged swallow, barn swallow, cliff swallow, chimney swift, common
nighthawk, eastern kingbird, eastern phoebe, eastern bluebird, eastern towhee, gray
catbird, hermit thrush, northern mockingbird, veery, and wood thrush. A qualitative
analysis of risk to these species indicates that the cliff swallow, eastern kingbird,
eastern bluebird, and eastern towhee have a similar to lower level of risk compared
to the tree swallow; barn swallow, common nighthawk, eastern phoebe, hermit
thrush, northern mockingbird, veery, and wood thrush have a similar level of risk; and
bank swallow, chimney swift, northern rough-winged swallow, and gray catbird have
a similar to higher level of risk compared to the tree swallow.
7.6 REFERENCES
Bell, G.P. 1990. Birds and mammals on an insect diet: A primer on diet composition analysis in
relation to ecological energetics. Studies in Avian Biology 13:391-415. Cited in EPA 1993.
Bierman, G.C. and S.G. Sealy. 1985. Seasonal dynamics of body mass of insectivorous
passerines breeding on the forested dune ridge, Delta Marsh, Manitoba. Canadian Journal of
Zoology 63:1675-1682.
Blancher, P.J. and D.K. McNicol. 1991. Tree swallow diet in relation to wetland acidity.
Canadian Journal of Zoology 69:2629-2637.
Borga, K., G.W. Gabrielsen, and J.U. Skaare. 2001. Biomagnification of organochlorines along
a Barents sea food chain. Environ. Pollut. 113(2): 187-198.
Bosveld, A.T.C. and M. Van den Berg. 1994. Effects of polychlorinated biphenyls, dibenzo-p-
dioxins, and dibenzofurans on fish-eating birds. Environmental Reviews 2:147-166.
Brehmer, P.M. and R.K. Andersen. 1992. Effects of urban pesticide application on nesting
success of songbirds. Bulletin of Environmental Contamination and Toxicology 48:352-359.
Britton, W.M. and T.M. Huston. 1973. Influence of polychlorinated biphenyls in the laying hen.
Poultry Science 52:1620-1624.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-92
-------
1 Brunstrom, B. and J. Lund. 1988. Differences between chick and turkey embryos in sensitivity
2 to 3,4'4,4'-tetrachlorobiphenyl and in concentration/affinity of the hepatic receptor for 2,3,7,8-
3 tetrachlorodibenzo-/>dioxin. Com. Biochem. Physiol. C 91(2):507-512.
4 Chapman, L.B. 1955. Studies of a tree swallow colony. Bird Banding. 26:45-70.
5 Clench, M.H. and R.C. Leberman. 1978. Weights of 151 Species of Pennsylvania Birds
6 Analyzed by Month, Age, and Sex. Bulletin of the Carnegie Museum of Natural History. 87 pp.
7 CCME (Canadian Council of Ministers of the Environment). 1999. Canadian Tissue Residue
8 Guidelines for the Protection of Wildlife Consumers of Aquatic Biota: Polychlorinated
9 Biphenyls (PCBs). In Canadian Environmental Quality Guidelines, 1999. Canadian Council of
10 Ministers of the Environment, Winnipeg.
11 Collopy, M.W. 1975. Behavioral and predatory dynamics of kestrels wintering in the Areata
12 Bottoms [master's thesis], Humboldt State University, Areata, CA.
13 Corbet, R.L., D.G. Muir, and G.R.B. Webster. 1983. Fate of carbon-14 labeled 1,3,6,8-
14 tetrachloro-dibenzo-p-dioxin in an outdoor aquatic system. Chemosphere 12:523-528.
15 Cummins, K.W. and J.C. Wuycheck. 1971. Caloric equivalents for investigations in ecological
16 energetics. International Association of Theoretical and Applied Limnology. Stuttgart, West
17 Germany, Cited in EPA 1993.
18 Custer, C.M. 2002. Exposure and Effects of Chemical Contaminants on Tree Swallows Nesting
19 Along the Housatonic River, Berkshire Co., Massachusetts, 1998-2000. Final Report to U.S.
20 Environmental Protection Agency. USGS, Upper Midwest Environmental Sciences Center, La
21 Crosse, WI.
22 Dunning, J.B. 1984. Body Weights of 686 Species of North American Birds. Eldon Publishing,
23 Cave Creek, Arizona.
24 EPA (U.S. Environmental Protection Agency). 1993. Wildlife Exposure Factors Handbook.
25 EPA/600/R-93/187a. Office of Research and Development. Washington, DC.
26 Fernie, K.J., J.E. Smits, G.R. Bortolotti, and D.M. Bird. 2001. Reproductive success of American
27 kestrels exposed to dietary polychlorinated biphenyls. Environmental Toxicology and Chemistry
28 20:776-781.
29 Fleutsch, K.M. and D.W. Sparling. 1994. Avian nesting success and iversity in conventionally
30 and organically managed apple orchards. Environmental Toxicology and Chemistry
31 13(10): 1651-1659.
32 Haffner, G.D., M. Tomczak, and R. Lazar. 1994. Organic contaminant exposure in the Lake St.
33 Clair food web. Hydrobiologia 281:19-27.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-93
-------
1 Hazelton, P.K., R. J. Robel and A.D. Dayton. 1984. Preferences and influences of paired food
2 items on energy intake of American robins (Turdus migratorius) and gray catbirds (Dumetella
3 carolinensis) J. Wildl. Manage. 48:198-202.
4 Heath, R.G., J.W. Spann, E.F. Hill, and J.F. Kreitzer. 1972. Comparative Dietary Toxicities of
5 Pesticides to Birds. U.S. Fish and Wildlife Service Special Scientific Report on Wildlife 152. 57
6 pp.
7 Henning, M.H. 2002. Robin Productivity in the Housatonic River Watershed, Berkshire County,
8 Massachusetts. Prepared for: General Electric Company, Pittsfield, Massachusetts.
9 Hoffman, D.J., M.J. Melancon, P.N. Klein, C.P. Rice, J.D. Eisemann, R.K. Hines, J.W. Spann,
10 and G.W. Pendleton. 1996. Developmental toxicity of PCB 126 (3,3,4,4,5-pentachlorobiphenyl)
11 in nestling American kestrels (Falco sparverius). Fundamental and Applied Toxicology 34:188-
12 200.
13 Howard, P.H., R.S. Boethling, W.F. Jarvis, W.M. Meylan, and E.M. Michalenko. 1991.
14 Handbook of Environmental Degradation Rates. Lewis Publishers, Chelsea, MI.
15 Hudson, R., R. Tucker, and M. Haegele. 1984. Handbook of Toxicity of Pesticides to Wildlife.
16 2nd Edition. U.S. Fish and Wildlife Service Resource Publication 153. Washington, DC.
17 Karasov, W.H. 1990. Digestion in birds: Chemical and physiological determinants and
18 ecological implications. Studies in Avian Biology 13:391-415.
19 Kemper, D.L. and J.M. Taylor. 1981. Seasonal reproductive changes in the American robin
20 (Turdus migratorius) of the Pacific Northwest. Canadian Journal of Zoology 59:212-217.
21 Kuehl, D.W., P.M. Cook, A.R. Batterman, and B.C. Butterworth. 1987. Isomer dependent
22 bioavailability of polychlorinated dibenzo-p-dioxins and dibenzofurans from municipal
23 incinerator fly ash to carp. Chemosphere 16:657-666.
24 Lillie, R.J., H.C. Cecil, J. Bitman, and G.F. Fries. 1974. Differences in response of caged white
25 leghorn layers to various polychlorinated biphenyls (PCBs) in the diet. Poultry Science 53:726-
26 732.
27 Marcum, H.A., W.E. Grant, and F. Chavez-Ramirez. 1998. Simulated behavioral energetics of
28 non-breeding American robins: The influence of foraging time, intake rate, and flying time on
29 weight dynamics. Ecological Modeling 106:161-175.
30 McCarty, J.P., and A.L. Secord. 1999a. Nest-building behaviour in PCB-contaminated tree
31 swallows. Auk 116:55-63.
32 McCarty, J.P., and A.L. Secord. 1999b. Reproductive ecology of tree swallows (Tachycineta
33 bicolor) with high levels of polychlorinated biphenyl contamination. Environ. Toxicol. Chem.
34 18:1433-1439.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-94
-------
1 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S. Petron,
2 B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-of-
3 Evidence Workgroup: A weight-of-evidence approach for evaluating ecological risks. Human
4 Ecological Risk Assessment 2:277-304.
5 Moore, D.R.J., B.E. Sample, G.W. Suter, B.R. Parkhurst, and R.S. Teed. 1999. A probabilistic
6 risk assessment of the effects of methylmercury and PCBs on mink and kingfishers along East
7 Fork Poplar Creek, Oak Ridge, Tennessee, USA. Environmental Toxicology and Chemistry
8 18(12):2941-2953.
9 Muir, D.C.G., R.J. Nostrom, and M. Simon. 1988. Organochlorine contaminants in Arctic
10 marine food chains: Accumulation of specific poly chlorinated biphenyls and chlordane-related
11 compounds. Environmental Science and Technology 22:1071-1079.
12 Nagy, K.A. 1987. Free metabolic rate and food requirement scaling in mammals and birds.
13 Ecological Monographs 57:111-128.
14 Nagy, K.A., I.A. Girard, and T.K. Brown. 1999. Energetics of free-ranging mammals, reptiles
15 and birds. Annu. Rev. Nutr. 19:247-277.
16 Newsted, J.L., J.P. Giesy, G.T. Ankley, D.E. Tillitt, R.A. Crawford, J.W. Gooch, P.D. Jones, and
17 M.S. Denison. 1995. Development of toxic equivalency factors for PCB congeners and the
18 assessment of TCDD and PCB mixtures in rainbow trout. Environmental Toxicology and
19 Chemistry 14(5): 861 -871.
20 Nosek, J.A., S.R. Craven, J.R. Sullivan, S.S. Hurley, and R.E. Peterson. 1992. Toxicity and
21 reproductive effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin in ring-necked pheasant. Journal of
22 Toxicology and Environmental Health 35:187-198.
23 NRCC (National Research Council of Canada). 1981. Poly chlorinated dibenzo-p-dioxins:
24 Criteria for their effects on man and his environment. Publication NRCC No. 18574. National
25 Research Council of Canada. Ottawa, Ontario. 251 p.
26 Prestt, I., D.J. Jefferies, and N.W. Moore. 1970. Polychlorinated biphenyls in wild birds in
27 Britain and their avian toxicity. Environmental Pollution 1:3-26.
28 Robertson, R.J., B.J. Stuchbury, and R.R. Cohen. 1992. Tree Swallow (Tachycineta bicolor). In
29 The Birds of North America, No. 11, A. Poole, P. Stettenheim, and F. Gill, Editors. Academy of
30 Natural Sciences, Philadelphia, and American Ornithologists Union, Washington, DC.
31 Roberston, R.J., and J. Jones. 2002. Spatial and Demographic Effects on Tree Swallow Nest
32 Quality and Reproductive Success. Department of Biology, Queen's University. Kingston,
33 Ontario. Prepared for General Electric Company, April 2002.
34 Safe, S.H. 1994. Polychlorinated biphenyls (PCBs): Environmental impact, biochemical and
35 toxic responses, and implications for risk assessment. Critical Reviews in Toxicology 24(2): 87-
36 149.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-95
-------
1 Secord, A.L., J.P. McCarty, K.R. Echols, J.C. Meadows, R.W. Gale, and D.E. Tillitt. 1999.
2 Polychlorinated biphenyls and 2,3,7,8-tetrachlorodibenzo-p-dioxin equivalents in tree swallows
3 from the upper Hudson River, New York State, USA. Environmental Toxicology and Chemistry
4 18:2519-2525.
5 Senthilkumar, K., N. Iseki, S. Hayama, J. Nakanishi, and S. Masunaga. 2001. Polychlorinated
6 dibenzo-p-dioxins, dibenzofurans, and dioxin-like polychlorinated biphenyls in livers of birds
7 from Japan. Arch. Environ. Contam. Toxicol. 42:244-255.
8 Scott, M.L. 1977. Effects of PCBs, DDT, and mercury compounds in chickens and Japanese
9 quail. Federation Proceedings 36:1888-1893.
10 Teather, K. 1996. Patterns of growth and asymmetry in nestling tree swallows. Journal of Avian
11 Biology 27:302-310.
12 Tsushimoto, G., F. Matsumura, and R. Sago. 1982. Fate of 2,3,7,8-tetrachlorodibenzo-p-dioxin
13 (TCDD) in an outdoor pond and in model aquatic ecosystems. Environmental Toxicology and
14 Chemistry 1:61-68.
15 Van den Berg, M., L. Birnbaum, A.T.C. Bosveld, B. Brunstrom, P. Cook, M. Feeley, J.P. Giesy,
16 A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C. Larsen, F.X.R. van Leeuwen,
17 A.K.D. Liem, C. Nolt, R.E. Petersen, L. Poellinger, S. Safe, D. Schrenk, D. Tillitt, M. Tysklind,
18 M. Younes, F. Waern, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs) for tPCBs,
19 PCDDs, PCDFs for humans and wildlife. Environ. Health Perspectives 106:775-792.
20 Wheelwright, N.T. 1986. The diet of American robins: An analysis of U.S. Biological Survey
21 records. The Auk 103:710-725.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_7.DOC
7-96
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
8. ASSESSMENT ENDPOINT—SURVIVAL, GROWTH, AND
REPRODUCTION OF PISCIVOROUS BIRDS
Highlights
Conceptual Model
Piscivorous birds, including the osprey and belted kingfisher, selected as representative
species for the ERA, are exposed to contaminants of concern (COCs) via diet and trophic
transfer. The assessment endpoint is the survival, growth, and reproduction of
piscivorous birds in the Housatonic River PSA.
Exposure
COC intake by ospreys and belted kingfishers was highest in Reaches 5 and 6 of the
PSA, while exposure in the reference areas was much lower.
Effects
No information was available specifically on the toxicity of tPCBs and TEQ to belted
kingfishers or ospreys. A threshold range spanning sensitive and tolerant surrogate
species was used instead for both tPCBs and TEQ.
Risk
Ospreys are at high risk as a result of exposure to tPCBs and intermediate risk as a
result of exposure to TEQ in the Housatonic River PSA. In these high-risk areas,
modeled exposure of ospreys to PCBs is greater than doses that cause adverse effects
in the most tolerant species studied. The weight-of-evidence (WOE) conclusion of high
risk is uncertain because other lines of evidence (e.g., field surveys, in situ or whole
media studies) were unavailable.
While modeled exposure and effects indicated high risk to belted kingfishers as a result
of exposure to tPCBs and intermediate risk as a result of exposure to TEQ, a field study
of kingfisher productivity indicated that the birds were reproducing in the PSA. The WOE
assessment for belted kingfishers concluded that this species is at low risk. This
conclusion, however, is uncertain because of the lack of concordance between the two
lines of evidence.
8.1 INTRODUCTION
This section summarizes the current and potential risks to piscivorous birds exposed to
contaminants of potential concern (COPCs) in the Housatonic River and floodplain. It focuses
on tPCBs and other COPCs originating from the General Electric Company (GE) facility in
Pittsfield, MA. The river is located in western Massachusetts and Connecticut, discharging to
Long Island Sound, with the GE facility located near the headwaters. The Primary Study Area
(PSA) includes the river and 10-year floodplain from the confluence of the East and West
Branches of the Housatonic River, downstream of the GE facility, to Woods Pond Dam.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC q -i 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
A Pre-ERA was conducted to narrow the scope of the ERA by identifying contaminants, other
than tPCBs, that pose potential risks to aquatic biota and wildlife in the PSA (Appendix B). A
three-tiered deterministic approach was used to screen COPCs. The deterministic assessments
compared conservative estimates of potential exposure with conservative adverse effects
benchmarks to identify which contaminants are of potential concern to piscivorous birds in the
Housatonic River. A hazard quotient (total daily intake/effect benchmark) greater than 1 for
piscivorous birds in the Housatonic River area resulted in the COPC being screened through to
the next tier assessment and to the probabilistic ERA, if necessary. Subsequent to the Pre-ERA,
several other COPCs (primarily organochlorine pesticides) were screened out because their
actual concentrations in the PSA were likely much lower than the measured values due to
laboratory interference (see Section 2.4).
In summary, the COCs that were retained for the probabilistic risk assessment for piscivorous
birds were total PCBs (tPCBs) and 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TCDD) toxic
equivalence (TEQ). Total PCBs detected in Housatonic River media closely resemble the
commercial PCB mixtures Aroclor 1260 and Aroclor 1254, which are similar in congener
makeup. TEQ is calculated from coplanar PCB and dioxin and furan congeners using the toxic
equivalency factor (TEF) approach developed by Van den Berg et al. (1998) (see Section 6.4 of
the ERA and Appendix C.10).
8.1.1 Overview of Approach
A step-wise approach was used to assess the risks of tPCBs and TEQ to piscivorous birds in the
Housatonic River watershed. The four main steps in this process include:
1. Derivation of a conceptual model (Figure 8.1-1).
2. Assessment of exposure of birds to COCs (Figure 8.1-2).
3. Assessment of the effects of COCs on birds (Figure 8.1-3).
4. Characterization of risk to the piscivorous avian species (Figure 8.1-4).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
This section is organized as follows:
¦ Section 8.2 (Conceptual Model) - Describes the conceptual model for piscivorous
birds, including selection of representative species and establishment of assessment
and measurement endpoints.
¦ Section 8.3 (Exposure Assessment) - Describes the exposure model, input variables,
and techniques to propagate uncertainty. Also presented in this section are the input
parameters and exposure modeling results for belted kingfishers and ospreys.
¦ Section 8.4 (Effects Assessment) - Describes the potential effects to birds exposed to
tPCBs and TEQ. This section reviews the belted kingfisher field study conducted in
the PSA, as well as toxicity thresholds found in the literature.
¦ Section 8.5 (Risk Characterization) - Presents the lines of evidence addressed in the
risk characterization, followed by a discussion of the sources of uncertainty in this
assessment, and the conclusions regarding risks of tPCBs and TEQ to piscivorous
birds in the Housatonic River PSA.
The detailed ecological risk assessment for piscivorous birds is
provided in Appendix H.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC g ^ 7/10/2003
-------
Releases
Historic releases of PCBs from the
Pittsfield GE Facility and
surrounding disposal areas
r ^
T 1
r
0)
o
o
<0
"D
0)
Legend
Direct uptake
Trophic transfer
Both
Potential effects
Wetland and
surface water
discharge
Groundwater
discharge
Contaminated soil, sediment, water and biota
£
o
•*->
Q.
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
EXPOSURE
Figure 8.1-2 Overview of Approach Used to Assess Modeled Exposure of
Piscivorous Birds to Contaminants of Concern (COCs) in the Housatonic River
PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC g ^ 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
EFFECTS
Figure 8.1-3 Overview of Approach Used to Assess the Modeled Effects of
Contaminants of Concern (COCs) to Piscivorous Birds in the Housatonic River
PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC g g 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
RISK CHARACTERIZATION
Figure 8.1-4 Overview of Approach Used to Characterize the Risks of
Contaminants of Concern (COCs) to Piscivorous Birds in the Housatonic River
PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC g j 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
8.2 CONCEPTUAL MODEL
The conceptual model presented in Figure 8.1-1 illustrates the exposure pathways for piscivorous
birds exposed to tPCBs and TEQ in the PSA. Total PCBs and TEQ are persistent, lipophilic, and
hydrophobic. Therefore, they are bioaccumulated by aquatic and terrestrial biota directly
through the consumption of contaminated prey as part of the food chain (Haffner et al. 1994;
Senthilkumar et al. 2001). Piscivorous birds that reside, or partially reside, within the study area
are exposed to tPCBs and TEQ principally through diet and trophic transfer. Other routes of
exposure, considered to be less important to overall exposure, include inhalation, water
consumption, and sediment ingestion (Moore et al. 1999).
The problem formulation (see Section 2 of the ERA) identified the belted kingfisher (Ceryle
alcyon\ Figure 8.2-1) and osprey (Pandion haliaetus; Figure 8.2-2) as the representative species
for piscivorous birds potentially exposed to tPCBs and TEQ from consumption of contaminated
prey in the PSA. Kingfishers have been observed nesting and breeding in the PSA, while
observations of ospreys suggest that birds foraging in the PSA are transients. The PSA contains
suitable habitat for ospreys, with abundant prey, so there is a high likelihood that as the
Massachusetts and Connecticut osprey population continues to expand, they may nest in the
PSA. Great blue herons were also considered as a representative species, but were not included
in the assessment because, although productivity data for herons in the vicinity of the PSA are
available (MDFW 1979, 1980, 1981, 1982, 1983, 1984, 1985, 1986a,b, 1987, 1989, 1991, 1996),
only a few of the birds from the rookery forage in the study area. Estimating exposure to these
individual birds, and estimating exposure to any contaminants for the remainder of the herons in
the area, would be difficult. Therefore, effects to great blue heron productivity, or lack thereof,
observed in the field would be difficult to attribute to specific COCs.
Life history profiles for belted kingfishers and ospreys are summarized in the following text
boxes. Additional life history information on these species is provided in Appendix H.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
-------
1 The assessment endpoint that is the subject of this section is the survival, growth, and
2 reproduction of piscivorous birds in the Housatonic River PSA. The measurement endpoints
3 used to evaluate the assessment endpoint were based on the determination of the extent to which
4 the concentrations of PCBs and TEQ ingested in the diet will impact the survival, reproduction,
5 or growth of piscivorous birds. The assessment for piscivorous birds includes both a site-specific
6 field study of kingfisher reproductive success, and comparisons of modeled exposures to doses
7 reported in the literature to cause adverse effects.
8
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-9
-------
Life History of Belted Kingfisher
The belted kingfisher is a pigeon-sized member of the family Alcedinidae and is a
common bird in North America, excluding the far north and the higher elevations of
the Rocky Mountains.
Habitat - Prefers foraging areas with Clearwater and visibility unobstructed by
turbidity or aquatic vegetation. Size of territory depends upon the availability of prey,
ranging from 0.5 to 1.36 miles (0.8 to 2.2 km) of shoreline.
Diet - Principal prey is fish, but may also feed on berries and other small animals,
including mollusks, crustaceans, insects, amphibians, reptiles, young birds, and
small mammals.
Figure 8.2-1 Belted Kingfisher (Ceryle alcyon)
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-10
-------
2
3
4
5
6
7
8
9
10
11
12
13
14
Figure 8.2-2 Osprey (Pandion haliaetus)
Life History of Osprey
Ospreys, also known as fish hawks or fishing eagles, are the only species in the
family Pandionidae, The range of these raptorial birds covers almost all of North
America, except for the extreme north.
Habitat - Use both fresh and saltwater ecosystems, but primarily the latter (Rattner
et al. 2001). Ospreys are tree nesters, but have also adapted to man-made
structures. Foraging ranges from 1.7 km to 15 km, depending on prey availability.
Diet - Almost exclusively piscivorous, preferring medium-sized fish (13 to 40 cm).
On rare occasions, osprey will take dead fish or prey on small mammals, reptiles,
and crustaceans.
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-11
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
8.3 EXPOSURE ASSESSMENT
The exposure assessment for piscivorous birds focuses on the PSA. Where possible, exposure
assessments were also conducted for two reference locations: East Branch of the Housatonic
River, upstream of Dalton (termed "upstream reference area"), and Threemile Pond in Sheffield,
MA. This section describes the general model used to estimate exposure of the two
representative species to tPCBs and TEQ in the PSA, as well as the inputs used for each
representative species. A summary of the exposure analyses results concludes the section.
8.3.1 Exposure Model
The exposure model for piscivorous birds focuses on the ingestion of tPCBs and 2,3,7,8-TCDD
TEQ through the diet. Other exposure routes (e.g., water, air) were considered to be of much
less importance for tPCBs and TEQ, and were excluded from the analyses. The equation used to
estimate exposure was adapted from the Wildlife Exposure Factors Handbook (EPA 1993) and
related publications:
n
TDI = FTx FIRYj C, x I] (Eq. 1)
2 = 1
where
TDI = Total daily intake (mg/kg bw/d tPCBs, ng/kg bw/d TEQ).
FT = Foraging time in the PSA (unitless).
FIR = Normalized food intake rate (kg/kg bw/d).
Ci = Concentration in the z'th prey item (mg/kg for tPCBs, ng/kg for TEQ).
Pi = Proportion of the z'th prey item in the diet (unitless).
Because of differences in the size of their foraging ranges, exposure analyses for kingfishers
were conducted separately for Reaches 5 and 6, and analyses for ospreys were conducted for the
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-12
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
PSA as a whole. The upstream reference area and Threemile Pond reference area were included
for comparative purposes.
Monte Carlo and probability bounds analyses were used to propagate input variable uncertainties
through the exposure model for each COC. Descriptions of these techniques and the methods
used to parameterize input variables are presented in Section 6.5. The results of the Monte Carlo
analysis are used to estimate the probability of exposure exceeding an effects threshold or levels
that cause adverse effects of differing magnitudes. The probability bounds analysis is conducted
to determine how uncertainty regarding the distributions of the input variables influences the
estimated exposure distribution. The results of these analyses are discussed in detail in
Appendix H.
Input distributions to the exposure analyses were generally assigned as follows:
¦ Lognormal distributions for variables that were right skewed with a lower bound of
zero and no upper bound (e.g., amount of COC transferred from mother to offspring
via egg tissue for tree swallows).
¦ Beta distributions for variables bounded by zero and one (e.g., proportion of a prey
item in the diet).
¦ Normal distributions for variables that were symmetric and not bounded by one (e.g.,
body weight).
¦ Point estimates for minor variables or variables with low coefficients of variation.
In certain situations (e.g., poor fit of data), other distributions were fit to the data or other
approaches were used.
The input variable distributions used in the exposure models for piscivorous birds are depicted in
Figures 8.3-1 and 8.3-2 and are summarized in the following sections. These distributions are
also presented in greater detail in Appendix H.
8.3.1.1 Foraging Time (FT)
The foraging ranges of the two representative species are within the size of the PSA. Prey
availability and an abundance of suitable foraging habitat suggest that the birds that forage in the
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
-------
1 PSA are able to meet their needs exclusively within this section of the river and floodplain. The
2 foraging range of kingfishers is relatively small (Salyer and Lagler 1946; Brooks and Davis
3 1987). Although the foraging radius of ospreys may be as large as 16 km (Clark 1995), foraging
4 range is reduced when prey are readily available near the nesting site (Clark 1995; Newton
5 1979). The assessment of risk to piscivorous birds inhabiting the PSA of the Housatonic River
6 therefore focuses on those birds that spend 100% of their time foraging within the PSA.
7 Foraging time is a point estimate and is not shown in Figures 8.3-1 and 8.3-2.
8 8.3.1.2 Body Weight (BW)
9 Body weights of belted kingfishers vary only slightly with sex (Hamas 1994). Dunning (1993)
10 reported a body weight range of 125 to 215 g with a mean of 148 g. Mean body weights have
11 been reported close to this value by other investigators as well (Alexander 1977; Salyer and
12 Lagler 1946; Brooks and Davis 1987). The distribution of the body weights of belted kingfishers
13 is depicted in Figure 8.3-1.
14 Female ospreys are generally larger than males, weighing an average of 1.6 kg and 1.4 kg,
15 respectively (Rattner et al. 2001). Poole (1985) studied ospreys in Massachusetts and found that
16 females ranged from 1.7 to 1.9 kg in weight during the breeding season, while males were about
17 1.4 kg. Brown and Amadon (1968) observed body weights of 1.2 to 1.9 kg for females and 1.2
18 to 1.6 kg for males in Nova Scotia. The distribution of the body weights of ospreys is depicted
19 in Figure 8.3-2.
20 8.3.1.3 Food Intake Rate (FIR)
21 The food intake rate of belted kingfishers has not been well characterized. Food ingestion rate
22 data were available in EPA (1993), however this information was not appropriate for use in this
23 exposure model. These data lacked body weight information, lacked statistical analyses, were
24 estimates themselves, and/or were collected from young or captive birds. The field-based
25 estimate of the daily food intake rate of free-living adult kingfishers (0.50 g/g-day, Alexander
26 1977) was close to the 30th percentile of the modeled food intake rate (see below) for these birds.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-14
-------
Body weight
Body Weight (g)
FMR slope term (a)
Proportion fish in diet
Proportion crayfish in diet
Proportion {"A)
Figure 8.3-1 Exposure Model Input Distributions for Belted Kingfisher
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
-------
Body weight
Body Weight (g)
FMR slope term (a)
FMR slope term (a)
Figure 8.3-2 Exposure Model Input Distributions for Osprey
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC g jg 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
Nagy (1987) and Nagy et al. (1999) derived allometric equations for estimating the metabolic
rate of free-living birds using the following general equation:
FMR (kJ/day) = a x BW(g)b (Eq. 2)
The slope (a) and power (b) terms in Equation 2 were based on the error statistics derived from
regression analysis of the data reported in Nagy et al. (1999). There were insufficient data to
generate an allometric equation for Coraciiformes, of which belted kingfishers are members, so
the equation for all birds was used. The slope term log a had a mean of 1.02 and a standard error
of 0.0393 in logio units, and the slope term b had a mean of 0.681 and a standard error of 0.0182
(Nagy et al. 1999).
The food intake rate of ospreys has not been well characterized either. Food ingestion rate data
were available in EPA (1993), however this information was not appropriate for use in this
exposure model. These data lacked body weight information, lacked statistical analyses, were
estimates themselves, and/or were collected from young or captive birds. The field-based
measurements of the daily food intake rate of adult male ospreys (0.21 g/g-day, Poole 1983)
were close to the 25th percentile of the modeled food intake rate described below. There were
insufficient data to generate an allometric equation for Falconiformes, of which ospreys are
members, so the equation for Charadriiformes was used. This Order includes many piscivorous
birds and was thought to be a suitable surrogate group. The slope term log a had a mean of
0.928 and a standard error of 0.197 in logio units, and the power term b had a mean of 0.768 and
a standard error of 0.0874 (Nagy et al. 1999). These input variable distributions are depicted in
Figures 8.3-1 and 8.3-2. The body weights (BW) for these birds are described above. The results
of the calculation were then converted to kcal/kg bw/d.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-17
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Food intake rate is derived from FMR using the following equation:
n
FIR {kg / kgbw / day) = FMR / ^ x GEj) (Eq. 3)
Z = 1
where AEt is the assimilation efficiency of the z'th food item (unitless) and GEt is the gross energy
in the rth food item (kcal/g). For kingfishers, mean assimilation efficiencies were 77% for
aquatic invertebrates and 79% for fish, and the mean gross energies were 1.1 kcal/g wet weight
for aquatic invertebrates and 1.2 kcal/g wet weight for fish. For ospreys, the mean assimilation
efficiency of fish was 79% and the mean gross energy of fish was 1.2 kcal/g wet weight (EPA
1993). Point estimates were used for AEt and GEj in the Monte Carlo and probability bounds
analyses because of their relatively small coefficients of variation (i.e., CV <10%). As a result,
these input variables are not included in Figures 8.3-1 and 8.3-2.
8.3.1.4 Proportion of Dietary Items (Pj)
The principal prey of kingfishers is fish, but they also feed on berries and a variety of other small
animals, including mollusks, crustaceans, insects, amphibians, reptiles, young birds, and small
mammals (Hamas 1994). Fish prey species are those that typically live in shallow water or near
the surface (Hamas 1994) and include trout, salmon, suckers, perch, minnows, killifish,
sticklebacks, and others (EPA 1993). Fish and crayfish are the primary prey items for
kingfishers, with other items expected to contribute little to the diet. The exposure model,
therefore, uses a diet with a mean composition of 86% fish and 14% crayfish in Reach 5.
Distributions of the proportion of kingfishers' dietary items in Reach 5 are depicted in Figure
8.3-1. In Woods Pond and Threemile Pond, crayfish were not included in the kingfisher diet.
The primary reasons for this include:
¦ The lack of observations of crayfish when conducting other field surveys during the
last 3 years at these locations.
¦ The presence of aquatic vegetation, which conceals crayfish from kingfishers.
¦ The abundance of cyprinids and centrachids of forage size, which live in the shallow
areas and are visually attractive to hunting kingfishers.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-18
-------
1 Given these factors, it was assumed that fish would replace crayfish in the diet of kingfishers
2 foraging in the areas around Woods Pond and Threemile Pond. Therefore, the percent
3 contribution of fish in the diet was assumed to be 100%, and is not depicted in Figure 8.3-1.
4 Ospreys prefer to forage in shallow waters in lakes and rivers where fish occur near the surface
5 and may be easily seen (DeGraaf and Yamasaki 2001). The birds are almost exclusively
6 piscivorous, preferring medium-sized fish (13 to 40 cm in length) (Vana-Miller 1987). On rare
7 occasions, ospreys will take dead fish or prey on small mammals, reptiles, and crustaceans
8 (Chubbs and Trimper 1998). For this exposure assessment, it was assumed that fish account for
9 100% of the osprey diet. Proportion of fish in the osprey diet is a point estimate and therefore, is
10 not shown in Figure 8.3-2.
11 8.3.1.5 Concentration of COCs in Dietary Items (Cj)
12 The concentrations of tPCBs and TEQ in the prey of piscivorous birds are illustrated in Figures
13 8.3-3 to 8.3-6. The bars in these figures depict the median concentration of each COC in each
14 dietary item in each of the areas for the risk assessment. The stars depict the arithmetic mean
15 and the vertical lines depict the interquartile ranges for the concentrations. The concentrations of
16 COCs used in the exposure analyses are shown in Tables H.2-4, H.2-5, H.2-12, and H.2-13 of
17 Appendix H. Rationales for the concentration variables are also provided in Appendix H. As
18 evident in these figures, the concentrations of tPCBs and TEQ in prey items of piscivorous birds
19 are highest within the PSA and substantially lower in the reference areas.
20 8.3.2 Exposure Model Results
21 The exposure model results for belted kingfishers and ospreys exposed to tPCBs and TEQ are
22 discussed in greater detail in Section 2 of Appendix H.
23 Belted kingfishers had the highest modeled exposure to tPCBs in Reach 5 (Figure 8.3-7) and
24 Reach 6 (Figure 8.3-8), whereas exposures in the reference areas (Figures 8.3-9 and 8.3-10) were
25 substantially lower. Belted kingfishers had the highest modeled exposure to TEQ in Reach 5
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-19
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
100
~d1d
E
M
U
0.01
0*
c?
0*
~ Median
X Arithmetic Mean
75 Percentile
25th Percentile
ea
#
y>
<#
Prey item and location
Notes: Error bars indicate interquartile range.
URA = Upstream reference area; TMP = Threemile Pond reference area.
Figure 8.3-3 Concentrations of tPCBs in the Prey of Belted Kingfishers in the
Housatonic River PSA and Reference Areas
1 0
(f
&
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
100
10
M
U
Ph
<#•
~ Median
X Arithmetic Mean
75 Percentile
25 th Percentile
*
Prey item and location
Notes: Error bars indicate interquartile range.
URA = Upstream reference area; TMP = Threemile Pond reference area.
Figure 8.3-5 Concentrations of tPCBs in the Prey of Ospreys in the Housatonic
River PSA and Reference Areas
10000
=* 1000
Dt)
100
$
~ Median
X Arithmetic Mean
75 Percentile
25th Percentile
4*
Prey item and location
Notes: Error bars indicate interquartile range.
URA = Upstream reference area; TMP = Threemile Pond reference area.
Figure 8.3-6 Concentrations of TEQ in the Prey of Ospreys in the Housatonic
River PSA and Reference Areas
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-21
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
(Figure 8.3-11) and Reach 6 (Figure 8.3-12), whereas exposures in the reference areas (Figures
8.3-13 and 8.3-14) were again substantially lower.
The exposure distributions for ospreys exposed to tPCBs in the PSA, upstream reference area,
and the Threemile Pond reference area are presented in Figures 8.3-15, 8.3-16, and 8.3-17,
respectively. The exposure distributions for ospreys exposed to TEQ in the PSA, upstream
reference area, and the Threemile Pond reference area are presented in Figures 8.3-18, 8.3-19,
and 8.3-20, respectively. Ospreys had the highest modeled exposure to tPCBs and TEQ in the
PSA, while exposures in the reference areas were substantially lower. The differences in the
exposure estimates between the PSA and reference areas may be explained in great part by the
differences in COC concentrations in prey items at the sites, as depicted in Figures 8.3-3 to 8.3-6.
Overall, ospreys foraging in the PSA are expected to have the highest exposure to tPCBs and
TEQ of the representative piscivorous bird species, followed by belted kingfishers. Tables H.2-6
and H.2-14 present a summary of tPCB exposure model results for belted kingfishers and
ospreys, respectively. Results for the TEQ exposure model for piscivorous birds are presented in
Tables H.2-7 and H.2-15, respectively. A complete account of the exposure model results,
including Monte Carlo and probability bounds analyses and figures and tables, is presented in
Appendix H.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-22
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Reach 5
Monte Carlo
Dose (mg/kg bw/d)
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-7 Exposure of Belted Kingfishers to tPCBs in Reach 5 of the
Housatonic River PSA
Reach 6
g,
-D
C3
-D
100 r
60
S 40
u 20
0.01
—I I I I I I I I I I I I I I I I I I I I I I I I I I
0.1
1
Dose (mg/kg bw/d)
10
Monte Carlo
LPB
UPB
• Low-intenned.
criterion
O Intermed.-high
criterion
_l I l l l l I
100
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-8 Exposure of Belted Kingfishers to tPCBs in Reach 6 of the
Housatonic River PSA
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-23
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Upstream Reference Area
100 r
g,
c
«
¦s
*
0.01
0.1
Monte Carlo
LPB
UPB
# Low-intemied. criterion
O Intermed.-high criterion
o
I I I I I I I I I
I I I I I I I I I
1
Dose (mg/kg bw/d)
10
100
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-9 Exposure of Belted Kingfishers to tPCBs in the Housatonic River
Upstream Reference Area
Threemile Pond
0.01
Monte Carlo
LPB
UPB
• Low-intenned. criterion
O Intermed.-high criterion
o
I I I I I I I I
I I I I I I I I I
o.i
i
Dose (mg/kg bw/d)
10
100
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-10 Exposure of Belted Kingfishers to tPCBs in the Threemile Pond
Reference Area
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-24
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Reach 5
100
g
Sfi
60
40
20
I I I I I I I I I
^"Monte Carlo
LPB
UPB
• Low-intermed. criterion
O Interned.-high criterion
o
I I I I I I I 11
I I I I I I I 11
10
100 1000
Dose (ng/kg bw/d)
10000
100000
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-11 Exposure of Belted Kingfishers to TEQ in Reach 5 of the
Housatonic River PSA
Reach 6
100 r
g
¦a
«
.a
o
W
60
40
20
10
100 1000
Dose (ng/kg bw/d)
LPB = Lower probability bound
UPB = Upper probability bound
Monte Carlo
LPB
UPB
• Low-intermed. criterion
O Intermed.-high criterion
o
10000
100000
Figure 8.3-12 Exposure of Belted Kingfishers to TEQ in Reach 6 of the
Housatonic River PSA
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-25
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Upstream Reference Area
Monte Carlo
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-13 Exposure of Belted Kingfishers to TEQ in the Housatonic River
Upstream Reference Area
Threemile Pond
100 r
£
-D
C3
-D
O
*
60
40
20
Monte Carlo
LPB
UPB
• Low-intenned. criterion
O Intenned.-liigh criterion
o
10
100 1000
Dose (ng/kg bw/d)
10000
100000
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-14 Exposure of Belted Kingfishers to TEQ in the Threemile Pond
Reference Area
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-26
-------
1
2
3
4
5
6
7
8
9
0
1
Primary Study Area
100
-D
-D
O
*
w
60
40
20
0
Monte Carlo
LPB
UPB
• Low-intemied. criterion
O Intermed.-high criterion
Area
0.001 0.01 0.1 1
Dose (mg/kg bw/d)
10
100
1000
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-15 Exposure of Ospreys to tPCBs in Reaches 5 and 6 of the
Housatonic River PSA
Upstream Reference Area
g,
-D
C3
-D
O
C
«
¦s
100
80
60
40
20
Monte Carlo
LPB
UPB
• Low-intermed. criterion
O Interned.-high criterion
o
o
o.ooi
0.01
0.1
1
Dose (mg/kg bw/d)
10
100
1000
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-16 Exposure of Ospreys to tPCBs in the Housatonic River Upstream
Reference Area
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-27
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
Threemile Pond
0.001 0.01 0.1 1 10 100 1000
Dose (mg/kg bw/d)
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-17 Exposure of Ospreys to tPCBs in the Threemile Pond Reference
Area
Primary Study Area
0.1 1 10 100 1000 10000 100000
Dose (ng/kg bw/d)
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-18 Exposure of Ospreys to TEQ in Reaches 5 and 6 of the Housatonic
River PSA
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-28
-------
1
2
3
4
5
6
7
8
9
10
11
12
Upstream Reference Area
£
-D
C3
-D
p
*
0.1
^"Monte Carlo
LPB
UPB
• Low-intermed. criterion
O Intermed.-high criterion
o
10 100 1000
Dose (ng/kg bw/d)
10000
100000
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-19 Exposure of Ospreys to TEQ in the Upstream Reference Area
Threemile Pond
0.1
Monte Carlo
LPB
UPB
# Low-intermed. criterion
O Intermed.-high criterion
o
I I I I I 111
I I I I 111 I I I I I 1111
10 100 1000
Dose (ng/kg bw/d)
10000
100000
LPB = Lower probability bound
UPB = Upper probability bound
Figure 8.3-20 Exposure of Ospreys to TEQ in the Threemile Pond Reference Area
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-29
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
8.4 EFFECTS ASSESSMENT
The purpose of the effects assessment is twofold. The first is to review the scientific literature for
effects of PCB mixtures (mainly Aroclor 1254 and 1260 mixtures) and TEQ to piscivorous birds.
Of primary interest are documented effects to the representative species in this assessment:
ospreys and belted kingfishers. In the absence of data for these species, other avian species were
considered. The other purpose is to derive the effects metrics that will be used, in conjunction
with the exposure assessment results, to estimate risks to piscivorous birds from exposure to
COCs in the Housatonic River PSA.
The COCs in this assessment are tPCBs and 2,3,7,8-TCDD toxic equivalence (TEQ). The
congeners used to estimate TEQ concentrations share the ability to bind to the aryl hydrocarbon
(Ah) receptor protein (Bosveld and Van den Berg 1994) and elicit an Ah-receptor-mediated
biochemical and toxic response. The toxic response to this group of chemicals is related to the
three-dimensional structure of the substance, including the degree of chlorination and positions
of the chlorine on the aromatic frame (Van den Berg et al. 1998; Newsted et al. 1995; Safe
1994).
Sensitivities of avian species to tPCBs and TEQ have been shown in the literature studies to vary
greatly. Wild turkey embryos were found to be 20 to 100 times less sensitive than chicken
embryos to the egg yolk injection of PCB-77. This difference in toxicity may be attributed to
differences in availability of Ah receptors. Ah receptors were found in hepatocytes of 7-day-old
chicken embryos, but not in liver cells of 9-day-old turkey embryos (Brunstrom and Lund 1988).
Figures 8.4-1 and 8.4-2 illustrate the ranges of effects of tPCBs and TEQ, respectively, to
various avian species.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-30
-------
1
Toxicity of tPCBs and TEQ to Avian Species
2
Mode of Action
3
Binding to the aryl-hydrocarbon (Ah) receptor, eliciting an Ah-receptor-mediated
4
biochemical and toxic response.
5
Types of Toxicity
Specific Effects
6
hepatotoxicity
mortality
7
immunotoxicity
decreased growth
8
neurotoxicity
weight loss
9
embryotoxicity
porphyria
10
teratogenicity
reduced hatching
11
embryo mortality
12
13 8.4.1 Total PCBs
14 Heath et al. (1972) studied the effects of Aroclor 1254 on mortality on four avian species. The
15 most sensitive (after oral dosing for 5 days with Aroclor 1254) was the bobwhite quail, with a
16 median lethality response occurring at a dietary PCB concentration of 604 mg/kg. Other species
17 tested, such as the Japanese quail, mallard duck, and ring-necked pheasant, were less sensitive,
18 with oral LC50S of 2,898, 2,699, and 1,091 mg/kg diet, respectively. Prestt et al. (1970)
19 estimated the median lethal dietary dose rate of Aroclor 1254 to adult Bengalese finches to be
20 256 mg/kg/d.
21 Reproductive impairment of birds caused by tPCBs has been investigated in several species, in
22 dietary and egg injection studies, as well as field studies examining egg and hatchling
23 concentrations and hatching success. The most commonly noted effects to the reproduction of
24 avian species are reduced egg productivity, egg hatchability, and chick growth rates (CCME
25 1999). Of the species studied, chickens appear to be the most sensitive, followed by pheasants,
26 turkeys, ducks, and herring gulls (Bosveld and Van den Berg 1994). Total PCBs appear to have
27 no adverse effects on total egg weight, eggshell weight, or eggshell thickness (Lillie et al. 1974;
28 Britton and Huston 1973; Scott 1977). Lillie et al. (1974) exposed hens to 2 mg/kg Aroclor 1254
29 for 63 days in feed, giving birds a daily PCB dose of approximately 0.12 mg/kg bw/d. At this
30
31
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC 8 3 1 7/10/2003
-------
1000
i253.6
253.6 Bengalese Finches estimated
LD50; Avg brain cone - 290 mg/kg;
avg liver cone - 345 mg/kg. Only
treatment.
Figure 8.4-1 Effects of Aroclor 1254/1260 on Avian Species (mg/kg bw/d)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-32
-------
10
Figure 8.4-2 Effects of 2,3,7,8-TCDD TEQ on Avian Species (ng TEQ/kg bw/d)
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-33
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
treatment level, no significant effects were observed on fertility, egg production, shell thickness,
or hatchability, but the growth rate of chicks was slightly reduced. American kestrels were given
an approximate tPCB dose of 7 mg/kg bw/d for 100 days (Fernie et al. 2001) and birds
experienced a slight, but statistically insignificant, decrease in clutch size and the numbers of
fertile eggs, hatchlings, and fledglings per breeding pair.
8.4.2 TEQ
In single, oral doses of TCDD, bobwhite quail, mallards, and ringed turtledoves were found to
have 37-day LD50s of 15,000, 108,000, and 810,000 ng/kg bw (Hudson et al. 1984). Ring-
necked pheasants given intraperitoneal TCDD doses weekly of 10, 100, or 1,000 ng/kg bw/week
for 10 weeks (Nosek et al. 1992) experienced no mortality or body weight effects in the two
lowest treatment groups, but the 1,000 ng/kg bw/week treatment group experienced 60%
mortality by the 23rd week (13 weeks after the dosing period).
Nosek et al. (1992a) also investigated reproductive effects of TCDD to ring-necked pheasant
hens. The two lower doses (10 and 100 ng/kg bw/week) caused no significant impairment to egg
production. The highest dose, 1,000 ng/kg bw/week, caused a decrease in cumulative egg
production of approximately 70% over 7 weeks. The geometric mean of the lowest observed
adverse effect level (LOAEL) and no observed adverse effect level (NOAEL) from this study
was 44 ng/kg bw/d. Hoffman et al. (1996) investigated the developmental effects of TEQ on
American kestrels and observed that skeletal growth was significantly reduced at a treatment
level of 25,000 ng TEQ/kg bw/d. This dose did not translate into significant effects on hatchling
success or weight gain.
8.4.3 Effects Metrics
Effects data are ideally summarized as dose-response curves for each representative species. For
this assessment, however, data were insufficient to generate dose-response curves, NOAELs and
LOAELs, or field based measures of effect. Therefore, a threshold range for surrogate species
was used to represent the effects of tPCBs and TEQ to piscivorous birds. This approach
establishes a range of toxic effects thresholds for the most sensitive and tolerant avian species
known and assumes that the thresholds for the representative species are within these bounds.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC Q ^ /I 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Further details on the decision criteria used in selecting effects metrics are presented below and
in Section 6.6 of the ERA.
8.4.3.1 Total PCBs
Based on a review of avian toxicity literature, white leghorn chickens were the most sensitive
avian species to the reproductive effects of tPCBs (Lillie et al. 1974) and the most reproductively
tolerant avian species to tPCBs was the American kestrel (Fernie et al. 2001). The resulting
threshold range for the reproductive success of piscivorous birds exposed to tPCBs selected for
this assessment is 0.12 to 7.0 mg/kg bw/d.
Decision Criteria for Derivation of Effects Metric
The following is the hierarchy of decision criteria used to characterize effects for each
receptor-COC combination:
¦ Have single-study bioassays with five or more treatments been conducted on the
receptor of interest or a reasonable surrogate? If yes, estimate the
concentration- or dose-response relationship. If not, go to step 2.
¦ Are multiple bioassays with similar protocols, exposure scenarios, and effects
metrics available that, when combined, have five or more treatments for the
receptor of interest or a reasonable surrogate? If yes, estimate the dose-
response relationship as in step 1. If not, go to step 3.
¦ Have bioassays with less than five treatments been conducted on the receptor of
interest or a reasonable surrogate? If yes, conduct or report results of
hypothesis testing to determine the NOAEL and LOAEL. If not, go to step 4.
¦ Are sufficient data available from field studies and monitoring programs to
estimate concentrations or doses of the COC that are consistently protective or
associated with adverse effects? If yes, develop field-based effects metrics. If
not, go to step 5.
¦ Derive a range where the threshold for the receptor of interest is expected to
occur. Because information on the sensitivity of the receptor of interest is
lacking, it is difficult to derive a threshold that is neither biased high nor low. If
bioassay data are available for several other species, however, calculate a
threshold for each to determine a threshold range that spans sensitive and
tolerant species. That range is likely to include the threshold for the receptor of
interest.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-35
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
8.4.3.2 2,3,7,8-TCDD Toxic Equivalence (TEQ)
The lower toxicological threshold for the effects of TEQ to sensitive birds is based on ring-
necked pheasants (Nosek et al. 1992) and the upper threshold for tolerant species is based on the
American kestrel (Hoffman et al. 1996). The resulting threshold range for the reproductive
success of piscivorous birds exposed to TEQ is 44 to 25,000 ng/kg bw/d.
8.4.4 Belted Kingfisher Field Study
A belted kingfisher reproduction study was performed in the PSA during the 2002 breeding
season (Henning 2002). The objective of the study was to evaluate the relationship between
reproduction of kingfishers and exposure of adult and nestling kingfishers to tPCBs. Nine belted
kingfisher burrows were monitored during this study, three of which were depredated before the
young could fledge. In the remaining six nests, there was an average of 4.8 nestlings, or 87%,
that survived from egg to 26 days. When depredated nests were excluded, fledging rates were
consistent with the results of the only other kingfisher study reported in the literature (Brooks
and Davis 1987). Total daily intake of tPCBs was estimated based on prey concentrations and
food ingestion rates. No significant relationships were observed between estimated tPCB dose
and reproductive output (p>0.05), although this does not necessarily support a conclusion of no
adverse effects to the reproductive success of belted kingfishers. See Section 8.5.2 and Section
H.4.2 of Appendix H for discussion of this study.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-36
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
8.5 RISK CHARACTERIZATION
This section characterizes risks to piscivorous birds exposed to tPCBs and 2,3,7,8-TCDD TEQ in
the PSA of the Housatonic River. The risk characterization includes a comparison of
probabilistic exposure estimates to relevant effect metrics, a review of the findings of the belted
kingfisher field study, a summary of the weight-of-evidence (WOE) assessment, a discussion of
the sources of uncertainty that may influence the estimates of risk, and a discussion of risks to
other piscivorous birds foraging in the PSA.
Risk Questions
¦ Are the concentrations of tPCBs and TEQ present in the prey of piscivorous birds
sufficient to cause adverse effects to individuals inhabiting the PSA of the
Housatonic River?
¦ If so, how severe are the risks and what are their potential consequences?
Lines of Evidence
¦ Probabilistic exposure and effects modeling.
¦ Field study of the reproductive success of belted kingfishers in the PSA.
8.5.1 Comparison of Estimated Exposures to Laboratory-Derived Effect Doses
For piscivorous birds, exposure was assessed separately in Reaches 5 and 6 for belted kingfishers
and in Reaches 5 and 6 combined for ospreys. For each receptor-COC-area combination, a
category of low, intermediate, or high risk was assigned using the guidance below, when
integrating the exposure and effects distributions.
Guidance for Integrating the Exposure and Effects Distributions
¦ If the probability of exceeding the lower toxicity threshold is less than 20%, the
risk is considered to be low.
¦ If the probability of exceeding the upper toxicity threshold is greater than 20%,
the risk is considered to be high.
¦ All other outcomes are considered to have intermediate risk.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC 8 37 7/10/2003
-------
1 This exercise was done separately for the results of the Monte Carlo analyses and the lower and
2 upper bounds from the probability bounds analyses. The "risk category" refers to the level of
3 risk based on the results of the Monte Carlo analysis. The "risk range" refers to the levels of risk
4 based on the results of the probability bounds analyses.
5 The results of the risk characterization, as described by the modeled exposure and effects line of
6 evidence, are summarized in Table 8.5-1. The highest risk to piscivorous birds is from exposure
7 to tPCBs in the PSA, with low-intermediate risk in the reference areas. As shown in Figure
8 8.3-7, the tPCB exposure curve for ospreys in the PSA is well above the upper toxicity threshold.
9 This means that the estimated daily intake of tPCBs by ospreys is greater than the intake known
10 to cause adverse effects in the most tolerant bird species studied. The risk category for ospreys is
11 high in the PSA and the risk range is also high (Table 8.5-1).
12 The highest risk to kingfishers is from exposure to tPCBs in Reaches 5 and 6. Both
13 representative piscivorous bird species were determined to be at intermediate risk to TEQ in the
14 PSA. The complete characterization of risks of piscivorous birds to tPCBs and TEQ is presented
15 in Appendix H.
16 8.5.2 Belted Kingfisher Field Study
17 A study of belted kingfisher reproduction in the PSA was performed by GE during the 2002
18 breeding season. The objective of the study was to determine the relationship between tPCB
19 dose and reproductive success. More information on this study can be found in Henning (2002).
20 Active kingfisher burrows were sought in the river banks and riparian zone of the PSA. In each
21 burrow, the number of nestlings and parental behavior were recorded. According to the report,
22 the total daily intake of tPCBs by kingfishers was estimated based on the concentration of COCs
23 in fish and crayfish samples taken within 1,200 m of each burrow and on prey ingestion rates of
24 adults and nestlings obtained from the literature. Crayfish samples were associated with specific
25 sampling locations, but fish samples were not. When developing the estimated dose, Henning
26 assumed a more precise location of a fish sample than is possible from the information
27 associated with the sample. River miles designated in the sample IDs in the database were
28
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC 8 38 7/10/2003
-------
1 Table 8.5-1
2
3 Summary of Qualitative Risk Statements for Piscivorous Birds from the
4 Housatonic River PSA
Bird / Location
Qualitative Risk Statements
tPCBs
TEQ
Risk Category
Risk Range
Risk Category
Risk Range
Belted Kingfisher
Reach 5
High
High
Intermediate
Intermediate
Reach 6
High
High
Intermediate
Intermediate
Upstream Reference
Area
Intermediate
Low/intermediate
Intermediate
Low/intermediate
Threemile Pond
Low
Low
Low
Low
Osprey
Reaches 5 and 6
High
High
Intermediate
Intermediate
Upstream Reference
Area
Intermediate
Intermediate
Low
Low
Threemile Pond
Low
Low
Low
Low
5
6
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-39
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
representative of larger areas of the river, roughly corresponding to a segment of Reach 5A
referred to in sampling records as "Shallow Reach" or Reaches 5B and 5C combined ("Deep
Reach").
The total daily intake calculation did not result in a dose gradient necessary to evaluate a dose-
response relationship for piscivorous birds consuming Housatonic River fish. First, the fish
sampling location was known only at the resolution of a range of river miles (i.e., a river reach
level); therefore, an actual river mile cannot be assumed for a sample. Second, fish species
sampled may be mobile between life stages and seasons, and have integrated contaminant
concentrations across these areas. These factors contribute to an averaging of the tPCB
concentrations in prey used to evaluate exposure by kingfishers, and result in a very narrow
exposure gradient, with total daily intake for adult birds ranging from 7.4 to 21 mg/kg bw/d in
the GE kingfisher field study (Henning 2002).
Although nine belted kingfisher burrows were monitored during this study, three of which were
depredated before the fledging date. For the remaining nests, there was an average of 4.8
nestlings, or 87%, that survived to 26 days. If depredated nests were included in the analysis, the
average decreased to 3.9 young per nest, or 58% surviving to 26 days. Estimated tPCB doses for
adults and young in the PSA were 13 and 35 mg/kg bw/d, respectively. Doses ranged from 7.4
to 21 mg/kg bw/d for adults and 20 to 57 mg/kg bw/d for nestlings. Henning (2002) reported
that there were no significant relationships between estimated tPCB dose and any of the
endpoints (p>0.05). The range of estimated total daily intakes in the PSA was narrow, however,
and could not be replicated from the data, providing insufficient basis on which to evaluate a
dose-response relationship. Multivariate models also indicated that combined independent
variables (e.g. nest density, tPCB dose) provided no significant relationship between stressors
and reproductive effects. The results were similar when depredated nests were included in the
analyses. The results of this analysis are similarly confounded as in the previously described
analyses, because of the same range of daily tPCB intakes in the PSA.
The kingfisher population in the Housatonic River appears to be breeding successfully, with
fledging rates and population densities that, when depredated nests are excluded, are similar to
what was reported in the only other comparable study from the literature (Brooks and Davis
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-40
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1987). The lack of data from reference areas for comparison to the observations in the PSA
introduces uncertainty in interpretation of this study. Small sample size also introduces some
uncertainty, as only nine nests, six of which were successful, were observed in the study. The
model used to estimate the total daily intake of tPCBs has limited applicability as it was not
possible to attain a sufficiently wide dose gradient. One additional nest was observed during
oversight of the study by EPA contract staff (Woodlot Alternatives, Inc. 2002), but was
apparently not included in the study.
8.5.3 Weight-of-Evidence Analysis
A WOE analysis procedure was used to assess risks of tPCBs and TEQ to piscivorous birds. The
goal of this analysis was to determine whether significant risk is posed to piscivorous birds in the
Housatonic River PSA as a result of exposure to tPCBs and TEQ. The three-phase approach of
Menzie et al. (1996) and the Massachusetts Weight-of-Evidence Workgroup was applied for this
purpose, in which WOE was reflected in the following three characteristics: (1) the weight
assigned to each measurement endpoint; (2) the magnitude of response observed in the
measurement endpoint; and (3) the concurrence among outcomes of the multiple measurement
endpoints (see Section 2.9 for details).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-41
-------
1 The rationale for evaluating measurement endpoints is provided in Section 2.9 and Appendix H,
2 along with a discussion of attributes considered in the WOE. The measurement endpoint
3 weighting scores are presented in Table 8.5-2 and evidence of harm and magnitudes of responses
4 are presented in Tables 8.5-3 and 8.5-4 for tPCBs and TEQ, respectively. For both tPCBs and
5 TEQ, the modeled exposure and effects line of evidence was given a moderate weighting. The
6 belted kingfisher field study was given a moderate/high weighting.
7 Table 8.5-2
8
9 Weighting of Measurement Endpoints for Piscivorous Birds Weight-of-Evidence
10 Evaluation
Attributes
Modeled Exposure
and Effects for
tPCBs and TEQ
GE Kingfisher Field
Study
(Henning 2002)
I. Relationship Between Measurement and Assessment Endpoints
1. Degree of Association
M
H
2. Stressor/Response
M
M
3. Utility of Measure
M
M
II. Data Quality
4. Data Quality
M/H
M
III. Study Design
5. Site Specificity
L/M
M/H
6. Sensitivity
L/M
L/M
7. Spatial Representativeness
M
M/H
8. Temporal Representativeness
M
M
9. Quantitative Measure
M/H
M
10. Standard Method
M
M/H
Overall Endpoint Value
M
M/H
11 L = low
12 M = moderate
13 H = high
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-42
-------
1 Table 8.5-3
2
3 Evidence of Harm and Magnitude of Effects for Piscivorous Birds Exposed to
4 tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled Exposure and
Effects
M
Yes (Kingfisher)
Yes (Osprey)
High (Kingfisher)
High (Osprey)
Belted Kingfisher Field
Study (Henning 2002)
M/H
No (Kingfisher)
Low (Kingfisher)
5
6 Table 8.5-4
7
8 Evidence of Harm and Magnitude of Effects for Piscivorous Birds Exposed to
9 TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled Exposure and
Effects
M
Yes (Kingfisher)
Yes (Osprey)
Intermediate (Kingfisher)
Intermediate (Osprey)
Belted Kingfisher Field
Study (Henning 2002)
M/H
No (Kingfisher)
Low (Kingfisher)
10
11 The magnitude of the response in the measurement endpoint is considered together with the
12 measurement endpoint weight in judging the overall WOE (Menzie et al. 1996). This requires
13 assessing the strength of evidence that ecological harm has occurred, as well as an indication of
14 the magnitude of response, if present. For both tPCBs (Table 8.5-3) and TEQ (Table 8.5-4), the
15 modeled exposure and effects line of evidence indicated that there was evidence of harm, and
16 that the magnitude was high. The belted kingfisher field study (Henning 2002) indicated that
17 there was no evidence of adverse effects to productivity, and the magnitude was low.
18 A graphical method was used for displaying concurrence among measurement endpoints. Tables
19 8.5-5 and 8.5-6 depict the outcome for belted kingfishers and ospreys exposed to tPCBs and
20 TEQ. The analyses were conducted separately for ospreys and belted kingfishers exposed to
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-43
-------
1 Table 8.5-5
2
3 Risk Analysis Summary for Piscivorous Birds Exposed to tPCBs in the
4 Housatonic River PSA
5
Assessment Endpoint:
Survival, growth, and reproduction of piscivorous birds.
Weighting Factors (increasing confidence of weight)
Harm/Magnitude
Low
Low/
Moderate
Moderate
Moderate/
High
High
Yes/High
MEE-KF
MEE-0
Yes/Intermediate
Yes/Low
A
i
J
Undetermined/High
Undetermined/Intermediate
Undetermined/Low
8
No/Low
FS-KF
No/Intermediate
No/High
l
~
9
10 MEE = Modeled exposure and effects
11 FS = Field study
12 KF = Kingfisher
13 O = Osprey
14
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-44
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Table 8.5-6
Risk Analysis Summary for Piscivorous Birds Exposed to TEQ in the Housatonic
River PSA
Assessment Endpoint:
Survival, growth, and reproduction of piscivorous birds.
Weighting Factors (increasing confidence of weight)
Harm/Magnitude
Low
Low/
Moderate
Moderate
Moderate/
High
High
Yes/High
Yes/Intermediate
MEE-KF
MEE-0
Yes/Low
Undetermined/High
Undetermined/Intermediate
Undetermined/Low
No/Low
FS-KF
No/Intermediate
No/High
I
MEE = Modeled exposure and effects
FS = Field study
KF = Kingfisher
O = Osprey
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-45
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
tPCBs and TEQ. There is moderate confidence, because only one line of evidence is available,
that ospreys foraging in the PSA are subject to risk from exposure to tPCBs and TEQ (Appendix
H, Section 4.3). Belted kingfishers were judged, with moderate confidence, to be at low risk
from exposure to tPCBs and TEQ in the PSA based on the Kingfisher field study. Based on the
modeled exposure and effects analysis, there is moderate confidence that Kingfishers are
adversely affected by exposure to tPCBs and TEQ. Confidence in this conclusion is not high
because the two lines of evidence (modeled exposure and effects, kingfisher field study)
produced conflicting risk estimates. Risks in the reference areas for these COCs are generally
low.
8.5.4 Sources of Uncertainty
The assessment of risk to piscivorous birds contains uncertainties. Each source of uncertainty
can influence the estimates of risk, therefore, it is important to describe and, when possible,
specify the magnitude and direction of such uncertainties. The sources of uncertainty associated
with the assessment of risks of tPCBs and TEQ to piscivorous birds are described as follows.
¦ The Monte Carlo sensitivity analyses suggested that the free metabolic rate (FMR)
slope and power terms were generally the most influential variables on predicted total
daily intakes of COCs. However, no suitable direct measurements of free metabolic
rate are available for the representative wildlife species. Similarly, suitable measured
food intake rates are not available for free-living belted kingfisher and osprey.
Therefore, free metabolic rates were estimated using allometric equations. The use of
allometric equations introduces some degree of uncertainty into the exposure
estimates because they are subject to model-fitting error, and are based on species
different from the representative species used in this assessment. Given the lack of
empirical data on species specific to this assessment, it is difficult to judge the
magnitude of the uncertainty introduced by the use of the allometric models. The
uncertainty due to model-fitting error was propagated in the uncertainty analyses by
using distributions as input for the allometric slope and power terms.
¦ Sample sizes were limited for the analyses of COC concentrations in some prey
items, specifically, crayfish. To address this uncertainty in the Monte Carlo analysis,
the UCL or data set maximum (see Section 6.4 and Appendix C.5) was used as an
estimate of COC concentrations in prey items. The potential magnitude of the
uncertainty associated with small sample sizes for COC concentrations is unknown,
but this approach likely overestimates exposure. The probability bounds analysis
used an unbiased approach (e.g., distribution free range from lower confidence limit
[LCL] to upper confidence limit [UCL]) to deal with sample size uncertainty.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-46
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
¦ PCB congeners 123 and 157 co-eluted with other congeners (PCB-123 with PCB-
149; PCB-157 with PCB-173 and PCB-201). As a result, decision criteria were
developed (see Section 6.4) for co-eluted congeners to determine TEQ concentrations
used as distribution parameters in the Monte Carlo and probability bounds analyses.
These criteria were designed to explicitly incorporate this source of uncertainty in the
probabilistic analyses. Thus, this source of uncertainty has been incorporated in this
risk assessment.
¦ The greatest source of uncertainty of the effects assessment was associated with the
lack of toxicity studies involving the representative species. There were no toxicity
studies available for belted kingfishers or ospreys exposed to tPCBs or TEQ. As a
result, laboratory studies involving other species were used to estimate effects to
piscivorous birds. This extrapolation introduced uncertainty in the effects assessment
because of the variations in physiological and biochemical factors that can alter the
potential toxicity of a contaminant. The sensitivity of wild birds to an environmental
contaminant may differ from that of a laboratory or domestic species due to
behavioral and ecological parameters including stress factors (e.g., competition,
seasonal changes in temperature or food availability), disease, and exposure to other
contaminants. Inbred laboratory animal strains may also have an unusual sensitivity
or resistance to a tested substance. To address uncertainty in the effects assessment, a
threshold range was used in which effects to tolerant and sensitive species were
considered. It is assumed that the toxicity thresholds for the representative species lie
within these ranges.
¦ The belted kingfisher field study methods appeared to generally follow accepted
protocols, however EPA was not provided with an opportunity to review these
protocols prior to receiving the study. There were several shortcomings of the
approach used. For example, there was no reference site, no information was
provided regarding nest search intensity, the researchers were unable to determine
clutch size, and there were too few visits to the nests during the reproductive cycle.
These shortcomings limit the ability to draw rigorous conclusions.
¦ The statistics used in the belted kingfisher field study were not clearly stated. Student
t-tests were apparently performed even though there were no reference sites to
compare to. A power analysis of the results would have been useful. The sample
sizes were very small (i.e., n=6) for the statistics used (i.e., t-test and regression).
¦ The approach used to estimate dose in the belted kingfisher field study had a number
of shortcomings. The investigators assumed a foraging radius of 1,200 m and
attempted to identify prey samples within this radius of each burrow. The fish
samples had a "river mile" location associated with each sample, but this is an
imprecise measure that does not allow the location of the sample with sufficient
precision to assign specific fish to a specific 1,200-m foraging radius. Fish are also
mobile within the PSA, meaning that they receive their total exposure from many
areas of the river and that concentrations in fish do not vary substantially within the
PSA. As a result, the dose gradient achieved by this approach is likely too narrow to
detect a significant dose-response relationship.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-47
-------
1 ¦ The belted kingfisher field study results do not definitively support the conclusions of
2 low risk because the data are limited. There are several conclusions drawn by the
3 authors that are not strongly supported by the information presented in the report.
4 The conclusion that the kingfisher population is consistent with the quality of habitat
5 present is speculative. Survival to 26 days and densities were compared with the
6 results from only one study (Brooks and Davis 1987). It is inappropriate to conclude
7 that the Housatonic River kingfishers fall within the range reported for other
8 kingfisher populations when only one study is referenced. Although the GE study
9 provides no evidence of impaired reproduction or population density attributable to
10 PCBs, it fails to acknowledge the limitations associated with the use of only one
11 metric to evaluate reproduction.
12 8.5.5 Extrapolation to Other Species
13 Belted kingfisher and osprey are the only piscivorous birds common to the area. Other, less
14 common, piscivorous birds (e.g., pied-billed grebe, great blue heron), are addressed via
15 extrapolation in Section 11 and Appendix K.
16 8.5.6 Summary and Conclusions
17 The WOE analysis indicated that exposure of piscivorous birds, such as the belted kingfisher and
18 osprey, to tPCBs and TEQ in the PSA, could lead to adverse reproductive effects in some
19 species. The two lines of evidence used to support this conclusion were the field study of
20 kingfisher productivity and the comparison of modeled exposure with effects to piscivorous
21 birds.
22 For the assessment of risks to kingfishers, both lines of evidence were employed. The modeled
23 exposure and effects line of evidence indicated that kingfishers in the PSA are likely to receive a
24 tPCB dose greater than what the most tolerant species known can endure. For TEQ, the risk is
25 less clear because the threshold range for this COC is very wide and the exposure estimates for
26 kingfishers fell within this range. Thus, without effects data specific to kingfishers, it is difficult
27 to make definitive conclusions about the risks of TEQ to this species. The field study of
28 kingfisher productivity, however, indicated that these birds are able to reproduce in the PSA.
29 This line of evidence was given a higher weighting than the exposure and effects modeling,
30 despite concerns about the field study. Therefore, kingfishers are considered to be at low risk in
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-48
-------
1 the PSA as a result of exposure to tPCBs and TEQ. The conclusion of low risk to kingfishers is
2 uncertain because the two lines of evidence did not give concordant results.
3 For ospreys, only the modeled exposure and effects line of evidence was available to assess risk
4 to these birds. As with kingfishers, this line of evidence indicated that ospreys in the PSA are
5 likely to receive a tPCB dose that is greater than what the most tolerant species known can bear.
6 The risks due to exposure to TEQ are unclear, as the estimates for exposure also fell within
7 toxicity threshold range. Ospreys, however, lack a site-specific study that investigated the
8 effects of COCs in the PSA. The PSA contains suitable habitat for ospreys, with abundant prey,
9 raising the possibility that they are not resident in the area because of contaminants. Ospreys are
10 therefore considered to be at risk in the PSA as a result of exposure to tPCBs and TEQ.
11
12
13
14
15
16
17 8.6 REFERENCES
18 Alexander, G.R. 1977. Food and vertebrate predators on waters in north central lower Michigan.
19 Michigan Academician 10:187-195.
20 Bosveld, A.T.C. and M. Van den Berg. 1994. Effects of polychlorinated biphenyls, dibenzo-p-
21 dioxins, and dibenzofurans on fish-eating birds. Environmental Reviews 2:147-166.
22 Britton, W.M. and T.M. Huston. 1973. Influence of polychlorinated biphenyls in the laying hen.
23 Poultry Science 52:1620-1624.
24 Brooks, R.P. and W.J. Davis. 1987. Habitat selection by breeding belted kingfishers. American
25 Midland Naturalist 111 .63-10.
ERA Results for Piscivorous Birds
The WOE analysis suggests that ospreys may be at high risk from exposure to tPCBs and
intermediate risk from exposure to TEQ in the Housatonic River PSA. In the PSA, exposure of
piscivorous birds to tPCBs is greater than concentrations that caused adverse effects in the most
tolerant species studied. The conclusion of high risk to ospreys is uncertain because only one
line of evidence was available.
Belted kingfishers are considered to be at low risk as a result of exposure to tPCBs and TEQ in
the Housatonic River PSA. While modeled exposure and effects indicated high risk for tPCBs
and intermediate risk for TEQ, a field study of kingfisher productivity indicated that the birds were
reproducing in the PSA. The conclusion of low risk to kingfishers is uncertain because the two
lines of evidence did not give concordant results.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-49
-------
1 Brown, L. and D. Amadon. 1968. Eagles, Hawks, and Falcons of the World. McGraw-Hill, New
2 York, NY, USA.
3 Brunstrom, B. and J. Lund. 1988. Differences between chick and turkey embryos in sensitivity to
4 3,4'4,4'-tetrachlorobiphenyl and in concentration/affinity of the hepatic receptor for 2,3,7,8-
5 tetrachlorodibenzo-/>dioxin. Com. Biochem. Physiol. C 91(2):507-512.
6 CCME (Canadian Council of Ministers of the Environment). 1999. Canadian Tissue Residue
7 Guidelines for the Protection of Wildlife Consumers of Aquatic Biota: Polychlorinated
8 Biphenyls (PCBs). In Canadian Environmental Quality Guidelines, 1999. Canadian Council of
9 Ministers of the Environment, Winnipeg.
10 Chubbs, T.E. and P.G. Trimper. 1998. The diet of nesting osprey, Pandion haliaetus, in
11 Labrador. Canadian Field-Naturalist 112:502-505.
12 Clark, K.E. 1995. Osprey. In Living Resources of the Delaware Estuary. L.E. Dove and R.M.
13 Nyman (eds.). The Delaware Estuary Program, DE, USA. p. 395-400.
14 DeGraaf, R.M. and M. Yamasaki. 2001. New England Wildlife: Habitat, Natural History, and
15 Distribution. University Press of New England, Hanover, NH, USA.
16 Dunning, J.C. Jr. 1993. CRC Handbook of Avian Body Masses. CRC Press, Boca Raton, FL.
17 EPA (U.S. Environmental Protection Agency). 1993. Wildlife Exposure Factors Handbook.
18 EPA/600/R-93/187a. Office of Research and Development, Washington, DC.
19 Fernie, K.J., J.E. Smits, G.R. Bortolotti, and D.M. Bird. 2001. Reproductive success of American
20 kestrels exposed to dietary polychlorinated biphenyls. Environmental Toxicology and Chemistry
21 20:776-781.
22 Haffner, G.D., M. Tomczak, and R. Lazar. 1994. Organic contaminant exposure in the Lake St.
23 Clair food web. Hydrobiologia 281:19-27.
24 Hamas, M.J. 1994. Belted Kingfisher (Ceryle alcyon). In: The Birds of North America, No. 84.
25 A. Poole and G. Gill, Editors. The Birds of North America, Inc., Philadelphia, PA.
26 Heath, R.G., J.W. Spann, E.F. Hill, and J.F. Kreitzer. 1972. Comparative Dietary Toxicities of
27 Pesticides to Birds. U.S. Fish and Wildlife Service Special Scientific Report on Wildlife 152.
28 57 pp.
29 Henning, M.H. 2002. Productivity and Density of Belted Kingfishers on the Housatonic River,
30 Berkshire County, Massachusetts. Prepared for: General Electric Company, Pittsfield,
31 Massachusetts.
32 Hoffman, D.J., M.J. Melancon, P.N. Klein, C.P. Rice, J.D. Eisemann, R.K. Hines, J.W. Spann,
33 and G.W. Pendleton. 1996. Developmental toxicity of PCB-126 (3,3,4,4,5-pentachlorobiphenyl)
34 in nestling American kestrels (Falco sparverius). Fundamental and Applied Toxicology
35 34:188-200.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-50
-------
1 Hudson, R., R. Tucker, and M. Haegele. 1984. Handbook of Toxicity of Pesticides to Wildlife.
2 2nd Edition. U.S. Fish and Wildlife Service Resource Publication 153. Washington, DC.
3 Lillie, R.J., H.C. Cecil, J. Bitman, and G.F. Fries. 1974. Differences in response of caged white
4 leghorn layers to various polychlorinated biphenyls (PCBs) in the diet. Poultry Science
5 53:726-732.
6 MDFW. 1979. Field Investigation Report: Great Blue Heron Rookery Inventory. Commonwealth
7 of Massachusetts Division of Fisheries and Wildlife. October 1.
8 MDFW. 1980. Field Investigation Report: Great Blue Heron Rookery Inventory, 1980.
9 Commonwealth of Massachusetts Division of Fisheries and Wildlife. June 30.
10 MDFW. 1981. Field Investigation Report: Great Blue Heron Rookery Inventory, 1981.
11 Commonwealth of Massachusetts Division of Fisheries and Wildlife. June 30.
12 MDFW. 1982. Field Investigation Report: Great Blue Heron Rookery Inventory, 1982.
13 Commonwealth of Massachusetts Division of Fisheries and Wildlife. July 16.
14 MDFW. 1983. Field Investigation Report: Great Blue Heron Rookery Inventory, 1983.
15 Commonwealth of Massachusetts Division of Fisheries and Wildlife. July 3.
16 MDFW. 1984. Field Investigation Report: Great Blue Heron Rookery Inventory Results.
17 Commonwealth of Massachusetts Division of Fisheries and Wildlife. October 16.
18 MDFW. 1985. Field Investigation Report: Great Blue Heron Rookery Inventory Results.
19 Commonwealth of Massachusetts Division of Fisheries and Wildlife. December 16.
20 MDFW. 1986a. Field Investigation Report: Great Blue Heron Rookery Inventory Results.
21 Commonwealth of Massachusetts Division of Fisheries and Wildlife. February 19.
22 MDFW. 1986b. Field Investigation Report: Great Blue Heron Rookery Inventory, 1986.
23 Commonwealth of Massachusetts Division of Fisheries and Wildlife. November 25.
24 MDFW. 1987. Field Investigation Report: Great Blue Heron Rookery Inventory, 1987.
25 Commonwealth of Massachusetts Division of Fisheries and Wildlife. October 9.
26 MDFW. 1989. Field Investigation Report: Great Blue Heron Rookery Inventory, 1989.
27 Commonwealth of Massachusetts Division of Fisheries and Wildlife.
28 MDFW. 1991. Memorandum: 1991 Great Blue Heronry Survey. Commonwealth of
29 Massachusetts Division of Fisheries & Wildlife. August 19.
30 MDFW. 1996. Memorandum: 1996 Great Blue Heronry Survey. Commonwealth of
31 Massachusetts Division of Fisheries & Wildlife. November 1.
32 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S. Petron,
33 B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-of-
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-51
-------
1 Evidence Workgroup: A weight-of-evidence approach for evaluating ecological risks. Human
2 Ecological Risk Assessment 2:277-304.
3 Moore, D.R.J., B.E. Sample, G.W. Suter, B.R. Parkhurst and R.S. Teed. 1999. A probabilistic
4 risk assessment of the effects of methylmercury and PCBs on mink and kingfishers along East
5 Fork Poplar Creek, Oak Ridge, Tennessee, USA. Environ. Toxicol. Chem. 18:2941-2953.
6 Nagy, K.A. 1987. Free metabolic rate and food requirement scaling in mammals and birds.
7 Ecological Monographs 57:111-128.
8 Nagy, K.A., I. A. Girard, and T.K. Brown. 1999. Energetics of free-ranging mammals, reptiles,
9 and birds. Annual Review of Nutrition 19:247-277.
10 Newsted J.L., J.P. Giesy, G.T. Ankley, D.E. Tillitt, R.A. Crawford, J.W. Gooch, P.D. Jones, and
11 M.S. Denison. 1995. Development of toxic equivalency factors for PCB congeners and the
12 assessment of TCDD and PCB mixtures in rainbow trout. Environmental Toxicology and
13 Chemistry 14:861-871.
14 Newton, I. 1979. Population Ecology of Raptors. Buteo Books, Vermillion, SD. 399 pp.
15 Nosek, J.A., S.R. Craven, J.R. Sullivan, S.S. Hurley, and R.E. Peterson. 1992. Toxicity and
16 reproductive effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin in ring-necked pheasant. Journal of
17 Toxicology and Environmental Health 35:187-198.
18 Poole, A.F. 1983. Poole, A. F. (1983) Courtship feeding, clutch size, and egg size in ospreys: A
19 preliminary report. In: Bird, D. M.; Seymour, N. R.; Gerrard, J. M., eds. Biology and
20 management of bald eagles and ospreys. St. Anne de Bellvue, Quebec: Harpell Press; pp.
21 243-256.
22 Poole, A.F. 1985. Courtship, feeding, and osprey reproduction. The Auk 102:479-492.
23 Prestt I., D.J. Jefferies, and N.W. Moore. 1970. Polychlorinated biphenyls in wild birds in Britain
24 and their avian toxicity. Environmental Pollution 1:3-26.
25 Rattner, B.A., N.H. Golden, J.L. Pearson, J.B. Cohen, L.J. Garrett, M.A. Ottinger, and R.M.
26 Erwin. 2001. Biological and Ecotoxicological Characteristics of Terrestrial Vertebrate Species
27 Residing in Estuaries, http://www.pwrc.usgs.gov/bioeco/default.htm.
28 Safe, S. 1994. Polychlorinated biphenyls (PCBs): Environmental impact, biochemical and toxic
29 responses, and implications for risk assessment. Critical Reviews in Toxicology 24:87-149.
30 Salyer, J.C. and K.F. Lagler. 1946. The eastern belted kingfisher, Megaceryle alcyon alcyon
31 (Linneaus), in relation to fish management. Transactions of the American Fisheries Society
32 76:97-117.
33 Scott, M.L. 1977. Effects of PCBs, DDT, and mercury compounds in chickens and Japanese
34 quail. Federation Proceedings 36:1888-1893.
MK0110:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-52
-------
1 Senthilkumar, K., N. Iseki, S. Hayama, J. Nakanishi, and S. Masunaga. 2001. Polychlorinated
2 dibenzo-p-dioxins, dibenzofurans, and dioxin-like polychlorinated biphenyls in livers of birds
3 from Japan. Arch. Environ. Contam. Toxicol. 42:244-255.
4 Vana-Miller, S. 1987. Habitat Suitability Index Models: Osprey. U.S. Fish and Wildlife Service
5 Biological Report 82 (10.154). 46 pp.
6 Van den Berg, M., L. Birnbaum, A.T.C. Bosveld., B. Brunstrom, P. Cook, M. Feeley, J.P. Giesy,
7 A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C. Larsen, F.X. Rolaf van Leeuwen,
8 A.K.D. Liem, C. Nolt, R.E. Peterson, L. Poellinger, S. Safe, D. Schrenk, D. Tillitt, M. Tysklind,
9 M. Younes, F. Waern, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs) for PCBs,
10 PCDDs, PCDFs for humans and wildlife. Environmental Health Perspectives 106(12):775-792.
11 Woodlot Alternatives, Inc. 2002. Draft Report, Oversight of GE Field Studies for Housatonic
12 River Primary Study Area. Prepared for Weston Solutions, Contract No. DACW33-00-D-
13 0006/004, DCN: GE-082902-ABEG.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_8.DOC
8-53
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
9. ASSESSMENT ENDPOINT - SURVIVAL, GROWTH, AND
REPRODUCTION OF PISCIVOROUS MAMMALS
Highlights
Conceptual Model
The assessment endpoint is the survival, growth, and reproduction of piscivorous
mammals in the Housatonic River PSA. Piscivorous mammals (mink and river otter),
selected as representative species for the ERA, are exposed to tPCBs and TEQ via diet
and trophic transfer.
Exposure
Exposure of the representative species to COCs (tPCBs and TEQ) was determined from
concentrations found in prey items and an estimation of the daily intake of COCs from
consumption of prey.
Effects
Data on toxicity of tPCBs to mink were used to derive a dose-response relationship. The
existing data on toxicity of TEQ to mink were insufficient to derive a dose-response
relationship; therefore, upper and lower toxicity thresholds were derived. No tPCB or
TEQ toxicity data were available for otter. River otter were assumed to have a similar
sensitivity to tPCBs and TEQ. A site-specific feeding study was conducted to evaluate
adverse effects to mink from Housatonic River COCs.
Risk
Mink and river otter are at high risk as a result of exposure to tPCBs and TEQ in the
Housatonic PSA. The risk remains high even for those individuals who forage only a
fraction of their time in the PSA.
9.1 INTRODUCTION
The purpose of this section is to characterize and quantify the current and potential risks posed to
piscivorous mammals exposed to contaminants of potential concern (COPCs) in the Housatonic
River and floodplain, focusing on total PCBs (tPCBs) and other COPCs originating from the
General Electric Company (GE) facility in Pittsfield, MA. The river is located in western
Massachusetts and Connecticut, discharging to Long Island Sound, with the GE facility located
near the headwaters. The Primary Study Area (PSA) includes the river and 10-year floodplain
from the confluence of the East and West Branches of the Housatonic River downstream of the
GE facility, to Woods Pond Dam (Figure 1.1-2).
A Pre-ERA was conducted to narrow the scope of the ecological risk assessment (ERA) by
identifying contaminants, other than tPCBs, that pose potential risks to aquatic biota and wildlife
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC q -i 7/10/2003
-------
1 in the PSA (Appendix B). A three-tiered deterministic approach was used to screen COPCs.
2 The deterministic assessments compared conservative estimates of potential exposure with
3 conservative adverse effects benchmarks to identify contaminants that are of potential concern to
4 piscivorous mammals in the Housatonic River. A hazard quotient (total daily intake/effect
5 benchmark) greater than one resulted in the COPC being screened through to the next Tier
6 assessment and to the probabilistic ecological risk assessment, if necessary.
7 Subsequent to the Pre-ERA, several other COPCs (primarily organochlorine pesticides) were
8 screened out because their actual concentrations in the PSA were likely much lower than the
9 measured values due to laboratory interference problems (see Section 2.4). These COPCs were
10 evaluated further for each assessment endpoint, and the contaminants of concern (COCs) that
11 were retained for the probabilistic risk assessment for piscivorous mammals were tPCBs and
12 2,3,7,8-TCDD toxic equivalence (TEQ). Total PCBs detected in Housatonic River samples
13 closely resemble the commercial PCB mixtures Aroclor 1260 and Aroclor 1254, which are
14 similar in congener makeup. TEQ is calculated from coplanar PCB and dioxin and furan
15 congeners using the toxic equivalency factor (TEF) approach developed by Van den Berg et al.
16 (1998)(see Section 6.4).
17 A step-wise approach was used to assess the risks of tPCBs and TEQ to piscivorous mammals in
18 the Housatonic River watershed. The four main steps in this process include:
19 1. Derivation of a conceptual model (Figure 9.1-1).
20 2. Assessment of exposure of piscivorous mammals to COCs (Figure 9.1-2).
21 3. Assessment of the effects of COCs on piscivorous mammals (Figure 9.1-3).
22 4. Characterization of risks to the piscivorous mammalian community (Figure 9.1-4).
23
24 This section is organized as follows:
25
26
27
¦ Section 9.2 (Conceptual Model)—Describes the conceptual model for piscivorous
mammals, including selection of representative species and establishment of
measurement and assessment endpoints.
28
29
30
¦ Section 9.3 (Exposure Assessment)—Describes the exposure model, input variables,
and techniques to propagate uncertainty. Also presented in this section are the
exposure modeling results for mink and river otter.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-2
7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
¦ Section 9.4 (Effects Assessment)—Describes the effects to mammals exposed to
tPCBs and TEQ and derives the effects metrics.
¦ Section 9.5 (Risk Characterization)—Integrates the exposure and effects
assessments to quantify risk to piscivorous mammals in the PSA for each line of
evidence. This section contains brief descriptions of field surveys, feeding study, and
modeled exposure and effects measurement endpoints. The feeding and field studies
were used in the risk characterization only and were not used to develop effect levels
for comparison to modeled results. The risk information from three lines of evidence
is combined in the weight-of-evidence analysis. This section also describes the
sources of uncertainty in the ERA for piscivorous mammals, followed by the
conclusions regarding risks of tPCBs and TEQ to piscivorous mammals in the
Housatonic River PSA.
The detailed ecological risk assessment for piscivorous mammals is provided in
Appendix I.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-3
-------
-------
EXPOSURE
Figure 9.1-2 Overview of Approach Used to Assess Modeled Exposure of
Piscivorous Mammals to Contaminants of Concern (COCs) in the Housatonic
River PSA
MK0110:\20123001.096\ERA_PB\ERA_PB_9.DOC q 7/10/2003
-------
EFFECTS
Figure 9.1-3 Overview of Approach Used to Assess the Modeled Effects of
Contaminants of Concern (COCs) to Piscivorous Mammals in the Housatonic
River PSA
MK0110:\20123001.096\ERA_PB\ERA_PB_9.DOC q r 7/10/2003
-------
RISK CHARACTERIZATION
Figure 9.1-4 Overview of Approach Used to Characterize the Risks of
Contaminants of Concern (COCs) to Piscivorous Mammals in the Housatonic
River PSA
MK0110:\20123001.096\ERA_PB\ERA_PB_9.DOC q q 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
9.2 CONCEPTUAL MODEL
The conceptual model presented in Figure 9.1-1 illustrates the exposure pathways for piscivorous
mammals exposed to tPCBs and TEQ in the PSA. Total PCBs and TEQ are persistent and
hydrophobic and lipophilic. Therefore, they are bioaccumulated by aquatic and terrestrial biota
through the consumption of contaminated prey as part of the food chain (Haffner et al. 1994;
Senthilkumar et al. 2001). Fish, small mammals, crayfish, waterfowl, and amphibians are the
major dietary items for piscivorous mammals. Piscivorous mammals that reside, or partially
reside, within the study area are exposed to tPCBs and TEQ principally through diet and trophic
transfer. Other routes of exposure, considered to be less important to overall exposure, include
inhalation, water consumption, and sediment ingestion (Moore et al. 1999).
The problem formulation (see Section 2) identified mink (Mustela vison) and river otter (Lutra
canadensis) (Figures 9.2-1 and 9.2-2) as the representative species for piscivorous mammals
exposed to tPCBs and TEQ from consumption of contaminated prey. Life history profiles for
mink and river otter are summarized in the following text boxes. Additional life history
information on these species is presented in Sections 1.2.1.5 and 1.2.1.6, respectively.
The assessment endpoint that is the subject of this section is the survival, growth, and
reproduction of piscivorous mammals in the Housatonic River PSA. The measurement
endpoints used to evaluate the assessment endpoint included: (1) determining the extent to
which the concentrations of tPCBs and TEQ ingested in the diet impact the survival,
reproduction, or growth of piscivorous mammals by comparisons to doses reported in the
literature to cause adverse effects; (2) determining, by conducting quantitative field surveys, the
abundance of piscivorous mammals in the Housatonic River relative to appropriate
uncontaminated reference areas within the watershed; and (3) determining, by conducting a
feeding study using fish collected from the PSA, whether a diet of site-specific fish has an
adverse effect on the survival and reproduction of farm-raised mink.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-8
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Figure 9.2-1 Mink (Mustela vison)
Life History of Mink
Mink are small, fur-bearing animals with characteristic elongated bodies, short legs,
and long tails. Mink are one of the most widespread mammalian carnivores, with a
range spanning much of the continental USA and Canada.
Habitat - Require access to open water such as streams, tidal flats, marshes,
shallow rivers, lakes, and swamps. Also suitable cover in the form of overhanging
vegetation, rock crevices, exposed roots, log jams, and undercut banks.
Home Range - Adult males occupy home ranges exclusive of other adult males, and
may include the home ranges of one or more females. Males range from 309 to 776
ha, and females range from 7.8 to 20.4 ha. Riverine home ranges are linear
(between 1.0 and 6.0 km of shoreline); those in marsh habitats tend to be more
circular.
Dietary Habits - Primary food items include fish, small mammals, benthic
invertebrates, birds, and amphibians. Opportunistic; diet varies depending upon the
availability of prey items. Mean percentage of prey items in diet: fish, 23%;
mammals, 15%; birds, 11.0%; invertebrates, 36%; and amphibians and reptiles,
15.0%.
MK0110:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-9
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Figure 9.2-2 River Otter (Lutra canadensis)
Life History of River Otter
River otter are long-bodied, short-legged, semi-aquatic mustelids that occur
throughout most of Canada and the continental United States. Male otter in the
eastern United States are quite large and range in weight from 8 to 11 kg. Females
range from 7.5 to 8 kg.
Habitat - Remain close to aquatic habitats such as lakes, marshes, streams,
seashores, rivers, creeks, and bayous. In New England, preferentially select riverine
and lacustrine systems. Have numerous denning and nesting sites within home
range, used over the course of the year. Denning and resting sites may be located in
log jams, riparian vegetation, snow or ice cavities, riprap, talus rock, boulders, brush
and log piles, undercut banks, and dens constructed by other animals.
Home Range - Average size of the home range for adult otter is about 30 km of
shoreline. Lactating females have the smallest home ranges. Other than family
groups, are typically solitary. Will form temporary associations that may consist of
related or unrelated individuals. Home ranges shown to overlap extensively, with
some otter sharing essentially the same home range.
Dietary Habits - Diet somewhat variable; primarily consists of aquatic animals,
particularly fish; other prey includes crayfish, amphibians, turtles, birds, small
mammals, and insects. Prefer to forage in shallow water and eat primarily slow-
moving, shallow-dwelling fish, such as chubs, suckers, catfish, daces, darters, and
schooling fish such as bluegill and other sunfish.
MK0110:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-10
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
9.3 EXPOSURE ASSESSMENT
This exposure assessment evaluates exposure of piscivorous mammals to tPCBs and 2,3,7,8-
TCDD toxic equivalence (TEQ) in Reaches 5 (confluence to Woods Pond) and 6 (Woods Pond),
together referred to as the Primary Study Area (PSA) of the Housatonic River. Exposure
assessments were also conducted for two reference areas for comparative purposes. One of the
reference areas is located upstream of the GE facility on the East Branch of the Housatonic River
in Dalton, MA (herein referred to as the "upstream reference area"). The other reference area is
Threemile Pond located in Sheffield, MA, which is in the Housatonic River drainage, but at a
higher elevation, draining to the river. The representative species for piscivorous mammals are
mink and river otter. These mammals occur in the Housatonic River watershed and feed on prey
exposed directly to tPCBs and TEQ and through trophic transfer. The ingestion of contaminated
prey is the major exposure pathway for piscivorous mammals exposed to tPCBs and TEQ.
Total PCBs and TEQ tend to bioaccumulate in the food chain because:
¦ Total PCBs and TEQ are persistent, hydrophobic, and lipophilic substances.
¦ When released to aquatic systems, the majority of these compounds form associations
with dissolved and/or particulate matter in the water column and settle to the
sediment bed; biodegradation is considered to be a relatively minor fate process in
water (NRCC 1981; Howard et al. 1991).
¦ Aquatic sediment provides a sink for these compounds and may represent long-term
sources to the aquatic food web (Kuehl et al. 1987; Muir 1988; Corbet et al. 1983;
Tsushimoto et al. 1982). Both of these COCs are bioaccumulated by aquatic and
terrestrial biota directly through the consumption of contaminated prey as part of the
food chain (Haffner et al. 1994; Senthilkumar et al. 2001; Borga et al. 2001).
In summary, piscivorous mammals that reside, or partially reside, within the PSA are exposed to
tPCBs and TEQ principally through diet.
The exposure analysis for mink was carried out separately for Reach 5 and Reach 6 of the PSA
because the foraging range of mink approximates the lengths of those river sections. However,
the foraging range of river otter is larger; therefore, the exposure analysis for river otter was
conducted with Reaches 5 and 6 combined.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-11
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
This section begins with a description of the exposure model used for the representative species.
Subsequent sections describe the inputs used in the exposure analyses for each representative
species. The section concludes with a presentation of the results of the exposure analyses.
9.3.1 Exposure Model
Exposure of the representative species, mink and river otter, to tPCBs and TEQ was estimated
using a total daily intake model adapted from the Wildlife Exposure Factors Handbook (EPA
1993) and related publications. The model used in the exposure analysis was:
TDI = FT -FIR ±CrP, (Eq- 1)
Z=1
where
TDI = Total daily intake (mg/kg bw/d tPCBs, ng/kg bw/d TEQ).
FIR = Normalized food intake rate (kg/kg bw/d).
FT = Foraging time in PSA (unitless).
Ci = Concentration in z'th food item (mg/kg tPCBs, ng/kg TEQ).
Pi = Proportion of the z'th food item in the diet (unitless).
The models consider the food intake rates of the representative species (FIR), the concentrations
of COCs in each food item (Ci), and the proportion of the diet accounted for by that food item
(Pi). For those input variables that are uncertain, variable, or both, distributions are used rather
than point estimates. Monte Carlo and probability bounds analyses are the methods used to
propagate uncertainties about input variables in the exposure model for each COC. A
description of these techniques and methods used to parameterize input variables is presented in
Section 6.5 and Appendix C. The results of the Monte Carlo analysis are used to estimate the
probability of exposure exceeding an effects threshold or doses that cause adverse effects of
differing magnitudes. The probability bounds analysis is conducted to determine how
uncertainty regarding the distributions of the input variables influences the estimated exposure
distribution. The results of these analyses are discussed in detail in Appendix I.
Two issues arose when calculating a TEQ concentration in prey:
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-12
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
¦ Congener concentrations may be below the method detection limit (DL) (i.e., non-
detects).
¦ Some congeners may not be resolved due to co-elution during analysis.
An approach was developed to address these issues. This approach is presented in Section 6.4
and Appendix C.2. Briefly, congeners detected at or below the DL were included in the TEQ
calculations by investigating three options:
¦ Setting the concentration for the congener equal to zero (0).
¦ Setting it to half the DL.
¦ Setting it equal to the DL.
A comparison of the results of this bounding analysis provides a description of the uncertainty
surrounding the TEQ value due to concentrations of one or more congeners being below the
detection limit.
To resolve the co-elution issue, the concentrations of congeners that co-eluted with other
congeners were assumed equal to the total concentration (likely an overestimate of TEQ
concentration) or zero (likely an underestimate of TEQ concentration). The decision criteria in
Section 6.4 were followed to deal with the uncertainty arising from co-elution or non-detection
of congeners when estimating exposure point concentrations (EPCs) for use in the exposure
analyses.
Input distributions to the exposure analyses were generally assigned as follows:
¦ Lognormal distributions were assigned to variables that were right skewed with a
lower bound of zero and no upper bound (e.g., amount of COC in fish).
¦ Beta distributions for variables bounded by zero and one (e.g., proportion of a prey
item in the diet).
¦ Normal distributions for variables that were symmetric and not bounded by one (e.g.,
body weight).
¦ Point estimates for minor variables or variables with low coefficients of variation.
In certain situations (e.g., poor fit of data), other distributions were fit to the data or other
approaches were used. To quantify uncertainty, two approaches were used as described in
Section 6.5.2 and Appendix C. The distributions used in the exposure analyses for mink and
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC 9 13
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
river otter are shown in Figures 9.3-1 and 9.3-2. A brief description of these variables is
provided below.
9.3.1.1 Body Weight (BW)
Average body weights (wet weight of wild animals) of female mink range from 550 g (Mitchell
1961) to 970 g (Hornshaw et al. 1983) and males range from 630 to 1,000 g (Whitaker and
Hamilton 1998). For the Monte Carlo analysis, the mean weight of females was estimated to be
685 g with a standard deviation of 122. Body weights were assumed to be distributed normally.
There is low uncertainty associated with this variable. The uncertainty in this variable is due to
variability, rather than lack of knowledge or data (i.e., the variable is easily measured and many
studies have been conducted that measured this variable). Accordingly, the same distribution
was used in the probability bounds analysis.
Body weight is not used in the model directly, but is a required variable in allometric models
(e.g., Nagy 1987) to estimate food intake or free metabolic rates. Whitaker and Hamilton (1998)
reported that body weights of river otter ranged from 8 to 11 kg (average of 9.2 kg) for males and
from 7.5 to 8.0 kg (average of 7.9 kg) for females in eastern United States populations. In the
Monte Carlo analysis, body weight was assumed to be normally distributed with a mean of 8,630
g and a standard deviation of 1,600 g. The same distribution was used in the probability bounds
analysis for this input variable. The uncertainty in this variable is small and is likely due to
variability, rather than lack of knowledge or data gaps.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-14
-------
FMRSIope Term(a)
FMR Power Term(b)
0.40 1.56 2.78 4.00 5.22 6.44
0.706 0.761 0.816 0.870 0.925 0.980
Body Weight
Proportion of Fish in Diet
0.628 0.748
Weight (kg)
0.222 0.325
Proporti on
0.429 0.532
Proportion of Inve rtebrates i n Diet
0.062 0.148
0.235 0.321
Proportion
0.407 0.494
Proportion of Birds in Diet
0.124 0.183
Prop ortion
0.243 0.302
Proportion of Mammals in Diet
0.008 0.054
0.099 0.145
Prop ortion
0.191 0.236
Proportion of Amphibians in Diet
0.006 0.068
0.130 0.193
Proportion
0.255 0.317
Figure 9.3-1 Input Distributions Used in Exposure Modeling for Mink
MK0110:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-15
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
FMR Slope Term (a)
FMR Powe rTerm(fc)
Body Weight
1
08
£ 0.6
1 0.4
o
0.2
1
_ 0.8
eS
>• 0.6
\q
« 0.4
o
0.2
Proportion of Invertebrates in Diet
Weight (kg)
0.164 0.186
Proportion
0
0.119
Figure 9.3-2 Input Distributions Used in Exposure Modeling for River Otter
9.3.1.2 Food Intake Rate (FIR)
The daily energy requirements of mink vary depending on environmental conditions and the
stage of the reproductive cycle. However, the long-term average daily consumption of dry
matter is approximately 0.040 kg/kg of body mass for males and 0.0530 kg/kg of body mass for
female captive mink (Bleavins and Aulerich 1981; Lariviere 1999). A 1.0-kg mink living in a
laboratory requires approximately 150 kilocalories (kcal) of digestible energy every day for
maintenance. A nursing female can require 3 times that amount at 3 weeks post-partum
(Lariviere 1999). However, a nursing female food intake rate was not considered in this
assessment because nursing is a short-term event relative to the extended time scale of this
assessment (1 year). The time scale of this exposure assessment was chosen to be approximately
1 year based on the extended reproductive cycle of mink (mating starts in early March) and the
duration of the mink feeding study, which evaluated the effects on young mink until they were 6
months old.
MK0110:\20123001.096\ERA_PB\ERA_PB_9.DOC q 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
For the purpose of the ERA for piscivorous mammals, FIR was estimated using an allometric
equation rather than using literature-reported values for captive mink. An allometric model-
derived FIR better approximates the increased energy demand of wild mink resulting from
higher activity levels incurred while foraging, defending and inspecting territory, and avoiding
predators (Lamprey 1964; Buechner and Golley 1967; Koplin et al. 1980).
Food intake rate (FIR) is derived using the following equation:
FMR (kJ/d) = a- BW (g)b (Eq. 2)
where FMR is the free-living metabolic rate and BW is the body weight. The slope (a) and
power (b) distributions were based on the error statistics from regression analysis of the data
reported in Nagy et al. (1999). For carnivorous mammals, the mean slope term log (a) had a
mean of 0.367 and a standard error of 0.223. The power term (b) had a reported mean of 0.850
and a standard error of 0.055(Nagy et al. 1999).
Food intake rate is derived from FMR using the following equation:
FMR
F1R=- (Eq. 3)
i=1
where AEt is the assimilation efficiency of z'th food item (unitless) and GE is the gross energy of
z'th food item (kcal/kg).
The gross energies of various wildlife food sources are summarized in the Wildlife Exposure
Factors Handbook (EPA 1993) and were as follows: fish and amphibians (assumed for
amphibians) 1.20 kcal/g (Thayer et al. 1973; EPA 1993), invertebrates 1.10 kcal/g (Jorgensen et
al. 1991; Minnich 1982; Thayer et al. 1973), and birds and mammals 1.8 kcal/g (EPA 1993).
These variables were treated as point estimates in Monte Carlo simulations because of their
relatively small coefficients of variation. Gross energy is easily measured and thus measurement
error is likely to be low.
Average assimilation efficiency for mammals consuming fish and amphibians is 0.91, for
invertebrates it is 0.87, and for birds and mammals it is 0.84 (EPA 1993; Grodzinski and Wunder
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-17
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
1975; Barrett and Stueck 1976). No data were available for assimilation efficiency of mammals
consuming amphibians, but it is likely to be similar to that for mammals consuming fish. These
variables were treated as point estimates in Monte Carlo simulations and probability bounds
analyses because of their relatively small coefficients of variation. As a result, these input
variables are not included in Figures 9.3-1 and 9.3-2.
9.3.1.3 Proportions of Dietary Items (Pi)
The primary food items in the mink diet include small mammals, fish, benthic invertebrates
(crayfish), birds (waterfowl), and amphibians (Alexander 1977; Burgess and Bider 1980; Cowan
and Reilly 1973; Gilbert and Nanckivell 1982; Hamilton 1959, 1940; Melquist et al. 1981;
Proulx et al. 1987) (Table 1.2-2). Combining the available data, an average of 23% (range of 0 to
64.7%) of the mink diet consists of fish. Mammals on average comprise 15% of the diet (range
of 0 to 25%). Reptiles and amphibians also constitute an average of 15% (range of 0 to 30%) of
the diet, and birds (i.e., waterfowl) 11% (range of 0 to 39%) of the diet. Invertebrates constitute
an average of 36% of the diet (range of 0 to 54%).
Melquist et al. (1981) found that fish taken by mink were mostly cyprinids between 7 and 12 cm
long. Similarly, Hamilton (1940) recorded that the average length of fish taken by mink ranged
from 7.6 to 10.2 cm. According to Alexander (1977), mink in rivers and streams in lower
Michigan and New York consume fish ranging from 15 to 18 cm. Based on this information,
fish used in the exposure analyses were limited to a minimum length of 7 cm and a maximum
length of 20 cm. Fish prey of river otter can range from 2 to 50 cm in length (Melquist and
Hornocker 1983). In some areas, fish captured were typically less than 15 cm (Hamilton 1961;
Lagler and Ostenson 1942; Alexander 1977). Greer (1956), however, indicated that fish
captured by otter ranged from 15 to 25 cm. Based on these observations, the exposure analysis
for otter included tissue samples for fish ranging in length from 2 to 50 cm.
The proportion of each prey type in the diet was assumed to follow a beta distribution in the
Monte Carlo analysis and was parameterized to approximate the above averages and ranges
(Table 1.2-3). The beta distribution is not an available option in RiskCalc, the software used for
conducting the probability bounds analyses. As an alternative, minimum, mean, and maximum
values were specified for each dietary item using the means and ranges described above. The
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC q -i q 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
minimum, mean, and maximum values were then included as a distribution-free statement in
RiskCalc. The results bound all possible distributions, given the minimum, mean, and maximum
values specified for the dietary items.
9.3.1.4 Concentrations of COCs in Prey
The median concentrations of tPCBs in mink prey from the PSA range from 2.45 mg/kg in
amphibians from Reach 5 to 29.9 mg/kg in fish from the same location. The 25th and 75th
percentiles are 1.13 and 5.37 mg/kg for amphibians and 24.6 and 39.2 mg/kg for fish from Reach
5. Median TEQ levels in mink prey range from 91.6 ng/kg in amphibians from Reach 5 to 858
ng/kg in birds from the same location. The 25th and 75th percentiles are 58.8 and 123 ng/kg for
amphibians and 532 and 1,596 ng/kg for birds from Reach 5. The distributions for
concentrations of tPCBs and TEQ in prey of mink are presented in Figures 9.3-3 and 9.3-4,
respectively. The distributions for concentrations of tPCBs and TEQ in prey of river otter are
presented in Figures 9.3-5 and 9.3-6, respectively. The input variables for concentrations of
COCs in prey of mink and river otter are shown in Tables 1.2-4,1.2-5,1.2-12, and 1.2-13.
£
c
o
c
CJ
fi
O
CJ
05
U
a.
100
10
to
*
~ Median
X Arithmetic Mean
75 Percentile
25th Percentile
0.1
Jf
Xr
&
&
J?
..?PV
rv*
&
&
^3
¦ cS>
&
4°
Prey item and location
Figure 9.3-3 Concentrations of tPCBs in Prey of Mink
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-19
-------
bJD
a
©
c
o
cj
c
o
cj
a
w
H
10000
1000
100
10
1
0.1
3?
•XT
_c>
&
-c>
_c>
,!p
v <&
&
«f
^ # #
c>
_c>
S
~ Median
X Arithmetic Mean
75 Percentile
25th Percentile
^ J?
V ^
r>
_c>
_c>
•2j> by>
^ #
1
2
Prey item and location
Note: Error bars indicate interquartile range.
Figure 9.3-4 Concentrations of TEQ in Prey of Mink
100
10
09
U
Ph
0.1
.a>
<
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
1000
M
£?) 100
A
=
01
(J
e
o
tj
a
W
H
10
0.1
&
Cf
Prey item and location
~ Median
* Arithmetic Mean
75 Percentile
25th Percentile
Figure 9.3-6 Concentrations of TEQ in Prey of River Otter
9.3.2 Results of Exposure Assessments
Exposure distributions for exposure of mink and river otter to tPCBs and TEQ in Reaches 5 and
6, and reference areas are presented in Figures 9.3-7 through 9.3-20.
Figure 9.3-7 depicts the cumulative distribution of tPCB intake rates for mink in Reach 5. The
Monte Carlo analysis indicated that exposure of mink to tPCBs could range from a minimum of
0.308 to a maximum of 82.5 mg/kg bw/d. The mean exposure was 5.29 mg/kg bw/d and the
median exposure was 3.97 mg/kg bw/d. Of the exposure estimates, 90% were between 1.15 and
13.6 mg/kg bw/d. The probability bounds estimated for mink foraging in Reach 5 are depicted in
Figure 9.3-7. The 10th percentile of the probability envelope formed by the lower and upper
bounds ranged between 0.0244 and 4.11 mg/kg bw/d. The 50th percentile ranged between 0.292
and 8.47 mg/kg bw/d, and the 90th percentile ranged between 1.72 and 22.5 mg/kg bw/d. In
comparison, the 10th percentile of the Monte Carlo output was 1.52, the 50th percentile was
3.97, and the 90th percentile was 10.4 mg/kg bw/d.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-21
-------
1
Reach 5
2
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 9.3-7 Exposure of Mink to tPCBs in Reach 5 of the Housatonic River
6
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-22
-------
Reach 6
1
2 LPB = Lower probability bound
3 UPB = Upper probability bound
4 Figure 9.3-8 Exposure of Mink to tPCBs in Reach 6 of the Housatonic River
5
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
-------
Upstream Reference Area
100
X 60
40
h
W
20
¦ Monte Carlo
-LPB
-UPB
LV
10 15 20 25 30
TDI (mg/kg bw/d)
35
40
45
50
2 LPB = Lower probability bound
3 UPB = Upper probability bound
4 Figure 9.3-9 Exposure of Mink to tPCBs in the Housatonic River Upstream
5 Reference Area
6
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-24
-------
Threemile Pond Reference Area
100
¦ Monte Carlo
-LPB
-UPB
X 60
40
20
10 15 20 25 30
TDI (mg/kg bw/d)
35
40
45
50
LPB = Lower probability bound
UPB = Upper probability bound
Figure 9.3-10 Exposure of Mink to tPCBs in the Threemile Pond Reference Area
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-25
-------
Reach 5
Monte Carlo
LPB
UPB
100 200 300 400 500 600
TDI (ng/kg bw/d)
700
800
900
1000
2
3
4
5
LPB = Lower probability bound
UPB = Upper probability bound
Figure 9.3-11
Exposure of Mink to 2,3,7,8-TCDD TEQ in Reach 5 of the
Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-26
-------
Reach 6
2 LPB = Lower probability bound
3 UPB = Upper probability bound
4 Figure 9.3-12 Exposure of Mink to 2,3,7,8-TCDD TEQ in Reach 6 of the
5 Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-27
-------
Upstream Reference Area
1
2 LPB = Lower probability bound
3 UPB = Upper probability bound
4 Figure 9.3-13 Exposure of Mink to 2,3,7,8-TCDD TEQ in the Housatonic River
5 Upstream Reference Area
6
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-28
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Threemile Pond Reference Area
LPB = Lower probability bound
UPB = Upper probability bound
Figure 9.3-14 Exposure of Mink to 2,3,7,8-TCDD TEQ in the Threemile Pond
Reference Area
Figure 9.3-15 depicts the cumulative distribution of tPCB intake rates for river otter in Reaches 5
and 6. The Monte Carlo analysis indicated that exposure of otter to tPCBs foraging in the PSA
100% of the time could range from a minimum of 0.251 to a maximum of 111 mg/kg bw/d. The
mean exposure was 8.42 mg/kg bw/d and the median exposure was 6.02 mg/kg bw/d (Table 1.2-
14). Of the exposure estimates, 90% were between 1.60 and 22.8 mg/kg bw/d. The probability
bounds estimated for river otter foraging in Reaches 5 and 6 are depicted in Figure 9.3-15. The
10th percentile of the probability envelope formed by the lower and upper bounds ranged
between 1.59 and 8.12 mg/kg bw/d. The 50th percentile ranged between 3.27 and 14.2 mg/kg
bw/d, and the 90th percentile ranged between 5.78 and 53.0 mg/kg bw/d. In comparison, the
10th percentile of the Monte Carlo output was 2.15, the 50th percentile was 6.03, and the 90th
percentile was 17.1 mg/kg bw/d.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC Q O Q 7/10/2003
-------
1
Reaches 5 and 6
3 LPB = Lower probability bound
4 UPB = Upper probability bound
5 Figure 9.3-15 Exposure of River Otter to tPCBs in Reaches 5 and 6 of the
6 Housatonic River
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-30
-------
Upstream Reference Area
TDI (mg/kg bw/d)
LPB = Lower probability bound
UPB = Upper probability bound
Figure 9.3-16 Exposure of River Otter to tPCBs in the Housatonic River
Upstream Reference Area
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-31
-------
Threemile Pond Reference Area
TDI (mg/kg bw/d)
LPB = Lower probability bound
UPB = Upper probability bound
Figure 9.3-17 Exposure of River Otter to tPCBs in the Threemile Pond Reference
Area
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-32
-------
Reaches 5 and 6
100
I «
C3
A
B
¦d
h
W
40
20
¦ Monte Carlo
-LPB
-UPB
20
40
60
100 120
TDI (ng/kg bw/d)
140
160
180
200
2 LPB = Lower probability bound
3 UPB = Upper probability bound
4 Figure 9.3-18 Exposure of River Otter to 2,3,7,8-TCDD TEQ in Reaches 5 and 6
5 of the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-33
-------
Upstream Reference Area
0 20 40 60 80 100 120 140 160 180 200
TDI (ng/kg bw/d)
1
2 LPB = Lower probability bound
3 UPB = Upper probability bound
4 Figure 9.3-19 Exposure of River Otter to 2,3,7,8-TCDD TEQ in the Upstream
5 Reference Area
6
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-34
-------
Threemile Pond Reference Area
100
¦ Monte Carlo
-LPB
-UPB
.Q
-O
2
a.
01
(J
S
03
¦a
60
m
w
20
20 40 60 80 100 120
TDI (ng/kg bw/d)
140
160
180
200
2 LPB = Lower probability bound
3 UPB = Upper probability bound
4 Figure 9.3-20 Exposure of River Otter to 2,3,7,8-TCDD TEQ in the Threemile Pond
5 Reference Area
6
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-35
-------
1 9.4 EFFECTS ASSESSMENT
2 The purpose of the effects assessment is to review the scientific literature and derive appropriate
3 effects metrics for effects of tPCBs and TEQ to piscivorous mammals. An effects metric can be
4 represented by a dose-response relationship or a daily dose for a COC that represents a threshold
5 beyond which toxic effects may appear in piscivorous mammals. The effects metrics are used, in
6 conjunction with the exposure assessment, to estimate risks to piscivorous mammals exposed to
7 tPCBs and TEQ in the Housatonic River PSA. This section focuses on effects that have an
8 influence on the maintenance of local populations (i.e., mortality, or impairment of reproduction
9 or growth). Studies involving multiple exposure treatments and where reported results were
10 evaluated statistically to identify significant differences from controls were preferred. This
11 section also presents the results of a study where farm-raised mink were exposed to a diet
12 containing fish collected from the PSA.
13 Studies that document effects of tPCBs and TEQ were available only for mink, not otter.
14 However, given the close similarities between mink and otter in their feeding preferences and
15 phylogeny, an assumption was made that toxicity data for mink can be used to approximate
16 toxicity to river otter.
17 9.4.1 Review of Toxicity from the Literature
18 Presented below is a brief review of the scientific literature on the effects of dietary tPCBs and
19 TEQ to piscivorous mammals. The discussion focuses on ecologically relevant effects endpoints
20 such as survival, growth, and reproduction. A summary of reproduction effects for tPCBs and
21 TEQ is presented in Figures 1.3-1 and 1.3-2 and Table 1.3-1.
22 9.4.1.1 Total PCBs
23 9.4.1.1.1 Mortality
24 In a study where the diet was prepared from cattle that had consumed feed contaminated with
25 Aroclor 1254 (Figure 1.3-1; Table 1.3-1; Platonow and Karstad 1973), a dose of 0.0896 mg/kg
26 bw/d consumed by female mink over 160 days of exposure caused 100% mortality in the
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-36
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
offspring. The treatment also caused 17% mortality in adult females, but not in males.
Hornshaw et al. (1983) fed female mink with contaminated carp containing a dose of 0.210
mg/kg bw/d of PCBs identified as Aroclor 1254. After 7 months of this feeding regime, the
mink were allowed to reproduce. None of the young were born alive. A slightly higher dose of
0.280 mg/kg bw/d caused 100% kit mortality 4 weeks after birth. Adult female mink
experienced 12% mortality after 10 months of continuous exposure to this treatment (Aulerich
and Ringer 1977).
Total mortality in adults was observed at a dose of 0.500 mg/kg bw/d (Platonow and Karstad
1973). Only 105 days of dietary exposure at this concentration were required to kill all the
adults. In another study, female mink were exposed to a dose of 0.700 mg/kg bw/d. Of these
individuals, 30% died after 9 months of exposure (Aulerich and Ringer 1977). Mortality
increased to 71% in response to a dose of 1.40 mg/kg bw/d.
Ranch-raised mink exposed to 0.140 mg/kg bw/d reported as Aroclor 1254 from field-collected
carp experienced lower survival in lactating offspring (Wren et al. 1987b). However, the carp
contained other contaminants that could have contributed to the toxic response. Dietary LC50
tests with mink performed by Hornshaw et al. (1986) using Aroclor 1254 revealed average LC50s
from 6.58 mg/kg bw/d to 8.12 mg/kg bw/d. One of the highest estimates of acute doses was
reported by Aulerich et al. (1973), who found a 48-hour LD50 of 140 mg/kg bw/d.
Dietary exposure of female mink to a dose of 0.004 mg/kg bw/d tPCBs (42 to 60 % chlorine) in
carp for 3 to 6 weeks resulted in 15% mortality in kits (Heaton et al. 1995). Mortality increased
to 69% at a dose of 0.1 mg/kg bw/d after 3 weeks of exposure and 71% after 6 weeks of
exposure. At the dose of 0.210 mg/kg bw/d, kit mortality was 71% after 3 weeks of exposure
and 89% after 6 weeks of exposure. Total kit mortality was observed at a dose of 0.360 mg/kg
bw/d, with death being observed in as little as 24 hours after receiving the dose. Jensen et al.
(1977) exposed female mink to a dose of 1.54 mg/kg bw/d of Aroclor 1254 for 66 days. After
the treatment, no live kits were born to exposed females. Ringer et al. (1972) exposed mink to a
diet spiked with 4.20 mg/kg bw/d PCBs (equal amounts of Aroclors 1242, 1248, and 1254). All
adult mink died prior to whelping.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-37
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
9.4.1.1.2 Reproduction
Farm-raised mink exposed to 0.140 mg/kg bw/d Aroclor 1254 experienced reduced survival of
lactating offspring. However, no declines in fertility, whelping, or fecundity were observed
(Wren et al. 1987b). Kihlstrom et al. (1992) exposed female mink to 1.64 mg of Aroclor 1254
per individual (1.28 mg/kg bw/d) in food for 105 days. The exposure caused all kits to be
stillborn. The dose also increased the number of interrupted pregnancies. Aulerich and Ringer
(1977) reported that exposure of mink to 0.280 mg/kg bw/d of Aroclor 1254 did not affect birth
rate, birth weight, or survival. However, a dose of 2.80 mg/kg bw/d caused reduced whelping
and growth rate of kits. At 0.7 mg/kg bw/d, no whelping was observed, although survival was
unaffected (Bleavins et al. 1980).
Decreased mink fecundity has been observed following exposure to 0.08 mg/kg bw/d (0.7 mg/kg
diet) (Brunstrom et al. 1991). In another study, Aulerich et al. (1985) exposed mink to dietary
concentrations of Aroclor 1254 over extended exposure periods (several weeks). A
concentration of 2.5 mg/kg was associated with reduced fecundity. Only one female whelped
and the kit that was born died after birth. This dietary exposure is equivalent to a dose of 0.288
mg/kg bw/d given the food intake rate of 115 g/day. Male and female mink fed PCB-
contaminated diets (Saginaw Bay carp) had decreased breeding performance. Kit body weight
and survival were reduced at birth following exposure to 0.140 mg/kg bw/d of tPCBs in the diet
(Restum et al. 1998).
9.4.1.2 2,3,7,8-TCDD Toxic Equivalence (TEQ)
Effects of TEQ
Types of effects to mammals from exposure to TEQ include:
¦ Hormone induction
¦ Decreases in body and organ weight
¦ Reduced fertility
¦ Reduced litter size
¦ Reduced survival at birth or weaning
¦ Mortality
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-38
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Mode of Action of TEQ Congeners
Congeners that comprise the TEQ group have the ability to bind with the aryl
hydrocarbon (Ah) receptor and elicit similar toxic responses. The most toxic
congeners tend to be those that have a planar shape and are chlorinated in the
2,3,7, and 8 positions fordioxins and furans, and in the meta and para positions for
PCBs.
This structural configuration best fits the receptor and leads to a common mechanism
of action in many animal species involving binding to the Ah receptor and elicitation
of an Ah-receptor-mediated biochemical and toxic response. The toxic response of
this group of chemicals is, therefore, related to the three-dimensional structure of the
substance, including the degree of chlorination and positions of the chlorine on the
aromatic frame.
Planar chlorinated hydrocarbons are found in the environment as a mixture of
congeners. The congeners can have different toxic potencies. To address this issue
and effectively estimate the relative toxicity of these mixtures, various systems have
been created involving the development and use of toxic equivalency factors (TEFs)
to derive toxic equivalence (TEQ). The approach used for this assessment is
described in Section 6.4.
9.4.1.2.1 Mortality
Mature female mink fed diets with 0.600, 16.0, 53.0, 180, and 1,400 ng/kg of 2,3,7,8-TCDD
(equivalent to a dose of 0.0840, 2.24, 7.42, 25.2, and 196 ng/kg bw/d) for a maximum of 132
days exhibited 17% mortality, as well as lethargy and bloody stools at the highest dose
concentration (Hochstein et al. 2001). Final body weights were inversely related to dietary
TCDD concentration and there was a dose-dependent drop in kit weight from birth to week three
of exposure. At the highest dose concentration of 196 ng/kg bw/d various physiological
functions were depressed. Hochstein et al. (1998) exposed female mink to 1, 10, 100, 1,000,
10,000, and 100,000 ng/kg of TCDD in the diet (daily dose equivalent of 0.14, 1.4, 14, 140,
1,400, and 14,000 ng/kg bw/d) for 125 days. A dose-dependent wasting syndrome (decrease in
body weight) was observed. Mortality reached 12.5%, 62.5%, and 100% after 28 days of
exposure to 140, 1,400, and 14,000 ng/kg bw/d, respectively. After 125 days of exposure,
mortality reached 100% in the 1,400 and 14,000 ng/kg exposure groups. Newborn mink given
doses (intraperitoneal injection) of 100 and 1,000 ng TCDD/kg bw experienced 100% mortality
at the higher dose after 12 days. The lower dose caused depressed body weight and 62%
mortality (Aulerich et al. 1988). Adult mink administered a single oral dose of 2,500 ng/kg bw
TCDD had significantly reduced body weights after 3 weeks (Hochstein et al. 1988). At 0.250
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC Q ^ Q 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
ng/kg bw/d, Heaton et al. (1995) observed 15% mortality to mink kits after exposure for 3
weeks. At a dose of 3.6 ng/kg bw/d, 69% mortality in kits was reported. Mortality increased to
100%) at a dose of 10.7 ng/kg bw/d.
9.4.1.2.2 Reproduction
Adult mink exposed to 0.6, 16, 53, 180, and 1,400 ng/kg TCDD in a diet of field-collected fish
(daily dose equivalent of 0.084, 2.24, 7.42, 25.2, and 196 ng/kg bw/d) for up to 132 days
produced offspring that had reduced survival from birth to week three of exposure (Hochstein et
al. 2001). There is some evidence that TCDD interferes with ovulation. Ushinohama et al.
(2001) administered a dose of 32,000 ng/kg by gavage to female rats. The treatment led to
reduced body weight gains as well as to reduced ovarian weights. Infertility and fetal loss have
been observed at a dose of 0.01 mg/kg/day (10 ng/kg bw/d TEQ) administered to rats (Murray et
al. 1979). A lowest observed adverse effect level (LOAEL) of 1 ng/kg/day was estimated by
Nisbet and Paxton (1982) using Murray et al. (1979) data. Ovulation was delayed and fewer ova
produced. A dose of 0.350 mg/kg bw/d of 2,3,6,2',3',6'-HxCB caused reduced litter size
(Aulerich et al. 1985). Female mink exposed to 0.00140 mg/kg bw/d of the isomer 3,4,5,3',4',5-
HxCB for 120 days did not experience adverse effects on reproduction (Aulerich et al. 1987). A
dose of 0.0140 mg/kg bw/d was associated with a total absence of whelping.
9.4.2 Mink Feeding Study
It was hypothesized, when developing the conceptual model for the ERA, that contaminants in
the prey of piscivorous mammals foraging in the PSA may have caused adverse effects on the
survival, reproduction, and/or growth of exposed individuals, based on a lack of observations of
mink or otter (or sign) during EPA field investigations. To test this hypothesis, a long-term
feeding study was performed by researchers at Michigan State University (MSU) (Bursian et al.
2002), the results of which are described below.
9.4.2.1 Methodology
In this study, fish were collected from the PSA, frozen, and sent to MSU. These fish were mixed
with ocean herring in varying proportions to derive a control diet formulated to meet the
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC q a rx 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
nutritional requirements of farm-raised mink (all diets were 30% fish, 70% formulated mink diet)
and five treatment diets containing target concentrations ranging from 0.25 to 4 mg/kg tPCBs.
These concentrations were established during the study design to span the range of known effects
thresholds from previous studies of the effects of PCBs on mink. The diets were fed to captive
adult female mink for approximately 160 days. The exposure period for adult females began
approximately 2 months prior to mating, and continued through mating and whelping of the kits.
Some kits were exposed for an additional 6 months following whelping. A variety of endpoints
were measured during the study including feed consumption rate, mating success, gestation
length, number of kits born, adult and kit survival, body weights, organ weights, and tissue
histology. Biochemical parameters and the histopathology of the jaws of mink kits were also
measured. The latter endpoints are discussed separately in Sections 9.4.2.3 and 9.4.2.4.
9.4.2.2 Results and Discussion
The presence of COCs in the diet did not have a significant effect on food intake rate of adult
female mink. Consumption of diets containing COCs derived from Housatonic River fish had no
significant effect on breeding success (number of females bred/total number of females) or
whelping success (number of females whelping/number of females bred) of female mink.
Gestation length was not significantly altered by exposure treatments. Average litter size and kit
survival at birth and 3 weeks of age were also not affected by the exposure treatments. However,
decreased survival of kits in the 3.7 mg/kg tPCB treatment group at 6 weeks of age (i.e., 46%
lower compared to controls) was statistically significant compared to kits in the control and 1.6
mg/kg tPCB treatment groups (Table 1.3-2).
There were no significant differences between treatments for adult female body weights at the
beginning of the study, during the pre-breeding period, and at 3 and 6 weeks post whelping.
However, there was a significant treatment by date interaction for kit body weights from birth to
6 weeks of age. At 3 weeks of age, kits in the 0.0.61 mg/kg tPCB treatment group had
significantly higher body weights when compared to kits in the other five groups. Kits in the 3.7
mg/kg tPCB treatment group had significantly lower body weights when compared to kits in the
other groups. At 6 weeks of age, however, mean kit body weight in the 3.7 mg/kg tPCB
treatment group was only slightly lower than mean kit body weights observed in the control
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-41
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
treatment (251 ± 16.2 g versus 293 ± 11.3 g, respectively). From 10 to 30 weeks of age,
differences in kit body weights between treatments were minor and did not have a dose-
dependent relationship with either tPCBs or TEQ.
Absolute and relative (expressed as a percentage of body weight) brain, heart, spleen, liver,
kidney, and adrenal gland weights of adult females and kits were not significantly different
between treatment groups at necropsy. Differences in absolute heart and liver weights of kits
between treatments were minor and did not have a dose-dependent relationship with either
tPCBs or TEQ. The results of the histological examination of the tissues of major internal
organs of the adult female mink and their kits did not show remarkable changes attributable to
the treatment diets.
9.4.2.3 Mink Jaw Lesion Study
The purpose of this study was to examine the histopathology of jaws of mink from the feeding
study by Bursian et al. (2002). The objective was to determine whether the dietary treatments
induced lesions that have been previously observed in other studies of mink fed PCB-126 and
TCDD. The evaluation was conducted on 6-month kits necropsied at the end of the mink
feeding study.
9.4.2.3.1 Methodology
The skulls of 6-month-old mink kits (36 kits collected) were fixed in a 10% formalin-saline
solution at necropsy, decalcified in 5% nitric acid, rinsed, trimmed, processed using a routine
histotechnologic method, and embedded in paraffin. Tissues were sectioned at 6 microns and
stained with hematoxylin and eosin. Jaws from 36 kits were examined for pathologies. There
were 6 jaw samples from each of the control and 0.34 0.61 0.96, 1.6, and 3.7 mg tPCBs/kg
treatments. The observed lesions were graded as mild, moderate, or severe based on the number
and size of foci of squamous cell proliferation in maxilla and mandibles.
9.4.2.3.2 Results and Discussion
While none of the mink kits had gross abnormalities of the maxilla and mandible, histological
evidence in the form of proliferation of periodontal squamous epithelial cells was present. Nests
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC O /IO 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
of squamous epithelium were found adjacent to the teeth and some had cystic centers. The
proliferation resulted in focal loss of alveolar bone. Squamous cell proliferation was apparent in
17%, 33%, and 100%) of kits in the 0.96, 1.6 and 3.7 mg tPCBs/kg treatments, respectively. No
lesions were observed in the controls, 0.34, and 0.61 mg tPCBs/kg treatments. The lesions
appeared to start from the caudal molar region of the jaw and advanced to the pre-molar, canine,
and incisor regions. The initial lesions in the molar region usually consisted of large cysts lined
with thick layers of stratified squamous epithelium and filled with floating, sloughed squamous
cells. The subsequent lesions in the pre-molar, canine, and incisor regions of the jaw were
characterized as multiple nodules of compact stratified squamous epithelium.
These results indicate that dietary concentrations of PCB-126 as low as 54 ng/kg in the diet (0.96
mg tPCBs/kg diet) can induce maxillary and mandibular squamous cell proliferation. Exposure
of mink to higher concentrations of PCB-126 for longer periods of time, as would be expected in
the Housatonic River ecosystem, would undoubtedly cause increased severity of the lesions
leading to erosion of the mandible and maxilla with concomitant loss of teeth. Such an effect
could ultimately cause the animal to die of starvation (Bursian et al. 2002; 2003).
9.4.2.4 Mink Enzyme Study
Tillitt et al. (2003) performed a study to measure hepatic O-dealkylase activities associated with
cytochrome P450 (CYP) isozymes induced in mink fed diets containing fish collected from the
PSA as part of the MSU feeding study by Bursian et al. (2002). Specific activities were
measured against four separate substrates to measure the induction of CYP enzymes in maternal
and F1 generations of the exposed mink. The induction of CYP enzymes is a good indicator of
exposure to coplanar PCBs, dioxins, and furans, and indicates a first level of toxicological
response (Aulerich et al. 2003). Hepatic activities were measured because the majority of
detoxification of xenobiotics occurs in the liver.
9.4.2.4.1 Methodology
In the feeding study by Bursian et al. (2002), 36 offspring (kits) along with the adults were used
at 6 weeks after whelping. Another 36 kits were used at 6 months post whelping. The livers
were removed and placed in 1.2-ml cryovials and frozen in liquid nitrogen. Frozen liver samples
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC O /I ^ 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
from the parental generation, 6-week-old offspring, and 6-month-old offspring were transmitted
to the Columbia Environmental Research Center (CERC) for analysis. The analyses consisted of
microsomal preparation and various O-dealkylase assays. All procedures were executed
according to CERC Standard Operating Procedures (SOPs) and QA/QC procedures.
9.4.2.4.2 Results and Discussion
Induction of CYP2B-related activity in mink (benzyloxyresorufin-O-deethylase or BROD and
pentoxyresorufin-O-deethylase or PROD) was not substantial at any of the doses of fish from the
Housatonic River. Only a few dose-age treatment combinations had significant inductions
toward BROD or PROD activities. Further, none of the increases in BROD or PROD activities
occurred in a dose-dependent fashion. Thus, the amounts of di- to tetra-ortho-chloro-substituted
PCBs (PCB congeners thought to be responsible for CYP2B-related enzyme inductions) were
either below a threshold of activation of these enzymes in the dietary treatments or the enzyme
induction pathways were saturated. Further analysis (protein content or message) would be
required to discern which of these occurred in these studies.
Induction of CYPlAl-related hepatic enzyme activities (ethoxycoumarin-O-deethylase or ECOD
and ethoxyresorufin-O-deethylase or EROD) was observed to occur in a dose-dependent manner
in all ages of mink examined. Significant increases in these Ah-receptor-regulated enzymes
were observed even in treatments with only a small amount of fish from the Housatonic River
(0.44%) in their diets. These results confirm the known sensitivity of mink to the effects of
tPCBs and other related dioxin-like compounds. The results also confirm that only a small
amount of fish (< 0.5%) from the Housatonic River would be required in the diets of mink to
activate Ah receptor pathways and processes in mink (Tillitt et al. 2003).
9.4.2.5 PCB Congener Comparison in Diet
The composition of PCB congeners in the diet-fish blend used in the mink feeding study was
compared to the congener composition measured in Housatonic River fish likely to be consumed
by mink, which were used to determine modeled exposure concentrations. This comparison was
performed to determine whether there were potential differences in toxicity between the diet
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-44
-------
1
2
blend used in the feeding experiment and the fish that would be consumed by wild mink in the
PSA. See Appendix C.7 for more detail.
3 9.4.2.5.1 Methodology
4 The fish component of the mink diet from the feeding study was analyzed by the USGS
5 Columbia Environmental Research Center (CERC) laboratory for 136 individual congeners plus
6 2 co-eluting pairs of congeners. These congeners collectively total over 95% of Aroclor 1260
7 (Frame et al. 2001). The fish used in the exposure analyses for mink were analyzed by the
8 CERC laboratory for 71 individual congeners, 22 co-eluting pairs, and 2 co-eluting triplets,
9 which also represented over 95% of Aroclor 1260. The individual congeners or congener groups
10 common to the two analyses were used for this evaluation, and included 61 individual congeners,
11 20 pairs, and 2 triplicate congener groups for a total of 83 congeners/congener groups.
12 9.4.2.5.2 Results and Discussion
13 The congener patterns in the feeding study diet were comparable to those in the fish used in the
14 exposure analyses (Figure 1.3-5). There were some exceptions, however. The following
15 congeners were higher (difference between means greater than 2 standard errors) in the feeding
16 study diet relative to at least one species of fish used in the exposure analyses:
26 The following congeners were lower in the feeding study diet relative to the fish samples used in
27 the exposure analyses:
17
18
19
20
21
22
23
24
25
PCB-149/123.
PCB-170/190.
PCB-174.
PCB-136.
PCB-42/3 7/5 9.
PCB-130.
PCB-22/51
PCB-209.
28
29
30
31
¦ PCB-82.
¦ PCB-56.
¦ PCB-67.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-45
7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
When the analysis was repeated for all fish species combined, there was an increase in the
number of congener concentrations that differed by more that two standard errors between the
feeding study diet and exposure analysis fish. That increase was attributed to a drop in standard
error due to the increased number of samples (N = 92). There were 15 congeners that were
higher and 27 congeners that were lower in the feeding study diet than in the exposure analysis
fish.
The percent contribution of several coplanar congeners to the tPCB mixture differs slightly in the
fish used in the feeding study versus the fish used in the exposure analyses. The mean percent
contribution of the most toxic congener, PCB-126, in the fish used in the exposure analyses was
approximately 0.022% compared to approximately 0.005% in the fish used in the feeding study
(Figure 1.3-6). However, because the error bars (+ 2 standard errors) for the two means overlap,
it is unlikely that these differences are statistically meaningful. Therefore, the PCB composition
in fish from the feeding study can be treated as similar to that in fish used in the exposure
analyses and the results from both studies are directly comparable.
9.4.3 Effects Metrics for Characterizing Risk
Effects data can be summarized in a variety of ways ranging from benchmarks designed to be
protective of most or all species to dose-response curves. A summary of the decision criteria
used to derive effects metrics is provided in the text box. Further details on the decision criteria
used in selecting effects metrics is provided in Section 6.6.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-46
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
Decision Criteria for Derivation of Effects Metrics
The following is the hierarchy of decision criteria used to characterize effects for each
receptor-COC combination:
1. Have single-study bioassays with five or more treatments been conducted on the
receptor of interest or a reasonable surrogate? If yes, estimate the
concentration- or dose-response relationships. If not, go to 2.
2. Are multiple bioassays with similar protocols, exposure scenarios and effects
metrics available that, when combined, have five or more treatments for the
receptor of interest or a reasonable surrogate? If yes, estimate the dose-
response relationship as in 1. If not, go to 3.
3. Have bioassays with less than five treatments been conducted on the receptor of
interest or a reasonable surrogate? If yes, conduct or report results of
hypothesis testing to determine the NOAEL and LOAEL. If not, go to 4.
4. Are sufficient data available from field studies and monitoring programs to
estimate concentrations or doses of the COC that are consistently protective or
associated with adverse effects? If yes, develop field-based effects metrics. If
not, go to 5.
5. Derive a range where the threshold for the receptor of interest is expected to
occur. Because information on the sensitivity of the receptor of interest is
lacking, it is difficult to derive a threshold that is biased neither high nor low. If
bioassay data are available for several other species, however, calculate a
threshold for each to determine a threshold range that spans sensitive and
tolerant species. That range is likely to include the threshold for the receptor of
interest.
In this ERA, data were available to derive dose-response curves for mink exposed to tPCBs.
There were insufficient data to derive dose-response relationships for TEQ. Field-based
threshold range was derived instead. There were no toxicity data for river otter. Mink toxicity
data were used as surrogate estimates of toxicity of tPCBs and TEQ to river otter.
9.4.3.1 Effects of tPCBs to Mink and River Otter
Derivation of a dose-response curve requires long-term feeding studies that singly or combined
have at least five dose treatments for sensitive endpoints such as mortality or reproductive
success. The acceptable studies that met these criteria were the Bleavins et al. (1980) and
Aulerich et al. (1985) studies. Because both studies used similar protocols, exposure duration,
and species (a similar strain of farm-raised mink), they were combined to yield a data set with
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-47
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
nine treatments for fecundity. Figure 9.4-1 presents the dose-response curve for reduced
fecundity of mink exposed to tPCBs. The dose-response curve indicates that 10% and 20%
declines in fecundity would be expected at doses of 0.0128 and 0.0272 mg/kg bw/d, respectively.
0 0.5 1 1.5 2
Dose (mg/kg bw/day)
Note: Symbols indicate raw data.
Figure 9.4-1 Dose Response Curve for Effects of tPCBs on Fecundity of Mink
9.4.3.2
Effects of TEQ to Mink and River Otter
The studies by Heaton et al. (1995), Hochstein et al. (1998, 2001), and Aulerich et al. (1988)
involved exposing mink to fish contaminated with TEQ and other contaminants. Based on a
review of these studies, adverse effects on growth in kits begin to occur at concentrations of 3.6
ng/kg bw/d (lower toxicity threshold; Heaton et al. 1995). The highest dose that did not cause
adverse effects was 36 ng/kg bw/d (upper toxicity threshold; Hochstein et al. 2001). Thus, the
threshold range, based on studies that used field-collected fish, is 3.6 to 36 ng/kg bw/d for
piscivorous mammals.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-48
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
9.5 RISK CHARACTERIZATION
This section characterizes the risk to piscivorous mammals exposed to tPCBs and TEQ in the
PSA of the Housatonic River. The risk characterization uses three lines of evidence to determine
ecological risks to this group of mammals. The three major lines of evidence are considered to
be independent and are combined in a weight-of-evidence (WOE) assessment. The key risk
questions and the three lines of evidence are summarized in the text box.
Key Risk Questions
¦ Are the concentrations of tPCBs and TEQ present in the prey of piscivorous
mammals sufficient to cause adverse effects to individuals inhabiting the PSA of
the Housatonic River?
¦ If so, how severe are the risks and what are their potential consequences?
Lines of Evidence
¦ Use of semi-qualitative field surveys.
¦ Probabilistic exposure and effects modeling.
¦ Feeding study using fish from the PSA.
Section 9.5.1 presents a brief overview of the methodology, results, and interpretation of the
mammal surveys conducted from 1998 to 2001 in the Housatonic PSA. A more detailed
presentation of this information can be found in Appendix A. In Section 9.5.2, the dose-response
curves are combined with the corresponding exposure distributions to derive risk curves that
characterize the relationship between probability and magnitude of effect. The results of the
mink feeding study are summarized in Section 9.5.3. A WOE assessment is presented is Section
9.5.4, along with sources of uncertainty (Section 9.5.5) and the overall findings of the risk
characterization (Section 9.5.8).
9.5.1 Field Surveys
9.5.1.1 EPA Study
The mammalian community in the PSA was studied by EPA over a 4-year period, from 1998 to
2001. Surveys were conducted to record presence, relative abundance, and habitat usage for
small and large mammals including mink and river otter. A variety of field survey techniques
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC Q /I Q 7/10/2003
-------
1 including small mammal trapping, snow tracking, and scent-post station surveys were used to
2 characterize the mammalian community. Mink tracks and scats were observed at several
3 locations in the PSA during snow-tracking surveys. Tracks were observed at the northern and
4 southern-most portions of the PSA; no observations occurred in the middle portions. Mink were
5 also observed at the Washington Mountain Lake and Ashley Lake reference areas during the
6 1999 and 2000 surveys. On a per effort basis, mink were observed in the reference areas twice
7 as frequently as in the PSA (Table 9.5-1).
8
Table 9.5-1 Results of Snow Tracking and Scent Post Station Surveys
in the PSA and Reference Areas
9
10
Primary Study Area
Mink
Otter
Hours of survey effort
260.5
260.5
11
Number of sightings
5
4
No. sightings/hour
0.019
0.015
12
Reference Areas
13
Hours of survey effort
108.0
108.0
Number of sightings
4
14
14
No. sightings/hour
0.037
0.130
15
16 River otter signs were observed on only four occasions in three locations in the PSA. Otter
17 tracks, slides, and scats were observed in the reference areas relatively frequently. A nearly 9:1
18 ratio of observations per unit effort occurred between the reference areas and the PSA during the
19 tracking surveys (Table 9.5-1) and, additionally, families of river otter were repeatedly observed
20 in the reference areas during the course of other field surveys.
21 The Housatonic River in the PSA offers an abundance of habitat that meets the requirements for
22 mink and river otter. Their occurrence in the PSA, however, was much lower than would be
23 expected, considering the large amount of available habitat and food resources. Despite
24 hundreds of hours specifically conducting track and scent post surveys for these species and
25 thousands of person-days spent conducting other field surveys and sampling, only a handful of
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-50
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
observations of each species occurred. Based on the experience of the field personnel and the
substantial number of hours spent in the study area from 1998 to 2001, far fewer observations of
mink and otter, or their sign, occurred than would be normally expected in a riverine system such
as the PSA.
9.5.1.2 GE Study
The General Electric (GE) company studied the presence or absence and possible distribution of
wild mink in the Housatonic River Study Area from the spring of 2001 to the spring of 2003
(Bernstein et al. 2003). River otter (Lutra canadensis) were included in the study of winter of
2002/2003. The methods were similar to those used in the EPA surveys, and consisted of
looking for tracks in soft sand at scent post stations in the spring, summer, and fall (mink only),
and snow tracking (mink and otter) during the winter months. The survey efforts were
concentrated in suitable mink habitat, irregular shorelines and backwaters with dense, wooded
cover near the water. Additional efforts included setting traps (100 in total), scented burrows,
and motion sensitive camera sites. The observations were conducted along the Housatonic River
between New Lenox Road and Woods Pond (the midpoint of Reach 5B, and Reach 5C and 6; no
work was performed in the upstream half of the PSA).
The study reported 35 sets of mink tracks between April 2001 and March 2002. A total of 33
mink track sets and 41 river otter track sets were observed in 2003. However, only 4 out of the
35 mink track sets were observed in the snow-free months. In 2003, all mink and river otter
tracks were observed in winter. This suggests that either the observation methods used in the
spring, summer, and fall were not effective or that mink were not present in the PSA. The
former could have been due to issues with methods used during the construction of scent posts
(e.g., failure to wear rubber boots and gloves during the construction of the scent posts)
(Woodlot Alternatives, Inc. 2003). If the lack of sightings during the spring, summer, and fall
was due to lack of mink in the PSA (a hypothesis supported by very few tracks at post stations,
no photographs recorded by the motion-sensitive camera in the snow-free months, and no tracks
in scented burrows), then the tracks detected in winter likely belonged to transient mink rather
than to local residents. The 2003 survey discovered one confirmed and one suspected river otter
den site in winter, suggesting that that some river otter might be present in the PSA for extended
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC q r -> 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
periods of time. However, the presence of one, perhaps a temporary den site, in winter is
insufficient as evidence for a healthy and reproducing river otter population in the PSA.
In summary, the GE report cited incidences of mink and river otter signs in the PSA. However,
the study had several limitations, which lead to conclusions that are not supported by the data.
These limitations included the failure to discuss the implications of the disproportionate number
of sightings in winter versus other seasons, apparent ineffectiveness of scent posts, no
established reference areas outside the PSA, lack of tracking expertise and experience, empty
traps, no results from motion-sensitive camera trials (i.e., no mink or river otter observed in the
snow-free months trials), uncertainty in determining sex of mink from tracks, and uncertainty in
attributing a different sets of tracks to separate individuals.
9.5.2 Comparison of Estimated Exposures to Laboratory-Derived Effects Doses
Exposure was assessed for mink and river otter in the PSA. Because Reaches 5 and 6 combined
roughly correspond to the size of the home range for otter, these reaches were combined for the
river otter analysis. For mink, the assessment was conducted separately for each reach because
of their smaller foraging range. For comparative purposes, exposure was also estimated for mink
and otter in two reference areas: the upstream reference area and Threemile Pond. Moreover,
exposure was also estimated for mink and river otter foraging 50, 25, and 10% of time in the
PSA. For each receptor-COC combination, a category of low, intermediate, or high risk was
assigned following integration of the exposure and effects distributions. This exercise was done
separately for the results of the Monte Carlo analyses and each of the lower and upper bounds
from the probability bounds analyses. The "risk category" refers to the level of risk based on the
results of the Monte Carlo analysis. The "risk range" refers to the levels of risk based on the
results of the probability bounds analyses.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-52
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Guidance for Determining Level of Risk to Representative Species
Risk Curves for Mink and River Otter Available
¦ If the probability of 10% or greater effect is less than 20%, then the risk to
piscivorous mammals is low.
¦ If the probability of 20% or greater effect is greater than 50%, then the risk to
piscivorous mammals is high.
¦ All other outcomes are considered to have intermediate risk.
Risk Curves for Mink and River Otter Not Available
¦ If the probability of exceeding the lower toxicity threshold was less than 20%, the
risk to piscivorous mammals was low.
¦ If the probability of exceeding the upper toxicity threshold was greater than 20%,
the risk to piscivorous mammals was high.
¦ All other outcomes for the lower and upper thresholds were considered to have
intermediate risk.
The results of the risk characterization are summarized in Table 9.5-2. Figures 9.5-1 through to
9.5-14 depict the risk curves for mink and river otter exposed to tPCBs in Reaches 5 and 6,
including the Monte Carlo estimate and both the lower and upper probability bounds (LPB and
UPB, respectively). The highest risk to mink and river otter is from exposure to tPCBs in
Reaches 5 and 6. Risks were much lower in the reference areas. The exposure in Reaches 5 and
6 is so high that individuals foraging 10% of their time at those locations would experience high
risk.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-53
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Table 9.5-2
Summary of Qualitative Risk Statements for Piscivorous Mammals from the
Housatonic River Study Area
Location
Qualitative Risk Statements
PCBs
TEQ
Risk Category
Risk Range
Risk Category
Risk Range
Mink
Reach 5
High
High
High
Intermediate/High
Reach 6
High
Intermediate/High
High
Intermediate/High
Upstream Reference
Area
Intermediate
Low/High
Intermediate
Low/High
Threemile Pond
High
Low/High
Low
Low/High
River Otter
Reaches 5 and 6
High
High
High
High
Upstream Reference
Area
Intermediate
Low/Intermediate
Intermediate
Low/Intermediate
Threemile Pond
Low
Low/Intermediate
Low
Low/Intermediate
The degree of risk for mink (and otter) associated with exposure to tPCBs downstream of Woods
Pond was assessed by comparing concentrations of tPCBs in prey fish of mink (5 to 20 cm) in
Reaches 7 to 16 to a maximum acceptable threshold concentration (MATC) developed
specifically for mink (Appendix I). Fish residue data were obtained from sampling efforts from
1998 to 2002. The MATC of 2.65 mg/kg tPCBs in fish (whole body, wet weight) was developed
as the geometric mean of the NOAEL and LOAEL developed by Bursian et al. (2002) in the site-
specific study of the toxicity of Housatonic River fish to mink. The LOAEL was based on the
observation of significantly reduced mink kit survivability at 6 weeks of age. The value of this
LOAEL was estimated at 3.7 mg/kg feed supplied to reproducing females. The NOAEL was
based on the same endpoint and its value was 1.6 mg/kg feed.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-54
-------
Reach 5
- 60
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
0 Low - Intermediate Criterion
o Intermediate - High Criterion
40 50 60
% Decline in Fecundity
Figure 9.5-1 Total PCB Risk to Mink Exposed to tPCBs in Reach 5 of the
Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-55
-------
1
Reach 5
100
a
C3
A
W
, Monte Carlo
, LPB - Lower Probability Bound
UPB - Upper Probability Bound
Low - Intermediate Criterion
Intermediate - High Criterion
2
3
4
10 20 30 40 50 60 70
% Decline in Fecundity
90
100
Figure 9.5-2 Total PCB Risk to Mink (10% Foraging Time in Reach 5)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-56
-------
Reach 6
100
-a
d
-fi
o
80
60
o
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
Low - Intermediate Criterion
Intermediate - High Criterion
a
C3
¦d
40
h
W
20
_l_
10 20 30 40 50 60 70
% Decline in Fecundity
80
90
100
Figure 9.5-3 Total PCB Risk to Mink Exposed to tPCBs in Reach 6 of the
Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-57
-------
Reach 6
100
£
h
W
60
40
20
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
0 Low - Intermediate Criterion
O Intermediate - High Criterion
10 20 30 40 50 60 70
% Decline in Fecundity
90
100
2
3
Figure 9.5-4 Total PCB Risk to Mink (10% Foraging Time in Reach 6)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-58
-------
Upstream Reference Area
Figure 9.5-5 Total PCB Risk to Mink Foraging in the Upstream Reference Area
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-59
-------
Upstream Reference Area
100
§ 60
h
W
40
20
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
9 Low - Intermediate Criterion
O Intermediate - High Criterion
¦
10 20 30 40 50 60
% Decline in Fecundity
70
90
100
1
2
3
4
Figure 9.5-6 Total PCB Risk to Mink (10% Foraging Time in the Upstream
Reference Area)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-60
-------
Threemile Pond Reference Area
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
0 Low - Intermediate Criterion
O Intermediate - High Criterion
10 20 30 40 50 60
% Decline in Fecundity
70
90
100
Figure 9.5-7 Total PCB Risk to Mink Exposed to tPCBs in the Threemile Pond
Reference Area
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-61
-------
Threemile Pond Reference Area
100
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
0 Low - Intermediate Criterion
O Intermediate - High Criterion
10 20 30 40 50 60
% Decline in Fecundity
70
90
100
Figure 9.5-8 Total PCB Risk to Mink (10% Foraging Time in the Threemile Pond
Reference Area)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-62
-------
Reaches 5 and 6
100
1
2
3
4
5
£
X
W
60
40
20
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
0 Low - Intermediate Criterion
O Intermediate - High Criterion
10 20 30 40 50 60
% Decline in Fecundity
70
90
100
Note: The LPB and UPB overlap the Monte Carlo line.
Figure 9.5-9 Total PCB Risk to River Otter Exposed to tPCBs in Reaches 5 and 6
of the Housatonic River
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-63
-------
Reaches 5 and 6
100
£
h
W
60
40
20
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
# Low - Intermediate Criterion
O Intermediate - High Criterion
10 20 30 40 50 60
% Decline in Fecundity
70
90
100
Figure 9.5-10
Total PCB Risk to River Otter (10% Foraging Time in Reaches
5 and 6)
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-64
-------
Upstream Reference Area
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
# Low - Intermediate Criterion
O Intermediate - High Criterion
10 20 30 40 50 60
% Decline in Fecundity
70
90
100
Figure 9.5-11 Total PCB Risk to River Otter Foraging in the Upstream
Reference Area
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-65
-------
Upstream Reference Area
% Decline in Fecundity
1
2 Figure 9.5-12 Total PCB Risk to River Otter (10% Foraging Time in the
3 Upstream Reference Area)
4
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-66
-------
Threemile Pond Reference Area
% Decline in Fecundity
1
2 Figure 9.5-13 Total PCB Risk to River Otter Exposed to tPCBs in the
3 Threemile Pond Reference Area
4
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-67
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Threemile Pond Reference Area
I 40
¦d
20
Monte Carlo
LPB - Lower Probability Bound
UPB - Upper Probability Bound
A Low - Intermediate Criterion
O Intermediate - High Criterion
o
I •
0 10 20 30 40 50 60 70 80 90 100
% Decline in Fecundity
Figure 9.5-14 Total PCB Risk to River Otter (10% Foraging Time in the
Threemile Pond Reference Area)
9.5.3 Mink Feeding Study
Consumption of diets containing tPCBs and TEQ derived from fish collected from the
Housatonic River did not have an adverse effect on adult mink reproduction as assessed by
breeding success, whelping success, and gestation length. Kit survival at 6 weeks of age,
however, was significantly decreased in the 3.7 mg/kg tPCBs (68.5 ng/kg TEQ) treatment group.
In this treatment, less than 4% of the diet was derived from Housatonic River fish, which is well
below what mink typically consume in the wild (23% on average with a range from 0 to 65%).
The enzyme induction analysis indicated that ECOD and EROD enzymes were induced even
when the Housatonic River fish content in the diet was very low (0.44%; 0.50 mg/kg tPCBs)
attesting to the high sensitivity of mink to tPCBs and TEQ. The histopathological examination
of kit jawbones revealed that jaw lesions were apparent at tPCB treatments as low as 0.96 mg/kg
diet (0.88%) Housatonic River fish)(Bursian 2002, 2003).
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-68
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
9.5.4 Weight-of-Evidence Analysis
A weight-of-evidence (WOE) analysis was used to combine the three major lines of evidence
described in the preceding sections for mink and river otter. The goal of this analysis was to
determine whether significant risk is posed to piscivorous mammals in the Housatonic River
PSA as a result of exposure to tPCBs and TEQ. The three-phase approach of Menzie et al.
(1996) and the Massachusetts Weight-of-Evidence Workgroup was used, in which WOE was
expressed with the following three characteristics: (1) the weight assigned to each measurement
endpoint; (2) the magnitude of response observed in the measurement endpoint; and (3) the
concurrence among outcomes of the multiple measurement endpoints.
Each measurement endpoint was evaluated and assigned a qualitative weight in Appendix I,
along with a discussion of the reason for the value assigned (Table 9.5-3). The EPA field survey
had a moderate to high value, the GE field survey had a moderate value, the feeding study had a
high value, and the modeled exposure and effects line of evidence for tPCBs and TEQ had a high
and moderate/high values in the overall WOE analysis.
The magnitude of the response in the measurement endpoint is considered together with the
measurement endpoint value in developing the overall WOE (Menzie et al. 1996). This requires
assessing the strength of evidence that ecological harm has occurred, as well as an indication of
the magnitude of response, if present. The weighting values, evidence of harm, and magnitude
of response were combined in a matrix format and are presented in Tables 9.5-4 and 9.5-5.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-69
-------
Table 9.5-3
Weighting of Measurement Endpoints for Piscivorous Mammals Weight-of-
Evidence Evaluation
Attributes
Field Surveys
Feeding
Study
Modeled Exposure and
Effects
EPA
GE
tPCBs
TEQ
I. Relationship Between Measurement and Assessment Endpoints
1. Degree of Association
L/M
L/M
H
H
H
2. Stressor/Response
M
L/M
M/H
H
M
3. Utility of Measure
M
L/M
H
M/H
M
II. Data Quality
4. Data Quality
M/H
L
H
M/H
M/H
III. Study Design
5. Site Specificity
H
M/H
M
L/M
L/M
6. Sensitivity
M
L/M
H
H
H
7. Spatial Representativeness
H
M/H
M/H
M/H
M/H
8. Temporal Representativeness
H
M/H
M/H
M/H
M/H
9. Quantitative Measure
L
L
H
H
H
10. Standard Method
H
M/H
H
M/H
M/H
Overall Endpoint Value
M/H
M
H
H
M/H
5 L = low; M = moderate; H = high
6
7
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC q qrs 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
Table 9.5-4
Evidence of Harm and Magnitude of Effects for Piscivorous Mammals Exposed to
tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No,
Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Surveys
EPA
Moderate/High
Yes
High
GE
Moderate
No
Low
Feeding Study
High
Yes
High
Modeled Exposure and
Effects
Moderate/High
Yes
High
Table 9.5-5
Evidence of Harm and Magnitude of Effects for Piscivorous Mammals Exposed to
TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Surveys
EPA
Moderate/High
Yes
High
GE
Moderate
No
Low
Feeding Study
High
Yes
High
Modeled Exposure and
Effects
Moderate/High
Yes
High
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC O 71 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
All three lines of evidence indicated that the elevated concentrations of tPCBs and TEQ in the
PSA of the Housatonic River are causing adverse effects of high magnitude to mink and river
otter. The field surveys indicated that mink and river otter are rarely present in the PSA, except
during winter, and likely have not established home territories close to the main channel despite
suitable mink and otter habitat. The MSU feeding study indicated that feeding adult female mink
with a diet containing as little as 3.51% fish from the PSA caused a statistically significant
reduction (46% compared to controls) in kit survival to 6 weeks of age. Because mink in the
wild typically consume between 20% or more fish in their diet, the associated risk is
correspondingly higher. In addition, other components of the mink diet in the PSA (e.g.,
crayfish) have high concentrations of tPCBs and TEQ. Further, the jaw lesion study indicated
that erosion of the jaw occurs at even lower doses and exhibits a dose-response relationship.
Such effects could eventually lead to starvation. The occurrence of jaw lesions coincides with
the induction of Ah-receptor-regulated enzymes (ECOD and EROD) also in a dose-response
manner.
The high risks evident from the feeding study are further supported by the modeled exposure and
effects line of evidence. The estimated potential for exposure is so high that even individual
mink and otter that only forage in the PSA for short periods of time (less than or equal to 10% of
foraging time) are at an intermediate or higher risk from tPCBs and TEQ.
A graphical method was used for displaying concurrence among measurement endpoints. Tables
9.5-6 and 9.5-7 depict the outcome for piscivorous mammals exposed to tPCBs and TEQ,
respectively. The three measurement endpoints have a high degree of concurrence.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-72
-------
1
2
3
4
5
6
7
8
9
10
11
Table 9.5-6
Risk Analysis Summary for Piscivorous Mammals Exposed to tPCBs
in the Housatonic River PSA
Assessment Endpoint: Survival, growth, and reproduction of piscivorous mammals
Harm/Magnitude
Weighting Factors (increasing confidence of weight)
Low
Low/Moderate
Moderate
Moderate/High
High
Yes/High
MEE, FS-EPA
MFS
Yes/Indeterminate
Yes/Low
Undetermined/High
Undetermined/Intermediate
Undetermined/Low
No/Low
FS-GE
No/Intermediate
No/High
FS-EPA = Field surveys by EPA
FS-GE = Field surveys by GE
MFS = Mink feeding study
MEE = Modeled exposure and effects
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-73
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Table 9.5-7
Risk Analysis Summary for Piscivorous Mammals Exposed to TEQ
in the Housatonic River PSA
Assessment Endpoint: Survival, growth, and reproduction of piscivorous mammals
~
Harm/Magnitude
Weighting Factors (increasing confidence of weight)
Low
Low/Moderate
Moderate
Moderate/High
High
Yes/High
MEE, FS-EPA
MFS
Yes/Intermediate
Yes/Low
Undetermined/High
Undetermined/Intermediate
Undetermined/Low
No/Low
FS-GE
No/Intermediate
No/High
FS-EPA = Field surveys by EPA
FS-GE = Field surveys by GE
MFS = Mink feeding study
MEE = Modeled exposure and effects
9.5.5 Sources of Uncertainty
The assessment of risk to piscivorous mammals contains uncertainties. Each source of
uncertainty can influence the estimates of risk. Therefore, it is important to describe, and when
possible, specify the magnitude and direction of such uncertainties. Some of the major sources
of uncertainty associated with the assessment of risks of tPCBs and TEQ to piscivorous
mammals are briefly described below. An expanded discussion is presented in Appendix I.
¦ In this assessment, it was assumed that dietary exposure represented the most
important pathway of exposure for piscivorous mammals exposed to COCs.
Although unlikely to provide a major contribution to the risk, other pathways could
increase the exposure and perhaps increase risk slightly (Moore et al. 1999). Other
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC q 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
pathways include drinking water intake, incidental ingestion of sediment, inhalation,
transdermal uptake, and preening activity. When drinking water was included in a
screening level analysis for piscivorous mammals, the results showed negligible
increases in exposure due to drinking water and their likely low importance. The
remaining pathways were not assessed due to the difficulty in quantifying intake via
those routes.
¦ The Monte Carlo sensitivity analyses indicated that the free metabolic rate (FMR)
slope and power terms were generally the most influential variables on predicted total
daily intakes of COCs. However, no measurements of free metabolic rate were
available for the representative wildlife species. Similarly, measured food intake
rates were not available for free-living mink and river otter or reasonable surrogate
species. Therefore, free metabolic rates were estimated using allometric equations.
The use of allometric equations introduces some uncertainty into the exposure
estimates because they have model-fitting error and are based on species different
from the representative species used in this assessment. For mink and river otter, the
carnivora model of Nagy et al. (1999) was selected as the most appropriate allometric
model to estimate free metabolic rate. Examples of other species used in the model
included the cat, fox, dog, and wolf. Given the lack of data on representative species
used in the current assessment, it is difficult to judge the magnitude of the uncertainty
introduced by the use of the allometric models. The uncertainty due to model-fitting
error was propagated in the uncertainty analyses by using distributions as inputs for
the allometric slope and power terms.
¦ Because no stomach contents or other dietary analyses were available for mink in the
PSA, dietary compositions were derived from those reported in the literature from
other similar geographical locations. The potential magnitude and direction of the
uncertainty associated with lack of information on diet are unknown. The uncertainty
due to lack of knowledge on diet of mink in the PSA was partially addressed by using
distributions to represent variability in diets observed at other similar sites. Small
mammals were the most contaminated prey, thus, any increases in the proportion of
this type of prey consumed would lead to increases in exposure. Conversely,
amphibians were the least contaminated prey, thus increases in the intake of this prey
item would lead to decreases in exposure to tPCBs and TEQ.
¦ Sample sizes were limited for the analyses of COC concentrations in some prey
items. For example, there were only four amphibian samples for tPCBs from the
upstream reference area, only one amphibian sample for TEQ from the upstream
reference area and the Threemile Pond reference area, and only three invertebrate
samples for TEQ from the upstream reference area. Uncertainty due to sample size
was explicitly addressed in the uncertainty analyses. In the Monte Carlo analysis,
sample size uncertainty was addressed by use of the 95% upper confidence limit
(UCL) on the mean. The use of the UCL addressed uncertainty, but it was biased
toward overestimating exposure. In the probability bounds analysis, uncertainty was
addressed by specifying concentration variables as a range from the 5% lower
confidence limit (LCL) to the UCL. This treatment of uncertainty was unbiased.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-75
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
¦ Data on concentrations of tPCBs and TEQ in crayfish and mammals were not
available for Reach 6. In those cases, the concentrations of tPCBs and TEQ in prey
were estimated using ratios between prey items at other locations for which full data
sets existed. This type of extrapolation introduces some uncertainty regarding the
concentration of COCs in prey tissue, although the magnitude and direction of this
uncertainty is difficult to judge.
¦ In some instances, data on concentrations of tPCBs and TEQ in crayfish, birds, and
mammals from reference areas were missing. In those cases, an assumption was
made that these prey items contained no detectable residues of the two contaminants.
The magnitude of uncertainty introduced by this assumption is likely to be small
because sediment data indicate that detectable tPCB and TEQ residues are rare at
those sites.
¦ The base exposure scenario for mink and river otter assumed that these animals
would forage 100% of their time in the PSA. This assumption is reasonable given the
similarity between the size of the PSA and the foraging ranges of these species.
However, some individuals might forage part of their time outside the PSA on less
contaminated prey. The exposure and risk analyses indicated that individual mink
and river otter that forage even a small fraction of time in the PSA (10%) are at high
risk, particularly for tPCBs.
¦ The effects metrics used to estimate risk to piscivorous mammals via exposure
models were derived for Aroclor 1254 mixtures. Some uncertainty is inherent in
extrapolating from studies using the Aroclor 1254 mixture to the specific congener
patterns observed in weathered mixtures in the PSA of the Housatonic River. The
feeding study with mink using fish from the PSA suggested that the PCB mixture in
fish (most closely resembling Aroclor 1260) was less toxic than the PCB mixture
reported in literature (Aroclor 1254). Thus, the risk to mink and river otter estimated
by the exposure model may be slightly overestimated. This overestimate, however,
does not affect the final risk conclusion due to the very high exposure rates for mink
and river otter.
¦ The comparison of PCB congeners in the diet-fish blend used in the MSU study to the
congener composition measured in Housatonic River fish revealed (with few
exceptions) that the congener patterns (and potency) in the feeding study diet were
comparable to those in the fish used in the exposure analyses. However, there was
some uncertainty as to the influence of PCB-126 on the toxicity of the treatment diet
(PCB-126 content of 0.005%) vs. exposure analysis fish (PCB-126 content of
0.022%). Although the percentages were within two standard errors of each other
(criteria for similarity), the difference might have contributed to the lower than
expected toxicity of tPCBs observed in the feeding study.
¦ There was some uncertainty whether other COCs present in Saginaw Bay fish
contributed to the increased toxicity of those fish compared to fish from the
Housatonic River.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-76
-------
1 "It was uncertain whether the food intake rates of the mink fed fish from the
2 Housatonic River were comparable to the corresponding rates observed for mink in
3 the Saginaw Bay study.
4 ¦ There was uncertainty whether the congener mixture in the Housatonic River fish has
5 the same potency as the mixture in the Saginaw Bay fish.
6 ¦ The GE mink and otter study lacks critical information needed to confirm track
7 identification (i.e., multiple measurements with a scale) and sex determination (i.e.,
8 photographs of tracks, most without a scale, were sent out of state to scientists in
9 Louisiana). In addition, data on spacing between paired tracks and on lope distance
10 were not presented to segregate male long-tailed weasels from female mink. Without
11 this supporting information, the results and interpretation remain questionable.
12 "In the GE study, the use of a study area that represents only a portion of Housatonic
13 River and adjacent floodplain known to be affected by PCB contamination, and the
14 lack of reference areas creates uncertainty, and limits the ability to draw inferences
15 about whether the number of individual mink and otter observed was comparable to
16 other uncontaminated sites in this area.
17 9.5.6 Comparison to Other Piscivorous Mammals
18 There are no piscivorous mammals other than mink and river otter in the PSA.
19 9.5.7 Risk Downstream of PSA
20 The risk for mink and river otter associated with exposure to tPCBs downstream of Woods Pond
21 was assessed by comparing concentrations of tPCBs in prey fish (5 to 20 cm) in Reaches 7 to 16
22 to a maximum acceptable threshold concentration (MATC) developed specifically for mink (also
23 used for river otter). For the downstream assessment for mink, it was assumed that the fish
24 constituted 23% of the diet and invertebrates (mostly crayfish) 36% of the diet. The remaining
25 41% of the diet consisted of other uncontaminated dietary items. No crayfish data were available
26 for the downstream reaches. However, crayfish residues in the PSA were similar to the fish
27 residues. Therefore, it was assumed that 59% of the diet of mink foraging downstream of Woods
28 Pond was composed of fish. For the downstream assessment of river otter, it was assumed that
29 100%) of the diet was composed of fish. On average, however, river otter consume
30 approximately 80% fish and 20% crayfish. Crayfish data were not available for the downstream
31 reaches and tissue residue concentrations in crayfish were approximated using fish residues
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-77
-------
1 (which were similar in the PSA). Thus, the assumption of 100% fish diet for downstream otter
2 was reasonable.
3 The MATC of 2.65 mg/kg tPCBs in fish (whole body, wet weight) was developed as the
4 geometric mean of the NOAEL and LOAEL developed by Bursian et al. (2002) in a site-specific
5 study of the toxicity of Housatonic River fish to mink. The LOAEL was based on the
6 observation of significantly reduced mink kit survivability at 6 weeks of age. The value of this
7 LOAEL was estimated at 3.7 mg/kg feed supplied to reproducing dams. The NOAEL was based
8 on the same endpoint and its value was 1.6 mg/kg feed.
9 To determine the extent and types of habitats available for mink and river otter downstream, U.S.
10 Fish and Wildlife Service National Wetland Inventory, U.S. Geological Survey Topographical
11 Quadrangle maps, and aerial photos of the river were examined in detail. The species-habitat
12 matrix in Appendix A.2 (Ecological Characterization of the Housatonic River Downstream of
13 Woods Pond) identified potential habitat for mink and otter. According to this analysis, potential
14 mink habitat is ubiquitous and includes all areas except high gradient stream, calcareous rock
15 cliff, cultural grassland, agricultural cropland, and residential/industrial development. Potential
16 river otter habitat is less abundant and centers more on larger wetland systems, with slower
17 flowing water, or with impounded water. Any places where the river is impounded, or near a
18 lake or pond, there is potential river otter habitat.
19 Fish tissue data were obtained from sampling efforts conducted during 1998 to 2002. The results
20 of the analysis are presented in Figures 1.5-15 and 1.5-16. Potential risk to mink and river otter
21 exists in river sections from Woods Pond to the end of Reach 10 (mink) and 12 (river otter).
22 9.5.8 Conclusions
23 For piscivorous mammals, data from three major lines of evidence were available, including
24 field surveys, mink feeding study, and exposure and effects modeling. In general, the weight-of-
25 evidence analysis indicates an intermediate to high risk for mink and river otter to tPCBs and
26 TEQ in the PSA.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-78
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Field surveys by EPA and GE were conducted to determine the presence of mink and river otter
in the PSA. The surveys were not designed to provide a quantitative evaluation of the
relationship between exposure to COCs and the survival, growth, and reproduction of
piscivorous mammals in the PSA. Instead, the surveys determined the presence and relative
abundance of piscivorous mammals. Signs of mink and river otter were observed in the PSA,
but mostly in winter, suggesting that mink and river otter that are present in the PSA are there on
a transient basis.
The mink feeding study was designed to determine the effects on growth and reproduction of
captive mink fed a diet containing fish from the PSA. The results from this study indicated that
feeding adult female mink with a diet containing as little as 3.51% fish from the PSA caused a
statistically significant reduction (46% compared to controls) in kit survival to 6 weeks of age.
Because mink in the wild typically consume between 0 and 65% fish in their diet (mean 23%),
the associated risk is correspondingly higher. Further, the jaw lesion study indicates that erosion
of the jaw occurs at even lower doses and exhibits a dose-response. Such effects could
eventually lead to starvation. The occurrence of jaw lesions coincides with the induction of Ah-
receptor-regulated enzymes (ECOD and EROD) also in a dose-responsive manner.
The modeling of exposure and effects line of evidence was used to determine the level of risk to
the representative mammal species, mink and river otter. The effects characterization developed
a dose-response curve to describe the potential effects of tPCBs to piscivorous mammals.
Toxicity benchmarks based on mink studies were developed for TEQ. The dose-response curve
for effects of tPCBs to piscivorous mammals indicated that 10% and 20% declines in fecundity
would be expected at doses of 0.0128 and 0.0272 mg/kg bw/d, respectively. For TEQ
benchmarks (reduction in kit growth), the lower threshold was set at 3.6 ng/kg bw/d and the
upper threshold was set at 36 ng/kg bw/d. The modeled exposure results indicated that the daily
intake rates of tPCBs by mink and river otter were far greater than the toxicity thresholds. This
means that mink and river otter feeding in the PSA receive tPCB doses that cause adverse
reproductive effects. A similar conclusion was reached for TEQ.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-79
-------
1
2
ERA Summary
3
4
The weight-of-evidence analysis indicates an intermediate to high risk for mink and
river otter exposed to tPCBs and TEQ in the PSA.
5
6
The risk continues to be elevated for individuals that forage only a small fraction of
their time in the PSA.
7
Downstream of Woods Pond (Reach 6), a screening level ERA indicated that mink
and river otter may be at risk from exposure to tPCBs and TEQ as far as Reach 10
and 12, respectively.
8
9
10
11
12 9.6 REFERENCES
13 Alexander, G.R. 1977. Food of vertebrate predators on trout waters in north central lower Michigan.
14 The Michigan Academician 10:181-195.
15 Allen, A.W. 1986. Habitat Suitability Index Models: Mink, Revised. U.S. Fish and Wildlife Service
16 Biological Report 82(10.127). 23 pp.
17 Anderson, E.A., and A. Woolf. 1987. River otter food habits in northwestern Illinois. Transactions
18 of the Illinois State Academy of Science 80:115-118.
19 Arnold, T.W., and E.K. Fritzell. 1987. Activity patterns, movements, and home ranges of prairie
20 mink. Prairie Naturalist 19:25-32.
21 Arnold, T.W., and E.K. Fritzell. 1990. Habitat use by male mink in relation to wetland
22 characteristics and avian prey abundance. Canadian Journal of Zoology 68:2205-2208.
23 Aulerich, R.J., R.K. Ringer, and S. Iwamoto. 1973. Reproductive failure and mortality in mink
24 fed on Great Lakes fish. Journal of Reproduction and Fertility Supplements 19:365-376.
25 Aulerich, R.J., S.J. Bursian, W.J. Breslin, B.A. Olson, and R.K. Ringer. 1985. Toxicological
26 manifestations of 2,4,5,2',4',5'-, 2,3,6,2',3',6'-, and 3,4,5,3',4',5'-hexachlorobiphenyl and
27 Aroclor® 1254 in mink. Journal of Toxicology and Environmental Health 15:63-79.
28 Aulerich, R.J., S.J. Bursian, M.G. Evans, J.R. Hochstein, K.A. Koudele, B.A. Olson, and A.C.
29 Napolitano. 1987. Toxicity of 3,4,5,3',4',5'-hexachlorobiphenyl to mink. Archives of
30 Environmental Contamination and Toxicology 16:53-60.
31 Aulerich, R.J., S.J. Bursian, and A.C. Napolitano. 1988. Biological effects of epidermal growth
32 factor and 2,3,7,8-tetrachlorodibenzo-p-dioxin on developmental parameters of neonatal mink.
3 3 Archives of Environmental Contamination and Toxicology 17:27-31.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC q qa 7/10/2003
-------
1 Aulerich, R.J., S. Bursian, and B. Yamini. 2003. Hepatic 0-dealkylase activities from induction of
2 cytochrome P-450 in mink and their offspring fed diets containing contaminated fish from the
3 Housatonic River. USGS Biochemistry and Physiology Branch Final Laboratory Report FY 2003-
4 30-4. 18 pp.
5 Barrett, G.W., and K.L. Stueck. 1976. Caloric ingestion rate and assimilation efficiency of the
6 short-tailed shrew, Blarina brevicauda. Ohio Journal of Science 76:25-26.
7 Birks, J.D.S., and N. Dunstone. 1985. Sex-related differences in the diet of the mink, Mustela
8 vison. Holarctic Ecology 8:245-252.
9 Birks, J.D., and I.J. Linn. 1982. Studies of home range of the feral mink, Mustela vison. Symp.
10 Zool. Soc. Lond. 49: 231-257.
11 Bjerke, D.L. and RE. Peterson. 1994. Reproductive toxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin
12 in male rats: different effects of in utero versus lactational exposure. Toxicol. Appl. Pharmacol.
13 127:241-249.
14 Bjerke, D.L., R.J. Sommer, R.W. Moore, and R.E. Peterson. 1994. Effects of in utero and
15 lactational 2,3,7,8-tetrachlorodibenzo-p-dioxin exposure on responsiveness of the male rat
16 reproductive system to testosterone stimulation in adulthood. Toxicol. Appl. Pharmacol. 127:250-
17 257.
18 Bleavins, M.R., R.J. Auerlich, and R.K. Ringer. 1980. Polychlorinated biphenyls (Aroclors 1016
19 and 1242): Effects on survival and reproduction in mink and ferrets. Archives of Environmental
20 Contamination and Toxicology 9:627-635.
21 Borga, K., G.W. Gabrielsen, and J.U. Skaare. 2001. Biomagnification of organochlorines along a
22 Barents Sea food chain. Environ. Pollut. 113(2): 187-198.
23 Brunstrom, B., H. Hakansson, and K. Lundberg. 1991. Effects of a technical PCB preparation and
24 fractions thereof on ethoxyresorufin O-deethylase activity, vitamin A levels and thyn development
25 in the mink (.Mustela vison). Annual Review of Pharmacology and Toxicology 69:421-26.
26 Buechner, H.K and F.B. Golley. 1967. Preliminary estimation of energy flow in Uganda kob. In:
27 Petrusiewicz, L. ed. Secondary Productivity of Terrestrial Ecosystems. Warszawa-Krakow; pp.
28 243-254.
29 Burgess, S.A. and J.R. Bider. 1980. Effects of stream habitat improvements on invertebrates,
30 trout populations, and mink activity. Journal of Wildlife Management 44(4):871-880.
31 Bursian, S.J., R.J. Aulerich, B. Yamini, and D.E. Tillitt. 2002. Dietary Exposure of Mink to Fish
32 from the Housatonic River: Effects on Reproduction and Survival. Michigan State University,
33 Department of Animal Science. Final Draft Report. March 28, 2002.
34 Bursian, S.J., and B. Yamini. 2003. Dietary Exposure of Mink to Fish from the Housatonic River.
35 Inducement of Manidibular and Maxillary Squamos Cell Profileration. Report submitted to Weston
36 Solutions, Inc. March 18, 2003.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-81
-------
1 Burt, W.H., and R.P. Grossenheider. 1976. A Field Guide to the Mammals. Field Marks of all
2 North American Species Found North of Mexico. Sponsored by the National Wildlife Federation
3 and National Audubon Society. Houghton Mifflin Company, Boston, MA. 290 pp.
4 Choate, J.R., J.K. Jones, Jr., and C. Jones. 1994.Handbook of Mammals of the South-Central
5 States. Louisiana State University Press, Baton Rouge, LA. 304 pp.
6 Corbet, R.L., D.G. Muir, and G.R.B. Webster. 1983. Fate of carbon-14 labeled 1,3,6,8-
7 tetrachloro-dibenzo-p-dioxin in an outdoor aquatic system. Chemosphere 12:523-528.
8 Cowan, W.F. and J.R. Reilly. 1973. Summer and fall foods of mink on the J. Clark Salyer
9 National Wildlife Refuge. Prairie Naturalist 5:20-24.
10 Eagle, T.C. and J. S. Whitman. 1987. Mink. In WildFurbearer Management and Conservation,
11 M. Novak, J. A. Baker, M.E. Obbarel, et al. eds. University of Pittsburgh Press, Pittsburgh, PA.
12 pp.615-624.
13 EPA (U.S. Environmental Protection Agency). 1993. Wildlife Exposure Factors Handbook
14 Volume 1 of 2. EPA/600/R-93/187a. U.S. Environmental Protection Agency, Office of Research
15 and Development, Washington, DC.
16 Frame, G. M. 2001. The current state-of-the-art of comprehensive, quantitative, congener-
17 specific PCB analysis, and what we now know about the distributions of individual congeners in
18 commercial Aroclor mixtures. In PCBs: Recent Advances in Environmental Toxicology and
19 Health Effects, L. W. Robertson and L.G. Hansen, Eds.
20 French, Jr., J.B., M.B. Voltura, and T.E. Tomasi. 2001. Effects of pre- and postnatal
21 poly chlorinated biphenyl exposure on metabolic rate and thyroid hormones of white-footed mice.
22 Environmental Toxicology and Chemistry 20:1704-1708.
23 Gerell, R. 1970. Home ranges and movements of the mink Mustela vison Schreber in southern
24 Sweden. Oikos 21(2): 160-173.
25 Giavini, E., M. Prati, and C. Vismara. 1983. Embryonic effects of 2,3,7,8-tetrachlorodibenzo-p-
26 dioxin administered to female rats before mating. Environ. Res. 31:105-110.
27 Gilbert, F. F. and E.G. Nancekivell. 1982. Food habits of mink (Mustela vison) and otter (Lutra
28 canadensis) in northeastern Alberta. Canadian Journal of Zoology 60:1282-1288.
29 Gray, L.E. Jr. and J.S. Ostby. 1995. In utero 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) alters
30 reproductive morphology and function in female rat offspring. Toxicol. Appl. Pharmacol. 133:285-
31 294.
32 Gray, L.E. Jr., W.R. Kelce, E. Monosson, J.S. Ostby, and L.S. Birnbaum. 1995. Exposure to TCDD
33 during development permanently alters reproductive function in male Long Evans rats and
34 hamsters: Reduced ejaculated and epididymal sperm numbers and sex accessory weights in
35 offspring with normal androgenic status. Toxicol. Appl. Pharmacol. 131:108-118.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-82
-------
1 Green, S., F.M. Sauro, and L. Friedman. 1975. Lack of dominant lethality in rats treated with
2 polychlorinated biphenyls (Aroclors 1242 and 1254). I'd. Cosmet. Toxicol. 13:507-510.
3 Greer, K.R. 1956. Fur Resources and Investigations: Study of the Otter Food Habits Along a
4 Segment of the Gallatin River. Montana Fish and Game Department, Job Comp. Rep. W-049-R-
5 06:35-59.
6 Greer, K.R. 1955. Yearly food habits of the river otter in the Thompson Lakes region,
7 northwestern Montana, as indicated by scat analysis. American Midland Naturalist 54:299-313.
8 Grodzinski, W., and B.A. Wunder. 1975. Ecological Energetics of Small Mammals. In Small
9 Mammals: Their Productivity and Population Dynamics. F.B. Golley, K. Petrusewicz, and L.
10 Ryszkowksi, Editors. Cambridge University Press, Cambridge, MA.
11 Haffner, G.D., M. Tomczak, and R. Lazar. 1994. Organic contaminant exposure in the Lake St.
12 Clair food web. Hydrobiologia 281:19-27.
13 Hamilton, W.J. Jr. 1940. The summer food of minks and raccoons on the Montezuma Marsh, New
14 York. Journal of Wildlife Management 4:80-84.
15 Hamilton, W.J. Jr. 1959. Foods of mink in New York. New York Fish and Game Journal
16 6(l):77-85.
17 Hamilton, W.J. Jr. 1961. Later fall, winter and early spring food of 141 otters from New York, NY.
18 Fish and Game Journal 8:106-109.
19 Heaton, S.N., S.J. Bursian, J.P. Giesy, D.E. Tillitt, J.A. Render, P.D. Jones, D.A. Verbrugge, T.J.
20 Kubiak, and R.J. Aulerich. 1995. Dietary exposure of mink to carp from Saginaw Bay,
21 Michigan. 1. Effects on reproduction and survival, and the potential risks to wild mink
22 populations. Archives of Environmental Contamination and Toxicology 28:334-343.
23 Hochstein, J.R., R.J. Aulerich, and S.J. Bursian. 1988. Acute toxicity of 2,3,7,8-
24 tetrachlorodibenzo-p-dioxin to mink. Archives of Environmental Contamination and Toxicology
25 17:33-37.
26 Hochstein, J.R., S.J. Bursian, and R.J. Aulerich. 1998. Effects of dietary exposure to 2,3,7,8-
27 tetrachlorodibenzo-p-dioxin in adult female mink (.Mustela vison). Archives of Environmental
28 Contamination and Toxicology 35:348-353.
29 Hochstein, J.R., J.A. Render, S.J. Bursian, and R.J. Aulerich. 2001. Chronic toxicity of dietary
30 2,3,7,8-tetrachlorodibenzo-p-dioxin to mink. Veterinary and Human Toxicology 43:134-39.
31 Hornshaw, T.C., R.J. Aulerich, and H.E. Johnson. 1983. Feeding Great Lakes fish to mink:
32 Effects on mink and accumulation and elimination of PCBs by mink. Journal of Toxicology and
33 Environmental Health 11:933-946.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-83
-------
1 Hornshaw, T.C., J. Safronoff, R.K. Ringer, and R.J. Aulerich. 1986. LC50 test results in
2 polychlorinated biphenyl-fed mink: Age, season, and diet comparisons. Archives of
3 Environmental Contamination and Toxicology 15:717-23.
4 Howard, P.H., R.S. Boethling, W.F. Jarvis, W.M. Meylan, and E.M. Michalenko. 1991.
5 Handbook of Environmental Degradation Rates. Lewis Publishers, Chelsea, MI.
6 Huuskonen, H., M. Unkila, R. Pohjanvirta, and J. Tuomisto. 1994. Developmental toxicity of
7 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in the most TCDD-resistant and -susceptible rat
8 strains. Toxicol. Appl. Pharmacol. 124:174-180.
9 Jensen, S., J.E. Kihlstrom, M. Olsson, C. Lundberg, and J. Orberg. 1977. Effects of PCB and
10 DDT on mink (.Mustela vison) during the reproductive season. Ambio 6:239.
11 Jorgensen, S.E., S.N. Nielsen, and L.A. Jorgensen. 1991. Handbook of Ecological Parameters
12 andEcotoxicology. Elsevier Science Publishers, Amsterdam, The Netherlands.
13 Khera, K.S. and J.A. Ruddick. 1973. Polychlorodibenzo-p-dioxins: Perinatal effects and the
14 dominant lethal test in Wistar rats. Toxicol. 120:70-84.
15 Kihlstrom, J.E., M. Olsson, S. Jensen, A. Johansson, and A. Bergman. 1992. Effects of PCB and
16 different fractions of PCB on the reproduction of the mink (.Mustela vison). Ambio 21:563-69.
17 Kimbrough, R.D., R.E. Linder, and T.B. Gaines. 1972. Morphological changes in livers of rats
18 fed polychlorinated biphenyls. Archives of Environmental Health 25:354-364.
19 Koplin, J.R., Collopy, M.W., and A.R. Bammann. 1980. Energetics of two wintering raptors.
20 Auk 97 : 795-806.
21 Kuehl, D.W., P.M. Cook, A.R. Batterman, and B.C. Butterworth. 1987. Isomer dependent
22 bioavailability of polychlorinated dibenzo-p-dioxins and dibenzofurans from municipal
23 incinerator fly ash to carp. Chemosphere 16:657-666.
24 Lagler, K.F., and B.T. Ostensen. 1942. Early spring food of the otter in Michigan. Journal of
25 Wildlife Management 6:244-254.
26 Lamprey, H.F. 1964. Estimation of the large mammal densities, biomass, and energy exchange
27 in the Tarangire Game Reserve and the Masai Steppe in Tanganyika. E. Afr. Wild. J. 2: 1-46.
28 Lariviere, S. 1999. Mustela vison. Mammalian Species 608:1-9.
29 Liers, E.E. 1951. Notes on the river otter {Lutra canadensis). Journal of Mammalogy 32:1-9.
30 Linn, I.J., and J.D. Birks. 1981. Observations on the home ranges of feral American mink
31 (.Mustela vison) in Devon, England, as revealed by radio-tracking. In Proceedings Worldwide
32 Furbearer Conference: Vol. 1. J.A. Chapman, and J.A. Pursley, Editors. Frostbury, MD. pp.
33 1088-1102.Lowery, G.H. 1974. The Mammals of Louisiana and Its Adjacent Waters. Louisiana
34 State University Press, Baton Rouge, LA. 565 pp.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-84
-------
1 Mably, T.A., R.W. Moore, and R.E. Peterson. 1992. In utero and lactational exposure of male
2 rats to 2,3,7,8-tetrachlorodibenzo-p-dioxin. 1. Effects on androgenic status. Toxicol. Appl.
3 Pharmacol. 144:97-107.
4 Melquist, W.E., and M.G. Hornocker. 1983. Ecology of river otters in west central Idaho.
5 Wildlife Monograph 83. 60 pp.
6 Melquist, W.E., J.S. Whitman, and M.G. Hornocker. 1981. Resource partitioning and coexistence of
7 sympatic mink and river otter populations. In Worldwide Furbearer Conference Proceedings, Vol.
8 I. J.A. Chapman and D. Pursley, Editors. Frostberg, MD.
9 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S.
10 Petron, B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-
11 of-Evidence Workgroup: A weight-of-evidence approach for evaluating ecological risks. Human
12 and Ecological Risk Assessment 2:277-304.
13 Minnich, J.E. 1982. The Use of Water. In Biology of the Reptilia. Vol. 12. C. Gans, F.H. Pough,
14 Editors. Academic Press, New York, NY.
15 Mitchell, J.L. 1961. Mink movements and populations on a Montana river. Journal of Wildlife
16 Management 25:48-54.
17 Moore, D.R.J., B.E. Sample, G.W. Suter, B.R. Parkhurst, and R.S. Teed. 1999. A probabilistic
18 risk assessment of the effects of methylmercury and PCBs on mink and kingfishers along East
19 Fork Poplar Creek, Oak Ridge, Tennessee, USA. Environmental Toxicology and Chemistry
20 18:2941-2953.
21 Muir, D.C.G., R.J. Nostrom, and M. Simon. 1988. Organochlorine contaminants in Arctic
22 marine food chains: Accumulation of specific polychlorinated biphenyls and chlordane-related
23 compounds. Environmental Science and Technology 22:1071-1079.
24 Murray, F.J., F.A. Smith, K.D. Nitschke, C.G. Humiston, R.J. Kociba, and B.A. Schwetz. 1979.
25 Three-generation reproduction study of rats given 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)
26 in the diet. Toxicology and Applied Pharmacology 50:241-252.
27 Nagy K.A. 1987. Field metabolic rate and food requirement scaling in mammals and birds.
2 8 Ecological Monographs 57:111-128.
29 Nagy, K.A., I. A. Girard, and T.K. Brown. 1999. Energetics of free-ranging mammals, reptiles,
3 0 and birds. Annual Reviews in Nutrition 19:247-77.
31 Nisbet, I.C.T., and M.B. Paxton. 1982. Statistical aspects of three-generation studies of the
32 reproductive toxicity of TCDD and 2,4,5-T. Journal American Statistical Association 36:290-
33 298.
34 NRCC (National Research Council of Canada). 1981. Polychlorinated dibenzo-p-dioxins:
35 Criteria for their effects on man and his environment. Publication NRCC No. 18574. National
36 Research Council of Canada, Ottawa, Ontario. 251 p.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-85
-------
1 Palmer, E.L. and H.S. Fowler. 1975. Fieldbook of Natural History. McGraw-Hill Book
2 Company, New York, NY.
3 Platanow, N.S., and L.H. Karstad. 1973. Dietary effects of polychlorinated biphenyls on mink.
4 Canadian Journal of Comparative Medicine 37:391-400.
5 Pohjanvirta, R., M. Unkila, and J. Tuomisto. 1993. Comparative acute lethality of 2,3,7,8-
6 tetrachlorodibenzo-p-dioxin (TCDD), 1,2,3,7,8-pentachlorodibenzo-p-dioxin and 1,2,3,4,7,8-
7 hexachlorodibenzo-p-dioxin in the most TCDD susceptible and the most TCDD-resistant rat
8 strain. Pharmacol, and Toxicol. 73:52-56.
9 Proulx, G., J.A. McDonnell, and F.F. Gilbert. 1987. The effect of water level fluctuations on
10 muskrat, Ondatra zibethicus, predation by mink, Mustela vison. Canadian Field-Naturalist
11 101 (1): 89-92.
12 Restum, J.C., S.J. Bursian, J.P. Giesy, J.A. Render, W.G. Helferich, E.B. Shipp, D.A. Verbrugge,
13 and R.J. Aulerich. 1998. Multigenerational study of the effects of consumption of PCB-
14 contaminated carp from Saginaw Bay, Lake Huron, on mink. 1. Effects on mink reproduction,
15 kit growth and survival, and selected biological parameters. Journal of Toxicology and
16 Environmental Health 54:343-375.
17 Ringer, R.K., R.J. Aulerich, and M. Zabik. 1972. Effects of dietary polychlorinated biphenyls on
18 growth and reproduction in mink. Proceedings of the 164th National Meeting, American Chemical
19 Society 12:149-154.
20 Safe, S.H. 1994. Polychlorinated biphenyls (PCBs): Environmental impact, biochemical and toxic
21 responses, and implications risk assessment. Critical Reviews in Toxicology 24:87-149.
22 Senthilkumar, K., N. Iseki, S. Hayama, J. Nakanishi, and S. Masunaga. 2001. Polychlorinated
23 dibenzo-p-dioxins, dibenzofurans, and dioxin-like polychlorinated biphenyls in livers of birds
24 from Japan. Arch. Environ. Contam. Toxicol. 42:244-255.
25 Sheldon, W.G., and W.G. Toll. 1964. Feeding habits of the river otter in a reservoir in central
26 Massachusetts. Journal of Mammalogy 45:449-455.
27 Stahl, B.U., A. Kettrup, and K. Rozman. 1992. Comparative toxicity of four chlorinated dibenzo-p-
28 dioxins (CDDs) and their mixture. Arch. Toxicol. 66:471 -477.
29 Thayer, G.W., W.E. Schaaf, J.W. Angelovic, and M.W. LaCroix. 1973. Caloric measurements of
30 some estuarine organisms. Fishery Bulletin 71:289-296.
31 Tillitt, D. E., S. R. Birke, and D. K. Nicks. 2003. Hepatic O-dealkylase activities from induction
32 of cytochrome P-450 in mink and their offspring fed diets containing contaminated fish from the
33 Housatonic River. USGS, Columbia Environmental Research Center, Biochemistry and
34 Physiology Branch, Final Laboratory Report FY 2003 - 30 - 4, 2003. 16 p.
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-86
-------
1 Ushinohama, K., D-S. Son, K.F. Roby, K.K. Rozman, and P.F. Terranova. 2001. Impaired
2 ovulation by 2,3,7,8 tetrachlorodibenzo-p-dioxin (TCDD) in immature rats treated with equine
3 chorionic gonadotropin. Reproductive Toxicology 15:275-280.
4 Van den Berg, M., L. Birnbaum, A.T.C. Bosveld, B. Brunstrom, P. Cook, M. Feeley, J.P. Giesy,
5 A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C. Larsen, F.X.R. van Leeuwen, A.K.D.
6 Liem, C. Nolt, R.E. Petersen, L. Poellinger, S. Safe, D. Scenk, D. Tillitt, M. Tysklind, M.
7 Younes, F. Waern, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs) for PCBs,
8 PCDDs, PCDFs for humans and wildlife. Environmental Health Perspectives 106:775-792.
9 Whitaker, J.O., Jr., and W.J. Hamilton Jr. 1998. Mammals of the Eastern United States. 3rd
10 edition, Cornell University Press, Ithaca, NY, USA.
11 Wren, C.D., D.B. Hunter, J.F. Leatherland, and P.M. Stokes. 1987b. The effects of
12 polychlorinated biphenyls and methylmercury, singly and in combination on mink. II.
13 Reproduction and kit development. Archives of Environmental Contamination and Toxicology
14 16:449-454.
15
MK01 |O:\20123001.096\ERA_PB\ERA_PB_9.DOC
9-87
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
10. ASSESSMENT ENDPOINT—SURVIVAL, GROWTH, AND
REPRODUCTION OF OMNIVOROUS AND CARNIVOROUS
MAMMALS
Highlights
Conceptual Model
The assessment endpoint is the survival, growth, and reproduction of omnivorous
and carnivorous mammals in the Housatonic River PSA. Common omnivorous and
carnivorous mammals, including red fox and northern short-tailed shrew, are
exposed to tPCBs and TEQ via trophic transfer. These two species were selected as
representative species for the ecological risk assessment (ERA).
Exposure
Exposure of the representative species to tPCBs and TEQ was determined from
concentrations found in prey items and an estimation of the daily intake of these
contaminants of concern (COCs) from consumption of prey.
Effects
No data were available on toxicity of tPCBs and TEQ to red fox and northern short-
tailed shrew. Surrogate species were used to estimate effects to these species.
Sufficient surrogate data were available to generate dose-response curves for each
species.
Risk
Modeled exposure and effects for red fox and short-tailed shrew suggest that they
are at an intermediate risk as a result of exposure to tPCBs and TEQ in the
Housatonic PSA. Other omnivorous and carnivorous mammal species common to
the PSA are expected to have either higher levels of risk (e.g., smoky shrew), similar
levels of risk (e.g., masked shrew, gray fox), and in one case (coyote), a lower level
of risk compared to the representative species.
10.1 INTRODUCTION
The purpose of this section is to characterize and quantify the current and potential risks posed to
omnivorous and carnivorous mammals exposed to contaminants of potential concern (COPCs) in
the Housatonic River and floodplain, focusing on total PCBs (tPCBs) and other COPCs
originating from the General Electric Company (GE) facility in Pittsfield, MA. The watershed is
located in western Massachusetts and Connecticut, discharging to Long Island Sound, with the
GE facility located near the headwaters of the watershed. The Primary Study Area (PSA)
includes the river and 10-year floodplain from the confluence of the East and West Branches of
the Housatonic River downstream of the GE facility, to Woods Pond Dam (Figure 1.1-2).
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC -> rx -> 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
A Pre-ERA was conducted to narrow the scope of the ecological risk assessment by identifying
contaminants, other than tPCBs, that pose potential risks to aquatic biota and wildlife in the PSA
(Appendix B). A three-tiered deterministic approach was used to screen COPCs. The
deterministic assessments compared potential conservative estimates of exposure with
conservative adverse effects benchmarks to identify which contaminants are of potential concern
to omnivorous and carnivorous mammals in the Housatonic River. A hazard quotient (total daily
intake/effect benchmark) greater than 1 in the Housatonic River area resulted in the COPC being
screened through to the next tier assessment, and to the probabilistic ecological risk assessment,
if necessary. In the COPC screening specific to this endpoint, several other COPCs (primarily
organochlorine pesticides) were screened out because their actual concentrations in the PSA
were likely much lower than the measured values due to laboratory interference (see Section
2.4). In summary, the COPCs that screened through to the probabilistic risk assessment for
omnivorous and carnivorous mammals were the contaminants of concern (COCs), tPCBs and
2,3,7,8-TCDD TEQ (TEQ). Total PCBs detected in Housatonic River media samples closely
resemble the commercial PCB mixtures Aroclor 1260 and Aroclor 1254, which are similar in
congener makeup. TEQ is calculated from coplanar PCB and dioxin and furan congeners using
the toxic equivalency factor (TEF) approach developed by Van den Berg et al. (1998)(see
Section 6.4).
A step-wise approach was used to assess the risks of tPCBs and TEQ to omnivorous and
carnivorous mammals in the Housatonic River watershed. The four main steps in this process
include the following:
1. Derivation of a conceptual model (Figure 10.1-1).
2. Assessment of exposure of mammals to COCs (Figure 10.1-2).
3. Assessment of the effects of COCs on mammals (Figure 10.1-3).
4. Characterization of risks to the omnivorous and carnivorous mammalian species
(Figure 10.1-4).
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-2
-------
1 This section is organized as follows:
2 ¦ Section 10.2 presents the conceptual model for assessing the ecological risk to
3 omnivorous and carnivorous mammals.
4 ¦ Section 10.3 describes the exposure model, input variables, and uncertainty
5 propagation techniques. Also presented in this section are the exposure modeling
6 results for red fox and northern short-tailed shrew.
7 ¦ Section 10.4 describes the effects to mammals exposed to tPCBs and TEQ and
8 derives the effects metrics.
9 ¦ Section 10.5 describes the lines of evidence, followed by a discussion of the sources
10 of uncertainty in this assessment, and the conclusions regarding risks of tPCBs and
11 TEQ to omnivorous and carnivorous mammals in the Housatonic River PSA.
12
13 This section provides a summary of the ERA for omnivorous and carnivorous mammals,
14 which is presented in detail in Appendix J.
15
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-3
-------
(0
0
»
re
0
v
01
(0
0
o
3
O
(0
Historic releases of PCBs from the
Pittsfield GE Facility and
surrounding disposal areas
r 1
r
Legend
^
Direct uptake
=~
Trophic transfer
—~
Both
Potential effects
Wetland and
surface water
discharge
Floodplain
runoff
Groundwater
discharge
Contaminated soil, sediment, water and biota
2
o
+¦»
Q.
V
O
CH
Terrestrial Vegetation
Plants, fungi
Terrestrial Invertebrates
Earthworms, snails,
insects
T
Mammals
Shrews,
white-footed
mice, rabbits
WF1
Birds
Omnivorous Mammals
Northern short-tailed shrew
Carnivorous Mammals
Red fox
LU
Decreased Survival, Growth, or Reproduction
Figure 10.1-1 Conceptual Model Diagram: Exposure Pathways for Omnivorous
and Carnivorous Mammals Exposed to Contaminants of Concern (COCs) in the
Housatonic River PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-4
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
EXPOSURE
Figure 10.1-2 Framework Used to Model Exposure of Wildlife Species to
Contaminants of Concern (COCs) in the Housatonic River PSA
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia r 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
EFFECTS
Compile Toxicity Data from
the Literature
Evaluate Against
Acceptability Criteria for
Effects Studies in the
Literature
Select Effects Data
Derive Effects Metric
Figure 10.1-3 Approach Used to Model Effects of Contaminants of Concern
(COCs) to Representative Species in the Housatonic River PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-6
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
RISK CHARACTERIZATION
Figure 10.1-4 Overview of Approach Used to Characterize the Risks of
Contaminants of Concern (COCs) to Omnivorous and Carnivorous Mammals in
the Housatonic River PSA
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia q 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
10.2 CONCEPTUAL MODEL
The conceptual model presented in Figure 10.1-1 illustrates the exposure pathways for
omnivorous and carnivorous mammals exposed to tPCBs and TEQ in the PSA. Total PCBs and
TEQ are persistent, hydrophobic, and lipophilic. Therefore, they are bioaccumulated by aquatic
and terrestrial biota directly through the consumption of contaminated prey as part of the food
chain (Haffner et al. 1994; Senthilkumar et al. 2001; Borga et al. 2001). Small mammals,
earthworms and other invertebrates, and plants comprise the major dietary items for omnivorous
mammals. Carnivorous mammals primarily feed on mammals, although fruits, birds, and
invertebrates can supplement their diet. In summary, omnivorous and carnivorous mammals that
reside, or partially reside, within the study area are exposed to tPCBs and TEQ principally
through diet as a result of trophic transfer. Other routes of exposure, considered to be less
important to overall exposure, include inhalation, water consumption, and soil/sediment
ingestion (Moore et al. 1999).
The problem formulation (see Section 2) identified the red fox (Vulpes vulpes) (Figure 10.2-1) as
the representative species for carnivorous mammals potentially exposed to tPCBs and TEQ from
consumption of contaminated prey. The northern short-tailed shrew (Blarina brevicauda)
(Figure 10.2-2) was selected as the representative species for omnivorous mammals. Life history
profiles for the red fox and short-tailed shrew are provided in the text boxes. Additional life
history information on these species is presented in Section J.2.1.
The assessment endpoint is the survival, growth, and reproduction of omnivorous and
carnivorous mammals in the Housatonic River PSA. The measurement endpoints used to
evaluate the assessment endpoint include: (1) determining, by comparisons to doses reported in
the literature to cause adverse effects, the extent to which the concentrations of tPCBs and TEQ
ingested in the diet will cause adverse effects to the survival, growth, or reproduction of
omnivorous and carnivorous mammals, and (2) determining, by conducting field surveys, the
relationship between the concentrations of tPCBs and TEQ and survival, reproduction, and
relative abundance of omnivorous and carnivorous mammals in the Housatonic River floodplain.
As part of the EPA field survey, placental scars in small mammals were analyzed as an
indication of past reproductive performance.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC 10 8 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Figure 10.2-1 Red Fox (Vulpes vulpes)
Life History of Red Fox
The red fox is a common dog-sized canine that occurs in many habitats throughout
its range and is the most widely distributed carnivore in the world. In North America,
the red fox is found throughout the United States and Canada, but not in the
southeast coastal region, extreme southwest, parts of the central states, or the
Pacific coastal regions. The typical pelage color of fox is red and it can be identified
by its characteristic bushy, white-tipped tail, pointed muzzle, and prominent ears.
Habitat - Occupies a variety of habitats, but preferred habitat is a matrix of forest,
cropland, and pastureland, habitats common in the PSA. The availability of suitable
prey as well as suitable den sites is also important. Prefer to locate dens in forested
areas, but within a short distance of open areas and usually within 100 meters of a
source of open water.
Home Range - Maintains territory throughout the year and is considered
nonmigratory. Average home range for adults in Maine was 14.7 km2 (range = 6.0-
27.5 km2), average home range in Ontario was 9 km2 with a range of 5 to 20 km2.
Mean territory sizes reported in EPA (1993) ranged from 100 to 2,000 hectares (1 to
20 km2). Adults traverse most of their territory on a routine basis, but focus activities
around dens, preferred hunting areas, food supplies, and resting areas.
Dietary Habit - Diet varies throughout the year depending on food availability.
Includes almost all available animals as prey, such as insects, fish, reptiles,
amphibians, birds, small mammals, and carrion. Although typically identified as
carnivores, can consume considerable amounts of plant materials, particularly in the
summer and fall. Plant material in the diet includes berries, apples, and nuts.
MK0110:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-9
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Figure 10.2-2 Northern Short-Tailed Shrew (Blarina brevicauda)
Life History of Short-Tailed Shrew
The northern short-tailed shrew is a small energetic mouse-like animal with dark
slate-colored pelage found throughout the northcentral and northeastern United
States extending into southern Canada. It is easily identified as a shrew by its long
pointed snout, small black eyes, concealed ears, and five toes on each foot. The
northern short-tailed shrew has a short tail, which is approximately 20% of total
animal length.
Habitat - Can be found in a variety of habitats, including wetlands and uplands, and
are common in areas with abundant vegetative cover, occur in both forested and
open habitats.
Home Range - Home range of 0.06 acre (0.024 ha) in central New York State.
Other estimates of home range size vary from 0.25 to 0.5 acres (0.1 to 0.2 ha) in
areas of low prey density in winter months during nonbreeding periods to 0.07 to
0.17 acres (0.03 to 0.07 ha) in areas of high prey density with a minimum of territory
overlap. Do not migrate seasonally, remaining in home range.
Dietary Habits - Earthworms and insects comprise most of the diet, earthworms
reported to be the most important item in the diet. Invertebrates in the diet are
mainly obtained from the leaf litter layer, and consist of millipedes, insect larvae,
spiders, slugs, snails, and other mollusks. Plant materials, including nuts, berries,
roots, and fungi, and occasional small mammals are also a component of the diet.
MK0110:\20123001,096\ERA_PB\ERA_PB_10.DOC j q j q
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
10.3 EXPOSURE ASSESSMENT
This exposure assessment for omnivorous and carnivorous mammals focuses on the PSA. The
representative species for omnivorous and carnivorous mammals are the northern short-tailed
shrew and the red fox. These mammals occur in the PSA and feed on prey exposed directly to
tPCBs and TEQ and through trophic transfer. Trophic transfer and exposure through ingestion
of contaminated prey are the major exposure pathways for omnivorous and carnivorous
mammals exposed to tPCBs and TEQ. Other routes of exposure, considered to be negligible
contributors to overall exposure, include inhalation, water consumption, and soil/sediment
ingestion (Moore et al. 1999). Total PCBs and TEQ tend to bioaccumulate in the food chain
because of the following:
¦ Total PCBs and TEQ are persistent, and hydrophobic and highly lipophilic
substances.
¦ When released to aquatic systems, the majority of these compounds form associations
with dissolved and/or particulate matter in the water column and remain in sediment
layers; biodegradation is considered to be a relatively minor fate process in water
(NRCC 1981; Howard etal. 1991).
¦ Aquatic sediment provides a sink for these compounds and may represent long-term
sources to the aquatic food web (Kuehl et al. 1987; Muir 1988; Corbet et al. 1983;
Tsushimoto et al. 1982). Both of these COCs are bioaccumulated by aquatic and
terrestrial biota directly through the consumption of contaminated prey as part of the
food chain (Haffner et al. 1994; Senthilkumar et al. 2001; Borga et al. 2001).
Foxes were observed throughout the PSA from 1998 to 2001 (Appendix A). The exposure
analysis was carried out for all of Reach 5 of the PSA because the foraging range of red fox is
fairly large. Because short-tailed shrews have a much smaller foraging range, the exposure
analysis was performed for three locations in the PSA (Locations 13, 14, and 15, Figure J.2-3)
that represent the range of COC concentrations found in the PSA.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-11
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Description of Sampling Locations 13, 14, and 15
Location 13 is a relatively flat area on the west shore of the river, in the floodplain adjacent to river mile
133.2, situated at an elevation of 965 ft (294 m). The community type is transitional floodplain forest that
is flooded seasonally and is moderately well drained, with extensive vegetation cover (80%) and alluvial
silt-loam soil. PCB concentrations in floodplain soil averaged 55.2 mg/kg dw.
Location 14 is a relatively flat low-lying area on the west shore of the river, in the floodplain adjacent to
river mile 129.9, situated at an elevation of 965 ft (294 m). The community type is transitional floodplain
forest that is flooded seasonally, with extensive vegetation cover (70%) and fluvial silt soil. PCB
concentrations in floodplain soil averaged 26.1 mg/kg dw.
Location 15 is a flat area on the west shore of the river, in the floodplain adjacent to river mile 126.7,
situated at an elevation of 965 ft (294 m). Community types are circumneutral hardwood swamp and
transitional floodplain forest that are flooded seasonally. This site has 60% vegetation cover, 40% leaf
litter cover, and a primarily mineral soil. PCB concentrations in floodplain soil averaged 0.484 mg/kg dw.
This section begins with a description of the exposure model used for both of the representative
species. Subsequent sections describe the inputs used in the exposure analyses for each
representative species. The section concludes with a presentation of the results of the exposure
analyses.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-12
-------
1 10.3.1 Exposure Model
2 Exposure of the representative species, red fox and northern short-tailed shrew, to tPCBs and
3 TEQ was estimated using a total daily intake model adapted from the Wildlife Exposure Factors
4 Handbook (EPA 1993) and related publications. The model used in the exposure analysis was:
5 TDI = FT ¦ FIR tCrP< (Eq 1)
2=1
6 where
7
TDI
= total daily intake (mg/kg bw/d tPCBs, ng/kg bw/d TEQ)
8
FIR
= normalized food intake rate (kg/kg bw/d)
9
FT
= foraging time in PSA (unitless)
10
c,
= concentration in z'th food item (mg/kg tPCBs, ng/kg TEQ)
11
P,
= proportion of the z'th food item in the diet (unitless)
12 The models consider the food intake rates of the representative species (FIR), the concentrations
13 of COCs in each food item (Ci), and the proportion of the diet accounted for by that food item
14 (Pj). For those input variables that are uncertain, variable, or both, distributions are used rather
15 than point estimates. Monte Carlo and probability bounds analyses are the methods used to
16 propagate uncertainties about input variables through the exposure model for each COC. A
17 description of these techniques and the methods used to parameterize input variables is presented
18 in Section 6.5 and Appendix C.4. The results of the Monte Carlo analysis are used to estimate
19 the probability of exposure exceeding an effects threshold or doses that cause adverse effects of
20 differing magnitudes. The probability bounds analysis is conducted to determine how
21 uncertainty regarding the distributions of the input variables influences the estimated exposure
22 distribution. The results of these analyses are discussed in detail in Appendix J.
23 Two issues often arise when calculating a TEQ concentration in prey:
24 ¦ Congener concentrations may be below the detection limit (DL) (i.e., non-detects).
25 ¦ Some congeners may not be resolved due to co-elution during analysis.
26
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
An approach was developed to address these issues and is presented in Section 6.4 and Appendix
C.2. Briefly, congeners detected at or below the DL were included in the TEQ calculations by
investigating three options:
¦ First, setting the concentration for the congener equal to zero (0),
¦ Setting it to half the DL, and,
¦ Finally, setting it equal to the DL.
A comparison of the results of this bounding analysis provides a description of the uncertainty
surrounding the TEQ value due to concentrations of one or more congeners being below the
detection limit.
To resolve the co-elution issue, the concentration of congeners that co-eluted with other
congeners was assumed equal to the total concentration of the co-elutes (overestimate of TEQ
concentration) or zero (underestimate of TEQ concentration). The decision criteria in Section
6.4 were followed to deal with the uncertainty arising from co-elution or non-detection of
congeners when estimating exposure point concentrations for use in the exposure analyses.
Input distributions to the exposure analyses were generally assigned as follows:
¦ Lognormal distributions for variables that were right skewed with a lower bound of
zero and no upper bound (e.g., amount of COC transferred from mother to offspring
via egg tissue for tree swallows).
¦ Beta distributions for variables bounded by zero and one (e.g., proportion of a prey
item in the diet).
¦ Normal distributions for variables that were symmetric and not bounded by one (e.g.,
body weight).
¦ Point estimates for minor variables or variables with low coefficients of variation.
In certain situations (e.g., poor fit of data), other distributions were fit to the data or other
approaches were used. To quantify uncertainty, two approaches were used as described in
Section 6.5.2 and Appendix C.4, Monte Carlo analyses and Probability Bounds analyses. The
distributions used in the exposure analyses for red fox and northern short-tailed shrews are
shown in Figures 10.3-1 and 10.3-2. A brief description of these variables is provided below. A
discussion of the concentrations of COCs in prey follows.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-14
-------
1
2
3
FMR Slope Term (a)
FMR Power Term (b )
0.8
5
0.6
¦§ 0.4
0.2
0.5
1.5
Body Weight
Proportion of Mammals in Diet
3000 3500 4000 4500
Weight (g)
5000
-§ 0.4
-
0.2
0
0.4
0.6 0.8
Proportion
Figure 10.3-1 Input Distributions for the Exposure Modeling of the Red Fox
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-15
-------
FMR Slope Term (a)
1
0.8
i 06
15
C5
-§ 0.4
Ph
0.2
0
a
0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Log (a)
0.8
g
^ 0.6
15
es
¦§ 0.4
Ph
0.2
0.2
FMR Power Term (b)
A
0.4
0.6
0.8
Body Weight
Proportion of Litter Inverte brates in Diet
20 25
Weight (g)
0.8
?
2 06
15
cs
¦§ 0.4
Ph
0.2
0
JflULu i Eel.
0.2 0.4 0.6 OA
Proportion
Proportion of Earthworms in Diet
Proportion of Mammals in Diet
0.8
g
& 0.6
15
cs
-§ 0.4
U.
Ph
0.2
y\
¦ \
o
0.05 0.1 0.15 0.2 0.25 0.3
Proportion
0.8
2 °-6
15
cs
¦§ 0.4
U.
Ph
0.2
0.05 0.1
Proportion
0.15
0.2
Figure 10.3-2 Input Distributions for the Exposure Modeling of Short-Tailed
Shrew
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-16
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
fO.3.-/.-/ Body Weight (BW)
As with many mammalian species, the red fox exhibits sexual dimorphism in body size. Males
are typically 10% larger and 20% to 30% heavier than females (Storm et al. 1976; Lariviere and
Pasitschniak-Arts 1996; Voigt 1987). In a study conducted in Indiana, males weighed an
average of 4.9 kg and females weighed an average of 4.0 kg (Lariviere and Pasitschniak-Arts
1996). Voigt (1987) found that male red fox in Ontario averaged 4.1 kg in weight (n = 37) and
females averaged 3.4 kg in weight (n = 37).
The northern short-tailed shrew can weigh over 22 grams (George et al. 1986; Burt and
Grossenheider 1980, as cited in EPA 1993). Whitaker and Hamilton (1998) reported a mean
body weight for adult males and females of 19.3 g. As part of the ecological characterization of
the PSA (Appendix A), 58 adult short-tailed shrews of both sexes were caught during small
mammal trapping in 1998 to 2001. The body weights ranged from 15 to 27 g (mean = 21.9 g).
The average weight of adult female shrews was 22.3 g (SD = 2.87 g).
10.3.1.2 Food Intake Rate (FIR)
The food intake rate of red fox and northern short-tailed shrew were measured in laboratory and
captive animals (Sargeant 1978; Barrett and Stuek 1976; Morrison et al. 1957). Because the
animals were captive or kept in a laboratory, the measured food intake rates likely
underestimated food intake rates of free-living fox and shrew (EPA 1993). Free-living fox and
shrew, unlike captive fox and shrew, expend energy foraging for prey, avoiding predators,
defending territories, etc. As a result, an allometric modeling approach, described below, was
used to estimate food intake rate for red fox and short-tailed shrew.
Nagy (1987) and Nagy et al. (1999) derived allometric equations for estimating the free
metabolic rate (FMR) of free-living mammals in kilojoules (kJ) per day using the following
general equation:
FMR {kJ / d) = a- BW (g)b (Eq. 2)
The slope (a) and power (b) distributions were based on the error statistics derived from
regression analysis of the data reported in Nagy et al. (1999). For red fox, the carnivore equation
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC 1^17 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
was used and had a mean slope term log a equal to 1.67 and a standard error of 2.65 in logio
units. The power term (b) had a reported mean of 0.869 and a standard error of 0.116 (Nagy et
al. 1999). For short-tailed shrew, the insectivore equation was used. The slope term log a had a
reported mean of 6.98 and a standard error of 1.32 in logio units, and the power term (b) had a
reported mean of 0.622 and a standard error of 0.0630 (Nagy et al. 1999). The body weight
(BW) distribution was described above. The results of the calculation were then converted to
kcal/kg bw/d.
Food intake rate (FIR) is derived from FMR using the following equation:
FMR
FIR(kg/kgbw/d)= (Eq. 3)
2>UvG£:
i-1
where AEt is the assimilation efficiency of z'th food item (unitless) and GEt is the gross energy of
z'th food item (kcal/kg).
The gross energies of various wildlife food sources are summarized in the Wildlife Exposure
Factors Handbook (EPA 1993). The gross energy of earthworms ranges from 780 to 830
kcal/kg (mean = 805; SD = 141) (Cummins and Wuycheck 1971; Thayer et al. 1973). The mean
gross energy for grasshoppers and crickets is 1,700 kcal/kg (SD = 260) (Cummins and
Wuycheck 1971; Collopy 1975; Bell 1990), and for adult beetles, the mean is 1,500 kcal/kg
(Cummins and Wuycheck 1971; Collopy 1975; Bell 1990). Grasshoppers, crickets, and beetles
were used as representatives of litter invertebrates; their mean gross energy is 1,600 kcal/kg.
Mammals have a gross energy of 1,700 kcal/kg (SD = 280) (Koplin et al. 1980).
The assimilation efficiency for mammals consumed by mammals is 84% (SD = 6.5%) (Castro
etal. 1989). The assimilation efficiency of earthworms consumed by mammals is not known.
The mean assimilation efficiency for insects consumed by small mammals is 87% (Bryant and
Bryant 1988). This value was used to represent the assimilation efficiency for earthworms
consumed by mammals. Point estimates were used for these variables in the Monte Carlo and
probability bounds analyses because of their relatively small coefficients of variation (i.e.,
CV<10%). As a result, these input variables are not included in Figures 10.3-1 and 10.3-2.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-18
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
10.3.1.3 Proportions of Dietary items (P,)
The red fox can occupy a variety of habitats in the PSA and may use a variety of food sources.
Available studies reporting the dietary composition of the red fox in North America show that
the proportion of dietary items varies according to season (Table J.2-1). Most studies found that
mammals constitute the majority of the diet of the red fox, with the percentage in the diet as high
as 92% in the spring (Knable 1974). For this assessment, mammals represent approximately
76% of the average diet for all seasons (Figure 10.3-1). However, the distributions used in the
exposure analyses for mammals were sufficiently wide to incorporate the range of variation
reported in the literature (Appendix J). Other food items including birds, invertebrates, and
vegetation were not included as part of the exposure model because the dietary items represent a
relatively small portion of the diet (e.g., birds and invertebrates) or the contribution to overall
exposure is negligible (e.g., vegetation).
As with the red fox, there is variation in the proportion of dietary items reported for the short-
tailed shrew (Table J.2-9). Earthworms comprised between 5% and 31% of the diet of short-
tailed shrew, whereas insects and small mammals were reported as high as 61% and 24% of the
diet, respectively (Hamilton 1941, as cited in EPA 1993; Linzey and Linzey 1973; Eadie 1944).
Averaging the available data for the winter and summer diets of the short-tailed shrew indicates
that earthworms, litter invertebrates (all combined), and small mammals comprise the major
dietary items for the short-tailed shrew. For this exposure assessment, the diet of the shrew was
on average 19% for earthworms, 60% for litter invertebrates, and 12% for small mammals
(Figure 10.3-2). However, the distributions used in the exposure analyses for each dietary
component were sufficiently wide to incorporate the range of variation reported in the literature
(Appendix J).
10.3.1.4 Foraging Time (FT)
The red fox visits all parts of its territory on a regular basis (Abies 1974). The home range used
in this assessment was 9 km2 (Voigt and Tinline 1980) and the width of the floodplain for Reach
5 ranges from 200 to 600 m. Therefore, the red fox is expected to spatially and temporally
average exposure inside and outside the PSA within its home range, potentially experiencing
areas of high contamination along with areas of low or no contamination. As a result, the PSA
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia iq 7/10/2003
-------
1 represents only a portion of the home range of red fox. It was estimated that red fox spend up to
2 50% of their time foraging in the PSA (see Appendix J.2).
3 The foraging range for northern short-tailed shrew is small. Whitaker and Hamilton (1998) and
4 Degraaf and Yamasaki (2001) found that the northern short-tailed shrew had a home range size
5 of 0.06 acre (0.024 ha), while Piatt (1976) reported that the home range size varies from 0.25 to
6 0.5 acres (0.1 to 0.2 ha) in areas of low prey density. The sizes of Locations 13, 14, and 15 are
7 approximately 2 to 3 acres. Therefore, shrews are expected to have 100% of their foraging range
8 within each of Locations 13, 14, and 15 in the PSA.
9 10.3.1.5 Concentrations of COCs in Prey
10 Mammals such as white-footed mouse and short-tailed shrew are the major dietary items for red
11 fox. The median concentration of tPCBs in mammals measured in Reach 5 is 4.98 mg/kg. The
12 25th and 75th percentiles are 1.78 and 29.3 mg/kg, respectively. The median, 25th and 75th
13 percentile concentrations of TEQ are 290, 179, and 1,107 ng/kg, respectively.
14 The diet for northern short-tailed shrew includes earthworms, litter invertebrates, and mammals.
15 Similar statistics for concentrations of tPCBs and TEQ in these prey at Locations 13, 14, and 15
16 are presented in Figures 10.3-3 and 10.3-4, respectively. TEQ concentration in earthworms was
17 measured in one composite sample of 20 to 45 earthworms at each location. Data on
18 concentrations of TEQ in litter invertebrate prey were not available to estimate exposure to short-
19 tailed shrew. In this case, the concentrations of TEQ in prey were extrapolated using measured
20 concentrations in earthworms. Concentrations of tPCBs and TEQ in all prey items were highest
21 at Locations 13 and 14. At these locations, earthworms had the highest concentrations of tPCBs,
22 whereas concentrations of TEQ were highest in mammals.
23
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-20
-------
100
10
0.1
H
¦«&>
3*
/•
J?
<
~ Median
X Arithmetic Mean
I 75 Percentile
25th Percentile
<
/•
Prey item and location
Note: Error bars indicate interquartile range.
Figure 10.3-3 Concentrations of tPCBs in Prey of Northern Short-Tailed Shrew
(n=1 for invertebrates and earthworms)
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-21
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
O
C*
ex
£
G
O
H
10000
1000
100
10
0.1
El
~ Median
X Arithmetic Mean
75 Percentile
25th Percentile
t
&
4?
3*
•s?>
V5
r
&
&
A®
Prey item and location
Note: Error bars indicate interquartile range.
Figure 10.3-4 Concentrations of TEQ in Prey of Northern Short-Tailed Shrew (n=1
for invertebrates and earthworms)
The input variables for concentrations of COCs in prey of red fox and short-tailed shrew are
shown in Tables J.2-12 and J.2-13.
10.3.2 Results of Exposure Assessments
Figures 10.3-5 to 10.3-12 present exposure distributions for red fox and short-tailed shrew to
tPCBs and TEQ.
Figure 10.3-5 depicts the cumulative distribution of tPCB intake rates for red fox in Reach 5.
The Monte Carlo analysis indicated that exposure of red fox to tPCBs could range from a
minimum of 0.0220 to a maximum of 82.5 mg/kg bw/d. The mean exposure was 6.25 mg/kg
bw/d and the median exposure 2.68 mg/kg bw/d. Ninety percent of the exposure estimates were
between 0.321 and 25.0 mg/kg bw/d.
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia rsrs 7/10/2003
-------
Reach 5
a>
-Q
at
-Q
o
CLh
a>
u
s
«
¦o
a>
a>
u
x
100
80
60
40
20
0
Monte Carlo
LPB
UPB
20 40
TDI (mg/kg bw/d)
60
Notes: LPB = Lower Probability Bound
UPB = Upper Probability Bound
Figure 10.3-5 Exceedance Probability Distribution for Red Fox Exposed to tPCBs
in Reach 5 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-23
-------
Reach 5
a>
-Q
at
-Q
o
CLh
a>
u
s
«
¦o
a>
a>
u
x
1
2
100
80
60
40
20
0
Monte Carlo
LPB
UPB
0
20 40
TDI (mg/kg bw/d)
Notes: LPB = Lower Probability Bound
UPB = Upper Probability Bound
60
3 Figure 10.3-6 Exceedance Probability Distribution for Red Fox Exposed to TEQ in
4 Reach 5 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-24
-------
Location 13
xi
«
xi
o
—
-
a>
u
=
¦a
a>
a>
u
M
W
100
80
60
40
20
-* -
Monte Carlo
LPB
UPB
*
0.01 0.1 1 10
TDI (mg/kg bw/d)
100
1000
Notes: LPB = Lower Probability Bound
UPB = Upper Probability Bound
Figure 10.3-7 Exceedance Probability Distribution for Short-Tailed Shrew
Exposed to tPCBs at Location 13 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-25
-------
1
Location 14
2
3
a
«
a
©
a.
4>
w
=
¦a
0)
0)
w
H
UJ
100
80
60
40
20
¦Monte Carlo
¦LPB
¦UPB
0.01
0.1
1 10
TDI (mg/kg bw/d)
100
1000
Notes: LPB = Lower Probability Bound
UPB = Upper Probability Bound
5
6
Figure 10.3-8 Exceedance Probability Distribution for Short-Tailed Shrew
Exposed to tPCBs at Location 14 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-26
-------
Location 15
Notes: LPB = Lower Probability Bound
UPB = Upper Probability Bound
Figure 10.3-9 Exceedance Probability Distribution for Short-Tailed Shrew
Exposed to tPCBs at Location 15 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-27
-------
Location 13
TDI (ng/kg bw/d)
1
2 Notes: LPB = Lower Probability Bound
3 UPB = Upper Probability Bound
4 Figure 10.3-10 Exceedance Probability Distribution for Short-Tailed Shrew
5 Exposed to TEQ at Location 13 of the PSA
6
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia 7/10/2003
-------
Location 14
100
0s
.a
CS
.a
o
•-
a.
u
H
W
80
60
40
20
Monte Carlo
LPB
UPB
10 100
TDI (ng/kg bw/d)
1000
10000
1
2 Notes: LPB = Lower Probability Bound
3 UPB = Upper Probability Bound
4 Figure 10.3-11 Exceedance Probability Distribution for Short-Tailed Shrew
5 Exposed to TEQ at Location 14 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-29
-------
Location 15
1
2 Notes: LPB = Lower Probability Bound
3 UPB = Upper Probability Bound
4 Figure 10.3-12 Exceedance Probability Distribution for Short-Tailed Shrew
5 Exposed to TEQ at Location 15 of the PSA
6
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-30
-------
1
2
3
4
5
6
7
8
9
10
11
12
The probability bounds estimated for red fox foraging in Reach 5 are depicted in Figure 10.3-5.
The 10th percentile of the probability envelope formed by the lower and upper bounds ranged
between 0.106 and 0.702 mg/kg bw/d. The 50th percentile ranged between 0.607 and 3.54 mg/kg
bw/d, and the 90th percentile ranged between 3.52 and 23.5 mg/kg bw/d. In comparison, the 10th
percentile of the Monte Carlo output was 0.434, the 50th percentile was 2.68, and the 90th
percentile was 16.7 mg/kg bw/d (Table J.2-6).
Short-tailed shrew living at Locations 13 and 14 had the highest exposure to tPCBs. Red fox
foraging in Reach 5 had slightly less exposure to tPCBs than shrews at Locations 13 and 14.
Red fox in Reach 5 and short-tailed shrew at Locations 13 and 14 had the highest exposures to
TEQ. For both tPCBs and TEQ, short-tailed shrew foraging at Location 15 had the lowest
exposure.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-31
-------
1 10.4 EFFECTS ASSESSMENT
2 The objective of the effects assessment is to review the scientific literature and derive the most
3 appropriate effects metrics for effects of tPCBs and TEQ to omnivorous and carnivorous
4 mammals. An effects metric can be represented by a dose-response relationship or a daily dose
5 for a COC that represents a threshold beyond which toxic effects may appear in omnivorous and
6 carnivorous mammals. The effects metrics will be used, in conjunction with the exposure
7 assessment, to estimate risks to omnivorous and carnivorous mammals exposed to tPCBs and
8 TEQ in the Housatonic River PSA. This section focuses on effects that have an influence on the
9 long-term maintenance of mammal populations (i.e., mortality, or impairment of reproduction or
10 growth). Studies involving multiple exposure treatments that employed statistical evaluations to
11 determine whether treatment results were different from controls are preferred. Studies that
12 document effects of tPCBs and TEQ on the representative species, red fox and northern short-
13 tailed shrew, were preferred but unavailable. As a result, laboratory studies involving surrogate
14 species were used to estimate effects to the representative species. For short-tailed shrew and red
15 fox, the rat was used as a surrogate species for effects due to exposure to tPCBs. In the case of
16 exposure to TEQ, the mouse was used as a surrogate species for short-tailed shrew, while the rat
17 was used for red fox.
18 Exposure of mammals to tPCBs and 2,3,7,8-TCDD (TEQ) causes a range of effects (see text
19 box). The congeners that comprise the TEQ group have the ability to bind to the aryl
20 hydrocarbon receptor protein (Bosveld and van den Berg 1994) and elicit an Ah-mediated
21 biochemical and toxic response. A discussion of the chemical features that elicit the toxic
22 response and the mode of action are shown below.
23
24
25
26
27
28
29
30
Effects of tPCBs and TEQ on Mammals
Types of effects to mammals from exposure to tPCBs and TEQ include:
¦ Hormone induction
¦ Decreases in body and organ weight
¦ Reduced fertility
¦ Reduced litter size
¦ Reduced survival at birth or weaning
¦ Mortality
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-32
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20 Presented below is a brief review of the scientific literature on the effects of dietary tPCBs and
21 TEQ to mammals. The discussion focuses on ecologically relevant effects endpoints such as
22 survival, growth, and reproduction. A summary of reproductive effects for tPCBs and TEQ is
23 presented in Figures J.3-1 and J.3-2 and Tables J.3-2 and J.3-3.
24 10.4.1 Review of Effects of tPCBs and TEQ
25 10.4.1.1 tPCBs—Mortality
26 Linder et al. (1974) studied the effects of Aroclor 1254 and 1260 on 3- to 4-week-old Sherman-
27 strain male rats. Oral LD50s were 1,295 and 1,315 mg/kg bw/d, respectively (Linder et al. 1974).
28 Under similar test conditions, groups of 10 female Sherman-strain rats were treated with Aroclor
29 1260 doses of 7.2, 38.2, and 72.4 mg/kg bw/d for 8 months (Kimbrough et al. 1972). During this
30 time period, one, two, and eight females died, respectively. Bruckner et al. (1973) estimated a
31 14-day LD50 of 4,250 mg/kg bw/d for Aroclor 1242 for rats.
Mode of Action of TEQ Congeners
Congeners that comprise the TEQ group have the ability to bind with the Ah receptor
and elicit similar toxic responses. The most toxic congeners tend to be those that
have a planar shape and are chlorinated in the 2,3,7, and 8 positions for dioxins and
furans, and in the meta and para positions for PCBs. This structural configuration
best fits the receptor and leads to a common mechanism of action in many animal
species involving binding to the aryl hydrocarbon (Ah) receptor and elicitation of an
Ah receptor-mediated biochemical and toxic response. The toxic response of this
group of chemicals is, therefore, related to the three-dimensional structure of the
substance, including the degree of chlorination and positions of the chlorine on the
aromatic frame.
Planar chlorinated hydrocarbons are found in the environment as a mixture of
congeners. The congeners can have different toxic potencies. To address this issue
and effectively estimate the relative toxicity of these mixtures, various systems have
been created involving the development and use of toxic equivalency factors (TEFs)
to derive toxic equivalence (TEQ). The approach used for this assessment is
described in Section 6.4.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
10.4.1.2 tPCBs—Reproduction
Many of the available studies focus on determining effects to offspring of mammals following in
utero and/or lactational exposure to tPCBs. Impaired reproductive performance as a result of
maternal PCB exposure has been reported for many mammals, including rats and mice. In
general, females are administered contaminants by gavage or in the diet prior to or during
gestation. Endpoints studied included pre- and post-natal survival and development, fertility,
and other effects (Linder et al. 1974; Brezner et al. 1984; Overmann et al. 1987; Linzey 1987,
1988; Masuda et al. 1979; Allen and Barsotti 1976; Bleavins et al. 1981).
A two-generation reproduction study was performed in which groups of 20 female and 10 male
Sherman rats were exposed to diets of Aroclor 1254 at doses of 0.06, 0.32, 1.5, and 7.6 mg/kg
bw/d (Linder et al. 1974). Exposure to Aroclor 1254 caused significantly reduced litter sizes at
the 1.5 and 7.6 mg/kg bw/d doses. Survival to weaning was reduced by 77.8% in the second
generation. Spencer (1982) investigated reproductive effects of Aroclor 1254 using eight
treatment concentrations on Sprague Dawley strain rats fed treated diets on days 6 through 15 of
gestation. Statistically significant reductions were observed in fetal weight at birth (11.8%;
p<0.05) and fetal survival (28%; p<0.05) for rats fed diets of Aroclor 1254 at 7.47 and 17.1
mg/kg bw/d, respectively. Reproductive impairment in white-footed mice was also observed in
several studies (Linzey 1987; McCoy et al. 1995; Merson and Kirkpatrick 1976). Effects
included longer intervals between births, smaller litter sizes at birth, smaller litter sizes at
weaning, reduced mean birth and weaning weight in offspring, and reduced litter production rate.
10.4.1.3 TEQ—Mortality
The acute lethality of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), 1,2,3,7,8-
pentachlorodibenzo-p-dioxin (PCDD), and 1,2,3,4,7,8-hexachlorodibenzo-p-dioxin (HCDD) was
investigated for Long-Evans (LE) rats and Han/Wistar (H/W) (Pohjanvirta et al. 1993). The
H/W rats were approximately 1,000-fold more resistant to TEQ than LE rats. For example, the
LD50 values for exposure of female and male LE rats were 9,800 and 17,700 ng/kg bw/d TEQ,
respectively, whereas LD50s for H/W female rats were greater than 7,200,000 ng/kg bw/d TEQ.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-34
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
In long-term exposure studies, 100% mortality occurred at 57.1 ng/kg bw/d TEQ when Sprague
Dawley rats were continuously exposed to TCDD for 31 weeks. Kociba et al. (1978) conducted
a 2-year study by feeding diets containing TCDD at 1, 10, and 100 ng/kg bw/d TEQ to male and
female Sprague Dawley rats. At 100 ng/kg bw/d TEQ, they observed a cumulative increase in
mortality in the latter half of the study period and a decrease in mean body weight from 6 to 24
months compared to controls.
10.4.1.4 TEQ—Reproduction
Among the 209 possible PCB congeners, the non-ortho-substituted (planar) congeners are the
most toxic due to their structural similarity to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). The
most toxic of these congeners include 3,3',4,4'-tetrachlorobiphenyl (PCB-77), 3,3',4,4',5,5'-
hexachlorobiphenyl (PCB-169), and 3,3',4,4',5-pentachlorobiphenyl (PCB-126) (Safe 1984).
The prenatal toxicity of PCB-77 was determined in rats and mice fed contaminated diets between
days 6 and 18 of gestation (Marks et al. 1989; d'Argy et al. 1987; Wardell et al. 1982; Rands et
al. 1982a; 1982b). Marks et al. (1989) reported a PCB-77 dose-related increase in the percentage
of implants that resorbed, at concentrations ranging from 400 (7% increase) to 6,400 (82.5%
increase) ng/kg bw/d TEQ; a significant increase (16.4%) was determined at 1,600 ng/kg bw/d
TEQ and above. In addition, the average number of live fetuses per female mouse was
significantly reduced (21.5%) at 1,600 ng/kg bw/d TEQ and above (Marks et al. 1989). Khera
and Ruddick (1973) treated pregnant Wistar rats with eleven doses of TCDD on gestation days 6
to 15. A dose-related decrease in live fetuses was observed; 100% embryonic lethality was
reported when animals were exposed to a dose of 4,000 ng/kg bw/d TEQ or higher. Similar
observations were made by Sparschu et al. (1971) in Sprague Dawley rats fed six doses of
TCDD on days 6 through 15 of gestation. The number of viable fetuses decreased and the total
number of resorptions increased starting at 125 ng/kg bw/d TEQ. Giavini et al. (1983) and
Huuskonen et al. (1994) also observed a reduction in the number of living fetuses per litter when
rats were fed TCDD.
In similar studies on rats and mice, other TEQ effects observed included significant increases in
mortality of offspring, resorptions, mortality of embryos and of offspring, reduced fecundity,
reduced litter size, reduced body weight at birth, reduced in vitro fertilizing ability of the eggs,
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia 7/10/2003
-------
1 and reduced survival to weaning (Rands et al. 1982a; Linzey 1987; d'Argy et al. 1987; Wardell
2 et al. 1982; Huang et al. 1998; Neubert and Dillman 1972; Murray et al. 1979; Bjerke and
3 Peterson 1994; Mably et al. 1992; Bjerke et al. 1994; Flaws et al. 1997; Gray and Ostby 1995;
4 Thomas and Hinsdill 1979; Khera and Ruddick 1973).
5 10.4.2 Effects Metrics for Characterizing Risk
6 Effects data can be characterized and summarized in a variety of ways ranging from benchmarks
7 designed to be protective of most or all species to concentration- or dose-response curves. A
8 summary of the decision criteria used to derive effects metrics is provided in the text box.
9 Further details on the decision criteria used in selecting effects metrics is provided in Section 6.6
10 of the ERA.
11
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-36
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Decision Criteria for Derivation of Effects Metric
The following is the hierarchy of decision criteria used to characterize effects for each
receptor-COC combination:
1. Have single-study bioassays with five or more treatments been conducted on the
receptor of interest or a reasonable surrogate? If yes, estimate the
concentration- or dose-response. If not, go to 2.
2. Are multiple bioassays with similar protocols, exposure scenarios and effects
metrics available that, when combined, have five or more treatments for the
receptor of interest or a reasonable surrogate? If yes, estimate the dose-
response relationship as in 1. If not, go to 3.
3. Have bioassays with less than five treatments been conducted on the receptor of
interest or a reasonable surrogate? If yes, conduct or report results of
hypothesis testing to determine the NOAEL and LOAEL. If not, go to 4.
4. Are sufficient data available from field studies and monitoring programs to
estimate concentrations or doses of the COC that are consistently protective or
associated with adverse effects? If yes, develop field-based effects metrics. If
not, go to 5.
5. Derive a range where the threshold for the receptor of interest is expected to
occur. Because information on the sensitivity of the receptor of interest is
lacking, it is difficult to derive a threshold that is neither biased high or low. If,
however, bioassay data are available for several other species, calculate a
threshold for each to determine a threshold range that spans sensitive and
tolerant species. That range is likely to include the threshold for the receptor of
interest.
In this ERA, data were available to derive dose-response curves using surrogate mammals for the
representative species.
10.4.2.1 Effects of tPCBs to Red Fox and Short-Tailed Shrew
The Spencer (1982) study was used for the derivation of a dose-response curve based on
mortality at birth. Figure 10.4-1 presents the dose-response curve for mortality of rats at birth.
The dose-response curve indicates that 10% and 20% declines in mortality at birth would be
expected at doses of 3.05 and 5.37 mg/kg bw/d, respectively.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-37
-------
1
2
3
4
5
6
7
8
9
10
11
100
~
0.1
10
1000
Dose (mg/kg bw/d) (logscale)
Note: Symbols indicate raw data.
Figure 10.4-1 Dose-Response Curve for Effects of tPCBs on Mortality
at Birth of Rats
10.4.2.2 Effects of TEQ to Red Fox
The Khera and Ruddick (1973) and Sparschu et al. (1971) studies were combined, because of the
similarity of the protocols, for the derivation of a dose-response curve based on reproductive
effects. Figure 10.4-2 presents the dose-response curve for reproductive fecundity of rats
exposed to TEQ. The dose-response curve indicates that 10% and 20% declines in reproductive
fecundity would be expected at doses of 156 and 330 ng/kg bw/d TEQ, respectively.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-38
7/10/2003
-------
Dose (ng TEQ/kg bw/d)
2 Note: Symbols indicate raw data.
3 Figure 10.4-2 Dose-Response Curve for Effects of TEQ on Reproductive
4 Fecundity of Rat
5 10.4.2.3 Effects of TEQ to Short-Tailed Shrew
6 The Marks et al. (1989) study was used for the derivation of a dose-response curve based on
7 reproductive effects. Figure 10.4-3 presents the dose-response curve for reproductive fecundity
8 of mice exposed to TEQ. The dose-response curve indicates that 10% and 20% declines in
9 reproductive fecundity would be expected at doses of 570 and 1207 ng/kg bw/d TEQ,
10 respectively.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-39
-------
1
2
3
4
5
10000 20000
Dose (ng TEQ/kg bw/d)
30000
Note: Symbols indicate raw data
Figure 10.4-3 Dose Response Curve for Effects of TEQ on Reproductive
Fecundity of Mouse
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-40
-------
1 10.5 RISK CHARACTERIZATION
2 This section characterizes risk to omnivorous and carnivorous mammals exposed to tPCBs and
3 TEQ in the PSA of the Housatonic River. The risk characterization uses two and three lines of
4 evidence to determine potential ecological risks to red fox and short-tailed shrew, respectively.
5 The major lines of evidence are considered to be independent and will be combined in a weight-
6 of-evidence assessment. The key risk questions and the lines of evidence are summarized in the
7 text box.
8
9
10
11
12
13
14
15
16
17
18 Section 10.5.1 presents a brief overview of the methodology, results, and interpretation of the
19 mammal surveys conducted from 1998 to 2001 in the Housatonic PSA. A more detailed
20 presentation of this information is presented in Appendix A. In Section 10.5.2, the dose-
21 response curves are combined with the corresponding exposure distributions to derive risk curves
22 that characterize the relationship between probability and magnitude of effect. A brief overview
23 of the population demographics field study is described in Section 10.5.3. A weight-of-evidence
24 assessment is presented is Section 10.5.4 along with sources of uncertainty (Section 10.5.5),
25 extrapolation of risk to other species (10.5.6) and the overall findings of the risk characterization
26 (Section 10.5.6).
27 10.5.1 Field Surveys (Performed by EPA)
28 The mammalian community in the PSA was studied by EPA over a 4-year period, from 1998 to
29 2001. Surveys were conducted to record presence, relative abundance, and habitat usage for
Key Risk Questions
¦ Are the concentrations of tPCBs and TEQ present in the prey of omnivorous and
carnivorous mammals sufficient to cause adverse effects to individuals inhabiting
the PSA of the Housatonic River?
¦ If so, how severe are the risks and what are their potential consequences?
Lines of Evidence
¦ Use of semiquantitative biological field surveys.
¦ Probabilistic exposure and effects modeling.
¦ Population demography field study for short-tailed shrew.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-41
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
small and large mammals including short-tailed shrew and red fox. A variety of field survey
techniques including small mammal trapping, snow tracking, and scent-post station surveys were
used to characterize the mammalian community.
Forty-two mammal species were documented in the PSA during the 4 years of field surveys (see
Appendix A for more details). Many species were observed throughout the PSA in a variety of
habitats. Forested communities, such as red maple swamp, black ash-red maple-tamarack,
transitional floodplain forest, and high-terrace floodplain forest supported the greatest number of
species. Observations of omnivorous and carnivorous mammals including coyotes, red fox,
raccoons, white-footed mice, short-tailed shrews, and little brown bats were common, all of
which were observed in forested and nonforested habitats as well as riverine, shoreline, wetland,
upland, and residential habitats. Other carnivorous mammals observed in the PSA included
bobcats, fishers, and long-tailed weasels. Omnivorous mammals were one of the most abundant
groups of mammals observed in the PSA. Common omnivores included white-footed mice,
raccoons, striped skunks, Virginia opossums, and black bears. The short-tailed shrew was the
most abundant insectivorous mammal observed in the PSA. The semiquantitative nature of the
field surveys and lack of reference locations in the surveys make it difficult to develop
relationships between abundance of representative species and concentrations of COCs.
During the small mammal surveys, females were checked for evidence of lactation; some
individuals were euthanized and the uterus was removed for placental scar analysis. Number of
placental scars has been used in a variety of mammals to estimate litter sizes (Hensel et al. 1969;
Sanderson 1950; Oleyar and McGinnes 1974; Nixon et al. 1975). Placental scar identifications
of small female mammals were performed on four species including short-tailed shrew, white-
footed mouse, meadow jumping mouse, and masked shrew at Locations 13, 14, and 15. Sample
sizes were small for each species. In some cases, placental scars were difficult to identify,
particularly for white-footed mouse. Although the data have uncertainties due to these
limitations, it appeared that there were no differences in numbers of placental scars between
locations (Table J.4-1). The white-footed mouse was the most frequently captured species with
sample sizes ranging from 6 to 11 females amongst the locations. The average number of
placental scars per white-footed mouse female was 6.33, 6.27, and 6.50 at Locations 13, 14, and
15, respectively. Mean soil tPCB concentrations vary over 50-fold among these three locations.
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia a rs 7/10/2003
-------
1 The number of females for the other three species combined ranged from two to six, and had a
2 lower average number of scars ranging from zero to 2.50 (Table J.4-1).
3 10.5.2 Comparison of Estimated Exposures to Laboratory-Derived Effects
4 Doses
5 Red fox exposure was assessed for all of Reach 5 and short-tailed shrew exposure was estimated
6 in three areas (Locations 13, 14, and 15, Figure J.2-3) in the PSA.
7 For each receptor-COC combination, a category of low, intermediate, or high risk was assigned
8 following integration of the exposure and effects distributions. This exercise was done
9 separately for the results of the Monte Carlo analyses and each of the lower and upper bounds
10 from the probability bounds analyses. The "risk category" refers to the level of risk based on the
11 results of the Monte Carlo analysis. The "risk range" refers to the levels of risk based on the
12 results of the probability bounds analyses. The 10% and 20% effects doses for each species and
13 COC are presented in Section 10.4.
14
15
16
17
18
19
20
21 The results of the risk characterization are summarized in Table 10.5-1. Figures 10.5-1 to
22 10.5-8 present the risk curves for red fox exposed in Reach 5 and short-tailed shrew exposed at
23 Locations 13, 14, and 15 to tPCBs and TEQ.
Guidance for Determining Level of Risk to Representative Species
¦ If the probability of 10% or greater effect is less than 20%, then the risk to
omnivorous and carnivorous mammals was considered low.
¦ If the probability of 20% or greater effect is greater than 50%, then the risk to
omnivorous and carnivorous mammals was considered high.
¦ All other outcomes were considered to have intermediate risk.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-43
-------
Table 10.5-1
Summary of Qualitative Risk Statements for Omnivorous and Carnivorous
Mammals from the Housatonic River Study Area
Mammal / Location
Qualitative Risk Statements
tPCBs
TEQ
Risk Category
Risk Range
Risk Category
Risk Range
Red Fox
Reach 5
Intermediate
Low-Intermediate
Intermediate
Low- Intermediate
Short-Tailed Shrew
Location 13
High
Intermediate -High
Low
Low-Intermediate
Location 14
High
Intermediate-High
Low
Low- Low
Location 15
Low
Low-Intermediate
Low
Low-Low
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia a a 7/10/2003
-------
Reach 5
100
S 60
40
20
Monte Carlo
LPB
UPB
• Low - Inter. Criterion
O Inter. - High Criterion
20 40 60
% Mortality at Birth
100
Figure 10.5-1 Risk Function for Red Fox Exposed to tPCBs in Reach 5 of the
Housatonic River
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-45
-------
Reach 5
100
80
.Q
03
.Q
0
-
a.
01
(J
S
03
¦a
01
x
w
60
40
20
Monte Carlo
— LPB
— UPB
9 Low - Inter. Criterion
O Inter. - High Criterion
• ¦ •
20 40 60
% Decline in Fecundity
80
100
Figure 10.5-2 Risk Function for Red Fox Exposed to TEQ in Reach 5 of the
Housatonic River
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-46
-------
Site 13
ct
o
a-
a>
u
=
¦S
V
V
W
H
W
100
80
60
40
20
— Monte Carlo
— LPB
— UPB
• Low - Inter. Criterion
O Inter. - High Criterion
20 40 60
% Mortality at Birth
80
100
2 Figure 10.5-3 Risk Function for Short-Tailed Shrew Exposed to tPCBs at Location
3 13 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-47
-------
Site 14
¦O
O
Ph
w
=
¦S
V
V
W
H
W
100
80
60
40
20
— Monte Carlo
— LPB
— UPB
• Low - Inter. Criterion
O Inter. - High Criterion
20 40 60
% Mortality at Birth
80
100
2 Figure 10.5-4 Risk Function for Short-Tailed Shrew Exposed to tPCBs at Location
3 14 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-48
-------
Site 15
2 Figure 10.5-5 Risk Function for Short-Tailed Shrew Exposed to tPCBs at Location
3 15 of the PSA
4
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-49
-------
Site 13
% Decline in Fecundity
Figure 10.5-6 Risk Function for Short-Tailed Shrew Exposed to TEQ at Location
13 of the PSA
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia r/-\ 7/10/2003
-------
Site 14
0 20 40 60 80 100
% Decline in Fecundity
Figure 10.5-7 Risk Function for Short-Tailed Shrew Exposed to TEQ at Location
14 of the PSA
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia pi 7/10/2003
-------
Site 15
¦O
O
Ph
w
=
¦S
V
V
W
H
W
100
80
60
40
20
Monte Carlo
LPB
UPB
• Low - Inter. Criterion
O Inter. - High Criterion
o
L
20 40 60
% Decline in Fecundity
80
100
Figure 10.5-8 Risk Function for Short-Tailed Shrew Exposed to TEQ at Location
15 of the PSA
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-52
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
The results of the risk characterization showed that the highest risk to omnivorous and
carnivorous mammals is from exposure to tPCBs at Locations 13 and 14. The risk category for
short-tailed shrew at Locations 13 and 14 was high, and the risk range, as determined by the
probability bounds analysis, ranged from intermediate to high. Risk to shrews at Location 15
was low. The risk category for exposure of short-tailed shrew to TEQ at Location 13 is low; the
risk range is low to intermediate. Short-tailed shrew exposed to TEQ at Locations 14 and 15
have a risk category of low. Both the upper and lower bound of the risk ranges for Locations 14
and 15 are low. Red fox had an intermediate risk category for both exposure to tPCBs and TEQ.
The risk range for both COCs for red fox is low to intermediate.
10.5.3 Population Demography Field Study (Performed by GE)
A population demography field study was performed in 2001 along a 16-km reach of the
Housatonic River between Pittsfield and Woods Pond (Boonstra 2002). The study objectives
included evaluating population density, survival, rates of reproduction, sex ratio, and growth
rates of short-tailed shrew. More information on this study is provided in Boonstra (2002).
Six sites were selected based on tPCB concentrations, habitat uniformity, and sufficient area to
permit a 1-hectare (ha) trapping grid to be located within each site. All sites were located within
the eastern deciduous temperate forest biome in primarily palustrine habitat, with portions of two
grids also including upland habitat. Two grid classes were selected, designated as high and low
concentrations of tPCBs, with three sites in each class. Habitat varied across the grids,
particularly between the northern sites and the southern sites (e.g., the former had much more
vegetative biomass than did the latter). Comparison of the northern and southern sites is an
indirect way of determining whether habitat quality has an effect on the survival, growth, and
reproduction of shrews. The six areas trapped in this study, in fact, varied in habitat quality.
Without habitat and microhabitat data at the six trapping sites, it is difficult to determine if
differences in habitat explain variation in population parameters among the six sites. Three
trapping sessions were conducted in spring, summer, and fall (trapping sessions one, two, and
three, respectively) over the course of the study, with each session lasting 3 consecutive days.
Boonstra (2002) suggested that population characteristics of short-tailed shrew living on more
contaminated tPCB sites in the PSA are not negatively affected compared to those living on less
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia ro 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
contaminated sites. The results showed that exposure of short-tailed shrew to tPCBs had no
apparent effect on population density, sex ratio, reproduction, and growth rate. In general, shrew
populations showed high monthly survival. Although there was grid-to-grid variability in
survival, this variability could not be explained by differences in tPCB concentrations among the
grids. The only significant effect was on body mass of males but, in this case, the males living in
the highly contaminated sites weighed more, not less, than those living in the low contaminated
sites. In summary, variations in tPCB concentrations among the sites resulted in no differences
in population demography of short-tailed shrew according to Boonstra (2002).
There were several confounding factors in the Boonstra (2002) study, including flooding of the
floodplain, which prevented trapping on three of the six grids during the first of the three
trapping events. While spring flooding is a natural phenomenon within the Housatonic River
PSA, and shrew populations are very likely accustomed to such events, it is difficult to determine
the impact of flooding on the study results. Habitat quality varied across the grids. Without
identifying habitat and microhabitat data at the six trapping sites, it is difficult to tell if
differences in habitat explain variation in population parameters between the six sites. The lack
of reproduction rate data in the Boonstra (2002) study creates uncertainty regarding population
maintenance. Even with high adult survivorship, without reproduction rate data, it is impossible
to know if shrew populations are maintaining themselves through natural production or
immigration. The use of body weight to imply reproductive status may not be appropriate
because it is insensitive to potential reproductive impairments. These factors limit the strength
of conclusions from the study.
Comparison of the spatially weighted concentrations of tPCBs in soil from Boonstra (2002) and
the spatially weighted mean concentrations of tPCBs in soil derived as part of this ERA (see
Appendix J) showed that the Boonstra (2002) estimates in the six grids appear to be in error.
Additional analyses were carried out to verify the relationship between shrew survival and
concentrations of tPCBs derived as part of this ERA. The results of the analyses indicated a
significant relationship between concentrations of tPCBs in soil and survival of shrews from
summer to autumn for males, females, and males and females combined. Although the results of
the analyses indicated a significant relationship between soil concentrations of tPCBs and shrew
survival, the confidence limits indicated that the relationships were not strong. Some of the
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-54
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
"noise" in the relationships may be attributed to the influence of habitat differences among the
grids, small sample sizes, the effects of flooding, the analytical methods used to measure tPCBs,
the selection of the correct soil samples for inclusion in the analyses, and the relatively small
number of treatments.
10.5.4 Weight-of-Evidence Analysis
A weight-of-evidence analysis was used to combine the two major lines of evidence described in
the preceding sections for red fox and short-tailed shrew. The goal of this analysis was to
determine whether significant risk is posed to omnivorous and carnivorous mammals in the
Housatonic River PSA as a result of exposure to tPCBs and TEQ. The three-phase approach of
Menzie et al. (1996) and the Massachusetts Weight-of-Evidence Workgroup was applied for this
purpose, in which weight-of-evidence was reflected in the following three characteristics: (a) the
weight assigned to each measurement endpoint, (b) the magnitude of response observed in the
measurement endpoint, and (c) the concurrence among outcomes of the multiple measurement
endpoints.
A discussion of attributes considered in the WOE is provided in Section 2, and the rationales for
weighting of measurement endpoints are provided in Appendix J. A summary of the derived
weightings is provided in Table 10.5-2. For both tPCBs and TEQ, the field surveys, the
population demography field study, and the modeled exposure and effects lines of evidence were
given a moderate/high value.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-55
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
Table 10.5-2 Weighting of Measurement Endpoints for Omnivorous and
Carnivorous Mammals Weight-of-Evidence Evaluation
Attributes
Field
Surveys
Population
Demography Field
Study*
Modeled Exposure
and Effects for
tPCBs and TEQ
I. Relationship Between Measurement and Assessment Endpoints
1. Degree of Association
L
M/H
M
2. Stressor/Response
M
M
M/H
3. Utility of Measure
L/M
M/H
M/H
II. Data Quality
4. Data Quality
H
M/H
M/H
III. Study Design
5. Site Specificity
H
H
L/M
6. Sensitivity
M
M
H
7. Spatial Representativeness
H
H
M
8. Temporal Representativeness
M/H
H
M
9. Quantitative Measure
M
M/H
H
10. Standard Method
H
H
M/H
Overall Endpoint Value
M/H
M/H
M/H
* Field study only for short-tailed shrew.
L = low; M = moderate; H = high
The magnitude of the response in the measurement endpoint is considered together with the
measurement endpoint weight in judging the overall weight-of-evidence (Menzie et al. 1996).
This requires assessing the strength of evidence that ecological harm has occurred, as well as an
indication of the magnitude of response, if present. The weighting values, evidence of harm, and
magnitudes of responses were combined in a matrix format and are presented in Tables 10.5-3
and 10.5-4. The field surveys indicated that red fox and short-tailed shrew are likely common in
the PSA. However, it is not known whether these species would be more abundant in the
absence of COCs, or if they are abundant because of immigration from less contaminated areas.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-56
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
The objectives of the population demography field study (Boonstra 2002) were to measure
population demography of short-tailed shrews directly, including reproduction, growth, and
survival in the PSA. The responses were quantitatively compared with magnitude of exposure.
However, confounding factors such as flooding, area (i.e., location within the floodplain), habitat
quality, and the use of body weight to imply reproductive status may have had significant effects
on population demographics and the results of the field study. Additional analyses of data
generated in the population demography field study showed that tPCBs may be having effects on
survival of short-tailed shrews. Other demographic parameters, however, do not appear to be
affected by tPCB concentrations in soil. The results from the modeled exposure and effects line
of evidence suggest that there is a high risk to short-tailed shrew exposed to tPCBs at Locations
13 and 14, and a low risk at Location 15. There is an intermediate risk for fox exposed to tPCBs
foraging in Reach 5 (Table J.4-8). There is an intermediate risk to red fox exposed to TEQ in the
PSA, and low risk to short-tailed shrew exposed to TEQ at Locations 13, 14, and 15 (Table
J.4-9).
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-57
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Table 10.5-3
Evidence of Harm and Magnitude of Effects for Omnivorous and Carnivorous
Mammals Exposed to tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Field Surveys
Moderate/High
Undetermined
Low
Population
Demography Field
Study
Moderate/High
Undetermined (Shrew)
Intermediate
Modeled Exposure and
Effects
Moderate/High
Yes (Shrew)
Undetermined (Red Fox)
High
Intermediate
Table 10.5-4
Evidence of Harm and Magnitude of Effects for Omnivorous and Carnivorous
Mammals Exposed to TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Field Surveys
Moderate/High
Undetermined
Low
Population
Demography Field
Study
Moderate/High
Undetermined (Shrew)
Intermediate
Modeled Exposure and
Effects
Moderate/High
No (Shrew)
Undetermined (Red Fox)
Low
Intermediate
A graphical method was used for displaying concurrence among measurement endpoints. Tables
10.5-5 and 10.5-6 depict the outcome for omnivorous and carnivorous mammals exposed to
tPCBs and TEQ, respectively. The uncertainty concerning the modeled exposure and effects line
of evidence for tPCBs and TEQ, particularly because surrogate species were used for estimating
effects, could mean that risks of these COCs are being under- or over-estimated for the PSA.
The field survey line of evidence, although inconclusive in terms of demonstrating effects and
risk, indicated that omnivorous and carnivorous mammals, including short-tailed shrew and red
fox, were commonly observed in hardwood forests, palustrine forested areas, and shorelines of
the PSA. In addition, according to Boonstra (2002), the population demography field study line
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC 10 58 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
of evidence suggests that no effects attributable to tPCBs are occurring to short-tailed shrews.
However, the results of the latter study are confounded by factors such as flooding, quality of
habitat, and using body weight to imply reproductive status, all of which likely introduced a
large amount of uncertainty. In addition, the soil concentration data used in the Boonstra (2002)
study appear to be in error. Additional analyses with revised soil concentration data do not
support the conclusion in the Boonstra (2002) study that shrew survival in the study grids was
not affected by soil tPCB concentration, but indicates that there is a statistically significant
relationship between PCB concentrations and survival, although not strong.
With the exception of modeled exposure and effects for shrew, the potential for harm to shrew
and red fox based on the remaining lines of evidence was undetermined. Based on the modeled
exposure and effects assessment for shrew, there is a high potential for adverse effects resulting
from tPCB exposure and very limited potential for adverse effects resulting from exposure to
TEQ.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-59
-------
1
2
3
4
5
6
7
8
9
10
11
12
Table 10.5-5
Risk Analysis Summary for Omnivorous and Carnivorous Mammals Exposed to
tPCBs in the Housatonic River PSA
. . t, . Survival, growth, and reproduction of omnivorous and carnivorous
Assessment Endpoint: ,
1 mammals
~
Harm/Magnitu de
Weighting Factors (increasing confidence of weight)
Low
Low/
Moderate
Moderate
Moderate/
High
High
Yes/High
MEE-S
Yes/Intermediate
Yes/Low
Undetermined/High
Undetermined/Intermediate
MEE-F,
PDFS
Undetermined/Low
FS
No/Low
No/Intermediate
No/High
FS = Field surveys
MEE-S = Modeled exposure and effects - shrew
MEE-F = Modeled exposure and effects - red fox
PDFS = Population demography field study for short-tailed shrew only
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-60
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Table 10.5-6
Risk Analysis Summary for Omnivorous and Carnivorous Mammals Exposed to
TEQ in the Housatonic River PSA
. , _ , . , Survival, growth, and reproduction of omnivorous and carnivorous
Assessment Endpoint: ,
mammals
Harm/Magnitude
Weighting Factors (increasing confidence of weight)
Low
Low/
Moderate
Moderate
Moderate/
High
High
Yes/High
Yes/Intermediate
Yes/Low
Undetermined/High
Undetermined/Intermediate
MEE-F,
PDFS
Undetermined/Low
FS
No/Low
MEE-S
No/Intermediate
No/High
~
I
_l
J
FS = Field surveys
MEE-S = Modeled exposure and effects - shrew
MEE-F = Modeled exposure and effects - red fox
PDFS = Population demography field study for short-tailed shrew only
10.5.5 Sources of Uncertainty
The assessment of risk to omnivorous and carnivorous mammals contains uncertainties. Each
source of uncertainty can influence the estimates of risk. Therefore, it is important to describe
and, when possible, specify the magnitude and direction of such uncertainties. Some of the
major sources of uncertainty associated with the assessment of risks of tPCBs and TEQ to
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-61
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
omnivorous and carnivorous mammals are briefly described below. A more complete list is
presented in Appendix J.
¦ The Monte Carlo sensitivity analyses suggest that the free metabolic rate (FMR) slope
and power terms were generally the most influential variables on predicted total daily
intakes of COCs. However, no measurements of free metabolic rate were available
for the representative wildlife species. Similarly, measured food intake rates were not
available for free-living red fox or northern short-tailed shrew or reasonable surrogate
species. Therefore, free metabolic rates were estimated using allometric equations.
The use of allometric equations introduces some uncertainty into the exposure
estimates because they have model-fitting error, and are based on species different
from the representative species used in this assessment. Given the lack of data on
representative species used in the current assessment, it is difficult to judge the
magnitude of the uncertainty introduced by the use of the allometric models. The
uncertainty due to model-fitting error was propagated in the uncertainty analyses by
using distributions as input for the allometric slope and power terms.
¦ Sample sizes were limited for the analyses of COC concentrations in some prey
items. Only one composite sample of earthworm (comprising between 20 and 45
worms) and four samples of mammals were available to estimate exposure of shrews
to TEQ at each location. Similarly, only two or three samples of litter invertebrates
were available to estimate concentrations of tPCBs at each of Locations 13, 14, and
15. Uncertainty due to sample size was explicitly incorporated in the uncertainty
analyses. In the Monte Carlo analysis, sample size uncertainty was addressed by use
of the 95 upper confidence limit (UCL) on the mean. Use of the UCL addressed
uncertainty, but is biased toward overestimating exposure. In the probability bounds
analysis, uncertainty was addressed by specifying concentration variables as a range
from the 95% lower confidence limit (LCL) to the UCL. This treatment of
uncertainty is unbiased.
¦ PCB congeners 123 and 157 co-eluted with other congeners (PCB-123 with PCB-
149; PCB-157 with PCB-173 and PCB-201), leading to uncertainty about TEQ
concentrations in prey sample. This source of uncertainty was addressed in the
uncertainty analyses by estimating prey concentrations assuming concentrations of
PCB-123 and PCB-157 were equal to zero, and assuming that concentrations of PCB-
123 and PCB-157 were equal to the doublet and triplet concentrations, respectively.
The resulting TEQ estimates were then compared. If the ratio of the upper to lower
bound TEQ estimates was less than 1.3, this source of uncertainty was deemed
unimportant and disregarded. If the ratio exceeded 1.3, the uncertainty due to co-
elution was propagated through the uncertainty analyses.
¦ The largest source of uncertainty in the effects assessment was associated with the
lack of toxicity studies involving the representative species. There were no toxicity
studies available for red fox and short-tailed shrew exposed to tPCBs or TEQ. As a
result, laboratory studies involving surrogate species were used to estimate effects to
these species. These extrapolations introduced uncertainty in the effects assessment
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-62
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
because of the variation in sensitivities of mammal species to tPCBs and TEQ. The
sensitivity of wildlife to an environmental contaminant may also differ from that of a
laboratory or domestic species due to behavioral and ecological parameters, including
stress factors (e.g., competition, seasonal changes in temperature or food availability),
disease, and exposure to other contaminants.
¦ For omnivorous and carnivorous mammals, data for two and three major lines of
evidence were available for red fox and short-tailed shrew, respectively. For these
assessments, feeding studies involving prey and food items from the PSA would have
improved the weight-of-evidence assessment. Such studies would have accounted
directly for the specific characteristics of the Housatonic River ecosystem, and the
toxicity of the PCB mixture found on-site.
¦ The lack of reproduction rate data in the Boonstra (2002) study creates uncertainty
regarding population maintenance. Even with high adult survivorship, without
reproduction rate data, it is impossible to know if shrew populations are maintaining
themselves through natural production or immigration.
¦ In the Boonstra (2002) study, the use of body weight to imply reproductive fitness is a
limitation because it is insensitive to potential reproductive impairments. For
example, 6 of 10 female short-tailed shrews greater than 18 grams in weight trapped
by EPA in August 1999 had no evidence of breeding history upon placental scar
analysis. Although some of these six could have been young animals that had not
bred yet, others could have been animals with reproductive impairments due to
various factors, including PCB exposure. Additionally, during the 1999 trapping, of
those specimens submitted for tissue analysis (8 of the 10 females over 18 grams),
five females with no evidence of breeding history had body burdens up to 135 mg/kg
(average 74 mg/kg), whereas those with direct evidence of breeding had body
burdens up to 93 mg/kg (average 57 mg/kg). This sample size is quite small;
however, along with laboratory evidence of PCB-induced reproductive impairments
to mammals, it suggests that animal weight alone may not be representative of the
reproductive status of individual animals.
10.5.6 Conclusions
For omnivorous and carnivorous mammals, data from three major lines of evidence were
available, including field surveys, a population demography field study of short-tailed shrew, and
exposure and effects modeling. The weight-of-evidence analysis indicates an intermediate risk
for short-tailed shrews exposed to tPCBs and TEQ in the PSA. This conclusion, however, is
uncertain because of the lack of definitive findings as to whether effects are occurring in the field
surveys and population demography field study, and the lack of species-specific measures of
effects. The WOE also suggests, based on two lines of evidence for red fox, an intermediate risk
to fox exposed to tPCBs and TEQ in the PSA. This finding is also uncertain because although
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-63
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
fox were commonly observed during the field surveys, a foraging rate of 50% in Reach 5 was
used, and species-specific measures of effects were not available.
Field surveys were conducted (in part) to determine which omnivorous and carnivorous mammal
species were present in the PSA. The surveys were not designed to provide a quantitative
evaluation of the relationship between exposure to COCs and the survival, growth, and reproduction
of omnivorous and carnivorous mammals in the PSA. Instead, the surveys determined the presence,
relative abundance, size, and reproductive status of omnivorous and carnivorous mammals. Red
fox, short-tailed shrew, and other omnivorous and carnivorous mammals were observed frequently
in several areas in the PSA.
The objectives of the population demography field study were to determine population density,
survival, rates of reproduction, sex ratio, and growth rates of short-tailed shrew measured at six
sites having different concentrations of tPCBs in the PSA (Boonstra 2002). The Boonstra (2002)
results found that variation in tPCB concentrations among the sites resulted in no differences in
population demographic parameters of short-tailed shrew. However, confounding factors such
as flooding, area (i.e., location within the floodplain), and habitat quality may have had a
significant impact on shrew population demography. The lack of reproduction rate data in the
Boonstra (2002) study creates uncertainty regarding population maintenance. Even with high
adult survivorship, without reproduction rate data it is impossible to know if shrew populations
are maintaining themselves through natural production or immigration. In addition, the use of
body weight to imply reproductive fitness may not be appropriate because it is insensitive to
potential reproductive impairments.
In contrast to the findings in the Boonstra (2002) study, the results of the supplemental analyses
conducted for this ERA indicated a significant relationship between tPCB spatially weighted and
measured concentrations in soil and survival of shrews from summer to autumn for males,
females, and males and females combined, although the confidence limits indicate that the
relationships are not strong. Some of the "noise" in the relationships may be attributed to factors
listed above. The additional analyses do not support the Boonstra (2002) conclusion that "there
is no evidence that this variability [in shrew survival] can be explained by differences in tPCB
concentrations among the grids." The modeling of exposure and effects line of evidence was
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-64
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
used to determine the level of risk to the representative mammal species, short-tailed shrew, and
red fox. The effects characterization developed dose-response curves to describe the potential
effects of tPCBs and TEQ to omnivorous and carnivorous mammals. There were no toxicity
studies available for red fox and short-tailed shrew. Surrogate species were used to estimate
effects with the assumption that representative omnivorous and carnivorous mammal species
would experience adverse effects similar to the surrogate species.
The dose-response curve for effects of tPCBs to omnivorous and carnivorous mammals indicated
that 10% and 20% declines in mortality at birth would be expected at doses of 3.05 and 5.37
mg/kg bw/d, respectively. The modeled exposure results indicated that the daily intake of tPCBs
by red fox fell within this range while the daily intake of tPCBs by northern short-tailed shrew
was greater than 5.37 mg/kg bw/d at Locations 13 and 14. This means that shrews, and possibly
red fox in the PSA, are likely to receive tPCB doses that would cause adverse reproductive
effects. The daily intakes of short-tailed shrews at Locations 15 were below the 10% effects
dose, meaning that shrews are likely not at risk from exposure to TEQ at that site.
For TEQ, the dose-response curve for red fox indicated that 10% and 20% declines in
reproductive fecundity would be expected at doses of 156 and 330 ng/kg bw/d, respectively. The
modeled exposure results indicated that the daily intake of TEQ by red fox fell within this range.
It is, therefore, difficult to make definitive conclusions about the risks of TEQ to this species.
For northern short-tailed shrews, the dose-response curve indicated that 10% and 20% declines
in reproductive fecundity would be expected at doses of 570 and 1207 ng/kg bw/d TEQ,
respectively. The daily intakes of short-tailed shrews at Locations 13, 14, and 15 were below the
10%) effects dose, meaning that shrews are likely not at risk from exposure to TEQ in the PSA.
The field surveys and conclusions made in the Boonstra (2002) study contradict the results from
the modeling of exposure and effects line of evidence. However, the results of the supplemental
analyses of the data from the Boonstra study (2002) on survival of short-tailed shrews are in
agreement with the modeling results, suggesting that there is a high potential for adverse effects
from exposure to COCs in the contaminated areas of the PSA.
Population dynamics of mammals are affected by processes such as growth, reproduction, death
of predators, immigration, and emigration. As a result, a number of mechanisms exist to
MK01|O:\20123001.096\ERA_PB\ERA_PB_10.DOC ia 7/10/2003
-------
1 possibly compensate for the adverse effects of a toxic chemical. For example, a toxic chemical
2 may lead to an increase of a mammal population by reducing abundance of competitors or by
3 eliminating predators. Other mechanisms could also compensate for the direct effects of a toxic
4 chemical (e.g., increased immigration from uncontaminated sites). In the Housatonic River PSA,
5 such compensating mechanisms could exist for the local populations of short-tailed shrew, and
6 red fox. Thus, a possible explanation for the lack of concordance between the field survey
7 results, the Boonstra (2002) field study results, the additional analyses on survival of shrew, and
8 the modeling results is that other mechanisms (e.g., reduced competition, elimination of
9 predators) compensated for the direct effects due to tPCBs. No information, however, is
10 available to support or test this supposition.
11 Other omnivorous and carnivorous species common to the area include smoky shrews, masked
12 shrews, coyotes, gray fox, fishers, short-tailed weasels, and long-tailed weasels (see Appendix
13 A). Exposure and sensitivity to COCs are the two factors used to estimate risk to omnivorous
14 and carnivorous mammals. As noted in this ERA, effects studies conducted on short-tailed
15 shrew and red fox are not available. Similarly, effects data are not available for other
16 omnivorous and carnivorous species living in the Housatonic River area. As a result, the same
17 surrogate effects data used to estimate effects to short-tailed shrew and red fox would be used for
18 other omnivorous and carnivorous species. A qualitative analysis was conducted to compare
19 exposure of representative species and other omnivorous and carnivorous mammals to tPCBs
20 and TEQ. The major factors that influence mammalian exposure to tPCBs and TEQ include the
21 following:
22 ¦ Foraging behavior and dietary composition.
23 ¦ Foraging and home range size.
24 ¦ Species body weight and other life history characteristics.
25
26 Representative species and other mammal species were compared using these factors. Results
27 are provided in the text box.
28
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-66
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
ERA Summary
The weight-of-evidence analysis indicates a potential risk for short-tailed shrews
exposed to tPCBs and TEQ in the PSA. This conclusion, however, is uncertain
because of the uncertainty about whether effects are occurring in two of the lines of
evidence (i.e., field survey, population demography field study).
Risk to carnivorous mammals, such as red fox, exposed to tPCBs and TEQ are
undetermined in the PSA.
Other omnivorous and carnivorous mammal species common to the PSA include
smoky shrew, masked shrew, coyote, gray fox, fisher, short-tailed weasel, and long-
tailed weasel. A qualitative analysis of risk on these species indicates that smoky
shrew, short-tailed weasel, and long-tailed weasel have higher levels of risk; masked
shrew, gray fox, and fisher have similar levels of risk; and coyote has a lower level of
risk compared to the representative species.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-67
-------
1 10.6 REFERENCES
2 Abies, E.D. 1974. Ecology of the red fox in North America. In: Fox, M.W. ed. The Wild Canids.
3 Van Nostrand Reinhold, New York, NY. pp 148-163.
4 Allen, J.R. and D.A. Barsotti. 1976. The effects of transplacental and mammary movement of
5 PCBs on infant rhesus monkeys. Toxicology 6:331-340.
6 Barret, G.W. and K.L. Stueck. 1976. Caloric ingestion rate and assimilation efficiency of the
7 short-tailed shrew, Blarina brevicauda. Ohio J. Sci. 76:25-26.
8 Bell, G.P. 1990. Birds and mammals on an insect diet: A primer on diet composition analysis in
9 relation to ecological energetics. Studies in Avian Biology 13:391-415. Cited in EPA 1993.
10 Bjerke, D.L. and R.E. Peterson. 1994. Reproductive toxicity of 2,3,7,8-tetrachlorodibenzo-p-
11 dioxin in male rats: different effects of in utero versus lactational exposure. Toxicol. Appl.
12 Pharmacol. 127:241-249.
13 Bjerke, D.L., R.J. Sommer, R.W. Moore and R.E. Peterson. 1994. Effects of in utero and
14 lactational 2,3,7,8-tetrachlorodibenzo-p-dioxin exposure on responsiveness of the male rat
15 reproductive system to testosterone stimulation in adulthood. Toxicol. Appl. Pharmacol.
16 127:250-257.
17 Bleavins, M.R., R.J. Aulerich and R.K. Ringer. 1981. Placental and mammary transfer of
18 poly chlorinated and polybrominated biphenyl in the mink and ferret. In Avian and mammalian
19 wildlife toxicology. D.W. Lamb and E.E. Kenaga, Editors, pp. 121-131. Second conference,
20 ASTM STP 757. American Society for Testing and Materials, 1916 Race Street, Philadelphia,
21 PA.
22 Boonstra, R. 2002. Demography of short-tailed shrew population living on PCB contaminated
23 sites. Division of Life Sciences, University of Toronto at Scarborough, Ontario, Canada.
24 Borga, K., G.W. Gabrielsen, and J.U. Skaare. 2001. Biomagnification of organochlorines along
25 a Barents sea food chain. Environ. Pollut. 113(2): 187-198.
26 Bosveld, A.T.C. and M. van den Berg. 1994. Effects of polychlorinated biphenyls, dibenzo-p-
27 dioxins and dibenzofurans on fish eating birds. Environmental Reviews 2:147-165.
28 Brezner, E., J. Terkel and A.S. Perry. 1984. The effect of Aroclor 1254 (PCB) on the physiology
29 of reproduction in the female rat-I. Comp. Biochem. Physiol. 77C:65-70.
30 Bruckner, J.V., K.L. Khanna, and H.H. Cornish. 1973. Biological responses of the rat to
31 polychlorinated biphenyls. Toxicol. Appl. Pharmacol. 24:434-448.
32 Bryant, D.M. and V. M. Bryant. 1988. Assimilation efficiency and growth of nestling
33 insectivores. Arch. Hydrobiol. 130:268-274.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC 10 68 7/10/2003
-------
1 Burt, W.H., and R.P. Grossenheider. 1980. A field guide to the mammals of North America north
2 of Mexico. Houghton Mifflin Co., Boston, MA.
3 Castro, G., N. Stoyan, and J.P. Myers. 1989. Assimilation efficiencies in birds: A function of
4 taxon or food type? Comp. Biochem. Physiol. 92:271-278.
5 Collopy, M.W. 1975. Behavioral and predatory dynamics of kestrels wintering in the Areata
6 Bottoms [master's thesis], Humboldt State University, Areata, CA.
7 Conaway, C.H. 1955. Embryo resorption and placental scar formation in the rat. J. Mammal.
8 36:516-532.
9
10 Corbet, R.L., D.G. Muir, and G.R.B. Webster. 1983. Fate of carbon-14 labeled 1,3,6,8-
11 tetrachloro-dibenzo-p-dioxin in an outdoor aquatic system. Chemosphere 12:523-528.
12 Cummins, K.W. and J.C. Wuycheck. 1971. Caloric equivalents for investigations in ecological
13 energetics. International Association of Theoretical and Applied Limnology. Stuttgart, West
14 Germany, Cited in EPA 1993.
15 d'Argy, R., L. Dencker, E. Klasson-Wehler, A Bergman, P.O. Darnerud, and I. Brandt. 1987.
16 3,3',4,4'-tetrachlorobiphenyl in pregnant mice: Embryotoxicity, teratogenicity and toxic effects
17 on the cultured embryonic thymus. Pharm. Toxicol. 61:53-57.
18 DeGraaf, R.M., and M. Yamasaki. 2001. New England Wildlife: Habitat, Natural History, and
19 Distribution. University Press of New England, Hanover, NH.
20 DiBello, F., S. Arthur, and W. Krohn. 1990. Food habits of sympatric coyotes, red foxes, and
21 bobcats in Maine. Canadian Field Naturalist 104:403-408.
22 Eadie, W.R. 1944. The short-tailed shrew and field mouse predation. Journal of Mammalogy
23 25:359-364.
24 Eadie, W.R. 1948. Shrew-mouse predation during low mouse abundance. Journal of
25 Mammalogy 2935-31.
26 EPA (U.S. Environmental Protection Agency). 1993. Wildlife Exposure Factors Handbook.
27 Volumes I and II. Office of Research and Development. Washington, DC. EPA EPA/600/R-
28 93/187a, EPA/600/R-93/187b. (99-0040)
29 Flaws, J.A., R.J. Sommer, E.K. Silbergeld. R.E. Peterson, and A. Hirshfield. 1997. In utero and
30 lactational exposure to 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) induces genital
31 dysmorphogenesis in the female rat. Toxicol. Appl. Pharmacol. 147:351-362.
32 French, Jr., J.B., M.B. Voltura, and T.E. Tomasi. 2001. Effects of pre- and postnatal
33 polychlorinated biphenyl exposure on metabolic rate and thyroid hormones of white-footed mice.
34 Environmental Toxicology and Chemistry 20:1704-1708.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-69
-------
1 George, S.B., J.R. Choate, and H.H. Genoways. 1986. Blarina brevicauda. p. 1-9. In T.E. Lawlor
2 (ed.) Mammalian Species Number 261. American Society of Mammalogists, Baltimore, MD.
3 Giavini, E., M. Prati, and C. Vismara. 1983. Embryonic effects of 2,3,7,8-tetrachlorodibenzo-p-
4 dioxin administered to female rats before mating. Environ. Res. 31:105-110.
5 Gray, L.E. Jr. and J.S. Ostby. 1995. In utero 2,3,7,8-tetrachrlorodibenzo-p-dioxin (TCDD) alters
6 reproductive morphology and function in female rat offspring. Toxicol. Appl. Pharmacol.
7 133:285-294.
8 Gray, L.E. Jr., W.R. Kelce, E. Monosson, J.S. Ostby and L.S. Birnbaum. 1995. Exposure to
9 TCDD during development permanently alters reproductive function in male Long Evans rats
10 and hamsters: Reduced ejaculated and epididymal sperm numbers and sex accessory weights in
11 offspring with normal androgenic status. Toxicol. Appl. Pharmacol. 131:108-118.
12 Green, S., F.M. Sauro and L. Friedman. 1975. Lack of dominant lethality in rats treated with
13 polychlorinated biphenyls (Aroclors 1242 and 1254). Fd. Cosmet. Toxicol. 13:507-510.
14 Haffner, G.D., M. Tomczak, and R. Lazar. 1994. Organic contaminant exposure in the Lake St.
15 Clair food web. Hydrobiologia 281:19-27.
16 Hamilton, W.J., Jr. 1941. The foods of small forest mammals in eastern United States. Journal of
17 Mammalogy 22:250-263.
18 Harrison, S.J., J.A. Bissonnette and J.A. Sherburne. 1989. Spatial relationships between coyotes
19 and red foxes in eastern Maine. J.Wild.Manag.53.1%1-1%5.
20 Howard, P.H., R.S. Boethling, W.F. Jarvis, W.M. Meylan, and E.M. Michalenko. 1991.
21 Handbook of environmental degradation rates. Lewis Publishers. Chelsea, Michigan.
22 Hensel, R.J., W.A. Troyer and A.W. Erickson. 1969. Reproduction in the female brown bear. J.
23 Wildl. Manage. 33:357-365.
24 Huang, Y-W. M.J. Melancon, R.E. Jung, and W.H. Karasov. 1998. Induction of cytochrome
25 P450-associated monooxygenases in northern leopard frogs, Rana pipiens, by 3,3',4,4',5-
26 pentachlorobiphenyl. Environ. Toxicol. Chem. 17: 1564-1569.
27 Huuskonen, H., M. Unkila, R. Pohjanvirta and J. Tuomisto. 1994. Developmental toxicity of
28 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) in the most TCDD-resistant and -susceptible rat
29 strains. Toxicol. Appl. Pharmacol. 124:174-180.
30 Keenan, R.J. 1981. Spatial use of home range among red foxes (Vulpes vulpes) in south-central
31 Ontario. In: Chapman, J.A.; Pursley, D. eds. Worldwide furbearer conference proceedings,
32 August 1980, Frostburg, Maryland, pp 1041-1063.
33 Khera, K.S. and J.A. Ruddick. 1973. Polychlorodibenzo-p-dioxins: Perinatal effects and the
34 dominant lethal test in Wistar rats. Toxicol. 120:70-84.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC j q jq
-------
1 Kimbrough, R.D., R.E. Linder and T.B. Gaines. 1972. Morphological changes in livers of rats
2 fed polychlorinated biphenyls. Archives of Environmental Health 25:354-364.
3 Knable, A.E. 1974. Seasonal trends in the utilization of major food groups by the red fox (vulpes
4 fulva) in Union County, Illinois. Trans. III. State Acad. Sci. 66:113-115.
5 Kociba, R.J., D.G. Keyes, J.E. Beyer, R.M. Carreon, C.E. Wade, D A. Dittenber, R.P. Kalnins,
6 L.E. Frauson, C.N. Park, S.D. Barnard, R.A. Hummel, and C.G. Humiston. 1978. Results of a
7 two-year chronic study and oncogenicity study of 2,3,7,8-tetrachlorodibenzo-p-dioxin in rats.
8 Toxicol. Appl. Pharmacol. 46:279-303.
9 Kolb, H. H. 1986. Some observations on the home ranges of vixens in the suburbs of Edinburg.
10 Journal of Zoology 210:636-639.
11 Koplin, J.R., M.W. Collopy, A.R. Bammann and H. Levenson. 1980. Energetics of two
12 wintering raptors. Auk 97:795-806.
13 Korschgen, L.J. 1959. Food habits of the red fox in Missouri. J. Wild. Manage. 23:168-176.
14 Kuehl, D.W., P.M. Cook, A.R. Batterman, and B.C. Butterworth. 1987. Isomer dependent
15 bioavailability of polychlorinated dibenzo-p-dioxins and dibenzofurans from municipal
16 incinerator fly ash to carp. Chemosphere 16:657-666.
17 Lariviere, S. and M. Pasitschniak-Arts. 1996. Vulpes vulpes. Mammalian Species No. 537.
18 American Society of Mammalogist. 11pp.
19 Linder, R.E., T.B. Gaines, and R. Kimbrough. 1974. The effect of polychlorinated biphenyls on
20 rat reproduction. Fd. Cosmet. Toxicol. 12:63-77.
21 Linzey, A.V. 1987. Effects of chronic polychlorinated biphenyls exposure on reproductive
22 success of white-footed mice (Peromyscus leucopus). Arch. Environ. Contam. Toxicol. 16:455-
23 460.
24 Linzey, A.V. 1988. Effects of chronic polychlorinated biphenyls exposure on growth and
25 reproduction of second-generation white-footed mice (Peromyscus leucopus). Arch. Environ.
26 Contam. Toxicol. 17:39-45.
27 Linzey, D.W., and A.V. Linzey. 1973. Notes on food of small mammals from Great Smokey
28 Mountains National Park, Tennessee-North Carolina. J. Elisha Mitchell Science Society 89:6-14.
29 Mably, T.A., R.W. Moore, and R.E. Peterson. 1992. In utero and lactational exposure of male
30 rats to 2,3,7,8-tetrachlorodibenzo-p-dioxin. 1. Effects on androgenic status. Toxicol. Appl.
31 Pharmacol. 144:97-107.
32 Marks, T.A., G.L. Kimmel and R.E. Staples. 1989. Influence of symmetrical polychlorinated
33 biphenyl isomers on embryo and fetal development in mice. Fund. Appl. Tox. 13:681-693.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-71
-------
1 Masuda, Y., R. Kagawa, H. Kuroki, S. Tokudome and M. Kuratsune. 1979. Transfer of various
2 polychlorinated biphenyls to the fetuses and offspring of mice. I'd. Cosmet. Toxicol. 17:623-627.
3 McCoy, G., M.F. Finlay, A. Rhone, K. James, and G.P. Cobb. 1995. Chronic polychlorinated
4 biphenyls exposure on three generations of oldfield mice (Peromyscus polionotus): Effects on
5 reproduction, growth, and body burdens. Arch. Environ. Contam. Toxicol. 28:431-435.
6 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S. Petron,
7 B. Potocki, S. Svirsky and P. Tyler. 1996. Special report of the Massachusetts weight-of -
8 evidence workgroup: A weight-of-evidence approach for evaluating ecological risks. Human Ecol.
9 Risk Assess. 2:277-304.
10 Merson, M.H. and R.L. Kirkpatrick. 1976. Reproductive performance of captive white-footed
11 micefedaPCB. Bull. Environ. Contam. Toxicol. 16(4):392-398.
12 Miller, H. and L.L. Getz. 1977. Factors influencing local distribution and species diversity of
13 forest small mammals in New England. Can. J. Zool. 55:806-814.
14 Moore, D.R.J., B.E. Sample, G.W. Suter, B.R. Parkhurst and R.S. Teed. 1999. A probabilistic
15 risk assessment of the effects of methylmercury and PCBs on mink and kingfishers along East
16 Fork Poplar Creek, Oak Ridge, Tennessee, USA. Environ. Toxicol. Chem. 18: 2941-2953.
17 Morrison, P.R., M. Pierce and F.A. Ryser. 1957. Food consumption and body weight in the
18 masked and short-tailed shrews (genus Blarina) in Kansas, Iowa, and Missouri. Ann. Carnegie
19 Mus. 51 : 157-180.
20 Muir, D.C.G., R.J. Nostrom, and M. Simon. 1988. Organochlorine contaminants in Arctic
21 marine food chains: Accumulation of specific polychlorinated biphenyls and chlordane-related
22 compounds. Environmental Science and Technology 22:1071-1079.
23 Mumford, R.E., and J.O. Whitaker Jr. 1982. Mammals of Indiana. Indiana University Press,
24 Bloomington, IN.
25 Murray, F.J., F.A. Smith, K.D. Nitschke, C.G. Humiston, R.J. Kociba, and B.A. Schwetz. 1979.
26 Three-generation reproduction study in rats given 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD)
27 in diet. Toxicol. Appl. Pharmacol. 50:241-252.
28 Nagy K.A. 1987. Free metabolic rate and food requirement scaling in mammals and birds.
29 Ecological Monographs 57:111-128.
30 Nagy, K.A., I. A. Girard, and T.K. Brown. 1999. Energetics of free-ranking mammal, reptiles,
31 and birds. Annu. Rev. Nutr. 19:247-277.
32 Neubert, D. and I. Dillmann. 1972. Embryonic effects in mice treated with 2,4,5-
33 trichlorophenoxyacetic acid and 2,3,7,8-tetrachlorodibenzo-p-dioxin. Arch. Pharmacol. 272:243-
34 264.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-72
-------
1 Newsted, J.L., J.P. Giesy, G.T. Ankley, D.E. Tillitt, R.A. Crawford, J.W. Gooch, P.D. Jones, and
2 M.S. Denison. 1995. Development of toxic equivalency factors for PCB congeners and the
3 assessment of TCDD and PCB mixtures in rainbow trout. Environmental Toxicology and
4 Chemistry 14(5): 861 -871
5 Nixon, C.M., M.W. McClain and R.W. Donohoe. 1975. Effects of hunting and mast crops on a
6 squirrel population. J. Wildl. Manage. 39:1-25.
7 NRCC (National Research Council of Canada). 1981. Poly chlorinated dibenzo-p-dioxins:
8 Criteria for their effects on man and his environment. Publication NRCC No. 18574. National
9 Research Council of Canada. Ottawa, Ontario. 251 p.
10 Oleyar, C.M. and B.S. McGiness. 1974. Field evaluation of diethulstillbestrol for suppressing
11 reproduction in foxes. J. Wildl. Manage. 38:101-106.
12 Overmann, S.R., J. Kostas, L.R. Wilson, W. Shain and B. Bush. 1987. Neurobehavioral and
13 somantic effects of perinatal PCB exposure in rats. Environ. Res. 44:56-70.
14 Piatt, W. J. 1976. The social organization and territoriality of short-tailed shrew (Blarina
15 brevicauda) populations in old-field habitats. Anim. Behav. 24: 305-318.
16 Pohjanvirta, R., M. Unkila and J. Tuomisto. 1993. Comparative acute lethality of 2,3,7,8-
17 tetrachlorodibenzo-p-dioxin (TCDD), 1,2,3,7,8-pentachlorodibenzo-p-dioxin and 1,2,3,4,7,8-
18 hexachlorodibenzo-p-dioxin in the most TCDD susceptible and the most TCDD-resistant rat
19 strain. Pharmacol, and Toxicol. 73:52-56.
20 Rands, P.L., R.D. White, M.W. Carter, S.D. Allen, and W.S. Bradshaw. 1982a. Indicators of
21 developmental toxicity following prenatal administration of hormonally active compounds in the
22 rat. I. Gestational length. Teratology 25:37-43.
23 Rands, P.L., C.L. Newhouse, J.L. Stewart, and W.S. Bradshaw. 1982b. Indicators of
24 developmental toxicity following prenatal administration of hormonally active compounds in the
25 rat. II. Pattern of maternal weight gain. Teratology 25:45-51.
26 Safe, S. 1984. Polychlorinated biphenyls (PCBs) and polybrominated biphenyls (PBBs):
27 Biochemistry, toxicology, and mechanism of action. CRC Critical Reviews in Toxicology 13:
28 319-395.
29 Samuel, D.E. and B.B. Nelson. 1982. In: Chapman, J. A., Feldhammer, G.A. eds. Wild Mammals
30 of North America. Johns Hopkins University Press, Baltimore, MD. pp 475-490.
31 Sanderson, G.C. 1950. Methods of measuring productivity in raccoons. J. Wildl. Manage.
32 14:389-402.
33 Sargeant, A.B. 1978. Red fox prey demands and implications to prairie duck production. J.
34 Wildl. Manage. 42(3):520-527.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-
-------
1 Sargeant, A. B., S. H. Allen, and R. T. Eberhardt. 1984. Red fox predation on breeding ducks in
2 mid-continent North America. Wildlife Monographs 89:1-41.
3 Senthilkumar, K., Iseki, N., Hayama, S., Nakanishi, J., and Masunaga, S. 2001. Polychlorinated
4 Dibenzo-p-Dioxins, Dibenzofurans, and Dioxin-like Polychlorinated Biphenyls in Livers of
5 Birds from Japan. Arch. Environ. Contam. Toxicol. 42:244-255.
6 Sparschu, G.L., F.L. Dunn, and V.K. Rowe. 1971. Study of the Teratogenicity of 2,3,7,8-
7 tetrachlorodibenzo-p-dioxin in the rat. Fd. Cosmet. Toxicol. 9:405-412.
8 Spencer, F. 1982. An assessment of the reproductive toxic potential of Aroclor 1254 in female
9 Sprague Dawley rats. Bull. Environ. Contam. Tox. 28-290-297.
10 Stahl, B.U., A. Kettrup, and K. Rozman. 1992. Comparative toxicity of four chlorinated dibenzo-
11 p-dioxins (CDDs) and their mixture. Arch. Toxicol. 66All-All.
12 Storm, G.L., R.D. Andrews, R.L. Phillips, R.A. Bishop, D.B. Siniff and J.R. Tester. 1976.
13 Morphology, reproduction, dispersal and mortality of Midwestern red fox populations. Wild.
14 Monogr. 49, 82 pp.
15 Thayer, G.W., W.E. Schaaf, J.W. Angelovic, and M.W. LaCroix. 1973. Caloric measurements of
16 some estuarine organisms. Fishery Bull. 71: 289-296.
17 Thomas, P.T. and R.D. Hinsdill. 1979. The effect of perinatal exposure to tetrachlorodibenzo-p-
18 dioxin on the immune response of young mice. Drug and Chemical Toxicol. 2(l-2):77-98.
19 Tsushimoto, G., F. Matsumura, and R. Sago. 1982. Fate of 2,3,7,8-tetrachlorodibenzo-p-dioxin
20 (TCDD) in an outdoor pond and in model aquatic ecosystems. Environmental Toxicology and
21 Chemistry 1:61-68.
22 Van den Berg, M., L. Birnbaum, A.T.C. Bosveld, B. Brunstrom, P. Cook, M. Freely, J.P. Giesy,
23 A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C. Larsen, F.X. Rolan van Leeuwen,
24 A.K. Djien Liem, C. Nolt, R.E. Peterson, L. Poellinger, S. Safe, D. Schrenk, D. Twilit, M.
25 Tickling, M. Younest, F. Warn, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs)
26 for PCBs, PCDDs, PCDFs for humans and wildlife. Environmental Health Perspectives
27 106(12):775-792.
28 Voigt, D.R. 1987. Red fox In M. Novak, J. A. Baker, M. E. Obbard, and B. Mallock, eds., Wild
29 Furbearer Management and Conservation in North America, pp. 378-392. Toronto: Ontario
30 Ministry of Natural Resources, Toronto, Ontario.
31 Voigt, D.R. and J. Broadfoot. 1983. Locating pup-rearing dens of red foxes with radio-equipped
32 woodchucks. J. Wild. Manage. 47:858-859.
33 Voigt, D.R. and R.L. Tinline. 1980. Strategies for analyzing radio tracking data. In: Amlaner,
34 C. J., Jr. and D.W. MacDonald eds. A Handbook on Biotelemetry and Radio Tracking. Permagon
35 Press, Oxford, United Kingdom, pp. 387-404.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-74
-------
1 Wardell, R.E., R.E. Seegmiller, and W.S. Bradshaw. 1982. Induction of prenatal toxicity in the
2 rat by diethylstilbestrol, zeranol, 3,4,3',4'-tetrachlorobiphenyl, cadmium and lead. 1982.
3 Teratology 26:229-237.
4 Whitaker, J.O., Jr., and W.J. Hamilton Jr. 1998. Mammals of the Eastern United States. 3rd ed.
5 Cornell University Press, Ithaca, NY. 538 p.
6 Whitaker, O.J., Jr., and M.G. Ferraro. 1963. Summer food of 220 short-tailed shrews from
7 Ithaca, New York. Journal of Mammalogy 44:419.
8 Wrigley, R.E., J.E. Dubois, and H.W.R. Copland. 1979. Habitat, abundance, and distribution of
9 six species of shrews in Manitoba. Journal of Mammalogy 60:505-520.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_10.DOC
10-75
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
11. ASSESSMENT ENDPOINT—SURVIVAL, GROWTH, AND
REPRODUCTION OF THREATENED AND ENDANGERED
SPECIES
Highlights
Conceptual Model
The assessment endpoint is the survival, growth, and reproduction of T&E species in
the Housatonic River PSA. The measurement endpoints include comparisons to
doses reported in the literature to cause adverse effects and conducting field surveys
to determine the abundance of T&E species in the Housatonic River floodplain. T&E
species, including bald eagle, American bittern, and small-footed myotis, selected as
representative species for the ERA, are exposed to these COCs via diet and trophic
transfer.
Exposure
Exposure of the representative species to tPCBs and TEQ was determined from
concentrations found in prey items and an estimation of the daily intake of these
COCs from consumption of prey.
Effects
Limited data were available on the toxicity of tPCBs and TEQ to bald eagle,
American bittern, and small-footed myotis. Surrogate species were used to develop
toxicity thresholds for bald eagle and American bittern. Sufficient surrogate data
were available to generate effects dose-response curves for small-footed myotis.
Risk
Bald eagle, American bittern, and small-footed myotis are at risk as a result of
exposure to tPCBs and TEQ in the Housatonic PSA. In particular, bald eagles are at
high risk in the PSA. Other similar, but not T&E, species common to the PSA have
either higher levels of risk (e.g., least bittern, green heron); similar levels of risk (e.g.,
great blue heron, Indiana bat, little brown bat); or a lower level of risk (e.g., sora)
compared to the representative species.
11.1 INTRODUCTION
The purpose of this section of the ecological risk assessment (ERA) is to characterize and
quantify the current and potential risks posed to rare, threatened, and endangered (T&E) species
exposed to contaminants of potential concern (COPCs) in the Housatonic River and floodplain,
focusing on total polychlorinated biphenyls (tPCBs) and other COPCs originating from the
General Electric Company (GE) facility in Pittsfield, MA. The Housatonic River watershed is
located in western Massachusetts and Connecticut, discharging to Long Island Sound, with the
GE facility located near the headwaters of the watershed. The Primary Study Area (PSA)
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC -> -> -> 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
includes the river and 10-year floodplain from the confluence of the East and West Branches of
the Housatonic River downstream of the GE facility to Woods Pond Dam (Figure 1.1-2).
A pre-ERA was conducted to narrow the scope of the ERA by identifying contaminants, other
than tPCBs, that pose potential risks to aquatic biota and wildlife in the PSA (Appendix B). A
three-tiered deterministic approach was used to screen COPCs. The deterministic assessments
compared potential conservative estimates of exposure with conservative adverse effects
benchmarks to identify which contaminants are of potential concern to T&E species in the
Housatonic River. A risk quotient (total daily intake/effect benchmark) for T&E species greater
than 1 in the Housatonic River area resulted in the COPC being screened through to the next tier
assessment, and to the probabilistic ERA, if necessary.
In summary, the contaminants of concern (COCs) that were retained in the probabilistic risk
assessment for T&E species were tPCBs and 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-
TCDD) toxic equivalence (TEQ). Total PCBs detected in Housatonic River media samples
closely resemble the commercial PCB mixtures Aroclor 1260 and Aroclor 1254, which are
similar in congener makeup. TEQ is calculated from coplanar PCB and dioxin and furan
congeners using the toxic equivalency factor (TEF) approach developed by Van den Berg et al.
(1998) (see Section 6.4 of the ERA).
11.1.1 Overview of Approach
A step-wise approach was used to assess the risks of tPCBs and TEQ to T&E species in the
Housatonic River watershed. The four main steps in this process include:
1. Derivation of a conceptual model (Figure 11.1-1).
2. Assessment of exposure of T&E species to COCs (Figure 11.1-2).
3. Assessment of the effects of COCs on T&E species (Figure 11.1-3).
4. Characterization of risks to the T&E species community (Figure 11.1-4).
The detailed ecological risk assessment for T&E species is provided in Appendix K. I
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC H 2
-------
1
2
3
in
0)
in
re
a>
UL
(A
a>
o
L_
3
O
<0
o
a>
a:
Historic releases of PCBs from the
Pittsfield GE Facility and
surrounding disposal areas
Wetland and
surface water
discharge
Legend
Direct uptake
^ Trophic Transfer
-~ Potential Effects
Groundwater
discharge
Contaminated soil, sediment, water and biota
Terrestrial Vegetation
Plants, fungi
Terrestrial Invertebrates
Earthworms, snails, insects
Mammals
Shrews, mice,
voles
Aguatic Vegetation
Plants, algae
Aguatic Invertebrates
Insects, crayfish, mussels
Amphibians
Birds
Waterfowl, wading
birds, seagulls
T & E Species
Bald Eagle
T & E Species
American Bittern
T & E Species
Small-footed Myotis
Decreased Survival, Growth, or Reproduction
Figure 11.1-1 Conceptual Model Diagram: Exposure Pathways for T&E Species
Exposed to COCs in the Housatonic PSA
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 3
-------
Exposure
2
3 Figure 11.1-2 Approach Used to Assess Modeled Exposure of T&E Species to
4 COCs
5
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC H 4
-------
Effects
2
3 Figure 11.1-3 Approach Used to Assess the Modeled Effects of COCs to T&E
4 Species
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC H 5
-------
Risk Characterization
2
* Qualitative survey designs; therefore, not considered in the Weight-of-Evidence Analysis.
3 Figure 11.1-4 Approach Used to Characterize Risks of COCs to T&E Species
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 6
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
11.1.2 Conceptual Model
The conceptual model presented in Figure 11.1-1 illustrates the exposure pathways for T&E
species exposed to tPCBs and TEQ in the PSA. Total PCBs and TEQ are persistent and highly
hydrophobic and lipophilic. Therefore, they are bioaccumulated by aquatic and terrestrial biota
directly through the consumption of contaminated prey as part of the food chain (Haffner et al.
1994; Senthilkumar et al. 2001; Borga et al. 2001). Fish, amphibians, invertebrates, mammals,
and birds comprise the major dietary items for T&E species. In summary, T&E species that
reside, or partially reside, within the study area are exposed to tPCBs and TEQ principally
through diet and trophic transfer. Other routes of exposure, considered to be less important to
overall exposure, include inhalation, water consumption, and sediment ingestion (Moore et al.
1999).
The problem formulation (Section 2) identified the bald eagle (Haliaeetus leucocephalus),
American bittern (Botaurus lentiginosus), and small-footed myotis (Myotis leibii) as the
representative T&E species potentially exposed to tPCBs and TEQ from consumption of
contaminated prey. American bitterns have been observed during the breeding season in suitable
nesting habitat; therefore, they were chosen for inclusion because of the potential for nesting.
Similarly, bald eagles nest downstream, have attempted to nest in the PSA, and have ample
habitat available for nesting in the PSA. Small-footed myotis may occur in the PSA as well
because of their known range and the suitability of habitat. Life history profiles for the bald
eagle, American bittern, and small-footed myotis are presented in the following text boxes.
Additional life history information on these species is provided in Appendix K, Sections K.2.1.5,
K.2.1.6, and K.2.1.7, respectively.
The assessment endpoint, which is the subject of this section, is the survival, growth, and
reproduction of T&E species in the Housatonic River PSA. The potential lines of evidence
considered in the evaluation of the assessment endpoint included (1) comparing modeled
exposure to doses of tPCBs and TEQ ingested in the diet reported in the literature to cause
adverse effects to the survival, reproduction, or growth of omnivorous and carnivorous
mammals; and (2) determining, by conducting field surveys, the qualitative abundance of T&E
species in the Housatonic River floodplain.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 7
-------
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Photo by: Karol Worden
Life History of Bald Eagle
The bald eagle is one of the largest and most conspicuous birds of prey in North
America. Weights of adults and juveniles vary from 3 kg to over 7 kg. The bald
eagle is currently federally listed as Threatened in all of the 48 lower states, but is
more restrictively listed as Endangered by several New England states, including
Massachusetts, Vermont, New Hampshire, and Connecticut.
¦ Habitat - Habitat use varies depending on the region, but proximity to large
bodies of water with suitable foraging opportunities is critical; thus bald eagles
are generally restricted to coastal areas, lakes, and rivers. Relatively open
canopies, some type of habitat edge, and the availability of super-story trees with
stout horizontal perching branches providing good access to nests are preferred
habitat features for breeding pairs.
¦ Home Range -Large home ranges, minimum size of 1,730 acres (700 ha) and
an average size of 4,645 ± 2,224 acres (1,879.76 ± 900.02 ha); linear (riverine)
foraging distances of 1.9 to 4.3 miles (3.1 to 6.9 km). Nesting bald eagles were
reported to generally forage within 0.3 mile (0.5 km) of the nest, ranging up to 1.9
miles (3.0 km), and as far as 5.0 miles (8.0 km) from their nest.
¦ Dietary Habits -Feed primarily over water on aquatic prey; opportunistic
feeders, consuming a variety of live prey and scavenging carrion. Fish taken
primarily from shallow water form the largest percentage of diet. Fish
consumption is 17.1% to 90.1%, depending on location, season, and prey
availability. Birds, particularly waterfowl, can form large portions of the diet, more
commonly during the winter and in coastal habitats. Mammal species average
4.9% of prey; but are reported to be as much as 11.7%, or as little as 0%.
M K01 \O:\20123001,096\ERA_PB\ERA_PB_11. DOC 1 1 8
-------
1
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
Photo by: Scott Robinson
Life History of American Bittern
The American bittern is a mid-sized, stocky heron of freshwater marshes, it is
identified by its heavily streaked breast with vertical brown and white stripes below.
American bittern populations have been declining since the 1960s primarily as a
result of habitat loss and wetland degradation. The Commonwealth of
Massachusetts has included the American bittern on its list of Endangered species.
¦ Habitat - Use a wide range of freshwater wetlands with diversity of vegetation
classes (i.e., aquatic bed, emergent, and scrub-shrub) and high interspersion of
open water and plant cover.
¦ Home Range - Varies with geographic area and availability of preferred habitat
and prey species. Average home ranges of 315 acres (127 ha) in Minnesota,
with the birds using a 61-acre (25 ha) core area more than 50% of the time. In
Massachusetts breeding occurs in scattered localities in Berkshire County.
Nests built in dense emergent vegetation over water with depths ranging from 5
to 20 cm (2 to 8 inches), consisting of a 15- to 25-cm (6- to 10-inch) high platform
of reeds, sedges, or grasses bent down and lined with fine grasses.
¦ Dietary Habits - Prey upon insects, crayfish, amphibians, fish, and small
mammals. Insect prey consists primarily of adult and nymphal dragonflies, giant
waterbugs, water scorpions, water beetles, and grasshoppers. Fish species vary
with availability and include eels, catfish, pickerel, sunfish, suckers, killifish,
sticklebacks, and perch, typically from 10 to 100 mm in length.
M K01 \O:\20123001,096\ERA_PB\ERA_PB_11. DOC 1 1 9
-------
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
Photo Credit: Merlin Tuttle (http://www.dcnr.state.pa.us/wrcf/myopic.htm)
Life History of Small-Footed Myotis
The small-footed myotis is a small bat. It is identified by its golden brown fur and
black mask. The small-footed myotis is listed as a Species of Special Concern by
the Massachusetts Natural Heritage and Endangered Species Program (MNHESP)
(1984).
¦ Habitat - Use buildings, overhanging rocks, and caves as summer roosts and
maternity sites. Females and young roost in small (typically less than 20
individuals) maternity colonies in rock crevices and crevice-like places on
buildings; males are solitary. Hibernate hanging from walls or underneath fallen
rock and rubble from November to March, usually in the foothills of mountains up
to 610 m (2,000 feet) in elevation, in coniferous woodlands.
¦ Home Range - Home range is unknown. It is assumed that home ranges are
similar to other Myotis species (Indiana bat that has a home range of 52 to 95 ha
for pregnant and lactating females).
¦ Dietary Habits - Little is known about feeding habits; however, believed to be
similar to other Myotis species. Fiies, beetles, bugs, leafhoppers, and flying ants
have been found in their stomachs. Many species are opportunistic feeders,
exploiting available food resources. They fly low to the ground (1 to 3 m) when
feeding, along forest openings, including waterways.
M K01 \O:\20123001,096\ERA_PB\ERA_PB_11. DOC j j j q
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
11.1.3 Organization
This section is organized as follows:
¦ Section 11.2 (Exposure Assessment) - Describes the exposure model, input
variables, and techniques to propagate uncertainty. This section also presents the
exposure modeling results for T&E species.
¦ Section 11.3 (Effects Assessment) - Describes the effects to T&E species exposed to
tPCBs and TEQ and summarizes the ranges of benchmarks (toxicity thresholds)
derived from the literature.
¦ Section 11.4 (Risk Characterization) - Integrates the exposure and effects
assessments, and makes conclusions regarding risk for T&E species in the Housatonic
River PSA using the two potential lines of evidence. A discussion of the sources of
uncertainty regarding risk estimates follows. This section also presents an
extrapolation of risks beyond the PSA to areas downstream of Woods Pond.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC j j j j
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
11.2 EXPOSURE ASSESSMENT
The exposure of T&E species to tPCBs and TEQ in the Housatonic River PSA is estimated in
this section. The representative T&E species are the bald eagle, American bittern, and small-
footed myotis. These species are T&E species that occur in the PSA, potentially breed within the
PSA, and feed on prey exposed directly to the COCs and through trophic transfer (see Appendix
A). Trophic transfer and exposure through ingestion of contaminated prey are the major
exposure pathways for T&E species exposed to tPCBs and TEQ. Other routes of exposure,
considered to be negligible contributors to overall exposure, include inhalation, water
consumption, and sediment ingestion (Moore et al. 1999).
All of the bald eagle sightings in the PSA occurred south of New Lenox Road, primarily at
Woods Pond and the backwaters north of Woods Pond (Appendix A). Bald eagles would not be
expected to regularly utilize the more shallow and narrow northern sections of the river.
Therefore, the exposure area assumed for bald eagles was the southern portion of the PSA, from
the more downstream portion of Reach 5B to Woods Pond. This entire area was not subdivided
because individual bald eagles would likely forage throughout this area.
American bittern habitat occurs throughout the PSA, and bitterns have been observed from
Canoe Meadows Wildlife Sanctuary south to Woods Pond (Appendix A). In addition, home
range sizes and habitat requirements for this species are such that individuals forage
predominantly within one subreach. Therefore, for the PCB analyses, the PSA was split into
four reaches: Reach 5A, Reach 5B, Reach 5C, and Reaches 5D and 6 combined. For the TEQ
analyses, samples from the PSA were combined into one analysis because the smaller sample
sizes did not allow for statistically robust analyses to be conducted for each subreach.
Little is known about the home range size of the small-footed myotis. The Indiana bat (Myotis
sodalis), a similar Myotis species, has a range averaging 128 acres (52 ha), but this range may be
as large as 232 acres (94 ha) for lactating female bats (Kurta 1995; DeGraaf and Yamasaki
2001). The exposure area assumed for small-footed myotis was the entire PSA because of the
small number of dietary samples from each reach.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1112
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
This section begins with a description of the exposure model used for the representative species.
Subsequent sections describe the inputs used in the exposure analyses for each species. The
section concludes with a presentation of the results of the exposure analyses.
11.2.1 Exposure Model
Exposure of T&E species to tPCBs and TEQ was estimated using a total daily intake model
adapted from the Wildlife Exposure Factors Handbook (EPA 1993) and related publications.
The model used in the exposure analysis was:
TDI = FT ¦ FIR tC>-P< (Eq.l)
2=1
where
TDI = total daily intake (mg/kg bw/d tPCBs, ng/kg bw/d TEQ)
FIR = normalized food intake rate (kg/kg bw/d)
FT = foraging time in Primary Study Area (unitless)
Cj = concentration in z'th food item (mg/kg tPCBs, ng/kg TEQ)
Pi = proportion of the z'th food item in the diet (unitless)
The models consider the food intake rates (FIRs) of the representative species (FIR), the
concentrations of COCs in each food item (Ci), and the proportion of the diet accounted for by
that food item (Pj). For those input variables that are uncertain, variable, or both, distributions
are used rather than point estimates. Monte Carlo and probability bounds analyses are the
methods used to propagate uncertainties about input variables through the exposure model for
each COC. A description of these techniques and the methods used to parameterize input
variables is presented in Section 6.5. The results of the Monte Carlo analysis are used to
estimate the probability of exposure exceeding an effects threshold or doses that cause adverse
effects of differing magnitudes. The probability bounds analysis is conducted to determine how
uncertainty regarding the distributions of the input variables influences the estimated exposure
distribution. The results of these analyses are discussed in detail in Appendix K.
Two circumstances often arose when calculating a TEQ concentration in prey:
¦ Congener concentrations may be below the method detection limit (i.e., non-detects).
¦ Some congeners may not be resolved due to co-elution during analysis.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 13
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
An approach was developed to address these circumstances and is presented in Section 6.4 and
Appendix C.2. Briefly, congeners detected at or below the detection limit (DL) were included in
the TEQ calculations by investigating three options: first, setting the value for the congener equal
to zero (0); setting it to half the DL; and, finally, setting it equal to the DL. A comparison of the
results of this bounding analysis provides a description of the uncertainty surrounding the TEQ
value due to the concentrations of one or more congeners being below the detection limit. To
resolve the co-elution issue, the concentration of congeners that co-eluted with other congeners
were assumed to equal to the total concentration of the co-elutes (potential overestimate of TEQ
concentration) or zero (potential underestimate of TEQ concentration).
Input distributions to the exposure analyses were generally assigned as follows:
¦ Lognormal distributions for variables that were right skewed with a lower bound of
zero and no upper bound (e.g., amount of COC transferred from mother to offspring
via egg tissue).
¦ Beta distributions for variables bounded by zero and one (e.g., proportion of a prey
item in the diet).
¦ Normal distributions for variables that were symmetric and not bounded by one (e.g.,
body weight).
¦ Point estimates for minor variables or variables with low coefficients of variation.
In certain situations (e.g., poor fit of data), other distributions were fit to the data or other
approaches were used. To quantify uncertainty, two approaches were used as described in
Section 6.5.2, Monte Carlo Analysis and Probability Bounds Analysis. The distributions used in
the exposure analyses for bald eagle, American bittern, and small-footed myotis are shown in
Figures 11.2-1, 11.2-2, and 11.2-3. A brief description of these variables is provided below.
11.2.1.1 Input Variables
11.2.1.1.1 Body Weight (BW)
The typical weight of an adult bald eagle ranges from 3.0 kg to over 7.0 kg. Adult males average
4.13 kg and adult females average 5.4 kg (Dunning 1992; EPA 1993; Buehler 2000; Canadian
Wildlife Service 2000). For this risk assessment, the weight of female bald eagles was used
because the effects endpoint is reproductive impairment, and the female will have the greatest
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC -> -> -> A 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
effect on survival of the young through transfer of PCBs to the egg. The mean weight was 5.35
kg with a standard deviation of 0.40.
The typical weight of an adult American bittern ranges from 370 g to >800 g (Gibbs et al. 1992;
Dunning 1992). In the Monte Carlo analyses, the mean BW of 707 g with a standard deviation
of 183 was used (Dunning 1992).
Small-footed myotis typically weigh 5 to 7 g, although their weights can range from 3 to 8 g
(Kurta 1995). A mean BW of 6 g with a standard deviation of 0.7 g was used in the exposure
analyses for small-footed myotis.
11.2.1.1.2 Food Intake Rate (FIR)
In the EPA Wildlife Exposure Factors Handbook (EPA 1993), three studies were used to
determine FIR. The first study (Stalmaster and Gessaman 1982) used captive eagles (obtained
from a zoological garden), housed in 3x3x2.5-m chambers, in their feeding study. The second
study (Stalmaster and Gessaman 1984) estimated feeding rates by remote observation of the
amount of food consumed at feeding stations. Estimates were averages of the total estimated
food consumed by the total number of eagles observed feeding, and assumed that the eagles fed
exclusively at the stations, although the authors acknowledged that some birds fed elsewhere. In
the third study (Craig et al. 1988), Stalmaster and Gessaman's (1984) data were used to estimate
prey consumption. Because of the issues associated with estimating food consumption in each of
these studies, the FIR was derived from the estimated metabolic rate of free-living eagles using
data from Nagy (1987) and Nagy et al. (1999).
The FIR developed from the Nagy studies was compared with those from EPA (1993). The
measured FIRs reported in the Wildlife Exposure Factors Handbook (EPA 1993) are consistent
with the FIR distribution estimated from the allometric equation. Estimated values for free-
flying eagles from Connecticut (Craig et al. 1988) were 0.12 to 0.14 g/g bw/d, while the median
FIR from the allometric equation was 0.158 g/g bw/d.
Nagy (1987) and Nagy et al. (1999) derived allometric equations for estimating the free
metabolic rate (FMR) of free-living mammals in kilojoules per day using the following general
equation:
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC -> -> -> c 7/10/2003
-------
FMR Slope Term (a)
FMR Power Term (b)
0.2
0
Log (a)
Body Weight
Proportion of Bottom Fish in Diet
0.8
0.6
0 1000 2000 3000 4000 5000 6000 7000
Weight (g)
0.6
0.4
Proportion
Proportion of Predatory Fish in Diet
Proportion of Forage Fish in Diet
2
3
Proportion of Birds in Diet
0.05 0.1
0.25 0.3
Proportion of Mammals in Diet
0.02 0.04 0.06 0.08 0.1
Proportion Proportion
Figure 11.2-1 Input Distributions for the Exposure Modeling of Bald Eagle
MK01\O:\20123001,096\ERA PB\ERA PB 11.DOC
11-16
-------
FMR Slope Term (a)
FMR Power Term (b)
0.4 0.6
Log (a)
~ 0.6
S
S 0.4
o
£
0.2
0
0.2 0.4
0.6
b
Body Weight
0 200 400 600 800 1000 1200 1400
Weight (g)
Figure 11.2-2 Input Distributions for the Exposure Modeling of American Bittern
FMR Slope Term (a)
FMR PowerTerm (b)
1 1.2 1.4
Body Weight
0.6
0.4
4 6
Weight (g)
8 10
Figure 11.2-3 Input Distributions for the Exposure Modeling of Small-Footed
Myotis
MK01\O:\20123001,096\ERA PB\ERA PB 11.DOC
11-17
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
FMR (kJ / d) = a ¦BW(g)
(Eq. 2)
where FMR (kcal/kg bw/d) is the free-living metabolic rate, and BW (g) is the body weight in
grams. The slope (a) and power (b) distributions were based on the error statistics reported in
Nagy et al. (1999). For birds, the equation was used and had a mean slope term (a) equal to 8.47
and a standard error of 1.57. The power term (b) had a reported mean of 0.768 and a standard
error of 0.087 (Nagy et al. 1999). For small-footed myotis, the insectivore equation was used.
The slope term (a) had a reported mean of 6.98 and a standard error of 4.19, and the power term
(b) had a reported mean of 0.622 and a standard error of 0.0630 (Nagy et al. 1999). The BW
distribution was described above.
FIR is derived from FMR using the following equation:
FMR
F1R=- (Eq. 3)
ZAE,-GE,
i=1
where AE is the assimilation efficiency of z'th food item (unitless) and GEj is the gross energy of
z'th food item (kcal/kg).
The gross energies of various wildlife food sources are summarized in the Wildlife Exposure
Factors Handbook (EPA 1993). The mean gross energy is 1.6 for invertebrates, 1.2 for fish and
amphibians, and 1.8 for birds and mammals.
The mean assimilation efficiency for fish and amphibians consumed by birds is 79%. For the
consumption of mammals by birds, the mean assimilation efficiency is 78%. For the
consumption of invertebrates by birds, the mean assimilation efficiency is 72%. Point estimates
were used for these variables in the Monte Carlo and probability bounds analyses because of
their relatively small coefficients of variation (i.e., CV<10%).
11.2.1.1.3 Proportions of Dietary Items (P,)
The proportions of prey items in bald eagle diets are listed in Table K.2-1. Studies of bald eagles
in habitat similar to the PSA have found mean fish consumption to be 77.5% (range 71.0 to 90.1)
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC j j jg
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
of the prey species taken during the breeding season. Bird species comprise on average 16.9% of
the bald eagle diet in habitats similar to the PSA, but can be as little as 7.9% and up to 26.1%.
Consumption of mammals averages 4.8% and can range from 1.3% to 11.7%. Reptiles make up
0.24% of the diet and can range from 0 to 0.6%. Invertebrates such as crayfish, crabs, and
mussels make up 0.12% of the diet on average and can range from 0 to 0.6% (Haywood and
Ohmart 1986; Dunstan and Harper 1975; Todd et al. 1982; Watson et al. 1991; Stratus 1999).
Reptiles and invertebrates were not included in the exposure analysis, however, because of their
small contribution to the overall diet. The proportion in the diet for the Monte Carlo analysis
was parameterized to allow the diet to equal 1. This resulted in a diet of 50.3% bottom fish,
16.1%) predatory fish, 11.8% forage fish, 16.3% birds, and 5% mammals (Table K.2-1). For the
probability bounds analysis, the minimum, mean, and maximum values were used as specified
above (Table K.2-3).
An analysis of the stomach contents of 160 individuals reported that the American bittern diet
consisted of invertebrates (23%), amphibians (21%), fish (21%), crayfish (19%), small mammals
(10%>), and snakes (5%) (Cottam and Uhler 1945, as cited in Gibbs et al. 1992). It was assumed
that American bitterns would consume the same proportion of prey items in each reach for this
assessment. Reptiles were not included in the exposure analysis, however, because of their small
contribution to the overall diet. The proportion in the diet for the Monte Carlo and probability
bounds analysis was parameterized to allow the diet to equal 1. This resulted in a diet of 24.5%
invertebrates, 20.2% macroinvertebrates (crayfish), 22.3% fish, and 22.3% amphibians (Table
K.2-10).
Along the Housatonic River, the small-footed myotis likely forages on small emergent aquatic
insects, as does the little brown bat. Adult little brown bats in New York were found to consume
Chironomidae (76.4% of food volume), Trichoptera (18.2%), Lepidoptera (4.2%), and
Coleoptera (1.2%) (Belwood and Fenton 1976). Other studies conducted in the northeast found
these to be commonly consumed species along with other Diptera, such as Tipulidae, Culicidae,
Homoptera, Hymenoptera, Neuroptera, Plecoptera, and Ephemeroptera (Griffith and Gates 1985;
Anthony and Kunz 1977). For this exposure assessment, the proportion of invertebrates in the
diet was assumed to be a point estimate, with invertebrates accounting for 100% of the small-
footed myotis diet.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC -> -> -> <~v 7/10/2003
-------
1 11.2.1.1.4 Foraging Time (FT)
2 Bald eagles nesting in the PSA would be expected to forage entirely within the PSA, as they
3 generally forage within 0.3 miles (0.5 km) of the nest, with a maximum reported foraging
4 distance of up to 5.0 miles (8.0 km) from their nest (Bowerman et al. 1995; Stratus 1999). As a
5 result, for the purpose of modeling COC exposure, it was assumed that bald eagles would spend
6 100% of their time foraging in the PSA, based on their feeding habits and availability of fish,
7 waterfowl, and mammals in the PSA.
8 American bittern nesting in the PSA would be expected to forage entirely within the PSA, as
9 they have territories averaging 315 acres of wetlands (DeGraaf and Yamasaki 2001). As a result,
10 it was assumed that American bitterns would spend 100% of their time foraging in the PSA. The
11 foraging time was specified as a point estimate.
12 Little is known about home range size of the small-footed myotis. The Indiana bat, a similar
13 Myotis species, has a range averaging 128 acres (52 ha), but this range may be as large as 232
14 acres (94 ha) for lactating female bats (Kurta 1995; DeGraaf and Yamasaki 2001). Small-footed
15 myotis feed predominantly over water on emergent insects; therefore, it was assumed that small-
16 footed myotis would forage 100% of the time in the PSA.
17 11.2.1.2 Concentrations of COCs in Prey
18 Fish, birds, and mammals are the major dietary items for bald eagles. The median concentration
19 of tPCBs in bottom feeding fish is 59.1 mg/kg (mean = 88.7 mg/kg), 64.8 mg/kg (mean = 79.1
20 mg/kg) for predatory fish, and 34.5 mg/kg (mean = 36.9 mg/kg) for forage fish. The median
21 concentration of tPCBs in birds is 6.09 mg/kg (mean = 7.18 mg/kg). The median concentration
22 of tPCBs in mammals is 4.98 mg/kg (mean = 28.2 mg/kg). The median, 25th and 75th percentile
23 concentrations of tPCBs and TEQ are presented in Figures 11.2-4 and 11.2-5.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 2Q
-------
130
120
110
100
90
i£
80
g1 70
V
£ 60
B
I 50
5 40
30
20
10
0
~ Bottom Fish (mean 88.7)
~ Predatory Fish (mean 79.1)
~ Forage Fish (mean 36.9)
M Birds (mean7.18 0
~ Mammals (mean 28.2)
ELj:
1
2
3
4
Figure 11.2-4 Median Concentrations of tPCBs in Prey of Bald Eagles
~ Bottom Fish (mean .00122)
~ Predatory Fish (mean .00131)
~ Forage Fish (mean .000647)
~ Birds (mean .00133)
~ Mammals (mean .000921)
5
6
Figure 11.2-5 Median Concentrations of TEQ in Prey of Bald Eagles
MK01\0:\20123001,096\ERA_PB\ERA_PB_11 .DOC 1121
-------
1
2
3
4
5
1
2
3
4
5
6
7
The diet for American bittern includes fish, amphibians, small mammals, invertebrates, and
macroinvertebrates. Similar statistics for concentrations of tPCBs and TEQ in American bittern
prey from Reaches 5A, 5B, 5C, and 5D and 6 are presented in Figures 11.2-6 and 11.2-7,
respectively.
"BU
S 30
S 20
£
~ Fish
~ Invertebrates
~ Crayfish
~ Amphibians
~ Mammals
Fish
Reach 5A, mean 43.03
Reach 5B, mean 27.68
Reach 5C, mean 22.62
Reach 5D&6, mean 25.43
Invertebrates
Reach 5 A, mean 10.51
Reach 5B, mean 30.56
Reach 5C, mean 21.37
Reach 5D&6, mean 9.081
Crayfish
Reach 5A, mean 21.67
Reach 5B, mean 12.28
Reach 5C, mean 8.223
Reach 5D&6, mean 6.799
Amphibians
Reach 5A, mean 2.774
Reach 5B, mean 2.375
Reach 5C, mean 2.586
Reach 5D&6, mean 4.242
Mammals
Reach 5 A, mean 28.21
Reach 5B, mean 28.21
Reach 5C, mean 28.21
Reach 5D&6, mean 28.21
Figure 11.2-6 Median Concentrations of tPCBs in Prey of American Bittern
~ Fish (mean .000556)
~ Invertebrates (mean .000720)
~ Crayfish (mean .000347)
¦ Amphibians (mean .0000576)
~ Mammals (mean .000921)
Figure 11.2-7 Median Concentrations of TEQ in Prey of American Bittern
MK01\0:\20123001,096\ERA_PB\ERA_PB_11 .DOC j j 22
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
Small-footed myotis prey items were not directly sampled in the PSA. Concentrations of tPCBs
in invertebrates were obtained from samples of tree swallow (Tachycineta bicolor) gut contents
(Custer 2002). Tree swallow gut content samples were used for small-footed myotis prey
because these samples contain prey species that are more representative of actual myotis prey
items than are the benthic invertebrate samples. Small-footed myotis and tree swallows are both
aerial insectivores that forage primarily over open water and consume similar types of
invertebrates; therefore, gut contents of tree swallows are likely to be similar to that of the small-
footed myotis. The median concentration of these samples was 7.10 mg/kg for tPCBs and 564
ng/kg for TEQ.
In the Monte Carlo analysis, it was assumed that the spatially and temporally averaged exposure
estimate did not vary between individuals foraging in the same area. Thus, the point estimate of
centrality was the minimum of:
1. The 95% upper confidence limit (UCL) calculated using the Land H-statistic
(assuming data are lognormally distributed), or
2. The maximum concentration measured. In the probability bounds analyses, however,
the uncertainty regarding the arithmetic mean was accounted for with a different
procedure.
The procedure generally involved using the Land H-statistic to estimate the lower and upper 95%
confidence limits on the mean (Gilbert 1987), and then using these lower and upper confidence
limits to derive bounds on all possible distributions that exist within this range. This approach
results in an expression of the uncertainty about the true value of the arithmetic mean that arises
due to the small sample size.
The input variables for concentrations of COCs in prey of bald eagle, American bittern, and
small-footed myotis are shown in Tables K.2-4, K.2-5, K.2-12, K.2-13, K.2-20, and K.2-21.
11.2.2 Results of Exposure Assessments
Examples of exposure distributions for exposure to tPCBs and TEQ for bald eagles in the
southern PSA, American bittern in Reaches 5 and 6, and small-footed myotis in the PSA are
presented in Figures 11.2-8 through 11.2-23.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC j j
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
ff.2.2.7 Bald Eagle
Figure 11.2-8 depicts the cumulative distribution of tPCB intake rates by bald eagles in the
southern PSA. The Monte Carlo analysis indicated that exposure of bald eagles to tPCBs ranges
from a minimum of 6.23 to a maximum of 25.4 mg/kg bw/d. The mean exposure was 13.2
mg/kg bw/d, and the median exposure 13.0 mg/kg bw/d. Eighty percent of the exposure
estimates were between 10.2 and 16.6 mg/kg bw/d (Table K.2-6).
The probability bounds estimated for bald eagles foraging in the southern PSA are depicted in
Figures 11.2-8 and 11.2-9. The 10th percentile of the probability envelope formed by the lower
and upper bounds ranged between 3.73 and 13.4 mg/kg bw/d. The 50th percentile ranged
between 5.41 and 17.9 mg/kg bw/d, and the 90th percentile ranged between 8.27 and 24.2 mg/kg
bw/d. In comparison, the 10th percentile of the Monte Carlo output was 10.2, the 50th percentile
was 13.0, and the 90th percentile was 16.6 mg/kg bw/d (Table K.2-6).
Female bald eagles present in the PSA for 30 days prior to egg laying are estimated to have a
mean tPCB egg concentration of 35.3 mg/kg, a low egg concentration of 23.0 mg/kg, and a high
concentration of 51.5 mg/kg (Figure 11.2-10). Female bald eagles present in the PSA for 30
days prior to egg laying are estimated to lay eggs with a mean TEQ concentration of 683 ng/kg, a
low concentration of 440 ng/kg, and a high concentration of 997 ng/kg (Figure 11.2-11).
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 24
-------
Primary Study Area
11)1 (mg/kg bw/d)
1
2 Figure 11.2-8 Total Daily Intake (TDI) of tPCBs by Bald Eagles in the Housatonic
3 River Primary Study Area
4
Primary Study Area
TDI (ng/kg bw/d)
6 Figure 11.2-9 Total Daily Intake (TDI) of TEQ by Bald Eagles in the Housatonic
7 River Primary Study Area
8
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC -> -> oc 7/10/2003
-------
Primary study area
^ tPCB egg concentration (mg/kg)
2 Figure 11.2-10 Bald Eagle Egg Exposure to PCBs in the Housatonic River
3 Primary Study Area
4
Primary Study Area
TEQ egg concentration (ng/kg)
5
6 Figure 11.2-11 Bald Eagle Egg Exposure to TEQ in the Housatonic River Primary
7 Study Area
8
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 2g 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
11.2.2.2 American Bittern
The Monte Carlo analysis indicated that exposure of American bittern to tPCBs ranges from a
minimum of 4.70 to a maximum of 18.6 mg/kg bw/d. The mean exposure was 9.24 mg/kg bw/d,
and the median exposure was 9.07 mg/kg bw/d. Eighty percent of the exposure estimates were
between 7.30 and 11.4 mg/kg bw/d (Table K.2-14).
The probability bounds estimated for American bittern foraging in the PSA are depicted in
Figures 11.2-12 through 11.2-21. In Reach 5A, the 10th percentile of the probability envelope
formed by the lower and upper bounds ranged between 3.53 and 7.44 mg/kg bw/d. The 50th
percentile ranged between 4.42 and 9.19 mg/kg bw/d, and the 90th percentile ranged between
5.49 and 11.7 mg/kg bw/d. In comparison, the 10th percentile of the Monte Carlo output was
7.30, the 50th percentile was 9.07, and the 90th percentile was 11.4 mg/kg bw/d (Table K.2-14).
Exposures of American bittern to tPCBs in Reaches 5B, 5C, and 5D, and 6 were similar or lower
than in Reach 5A, having a mean TDI of 7.84, 9.03, and 6.53 mg/kg bw/d, respectively (Table
K.2-14). Thus, exposure of American bittern to tPCBs is similar for all reaches of the PSA. The
uncertainty of these exposure estimates, as illustrated by the probability bounds distributions,
indicates a similar degree of uncertainty for all four reaches.
Mean exposure of American bittern to TEQ was 372 for the PSA (Table K.2-16). Figures 11.2-
16 and 11.2-21 depict the cumulative distribution for TEQ intake, as well as the probability
bounds.
The lowest egg concentrations after 45 days in the PSA were 24.7, 21.0, 24.2, and 17.5 mg/kg
for Reaches 5A, 5B, 5C, and 6, respectively. Mean egg concentrations in the PSA after 45 days
were 37.0, 31.4, 36.2, and 26.2 mg/kg for Reaches 5A, 5B, 5C, and 6, respectively. High egg
concentrations after 45 days in the PSA were 53.0, 44.9, 51.8, and 37.4 mg/kg for Reaches 5A,
5B, 5C, and 6, respectively. The estimated TEQ egg concentration for American bitterns over
time is shown in Figure 11.2-21. The lowest egg concentrations after 45 days in the PSA was
898 ng/kg, the mean concentration was 1,490 ng/kg, and the high concentration was 2,290 ng/kg.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
Reach 5a
1
2
3
w
Monte Carlo
LPB
UPB
10 15
TDI (mg/kg bw/d)
Figure 11.2-12 Total Daily Intake (TDI) of tPCBs by American Bittern in Reach 5A
of the Housatonic River Primary Study Area
Reach 5B
5
6 Figure 11.2-13 Total Daily Intake (TDI) of tPCBs by American Bittern in Reach 5B
7 of the Housatonic River Primary Study Area
8
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC -> -> 7/10/2003
-------
Reach 5C
Monte Carlo
LPB
UPB
10 15
TDI (mg/kg bw/d)
2 Figure 11.2-14 Total Daily Intake (TDI) of tPCBs by American Bittern in Reach 5C
3 of the Housatonic River Primary Study Area
Reaches 5D and 6
100
.Q
.Q
O
W
¦Monte Carlo
-LPB
-UPB
10 15
TDI (mg/kg bw/d)
20
25
6 Figure 11.2-15 Total Daily Intake (TDI) of tPCBs by American Bittern in Reaches
7 5D and 6 of the Housatonic River Primary Study Area
MK01\O:\20123001,096\ERA PB\ERA PB 11.DOC
11-29
-------
Primary Study Area
1
2 Figure 11.2-16 Total Daily Intake (TDI) of TEQ by American Bittern in the
3 Housatonic River Primary Study Area
4
Reach 5A
1 10 100
tPCB egg concentration (mg/kg)
5
6 Figure 11.2-17 American Bittern Egg Exposure to tPCBs in Reach 5A of the
7 Housatonic River Primary Study Area
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
Reach 5B
100
bx
C
¦3
o
o
im
-O
©
<
V)
0N
10 100
tPCB egg concentration (mg/kg)
2 Figure 11.2-18 American Bittern Egg Exposure to tPCBs in Reach 5B of the
3 Housatonic River Primary Study Area
4
Reach 5C
tPCB egg concentration (mg/kg)
5
6 Figure 11.2-19 American Bittern Egg Exposure to tPCBs in Reach 5C of the
7 Housatonic River Primary Study Area
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 0 1 7/10/2003
-------
1
Reach 5D and 6
2
3 Figure 11.2-20 American Bittern Egg Exposure to tPCBs in Reaches 5D and 6 of
4 the Housatonic River Primary Study Area
5
Primary Study Area
TEQ egg concentration (ng/kg)
6
7 Figure 11.2-21 American Bittern Egg Exposure to TEQ in Reaches 5 and 6 of the
8 Housatonic River Primary Study Area
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC ^2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
11.2.2.3 Small-Footed Myotis
11.2.2.3.1 Total PCBs
The Monte Carlo analysis indicated that exposure of small-footed myotis to tPCBs could range
from a minimum of 2.05 to a maximum of 96.0 mg/kg bw/d. The mean exposure was 16.7
mg/kg bw/d, and the median exposure was 14.5 mg/kg bw/d (Table K.2-22). Eighty percent of
the exposure estimates were between 7.17 and 28.8 mg/kg bw/d. Figure 11.2-22 depicts the
cumulative distribution for small-footed myotis in Reach 5.
The probability bounds estimated for small-footed myotis foraging in Reach 5 are depicted in
Figure 11.2-22. The 10th percentile of the probability envelope formed by the lower and upper
bounds ranged between 1.96 and 7.67 mg/kg bw/d. The 50th percentile ranged between 4.05 and
15.0 mg/kg bw/d, and the 90th percentile ranged between 8.09 and 32.2 mg/kg bw/d. In
comparison, the 10th percentile of the Monte Carlo output was 7.17, the 50th percentile was 14.5,
and the 90th percentile was 28.8 mg/kg bw/d (Table K.2-22).
Reach 5
0 20 40 60 80 100
Dose (mg/kg bw/d)
Figure 11.2-22 Total Daily Intake (TDI) of tPCBs by Small-Footed Myotis in Reach
5 of the Housatonic River Primary Study Area
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC j j
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
11.2.2.3.2 TEQ
The Monte Carlo analysis indicated that exposure of small-footed myotis to TEQ ranges from a
minimum of 61.4 to a maximum of 7,020 ng/kg bw/d. The mean exposure was 1,130 mg/kg
bw/d, and the median exposure was 936 ng/kg bw/d (Table K.2-24). Eighty percent of the
exposure estimates were between 381 and 2,120 ng/kg bw/d. Figure 11.2-23 depicts the
cumulative distribution for small-footed myotis in Reach 5.
The probability bounds estimated for small-footed myotis foraging in the PSA are depicted in
Figure 11.2-23. The 10th percentile of the probability envelope formed by the lower and upper
bounds ranged between 18.8 and 985 ng/kg bw/d. The 50th percentile ranged between 38.1 and
1,910 ng/kg bw/d, and the 90th percentile ranged between 74.8 and 3,900 ng/kg bw/d. In
comparison, the 10th percentile of the Monte Carlo output was 381, the 50th percentile was 936,
and the 90th percentile was 2,120 ng/kg bw/d (Table K.2-24).
Reach 5
TDI (ng TEQs/kg bw/d)
Figure 11.2-23 Total Daily Intake (TDI) of TEQ by Small-Footed Myotis in Reach 5
of the Housatonic River Primary Study Area
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
1 11.3 EFFECTS ASSESSMENT
2 The purpose of the effects assessment is to review the scientific literature and to derive the most
3 appropriate metrics for effects of tPCBs and TEQ to T&E species. An effects metric can be
4 represented by a dose-response relationship or a daily dose of a COC that represents a threshold
5 beyond which toxic effects may appear in T&E species. The effects metric is used, in
6 conjunction with the exposure assessment, to estimate risks to T&E species exposed to tPCBs
7 and TEQ in the Housatonic River PSA. This section focuses on effects that have an influence on
8 the long-term maintenance of T&E species populations (i.e., mortality or impairment of
9 reproduction or growth).
10
Toxicity of tPCBs and TEQ to Avian Species
11
Mode of Action
12
Binding to the aryl-hydrocarbon (Ah) receptor, eliciting an Ah receptor-mediated
13
biochemical and toxic response.
14
Types of Toxicity
Specific Effects
15
hepatotoxicity
mortality
16
immunotoxicity
decreased growth
17
neurotoxicity
weight loss
18
embryotoxicity
porphyria
19
teratogenicity
reduced hatching
20
embryo mortality
21
22
Toxicity of tPCBs and TEQ to Mammal Species
23
Mode of Action
24
Binding to the Ah receptor, eliciting an Ah receptor-mediated biochemical and toxic
25
response.
26
Types of Toxicity
Specific Effects
27
hepatotoxicity
mortality
28
immunotoxicity
decreased growth
29
neurotoxicity
decreased body and organ weight
30
embryotoxicity
reduced survival at birth and weaning
31
teratogenicity
reduced fertility
32
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC -> -> C 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12 A brief review of the scientific literature on the effects of tPCBs and TEQ to T&E species from
13 dietary exposure is presented in the following sections. The discussion focuses on ecologically
14 relevant effect endpoints such as survival, growth, and reproduction of T&E species. A
15 summary of reproduction effects for tPCBs and TEQ is presented in Figures K.3-1 and K.3-2 and
16 Table K.3-1. The effects metrics used for this assessment are also presented.
17 11.3.1 Total PCBs
18 Laboratory studies on the toxicity of PCBs to bald eagles and American bittern have not been
19 conducted. However, studies using other avian species were available. Appendix H provides
20 detailed descriptions of dietary and in ovo exposures of PCBs and TEQ to surrogate bird species.
21 Laboratory studies on raptor species demonstrated that PCBs cause adverse effects. American
22 kestrels (Falco sparverius) dosed in ovo to produce a mean PCB tissue concentration of 34.1
23 mg/kg on a whole egg wet weight (ww) basis (PCBs were a 1:1:1 mixture of Aroclors
24 1248:1254:1260) had decreased reproductive success, including suppression of egg laying,
25 delays in clutch initiation, smaller clutch sizes, and reduced fledgling survival (Fernie et al.
26 2001a). Twenty-five percent of exposed females failed to lay any eggs compared to 9% of the
27 control females. PCB-exposed females also had lower fledgling success: 55% compared to
28 93.3% in the control group. Males exposed to PCBs in ovo also showed reduced reproductive
29 success, with 63.5% of their broods experiencing complete mortality compared to 0% complete
30 mortality in the control group (Fernie et al. 2001a).
Mode of Action of TEQ Congeners
Congeners that have been assigned a 2,3,7,8-TCDD TEF have the ability to bind
with the Ah receptor and elicit similar toxic responses. The most toxic congeners
tend to be those that have a planar shape and are chlorinated in the 2,3,7, and 8
positions for dioxins and furans, and in the meta and para positions for PCBs. This
leads to a common mechanism of action in many animal species involving binding to
the Ah receptor and elicitation of an Ah receptor-mediated biochemical and toxic
response. The toxic response of this group of chemicals is, therefore, related to the
three-dimensional structure of the substance, including the degree of chlorination
and positions of the chlorine on the aromatic frame.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
Numerous field studies have found that organochlorine compounds negatively impact the
reproductive success of raptors and piscivorous birds (see overview in Donaldson et al. 1999).
Toxicological effects include reduced hatching success, malformation, edema, and reduced organ
and body weight (Elliott et al. 1996).
Wiemeyer et al. (1993) reported a significant reproductive decline in bald eagles with egg
concentrations greater than 13 mg/kg. However, PCB concentrations were highly correlated
with DDE concentrations, thus this threshold must be considered with caution. For sensitivity to
PCBs, American kestrels can also be considered as a surrogate species for bald eagles. A daily
intake rate of 7 mg/kg bw/d was shown to cause an increase in laying lag and a decrease in the
number of fledglings per breeding pair (Fernie et al. 2001a and b). These birds had a long
exposure period (100 days) and the study covered a sensitive life stage.
Hoffman et al. (1986) found a negative correlation between embryonic weight and tPCB residues
in eggs of black-crowned night herons nesting in the San Francisco Bay. Heron eggs had a mean
PCB concentration of 4.1 mg/kg (ww). PCB-contaminated embryos, with the yolk sac removed,
had a significantly lower (15%) weight than embryos from clean sites. Concentrations of other
organochlorines were low (mean DDE concentration of 1.7 mg/kg). Other effects associated
with PCB exposure in black-crowned night herons included reduced femur to body weight ratio,
increased edema, and increased hepatic aryl hydrocarbon hydroxylase activity (Hoffman et al.
1993); some of these effects may have been related to the presence of other contaminants,
although the authors stated that concentrations of these other contaminants were not high enough
to account for the observed effects. Laporte (1982) reported that mean tPCB concentrations of
15 mg/kg in eggs negatively impacted great blue heron reproductive success in Quebec.
Great blue herons in Indiana showed no observable effects at 4.9 mg/kg tPCBs in their eggs
(predominantly PCB congeners 118/106, 105, and 156) (Custer et al. 1998). Great blue herons
in Texas had a mean of 6.2 mg/kg PCBs in their eggs and fledged 1.6 young per nest, a value
within the range of stable populations (Mitchell et al. 1981). Total egg PCB concentrations of 1
mg/kg were found to have no effect on great blue heron productivity (Elliott et al. 1989).
No PCB toxicology studies have been conducted on small-footed myotis; however, studies have
been conducted for little brown bats and big brown bats. The little brown bat is a closely related
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC -> -> n 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
species that has similar habitat, behavior, diet, and size. The big brown bat also shares many
traits with the small-footed myotis (see Appendix A).
Studies have shown that bats accumulate PCBs from their diet (Clark and Lamont 1976a; Clark
and Lamont 1976b; Clark and Prouty 1976; Clark 1978; Clark and Stafford 1981). Clark and
Prouty (1976) found that little brown bats accumulated higher concentrations of organochlorine
compounds (including PCBs, DDT, and DDE) than did other bat species at the same location.
Little brown bats also accumulated brain PCB concentrations that were higher than brain
concentrations of other bat species, which may make them more susceptible to PCB poisoning
(Clark and Prouty 1976). Like the little brown bat, small-footed myotis may also be more
susceptible to PCBs because they have similar physiology, diet, and habitat.
Adult little brown bats fed mealworms with 15 mg/kg PCBs (Aroclor 1260) for 40 days
accumulated a mean PCB concentration of 92 mg/kg (carcass ww) (Clark and Stafford 1981). In
the same study five little brown bats were fed a diet of mealworms containing 1,000 mg/kg
PCBs; four of the bats died before 40 days. These four bats had a mean PCB concentration of
3,300 mg/kg (ww). The one surviving bat had a PCB concentration of 940 mg/kg (Clark and
Stafford 1981).
Residues of PCBs in bat brains are a linear function of the amount of fat and residues in the
carcass (Clark and Prouty 1977; Clarke et al. 1978; Clark and Stafford 1981). During
hibernation, the percent of lipids in the body decreases, but the lipid percentage in the brain does
not change, which results in elevated concentrations of PCBs in the brain (Clark and Prouty
1977; Clark and Stafford 1981). Elevated concentrations of PCBs in the brain may lead to
tremoring, a common symptom of organochlorine poisoning (Clark and Stafford 1981). Bats in
hibernation have energy stores that are closely balanced against needs, and any disturbance, such
as tremoring, that increases metabolic rates can cause mortality through starvation (Clark and
Stafford 1981).
PCBs are also known to have adverse reproductive effects on bats (Clark and Lamont 1976a;
Clark and Lamont 1976b; Clark 1978). Female bats transfer PCBs to their young through the
placenta (Clark et al. 1975; Clark 1978). Clark and Lamont (1976a) found that neonates contain
16.8% to 31.8% as much PCBs as their parents. Organochlorine contaminants are also passed to
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC H 3 8 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
the young through the mother's milk (Clark et al. 1975; Clark and Lamont 1976a). Milk
collected from the stomachs of young big brown bats contained a PCB (Aroclor 1260)
concentration of 13 mg/kg. The young had a mean PCB concentration of 0.7 mg/kg (Clark and
Lamont 1976a). Wild captured female big brown bats that produced dead young contained
significantly higher concentrations of PCBs (1.99 mg/kg ww) than those that produced live
young (0.56 mg/kg ww) (Clark and Lamont 1976b).
11.3.2 2,3,7,8-TCDD Toxic Equivalence (TEQ)
Several researchers estimated NOAEL and LOAEL values for TEQ for bald eagles (Giesy et al.
1995, Bowerman et al. 1995, Elliott et al. 1996). Giesy et al. (1995) and Bowerman et al. (1995)
sampled the impact of contaminated prey fish on bald eagles in the Great Lakes region and
derived a NOAEL for bald eagle eggs of 7 ng/kg TEQ. This value is based on toxicity studies
conducted using other avian species, including the chicken, wood duck, and American kestrel.
The bald eagle is less sensitive to TEQ compared to the chicken, ducks, and other gallinaceous
species; therefore, this value may be low (Elliott et al. 1996; Elliott and Harris 2002 in press).
Elliott et al. (1996) reported a NOAEL of 135 ng TEQ/kg egg and a LOAEL of 400 ng TEQ/kg
egg based on studies of incubated bald eagle eggs taken from nests in British Columbia. Studies
conducted on ospreys (Pandion haliaetus) found a similar NOAEL on hatching success of 136
ng/kg (Elliott et al. 2001), and Woodford et al. (1998) found that 162 ng/kg had no effect on
productivity, but may have been influencing growth of young. Using the Elliott et al. (1996)
study, the NOAEL is 135 ng/kg TEQ in eggs and the corresponding LOAEL is 400 ng/kg TEQ
in eggs. A chronic LOAEL of 25,000 ng/kg bw/d can be derived from the oral dose of PCB-126
to American kestrels, which caused significantly reduced skeletal growth in hatchlings.
Hatchlings at this exposure dose also had decreased spleen weight, increased liver weight, and
lymphoid depletion in the spleen and bursa (Hoffman et al. 1996).
Black-crowned night heron pipping embryos had TEQ concentrations of 30, 622, and 272 ng/kg
(Rattner et al. 2000). Benzyloxyresorufin-O-deethylase (BROD) and 7-ethoxyresorufin-O-
deethylase (EROD) activity were elevated at the 272 and 622 ng/kg TEQ concentration. Elliott
et al. (1989) found that a TEQ of 230 ng/kg in the eggs of great blue herons caused reduced
reproductive success. The same study found TEQ concentrations of 11, 14, 34, 64, and 79 ng/kg
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 39
-------
1 in eggs to have no effect on hatching success. The NOAEL for TEQ in eggs is 79 ng/kg, and the
2 LOAEL is 230 ng/kg TEQ in eggs.
3 The dose-response curve for TEQ is derived using the results of Khera and Ruddick (1973) and
4 Sparschu et al. (1971). Khera and Ruddick (1973) treated pregnant Wistar rats with several
5 doses of TEQ on gestation days 6 to 15. Animals were sacrificed on day 22 of gestation. A
6 dose-related decrease in live fetuses was observed; 100% embryonic lethality was reported when
7 animals were exposed to a dose of 4,000 ng TEQ/kg bw/d. Sparschu et al. (1971) made similar
8 observations in Sprague Dawley rats fed several doses of TCDD on days 6 to 15 of gestation.
9 The number of viable fetuses decreased and the total number of resorptions increased dose
10 dependently, starting at 125 ng TEQ/kg bw/d.
11 11.3.3 Effects Metrics for Characterizing Risk
12 Effects data can be characterized and summarized in a variety of ways ranging from benchmarks
13 designed to be protective of most or all species to concentration- or dose-response curves. A
14 summary of the decision criteria used to derive effects metrics is provided in the text box.
15 Further details on the decision criteria used in selecting effects metrics is provided in Section 6.6
16 of the ERA.
17 In this ERA, data were available to derive dose-response curves using surrogate mammals for the
18 small-footed myotis. Toxicity threshold ranges were developed for bald eagles and American
19 bittern.
20 11.3.3.1 Effects of tPCBs to Bald Eagle
21 American kestrels can be considered as a surrogate species for bald eagles when evaluating
22 toxicity studies. A daily intake rate of 7 mg/kg bw/d was shown to cause an increase in laying
23 lag, and a decrease in the number of fledglings per breeding pair of kestrels (Fernie et al. 2001a
24 and b). These birds had a long exposure period (100 days) and the study covered a sensitive life
25 stage. Therefore, this dose is considered to be the LOAEL for tPCBs. A chronic NOAEL was
26 estimated by applying a factor of 10 to the LOAEL, resulting in a NOAEL of 0.7 mg/kg bw/d.
27 This NOAEL is used as the toxicity threshold for bald eagles.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26 Wiemeyer et al. (1993) reported a significant reproductive decline in bald eagles with egg
27 concentrations greater than 13 mg/kg. However, PCB concentrations were highly correlated
28 with DDE concentrations, thus this threshold must be considered with caution. A threshold
29 value of 20 mg/kg in bald eagle eggs was suggested in the recent assessment of the Fox
30 River/Green Bay system (Stratus 1999). That value is consistent with other raptor studies that
31 suggest tPCBs have higher egg thresholds for reproductive effects than does DDE (Helander et
32 al. 1982; Peakall et al. 1990; Nobel and Elliott 1990). Therefore, the field-based threshold
33 selected for tPCB in bald eagle eggs was 20 mg/kg. If the probability of exceeding the toxicity
34 threshold was less than 20%, the risk to T&E species was considered to be low. If the
35 probability of exceeding the toxicity threshold was greater than 50%, the risk to T&E species
36 was considered to be high. All other outcomes are considered to have intermediate risk.
Decision Criteria for Derivation of Effects Metric
The following is the hierarchy of decision criteria used to characterize effects for each
receptor-COC combination:
1. Have single-study bioassays with five or more treatments been conducted on the
receptor of interest or a reasonable surrogate? If yes, estimate the
concentration- or dose-response. If not, go to 2.
2. Are multiple bioassays with similar protocols, exposure scenarios and effects
metrics available that, when combined, have five or more treatments for the
receptor of interest or a reasonable surrogate? If yes, estimate the dose-
response relationship as in 1. If not, go to 3.
3. Have bioassays with less than five treatments been conducted on the receptor of
interest or a reasonable surrogate? If yes, conduct or report results of
hypothesis testing to determine the NOAEL and LOAEL. If not, go to 4.
4. Are sufficient data available from field studies and monitoring programs to
estimate concentrations or doses of the COC that are consistently protective or
associated with adverse effects? If yes, develop field-based effects metrics. If
not, go to 5.
5. Derive a range where the threshold for the receptor of interest is expected to
occur. Because information on the sensitivity of the receptor of interest is
lacking, it is difficult to derive a threshold that is neither biased high or low. If
bioassay data are available for several other species, however, calculate a
threshold for each to determine a threshold range that spans sensitive and
tolerant species. That range is likely to include the threshold for the receptor of
interest.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC H 41
-------
1 11.3.3.2 Effects of TEQ to Bald Eagle
2 A chronic LOAEL of 25,000 ng/kg bw/d can be derived from the oral dose of PCB-126 to
3 American kestrels, which caused significantly reduced skeletal growth in hatchlings. Hatchlings
4 at this exposure dose also had decreased spleen weight, increased liver weight, and lymphoid
5 depletion in the spleen and bursa (Hoffman et al. 1996). This dose is considered the chronic
6 LOAEL. A dose of 5,000 ng/kg bw/d TEQ did not produce any adverse effects on American
7 kestrel chicks, and therefore, is the NOAEL. This NOAEL is used as the toxicity threshold for
8 bald eagles exposed to TEQ. Using the Elliott et al. (1996) study, the toxicity threshold for TEQ
9 in eggs is the NOAEL of 135 ng/kg.
10 11.3.3.3 Effects of tPCBs to American Bittern
11 There were insufficient data available to develop a field-based threshold for American bitterns
12 exposed to tPCBs. In the absence of such data, the threshold range estimated for sensitive and
13 tolerant species, developed for piscivorous birds (Appendix H), exposed to tPCBs was applied.
14 For American bitterns, the NOAEL for sensitive species (0.12 mg/kg bw/d) was the lower
15 toxicity threshold, and the NOAEL for tolerant species (0.7 mg/kg bw/d) was the upper toxicity
16 threshold.
17 This review indicates that the threshold for toxic effects to herons is in the range of 4 to >6
18 mg/kg PCBs in eggs. American bitterns can be reasonably represented by black-crowned night
19 and great blue herons. For this assessment, a NOAEL of 4.9 mg/kg PCBs in eggs was selected
20 for American bitterns.
21 11.3.3.4 Effects of TEQ to American Bittern
22 There were insufficient data to develop a field-based threshold for American bitterns exposed to
23 TEQ. In the absence of such data, a range of toxic effects was estimated to span the range for
24 sensitive to tolerant avian species, as was done for piscivorous birds (Appendix H). The
25 threshold range estimated for American bittern exposed to TEQ was 14 ng/kg bw/d to 5,000
26 ng/kg bw/d.
27 Elliott et al. (1989) found that a TEQ of 230 ng/kg in the eggs of great blue herons caused
28 reduced reproductive success. The same study found TEQ concentrations of 11, 14, 34, 64, and
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC ^
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
79 ng/kg in eggs to have no effect on hatching success. The NOAEL selected for TEQ in eggs
was 79 ng/kg and the LOAEL was 230 ng/kg TEQ in eggs. The toxicity threshold for American
bittern eggs exposed to TEQ is the NOAEL, 79 ng/kg.
11.3.3.5 Effects of tPCBs to Small-Footed Myotis
The Spencer (1982) study was used for the derivation of a dose-response curve based on
mortality at birth. Figure 11.3-1 presents the dose-response curve for mortality of rats at birth.
The dose-response curve indicates that 10% and 20% declines in mortality at birth would be
expected at doses of 3.05 and 5.37 mg/kg bw/d, respectively.
0.1
10
Dose (mg/kg bw/day)
1000
Figure 11.3-1 Dose-Response Curve for Effects of tPCBs on Mortality at Birth of
Rats
11.3.3.6 Effects of TEQ to Small-Footed Myotis
Because of the similarity of the protocols, the Khera and Ruddick (1973) and Sparschu et al.
(1971) studies were combined for the derivation of a dose-response curve based on reproductive
effects. Figure 11.3-2 presents the dose-response curve for reproductive fecundity of rats
exposed to TEQ. The dose-response curve indicates that 10% and 20% declines in reproductive
fecundity would be expected at doses of 156 and 330 ng/kg bw/d TEQ, respectively.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 43
-------
1
2
3
Figure 11.3-2
0 1 10 100
Dose (mg TEQ/kg bw/day)
Dose-Response Curve for Effects of TEQ on Mortality at Birth of
Rats
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC ^
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
11.4 RISK CHARACTERIZATION
This section characterizes risk to T&E species exposed to tPCBs and TEQ in the PSA of the
Housatonic River. The risk characterization discusses two potential lines of evidence, field
surveys and modeled exposure and effects, to determine potential ecological risks to T&E
species. The key risk questions and the two potential lines of evidence are summarized in the
text box.
Key Risk Questions
¦ Are the concentrations of tPCBs and TEQ present in the prey of T&E species
sufficient to cause adverse effects to individuals inhabiting the PSA of the
Housatonic River?
¦ If so, how severe are the risks and what are their potential consequences?
Lines of Evidence
¦ Use of qualitative field surveys (not considered in the weight-of-evidence
analysis).
¦ Probabilistic exposure and effects modeling.
Section 11.4.1 presents a brief overview of the methodology, results, and interpretation of the
bird and bat surveys conducted from 1998 to 2001 in the Housatonic PSA. A more detailed
presentation of this information is provided in Appendix A. In Section 11.4.2, the dose-response
curves are combined with the corresponding exposure distributions to derive risk curves that
characterize the relationship between probability and magnitude of effect. A weight-of-evidence
analysis is presented is Section 11.4.3 along with sources of uncertainty (Section 11.4.4) and the
overall findings of the risk assessment (Section 11.4.5).
11.4.1 Field Survey
T&E species in the Housatonic River study area were surveyed from 1998 to 2001. Field data
were collected using methods targeted at specific species or family groups, as well as more
general, reconnaissance-level investigations of species' presence, relative abundance, and habitat
use. Surveys for T&E species were conducted as part of broader survey efforts. Throughout this
period, any observations were recorded along with notes on habitat use, breeding signs, and
behavior.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 45
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
The avian community in the PSA was studied over a 4-year period, from 1998 to 2001. Surveys
were conducted to record presence, abundance, and habitat usage for each major group of birds.
These surveys included wading and marsh bird surveys, hawk and owl surveys, and forest bird
surveys. Additional studies were conducted to sample animal tissues (i.e., waterfowl sampling,
tree swallow study). Observations recorded in the field were used to refine the matrix to depict
habitat use and seasonality of occurrence. Marsh and wading bird surveys were conducted in
1998 using playback point counts to identify species using the PSA wetlands and reference areas
(Appendix A). Playback point counts were also used in 1999 to survey hawks and owls (raptors)
in the PSA and in three reference areas. In the PSA, raptor transects were positioned along the
Housatonic River from the confluence of the East and West Branches to Woods Pond (Appendix
A).
The mammalian community in the PSA was studied from 1998 to 2001. Field data included
methods targeted at specific species, as well as more general, reconnaissance-level investigations
of species presence, relative abundance, and habitat use.
Bat surveys were conducted to determine presence in the PSA by recording their echolocation
calls. Three transects were established along the river in the northern, central, and southern
sections of the PSA (Appendix A). There is a large amount of overlap between the call
characteristics of the little brown bat, small-footed myotis, and Indiana bat, which makes it
difficult to distinguish between these Myotis species using echolocation. When recording the
results, these three species were all labeled as Myotis sp. The majority of these calls were likely
little brown bat; however, a small number of the calls had parameters that suggested small-footed
myotis rather than little brown bats or Indiana bats. Small-footed myotis cannot be confirmed
without having animals in hand for visual identification.
Bald eagles were not observed during raptor surveys in the PSA or any of the three reference
areas. However, incidental bald eagle observations were made in the PSA (primarily in the
vicinity of Woods Pond) and at the Threemile Pond reference area. The PSA provides suitable
nesting and foraging habitat for bald eagles. In the mid 1990s, a pair of bald eagles constructed a
nest at Woods Pond (T. Gulo, MDFW, personal communication 2001). The nest was reportedly
destroyed during an April snowstorm and the pair did not attempt to re-nest. In 2001, a bald
eagle pair nested along the Housatonic River in Connecticut, below Interstate 84, and raised one
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 7/10/2003
-------
1 chick. The pair returned in 2002 and displayed breeding activity (i.e., adding sticks to previous
2 year's nest); however, no nesting was observed (J. Bictoria, CTDEP, personal communication
3 2002).
4 American bitterns were not observed during marsh bird surveys in the PSA, and no marsh bird
5 surveys were conducted outside of the PSA. Incidental observations of American bitterns
6 occurred in the PSA and at Washington Mountain Lake during the breeding season, and one
7 individual was heard calling in the PSA, indicating intent to breed in the area. These
8 observations were incidental observations, occurring while researchers were on-site for other
9 surveys. During timed bird surveys (i.e., playback surveys), the results in the PSA were the
10 same for reference areas; no bald eagles or American bitterns were observed in either location.
11 Small-footed myotis observations in the PSA have not been confirmed, and no bat surveys were
12 conducted in reference areas. Suitable summer habitat for small-footed myotis is present in and
13 adjacent to the study area, and it is likely that the species occurs there. The small-footed myotis
14 has been recorded in western Massachusetts and has been documented twice since 1978 in
15 Hampden County, MA (MNHESP 1984; Godin 1977), making their presence in the study area
16 possible. It is believed that this species was recorded during bat surveys; however, as previously
17 mentioned, limitations of echolocation technology prevent this species from being definitively
18 identified. Other studies conducted in the region have reported small-footed myotis observations
19 (Zimmerman and Glanz 2000; Krusic et al. 1996).
20 Any differences in population structure between PSA and reference locations cannot be
21 evaluated due to the overall low population size of T&E species, few numbers of sightings, and
22 qualitative study design. The number of observations for these species by definition is expected
23 to be low and the lack of observations in one location does not necessarily reflect the suitability
24 of that habitat or the absence of an individual.
25 11.4.2 Comparison of Estimated Exposures to Laboratory-Derived Effects
26 Doses
27 For the bald eagle, exposure was assessed for the southern PSA, while exposure of American
28 bittern was estimated for each reach, and exposure to small-footed myotis was estimated for the
29 entire PSA.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
For each receptor-COC combination, a category of low, intermediate, or high risk was assigned
using the guidance in the text box following integration of the exposure and effects distributions.
This exercise was done separately for the results of the Monte Carlo analyses and each of the
lower and upper bounds from the probability bounds analyses. The "risk category" refers to the
level of risk based on the results of the Monte Carlo analyses. The "risk range" refers to the
levels of risk based on the results of the probability bounds analyses. The toxicity thresholds or
the 10% and 20% effects doses for each species and COC are presented in Section 11.3.
Guidance for Determining Level of Risk to Bald Eagle
¦ If the probability of exceeding the toxicity threshold was less than 20%, the risk to
bald eagle was considered to be low.
¦ If the probability of exceeding the toxicity threshold was greater than 50%, the
risk to bald eagle was considered to be high.
¦ All other outcomes are considered to have intermediate risk.
Guidance for Determining Level of Risk to American Bittern
¦ If the probability of exceeding the lower toxicity threshold was less than 20%, the
risk to American bittern was considered to be low.
¦ If the probability of exceeding the upper toxicity threshold was greater than 20%,
the risk to American bittern was considered to be high.
¦ All other outcomes are considered to have intermediate risk.
Guidance for Determining Level of Risk to Small-Footed Myotis
¦ If the probability of 10% or greater effect is less than 20%, then the risk to small-
footed myotis is low.
¦ If the probability of 20% or greater effect is greater than 50%, then the risk to
small-footed myotis is high.
¦ All other outcomes are considered to have intermediate risk.
The results of the risk characterization are summarized in Table 11.4-1. Figures 11.4-1 through
11.4-4 are risk curves for bald eagles exposed to tPCBs and TEQ in the PSA. Figures 11.4-5
through 11.4-14 show American bittern exposed to tPCBs and TEQ in the PSA, and Figures
11.4-15 and 11.4-16 show small-footed myotis exposed to tPCBs and TEQ in the PSA.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
1 Table 11.4-1
2
3 Summary of Qualitative Risk Statements for T&E Species from the Housatonic
4 River Study Area
Bird / Location
Qualitative Risk Statements
tPCBs
TEQ
Risk Category"
Risk Rangeb
Risk Category"
Risk Rangeb
Bald Eagle
Southern PSA
High
High
Low
Low
American Bittern
Reach 5A
High
High
Intermediate
Intermediate
Reach 5B
High
High
Intermediate
Intermediate
Reach 5C
High
High
Intermediate
Intermediate
Reach 5D and 6
High
High
Intermediate
Intermediate
Small-Footed Myotis
Reaches 5 and 6
High
Intermediate - High
High
Low - High
5 aRisk category is the risk level based on First Order Monte Carlo (FOMC).
6 bRisk range is the range of risk encompassed by the upper and lower probability bounds (UPB and LPB).
7
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC ^g
-------
1
Primaiy Study Area
0 10 20 30 40
TDI (mg/kg bw/d)
2
3 Figure 11.4-1 Risk Curves for Bald Eagles Exposed to tPCBs in the Housatonic
4 River Primary Study Area
Primary Study Area
TDI (ng/kg bw/d)
5
6 Figure 11.4-2 Risk Curves for Bald Eagles Exposed to TEQ in the Housatonic
7 River Primary Study Area
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 50
-------
Primary Study Area
100
W)
e
-
&
¦-
#o
'£
A
<
CM
a
VI
C3
Q
High
~Mean
Low
Toxicity threshold
1 100
tPCB egg concentration (mg/kg)
Figure 11.4-3 Risk for Bald Eagle Eggs Exposed to tPCBs in the Housatonic
River Primary Study Area
100
Primary Study Area
High
Mean
Low
Toxicity threshold
1 10 100 1000 10000
TEQ egg concentration (ng/kg)
Figure 11.4-4 Risk for Bald Eagle Eggs Exposed to TEQ in the Housatonic River
Primary Study Area
MK01\O:\20123001.096\ERA PB\ERA PB 11.DOC
11-51
-------
Reach 5A
100
-Q
-O
O
-
a.
Monte Carlo
LPB
UPB
Low - Intermediate Criterion
O Intermediate - High Crierion
•
5 10
TDI (mg/kg bw/d)
15
20
Figure 11.4-5 Risk Curves for American Bittern Exposed to tPCBs in Reach 5A
of the Housatonic River Primary Study Area
Reach 5B
Monte Carlo
LPB
UPB
Low - Intermediate Criterion
2
ft
o O Intermediate - High Criterion
*
5 10
TDI (mg/kg bw/d)
15
20
Figure 11.4-6 Risk Curves for American Bittern Exposed to tPCBs in Reach 5B
of the Housatonic River Primary Study Area
MK01\O:\20123001,096\ERA PB\ERA PB 11.DOC
11-52
-------
Reach 5C
Monte Carlo
— LPB
— UPB
# Low - Intermediate Criterion
O Intermediate - High Criterion
5 10
TDI (mg/kg bw/d)
15
20
Figure 11.4-7
Risk Curves for American Bittern Exposed to tPCBs in Reach 5C
of the Housatonic River Primary Study Area
Reaches 5D & 6
.Q
O
W
Monte Carlo
— LPB
— UPB
# Low - Intermediate Criterion
O Intermediate - High Criterion
5 10
TDI (mg/kg bw/d)
15
20
Figure 11.4-8 Risk Curves for American Bittern Exposed to tPCBs in Reaches
5D and 6 of the Housatonic River Primary Study Area
MK01\O:\20123001.096\ERA PB\ERA PB 11.DOC
11-53
-------
Primary Study Area
Monte Carlo
0 1000 2000 3000 4000 5000
TDI (ng/kg bw/d)
1
2 Figure 11.4-9 Risk Curves for American Bittern Exposed to TEQ in the
3 Housatonic River Primary Study Area
Reach 5A
1 10 100
tPCB egg concentration (mg/kg)
4
5 Figure 11.4-10 Risk for American Bittern Eggs Exposed to tPCBs in Reach 5A of
6 the Housatonic River Primary Study Area
MK01\O:\20123001,096\ERA_PB\ERA_PB_11 .DOC j j ^
-------
Reach 5B
High
Mean
Low
Toxicity threshold
1 10 100
tPCB egg concentration (mg/kg)
Figure 11.4-11 Risk for American Bittern Eggs Exposed to tPCBs in Reach 5B of
the Housatonic River Primary Study Area
Reach 5C
100
bD
C
¦3
u
to
&
<
a
&
cs
o
80
60
40
20
High
Mean
Low
Toxicity threshold
1 10 100
tPCB egg concentration (mg/kg)
Figure 11.4-12 Risk for American Bittern Eggs Exposed to tPCBs in Reach 5C of
the Housatonic River Primary Study Area
MK01\O:\20123001.096\ERA PB\ERA PB 11.DOC
11-55
-------
Reach 5D & 6
100
bD
a
¦3
0>
-Q
o
&
<
a
&
>%
cs
o
80
60
40
20
High
Mean
Low
Toxicity threshold
10 100
tPCB egg concentration (mg/kg)
Figure 11.4-13 Risk for American Bittern Eggs Exposed to tPCBs in Reaches 5D
and 6 of the Housatonic River Primary Study Area
Primary Study Area
100
b£
G
-C
o
-
A
$
a.
a
VI
C3
Q
10 100 1000
TEQ egg concentration (ng/kg)
High
Mean
Low
Toxicity threshold
Figure 11.4-14 Risk for American Bittern Eggs Exposed to TEQ in the Housatonic
River Primary Study Area
MK01\O:\20123001.096\ERA PB\ERA PB 11.DOC
11-56
-------
Reach 5
% Decline in Fecundity
1
2 Figure 11.4-15 Risk Curves for Small-Footed Myotis Exposed to tPCBs in Reach
3 5 of the Housatonic River Primary Study Area
Reach 5
100
4
5
6
g,
c«
-C
o
h
0-
O
u
e
o
¦a
o
o
u
fcj
80
60
40
20
Monte Carlo
LPB
UPB
• Low-Intermediate Criterion
O Intermediate - High Criterion
20 40 60
% Decline in Fecundity
80
100
Figure 11.4-16 Risk Curves for Small-Footed Myotis Exposed to TEQ in Reach 5
of the Housatonic River Primary Study Area
MK01\O:\20123001,096\ERA PB\ERA PB 11.DOC
11-57
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
The results of the risk characterization showed that the highest risk to T&E species is to bald
eagles and American bitterns from exposure to tPCBs. The risk for bald eagles exposed to TEQ
was low; however, risk to bald eagle eggs exposed to TEQ was high. The analysis for bald
eagles associated with exposure to tPCBs downstream of Woods Pond indicated that bald eagles
would only potentially be at risk in Reach 8 (Rising Pond). The risk to bald eagles nesting and
wintering downstream of the PSA is low. The risk category for American bittern was
intermediate for TEQ and high for eggs exposed to TEQ. The risk category for small-footed
myotis was high for both tPCB and TEQ. The risk range for small-footed myotis, as determined
by the probability bounds analysis, ranged from intermediate to high for tPCBs and low to high
for TEQ.
11.4.3 Weight-of-Evidence Analysis
A weight-of-evidence analysis was used to evaluate the lines of evidence described in the
preceding sections for T&E species. The goal of this analysis was to determine whether
significant risk is posed to T&E species in the Housatonic River PSA as a result of exposure to
tPCBs and TEQ. The three-phase approach of Menzie et al. (1996) and the Massachusetts
Weight-of-Evidence Workgroup was applied for this purpose, in which weight-of-evidence was
reflected in the following three characteristics: (a) the weight assigned to each measurement
endpoint, (b) the magnitude of response observed in the measurement endpoint, and (c) the
concurrence among outcomes of the multiple measurement endpoints. As noted previously, field
surveys were qualitative and therefore not used in this analysis.
The rationale for weighting of measurement endpoints is provided in Appendix K, along with a
discussion of attributes considered in the weight-of-evidence. A summary of how attributes were
weighted for the bald eagle, American bittern, and small-footed myotis lines of evidence is
provided in Table 11.4-2. For both tPCBs and TEQ, the modeled exposure and effects lines of
evidence were given a moderate/high value.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 5 8 7/10/2003
-------
1 Table 11.4-2
2
3 Weighting of Measurement Endpoints for T&E Species Weight-of-Evidence Evaluation
Attributes
Modeled Exposure and
Effects for Bald Eagles
Exposed to tPCBs and
TEQ
Modeled Exposure and
Effects for American
Bitterns Exposed to
tPCBs and TEQ
Modeled Exposure and
Effects for Small-
Footed Myotis Exposed
to tPCBs and TEQ
Rationale
I. Relationship Between Measurement and Assessment Endpoints
1. Degree of Association
H (tPCBs)
M/H (TEQ)
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
Exposure models were species-specific, but effects metrics
for bald eagle (body burden), American bitterns, and small-
footed myotis were derived from studies of surrogate
species. Effects metrics for bald eagle eggs were species-
specific for tPCBs.
2. Stressor/Response
M/H (tPCBs)
M (TEQ)
M (tPCBs and TEQ)
M/H (tPCBs and TEQ)
Exposure modeling was species- and stressor-specific.
Effects metrics for representative species were available
only for bald eagle eggs exposed to tPCBs. A dose-
response curve was used for small-footed myotis, rather
than thresholds.
3. Utility of Measure
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
Modeled exposure and effects procedures used were
standardized and widely accepted; the primary limitation
was lack of species-specific effects data, except for bald
eagle eggs exposed to tPCBs.
II. Data Quality
4. Data Quality
H (tPCBs and TEQ)
H (tPCBs and TEQ)
H (tPCBs and TEQ)
The field surveys were performed according to well-
defined and documented protocols. The low numbers of
individuals observed, and the inability to confirm
identification of the Myotis without handling, limited the
ability to observe site-specific effects. The DQOs for the
sampling analysis and tissue samples were met for the
tissue residue data used in the exposure analysis for both
tPCBs and TEQ.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 59
-------
Table 11.4-2
Weighting of Measurement Endpoints for T&E Species Weight-of-Evidence Evaluation
(Continued)
Attributes
Modeled Exposure and
Effects for Bald Eagles
Exposed to tPCBs and
TEQ
Modeled Exposure and
Effects for American
Bitterns Exposed to
tPCBs and TEQ
Modeled Exposure and
Effects for Small-
Footed Myotis Exposed
to tPCBs and TEQ
Rationale
III. Study Design
5. Site Specificity
M (tPCBs and TEQ)
M (tPCBs and TEQ)
M (tPCBs and TEQ)
Biological tissue data used in exposure models were site
specific, and other exposure parameters were representative
of site conditions. However, effects measures were not site
specific.
6. Sensitivity
M/H (tPCBs)
H (TEQ)
H (tPCBs and TEQ)
H (tPCBs and TEQ)
Modeled exposure and effects directly assessed exposure-
response relationship. Laboratory studies from which
effects data were derived were stressor-specific.
7. Spatial Representativeness
H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
H (tPCBs and TEQ)
Modeled exposures relied on tissue data collected
throughout the study area and areas of actual exposure.
American bittern exposure based on small mammals
trapped in unfavorable foraging area.
8. Temporal Representativeness
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
Modeled exposure and effects lines of evidence spanned
critical life stages and, in general, tissue data used were
collected when exposure was expected to be high.
9. Quantitative Measure
H (tPCBs and TEQ)
H (tPCBs and TEQ)
H (tPCBs and TEQ)
Probabilistic exposure and effects modeling were highly
quantitative and propagated uncertainty associated with
modeling procedures.
10. Standard Method
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
Generally accepted exposure and effects modeling
procedures were followed, but probability bounds analysis
is a relatively new technique for propagating uncertainty.
Overall Endpoint Value
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
M/H (tPCBs and TEQ)
—
M = Moderate
H = High
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC gQ 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
The magnitude of the response in the measurement endpoint is considered together with the
measurement endpoint weight in judging the overall weight-of-evidence (Menzie et al. 1996).
This requires assessing the strength of evidence that ecological harm has occurred, as well as an
indication of the magnitude of response, if present. For bald eagles and American bitterns,
exposure and effects for both tPCBs and TEQ were estimated for body burden and eggs
separately. In the weight-of-evidence analysis, the risks for both of these factors were combined
and presented together because ecologically if there is risk to either life stage, there is risk to the
organism.
The results from the modeled exposure and effects line of evidence indicate that there is no
evidence of harm to adult bald eagles exposed to TEQ in the PSA, but high risk to bald eagle
eggs. There is, however, evidence of harm to bald eagles and American bitterns exposed to
tPCBs, and an undetermined risk for American bittern exposed to TEQ and small-footed myotis
exposed to tPCBs and TEQ in the PSA.
The weighting, evidence of harm, and magnitudes of responses were combined in a matrix
format and are presented in Tables 11.4-3 and 11.4-4.
Table 11.4-3
Evidence of Harm and Magnitude of Effects for T&E Species Exposed to tPCBs in
Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled exposure and
effects, Bald Eagle
Moderate/High
Yes
High
Modeled exposure and
effects, American
Bittern
Moderate/High
Yes
High
Modeled exposure and
effects, Small-Footed
Myotis
Moderate/High
Undetermined
High
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC H 6 1 7/11/2003
-------
1 Table 11.4-4
2
3 Evidence of Harm and Magnitude of Effects for T&E Species Exposed to TEQ in
4 the Housatonic River PSA
Measurement Endpoints
Weighting Value
(High, Moderate, Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled exposure and effects,
Bald Eagle
Moderate/High
Yes
Intermediate
Modeled exposure and effects,
American Bittern
Moderate/High
Undetermined
High
Modeled exposure and effects,
Small-Footed Myotis
Moderate/High
Undetermined
High
5
6 A graphical method was used for displaying concurrence among measurement. Tables 11.4-5
7 and 11.4-6 depict the outcome for T&E species exposed to tPCBs and TEQ, respectively. The
8 field survey line of evidence was not included as it is inconclusive.
9 Table 11.4-5
10
11 Risk Analysis Summary for T&E Species Exposed to tPCBs in the Housatonic
12 River PSA
Assessment Endpoint:
Survival, growth, and reproduction of T&E species
13 ~
Harm/Magnitude
Weighting Factors (increasing confidence of weight)
Low
Low/
Moderate
Moderate
Moderate/
High
High
Yes/High
BE, AB
Yes/Intermediate
Yes/Low
Undetermined/High
SFM
Undetermined/Intermediate
Undetermined/Low
No/Low
No/Intermediate
No/High
14 BE = bald eagle
15 AB = American bittern
16 SFM = small-footed myotis
17
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC g2 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Table 11.4-6
Risk Analysis Summary for T&E Species Exposed to TEQ in the Housatonic River
PSA
Assessment Endpoint:
Survival, growth, and reproduction of T&E species
~
Harm/Magnitude
Weighting Factors (increasing confidence of weight)
Low
Low/Moderate
Moderate
Moderate/High
High
Yes/High
Yes/Intermediate
BE
Yes/Low
Undetermined/High
AB, SFM
Undetermined/Intermediate
Undetermined/Low
No/Low
No/Intermediate
No/High
BE = bald eagle
AB = American bittern
SFM = small-footed myotis
11.4.4 Sources of Uncertainty
The assessment of risk to T&E species contains uncertainties. Each source of uncertainty can
influence the estimates of risk. Therefore, it is important to describe and, when possible, specify
the magnitude and direction of such uncertainties. Some of the major sources of uncertainty
associated with the assessment of risks of tPCBs and TEQ to T&E species are briefly described
below. A more complete list is presented in Appendix K.
¦ The Monte Carlo sensitivity analyses suggest that the FMR slope and power terms
were generally the most influential variables on predicted total daily intakes of COCs.
However, no measurements of free metabolic rate were available for the
representative wildlife species. Similarly, measured food intake rates were not
available for bald eagles, American bitterns, small-footed myotis, or reasonable
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 63 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
surrogate species. Therefore, free metabolic rates were estimated using allometric
equations. The use of allometric equations introduces some uncertainty into the
exposure estimates because they have model-fitting error, and are based on species
different from the representative species used in this assessment. Given the lack of
data on representative species used in the current assessment, it is difficult to judge
the magnitude of the uncertainty introduced by the use of the allometric models. The
uncertainty due to model-fitting error was propagated in the uncertainty analyses by
using distributions as input for the allometric slope and power terms.
¦ Sample sizes were limited for the analyses of COC concentrations in some prey
items. Only two to four invertebrate samples for Reaches 5B, 5C, and 6 were
available to estimate exposure of American bittern to PCBs. Uncertainty due to
sample size was explicitly addressed in the uncertainty analyses. In the Monte Carlo
analysis, sample size uncertainty was addressed by use of the 95% UCL on the mean.
Use of the UCL addressed uncertainty, but is biased towards overestimating
exposure. In the probability bounds analysis, uncertainty was addressed by
specifying concentration variables as a range from the 95% LCL to the 95% UCL.
This treatment of uncertainty is unbiased.
¦ PCB congeners 123 and 157 co-eluted with other congeners (PCB-123 with PCB-
149; PCB-157 with PCB-173 and PCB-201) leading to uncertainty about TEQ
concentrations in prey sample. This source of uncertainty was addressed in the
uncertainty analyses by estimating prey concentrations assuming concentrations of
PCB-123 and PCB-157 were equal to zero, and assuming that concentrations of PCB-
123 and PCB-157 were equal to the doublet and triplet concentrations, respectively.
The resulting TEQ estimates were then compared. If the ratio of the upper to lower
bound TEQ estimates was less than 1.3, this source of uncertainty was deemed
unimportant and disregarded. If the ratio exceeded 1.3, the uncertainty due to the co-
elution was propagated through the uncertainty analyses.
¦ The adult body burden and associated egg concentration was estimated assuming that
avian species do not metabolize PCBs, to simplify the estimated accumulation by
bald eagles. This assumption may result in an overestimate of the amount of PCBs
accumulated.
¦ The adult body burden was estimated assuming that a breeding adult would arrive in
the PSA with no tPCBs in the body. New breeding pairs colonizing the PSA would
be expected to have low concentrations of tPCBs in their tissue. However, eagles
return to the same breeding area, and often the same nest, each year. Bald eagles
returning to the PSA during subsequent years would have COC body burdens
accumulated during previous breeding seasons. Therefore, body burdens and egg
concentrations are likely underestimated for eagles that have previously bred in the
PSA.
¦ The largest source of uncertainty in the effects assessment was associated with the
lack of, or limited, toxicity studies involving the representative species. There were
no toxicity studies available for American bittern and small-footed myotis exposed to
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
1 tPCBs or TEQ. As a result, laboratory studies involving surrogate species were used
2 to estimate effects to these species. These extrapolations introduced uncertainty in
3 the effects assessment because of the variation in sensitivities of species to tPCBs and
4 TEQ. The sensitivity of wildlife to an environmental contaminant may also differ
5 from that of a laboratory or domestic species due to behavioral and ecological
6 parameters including stress factors (e.g., competition, seasonal changes in
7 temperature, or food availability), disease, and exposure to other contaminants.
8 ¦ For T&E species, data for two potential lines of evidence were available. For these
9 assessments, toxicity studies performed in situ in the PSA of the Housatonic River or
10 feeding studies involving prey and food items from the PSA would have improved
11 the weight-of-evidence assessment. Such studies would have accounted directly for
12 the specific characteristics of the Housatonic River ecosystem and the toxicity of the
13 PCB mixture found on site.
14 11.4.5 Conclusions
15 11.4.5.1 Risks in the PSA
16 The weight-of-evidence analysis indicates that T&E species such as bald eagles, American
17 bitterns, and small-footed myotis are at some risk in the PSA as a result of exposure to tPCBs
18 and TEQ. The risks for bald eagles and American bitterns exposed to tPCBs are high. The risk
19 for small-footed myotis exposed to tPCBs and TEQ are undetermined.
20
21
22
23
24
25
26
27
28 The bald eagle, American bittern, and small-footed myotis were chosen to represent T&E species
29 inhabiting the Housatonic River PSA. Other T&E species that occur in the area include one
30 mussel (triangle floater); three dragonflies (riffle snaketail, zebra clubtail, and arrow clubtail); a
31 turtle (wood turtle); three salamanders (Jefferson salamander, four-toed salamander, and northern
32 spring salamander); three hawks (northern harrier, sharp-shinned hawk, and Cooper's hawk); two
33 warblers (northern parula and blackpoll warbler); a wading bird (common moorhen); and a shrew
34 (northern water shrew). Some of these species were assessed in other appendices, and the risks
ERA Results for Representative T&E Species
The weight-of-evidence analysis indicates that T&E species such as bald eagle and
American bittern are at risk in the PSA as a result of exposure to tPCBs. Risks to
bald eagles and American bittern exposed to tPCBs are high. There are intermediate
risks to bald eagles exposed to TEQ, and risks to American bittern exposed to TEQ
are undetermined. Risks to small-footed myotis exposed to tPCBs and TEQ are
undetermined.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 65
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
were compared to other, more appropriate assessment endpoints (i.e., salamanders assessed in
Appendix E, Amphibians).
A qualitative analysis was conducted to compare exposure of representative species and other
similar species to tPCBs and TEQ. The major factors that influence exposure to tPCBs and TEQ
include the following:
¦ Foraging behavior and dietary composition.
¦ Foraging and home range size.
¦ Species body weight and other life history characteristics.
As noted in this ERA, effects studies conducted on bald eagles, American bittern, and small-
footed myotis are not available. Similarly, effects data are not available for other similar species
living in the Housatonic River area. As a result, the surrogate effects data used to estimate
effects to bald eagles were also used to estimate risk for other piscivorous raptors, data for
American bittern were used for other wading birds, and data from small-footed myotis for other
bat species.
Results are provided in the following text box.
ERA Results for Other Piscivorous Raptors, Wading Birds, and Bats
Living in the PSA
The other piscivorous raptor that occurs in the PSA is the osprey. Risk to osprey is
characterized in Appendix H.
Other piscivorous wading bird species that could occur in the PSA include the least
bittern, green heron, great blue heron, king rail, least rail, sora, and pied-billed grebe.
A qualitative analysis of risk to these species indicates that the great blue heron and
king rail are expected to have a similar level of risk compared to the American bittern.
The wading birds that have similar diets but are smaller and have higher
metabolisms—such as least bittern, green heron, Virginia rail, and pied-billed
grebe—are expected to have a higher level of risk than the American bittern.
Wading birds that consume plant material, such as the sora, are expected to have
low levels of risk.
Other bat species, especially those in the myotis family (little brown bat, Indiana bat,
and northern myotis) are expected to have a similar level of risk as the small-footed
myotis.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC gg 7/10/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
11.4.5.2 Risk Estimates Downstream of Woods Pond
Risks to bald eagles due to contaminants in the river and floodplain below Woods Pond were
also assessed. Total PCBs measured in sediment, floodplain soil, and fish tissue from the river
above and below Woods Pond are presented in Appendix H. The data indicate that
contamination in these media declines substantially below Woods Pond Dam.
11.4.5.2.1 Risk for Bald Eagles Wintering Downstream of Woods Pond
The risk for bald eagles from exposure to tPCBs downstream of Woods Pond was assessed by
comparing concentrations of tPCBs in prey fish in Reaches 7 to 16 to a maximum acceptable
threshold concentration (MATC) developed specifically for bald eagles. The MATC of 30.41
mg/kg tPCBs in fish (whole body, wet weight) was developed as the concentration at which bald
eagle TDI would exceed the toxicity threshold for eggs. The TDI was calculated assuming that
eagles wintering downstream of Woods Pond would consume 83.4% fish and 16.1% waterfowl
(Stalmaster and Plettner 1992). The waterfowl concentration was assumed to be zero, as
waterfowl wintering on the Housatonic are likely to have migrated there from northern locations
outside the study area. The results of the analysis that indicate that bald eagles would be at risk
only in Reach 8 (Rising Pond), are presented in Figure K.4-5. This conclusion is conservative, in
that it assumes bald eagles would consume fish only from Rising Pond. However, this is
unlikely because Rising Pond is considerably smaller than a typical bald eagle foraging area.
11.4.5.2.2 Risk for Bald Eagles Breeding Downstream of Woods Pond
Figure K.4-5 presents the assessment of risk to bald eagles exposed to tPCBs downstream of
Woods Pond. Bald eagles are known to breed downstream of Woods Pond. In particular, one
bald eagle pair nested and raised one chick in Reach 15, just south of Interstate 84, in 2001. In
2002, the pair returned to the nest and displayed breeding activity but did not nest (J. Bictoria,
CTDEP, personal communication 2002).
Risk from exposure to tPCBs was estimated for bald eagles nesting at this location. Bald eagles
have a linear (riverine) foraging distance of 1.9 to 4.3 miles (3.1 to 6.9 km) (Craig et al. 1988).
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC gy
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
Therefore, bald eagles nesting near Interstate 84 could potentially be foraging in Reach 15 (Lake
Zoar) and the southern section of Reach 14 (Lake Lillinonah).
Total PCB concentrations for prey items from Reaches 14 and 15 were available only for fish.
Because of the small number of fish samples for Reaches 14 and 15, all fish were combined
instead of separating them into classes (i.e., predatory fish, bottom feeder, forage fish). Fish in
Reaches 14 and 15 had an average concentration of 0.717 mg/kg. Bald eagles on average
consume a summer diet consisting of 78.6% fish, 16.8% birds, and 5.1% mammals (see
Appendix K). Mammal and bird tPCB concentrations were not available for downstream
reaches. Total PCB concentrations for these prey items were estimated in three ways to give
high, moderate, and low concentrations. High concentrations assumed that waterfowl and
mammals from downstream would have tPCB concentrations equal to those in the PSA. Low
concentrations assumed that waterfowl and mammals from downstream would have tPCB
concentrations of zero. A moderate concentration was developed by determining fish-to-
mammal and fish-to-bird ratios based on concentrations in the PSA. Mammal tPCB
concentrations in the PSA are on average 75% of the total fish concentration, and waterfowl
tPCB concentrations averaged 15% of the total fish concentration. Therefore, moderate tPCB
concentrations downstream were 0.538 mg/kg for mammals and 0.108 mg/kg for birds.
The estimated low, moderate, and high tPCB intake rates averaged 0.022 mg/kg bw/d, 0.025
mg/kg bw/d, and 0.243 mg/kg bw/d, respectively. These values fall below the lower toxicity
threshold of 0.7 mg/kg bw/d.
Risks from TEQ to adult bald eagles in the PSA were intermediate. Because TEQ concentrations
in downstream prey species are reduced to the same degree as PCB concentrations, it is assumed
that risk from TEQ to bald eagles breeding downstream would be low.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 1 1 68
-------
1 11.5 REFERENCES
2 Anthony, E.P.L. and T.H. Kunz. 1977. Feeding strategies of the little brown bat in southern New
3 Hampshire. Ecology 58:775-786.
4 Belwood, J.J. and M.B. Fenton. 1976. Variation in the diet of Myotis lucifugus. Canadian
5 Journal of Zoology 54:1674-1678.
6 Bictoria, J. 2002. Connecticut Department of Environmental Protection. Personal
7 Communication to Woodlot Alternatives, Inc. 12 December 2002.
8 Borga, K., G.W. Gabrielsen, and J.U. Skaare. 2001. Biomagnification of organochlorines along a
9 Barents sea food chain. Environmental Pollution 113(2): 187-198.
10 Bowerman, W.W., J.P. Giesy, D.A. Best, and V.J. Kramer. 1995. A review of factors affecting
11 productivity of bald eagles in the Great Lakes region: Implications for recovery. Environmental
12 Health Perspectives 103(Suppl 4):51-59.
13 Buehler, D.A. 2000. Bald Eagle {Haliaeetus leucocephalus). pp. 1-39. In Poole, A. and F. Gill
14 (ed.) The Birds of North America. Vol. 506. The Birds of North America, Inc., Philadelphia, PA,
15 USA.
16 Canadian Wildlife Service. 2000. Hinterland Who's Who: Bald Eagle. URL http://www.cws-
17 scf.ec.gc.ca/hww-fap/bald/bald.html (accessed 13 May 2002).
18 Clark, D.R., Jr. 1978. Uptake of dietary tPCB by pregnant big brown bats and their fetuses.
19 Bulletin of Environmental Contamination and Toxicology 19:707-714.
20 Clark, D.R., Jr., and C.J. Stafford. 1981. Effects of DDE and PCB (Aroclor 1260) on
21 experimentally poisoned female little brown bats (Myotis lucifugus): Lethal brain concentrations.
22 Journal of Toxicology and Environmental Health 7:925-934.
23 Clark, D.R., Jr., C.O. Martin, and D.M. Swineford. 1975. Organochlorine insecticide residues in
24 the free-tailed bat (Tadarida brasiliensis) at Braken Cave, Texas. Journal of Mammalogy
25 56(2):429-443.
26 Clark, D.R., Jr., and R.M. Prouty. 1976. Organochlorine residues in three bat species from four
27 localities in Maryland and West Virginia, 1973. Pesticides Monitoring Journal 10:44-53.
28 Clark, D.R., Jr., and RM. Prouty. 1977. Experimental feeding of DDE and PCB to female big
29 brown bats (Eptesicus fuscus). Journal of Toxicology and Environmental Health 2:917-928.
30 Clark, D.R., Jr., and T.G. Lamont. 1976a. Organochlorine residues in females and nursing young
31 of the big brown bat. Bulletin of Environmental Contamination and Toxicology 15(1): 1-8.
32 Clark, D.R., Jr., and T.G. Lamont. 1976b. Organochlorine residues and reproduction in the big
33 brown bat. Journal of Wildlife Management 40(2):249-254.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC 7/10/2003
-------
1 Clark, D.R., Jr., T.H. Kunz, and T.E. Kaiser. 1978. Insecticides applied to a nursery colony of little
2 brown bats (Myotis lucifugus): lethal concentrations in brain tissues. Journal of Mammalogy
3 59:84-91.
4 Cottam, C. and F.M. Uhler. 1945. Birds in Relation to Fishes. U.S. Fish and Wildlife Service
5 Leaflet, No. 272.
6 Craig, R.J., E.S. Mitchell, and J.E. Mitchell. 1988. Time and energy budgets of bald eagles
7 wintering along the Connecticut River. Journal of Field Ornithology 59:22-32.
8 Custer, C.M. 2002. Exposure and effects of chemical contaminants on tree swallows nesting
9 along the Housatonic River, Berkshire Co., Massachusetts, 1998-2000. Final Report to U.S.
10 Environmental Protection Agency. USGS, Upper Midwest Environmental Sciences Center, La
11 Crosse, WI.
12 Custer, T.W., R.K. Hines, P.M. Stewart, M.J. Melancon, D.S. Henshel, and D.W. Sparks. 1998.
13 Organochlorines, mercury, and selenium in great blue heron eggs from Indiana Dunes National
14 Lakeshore, Indiana. Journal of Great Lakes Research 24(1):3-11.
15 DeGraaf, R.M. and M. Yamasaki. 2001. New England Wildlife: Habitat, Natural History, and
16 Distribution. University Press of New England, Hanover, NH, USA. 482 pp.
17 Donaldson, G.M., J.L. Shutt, and P. Hunter. 1999. Organochlorine contamination in bald eagle
18 eggs and nestlings from the Canadian Great Lakes. Archives of Environmental Contamination and
19 Toxicology 36:70-80.
20 Dunning, J.B., Jr. (ed.) 1992. CRC Handbook of Avian Body Mass. CRC Press, Boca Raton, FL,
21 USA.
22 Dunstan, T.C. and J.F. Harper. 1975. Food habits of bald eagles in north-central Minnesota.
23 Journal of Wildlife Management 39:140-143.
24 Elliott, J.E., RW. Butler, R.J. Norstrom, and P.E. Whitehead. 1989. Environmental contaminants
25 and reproductive success of great blue herons in British Columbia 1986-87. Environmental
26 Pollution 59:91-114.
27 Elliott, J.E., R.J. Norstrom, A. Lorenzen, L.E. Hart, H. Philibert, S.W. Kennedy, J.J. Stegeman,
28 G.D. Bellward, and K.M. Cheng. 1996. Biological effects of polychlorinated dibenzo-p-dioxins,
29 dibenzofurans, and biphenyls in bald eagle {Haliaeetus leucocephalus) chicks. Environmental
3 0 Toxicology and Chemistry 15:782-793.
31 Elliott, J.E., M.L. Harris, L.K. Wilson, P.E. Whitehead, and R.J. Norstrom. 2001.
32 Monitoring Temporal and Spatial Trends in Polychlorinated Dibenzo-p-dioxins (PCDDs) and
33 Dibenzofurans (PCDFs) in Eggs of Great Blue Heron (Ardea herodias) on the Coast of British
34 Columbia, Canada, 1983-1998. Amfoo 30(7):416-428.
35 Elliott, J.E., and M.L. Harris. In Press. An ecotoxicological assessment of chlorinated
36 hydrocarbon effects on bald eagle populations. Reviews in Toxicology.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
1 EPA (U.S. Environmental Protection Agency). 1993. Wildlife Exposure Factors Handbook.
2 Volume 1 of 2. EPA/600/R-93/187a. Office of Research and Development, Washington, DC,
3 USA.
4 Fernie, K.J., J.E. Smits, G.R. Bortolotti, and D.M. Bird. 2001a. In ovo exposure to
5 polychlorinated biphenyls: reproductive effects on second-generation American kestrels.
6 Archives of Environmental Contamination and Toxicology 40:544-550.
7 Fernie, K.J., J.E. Smits, G.R. Bortolotti, and D.M. Bird. 2001b. Reproductive success of
8 American kestrels exposed to dietary polychlorinated biphenyls. Environmental Toxicology and
9 Chemistry 20:776-781.
10 Gibbs, J.P., S. Melvin, and F.A. Reid. 1992. American Bittern (Botaurus lentiginosus). In The
11 Birds of North America, No. 18. A. Poole, P. Stettenheim, and F. Gill (eds). The Birds of North
12 America, Inc. Philadelphia, PA, USA.
13 Giesy, J.P., WW. Bowerman, M.A. Mora, D.A. Verbrugge, R.A. Othoudt, J.L. Newsted, C.L.
14 Summer, R.J. Aulerich, S.J. Bursian, J.P. Ludwig, G.A. Dawson, T.J. Kubiak, D.A. Best, and
15 D.E. Tillitt. 1995. Contaminants of fishes from Great Lakes-influenced sections and above dams
16 of three Michigan rivers: III. Implications for health of bald eagles. Archives of Environmental
17 Contamination and Toxicology 29(3):309-321.
18 Gilbert, R.O. 1987. Statistical Methods for Environmental Pollution Monitoring. Van Nostrand
19 Reinhold Company, New York, NY. 320 pp.
20 Godin, A.J. 1977. Wild Mammals of New England. The John Hopkins University Press, Baltimore,
21 MD, USA, and London, UK.
22 Griffith, L.A. and J.E. Gates. 1985. Food habits of cave-dwelling bats in the central
23 Appalachians. Journal of Mammalogy 66:451-460.
24 Gulo, T. 2001. Massachusetts Division of Fisheries and Wildlife. Personal Communication to
25 Woodlot Alternatives, Inc.
26 Haffner, G.D., M. Tomczak, and R. Lazar. 1994. Organic contaminant exposure in the Lake St.
27 Clair food web. Hydrobiologia 281:19-27.
28 Haywood, D.D. and R.D. Ohmart. 1986. Utilization of benthic-feeding fish by inland breeding
29 bald eagles. Condor 88:35-42.
30 Helander, B., M. Olsson, and L. Reuterhard. 1982. Residue levels of organochlorine and mercury
31 compounds in unhatched eggs and their relationship to breeding success in white-tailed sea
32 eagles in Sweden. Holarctic Ecology 5:349-366.
33 Hoffman, D.J., B.A. Rattner, C.M. Bunck, A. Krynitsky, H.M. Ohlendorf, and R.W. Lowe. 1986.
34 Association between PCBs and lower embryonic weight in black-crowned night herons in San
35 Francisco Bay. Journal of Toxicology and Environmental Health 19:383-391.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC H 71
-------
1 Hoffman, D.J., G.J. Smith, and B.A. Rattner. 1993. Biomarkers of contaminant exposure in
2 common terns and black-crowned night herons in the Great Lakes. Environmental Toxicology and
3 Chemistry 12:1095-1103.
4 Hoffman, D.J., M.J. Melancon, P.N. Klein, C.P. Rice, J.D. Eisemann, R.K. Hines, J.W. Spann,
5 and G.W. Pendleton. 1996. Developmental toxicity of PCB 126 (3,3,4,4,5-pentachlorobuphenyl)
6 in nestling American kestrels (Falco sparverius). Fundamental and Applied Toxicology 34:188-
7 200.
8 Khera, K.S. and J.A. Ruddick. 1973. Polychlorodibenzo-p-dioxins: Perinatal effects and the
9 dominant lethal test in Wistar rats. Toxicology 120:70-84.
10 Krusic, R.A., Y. Mariko, C.D. Neefus, and P.J. Pekins. 1996. Bat habitat use in the White
11 Mountain National Forest. Journal of Wildlife Management 60(3):625-631.
12 Kurta, A. 1995. Mammals of the Great Lakes Region. The University of Michigan Press. Ann
13 Arbor, MI. 392 pp.
14 Laporte, P. 1982. Organochlorine residues and eggshell measurements of great blue heron eggs
15 from Quebec. Colonial Waterbirds 5:95-103.
16 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S. Petron,
17 B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-of-
18 Evidence Workgroup: A weight of evidence approach for evaluating ecological risks. Human and
19 Ecological Risk Assessment 2(2):277-304.
20 Mitchell, C.A., D.H. White, and T.E. Kaiser. 1981. Reproductive success of great blue herons at
21 Nueces Bay, Corpus Christi, Texas. Bulletin of Texas Ornithological Society 14:18-21.
22 MNHESP (Massachusetts Natural Heritage and Endangered Species Program). 1984. Special
23 Concern Species of Massachusetts - Small-footed Myotis Fact Sheet prepared by the
24 Commonwealth of Massachusetts, Division of Fisheries and Wildlife. Westborough, MA, USA.
25 Moore, D.R.J., B.E. Sample, G.W. Suter, B.R. Parkhurst, and R.S. Teed. 1999. A probabilistic
26 risk assessment of the effects of methylmercury and PCBs on mink and kingfishers at East Fork
27 Poplar Creek, Oak Ridge, TN. Environmental Toxicology and Chemistry 18:2941-2953.
28 Nagy, K.A., I. A. Girard, and T.K. Brown. 1999. Energetics of free-ranging mammals, reptiles,
29 and birds. Annu. Rev. Nutr. 19:247-277.
30 Nagy, K.A. 1987. Free metabolic rate and food requirement scaling in mammals and birds.
31 Ecological Monographs 57:111-128.
32 Nobel, D.G. and J.E. Elliott. 1990. Levels of contaminates in Canadian raptors, 1966 to 1988;
33 effects and temporal trends. Canadian Field Naturalist 104:222-243.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC j^
-------
1 Peakall, J.D., D.G. Noble, J.E. Elliott, J.D. Somers, and G. Erickson. 1990. Environmental
2 contaminants in Canadian Peregrine Falcons, Falco peregrines: a toxicological assessment.
3 Canadian Field-Naturalist 104:244-254.
4 Rattner, B.A., D.J. Hoffman, M.J. Melancon, G.H. Olsen, S.R. Schmidt, and K.C. Parsons. 2000.
5 Organochlorine and metal contaminant exposure and effects in hatching black-crowned night
6 herons (Nycticorax nycticorax) in Delaware Bay. Archives of Environmental Contamination and
7 Toxicology 39:38-45.
8 Senthilkumar, K., N. Iseki, S. Hayama, J. Nakanishi, and S. Masunaga. 2001. Polychlorinated
9 Dibenzo-p-Dioxins, Dibenzofurans, and Dioxin-like Polychlorinated Biphenyls in Livers of
10 Birds from Japan. Archives of Environmental Contamination and Toxicology 42:244-255.
11 Sparschu G.L., F.L. Dunn, R.W. Lisowe, and V.K. Rowe. 1971. Study of the effects of high
12 levels of 2,4,5-trichlorophenoxyacetic acid on fetal development in the rat. Food and Cosmetics
13 Toxicology 9:527-530.
14 Spencer, F. 1982. An assessment of the reproductive toxic potential of Aroclor 1254 in female
15 Sprague-Dawley rats. Bulletin of Environmental Contamination and Toxicology 28:290-297.
16 Stalmaster, M.V. and J.A. Gessaman. 1982. Food consumption and energy requirements of
17 captive bald eagles. Journal of Wildlife Management 46(3):646-654.
18 Stalmaster, M.V. and J.A. Gessaman. 1984. Ecological energetics and foraging behavior of
19 overwintering bald eagles. Ecological Monographs 54(4):407-428.
20 Stalmaster, M.V. and R.G. Plettner. 1992. Diets and foraging effectiveness of bald eagles during
21 extreme winter weather in Nebraska. Journal of Wildlife Management 56:355-367.
22 Stratus (Stratus Consulting, Inc.). 1999. Injuries to Avian Resources, Lower Fox River/Green Bay
23 Natural Resource Damage Assessment. Final Report. Prepared for U.S. Fish & Wildlife Services,
24 U.S. Department of the Interior, and the U.S. Department of Justice. May 7, 1999.
25 Todd, C.S., L.S. Young, RB. Owen, and F.J. Gramlich. 1982. Food habits of bald eagles in Maine.
26 Journal of Wildlife Management 46:636-645.
27 Van den Berg, M., L. Birnbaum, A.T.C. Bosveld, B. Brunstrom, P. Cook, M. Feeley, P. Giesy,
28 A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C. Larson, F.W.R. Leeuwen, D. Tillitt,
29 M. Tysklind, M. Younes, F. Waern, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs)
30 for PCBs, PCDDs, PCDFs for humans and wildlife. Environmental Health Perspectives
31 106:775-792.
32 Watson, J.W., M.G. Garrett, and R.G. Anthony. 1991. Foraging ecology of bald eagles in the
33 Columbia River estuary. Journal of Wildlife Management 55:492-499.
34 Wiemeyer, S.N., C.M. Bunck, and C.J. Stafford. 1993. Environmental contaminants in bald
35 eagle eggs — 1980-84 — and further interpretations of relationships to productivity and shell
36 thickness. Archives of Environmental Contamination and Toxicology 24:213-227.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC j j
-------
1 Woodford, J.E., W.H. Karasov, M.W. Meyer and L. Chambers. 1998. Impacts of 2,3,7,8-TCDD
2 exposure on survival, growth, and behavior of ospreys breeding in Wisconsin, USA.
3 Environmental Toxicology and Chemistry 17:1323-1331.
4 Zimmerman, G.S. and W.E. Glanz. 2000. Habitat use by bats in eastern Maine. Journal of Wildlife
5 Management 64:1032-1040.
MK01\0:\20123001.096\ERA_PB\ERA_PB_11.DOC
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
12. RISK SUMMARY
Highlights of Risk Summary
¦ Total PCBs and other COCs in the PSA of the Housatonic River pose
unacceptable risks to some assessment endpoints.
¦ Risk is high for benthic invertebrates, amphibians, and piscivorous mammals.
Confidence in this conclusion is high because (1) multiple lines of evidence with
concordant results were available; (2) models used to estimate risk were not
conservative; and (3) consideration of uncertainty indicates a high probability of
effects.
¦ Risk is moderate to high for some piscivorous birds, omnivorous and carnivorous
mammals, and high for selected threatened and endangered bird and mammal
species. There is uncertainty regarding these conclusions because corroborating
lines of evidence were generally not available.
¦ Risk is low to moderate for fish and confidence in this conclusion is moderate.
¦ Risk is low for insectivorous birds, but confidence in this conclusion is not high.
¦ Other species not included in the quantitative risk assessments may also be at
risk in the PSA.
¦ Assessment of risks to the most susceptible endpoints downstream of the PSA
indicates that benthic invertebrates, amphibians, warmwater and coldwaterfish,
mink, river otter, and bald eagles may be at risk.
12.1 OVERVIEW
The assessment of ecological risks of COCs in the Housatonic River to aquatic life and wildlife
is described in Sections 3 through 11 and in more detail in Appendices D through K. The
amount of information considered in this assessment is large, and the analyses and interpretation
complex. The purpose of this section is to summarize the major findings of the ERA and to
discuss the implications of these findings for biota in the Primary Study Area (PSA) and
downstream of the PSA.
Section 12.2 summarizes the risk assessment findings for each assessment endpoint. The first
part of this presentation (Section 12.2.1) discusses the results of the weight-of-evidence approach
for each of the 8 assessment endpoints evaluated in the risk assessment. The WOE approach is a
process by which measurement endpoints are related to an assessment endpoint to evaluate
whether significant risk is posed to the environment (Menzie et al. 1996). A formal WOE can
range from a simple qualitative assessment to a highly quantitative evaluation; however, no
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC -> rs -i 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
matter what form the WOE takes, it should provide documentation of the thought process used
when assessing potential ecological risk.
The term "line of evidence" as used in this ERA follows the definition of "Information derived
from different sources or by different techniques that can be used to describe and interpret risk
estimates" provided in the Guidelines for Ecological Risk Assessment (EPA 1998). Unlike the
term "weight-of-evidence," this definition does not imply assignment of qualitative or
quantitative weightings to information. The three general lines of evidence under which most
measurement endpoints fall are (Hull and Suter 1994; Suter et al. 1995):
¦ Biological (field) survey data that indicate the state of the receiving environment.
¦ Media toxicity data that indicate whether the contaminated media are toxic (i.e.,
laboratory or in situ toxicity testing).
¦ Single contaminant toxicity data that indicate the toxic effects of the concentration
measured in site media (e.g., exposure modeling).
Two or three general lines of evidence were considered in evaluating potential risk for most
assessment endpoints. The WOE approach used in this ERA for each of the assessment
endpoints follows the approach originally described in the Massachusetts Weight-of-Evidence
Special Report (Menzie et al. 1996).
Following the WOE discussion, Section 12.2.2 presents a discussion of hazard quotients (HQs)
that were calculated for each receptor for the two COCs of greatest concern, tPCBs and TEQ, to
facilitate comparison of risks between assessment endpoints in the PSA. The HQ analysis
includes estimates of uncertainty to provide an indication of both the magnitude of risk for each
COC receptor combination and the amount of uncertainty about each risk estimate.
Following the HQ analysis, the assessment of risks conducted for areas downstream of the PSA
is described in Section 12.2.3. As is apparent from the preceding sections of the ecological risk
assessment, risks to some assessment endpoints vary within the PSA as well as downstream, due
to the small-scale variability in sediment and, to a lesser degree, floodplain soil concentrations.
Section 12.2 concludes with a brief discussion of possible reasons for the differences in risk
between assessment endpoints for the most influential COCs, tPCBs and TEQ.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-2
-------
1 Section 12.3 discusses the broader implications of the risk assessment findings summarized in
2 Section 12.2. Issues addressed include:
3
4
5
6
7
¦ The risk assessment described in Sections 3 through 11 and Appendices D through K
focused the majority of quantitative analyses on selected species, termed
"representative species." There are, however, many other species that occur in the
watershed of the Housatonic River. Section 12.3 begins with a discussion of
estimates of the potential risks posed by COCs to these other species.
9
10
11
12
8
¦ In addition to effects on survival, growth, and reproduction of individuals in the
Housatonic River, there are a number of other possible impacts of COCs on aquatic
life and wildlife that were not addressed in the individual assessments (i.e. indirect
effects, narrowing of the genetic pools for exposed species). These topics are briefly
addressed in Section 12.3.
13 Section 12.4 provides a discussion acknowledging that there are many sources of uncertainty in
14 an ecological risk assessment, even in assessments (such as this ERA) that have a great deal of
15 available information. These sources of uncertainty can have important consequences during the
16 risk management process; therefore, it is important to describe them. The preceding sections and
17 the appendices described the sources of uncertainty for each assessment endpoint and their
18 potential influence on risk estimates and the confidence in those risk estimates. Section 12.4
19 summarizes the most important sources of uncertainty, particularly those that were common to
20 many assessment endpoints.
21 Section 12.5 concludes with a listing of the major findings of the Housatonic River ecological
22 risk assessment.
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs ^ 7/11/2003
-------
1 12.2 SUMMARY OF THE ASSESSMENT ENDPOINT CONCLUSIONS
2 The problem formulation stage of the ERA (Section 2.8) identified the assessment endpoints
3 considered important in the Housatonic River ERA. Each of the assessment endpoints was
4 evaluated and conclusions made regarding the potential for adverse effects (see Sections 3
5 through 11 and Appendices D through K). Table 12.2-1 provides a short summary for each of
6 the assessment endpoints and the conclusions reached in the ERA for the PSA. Tables 12.2-2 to
7 12.2-16 indicate the results of the weight-of-evidence assessments for each assessment endpoint.
8 12.2.1 Results of Weight-of-Evidence Evaluation
9 12.2.1.1 Benthic Invertebrates
10 The WOE results for the benthic invertebrate assessment endpoint are shown in Table 12.2-2. In
11 this WOE table, the measurement endpoints for the three lines of evidence: water, sediment, and
12 tissue chemistry (C), toxicity tests (T), benthic community measures (B) are listed, as are the
13 weighting of the measurement endpoint, evidence of harm, and magnitude of response. This
14 table indicates that the majority of endpoints suggest some risk for benthic communities in both
15 coarse- and fine-grained sediment. The conclusion is that there is an intermediate to high risk to
16 much of the benthic community, as indicated by the WOE evaluation.
17 12.2.1.2 Amphibians
18 The results of the WOE assessment for amphibians are presented in Table 12.2-3. In the
19 amphibian WOE matrix, the measurement endpoints for the three lines of evidence: tissue
20 chemistry (C); wood frog toxicity tests (W) and leopard frog toxicity tests (L); and field surveys
21 (B) are listed. As shown in the table, many of the measurement endpoints indicated some degree
22 of risk. The potential for two amphibian studies conducted for GE to determine risk to
23 amphibians was judged to be undetermined due to limitations in the study designs. The only
24 endpoint that did not indicate potential risk was the earliest life stage wood frog toxicity
25 endpoint, for which there is a mechanistic explanation for the lack of response. Four endpoints
26 exhibited a high degree of risk combined with a moderate to high confidence rating.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-4
-------
1 Table 12.2-1
2
3 Ecological Assessment Endpoints and Conclusions for the
4 Primary Study Area Portion of the Lower Housatonic River
Receptor
Assessment
Endpoint
Conclusions
Benthic
Invertebrates
Community structure,
survival, growth, and
reproduction
The benthic invertebrate ERA demonstrates significant risk from
tPCBs based on a weight-of-evidence evaluation of multiple effects
endpoints. The pronounced toxicity (laboratory and in situ)
observed in PSA sediment was supported by toxicity identification
evaluation (TIE) findings, alterations to macroinvertebrate
community structure, and large exceedances of effects benchmark
values for invertebrate tissues, sediment, and water.
Amphibians
Community
condition, survival,
reproduction,
development, and
maturation
The amphibian ERA indicates significant risk to frogs from tPCBs
based on a weight-of-evidence evaluation of multiple effects
endpoints. The literature-derived tissue thresholds for tPCBs were
supported by site-specific toxicity studies, skewed sex ratios,
malformations, and other effects that implicated tPCBs as the causal
agent. Sediment toxicity tests indicated a correlation between level
of effect and tPCB concentration.
Fish
Survival, growth, and
reproduction
The fish ERA found significant potential risk to fish from tPCBs
and TEQ based on a weight-of-evidence evaluation. The tissue
thresholds identified in the literature and from site-specific toxicity
studies were exceeded by fish tissue concentrations measured in the
PSA for all representative species. For some species (e.g., yellow
perch), the majority of individual fish concentrations exceeded the
respective benchmarks for both tPCBs and TEQ. Despite the high
probability of risk, the magnitude of adverse responses is expected
to be low to intermediate. The observed fish tissue tPCB
concentrations did not exceed the derived effects levels by a large
factor (i.e., factor of 5 or more) in many samples, and studies
suggest that current effects to fish are not severe. The findings from
the literature reviews and site-specific toxicity studies are consistent
with the results of field studies, which indicate that populations of
fish in the PSA are not experiencing catastrophic effects.
Insectivorous Birds
Survival, growth, and
reproduction
The weight-of-evidence analysis indicates that insectivorous birds,
such as tree swallows and American robins, are likely at low risk in
the PSA as a result of exposure to tPCBs and TEQ. Risks to tree
swallows and robins in the PSA are predicted to be intermediate to
high based on modeling of exposure and effects, but field studies of
tree swallows and American robins detected no obvious adverse
reproductive effects to these birds in the PSA. The weight-of-
evidence assessment relied more heavily on the results of the site-
specific field studies, but the conclusion of low risk is uncertain
because the lines of evidence did not give concordant results.
O:\20123001,096\ERA_PB\ERA_PB_12. DOC 12 5
-------
Table 12.2-1
Ecological Assessment Endpoints and Conclusions for the
Primary Study Area Portion of the Lower Housatonic River
(Continued)
Receptor
Assessment
Endpoint
Conclusions
Piscivorous Birds
Survival, growth, and
reproduction
The weight-of-evidence analysis indicates that ospreys are at high
risk from exposure to tPCBs and intermediate risk from exposure to
TEQ in the Housatonic River PSA. In the PSA, exposure of ospreys
to tPCBs is greater than doses that caused adverse effects in the
most tolerant bird species studied. The conclusion of high risk to
ospreys is uncertain because only one line of evidence was
available. Belted kingfishers are considered to be at low risk as a
result of exposure to tPCBs and TEQ in the Housatonic River PSA.
While modeled exposure and effects indicated high risk for tPCBs
and intermediate risk for TEQ, a field study of kingfisher
productivity indicated that the birds were reproducing in the PSA.
The conclusion of low risk to kingfishers is uncertain because the
two lines of evidence did not give concordant results.
Piscivorous
Mammals
Survival, growth, and
reproduction
The weight-of-evidence analysis indicates that piscivorous
mammals (i.e., mink and river otter) are at intermediate to high risk
in the PSA as a result of exposure to tPCBs and TEQ. Evidence for
this conclusion includes limited sightings of mink and otter in the
PSA, except during winter, despite availability of appropriate
habitat and evidence that they are common in nearby reference
areas; results of the feeding study which showed effects on kit
survival and jaw lesions in surviving mink at a much smaller
fraction of fish in the diet (3.5%) than would be expected of mink
foraging in the PSA; and modeling of exposure and effects which
predicted severe risks to mink and otter foraging in the PSA. Risks
to mink and otter are likely to be elevated even for mink and otter
that forage only a small fraction (e.g., 10%) of their time in the
PSA.
Omnivorous and
Carnivorous
Mammals
Survival, growth, and
reproduction
The weight-of-evidence analysis indicates an intermediate risk for
red fox and short-tailed shrews exposed to tPCBs and TEQ in the
PSA. The field survey indicated that omnivorous and carnivorous
mammals, including short-tailed shrew and red fox, are common in
some areas of the PSA. In contrast, modeling of exposure and
effects predicts these animals to be at low to high risk as a result of
exposure to tPCBs and TEQ in the PSA. The population
demography field study suggested that short-tailed shrews are not
seriously affected by tPCB contamination, although a reanalysis of
the data did not fully support this conclusion. The conclusion of
intermediate risk is uncertain because of the uncertainty about
whether effects are occurring for two of the three lines of evidence.
Threatened and
Endangered
Species
Survival, growth, and
reproduction
Based on modeling of exposure and effects, bald eagles and
American bitterns are at risk as a result of exposure to tPCBs. Risk
from TEQ is intermediate for bald eagles and undetermined for
American bittern. The risks to small-footed myotis exposed to
tPCBs and TEQ are undetermined.
1
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs ^ 7/11/2003
-------
Table 12.2-2
Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related to Maintenance of a Healthy
Benthic Community
Measurement Endpoints
Weighting
Value (High,
Moderate,
Low)
Coarse-Grained Sediment
Fine-Grained Sediment
Evidence of
Harm (Yes, No,
Undetermined)
Magnitude (High,
Intermediate, Low)
Evidence of
Harm (Yes, No,
Undetermined)
Magnitude (High,
Intermediate, Low)
C. Chemical Measures
C-l. Concentration of PCB in overlying water in relation to
levels reported to be harmful to benthic invertebrates
Low/Moderate
Yes
Intermediate
Yes
Intermediate
C-2. Concentration of PCB in the sediment in relation to
levels reported to be harmful to benthic invertebrates
Low/Moderate
Yes
High
Yes
High
C-3. Concentration of PCB in invertebrate tissues in relation
to levels reported to be harmful to benthic invertebrates
Moderate
Yes
Intermediate
Yes
Intermediate
T. Toxicological Measures
T-l. Sediment toxicity to multiple invertebrate species, as
measured in laboratory toxicity tests
Moderate/
High
Yes
High
Yes
High
T-2. Sediment toxicity to multiple invertebrate species, as
measured in in situ toxicity tests
Moderate/
High
Yes
Intermediate
Yes
High
T-3. Indications of PCB as toxicity driver in toxicity
identification evaluations
Moderate
Undetermined
—
Yes
Intermediate
B. Benthic Community Measures
B-l. Abundance, richness, and biomass of invertebrates,
relative to reference stations of comparable substrate and
habitat (ANOVA)
Moderate
Yes
Intermediate
No
B-2. Benthic community structure, as assessed using a
multivariate assessment of key benthic metrics (rank analysis
and MDS)
Moderate
Yes
Intermediate
No
B-3. Water quality assessment using modified Hilsenhoff
Biotic Index (MHBI) indicator of organic pollution
Moderate
No
—
No
—
O:\20123001.096\ERA_PB\ERA_PB_12.DOC 10 n 7/11/2003
-------
Table 12.2-3
Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related to Maintenance of Amphibian
Populations in the Housatonic River PSA
Measurement Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No,
Undetermined)
Magnitude
(High, Intermediate,
Low)
C. Chemical Measures
C. Concentration of PCB in frog tissues in relation to levels reported to be harmful to
amphibians.
Moderate
Yes
Low
W. Wood Frog Toxicological Measures
W-l. Sediment toxicity to hatchling/late embryo life stages.
Mod/High
No
-
W-2. Sediment toxicity to larval life stages.
Mod/High
Yes
Intermediate
W-3. Sediment toxicity to late larval/metamorph life stage.
Mod/High
Yes
High
W-4. GE Study (juvenile wood frogs)
Low
Undetermined
-
L. Leopard Frog Toxicological Measures
L-l. Sediment toxicity to hatchling/late embryo life stages.
Mod/High
Yes
Low
L-2. Sediment toxicity to larval life stages.
Mod/High
Yes
High
L-3. Sediment toxicity to late larval/metamorph life stage.
Mod/High
Yes
High
L-4. Sediment toxicity to adult leopard frogs (reproductive health).
Mod/High
Yes
High
B. Biology
B-l. Vernal pool community study.
Mod/High
Yes
Low
B-2. GE leopard frog egg mass survey
Low/Mod
Undetermined
-
B-3. Anecdotal observations during collections for reproductive study.
Moderate
Yes
Low
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC 12 8 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
In addition, a population model was constructed for wood frogs to determine whether effects
from PCBs on individual wood frogs influence the populations within the PSA. A ten-year
simulation both with and without the effects of PCBs was conducted. The model demonstrated
that effects observed in the toxicity studies would result in population level impacts.
The conclusion is that there is a significant risk to amphibians as indicated by the preponderance
of the evidence, the relative weights of the measurement endpoints, and the population modeling.
The "no risk" value of measurement endpoint W-l does not diminish the overall conclusion,
because the study demonstrated that the embryo/early larval life stages are fairly insensitive to
the effects of maternally transferred PCBs relative to later juvenile life stages exposed to
contaminated media.
12.2.1.3 Fish
The WOE results for fish in the PSA are shown in Table 12.2-4. In the fish WOE matrix, the
measurement endpoints for the three lines of evidence: site-specific toxicity tests (A); fish tissue
chemistry (B); and field surveys (C) are listed. This table illustrates that the majority of
endpoints indicate, with a moderately high degree of confidence, that there are low magnitude
risks to fish in the PSA. Although a high probability of adverse impacts to fish from tPCBs
and/or TEQ is predicted throughout the PSA, the impacts predicted are for sensitive sublethal
endpoints (reproduction and development); mortality of adults is unlikely. Therefore, the
magnitude of impact is not predicted to be catastrophic in any reach; adverse effects, although
high in probability, are generally expected to be subtle. The field studies conducted in the PSA
(fish community and reproduction studies) support lack of catastrophic effects, but cannot be
used to assess lesser impacts.
12.2.1.4 Insectivorous Birds
The WOE results for exposure of insectivorous birds to tPCBs are presented in Table 12.2-5 and
for exposure to TEQ in Table 12.2-6. Two lines of evidence are presented in the table; the field
studies, and modeled exposure and effects. The results from the modeled exposure and effects
line of evidence suggest that tPCBs and TEQ pose intermediate to high risk to tree swallows
living in the PSA. However, the field study line of evidence suggests that, if effects are
O:\20123001.096\ERA_PB\ERA_PB_12.DOC 10 n 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Table 12.2-4
Evidence of Harm and Magnitude of Effects for Measurement Endpoints Related
to Maintenance of a Healthy Fish Community
Measurement Endpoints
Weighting
Value (High,
Moderate, Low)
Evidence of Harm (Yes,
No, Undetermined)
Magnitude (High,
Intermediate, Low)
A. Site-Specific Toxicity
Al. Reproductive success relative to reference
Mod/High
Yes
Low
A2. Reproductive success dose-response
High
Yes
Intermediate
B. Fish Tissue Chemistry
Bl. Observed fish tissue/Literature toxic levels
Mod
Yes
Low
B2. Observed fish tissue/Phase I toxic levels
Mod/High
Yes
Low
B3. Observed fish tissue/Phase II toxic levels
Mod/High
Yes
Low
C: Fish Community and Reproduction Studies
CI: EPA Study and GE Community Study
Low/Mod
Undetermined
-
C2: GE Reproduction Study
Low/Mod
Undetermined
-
Table 12.2-5
Evidence of Harm and Magnitude of Effects for Insectivorous Birds Exposed to
tPCBs in the Housatonic River PSA
Measurement Endpoints
Weighting Value
(High, Moderate, Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Field Study
High (Tree swallow)
Moderate/High
(American robin)
No (Tree swallow)
No (American robin)
Low (Tree swallow)
Low (American robin)
Modeled Exposure and
Effects
Moderate
Yes
High
Table 12.2-6
Evidence of Harm and Magnitude of Effects for Insectivorous Birds Exposed to
TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value (High,
Moderate, Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Field Study
High (Tree swallow)
Moderate/High
(American robin)
No (Tree Swallow)
No (American robin)
Low (Tree Swallow)
Low (American robin)
Modeled Exposure and
Effects
Moderate
Yes
Intermediate
O:\20123001.096\ERA_PB\ERA_PB_12.DOC lO 1A 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
occurring, they are minor. The uncertainty concerning the field-based threshold range for tPCBs
likely means that risks of this COC are overestimated for the PSA. Even the upper end of the
tPCB range is associated with equivocal evidence for adverse effects to tree swallows. For TEQ,
the threshold range is quite broad. The available evidence from field studies indicates that tree
swallows are tolerant to exposure to persistent organochlorines such as tPCBs and TEQ. If the
tree swallow threshold is near the upper end of the threshold range, then the current modeled
exposure and effects line of evidence is overestimating risks of TEQ to tree swallows. Thus, the
WOE assessment supports a finding of low risk for tree swallows exposed to tPCBs and TEQ in
the PSA. This conclusion, however, is uncertain because of the conflicting results in the WOE
assessment.
The results from the modeled exposure and effects lines of evidence suggest that tPCBs and TEQ
pose an intermediate to high risk to American robins inhabiting the PSA of the Housatonic River.
The American robin field study, however, suggests that reproductive success is not being
impaired by the tPCBs and TEQ in the PSA. The uncertainty in the modeled exposure and
effects line of evidence, outlined below, likely means the approach overestimates the risks of
tPCBs and TEQ to American robins in the PSA. The WOE assessment therefore supports a
conclusion of low risk to American robins exposed to tPCBs and TEQ in the PSA. This
conclusion, however, is uncertain because of the conflicting results in the WOE assessment.
12.2.1.5 Piscivorous Birds
The WOE analysis indicates that exposure of piscivorous birds, such as the belted kingfisher and
osprey (Tables 12.2-7 and 12.2-8), to tPCBs and TEQ in the PSA, could lead to adverse
reproductive effects in some species. The two lines of evidence used to support this conclusion
were the field study of kingfisher productivity and the comparison of modeled exposure with
effects to piscivorous birds.
For the assessment of risks to kingfishers, both lines of evidence were available. The modeled
exposure and effects line of evidence indicated that kingfishers in the PSA are likely to receive a
tPCB dose greater than what the most tolerant species known from the literature can be exposed
to without effects. For TEQ, the risk picture is less clear because the threshold range for this
COC is very wide and the exposure estimates for kingfishers fell within this range. Thus,
O:\20123001.096\ERA_PB\ERA_PB_12.DOC lO 1 1 7/11/2003
-------
1 Table 12.2-7
2
3 Evidence of Harm and Magnitude of Effects for Piscivorous Birds Exposed to
4 tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled Exposure and
Effects
M
Kingfisher - Yes
Osprey - Yes
Kingfisher - High
Osprey - High
Belted Kingfisher Field
Study (Henning 2002)
MZH
Kingfisher - No
Kingfisher - Low
5
6 Table 12.2-8
7
8 Evidence of Harm and Magnitude of Effects for Piscivorous Birds Exposed to
9 TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled Exposure and
Effects
M
Kingfisher - Yes
Osprey - Yes
Kingfisher - Intermediate
Osprey - Intermediate
Belted Kingfisher Field
Study (Henning 2002)
MZH
Kingfisher - No
Kingfisher - Low
10
11 without effects data specific to kingfishers, it is difficult to make definitive conclusions about the
12 risks of TEQ to this species. The field study of kingfisher productivity, however, indicated that
13 these birds are able to reproduce in the PSA. This line of evidence was given a higher weighting
14 than the exposure and effects modeling, despite concerns about the field study. Therefore,
15 kingfishers are considered to be at low risk in the PSA as a result of exposure to tPCBs and TEQ.
16 The conclusion of low risk to kingfishers is uncertain because the two lines of evidence did not
17 give concordant results.
18 For ospreys, only the modeled exposure and effects line of evidence was available to assess risk
19 to these birds. As with kingfishers, this line of evidence indicated that ospreys in the PSA are
20 likely to receive a tPCB dose that is greater than what the most tolerant species known in the
21 literature can be exposed to without effects. The risks due to exposure to TEQ are unclear, as the
22 estimates for exposure also fell within the toxicity threshold range. Ospreys, however, lack a
O:\20123001,096\ERA_PB\ERA_PB_12.DOC j2 ^2
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
site-specific study that investigated the effects of COCs in the PSA. The PSA contains suitable
habitat for ospreys, with abundant prey, raising the possibility that they are not resident in the
area because of contaminants. Ospreys are, therefore, considered to be at risk in the PSA as a
result of exposure to tPCBs and TEQ.
12.2.1.6 Piscivorous Mammals
The results of the WOE assessment for piscivorous mammals are presented for tPCB and TEQ,
respectively, in Tables 12.2-9 and 12.2-10. All three lines of evidence—field studies, feeding
study, and modeled exposure and effects—indicate that the elevated concentrations of tPCBs and
TEQ in the PSA of the Housatonic River are causing adverse effects of high magnitude to mink
and river otter. The field surveys indicate that mink and river otter are rarely present in the PSA,
except during winter, and likely have not established home territories close to the main channel
despite suitable mink and otter habitat. The MSU site-specific feeding study indicated that
feeding adult female mink with a diet containing as little as 3.51% fish from the PSA caused a
statistically significant reduction (46% compared to controls) in kit survival to 6 weeks of age.
Because mink in the wild typically consume between 20% or more fish in their diet, the
associated risk is correspondingly higher. In addition, other components of the mink diet in the
PSA (e.g., crayfish) have high concentrations of tPCBs and TEQ. Further, the jaw lesion study
indicated that erosion of the jaw occurs at even lower doses and exhibits a dose-response
relationship. Such effects could eventually lead to starvation. The occurrence of jaw lesions
coincides with the induction of Ah-receptor-regulated enzymes (ECOD and EROD) also in a
dose-response manner.
The high risks evident from the feeding study are further supported by the modeled exposure and
effects line of evidence. The estimated potential for exposure is so high that even individual
mink and otter that only forage in the PSA for short periods of time (less than or equal to 10% of
foraging time) are at an intermediate or higher risk from tPCBs and TEQ.
12.2.1.7 Omnivorous and Carnivorous Mammals
The WOE results for omnivorous and carnivorous mammals are shown in Table 12.2-11 for
tPCB and Table 12.2-12 for TEQ. Data from three lines of evidence were available, including
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC 1 O 1 ^ 7/11/2003
-------
1 Table 12.2-9
2
3 Evidence of Harm and Magnitude of Effects for Piscivorous Mammals Exposed to
4 tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No,
Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Surveys
EPA
Moderate/High
Yes
High
GE
Moderate
No
Low
Feeding Study
High
Yes
High
Modeled Exposure and
Effects
Moderate/High
Yes
High
5
6 Table 12.2-10
7
8 Evidence of Harm and Magnitude of Effects for Piscivorous Mammals Exposed to
9 TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate,
Low)
Field Surveys
EPA
Moderate/High
Yes
High
GE
Moderate
No
Low
Feeding Study
High
Yes
High
Modeled Exposure and
Effects
Moderate/High
Yes
High
10
11
O:\20123001,096\ERA_PB\ERA_PB_12.DOC j 2 J 4
-------
1 Table 12.2-11
2
3 Evidence of Harm and Magnitude of Effects for Omnivorous and Carnivorous
4 Mammals Exposed to tPCBs in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Field Surveys
Moderate/High
Undetermined
Low
Population
Demography Field
Study
Moderate/High
Undetermined (Shrew)
Intermediate
Modeled Exposure and
Effects
Moderate/High
Yes (Shrew)
Undetermined (Red Fox)
High
Intermediate
5
6 Table 12.2-12
7
8 Evidence of Harm and Magnitude of Effects for Omnivorous and Carnivorous
9 Mammals Exposed to TEQ in the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Field Surveys
Moderate/High
Undetermined
Low
Population
Demography Field
Study
Moderate/High
Undetermined (Shrew)
Intermediate
Modeled Exposure and
Effects
Moderate/High
No (Shrew)
Undetermined (Red Fox)
Low
Intermediate
10
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-15
-------
1 field surveys, a population demography field study of short-tailed shrew and exposure and
2 effects modeling. The weight-of-evidence analysis indicates an intermediate risk for short-tailed
3 shrews exposed to tPCBs and TEQ in the PSA. This conclusion, however, is uncertain because
4 of the lack of definitive findings as to whether effects are occurring in the field surveys and
5 population demography field study, and the lack of species-specific measures of effects in the
6 literature. The WOE also suggests, based on one line of evidence for red fox an intermediate
7 risk to fox exposed to tPCBs and TEQ in the PSA using a foraging rate of 50% in Reach 5, in
8 addition, measures of effects for fox were not available in the literature.
9 The field surveys, and conclusions made in a study of short-tailed shrew populations at the site
10 conducted for GE are not in agreement with the results from the modeling of exposure and
11 effects line of evidence. However, the results of the supplemental analyses of the data from the
12 GE study on survival of short-tailed shrews are in agreement with the modeling results,
13 suggesting that there are intermediate effects from exposure to COCs in the contaminated areas
14 of the PSA.
15 12.2.1.8 Threatened and Endangered Species
16 The results of the WOE evaluation for threatened and endangered species using a single line of
17 evidence, modeled site-specific exposures and effects, are shown in Table 12.2-13 and Table
18 12.2-14 for tPCBs and TEQ, respectively. The results of the risk characterization showed that
19 the highest risk to T&E species is to bald eagles and American bitterns from exposure to tPCBs.
20 The risk for adult bald eagles exposed to TEQ is low; however, risk to bald eagle eggs exposed
21 to TEQ is high.
22 The risk to American bittern was high for TEQ and high for eggs exposed to TEQ. The risk
23 category for small-footed myotis was high for both tPCB and TEQ. The risk range for small-
24 footed myotis, as determined by the probability bounds analysis, ranged from intermediate to
25 high for tPCBs and low to high for TEQ.
26 The weight-of-evidence analysis indicates that T&E species such as bald eagles, American
27 bitterns, and small-footed myotis are at risk in the PSA as a result of exposure to tPCBs
O:\20123001,096\ERA_PB\ERA_PB_12.DOC j 2 J g
-------
Table 12.2-13
Evidence of Harm and Magnitude of Effects for T&E Species Exposed to tPCBs in
Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled exposure and
effects, Bald Eagle
Moderate/High
Yes
High
Modeled exposure and
effects, American
Bittern
Moderate/High
Yes
High
Modeled exposure and
effects, Small-Footed
Myotis
Moderate/High
Undetermined
High
Table 12.2-14
Evidence of Harm and Magnitude of Effects for T&E Species Exposed to TEQ in
the Housatonic River PSA
Measurement
Endpoints
Weighting Value
(High, Moderate,
Low)
Evidence of Harm
(Yes, No, Undetermined)
Magnitude
(High, Intermediate, Low)
Modeled exposure and
effects, Bald Eagle
Moderate/High
Yes
Intermediate
Modeled exposure and
effects, American
Bittern
Moderate/High
Undetermined
High
Modeled exposure and
effects, Small-Footed
Myotis
Moderate/High
Undetermined
High
and TEQ. The risks for bald eagles and American bitterns exposed to tPCBs are high. The risks
to small-footed myotis exposed to tPCBs and TEQ are undetermined. Risk from exposure to
TEQ is intermediate for bald eagles and undetermined for American bittern.
12.2.2 Hazard Quotient Analyses
The assessments described in Sections 3 through 11 and Appendices D through K were
conducted using various lines of evidence including, in many cases, different measurement
O:\20123001.096\ERA_PB\ERA_PB_12.DOC 1 O 1 ^7 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
endpoints and effects metrics. It is clear that risks posed by COCs in the PSA vary between
species; however, the degree of variability is not clear from these discussions because of the
differing endpoints and metrics used. To facilitate comparison of risks among aquatic life and
wildlife receptors and to give a broad overview of the findings of the risk assessment, assessment
results were converted to probabilistic hazard quotients (HQs). An HQ is the expected
environmental concentration or dose of a contaminant divided by its estimated low or no toxic
effect concentration or dose. Higher quotients indicate greater risk. The methods used to
calculate the probabilistic HQs and the results of these analyses for each endpoint are discussed
in this section.
12.2.2.1 Aquatic Assessment Endpoints
12.2.2.1.1 Benthic Invertebrates
For benthic invertebrates, HQs were calculated for Reaches 5A, 5B, 5C, 5D, and 6. Using data
on concentrations of tPCBs in sediment, medians, means, 25th and 50th percentiles, and minimum
and maximum concentrations were calculated for each reach. Hazard quotients were calculated
by dividing each of these summary statistics by the effects benchmark for benthic invertebrates
exposed to tPCBs in sediment. The sediment effects threshold used in the derivation of the HQs
was 3 mg/kg tPCB, which represents a concentration above which significant adverse responses
were observed in site-specific toxicity tests (Section 3 and Appendix D). The results are plotted
in Figure 12.2-1. These results indicate that significant risk was observed in all reaches of the
PSA, with HQs for tPCBs above 1 for both mean and median tPCB concentrations. Predicted
risks were greatest in the upstream (Reach 5A) and Woods Pond (Reach 6) sediment. Due to the
considerable small-scale variation in sediment tPCB concentrations, HQs for the reaches span
about 4 orders of magnitude (approximately 0.01 to 100). Because benthic invertebrates are
much less mobile than fish and wildlife, they do not spatially and temporally integrate their
exposures. Thus, the hazard quotient results for benthic invertebrates indicate that the majority
of individuals are at risk (i.e., HQ > 1), but that some individuals in less contaminated areas are
not.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-18
-------
1000
100
e
o
s
O
¦a
•-
«
N
«
K
0.001
Max
Monte Carlo 75th
-1-
Mean
X
Median
—
Monte Carlo 25th
Min
T
0.01
n=2
Fish HQs are based on tissue concentrations;
others based on sediment tPCB.
<
in
m
in
u
in
Q
in
1
75
75
o
75
75
75
£
¦S
>
£
¦S
>
£
-£
>
o
t:
o
>
.a
o
O
o
o
o
s
s
C
o
m
1
o
1
o
m
1
o
m
C
o
CQ
<
GO
fin
I
£
T3
c3
2
<
Qh
I
£
T3
c3
2
<
Qh
i
£
2
C/2
<
Qh
i
£
m
T3
o
0
75
M
Oh
1
Qh
<
Qh
U
T3
o
0
75
M
1
Qh
<
Qh
i
p?
m
fin
>H
<
<
<
<
<
m
o
C/D
in
in
in
Qh
Qh
i
Qh
i
Qh
i
i
75
C3
75
C
75
C
N
2
2
Ph
m
2
£
£
£
o
m
£
£
o
4=
Oh
&
&
75
75
75
c3
a
a
a
o
S-H
75
75
c3
<
<
<
O
Qh
£
O
75
c3
m
a
m
a
o
m
o
a
£
1
o
OX)
0)
pXj
o
OX)
c3
c3
¦3
ft
a
<
WB-R = Whole body,
reconstituted fish tissue.
CM = Composite
WB = Whole body fish tissue.
Figure 12.2-1 Hazard Quotients for Aquatic Biota Exposed to tPCBs in the Housatonic River PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_12. DOC
12-19
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
12.2.2.1.2 Amphibians
For amphibians, HQs were calculated for Reaches 5A, 5B, 5C, 5D, and 6, using methods similar
to those used for benthic invertebrates. Hazard quotients were calculated by dividing summary
statistics for vernal pool sediment concentrations by the effects benchmark for amphibians
exposed to tPCBs in sediment (3 mg/kg tPCBs) (Section 4 and Appendix E). This approach does
not address adult leopard frog exposures that likely occur in river and backwater sediment. The
results are plotted in Figure 12.2-1. These results indicate significant risk in all reaches of the
PSA, with HQs above 1 for both mean and median tPCB concentrations. Predicted risks were
greatest in the upstream (Reach 5A) vernal pool habitats. Because of the variation in sediment
PCB concentrations between the vernal pools, HQs for the subreaches span about 4 orders of
magnitude (approximately 0.01 to 100). The hazard quotient results for amphibians indicate that
the majority of individuals are at risk (i.e., HQ > 1), with higher levels of risk (i.e., HQ > 5) in a
large percentage of vernal pools (about 50% in Reaches 5A and 5B).
12.2.2.1.3 Fish
For fish, HQs were calculated separately for the five representative warmwater species (Section
5 and Appendix F) by dividing summary statistics for exposure by the tissue effects benchmark
protective of all species of PSA fish (49 mg/kg tPCB; derived from site-specific toxicity studies).
The results are plotted in Figure 12.2-1. These results indicate that risk occurs in all reaches of
the PSA, with both mean and median HQs for tPCBs above 1 for adult fish (i.e., whole body
reconstituted tissue concentrations of some species). Predicted risks were greatest in adult fish
and in predator fish at the top of the food web. Due to the variation in fish tissue tPCB
concentrations, hazard quotients for the reaches span about an order of magnitude, with most HQ
values between 0.3 and 3; the lower bound of this range represents primarily younger age classes
that have not yet accumulated their maximum tPCB burdens and fish species near the lower end
of the food chain. Thus, the hazard quotient results for fish indicate that predatory species are at
risk (i.e., HQ > 1) once they reach their maximum adult tPCB concentrations. The ERA
indicates that these HQs are indicative of sublethal (e.g., reproductive and developmental)
responses to offspring; the pathway for the manifestation of effects is through the maternal
transfer of tPCBs to eggs. Acute mortality to adults is not expected for most fish.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-20
-------
1 In addition to tPCBs, fish HQs were derived and plotted for TEQ (Figure 12.2-2). The effects
2 benchmark applied in this analysis was derived from the site-specific toxicity studies (42 ng/kg
3 TEQ) (Appendix F). The magnitudes and probabilities of risk for TEQ are generally similar to
4 tPCB risks.
5 12.2.2.1.4 Summary
6 For aquatic receptors (benthic invertebrates, amphibians, and fish), the HQs presented in Figures
7 12.2-1 and 12.2-2 are not conservative. Although sensitive species were considered in the
8 derivation of the effects thresholds, no additional safety factors were used to estimate the effects
9 metrics. The thresholds used in HQ calculations represent levels demonstrated to cause adverse
10 responses to organisms in site-specific studies. Thus, HQ exceedances of 1 are cause for
11 concern.
12 12.2.2.2 Wildlife Assessment Endpoints
13 For wildlife, probabilistic HQs were calculated as follows:
14 ¦ The distributions from the Monte Carlo analyses for total daily intake of COCs by
15 representative species were each divided by the corresponding effects metrics used to
16 estimate risks in Sections 7 through 11 and Appendices G through K. In the case of a
17 dose-response curve effects metric (e.g., mink exposed to tPCBs), the effects metric
18 was specified as a uniform distribution of dose ranging from 10 to 20% effect. A
19 similar approach was used for NOAEL-LOAEL ranges (e.g., bald eagles exposed to
20 tPCBs), field-based effects metrics (e.g., tree swallows exposed to tPCBs), and
21 threshold ranges (e.g., kingfishers exposed to TEQ).
22 "A similar approach was used with the results of the probability bounds analysis,
23 except that the effects metric was specified as a distribution-free range.
24 ¦ The analyses were done for both tPCBs and TEQ.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-21
-------
10
a
o
s
O
¦a
«
K
0.1
Max
Monte Carlo 75th
-1-
Mean
X
Median
—
Monte Carlo 25th
Min
T
CfT'
-o
to y—V
I s
e
CLh
X
o
!-h
(L)
CLh
£
jd
"53
>H
u
3
o
U
CO
OD co
b cG
j m
WB-R = Whole body,
reconstituted fish tissue.
WB = Whole body fish tissue.
Figure 12.2-2 Hazard Quotients for Fish Exposed to 2,3,7,8-TCDD TEQ in the Housatonic River PSA
MK01 |O:\20123001.096\ERA_PB\ERA_PB_12. DOC
12-22
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
¦ Modified box-and-whisker plots were developed for each representative species in the
PSA. For species with smaller foraging ranges, the analyses were done for different
areas within the PSA. Included in the plots (Figures 12.2-3 and 12.2-4) are the
median HQ from the Monte Carlo analysis (the thick line bisecting each box), the
mean HQ from the Monte Carlo analysis (star symbol), the 25th and 75th percentile
HQs from the Monte Carlo analysis (the bottom and top of each box), the 10th
percentile HQ from the lower bound of the probability bounds analysis (bottom
whisker), and the 90th percentile from the upper bound of the probability bounds
analysis (top whisker).
¦ Probabilistic HQ plots were also developed for tree swallows exposed to tPCBs using
measured concentrations in 12-day nestlings (data from Custer 2002) as an estimate
of exposure. This was done to facilitate comparison of risks using the microexposure
modeling approach and using measured concentrations from birds in the PSA.
Insufficient data were available to perform a similar analysis for TEQ. For the
measured concentrations data, an empirical histogram was specified for each PSA
location using the available 3 years of data for the Monte Carlo portion of the
analyses. The empirical histograms were divided by the same threshold range as used
for HQs based on the microexposure model outputs. For the probability bounds
analyses, 95% Kolmogorov-Smirnov confidence limits were calculated for each
empirical histogram. Both the upper and lower confidence limits were divided by the
same threshold range as used for HQs based on the microexposure model outputs.
Using the results of these analyses, modified box-and-whisker plots were developed
as previously described (Figures 12.2-3 and 12.2-4).
¦ In addition to plots developed for mink and otter exposed to tPCBs and TEQ using
the results of literature-based dose-response curves, plots were also developed using
the results of the mink feeding study conducted by Bursian et al. (2003). In this case,
the denominator was the NOAEL to LOAEL range from Bursian et al. (2003), rather
than the 10 and 20% effects doses from the literature-based dose-response curve.
Otherwise, the approach was as described above. The results are shown in Figures
12.2-3 and 12.2-4.
For wildlife species other than tree swallow, mink, and river otter, it was not possible to derive
field-based or feeding-study-based HQs because the required data were not available.
Unlike traditional HQs, the probabilistic HQs presented in Figures 12.2-3 and 12.2-4 are not
biased to be conservative. No safety factors were used to estimate the effects metrics (except in
the case of the bald eagle), and uncertainties regarding exposure model inputs were explicitly
propagated through the probability bounds and Monte Carlo analyses. Thus, HQ exceedances of
1 are cause for concern.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-
-------
10000
1000
100
s
a
•-P
o
s
a
¦s
10
0.1
0.01
¥
US
T~<#
*
Upper P-Bound 90th Percentile
Monte Carlo 75th Percentile
Monte Carlo Mean
Monte Carlo Median
Monte Carlo 25th Percentile
Lower P-Bound 10th Percentile
? B
$
&
E
&
K
3
£
Pd
J
Z
m
&
CQ
&
5
6
s
&
s
&
Pd
00
c
&
00
c
o
Q
in
c
tZ5
tin
cf
Ji
c
U
&
Q
c
c
O
Q
<
tZ5
Oh
S-
-------
1000
100
.H
"-C
o
¦o
N
X
0.1
0.01
0.001
o.oooi
Upper P-Bound 90th Percentile
Monte Carlo 75th Percentile
Monte Carlo Mean
Monte Carlo Median
Monte Carlo 25th Percentile
10
04
04
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
In summary, the wildlife HQs presented in Figures 12.2-3 and 12.2-4 indicate the following:
¦ Wildlife risks from tPCBs are generally higher than risks from TEQ by 1 to several
orders of magnitude.
¦ The comparison of HQ plots for tree swallows indicates that risks are higher using the
results of the microexposure model than is the case using measured concentrations in
tree swallow nestlings. Thus, the microexposure model appears to be overestimating
exposure for tPCBs. The HQ plots using measured concentrations in nestlings
supports the weight-of-evidence conclusion that risks of tPCBs to tree swallows is
low (see Section 7, Appendix G and Table 12.2-1).
¦ Wildlife risks from tPCBs and TEQ are highest for mink and river otter, with mean
and median HQs between 100 and 500 for tPCBs, and between 5 and 10 for TEQ
using the results of the literature-based dose-response curve. The HQs for tPCBs
were lower when the results of the feeding study were used to derive the effects
range. However, this difference is no longer apparent for TEQ, supporting the
hypothesis presented in Section 9 and Appendix I that the congener mixture in
Housatonic River prey in the PSA is less toxic than the commercial mixtures Aroclor
1260 and 1254.
¦ The risks from tPCBs and TEQ to most wildlife species are uncertain to the extent
that the bottom or top whiskers span 1. The whiskers are the extreme representations
of uncertainty because they are tail outputs from the probability bounds analyses, a
technique designed to propagate all forms of uncertainty (e.g., inability to precisely
specify distribution type or parameter values for a distribution). Thus, the boxes in
Figures 12.2-3 and 12.2-4 should be interpreted as representing a reasonable range
within which the HQ estimate occurs for the receptor of interest, and the whiskers
should be interpreted as representing the extremes within which the HQ could occur.
The HQ analyses for aquatic biota and wildlife indicate that risks of tPCBs vary widely between
representative species and assessment endpoints. Section 12.2.3 explores the fundamental
reasons why this might occur.
12.2.3 Risk Assessment Downstream of Woods Pond
Because of the reduced levels of contaminant concentrations downstream of the PSA and
significant shifts in aquatic habitat types associated both with river gradient and location of
dams, a different approach than that applied in the PSA was followed to assess potential
ecological risks of tPCBs to biota in areas downstream of Woods Pond. The assessment of
potential ecological risks was conducted using mapping (GIS) techniques and threshold
concentrations that indicate potential risk for six taxonomic groups selected based on the
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-26
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
outcome of the evaluations performed in the PSA and the habitat characteristics found
downstream. These groups are benthic invertebrates, amphibians, warmwater fish, trout, mink,
otter, and bald eagles.
For each of these groups, a maximum acceptable threshold concentration (MATC) for tPCBs
was developed based primarily on the detailed risk assessment performed for the PSA. Each
MATC was then compared to media-specific data for areas downstream of Woods Pond to Long
Island Sound. Areas with MATC exceedances, indicating potential risk, were plotted on maps of
the river. The methods used for each of the six representative groups and the results of the
analyses are discussed in the following sections (see Table 12.2-15 for a summary of results).
Where insufficient data for the medium of interest were available for a reach to estimate the risk
for a species, other available data were examined to determine the likelihood of the
concentrations of tPCBs in the medium of interest being high enough to pose a risk. In all cases,
concentrations were lower than in other reaches with available data for the medium of interest;
therefore, it is unlikely that the receptor is at risk in those reaches for which data were not
available.
12.2.3.1 Benthic Invertebrates
For benthic invertebrates, the medium of interest was river sediment. An MATC of 3 mg/kg
tPCBs was used as a conservative measure of the potential for adverse affects to benthic
invertebrates downstream of Woods Pond (see Section 3 and Appendix D for a description of the
derivation of the MATC). This concentration was developed using multiple lines of evidence
(e.g., benthic community studies, in situ and laboratory toxicity testing, bioaccumulation testing,
Sediment Quality Triad) and was selected as the concentration at which some sensitive endpoints
exhibited adverse responses, but the magnitude of responses was not large. Above a
concentration of 3 mg/kg tPCBs, numerous endpoints indicated ecologically significant
responses, with many LC50/EC50 values falling in the range of 3 to 30 mg/kg tPCBs.
The MATC of 3 mg/kg tPCBs was compared to recent surficial sediment data downstream of
Woods Pond, and the results were plotted (Figure 12.2-5) to indicate samples above and below
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-27
-------
1 Table 12.2-15
2
3 Summary of the Assessment of Risks Conducted for Biota Exposed to tPCBs in
4 the Lower Housatonic River Below Woods Pond
Reach
Potentially at Risk (MATC Exceeded)
Benthic
Invertebrates
Amphibians
Warmwater
Fish
Trout
Mink
Otter
Eagle
7
Yes
Yes
No
Yes
Yes
Yes
No
8
Yes
Yes
No
No
suitable
habitat
Yes
Yes
Yes*
9
No
Yes
No
Yes
Yes
Yes
No
10
No
Limited
suitable
habitat
No
No
suitable
habitat
Yes
Yes
No
11
No
Limited
suitable
habitat
No
No
No
Yes
No
12
No
Limited
suitable
habitat
No
No
No
Yes
No
13
No
Limited
suitable
habitat
No
No
suitable
habitat
No
No
No
14
No
Limited
suitable
habitat
No
No
suitable
habitat
No
No
No
15
No
Limited
suitable
habitat
No
No
suitable
habitat
Insufficient
data but
unlikely
No
No
16
No
Limited
suitable
habitat
Insufficient
data but
unlikely
No
suitable
habitat
Insufficient
data but
unlikely
Insufficient
data but
unlikely
Insufficient
data but
unlikely
5 * Reach 8 is Rising Pond. Although eagles would be at risk foraging there, Rising Pond is smaller than an eagle's
6 foraging area; therefore, there is uncertainty about this risk estimate.
7
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-28
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
the MATC. Inverse distance weighting was used to interpolate sediment concentrations between
discrete sampling points, and the potential for risk to benthic invertebrates was shown by shading
the corresponding sections of the river channel (Figure 12.2-5).
In general, potential risks to benthic invertebrates occur in limited areas downstream of Woods
Pond to Rising Pond. These areas are depositional and tend to have higher concentrations of
tPCBs. Below Rising Pond, sediment does not contain concentrations of tPCBs that represent a
potential risk to benthic invertebrates.
12.2.3.2 Amphibians
Many species of amphibians are primarily exposed to PCBs in the floodplain, particularly in
vernal pools and other low-lying wet areas. A sediment/floodplain soil MATC of 3.0 mg/kg
tPCB (dry weight) was derived from the amphibian risk assessment conducted for the PSA
(Section 4 and Appendix E). Inverse distance weighting was used to interpolate tPCB
concentrations to the limit of the 100-year floodplain (10-year floodplain contours are not
available downstream of Woods Pond) using the 0 to 6 inch (0 to 15 cm) depth data from the
floodplain downstream of Woods Pond; a separate analysis was conducted for sediment in a
manner analogous to that described above for benthic invertebrates.
Areas where the 3.0 mg/kg tPCB threshold was exceeded, indicating potential risks to
amphibians due to PCBs in floodplain soil and sediment, are indicated in Figures 12.2-6 and
12.2-7. Several large floodplain areas of potential risk are located in the area between Woods
Pond and Rising Pond, with only small isolated areas of potential risk downstream of Rising
Pond. Downstream of the Massachusetts/Connecticut state line the risk mapping for amphibians
was conducted for sediment only because the concentrations in floodplain soil had decreased
below the MATC throughout Reach 9 upstream, and the extent of the floodplain is limited in
Connecticut.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-29
-------
NEW
YORK
MASSACHUSETTS
NOTES:
Risk to benthic invertebrates is based on a maximum acceptable threshold
concentration (MATC) of 3.0 mg/kg total PCB (tPCB) (dry weight) in surficial
sediments (0-6 inches).
LEGEND:
~ Town
Reach Break
BENTHIC INVERTEBRATE RISK
Roads
O <3 mg/kg
• >=3 mg/kg
I Housatonic River Basin Hydrology
State Boundary
w^prE
s
0 2 4 6 8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 12.2-5
ASSESSMENT OF RISK TO
BENTHIC INVERTEBRATES
EXPOSED TO tPCBs
DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - benthos d4-71 o:\gepitt\epsfiles\plots\in\benthos_risk_12-2-5.eps 110:42 AM, 7/10/2003 |
-------
NOTES:
Risk to amphibians is based on a maximum acceptable threshold concentration
(MATC) of 3.0 mg/kg total PCB (tPCB) concentrations (dry weight) in surficial
floodplain soils (0-6 inches). All available sample data were used to calculate
interpolated PCB concentrations in unsampled areas within the 100-year
floodplain, using an inverse distance weighting (IDW) approach.
LEGEND:
~
Town
AMPHIBIAN
RISK
Reach Breaks
O
•
< 3 mg/kg
>= 3 mg/kg
< 3 mg/kg
Roads
Housatonic River Basin Hydrology
>= 3 mg/kg
~
State Boundary
S
0.5 0 0.5 1 1.5 2 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 12.2-6
ASSESSMENT OF RISK
TO AMPHIBIANS IN
FLOODPLAIN EXPOSED TO
tPCBs DOWNSTREAM OF
WOODS POND IN
MASSACHUSETTS
| O:\gepitt\aprs\era_species.apr | layout - amphibs FLDe4-41 o:\gepitttepsfiles\plots\in\amphibs_fld_risk_12-2-6.eps 110:38 AM, 7/10/20031
-------
MASSACHUSETTS
NEW
YORK
NOTES:
Risk to amphibians is based on a maximum acceptable
threshold concentration (MATC) of 3.0 mg/kg total PCB (tPCB)
concentrations (dry weight) in surficial sediment samples (0-6 inches).
LEGEND:
AMPHIBIAN RISK
O <3 mg/kg
• >=3 mg/kg
< 3 mg/kg
>= 3 mg/kg
~ Town
Reach Breaks
Roads
Housatonic River Basin Hydrology
State Boundary
w^prE
s
0 2 4
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 12.2-7
ASSESSMENT OF RISK
TO AMPHIBIANS IN SEDIMENT
EXPOSED TO tPCBs
DOWNSTREAM OF WOODS
POND IN CONNECTICUT
| O:\gepitt\aprs\era_species.apr | layout - amphibs SEDS e4-51 o:\gepitttepsfiles\plots\in\amphibs_sed_risk_12-2-7.eps 110:40 AM, 7/10/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
12.2.3.3 Warmwater Fish
As was done for the PSA, risk to fish was evaluated based on concentrations of tPCBs in fish
tissue. An MATC of 49 mg/kg tPCB in tissue (whole body, wet weight) developed for the PSA
based on site-specific effects (Section 5 and Appendix F) was applied to areas downstream of
Woods Pond using the available (e.g., bass, perch, sunfish) tissue data for warmwater species.
Each downstream reach (Reaches 7 through 16) was evaluated as a unit, and the mean adult fish
tissue concentration in each reach was compared with the MATC to determine potential risk.
Only data collected since 1998 were used in this analysis. The results are shown in Figure
12.2-8.
In some cases, it was necessary to extrapolate from the available raw data to develop an adult
tissue concentration for comparison with the MATC. Young-of-year (YOY) largemouth bass
data were extrapolated up to estimated adult concentrations by applying a factor of 3.5; similarly,
yellow perch YOY data were scaled up by a factor of 2.5. Both ratios were calculated from the
more extensive database available for the PSA. In addition, the majority of filet data were
extrapolated to estimated whole body concentrations using the factor of 2.3 developed by
Bevelhimer et al. (1997). Brown bullhead filet data were extrapolated to estimated whole body
concentrations using the factor of 1.5 developed by EPA (1999a).
The evaluation of risk to warmwater fish species downstream of Woods Pond indicated that no
risks were indicated in any of the reaches below the PSA.
12.2.3.4 Trout
Trout were evaluated separately from warmwater fish species because of significant differences
in habitat requirements and in the sensitivity of some trout species to tPCBs documented in the
literature. Trout also tend to have higher tPCB concentrations because of their higher lipid
content. The site-specific studies did not indicate large differences between the effects threshold
for rainbow trout and warmwater species (Section 5 and Appendix F), but the strain of rainbow
trout used in the site-specific toxicity tests (Buckler 2002) is less sensitive than other strains used
widely in toxicity testing. Furthermore, there are other trout species found downstream of the
PSA (e.g., brown trout) for which sensitivity has not been assessed. Given
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs ^ 7/11/2003
-------
NEW
YORK
MASSACHUSETTS
CONNECTICUT
NOTES:
Risk to warmwater fish is based on a maximum acceptable threshold
concentration (MATC) of 49 mg/kg total PCB (tPCB) concentrations
(wet weight) in whole body tissue.
* Only fish collected in 1998 to the present (2002) were included.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
* Brown bullhead filet samples were scaled by 1.5.
* Warm water filet samples were scaled by 2.3.
LEGEND:
~
Town
WARM WATER FISH RISK
A/
Reach Break
< 49 mg/kg
Roads
>= 49 mg/kg
r~^i
Housatonic River Basin Hydrology
~
State Boundary
w^prE
s
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 12.2-8
ASSESSMENT OF RISK
TO WARM WATER FISH
EXPOSED TO tPCBs
DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - fish warm f4-101 o:\gepitt\epsfiles\plots\in\warm_fish_12-2-8.eps 110:48 AM, 7/10/20031
-------
1 that some trout species have been documented to have greater sensitivity to PCBs and dioxins
2 (Walker et al. 1994; Zabel et al. 1995), relative to the warmwater species considered in the
3 development of the 49 mg/kg tPCB warmwater MATC, a factor of 4 was applied in recognition
4 of these potential interspecies differences. Therefore, a tissue MATC of 12 mg/kg tPCBs (whole
5 body, wet weight) was derived for trout.
6 Because of the more limited database for trout, a number of extrapolations were necessary to
7 convert available warmwater fish data and/or trout filet data to estimated whole body
8 concentrations for trout. These extrapolations included all of the extrapolation factors discussed
9 above for warmwater fish species and an additional extrapolation factor of 2.0 to estimate trout
10 whole body tPCB concentrations from measured tPCB concentrations in warmwater fish
11 samples. This factor was developed from the available smallmouth bass and trout data from the
12 Trout Management Area (Reach 11, in part) in Connecticut. The Amrhein et al. (1999) factor of
13 1.47 was used to convert the trout filet data to estimated whole body concentrations.
14 The results of this evaluation are shown in Figure 12.2-9. Trout are potentially at risk in Reaches
15 7 and 9, but not in reaches with suitable habitat downstream. This assessment has high
16 uncertainty due to the number of extrapolations required and the low magnitude of exceedance of
17 the MATC value. Potential risk to trout was not evaluated in Reach 8, Reach 10, and
18 downstream of Reach 12 due to lack of suitable trout habitat.
19 12.2.3.5 Mink
20 An MATC for mink of 2.65 mg/kg tPCBs in fish (whole body, wet weight) was derived from the
21 the geometric mean of the NOAEL and LOAEL developed by Bursian et al. (2003) in a site-
22 specific study of the toxicity of a diet containing Housatonic River fish to mink (Section 9 and
23 Appendix I).
24 Habitat suitability for mink was determined for the reaches downstream of the PSA (Appendix
25 A). According to this analysis, potential mink habitat is ubiquitous along Reaches 7 to 16 and
26 includes all areas except high gradient stream, calcareous rock cliff, cultural grassland,
27 agricultural cropland, and residential/industrial development.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-35
-------
Vest St<
BncT
£enox Dale
(41)
Glend
Btoekbridge
Sou
Berksare I
bremont Plain
ghts^
(23)
Hartsvilt
MASSACHUSETTS
Drough
\41\
NEW
YORK
Falls\VittaaS»
%r
CONNECTICUT
VesrOornwa^
202
(55)
ICandlewood
LakeJ
ki.
my
NOTES:
Risk to coldwater fish is based on a maximum acceptable threshold
concentration (MATC) of 12 mg/kg total PCB (tPCB) concentrations
(wet weight) in trout tissue (whole body).
* Only fish collected in 1998 to the present (2002) were included.
* Fish fillet samples were scaled by a factor of 2.3 to convert to whole
body.
* Where trout data were unavailable, averages by reach for warmwater
species were calculated and scaled by 2 for trout. In some reaches,
only warmwater fillets were available for conversions. The
fillets were first scaled up by a factor of 2.3, then 2 for coldwater fish.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
* Brown bullhead filet samples were scaled by 1.5.
* Warmwater filet samples were scaled by 2.3.
* Trout were not evaluated downstream of Bulls Bridge Dam based on
insufficient trout data and no suitable trout habitat in downstream reaches.
f V
LEGEND:
~
Town
COLD WATER FISH RISK
A/
Reach Break
<12 mg/kg
Roads
>=12 mg/kg
r~^i
Housatonlc River Basin Hydrology
~
State Boundary
w^prE
s
2 4 6
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 12.2-9
ASSESSMENT OF RISK
TO TROUT EXPOSED TO
tPCBs DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - fish cold f4-111 o:\gepitttepsfiles\plots\in\trouLrisk_12-2-9.eps 110:46 AM, 7/10/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
Mean fish concentrations were calculated for each river reach downstream of Woods Pond using
available whole body fish tissue data from samples collected since 1998. The analysis was
restricted to fish with an overall body length between 7 and 20 cm, corresponding to the size
commonly preyed on by mink. In some cases, it was necessary to use YOY data for bass and
perch, and these were extrapolated to adult concentrations using the factors discussed above.
For this analysis, it was assumed that the mink were exposed to the mean fish concentration in
the downstream reaches for 59 % of the diet because the mink diet contains on average 23% fish
and 36% invertebrates, the majority of which are crayfish. No crayfish data were available for
the downstream reaches; however, within the PSA, crayfish tPCB concentrations were similar to
fish concentrations in the size range consumed by mink. Therefore, the assumption of 59% of
the dietary exposure at the mean fish concentrations in the downstream reaches is reasonable.
This risk estimate likely underestimates the mink exposure in the downstream reaches, as it was
assumed that the concentration of tPCBs in the remaining 41% of the diet was 0.
The results of the evaluation for mink are shown in Figure 12.2-10. Potential risk to mink due to
consumption of contaminated fish occurs from the Woods Pond Dam downstream to the Great
Falls Dam, corresponding to Reaches 7 through 10.
12.2.3.6 River Otter
The mink MATC of 2.65 mg/kg tPCB in fish (whole body, wet weight) was also used for river
otter. Potential river otter habitat downstream of Woods Pond is less abundant than for mink and
is associated with larger wetland systems, with slower flowing water, or with impounded water.
Any places where the river is impounded, or near a lake or pond, there is potential river otter
habitat. Mean fish concentrations were calculated for such areas in river reaches downstream of
Woods Pond using available whole body fish tissue data from samples collected since 1998. The
analysis was restricted to fish with an overall body length between 5 and 50 cm, corresponding
to the size commonly preyed on by otter. In some cases, it was necessary to use YOY data for
bass and perch, and these were extrapolated to adult concentrations using the factors discussed
above.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-37
-------
Vest St<
BncT
£enox Dale
(41)
Glend
Btoekbridge
Sou
Berksare I
bremont Plain
(23)
Hartsvilt
MASSACHUSETTS
Drough
\41\
NEW
YORK
Saljsbun
Falls Village*
%r
CONNECTICUT
vestuomwaJI
202)
(55)
ICandlewood
Lake/
KMJ
f V
NOTES:
Risk to mink is based on a maximum acceptable threshold concentration
(MATC) of 2.65 mg/kg total PCB (tPCB) concentration (wet weight) in whole
body fish tissue 7-20 cm in length.
* Only fish collected in 1998 to the present (2000) were included.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
* Trout filet samples were scaled by 1.47.
* Brown bullhead filet samples were scaled by 1.5.
* Warmwater filet samples were scaled by 2.3.
LEGEND:
~
Town
MINK RISK
A/
Reach Break
< 2.65 mg/kg
Roads
>= 2.65 mg/kg
r~^i
Housatonlc River Basin Hydrology
~
State Boundary
w^prE
s
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 12.2-10
ASSESSMENT OF RISK
TO MINK EXPOSED TO
tPCBs DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - mink i4-151 o:\gepitttepsfiles\plots\in\mink_risk_12-2-10.eps 110:47 AM, 7/10/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
For this analysis it was assumed that otter were exposed to the mean fish concentrations in the
downstream reaches for 100% of the diet because the majority (80%) of the otter diet is fish,
with most of the remainder (8 to 20%) composed of crayfish. No crayfish data were available
for the downstream reaches; however, within the PSA, crayfish tPCB concentrations were
similar to fish concentrations in the size range consumed by otter. Therefore, the assumption of
100%) of the dietary exposure at the mean fish concentrations in the downstream reaches was
reasonable.
The results of this evaluation for otter are shown in Figure 12.2-11. Potential risk to otter due to
consumption of contaminated fish occurs from the Woods Pond Dam downstream to the Bulls
Bridge Dam, corresponding to Reaches 7 through 12.
12.2.3.7 Bald Eagle
An MATC of 30.4 mg/kg tPCBs (whole body fish tissue, wet weight) (Appendix K) was
developed for wintering bald eagles, assuming that the eagle diet was composed of 78% fish, and
that the remainder of the diet included other non-aquatic species that were assumed, for the
purpose of this analysis, to be uncontaminated.
This concentration was compared with the available fish tissue concentration data from areas
downstream of the PSA, in some cases applying scaling factors as discussed above for other
receptors. Only data from samples collected since 1998 were used, and fish less then 12 cm total
length were excluded from the analysis (unless appropriately scaled) to reflect the common size
of fish preyed on by eagles.
A more in-depth analysis was performed for Reaches 14 and 15 where bald eagles have nested.
Bald eagles on average consume a summer diet consisting of 78.2% fish, 16.3% birds, and 5%
mammals (see Section K2.1.5). Mammal and bird tPCB concentrations were not available for
downstream reaches. Total PCB concentrations for these prey items were estimated in three
ways to give high, moderate, and low concentrations. High concentrations assumed that
waterfowl and mammals from downstream would have tPCB concentrations equal to those in the
PSA. Low concentrations assumed that waterfowl and mammals from downstream would have
tPCB concentrations of zero. A moderate concentration was developed by determining fish-to-
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs 7/11/2003
-------
Vest St<
BncT
£enox Dale
(41)
Glend
Btoekbridge
Sou
Berksare I
bremont Plain
(23)
Hartsvilt
MASSACHUSETTS
Drough
\41\
NEW
YORK
Saljsbun
Falls Village*
%r
CONNECTICUT
vestuomwaJI
202)
(55)
ICandlewood
Lake/
KMJ
f V
NOTES:
Risk to otter is based on a maximum acceptable threshold concentration
(MATC) of 2.65 mg/kg total PCB (tPCB) concentration (wet weight) in
whole body fish tissue 5-50 cm in length.
* Only fish collected in 1998 to the present (2000) were included.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
* Trout filet samples were scaled by 1.47.
* Brown bullhead filet samples were scaled by 1.5.
* Warmwater filet samples were scaled by 2.3.
LEGEND:
~
Town
OTTER RISK
A/
Reach Break
< 2.65 mg/kg
Roads
>= 2.65 mg/kg
r~^i
Housatonlc River Basin Hydrology
~
State Boundary
w^prE
s
2 4 6
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 12.2-11
ASSESSMENT OF RISK
TO OTTER EXPOSED TO
tPCBs DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - otter i4-16 | o:\gepitt\epsfiles\plots\in\otter_risk_12-2-11.eps 110:43 AM, 7/10/20031
-------
1 mammal and fish-to-bird ratios based on concentrations in the PSA. Mammal tPCB
2 concentrations in the PSA are on average 75% of the total fish concentration, and waterfowl
3 tPCB concentrations averaged 15% of the total fish concentration. Therefore, moderate tPCB
4 concentrations downstream were 0.539 mg/kg for mammals and 0.108 mg/kg for birds.
5 The results of the evaluation are shown in Figure 12.2-12. Potential risks to bald eagles from
6 consuming contaminated fish in areas downstream of Woods Pond are restricted to Reach 8,
7 corresponding to Rising Pond. However, Rising Pond is smaller than the typical eagle foraging
8 area so this estimate of risk is conservative. In addition, the more in-depth analysis specific to
9 Reaches 14 and 15 also did not show risk in the foraging area of the nesting bald eagles.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-41
-------
MASSACHUSETTS
NEW
YORK
CONNECTICUT
NOTES:
Risk to eagles Is based on a maximum acceptable threshold concentration
(MATC) of 30.4 mg/kg total PCB (tPCB) concentration (wet weight) in whole
body fish tissue greater than or equal to 12 cm.
* Only fish collected in 1998 to the present (2000) were included.
* Young-of-year bass composites were scaled by a factor of 3.5.
* Young-of-year perch composites were scaled by a factor of 2.5.
LEGEND:
~
Town
EAGLE RISK
/V
Reach Break
< 30.4 mg/kg
Roads
>= 30.4 mg/kg
r~^i
Housatonic River Basin Hydrology
~
State Boundary
w^prE
s
8 Miles
Ecological Risk Assessment
GE/Housatonic River Site
Rest of River
FIGURE 12.2-12
ASSESSMENT OF RISK
TO BALD EAGLE EXPOSED
TO tPCBs DOWNSTREAM
OF WOODS POND
| O:\gepitt\aprs\era_species.apr | layout - eagle k4-51 o:\gepitt\epsfiles\plots\in\eagle_risk_12-2-12.eps 110:49 AM, 7/10/20031
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
12.3 SPECIES SENSITIVITY AND MECHANISMS OF TOXICITY
There is a large amount of variability in the toxicokinetic responses of different species to tPCBs
and 2,3,7,8-TCDD equivalence (TEQ). Numerous studies have shown that tPCBs and TEQ may
cause a variety of adverse effects (e.g., Bosveld and Van den Berg 1994; Newsted et al. 1995;
Van den Berg et al. 1998). Effects may include:
¦ Lethality.
¦ Hepatic lesions.
¦ Immunotoxicity.
¦ Tumor promotion.
¦ Adverse effects on reproduction.
¦ Induction of drug-metabolizing enzymes.
How PCB and TEQ congeners cause these effects, and the ability of different species to defend
against these contaminants is less clear. A brief description of the primary toxic mechanism of
coplanar PCBs, dioxins, and furans is provided in Section 12.2.3.1 (see also Section 2), as is
information on the relative sensitivities of biota to tPCBs and TEQ.
Although sensitivity to COCs undoubtedly explains some of the differences in effects and
resulting risk experienced by biota in the PSA (see Figures 12.2-1 through 12.2-4), other factors
also play a role. For example, higher trophic level biota that may forage exclusively in the PSA
(e.g., kingfishers, mink) have higher exposures to tPCBs and TEQ than do biota with foraging
areas of which only a portion is in the PSA (e.g., red fox).
Also, the composition and toxicity of the congener mixture that biota are exposed to changes
with trophic level. The latter issue is briefly discussed in Section 12.2.3.2 and in more detail in
Appendix C.7.
12.3.1 Mechanism of Action and Sensitivity of Species to tPCBs and TEQ
Some chlorinated PCBs, dioxins, and furans belong to a large class of chemicals called planar
chlorinated hydrocarbons (PCHs) that are regularly detected in the environment. PCHs have a
common structural relationship that includes lateral halogenation (i.e., the addition of chlorine to
the compound) and the ability to assume a planar conformation. This structure is important as it
leads to a common mechanism of action in many animal species involving binding to the aryl
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
hydrocarbon (Ah) receptor and elicitation of an Ah-receptor-mediated biochemical and toxic
response (Van den Berg et al. 1998; Newsted et al. 1995; Safe 1994). The planar conformation
is the factor that controls the ability of the the chemical to bind with the Ah receptor (Birnbaum
and Devito 1995; Newsted et al. 1995). The Ah receptor facilitates the translocation of PCHs
into the nucleus of affected cells and the binding of the PCH-Ah receptor complex to sites on the
DNA (Newsted et al. 1995).
Exposure of PCBs and other organic toxins in vertebrates elicits a response of the cytochrome
P450 system with associated mixed function oxidases (MFO). MFOs enhance the elimination of
some hydrophobic chemicals through a series of oxidative reactions (Eisler 2000). However, the
MFO system is less capable of breaking down congeners with chlorine substitution at the 2, 3, 7,
and 8 positions. As a result, these coplanar congeners show resistance to metabolic breakdown
in many higher organisms (Bosveld and Van den Berg 1994).
The development of the cytochrome P450 system varies between species of vertebrates.
Therefore, some species may be more sensitive to tPCBs and TEQ than others, even within
taxonomic families. For example, mustelids may vary widely in sensitivity to PCBs (Leonards et
al. 1997). Mink are among the most sensitive species to PCBs known (Aulerich et al. 1985;
Giesy and Kannan 1998). Conversely, ferrets are much less sensitive to PCBs than mink
(Bleavins et al. 1980). There are, however, few data available for other mustelid species. Foxes
and dogs have been shown to have an unusual P450 isoenzyme that allows them to degrade
PCB-153 more efficiently than rats and monkeys (Georgii et al. 1994). In general, fish are less
capable of metabolizing PCBs than most birds and mammals (Van den Berg et al. 1998).
Despite their reduced ability to metabolize PCBs, fish are relatively insensitive to mono-ortho
PCBs, compared to birds and mammals (Van den Berg et al. 1998).
Fewer studies have been conducted on amphibians than on mammals, fish, and birds. The ability
of amphibians to metabolize organic contaminants appears to be comparable to that of fish, but
lower than that of rats (Eisler 2000). In amphibians, effects on the neutrophil function (i.e.,
immunosuppression) may be important (Angermann and Matsumura 1999). There is evidence
that indicates that amphibians contain the Ah receptor, but it is not as well described because of
limited research (Jung 1997). PCB-126 induces cytochrome P450 activity in both leopard frogs
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-44
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
and green frogs (Huang et al. 2001). Amphibians have a cytochrome P450 mixed function
oxidase that is less active and well developed than mammals, but that does not appear to be
significantly different from other vertebrates (Eisler 2000). Benthic invertebrate toxicity is not
mediated by an Ah receptor mechanism, and TEF systems have not been developed for
amphibians or benthic invertebrates.
CCME (1999) reviewed the toxicology literature for mammals and birds and suggested that bird
species may be less sensitive to the effects of PCBs than mammals. Some bird species such as
tree swallows appear to be quite tolerant of elevated exposures to tPCBs and TEQ (Custer et al.
1998; Bishop et al. 1995, 1999; McCarty and Secord 1999). Substantial differences in sensitivity
to PCBs between bird species have also been noted. Barron et al. (1995) determined that
differences in the genetic expression of the Ah receptor were the dominant factor explaining
differences in PCB sensitivity of the bird species examined. Brunstrom (1988, 1989) examined
the sensitivity of numerous species of avian embryos to coplanar PCBs and concluded that
interspecies differences in sensitivity were due to differences in the Ah receptor ligands.
Therefore, the available literature indicates that there are differences in sensitivity of biota, and
that these differences may be partially attributed to differences in development of the
cytochrome P450 system and other factors. The differences in sensitivity of biota partially
explain why, for example, mink are experiencing much greater effects from exposure to tPCBs
and TEQ in the PSA than are kingfishers (Figures 12.2-3 and 12.2-4), despite the two species
having similar diets. Similarly, tree swallows and small-footed myotis have similar diets, yet the
more tolerant tree swallows are likely to experience lower risks from exposure to tPCBs and
TEQ than does small-footed myotis (Figures 12.2-3 and 12.2-4).
Although differences in sensitivity of biota can partly be explained on the basis on differential
development of Ah receptor and cytochrome P450 systems, toxicity from non-coplanar
congeners not associated with these systems is also important. Observations of effects in aquatic
receptors that are either unrelated to Ah receptor interactions (e.g., benthic invertebrates) or are
greater than would be predicted on the basis of congeners with TEF values only (e.g., fish) are
evidence of this. Detailed knowledge of PCB toxicity for all 209 PCB congeners does not exist;
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-45
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
differential toxicity to non-coplanar congeners may explain some of the interspecies differences
in toxicity observed.
12.3.2 Congener Composition and Toxicity to Biota
Environmental degradation (or weathering) of PCH congeners varies due to the unique
physical/chemical properties of each congener (Cogliano 1998). This can cause differences
between the congeners detected in environmental samples and the congener makeup of the
original product or Aroclor (Cogliano 1998; Van den Berg et al. 1998). In the Housatonic River
PSA, PCB composition exhibits little spatial variability within a medium (e.g., in sediment
between reaches), although there are shifts in composition across media (Appendix C.7).
Between receptor groups, PCB congener composition may exhibit considerable variation. The
change in the congener composition in prey tissue can produce differences in toxicological
responses in exposed predator species relative to the effects observed in the laboratory for
species exposed to technical mixtures (e.g., Aroclor 1254 or 1260).
In the site-specific fish studies conducted for this assessment, the congener composition in fish
was found to be more toxic than would be expected from studies exposing fish to Aroclor 1254
and 1260 commercial mixtures (Appendix F). Conversely, the results of the mink feeding study
(Appendix I) indicated that the congener composition in fish tissue from the PSA was less toxic
than would be expected from mink toxicity studies conducted with mixtures expected to be
similar. For example, the highest dose treatment in the mink feeding study caused effects on
survival of mink kits that would have been expected at about a 4-fold lower dose based on
chronic feeding studies conducted with Aroclor 1254 (see Figure 1.3-5 in Appendix I).
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-46
-------
1 12.4 BROADER IMPLICATIONS
2 The weight-of-evidence assessments briefly described in Sections 3 through 11, and in more
3 detail in Appendices D through K, indicate that COCs in the PSA of the Housatonic River,
4 particularly tPCBs, are causing risks to many of the species chosen to represent the assessment
5 endpoints (see Figures 12.2-1 through 12.2-4, Table 12.2-1). Risks from COCs, however, may
6 potentially extend beyond adverse effects to survival, growth, and reproduction of representative
7 species. The purpose of this section is to explore the implications of the risks of COCs to
8 representative species demonstrated in the preceding sections. This section begins by extending
9 the ecological risk assessment to species that occur in the Housatonic River watershed, but that
10 had not been considered explicitly in the quantitative ecological risk assessments previously
11 described in Sections 3 through 11. This section is followed by a general discussion of the
12 possible broader ecological implications of this risk assessment.
13 12.4.1 Implications for Other Species in the Primary Study Area
14 The purpose of this section is to qualitatively compare exposure of the representative species to
15 other species that were identified in the Ecological Characterization (Appendix A) to occur in the
16 PSA for tPCBs and TEQ. The major factors that influence exposure to tPCBs and TEQ and that
17 were considered in the analysis include:
18 ¦ Dietary composition.
19 ¦ Foraging behavior and home range.
20 ¦ Size, metabolism, and life history characteristics.
21 ¦ Sensitivity to COCs.
22
23 The following sections briefly compare these factors between the representative species and
24 other species in their foraging groups. The comparison highlights similarities and differences,
25 and their potential to influence exposure and thus risks to tPCBs and TEQ. This comparison
26 does not consider differences in sensitivity between representative species and other species in
27 their foraging groups because toxicity data to make this comparison are lacking. Table 12.4-1
28 summarizes this qualitative risk assessment. More discussion is presented in Appendices D
29 through K.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-47
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Top Predator
Fish
Largemouth
bass/
Intermediate to
High
10-16 inches
(21-41 cm)
Year-
round
n/a
Aquatic
insects, fish,
crayfish
n/a
Hartel et al. 2002
Smallmouth
bass
8-13 inches
(20-33 cm)
Year-
round
n/a
Aquatic
invertebrates,
fish
Prefers cooler,
clearer, rockier areas
than largemouth
bass (LMB), but
similar
X
Hartel et al. 2002
Black
crappie
8-12 inches
(20-30 cm)
Year-
round
n/a
Aquatic
invertebrates,
fish
Prefers cooler,
clearer, rockier areas
than LMB, but diet
is similar
X
Hartel et al. 2002
Rock bass
6-8 inches
(15-20 cm)
Year-
round
n/a
Aquatic
invertebrates,
fish, crayfish
Diet and some
habitat preferences
are similar,
particularly to young
(3-4 y.o.) LMB
X
Hartel et al. 2002
Chain
pickerel
15-24 inches
(38-61 cm)
Year-
round
n/a
Invertebrates,
fish
Similar diet
compared to LMB
X
Hartel et al. 2002
Omnivorous
Bottom
Feeders
Brown
bullhead/
Intermediate to
High
8-14 inches
(20-36 cm)
Year-
round
n/a
Animal and
plant material
(up to 40%
plants, up to
60%
filamentous
algae)
n/a
Hartel et al. 2002;
Gunn et al. 1977
Yellow
bullhead
8-12 inches
(20-30 cm)
Year-
round
n/a
Insects,
crustaceans,
mollusks, small
fish, plant
material (up to
40% plants)
Similar habitat,
although typically
prefers clearer water
than brown
bullhead. Diet is
similar
X
Hartel et al. 2002
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Representative
Level of Risk Compared to Representative Species
Category
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
References
Common
carp
24 inches
(61 cm)
Year-
round
n/a
Animal and
plant material
Habitat and diet
similar to brown
bullhead
X
Hartel et al. 2002
Goldfish
5-13 inches
(13-33 cm)
Year-
round
n/a
Animal and
plant material
Habitat and diet
similar to brown
bullhead
X
Hartel et al. 2002
White sucker/
Intermediate to
High
12-24 inches
(30-61 cm)
Year-
round
n/a
Aquatic
invertebrates,
larval insects,
detritus
n/a
Hartel et al. 2002
Longnose
sucker
12-15 inches
(30-38 cm)
Year-
round
n/a
Aquatic
invertebrates,
algae
Prefers cooler,
cleaner stream
reaches than white
sucker. Diet is
similar to white
sucker
X
Hartel et al. 2002
Creek
chubsucker
9 inches
(23 cm)
Year-
round
n/a
Omnivorous
Habitat and diet
similar to white
sucker
X
Hartel et al. 2002
Forage Fish
Pumpkinseed/
Intermediate to
High
4-5 inches
(10-13 cm)
Year-
round
n/a
Aquatic
invertebrates
n/a
Hartel et al. 2002
Bluegill
5-7 inches
(13-18 cm)
Year-
round
n/a
Aquatic
invertebrates,
some small fish
Habitat and diet
similar to
pumpkinseed
X
Hartel et al. 2002
Redbreast
sunfish
4-8 inches
(10-20 cm)
Year-
round
n/a
Larvae and
adult aquatic
insects,
terrestrial
insects, some
small fish
Habitat and diet
similar to
pumpkinseed
X
Hartel et al. 2002
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Amphibians
Wood frog/
High
18 g
Year-
round,
larvae: 3
months
64.5 m2
Insects, beetles,
flies, slugs,
snails, spiders,
bugs, moth
larvae, and
earthworms
Small home range,
but establishes
territories >1,000 m
from breeding pools
n/a
DeGraaf and
Yamasaki 2001;
Hunter et al. 1999
Northern
spring
peeper
1.0-2.7 g
Year-
round,
larvae: 3
months
23 m2
Spiders (up to
50%), ants,
beetles, mites,
ticks,
springtails,
caterpillars,
slugs, and
snails
Territories usually
established near
suitable breeding
sites. Late
summer/fall
migrations to
hibernation sites
may be further away
X
DeGraaf and
Yamasaki 2001;
Hunter et al. 1999
Spotted
salamander
14 g
Year-
round,
larvae: 3
months
10-14 m2
Forest floor
invertebrates:
earthworms,
slugs, snails,
spiders,
millipedes,
centipedes,
larval and adult
insects
Home ranges
usually established
within 200 m of
breeding site
X
Petranka 1998;
DeGraaf and
Yamasaki 2001;
Hunter et al. 1999
Amphibians
(cont.)
Wood frog/
High (cont.)
Jefferson
salamander
ii g
Year-
round,
larvae: 3
months
Small
invertebrates,
worms, spiders,
insects,
crustaceans
Home ranges
typically within 250
m from breeding
pond, but have been
recorded up to 624
m away
X
Petranka 1998;
DeGraaf and
Yamasaki 2001;
Hunter et al. 1999
Northern
leopard frog/
High
38 g
Year-
round,
larvae: 3
months
5 to 53 m
nightly
movement
Beetles (up to
50%),
lepidopteran
larvae, bugs,
grasshoppers,
ants, spiders,
crayfish, snails
Semi-terrestrial
spending summer
month in damp
fields and woods,
hibernates and
breeds in permanent
bodies of water
n/a
DeGraaf and
Yamasaki 2001;
Hunter et al. 1999
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Pickerel frog
Not found
Year-
round,
larvae: 3
months
Not found
Insects, beetles,
caterpillars,
true bugs, ants,
spiders, snails,
crayfish, and
amphipods
Habitat preferences
and feeding habits
similar to leopard
frog
X
DeGraaf and
Yamasaki 2001;
Hunter et al. 1999
Green frog
30-70 g
Year-
round,
larvae: 1-2
years
61 m2
home
range, 2-3
m
breeding
territory
Adults: plants,
spiders, beetles,
true bugs,
wasps and
bees,
mosquitoes,
flies, midges
and gnats,
mayflies,
moths, and
butterflies
Green frogs are
more aquatic than
leopard frogs. They
do enter the
floodplain to access
seasonally available
food resources
X
DeGraaf and
Yamasaki 2001;
Hunter etal. 1999;
Stewart and
Sandison 1973;
Jenssen and
Klimstra 1966.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Representative
Level of Risk Compared to Representative Species
Category
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
References
Amphibians
(cont.)
Northern
leopard frog/
High (cont.)
Eastern newt
2-3 g
Year-
round,
larvae: 2
months,
efts 3-7
years
Adults:
captured
within 7
m of
original
capture
sites, efts:
270 m2 up
to 800 m
from
breeding
sites
Adults:
mayflies,
damselflies,
dragonflies,
mosquitoes,
midges, gnats,
water fleas,
amphipods,
bivalves, and
clams
Newt larvae
efts: snails,
slugs, mites,
ticks, beetles,
beetle larvae,
flies,
mosquitoes,
midges, gnats,
maggots,
wasps, and
bees
Larval period is
spent in pools;
metamorph into
terrestrial eft stage
that lasts 2-7 years.
A second
metamorphosis
occurs when adults
return to pools and
transform into
aquatic adults
X
Petranka 1998;
DeGraaf and
Yamasaki 2001;
Hunter etal. 1999;
Burton 1977;
MacNamara 1977;
Bellis 1968; Healy
1975
Insectivorous
Birds
Tree swallow/
Intermediate to
High
20.1 g (range
16.5-25.5 g)
6 months
Defend
10-15 m
around
nest; feed
within
300-400
m of nest
Primarily
emergent
insects, such as
flies,
mosquitoes,
midges, gnats,
mayflies, and
beetles
Feed on emergent
insects over bodies
of water
n/a
DeGraaf and
Yamasaki 2001;
Robertson et al.
1992; Ehrlich et
al. 1988; Martin et
al. 1951
Bank
swallow
13.5 g
5 months
Territory
limited to
immediate
vicinity of
the nest
entrance
Almost entirely
insects
Nest in exposed and
eroding riverbanks
and in gravel pits
X
DeGraaf and
Yamasaki 2001;
Ehrlich etal. 1992
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Insectivorous
Birds (cont.)
Tree swallow/
Intermediate to
High (cont.)
Northern
rough-
winged
swallow
16 g
5 months
Nest
within 1
km of
water
Entirely insects
Nest in exposed and
eroding riverbanks
and in gravel pits
X
DeGraaf and
Yamasaki 2001;
Ehrlichetal. 1992
Barn
swallow
19 g
5 months
Seldom
feed more
than 0.8
km from
nest site
Insects,
occasionally
berries and
seeds
Nest under bridges
in PSA and feed
over river and fields
X
DeGraaf and
Yamasaki 2001;
Ehrlichetal. 1992
Cliff
swallow
21 g
5 months
Foraging
range
typically
within 1
km
Almost entirely
insects, but
occasionally
gorge on
berries
Nest under bridges
in PSA and feed
over river and fields
X
DeGraaf and
Yamasaki 2001;
Ehrlichetal. 1992
Chimney
swift
23 g
5 months
Foraging
range up
to several
kilometers
Flying insects
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Sibley 2001
Common
nighthawk
62 g
5 months
Territory
2-23 ha
Flying insects,
especially
flying ants,
mosquitoes,
moths,
grasshoppers
Nests on rooftops in
town, feeds over
fields and water
X
DeGraaf and
Yamasaki 2001;
Ehrlichetal. 1992
Eastern
kingbird
40 g
6 months
Territory
5.7-14.2
ha
Flying insects,
some fruit
Commonly nests in
trees overhanging
the river in the PSA,
capture insects over
the river. Also
occurs in
agricultural areas,
over fields
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Insectivorous
Birds (cont.)
Tree swallow/
Intermediate to
High (cont.)
Eastern
phoebe
20 g
7 months
0.3 ha in
an Illinois
flood-
plain; 1.3-
2.9 ha in
settled
area
92-97% flying
insects from
spring through
fall, mostly
berries and
seeds in winter
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
American
robin/High
77 g
Year-
round
Territory:
0.4 ha;
foraging
range for
nestlings
and fledg-
lings; 0.15
and 0.81
ha
50-90% animal
matter in spring
and summer,
switches to
plants (berries)
in fall and
winter. Prey
includes
earthworms,
butterflies,
moths, beetles,
and ants
n/a
DeGraaf and
Yamasaki 2001;
Sallabanks and
James 1999;
Ehrlich et al.
1992;
Weatherhead and
McRae 1990;
Martin et al. 1951
Eastern
bluebird
31 g
8 months
Territory
2.2-3.5 ha
60% or more
animal matter
year-round, up
to 80-95% in
spring and
summer. Prey
includes
beetles,
grasshoppers,
crickets, and
caterpillars
Diet similar to robin
except earthworms
rarely consumed
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
Eastern
towhee
40 g
6 months
Territory
0.26-2.4
ha
50:50 plant and
animal in
summer and
fall. Mostly
terrestrial
insects, seeds,
and berries
Tends to consume
considerably less
animal matter than
robin
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Representative
Level of Risk Compared to Representative Species
Species and
Risk Category
Other
Foraging/
Home
Life History/
Lower
to
Similar
to
Category
in PSA*
Species
Size
Residency
Range
Diet
Miscellaneous
Lower
Similar
Similar
Higher
Higher
References
American
Gray catbird
37 g
7 months
Territory
40-80% animal
Diet similar to robin
X
DeGraaf and
robin/High
0.06-0.32
matter in spring
except earthworms
Yamasaki 2001;
(cont.)
ha
and summer,
fall diet nearly
80% plants
(berries). Prey
includes largely
terrestrial
insects (ants,
beetles),
caterpillars,
and
grasshoppers
rarely consumed
Ehrlich et al.
1992; Martin et al.
1951
Hermit
31 g
7 months
Territory
93 and 85%
Diet similar to robin
X
DeGraaf and
thrush
0.06-3.34
ha
animal matter
in spring and
summer,
respectively.
Dominant prey
includes
beetles, ants,
caterpillars,
flies, and
insects
with the exception
of earthworms
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
Northern
49 g
Year-
Territory
70-85% animal
Diet similar to robin
X
DeGraaf and
mockingbird
round
0.25-0.5
ha in
summer
matter in spring
and summer.
Dominant prey
includes
beetles, ants,
bees, wasps,
and
grasshoppers
except earthworms
rarely consumed
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
American
robin/High
(cont.)
Veery
31 g
5 months
Territory
0.1-3+ha
60-95% animal
matter in spring
and summer.
Dominant prey
includes
beetles, ants,
caterpillars,
spiders, and
grasshoppers
Diet similar to robin
except earthworms
rarely consumed
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
Wood thrush
47 g
6 months
Territory
0.08-2.8
ha
60-95% animal
matter in spring
and summer.
Dominant prey
includes
beetles, ants,
caterpillars,
spiders, and
grasshoppers
Diet similar to robin
except earthworms
rarely consumed
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
Omnivorous
and
Carnivorous
Mammals
Northern short-
tailed
shrew/High
20.5 g
Year-
round
0.024-0.2
ha
Common prey
includes insects
and
earthworms.
Also forage on
snails, slugs,
crustaceans,
small
mammals,
fungi, and,
rarely,
vegetation
Occurs in damp
woodlands and
fields
n/a
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995; EPA
1993; Whitaker
and Ferraro 1963;
Hamilton 1941;
Linzey and Linzey
1973; Eadie 1944;
Eadie 1948
Smoky
shrew
6.1-11 g
Year-
round
Not found
Insectivorous;
also
salamanders,
young mice,
vegetable
matter
Habitat preferences
and diet similar to
short-tailed shrew
X
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Representative
Level of Risk Compared to Representative Species
Category
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
References
Omnivorous
and
Carnivorous
Mammals
(cont.)
Short-tailed
shrew/High
(cont.)
Masked
shrew
4.0-6.5 g
Year-
round
0.16-0.28
ha
Predominantly
insectivorous;
also mollusks,
annelids, dead
bodies of larger
animals,
salamanders,
young mice,
Endogone,
vegetable
matter
Similar diet but
more commonly in
dryer uplands,
meadows, old fields,
and fencerows
X
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995
Red fox/
Intermediate
3.4-6.4 kg
Year-
round
60-600 ha
Up to 30%
plants in
summer and
fall, the
remainder
being small
mammals,
birds, and
insects.
Mammals
average 76% of
diet for all
seasons
Prefers open
agricultural land and
forest edges
n/a
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995; EPA
1993; Martin et al.
1951; Powell and
Case 1982;
Knable 1974;
Korschgen 1959;
Hockman and
Chapman 1983;
Dibello et al. 1990
Coyote
9.1-22.7 kg
Year-
round
1000-
4000 ha
78% mammals,
21% fruit, 10%
insects, and 3%
birds by
frequency in
1,500 scats
from
Adirondacks
Broad habitat
requirements, open
fields, agricultural
land, forested areas
X
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995;
Martin et al. 1951
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Omnivorous
and
Carnivorous
Mammals
(cont.)
Red fox/
Intermediate
(cont.)
Gray fox
3.2-5.9 kg
Year-
round
85-3200
ha
85-95% animal
matter
throughout the
year (e.g.,
rabbit, squirrel)
Most common in
forested areas
X
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995;
Martin et al. 1951
Fisher
3.6-5.5 kg
Year-
round
1500-
3500 ha
Nearly 100%
animal matter,
including small
mammals,
squirrels,
rabbits,
porcupine,
birds, reptiles,
and amphibians
Prefer forested areas
with closed canopies
X
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995;
Martin et al. 1951
Long-tailed
weasel
85-270 g
Year-
round
31.9-160
ha
78% small
mammals
(mice, voles,
shrews), 17%
rabbits; also
birds (up to
10%), squirrels,
snakes,
invertebrates
Terrestrial
X
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995
Short-tailed
weasel
50-150 g
Year-
round
Males:
17.0-25.0
ha
Females:
10.1-14.9
ha
75% small
mammals
(mice, voles,
shrews); also
squirrels,
rabbits, birds,
amphibians,
snakes,
invertebrates
Terrestrial
X
DeGraaf and
Yamasaki 2001;
Whitaker and
Hamilton 1998;
Kurta 1995
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Threatened
and
Endangered
Species**
American
bittern/High
370-500 g
5 months
Varies
with geo-
graphic
area,
preferred
habitat
avail-
ability
and prey
species
>3 ha
reported
for
Michigan
and 24.7
ha core
use
reported
for
Minnesota
Amphibians,
small snakes,
crayfish,
insects, small
fish
Consume a wide
variety of prey items
allowing them to
hunt in varying
habitats
n/a
Gibbs etal. 1992;
Brown and
Dinsmore 1986;
Azure 1998 in
Deschant et al.
2001; Gibbs et al.
1992 in DeGraaf
and Yamasaki
2001
Great blue
heron
2,400 g
Year-
round
Will feed
kilometers
from nest
Approximately
70% fish, also
insects,
crayfish, small
mammals, and
amphibians
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
Green heron
210 g
5 months
Defends
territory a
few feet
from nest
Crayfish
approximately
50% of diet,
25% aquatic
insects, and
20% small fish
Separate feeding
territories may be
vigorously defended
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Threatened
and
Endangered
Species
(cont.)
American
bittern/High
(cont.)
Pied-billed
grebe
450 g
7 months
Defends
area 50 m
around
nest
Primarily
animal matter,
including
crayfish, small
fish, mollusks,
aquatic insects
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
Sora
75 g
5 months
Distances
between
nest
ranges
from 1.2
to 25 m
60% animal
matter spring
and fall. Prey
includes
beetles, snails,
spiders, and
crustaceans
X
DeGraaf and
Yamasaki 2001;
Odum 1977;
Tanner and
Hendrickson 1956
in DeGraaf and
Yamasaki 2001;
Berger 1951 in
Degraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951
Virginia rail
85 g
5 months
Territorial
during
pair
formation
and nest
establish-
ment
90-95% animal
matter spring
and fall. Prey
includes
beetles, snails,
spiders, true
bugs, and
diptera larvae.
Also
crustaceans,
dragonfly and
damselfly
larvae, ants,
grasshoppers,
crickets, and
small fish
X
DeGraaf and
Yamasaki 2001;
Ehrlich et al.
1992; Martin et al.
1951.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Threatened
and
Endangered
Species
(cont.)
Small-footed
myotis/High
6 g, range 5-
7g
Year-
round, but
hibernate
Travel <
40 km
between
summer
and
winter
grounds
Primarily
midges,
caddisflies,
moths,
butterflies, and
beetles
Observed longevity
of 12 years in the
wild
n/a
DeGraaf and
Yamasaki 2001;
Griffith and Gates
1985; Anthony
and Kunz 1977;
Belwood and
Fenton 1976;
Kurta 1995;
van Zyll de Jong
1985 in DeGraaf
and Yamasaki
2001
Big brown
bat
15-24 g
Year-
round, but
hibernate
Forages 1
to 2 km
from day
roosts
Specialize in
capturing
beetles, but
also take true
bugs, wasps,
bees, flies,
mosquitoes,
midges, gnats,
moths, and
butterflies
Travels short
distances, usually no
more than 48 to 80
km from maternity
colony to
hibernaculum site
X
DeGraaf and
Yamasaki 2001;
Whitaker 1995;
Griffith and Gates
1985; Barbour and
Davis 1969 in
DeGraaf and
Yamasaki 2001;
Mills et al. 1975 in
DeGraaf and
Yamasaki 2001;
Kurta 1995; Kurta
and Baker 1990 in
Degraaf and
Yamasaki 2001;
Kurta 1995
Indiana
myotis
6-11 g
Year-
round, but
hibernate.
Female
range
over 52-
95 ha
Includes
terrestrial
insects as well
as emergent
aquatic insects
Forages in foliage of
tree crowns along
river and lake shores
X
DeGraaf and
Yamasaki 2001;
Kurta 1995
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12.DOC
-------
Table 12.4-1
Comparison of Risks of tPCBs and TEQ to Representative and Other Species in the Housatonic River PSA
(Continued)
Category
Representative
Species and
Risk Category
in PSA*
Other
Species
Size
Residency
Foraging/
Home
Range
Diet
Life History/
Miscellaneous
Level of Risk Compared to Representative Species
References
Lower
Lower
to
Similar
Similar
Similar
to
Higher
Higher
Threatened
and
Endangered
Species
(cont.)
Small-footed
myotis/High
(cont.)
Little brown
bat
6-12 g
Year-
round, but
hibernate
Unknown
Midges,
mayflies,
caddisflies, and
mosquitoes.
Also beetles,
moths,
stoneflies, true
bugs, and
termites
Consume 1.8-3.7 g
of food per night
X
DeGraaf and
Yamasaki 2001;
Griffith and Gates
1985; Anthony
and Kunz 1977;
Belwood and
Fenton 1976;
Kurta 1995
For representative species with multiple assessments in the PSA, the highest risk category is listed here.
** Several of the species included in this section (i.e., sora, great blue heron, green heron, Virginia rail, northern myotis, little brown bat) are not threatened and endangered species
either federally or in Massachusetts and Connecticut (Appendix A). They are included in the discussion of T&E species because they are taxonomically and ecologically similar
to either American bittern or to small-footed myotis.
n/a = not applicable.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-62
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
12.4.1.1 Benthic Invertebrates
The benthic invertebrate ERA included the entire benthic community; benthic community
composition analysis was a measurement endpoint considered in the weight-of-evidence
assessment. Individual species were also used in toxicity tests as surrogates for the Housatonic
River freshwater benthic community (i.e., Chironomus tentans, Hyalella azteca, Lumbriculus
variegatus, Daphnia magna, Ceriodaphnia dubia). Both the status of sensitive taxa and
community composition are considered indicators of the overall health and productivity of the
benthic community. Thus, there is no need to extrapolate the findings of the benthic invertebrate
assessment described previously to other benthic invertebrate species in the PSA.
12.4.1.2 Amphibians
Certain amphibian species that were not studied may be more susceptible to the effects of tPCBs
because of their life history characteristics. For example, blue-spotted (,Ambystoma laterale) and
spotted salamanders (,Ambystoma maculatum) have a lifestage as aquatic carnivorous, bottom-
dwelling larvae. Thus, they could potentially bioaccumulate PCBs more quickly than
herbivorous amphibians. The larvae of these two species forage on the bottom of vernal pools,
and have greater opportunities to be in contact with contaminated sediment than do pelagic frog
tadpoles. The salamander larvae feed on an assortment of planktonic animals, and then shift to
larger aquatic worms, insect larvae, small crustaceans, and tadpoles as they grow larger (Hunter
etal. 1999).
Blue-spotted and spotted salamanders also have longer larval periods, lasting between 70 to 100
days (Whitford and Vinegar 1966), than do wood frogs, which have a larval period averaging 67
days (Hunter et al. 1999). Salamanders appeared in lower numbers in vernal pools with high
sediment tPCB concentrations (Woodlot Alternatives, Inc. 2003). Several salamander species
occur in contaminated habitat in the PSA, including the spotted salamander, the Jefferson
salamander (Ambystoma jeffersonianum, formerly considered a variety of blue-spotted
salamander), and the four-toed salamander (Hemidactylium scutatum), the latter two of which are
Species of Special Concern (www.state.ma.us/dfwele/dfw/nhesp/nhrare.htm).
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-63
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
72.4.7.3 F/s/i
Five fish species—largemouth bass (Micropterus salmoides), yellow perch (Perca flavescens),
brown bullhead (Ameiurus nebulosis), white sucker (Catostomus commersoni), and pumpkinseed
(Lepomis gibbosus)—were selected as the representative species for the ERA. The fish species
were selected to include representatives of the principal trophic levels and exposure routes for
fish in the PSA.
There is evidence in the literature that salmonid species may have a higher sensitivity to the
effects of PCBs and other dioxin-like COPCs (Walker et al. 1994; Zabel et al. 1995). The use of
rainbow trout in the site-specific toxicity testing program (Phase II), combined with effects data
from the literature (Appendix F), provides a high degree of confidence that the ERA included an
evaluation of fish species with equal or greater sensitivities than the representative species listed
above. However, the procedure used to establish MATCs for fish in the PSA placed a low
weight on studies conducted with fish species known to be highly sensitive (e.g., lake trout) to
avoid an overly conservative assessment. Risks to coldwater fisheries (e.g., trout) downstream
of the PSA were explicitly evaluated using benchmarks developed for salmonids; the uncertainty
in these downstream risk estimates is high due to the number of extrapolations required. The
risk of COCs to the occasional salmonid occurring within the PSA is considered to be moderate.
The PSA, however, is considered to be a warmwater fishery, and thus salmonid abundance is
expected to be low in this portion of the river, even in the absence of chemical stressors.
12.4.1.4 Insectivorous Birds
The weight-of-evidence assessment indicated that exposure of insectivorous birds, such as tree
swallows and American robins, to tPCBs and TEQ is high but unlikely to lead to adverse
reproductive effects. Confidence in this conclusion, however, is not high because the two
available lines of evidence for both species did not produce concordant results.
The tree swallow was chosen as the one of the representative species for insectivorous birds.
This species is common in the PSA and other areas in the watershed of the Housatonic River
(Appendix A). Other insectivorous bird species that are comparable to tree swallows and are
common to the PSA include the bank swallow (Riparia riparia), northern rough-winged swallow
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-64
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
(,Stelgidopteryx serripennis), barn swallow (.Hirundo rustica), cliff swallow (.Hirundo
pyrrhonota), chimney swift (Chaetura pelagica), common nighthawk (iChordeiles minor),
eastern kingbird (Tyrannus tyrannus), and eastern phoebe (Sayornisphoebe).
Compared to the tree swallow, the cliff swallow is expected to have a similar to lower level of
risk from exposure to tPCBs and TEQ. The cliff swallow has a similar diet and is of similar size,
but has a much larger foraging range that may dilute exposure compared to tree swallows. The
eastern kingbird is also expected to have a similar to lower level of risk because it is twice the
size of the tree swallow and, therefore, has a lower metabolic and food intake rate. The lower
food intake rate will likely lead to reduced exposure. Eastern kingbirds and tree swallows have a
similar diet and foraging range. The barn swallow and eastern phoebe have a similar level of
risk compared to tree swallows because they have similar body sizes and diet. The bank
swallow, chimney swift, and northern rough-winged swallow have a similar to higher level of
risk from exposure to tPCBs and TEQ because they have a higher proportion of insects in the
diet and/or are smaller than tree swallows.
Insectivorous birds that are more comparable to American robins and are common to the area
include the eastern bluebird (Sialia sialis), eastern towhee (Pipilo erythrophthalmus), gray
catbird (Dumetella carolinensis), hermit thrush (iCatharus guttatus), northern mockingbird
(Mimus polyglottos), veery (Catharus fuscescens), and wood thrush (Hylocichla mustelina) (see
Appendix A).
Compared to American robins, eastern bluebirds and eastern towhees are expected to experience
lower to similar levels of risk from exposure to tPCBs and TEQ. Eastern bluebirds consume
similar prey compared to American robins, but have a larger foraging range that would dilute
exposure to tPCBs and TEQ in the PSA. Eastern towhees consume less animal matter than
American robins. Because animal matter contains higher concentrations of COCs than
vegetation, eastern towhee exposure to tPCBs and TEQ will likely be lower.
The level of risk for the hermit thrush, northern mockingbird, veery, and wood thrush is expected
to be similar to American robins. With the exception of earthworms in the robin diet, the dietary
preferences of these birds are similar to the American robin. The absence of earthworms, a
major dietary source of contaminants, will decrease their exposure to tPCBs and TEQ. However,
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs 7/11/2003
-------
1 their smaller body sizes result in higher food intake rates and thus greater exposure to tPCBs and
2 TEQ through diet compared to American robins.
3 Gray catbirds are expected to experience similar to higher levels of risk compared to American
4 robins. With the exception of earthworms in the robin diet, gray catbirds consume similar prey
5 and have a foraging range comparable to American robins. Their smaller body size results in a
6 higher food intake rate and greater exposure to tPCBs and TEQ through diet compared to
7 American robins.
8 12.4.1.5 Piscivorous Birds
9 The weight-of-evidence assessment indicates that risks of tPCBs and TEQ to belted kingfisher
10 are low; however, risks of these COCs to osprey are high and could lead to adverse reproductive
11 effects.
12 The belted kingfisher and osprey were chosen to represent piscivorous birds inhabiting the
13 Housatonic River area. Belted kingfisher and osprey are common piscivorous birds in the PSA.
14 Great blue herons are also found in the PSA, and are discussed in Appendix K with other
15 piscivorous birds (e.g., American bittern).
16 12.4.1.6 Piscivorous Mammals
17 Mink and river otter, the representative species for piscivorous mammals, are the only
18 piscivorous mammals commonly found in the watershed of the Housatonic River (EPA 2001)
19 (Appendix A).
20 12.4.1.7 Omnivorous and Carnivorous Mammals
21 The weight-of-evidence assessment indicates that omnivorous and carnivorous mammals, such
22 as red fox and short-tailed shrew, are at risk in the PSA as a result of exposure to tPCBs and
23 TEQ. Risks to short-tailed shrews exposed to tPCBs at Sites 13 and 14 are high.
24 The northern short-tailed shrew and red fox were chosen to represent omnivorous and
25 carnivorous mammals inhabiting the Housatonic River area. Other omnivorous and carnivorous
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs 7/11/2003
-------
1 species common to the area include the smoky shrew (Sorex fumeus), masked shrew (Sorex
2 cinereus), coyote (Canis latrans), gray fox (Urocyon cinereoargenteus), fisher (Martes
3 pennanti), long-tailed weasel {Mustela frenata), and short-tailed weasel {Mustela erminea) (see
4 Appendix A).
5 Masked shrews are expected to experience a level of risk similar to northern short-tailed shrews.
6 Both animals have similar foraging behaviors and ranges. Masked shrews are smaller than
7 northern short-tailed shrews; therefore, they have a higher metabolism that increases exposure to
8 contaminants. Masked shrews prefer to inhabit dry upland areas that are less contaminated than
9 the damp woodlands and fields of the PSA where northern short-tailed shrews are found.
10 Compared to northern short-tailed shrews, smoky shrews are expected to experience higher
11 levels of risk from exposure to tPCBs and TEQ. These shrews have similar foraging preferences
12 and life history characteristics. Smoky shrews are much smaller than northern short-tailed
13 shrews and thus have a higher metabolic rate that increases exposure to tPCBs and TEQ.
14 Coyotes have a larger body size and foraging range that decreases their exposure to tPCBs and
15 TEQ. Considering these characteristics, coyotes are expected to experience lower risks from
16 exposure to tPCBs and TEQ.
17 Gray and red foxes are expected to experience similar risks from exposure to tPCBs and TEQ.
18 Gray fox have a larger foraging range than red fox and that may decrease their exposure to
19 tPCBs and TEQ. Gray fox, however, have a greater reliance on animal matter and thus greater
20 exposure to tPCBs and TEQ.
21 Fisher, long-tailed weasels, and short-tailed weasels are expected to experience similar to higher
22 levels of risk from exposure to tPCBs and TEQ compared to the red fox. Fisher and red fox have
23 similar body weights. Animal matter constitutes nearly 100% of the fisher diet compared to an
24 average of 76% of the red fox diet. This greater consumption of animal matter increases the
25 fisher's exposure to tPCBs and TEQ. The diets of long- and short-tailed weasels are similar to
26 red fox, but weasels have smaller foraging ranges, which increases their exposure to tPCBs and
27 TEQ in the PSA. Their smaller body weight results in a higher metabolism that further increases
28 exposure to tPCBs and TEQ.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-67
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
12.4.1.8 Threatened and Endangered Species
The weight-of-evidence assessment indicates that threatened and endangered (T&E) species such
as bald eagles, American bitterns, and small-footed myotis are at risk in the PSA as a result of
exposure to tPCBs and, to a lesser extent, TEQ. The risk for bald eagles exposed to tPCBs is
high. The risks for American bittern and small-footed myotis exposed to tPCBs and TEQ are
intermediate.
The bald eagle, American bittern, and small-footed myotis were chosen to represent T&E species
that are likely to be highly exposed to COCs in the Housatonic River PSA. Other T&E species
that occur in the area include one mussel (triangle floater); three dragonflies (riffle snaketail,
zebra clubtail, and arrow clubtail); a turtle (wood turtle); three salamanders (Jefferson
salamander, four-toed salamander, and northern spring salamander); three hawks (northern
harrier, sharp-shinned hawk, and Cooper's hawk); two warblers (northern parula and blackpoll
warbler); a wading bird (common moorhen); and a shrew (northern water shrew). Some of these
species were qualitatively assessed in other appendices and compared to other, more appropriate,
assessment endpoints (e.g., amphibians for salamanders).
The level of risk for soras1 is expected to be lower than for American bitterns. The sora and
American bittern have similar life history characteristics and habitat preferences. However, the
sora consumes more vegetable matter and less animal matter than the American bittern. This
decreases the sora's exposure to tPCBs and TEQ.
Great blue herons and king rails are expected to experience a similar level of risk as American
bitterns. Great blue herons consume more fish than American bitterns. Fish contain higher
concentrations of tPCBs than other prey. However, great blue herons have a larger body size
than American bitterns resulting in a slower metabolism and lower accumulation of
contaminants. In addition, the reproductive strategy for great blue herons suggests that few
individuals from an entire rookery would be exposed in the PSA, lessening the risk to the local
1 Several of the species included in this section (i.e., sora, great blue heron, green heron, Virginia rail, northern
myotis, little brown bat) are not threatened and endangered species either federally or in Massachusetts and
Connecticut (Appendix A). They are included in the discussion of T&E species because they are taxonomically
and ecologically similar to either American bittern or to small-footed myotis.
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC ^ gg 7/11/2003
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
population. King rails consume prey that would have lower concentrations of COCs than
American bitterns. Therefore, although they have a smaller body size and higher metabolism,
their exposure to tPCBs and TEQ is expected to be lower.
The least bittern, green heron, Virginia rail, and pied-billed grebe are expected to experience
higher levels of risk compared to the American bittern. The foraging and life history
characteristics of these birds are similar to the American bittern. However, these birds are much
smaller than the American bittern. Their smaller body sizes result in a higher metabolism and
greater exposure to tPCBs and TEQ.
The Indiana bat, northern myotis, and little brown bat are expected to have similar levels of risk
as the small-footed myotis. These species belong to the same genus (Myotis) and have similar
foraging behaviors and life histories.
All of the above information is summarized in greater detail in Table 12.4-1.
12.4.2 Ecological Implications and Other Concerns
As with most ecological risk assessments of contaminated sites, the ecological risk assessment
for the Housatonic River is an assessment of the direct effects of COCs on a species-by-species
basis. The following discussion places the ecological risk assessment in a broader ecological
context by examining populations, ecological interactions and functions, and other issues of
concern to decision makers and the public. The section begins with a brief discussion of the
regulatory objectives of EPA and other agencies as they pertain to ecological protection goals for
contaminated sites.
12.4.2.1 Protection Goals
Recently, there has been considerable interest in regulatory agencies and elsewhere for assessing
risks at higher levels of organization, such as the population, community, or ecosystem (e.g.,
Environment Canada 1997; EPA 1997; Landis et al. 1998; EPA 1999b). Assessment of risks at
higher levels of organization is useful because it furthers the understanding of the seriousness of
risks posed by COCs, an important consideration in developing appropriate risk management
responses. Assessment of risk at higher levels of organization, however, is not an easy task
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs ^q 7/11/2003
-------
1 (Moore and Bartell 2000). The desire to understand risk at higher levels of organization should
2 not be misinterpreted to mean that risks must be demonstrated at higher levels of organization
3 (e.g., population or higher) to be of concern to EPA and other agencies at the Housatonic River,
4 or other assessments of contaminated sites. As stated in EPA 1999b:
5 Levels that are expected to protect local populations and communities can be
6 estimated by extrapolating from effects on individuals and groups of individuals
7 using a line-of-evidence approach. The performance of multi-year field studies at
8 Superfund sites to try to quantify or predict long-term changes in local
9 populations is not necessary for appropriate risk management decisions to be
10 made. Data from discrete field and laboratory studies, if properly planned and
11 appropriately interpreted, can be used to estimate local population or community
12 level effects.
13 In addition, the Massachusetts Contingency Plan at 310 CMR 40.0995 (4) states:
14 (b) The Stage II Environmental Risk Characterization shall identify
15 environmental resources associated with the disposal site, such as wetlands,
16 aquatic and terrestrial habitat, fisheries, or rare and endangered species, and shall
17 evaluate whether the release of oil and/or hazardous material has adversely
18 impacted, or may adversely impact the ecological functions which support those
19 resources.
20 1. The evaluation shall focus on ecological functions at the spatial scale of the
21 disposal site.
22 2. The relevance of potential impacts shall be judged at the spatial scale of the
23 disposal site (e.g., effects on subpopulations that use the site as habitat) rather
24 than the proportional significance of the site to regional environmental
25 resources.
26 The concentrations of contaminants at this site, compared to most sites assessed under hazardous
27 waste regulatory standards, are very high. By assessing aquatic life and wildlife that are exposed
28 to COCs, the risk assessment evaluated whether the contaminated habitats (i.e., the Primary
29 Study Area) are functioning as would normal habitats in the absence of contaminants. Vital
30 functions include providing adequate food and shelter and sustaining normal reproductive
31 success. The central question, for purposes of this assessment, is whether the exposed local
32 populations are thriving in the contaminated habitat, not whether the larger regional population is
33 surviving in spite of it.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-70
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
12.4.2.2 Ecological Implications
Populations of organisms are controlled by the balance between positive processes (e.g., growth,
reproduction, immigration) and negative processes (e.g., starvation, death from predation,
toxicant effects, emigration) (Taub 1989). Growth and reproduction are often controlled by food
supply and availability of adequate habitat. Mortality is often caused by predators and other
stressors in the environment. The dynamics of populations exposed to COCs have important
implications toward other interacting organisms and functions (Landis et al. 1998). Examples
include:
¦ Removal of Predators Compensates for Direct Effects of Contaminants—A toxic
chemical may increase food supply by reducing abundance of competitors or by
eliminating predators. In the Housatonic River ERA, the modeling of exposure and
effects line of evidence indicated that some species are experiencing risk from
exposure to tPCBs and TEQ (e.g., short-tailed shrews, largemouth bass), yet are fairly
abundant in the PSA. One possible explanation for this lack of concordance between
measurement endpoints is that elimination of predators (e.g., mink, river otter) may
be compensating for the direct effects due to tPCBs and TEQ.
For fish, the fishing ban imposed on the Housatonic River limits human predation on
the fish stocks, therefore likely compensating for the effects of tPCBs on recruitment
to older age classes. Elimination of predators, however, does not necessarily benefit
each prey population. For example, some prey populations may decline as a result of
increased competition from other prey populations (Bartell et al. 1992).
¦ Immigration Compensates for Direct Effects of Contaminants—The elimination
of individuals from a habitat creates openings that may be exploited by individuals
emigrating from surrounding habitats. Many species can migrate to the PSA or
within the PSA from areas of lower contamination to areas of higher contamination
and compensate for losses of organisms due to toxic effects; therefore, the presence of
a normal number of animals in a contaminated area does not necessarily demonstrate
that the population is unaffected or that the habitat is providing normal support
functions. If contaminant concentrations are such that organisms cannot reproduce
normally or thrive in the affected area without immigration from other areas, then
those effects are viewed by EPA as unacceptable. Immigration is likely the process
that explains the infrequent sightings of mink tracks in winter on the edges of the
PSA. In winter, juvenile and young adult mink often emigrate to other habitats,
particularly those that are not already occupied by mink, have an abundant food
supply, and offer ideal habitat (i.e., the PSA). The risk assessment for the PSA
indicates, however, that these mink are unlikely to survive and reproduce in the PSA.
¦ Populations Exposed to COCs May Be More Vulnerable to Other Stressors in
the Future—The studies conducted to support the ERA were done during a period of
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-71
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
regulatory restrictions, such as the fish, turtle, and frog advisories that have been in
place in the Housatonic River in Massachusetts since 1982. Several authors (e.g.,
Evans et al. 1990; Edwards et al. 1990) have surmised that lake trout populations in
the Great Lakes remained stable in the first half of the century despite fishing
pressure and the influence of other anthropogenic stressors. Lake trout populations
declined to very low levels by the 1970s, however, because of additional stressors
such as contaminants and introduction of sea lampreys. Thus, a fish population can
often tolerate some anthropogenic stresses, but if the combination of stresses becomes
too great, the population crashes. Density-independent stressors, such as toxic
chemicals that reduce fitness at all population densities, lower the capacity of
populations to respond to otherwise favorable conditions or to tolerate other stressors
(Evans et al. 1990). Control of sea lampreys, reductions in contaminant
concentrations, and reduced fishing pressure have not restored lake trout populations
to their historical levels (Edwards et al. 1990).
Many other indirect effects may occur as a result of the presence of tPCBs and TEQ in the PSA.
For example, Wu et al. (1993) and Spromberg et al. (1998) have shown that subpopulations in
patches removed from contamination may be affected by contaminated patches even if there is
no transfer of contaminant (the so-called "action at a distance"). Alternatively, COCs transferred
outside the PSA (e.g., by downstream transport, migration of birds) can augment exposures or
cause effects to organisms outside the PSA (e.g., hawks preying on migrating birds).
12.4.2.2.1 Genetic Diversity
Chemical adaptation has been cited as a compensatory mechanism that can enable populations to
survive at a site (Shugart 1996). This mechanism leads to selection of genotypes that are
resistant to a COC as a result of the elimination of sensitive individuals. The result is a
population with an altered genetic makeup, generally with a less diverse genetic pool. It is likely
that such reductions in the genetic pool cause alterations that may reduce a population's
resilience, making it more susceptible to other stresses in the future.
12.4.2.2.2 Immune System Effects
Non-coplanar PCBs can influence the activity of neutrophils (a type of white blood cell) through
mechanisms unrelated to the Ah receptor. In studies with rat- and human-derived neutrophils,
researchers have found that non-coplanar PCBs can activate biochemical pathways that lead to
the production of reactive oxygen species (ROS)(Fischer et al. 1998). Although the production
of ROS is a normal function of neutrophils (it is designed to destroy bacteria and viruses and to
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs qrs 7/11/2003
-------
1 break down tissue damaged by burns, chemicals, and physical injuries), when inappropriately
2 activated by PCBs, this neutrophil function can initiate harmful effects on healthy tissues
3 because of the destructive nature of ROS. Because neutrophils are among the first white blood
4 cells sent to sites of infection or inflammation, exposure to PCBs may weaken an animal's
5 immune and inflammatory responses.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
12.5 SOURCES OF UNCERTAINTY
The assessment of risks of COCs to aquatic and wildlife species in the Housatonic River contains
uncertainties. Each source of uncertainty can influence the estimates of risk; therefore, it is
important to describe and, when possible, specify the magnitude and direction of such
uncertainties. The sources of uncertainty associated with the assessment of risks of tPCBs and
TEQ to each assessment endpoint were summarized in Sections 3 through 11 and Appendices D
through K. This material is not repeated here. In this section, the most significant sources of
uncertainty commonly encountered throughout the ERA are described. The sources of
uncertainty are grouped by phase of the ERA (i.e., problem formulation, exposure assessment,
effects assessment, risk assessment).
12.5.1 Problem Formulation
The problem formulation is intended to define the linkages between stressors, potential exposure,
and predicted effects on ecological receptors. As such, the conceptual model provides the
scientific basis for selecting assessment and measurement endpoints to support the risk
assessment process. Potential uncertainties arise from lack of knowledge regarding ecosystem
functions, failure to adequately address spatial and temporal variability in the evaluations of
sources, fate and effects, omission of stressors, and overlooking secondary effects (EPA 1998).
The types of uncertainties associated with the conceptual model that links contaminant sources to
effects include those associated with the identification of COCs, environmental fate and transport
of COCs, exposure pathways, receptors at risk, and ecological effects. Of these, the
identification of exposure pathways probably represents the primary source of uncertainty in the
conceptual model. The detailed ecological characterization performed at this site has greatly
reduced the uncertainties associated with problem formulation, yet some remain and are
described below:
¦ The Housatonic River and surrounding floodplain have received chemical inputs
since the industrialization of the area. In addition to tPCBs and TEQ, other
contaminants identified in water, soil, and sediment samples include metals (e.g.,
mercury, lead, chromium); pesticides (e.g., aldrin, DDT, toxaphene, parathion, 2,4-
D); and semivolatile organic compounds (e.g., PAHs, chlorinated benzenes, anilines,
phenols). The conservative screening level assessment indicated that only tPCBs and
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-74
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
TEQ present a potential risk to wildlife species; therefore, only these COCs were
included in the probabilistic risk assessments for wildlife. Several other COCs were
screened through for aquatic life, although none were as influential as tPCBs.
Additive, synergistic, and antagonistic effects due to exposure to multiple
contaminants were not considered.
¦ In this assessment, it was assumed that dietary exposure represented the most
important pathway for the exposure of wildlife to COCs. Although unlikely to
provide a major contribution to the risk, other pathways could increase exposure and
perhaps increase risk slightly (Moore et al. 1999). Deterministic calculations were
conducted in which estimates of exposure to COCs via drinking water and inhalation
were included in the exposure model, but were not included in the assessment
because inclusion of these routes did not substantially increase overall exposure of
wildlife to the COCs. This issue was less important for aquatic life because these
assessments were conducted for multiple media exposures. The aquatic life endpoints
considered tissue burdens that integrated exposures from all sources and/or evaluated
exposures from the abiotic media (sediment, overlying water, porewater) deemed
most relevant to exposure.
12.5.2 Exposure Assessment
The exposure assessment is intended to describe the actual or potential co-occurrence of stressors
with receptors. As such, the exposure assessment identifies the exposure pathways and the
intensity and extent of contact with stressors for each receptor or group of receptors at risk. The
exposure models for wildlife were energetics-based models requiring information on body
weight, free living metabolic rate, proportions of food items in the diet, and the concentrations of
COCs in these food items. Each of these variables has associated uncertainties, most of which
were propagated through the exposure models. Exposure of fish species to COCs was assessed
using measured wet weight whole body tissue concentrations. Exposure of benthos to COCs was
assessed as either the COC concentrations in abiotic site media (i.e., sediment, water) or as the
tissue body burdens that represent integrated exposure from all sources.
¦ The greatest uncertainty in the benthic invertebrate and amphibian exposure
assessments was the potential for small-scale variability in exposure concentrations to
complicate the development of concentration-response relationships, additionally
confounded by analytical variability (Appendix C.ll). For studies that had replicate
measurements of tPCBs at a given station over a short period (e.g., benthic
macroinvertebrate sampling), the spatial variability was quantified and considered
explicitly in the derivation of concentration-response relationships. Where spatial
replication was not available, characterization of variability required the incorporation
of additional data sets.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-75
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
¦ Tissue chemistry data (tPCBs, TEQ) were relied upon in the characterization of
exposures to fish species and piscivorous wildlife. Total PCB concentrations in fish
exhibit seasonal fluctuations; these are sometimes related to lipid content changes that
occur during reproductive life history stages. To minimize confounding effects of
short-term variations in PCB concentrations due to spawning events, the vast majority
of the PCB data for this project were collected in the late summer and early fall.
Potential uncertainties (over time and space) in fish tissue chemistry were ameliorated
through the collection of a large number of samples (multiple species, ages, and
spanning multiple years of collections). Therefore, the ERA was conducted with a
robust data set (including a confirmatory analysis with non-EPA data sets) that
limited the probability of spurious outcomes in the exposure assessments.
¦ The Monte Carlo sensitivity analyses suggested that the parameters of the free
metabolic rate (FMR) allometric equation were generally the most influential
variables on predicted total daily intakes of COCs. However, no direct measurements
of free metabolic rate or food intake rate (other than for captive animals) were
available for most of the representative wildlife species. Therefore, free metabolic
rates were estimated using allometric equations. The use of allometric equations
introduces some degree of uncertainty into the exposure estimates because they are
subject to model-fitting error and are based on species different from the
representative species used in this assessment. Given the lack of data on species
specific to this assessment, it is difficult to judge the magnitude of the uncertainty
introduced by the use of the allometric models. The uncertainty due to model-fitting
error was propagated in the uncertainty analyses by using distributions as inputs for
the allometric slope and power terms.
¦ Sample sizes, while composites, were limited for the analyses of COC concentrations
in some prey items, including earthworms, litter invertebrates, and benthic
invertebrates. To address this uncertainty in the Monte Carlo analyses for wildlife,
the upper confidence limit (UCL) or data set maximum (see Section 6.4 and
Appendix C.5) was used as an estimate of COC concentrations in prey items. The
potential magnitude of the uncertainty associated with small sample sizes for COC
concentrations is unknown, but this approach likely overestimated exposure. The
probability bounds analysis used an unbiased approach (e.g., distribution free range
from LCL to UCL) to deal with sample size uncertainty.
12.5.3 Effects Assessment
The effects assessment is intended to describe the effects caused by stressors, link them to the
assessment endpoints, and evaluate how effects change with fluctuations in the levels (i.e.,
concentrations or doses) of the various stressors. In this assessment, the effects of tPCBs and
other COCs to representative species were assessed. There are several sources of uncertainty in
the assessment of effects, including measurement errors, extrapolation errors, and data gaps.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-76
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
¦ For benthos and amphibians, the effects benchmarks derived from the literature had a
high degree of uncertainty due to the need to extrapolate across sites and species. This
uncertainty was explicitly addressed in the weight-of-evidence evaluation.
¦ The site-specific fish toxicity studies indicated variations in the concentration-
response relationships observed across species, reaches, and treatments, and
introduced uncertainty into the development of effects thresholds.
¦ The methodology used to mimic maternal transfer used in the fish Phase II studies has
been recently developed and has not been widely applied as an environmental
monitoring technique; therefore, there are potential uncertainties inherent to
extrapolating these laboratory-based results to Housatonic River fish. Similarly, the
extrapolation of concentrations of tPCBs in egg to whole body concentrations has a
degree of uncertainty associated with it. The magnitude and direction of the
uncertainty is unknown.
¦ The greatest potential source of uncertainty for the fish and wildlife effects
assessments was associated with the lack of toxicity studies involving the
representative species. The direct assessment of effects to benthos, amphibians,
largemouth bass, and mink included studies on the effects of tPCBs to reproduction
and/or survival for these representative species. There were, however, no toxicity
studies available for many other representative species exposed to tPCBs or TEQ. As
a result, laboratory studies involving other species were often used to estimate effects
to representative species. To address uncertainty in the effects assessments, threshold
ranges were used in which effects to tolerant and sensitive species were considered.
It was assumed that the toxicity thresholds for the representative species lie within
these ranges.
¦ The effects metrics used to estimate risk from the literature were derived for Aroclor
1254 and Aroclor 1260 mixtures, and more information was available for Aroclor
1254 than for Aroclor 1260. Some uncertainty is inherent in extrapolating from
studies using commercial Aroclor mixtures to the specific congener patterns observed
in weathered mixtures in the PSA of the Housatonic River. The potential magnitude
and direction of the uncertainty associated with this extrapolation are unknown.
¦ TEQ is an expression of the planar chlorinated hydrocarbons (PCHs). TEQ are
derived from an equation that combines the relative potency of each congener into a
single concentration. The potencies of individual congeners are not known precisely
and were estimated based on a combination of data and professional judgment (Van
den Berg et al. 1998). Although there is uncertainty in these calculations, this
approach has been accepted and applied in numerous jurisdictions worldwide. The
potential magnitude and direction of the uncertainty associated with this approach are
unknown. The potential toxic effects of congeners not evaluated by this method are
poorly understood and can not be quantified.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-77
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
12.5.4 Risk Characterization
A weight-of-evidence procedure was used to assess risks of tPCBs and TEQ to the assessment
endpoints in the Housatonic River PSA. The analysis follows the methodology proposed by the
Massachusetts Weight-of-Evidence Workgroup (Menzie et al. 1996; see Section 2.9 for details).
In general, the weight-of-evidence approach is an inclusive process whereby multiple lines of
evidence are considered prior to determining risk. For the wildlife risk assessments, these lines
of evidence included the exposure and effects modeling results and, in some cases, field survey
results, and/or in situ or whole media toxicity test results. For the fish and benthic invertebrate
risk assessments, available lines of evidence included field survey results (e.g., community
evaluation for benthos), site-specific toxicity tests, and comparison of tissue and sediment
concentrations to benchmarks (both from the literature and site-specific benchmarks).
¦ Uncertainty in the risk characterization arises from the absence of one or more of the
available lines of evidence. In the case of piscivorous birds, data on two of the three
major lines of evidence were available, i.e., comparison of modeled exposure and
effects and the field study. Threatened and endangered species had only the modeled
exposure and effects line of evidence to support the risk assessment. The
consequence of the lack of multiple concurring lines of evidence is less confidence in
the conclusion regarding risk. For example, the risk characterization of piscivorous
mammals had three major lines of evidence available, thus providing high confidence
in the risk conclusions. The risk characterization for fish and benthic invertebrates
also had three major lines of evidence available.
¦ Uncertainty for individual lines of evidence was sometimes sufficiently large to
render the line of evidence of limited use in the ERA. For example, the community
evaluation for fish was confounded by large habitat variations combined with small
overall gradients and large small-scale variation in PCB concentrations. These factors
made derivations of concentration-response relationships unfeasible (i.e., due to very
low statistical power for determining contaminant-induced effects) and limited the
studies to qualitative assessments.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-78
-------
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
12.6 ERA CONCLUSIONS
¦ Weight-of-evidence assessments indicated that aquatic life and wildlife in the Primary
Study Area of the Housatonic River are experiencing unacceptable risks as a result of
exposure to tPCBs and other COCs. Confidence in this conclusion is high for benthic
invertebrates, amphibians, and piscivorous mammals because multiple lines of
evidence gave concordant results.
¦ The risks of tPCBs and other COCs likely extend beyond the representative species
considered in the quantitative risk assessments described herein. Qualitative risk
assessments indicated that many other species in the PSA are potentially at risk.
Further, there are likely indirect effects (e.g., changes in predator-prey relationships,
changes in metapopulation dynamics) occurring inside and outside the PSA as a result
of the direct impacts caused by tPCBs and other COCs.
¦ An assessment of risk downstream of the PSA indicated that tPCBs could potentially
be causing adverse effects to benthic organisms in depositional areas as far as Reach
8, amphibians in floodplain areas as far as Reach 8, trout in Reaches 7 and 9, mink as
far as Reach 10, otter as far as Reach 12, and bald eagle in Reach 8.
MK01|O:\20123001.096\ERA_PB\ERA_PB_12.DOC i rs 7/11/2003
-------
1 12.7 REFERENCES
2 Amrhein, J.F., C.A. Stow and C. Wible. 1999. Whole-fish versus filet polychlorinated-biphenyl
3 concentrations: An analysis using classification and regression tree models. Environmental
4 Toxicology and Chemistry 18:1817-1823.
5 Angermann, J.E., and F. Matsumura. 1999. Effects of PCBs and Cold Exposure on Immune
6 Function and Hepatic Glycogenolysis in Rana pipiens. Society of Environmental Toxicology and
7 Chemistry (SETAC). Presentation, SETAC 20th Annual Meeting, Nashville, Tennessee.
8 Anthony, E.P.L. and T.H. Kunz. 1977. Feeding strategies of the little brown bat in southern New
9 Hampshire. Ecology 58:775-786.
10 Aulerich, R.J., S.J. Bursian, W.J. Breslin, B.A. Olson, and R.K. Ringer. 1985. Toxicological
11 manifestations of 2,4,5-, 2',4',5'-, 2,3,6,2',3',6'-, and 3,4,5,3',4',5'-Hexachlorobiphenyl and
12 Aroclor 1254 in mink. Journal of Toxicology and Environmental Health 15:63-79.
13 Azure, D.A. 1998. Aspects of American Bittern ecology in northwest Minnesota. M.S. Thesis.
14 University of North Dakota, Grand Forks, North Dakota. 139 pp. Cited in: J. A. Deschant, M.L.
15 Sondreal, D.H. Johnson, L.D1, Igl, C.M. Goldade, A.L. Zimmerman, and B. R. Euliss. 2001.
16 Effects of Management Practices on Grassland Birds: American Bittern. Northern Prairie
17 Wildlife Research Center, Jamestown, ND, USA. Available at:
18 http://www.npwrc.usgs.gov/resource/literatr/grasbird/ambi/ambi.htm (Version 17FEB2000).
19 Barbour, R.W. and W.H. Davis. 1969. Bats of America. University of Kentucky Press.
20 Lexington, Kentucky. 286 pp. Cited in: R.M. DeGraaf, and M. Yamasaki. 2001. New England
21 Wildlife: Habitat, Natural History, and Distribution. University Press of New England, Hanover,
22 New Hampshire.
23 Barron, M.G., H. Galbraith, and D. Beltman. 1995. Comparative reproductive and developmental
24 toxicology of PCBs in birds. Comparative Biochemistry and Physiology 112C(1) 1 -14.
25 Bartell, S.M., R.H. Gardner and R.V. O'Neill. 1992. Ecological Risk Estimation. Lewis
26 Publishers, Boca Raton, FL.
27 Bellis, E.D. 1968. Summer movement of red-spotted newts in a small pond. Journal Herpetology
28 1:86-91.
29 Belwood, J.J. and M.B. Fenton. 1976. Variation in the diet of Myotis lucifugus. Canadian
30 Journal of Zoology 54:1674-1678.
31 Berger, A.J. 1951. Nesting density of Virginia and sora rails in Michigan. Condor 53:202. Cited
32 in: R.M. DeGraaf, and M. Yamasaki. 2001. New England Wildlife: Habitat, Natural History, and
33 Distribution. University Press of New England, Hanover, New Hampshire.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-80
-------
1 Bevelhimer, M.S., J.J. Bauchamp, B.E. Sample and G.R. Southworth. 1997. Estimation of
2 Whole-Fish Contaminant Concentrations From Fish Fillet Data. ES/ER/TM-202. Oak Ridge
3 National Laboratory, Oak Ridge, Tennessee.
4 Birnbaum, L.S., and M.J. DeVito. 1995. Use of toxic equivalency factors for risk assessment for
5 dioxins and related compounds. Toxicology 105:391-401.
6 Bishop, C.A., M.D. Koster, A.A. Chek, D.J.T. Hussell, and K. Jock. 1995. Chlorinated
7 hydrocarbons and mercury in sediments, red-winged blackbirds (Agelaius phoeniceus) and tree
8 swallows (Tachycineta bicolor) from wetlands in the Great Lakes-St. Lawrence River basin.
9 Environmental Toxicology and Chemistry 14:491-501.
10 Bishop, C.A., H.A. Mahony, S. Trudeau, and K.E. Pettit. 1999. Reproductive success and
11 biochemical effects in tree swallows (Tachycineta bicolour) exposed to chlorinate hydrocarbon
12 contaminants in wetlands of the Great Lakes and St. Lawrence River basin, USA and Canada.
13 Environmental Toxicology and Chemistry 18:263-271.
14 Bleavins, M.R., R.J. Aulerich, and R.K. Ringer. 1980. Polychlorinated biphenyls (Aroclor 1016
15 and 1242): effects on survival and reproduction in mink and ferret. Archives of Environmental
16 Contamination and Toxicology 9:627-63 5.
17 Bosveld, A.T.C, and M. Van den Berg. 1994. Effects of polychlorinated biphenyls: dibenzo-p-
18 dioxins and dibenzofurans on fish-eating birds. Environmental Reviews 2:147-166.
19 Brown, M. and J.J. Dinsmore. 1986. Implications of marsh size and isolation for marsh bird
20 management. Journal of Wildlife Management 50:392-397.
21 Brunstrom, B. 1988. Sensitivity of embryos from duck, goose, herring gull, and various chicken
22 breeds to 3,3', 4,4'-tetrachlorobiphenyl. Poultry Science 67:52-57.
23 Brunstrom, B. 1989. Toxicity of coplanar polychlorinated biphenyls in avian embryos.
24 Chemosphere 19(l-6):765-768.
25 Buckler, DR. 2002. Fish reproductive health assessment in PCB contaminated regions of the
26 Housatonic River, Massachusetts, USA: Investigations of causal linkages between PCBs and fish
27 health. Interim Report of Phase II Studies. Prepared by D.R. Buckler for U.S. Fish and Wildlife
28 Service, Concord, New Hampshire and U.S. Environmental Protection Agency, Boston,
29 Massachusetts. September 27, 2002.
30 Bursian, S.J., R.J. Aulerich, B. Yamini, and D.E. Tillitt. 2003. Dietary exposure of mink to fish
31 from the Housatonic River: Effects on reproduction and survival. Michigan State University,
32 Department of Animal Science. Final Report. June 10, 2003.
33 Burton, T.M. 1977. Population estimates, feeding habits, and nutrient and energy relationships of
34 Notophthalmus v. viridescens, in Mirror Lake, New Hampshire. Copeia. 1977:139-143.
35 CCME (Canadian Council of Ministers of the Environment). 1999. Canadian Tissue Residue
36 Guidelines for the Protection of Wildlife Consumers of Aquatic Biota: Polychlorinated
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-81
-------
1 Biphenyls (PCBs). In: Canadian Environmental Quality Guidelines 1999. Canadian Council of
2 Ministers of the Environment, Winnipeg.
3 Cogliano, V.J. 1998. Assessing the cancer risk from environmental PCBs. Environmental Health
4 Perspectives 106(6) :317-323
5 Conway, C.J. 1995. Virginia rail (Rallus limicola). In The Birds of North America. No. 421. A.
6 Poole and F. Gill, Editors. Academy of Natural Sciences, Washington, DC.
7 Custer, C.M., T.W. Custer, P.D. Allen, K.L. Stromborg, and M. Melancon. 1998.
8 Organochlorine contaminants and tree swallows nesting along the Fox River and in Green Bay,
9 Wisconsin. Environmental Toxicology and Chemistry 17:1786-1798.
10 DeGraaf, R.M. and M. Yamasaki. 2001. New England Wildlife: Habitat, Natural History, and
11 Distribution. University Press of New England, Hanover, New Hampshire.
12 Deschant, J.A., M.L. Sondreal, D.H. Johnson, L.D1, Igl, C.M. Goldade, A.L. Zimmerman, and B.
13 R. Euliss. 2001. Effects of Management Practices on Grassland Birds: American Bittern.
14 Northern Prairie Wildlife Research Center, Jamestown, North Dakota. Available at:
15 http://www.npwrc.usgs.gov/resource/literatr/grasbird/ambi/ambi.htm (Version 17FEB2000).
16 DiBello, F., S. Arthur, and W. Krohn. 1990. Food habits of sympatric coyotes, red foxes, and
17 bobcats in Maine. Canadian Field Naturalist 104:403-408.
18 Eadie, W.R. 1944. The short-tailed shrew and field mouse predation. Journal of Mammalogy
19 25:359-364.
20 Eadie, W.R. 1948. Shrew-mouse predation during low mouse abundance. Journal of
21 Mammalogy 29:35-37.
22 Edwards, C.J., R.A. Ryder, and T.R. Marshall. 1990. Using lake trout as a surrogate of
23 ecosystem health for oligotrophic waters of the Great Lakes. Journal of Great Lakes Research
24 16:591-608.
25 Ehrlich, P.R., D.S. Dobkin, and D. Wheye. 1988. The Birder's Handbook - A Field Guide to the
26 Natural History of North American Birds. Fireside Press, New York.
27 Ehrlich, P.R., D.S. Dobkin, and D. Wheye. 1992. The Birder's Handbook - A Field Guide to the
28 Natural History of North American Birds. Fireside Press. New York.
29 Eisler, R. 2000. Handbook of Chemical Risk Assessment: Health Hazards to Humans, Plants,
30 and Animals. Volume 2 - Organics. Lewis Publishers, Boca Raton, FL. ISBN 1-56670-506-1.
31 Environment Canada. 1997. Environmental Risk Assessments of Priority Substances und the
32 Canadian Environmental Protection Act: Guidance Manual. Chemicals Evaluation Division,
33 Commercial Chemicals Evaluation Branch, Environment Canada, Ottawa, Canada.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-82
-------
1 EPA (U.S. Environmental Protection Agency). 1993. Wildlife Exposure Factors Handbook.
2 Volumes I and II. Office of Research and Development, Washington, DC. EPA EPA/600/R-
3 93/187a, EPA/600/R-93/187b.
4 EPA (U.S. Environmental Protection Agency). 1997. Ecological Risk Assessment Guidance for
5 Superfund: Process for Designing and Conducting Ecological Risk Assessments. Interim Final
6 Report. Solid Waste and Emergency Response. EPA 540-R-97-006. June 1997.
7 EPA (U.S. Environmental Protection Agency). 1998. Guidelines for Ecological Risk Assessment.
8 EPA/630/R-95/002F. Risk Assessment Forum, Washington, DC.
9 EPA (U.S. Environmental Protection Agency). 1999a. Hudson River PCBs Reassessment RI/FS;
10 Volume 2E- Baseline Ecological Risk Assessment. Prepared for EPA, Region 2 and the US
11 Army Corps of Engineers, Kansas City District. Prepared by TAMS/MCA. August 1999.
12 EPA (U.S. Environmental Protection Agency). 1999b. Issuance of Final Guidance: Ecological
13 Risk Assessment and Risk Management Principles for Superfund Sites; Signed 10/7/99.
14 Memorandum from Stephen D. Luftig, Office of Emergency and Remedial Response,
15 Washington, DC.
16 EPA (U.S. Environmental Protection Agency). 2000. Methods for Measuring the Toxicity and
17 Bioaccumulation of Sediment-Associated Contaminants with Freshwater Invertebrates. EPA
18 600/R-99/064. 192 pp.
19 EPA (U.S. Environmental Protection Agency). 2001. Ecological Characterization of the
20 Housatonic River: Revised Preliminary Draft. DCN to be assigned. Boston, Massachusetts.
21 Prepared by Woodlot Alternatives, Inc. with Roy F. Weston under EPA contract No. DACW33-
22 94-D-0009/032.
23 Evans, D.O., G.J. Warren, and V.W. Cairns. 1990. Assessment and management of fish
24 community health in the Great Lakes: synthesis and recommendations. Journal of Great Lakes
25 Research 16:639-669.
26 Fischer, L., R.F. Seegal, P. Ganey, I.N. Pessah, and P.R.S. Kodavanti. 1998. Toxicity of non-
27 coplanar PCBs. Toxicological Sciences 41:49-61.
28 Georgii, S., G. Bachour, K. Failing, U. Eskins, I. Emadfa, and W.L. Budde. 1985. Mass
29 spectrometric determination of polychlorinated biphenyl congeners in foxes in Germany from
30 1983 to 1991. Archives of Environmental Contamination and Toxicology 26:1-6.
31 Gibbs, J.P., S. Melvin, and F.A. Ried. 1992. American Bittern (Botaurus lentiginosus). In: The
32 Birds of North America, No. 18. A. Poole, P. Stettenheim, and F. Gill, Editors. The Birds of
33 North America, Inc. Philadelphia, PA. pp. 1-11.
34 Giesy, J.P., and K. Kannan. 1998. Dioxin-like and non-dioxin-like toxic effects of
35 polychlorinated biphenyls (PCBs): implications for risk assessment. Critical Reviews in
36 Toxicology 28(6):511-569.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-83
-------
1 Griffith, L.A. and J.E. Gates. 1985. Food habits of cave-dwelling bats in the central
2 Appalachians. Journal of Mammalogy 66:451-460.
3 Gunn, J.M., S.U. Qadri, and D.C. Mortimer. 1997. Filamentous algae as a food source for the
4 brown bullhead. Journal of the Fisheries Research Board of Canada 34:396-401.
5 Hamilton, W.J., Jr. 1941. The foods of small forest mammals in eastern United States. Journal of
6 Mammalogy 22:250-263.
7 Hartel, K.E., D.B. Halliwell, and A.E. Launer. 2002. Inland Fishes of Massachusetts.
8 Massachusetts Audubon Society. Lincoln, Massachusetts.
9 Healy, W.R. 1975. Terrestrial activity and home range in efts of Notophthalmus v. viridescens.
10 American Midland Naturalist 93:131-138.
11 Hockman, J.G., and J.A. Chapman. 1983. Comparative feeding habits of red foxes (Vulpes
12 vulpes) and gray foxes (Urocyon cinereoargentus) in Maryland. American Midland Naturalist
13 110:267-285.
14 Huang, Y., D.J. Hoffman, and W.H. Karasov. 2001. Oxidative stress in PCB 126-exposed frogs,
15 Rana pipiens. Presentation, SET AC 22nd Annual Meeting, Baltimore, MD.
16 Hull, R.N. and G.W. Suter II. 1994. Using the Weight-of-Evidence Approach for Ecological Risk
17 Assessment at a DOE Facility. SET AC Denver, CO. November 1, 1994.
18 Hunter, M.L., A.J.K. Calhoun, and M. McCollough. 1999. Maine Amphibians and Reptiles. The
19 University of Maine Press, Orono, Maine. 252 pp.
20 Jenssen, T.A., and W.D. Klimstra. 1966. Food habits of the green frog, Rana clamitans, in
21 southern Illinois. American Midland Naturalist 76:169-182.
22 Jung, R.E., and M.K. Walker. 1997. Effects of 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) on
23 development of anuran amphibian. Environmental Toxicology and Chemistry 16:230-240.
24 Knable, A.E. 1974. Seasonal trends in the utilization of major food groups by the red fox (vulpes
25 fulva) in Union County, Illinois. Trans. III. State Acad. Sci. 66:113-115.
26 Korschgen, L.J. 1959. Food habits of the red fox in Missouri. Journal of Wildlife Management
27 23:168-176.
28 Kurta, A. 1995. Mammals of the Great Lakes Region. The University of Michigan Press. Ann
29 Arbor, MI. 392 pp.
30 Kurta, A. and R.H. Baker. 1990. Eptesicus fuscus. Mammalian Species No. 356:1-10. American
31 Society of Mammalogists. Shippensburg, Pennsylvania.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-84
-------
1 Landis, W.G., D.R.J. Moore, and S.B. Norton. 1998. Ecological risk assessment: looking in,
2 looking out. In: Pollution Risk Assessment and Management. P.E.T. Douben (Ed). John Wiley &
3 Sons, Toronto.
4 Leonards, P.E.G., Y. Zierikzee, U.A. Brinkman, W.P. Cofino, N.M. van Straalen, and B. van
5 Hattum. 1997. The selective dietary accumulation of planar poly chlorinated biphenyls in the
6 otter {Lutra lutra). Environmental Toxicology and Chemistry 16(9): 1807-1815.
7 Linzey, D.W. and A.V. Linzey. 1973. Notes on food of small mammals from Great Smokey
8 Mountains National Park, Tennessee-North Carolina. J. ElishaMitchell Science Society 89:6-14.
9 MacNamara, M.C. 1977. Food habits of terrestrial adult migrants and immature red efts of the
10 red-spotted newt, Notophthalmus viridescens. Herpetologica 33:127-132.
11 Martin, A.C., H.S. Zim, and A.L. Nelson. 1951. American Wildlife and Plants - A Guide to
12 Wildlife Food Habits. Dover Publications, Inc., New York.
13 McCarty, J.P., and A.L. Secord. 1999. Reproductive ecology of tree swallows (Tachycineta
14 bicolor) with high levels of polychlorinated biphenyl contamination. Environmental Toxicology
15 and Chemistry 18:1433-1439.
16 Menzie, C., M.H. Henning, J. Cura, K. Finkelstein, J. Gentile, J. Maughan, D. Mitchell, S.
17 Petron, B. Potocki, S. Svirsky, and P. Tyler. 1996. Special report of the Massachusetts Weight-
18 of-Evidence Workgroup: A weight-of-evidence approach for evaluating ecological risks. Human
19 and Ecological Risk Assessment 2:277-3 04.
20 Mills, R.S., G.W. Barrett, and M.P. Farrell. 1975. Population dynamics of the big brown bat
21 (Eptesicus fuscus) in southwestern Ohio. Journal of Mammalogy 56:591-604. Cited in: R.M.
22 DeGraaf, and M. Yamasaki. 2001. New England Wildlife: Habitat, Natural History, and
23 Distribution. University Press of New England, Hanover, New Hampshire.
24 Moore, D.R.J., and S.M. Bartell. 2000. Estimating ecological risk of multiple stressors: advanced
25 methods and difficult issues. In: Multiple Stressors in Ecological Risk and Impact Assessment:
26 Approaches to Risk Estimation. S.A. Ferenc and J.A. Foran (Eds). Society of Environmental
27 Toxicology and Chemistry (SETAC).
28 Moore, D.R.J., B.E. Sample, G.W. Suter, B.R. Parkhurst, and R.S. Teed. 1999. A probabilistic
29 risk assessment of the effects of methylmercury and PCBs on mink and kingfishers along East
30 Fork Poplar Creek, Oak Ridge, Tennessee. Environmental Toxicology and Chemistry 18:2941-
31 2953.
32 Newsted, J.L., J.P. Giesy, G.T. Ankley, D.E. Tillitt, R.A. Crawford, J.W. Gooch, P.D. Jones, and
33 M.S. Denison. 1995. Development of toxic equivalency factors for PCB congeners and the
34 assessment of TCDD and PCB mixtures in rainbow trout. Environmental Toxicology and
35 Chemistry 14(5): 861 -871
36 Odum, R.R. 1977. Sora (Porzana Carolina). In: Management of migratory shore and upland
37 game birds in North America, pp. 57-65. G.C. Sanderson, Editor. Washington, D.C.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-85
-------
1 Petranka, J.W. 1998. Salamanders from the United States and Canada. Smithsonian Institution
2 Press. Washington, DC.
3 Powell, D.G. and R.M. Case. 1982. Food habits of the red fox in Nebraska. Trans. Nebr. Acad.
4 Sci. andAffil. Soc. 10:13-16.
5 Robertson, R.J., B.J. Stutchbury, and R.R. Cohen. 1992. Tree Swallow (Tachycineta bicolor). In:
6 The Birds of North America, No. 77, pp. 1-26. A. Poole, P. Stettenheim, and F. Gill, Editors. The
7 Birds of North America, Inc., Philadelphia, Pennsylvania.
8 Safe, S.H. 1994. Polychlorinated Biphenyls (PCBs): Environmental impact, biochemical and
9 toxic responses, and implications for risk assessment. Critical Reviews in Toxicology 24(2):87-
10 149.
11 Sallabanks, R. and F.C. James. 1999. American Robin (Turdus migratorius) In: The Birds of
12 North America, No. 642, pp. 1-27. A. Poole, P. Stettenheim, and F. Gill, Editors. The Birds of
13 North America, Inc., Philadelphia, Pennsylvania.
14 Shugart, L.R. 1996. Molecular markers to toxic agents. In: Ecotoxicology: A Hierarchical
15 Treatment, pp. 154-156. M.C. Newman and C.H. Jagoe, Editors. CRC Lewis Publishers, New
16 York.
17 Sibley, D.A. The Sibley Guide to Bird Life & Behavior. 2001. Alfred A. Knopf, Inc., New York.
18 Spromberg, J.A., B.M. John, and W.G. Landis. 1998. Metapopulation dynamics: Indirect effects
19 and multiple discrete outcomes in ecological risk assessment. Environmental Toxicology and
20 Chemistry 17(8): 1640-1649.
21 Stewart, M.M. and P. Sandison. 1973. Comparative food habits of sympatric mink frogs,
22 bullfrogs, and green frogs. Journal of Herpetology 6:241-244.
23 Suter, G.W., II, B.E. Sample, D.S. Jones, T.L. Ashwood, and J.M. Loar. 1995. Approach and
24 Strategy for Performing Ecological Risk Assessments for the U.S. Department of Energy's Oak
25 Ridge Reservation. Oak Ridge National Laboratory, Oak Ridge, TN. ES/ER/TM-33/R2.
26 Tanner, W.D. and G.O. Hendrickson. 1956. Ecology of the sora in Clay County, Iowa. Iowa Bird
27 Life 26(4):78-81. Cited in: R.M. DeGraaf, and M. Yamasaki. 2001. New England Wildlife:
28 Habitat, Natural History, and Distribution. University Press of New England, Hanover, New
29 Hampshire.
30 Taub, F.B. 1989. Standardized aquatic microcosms. Environmental Science and Technology
31 23(9): 1064-1066.
32 Van den Berg, M., L. Birnbaum, A.T.C. Bosveld, B. Brunstrom, P. Cook, M. Freely, J.P. Giesy,
33 A. Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J.C. Larsen, F.X.R van Leeuwen, A.K.
34 Djien Liem, C. Nolt, R.E. Peterson, L. Poellinger, S. Safe, D. Schrenk, D. Tillitt, M. Tysklind,
35 M. Younes, F. Waern, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs) for PCBs,
36 PCDDs, PCDFs for humans and wildlife. Environmental Health Perspectives 106(12):775-792.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-86
-------
1 van Zyll de Jong, C.G. 1985. Handbook of Canadian Mammals - Bats. National Museum of
2 Natural Sciences. Ottawa, Canada. 212 pp. Cited in: R.M. DeGraaf, and M. Yamasaki. 2001.
3 New England Wildlife: Habitat, Natural History, and Distribution. University Press of New
4 England, Hanover, New Hampshire.
5 Walker, M.K., P.M. Cook, A.R. Batterman, B.C. Butterworth, C. Berini, J.J. Libal, L.C.
6 Hufnagle, and R.E. Peterson. 1994. Translocation of 2,3,7,8-tetrachlorodibenzo-p-dioxin from
7 adult female lake trout {Salvelinus namaycush) to oocytes: effects on early life stage
8 development and sac fry survival. Can. J. Fish. Aquat. Sci. 51:1410-1419.
9 Weatherhead, P.J. and S.B. McRae. 1990. Brood care in American robins: implications for
10 mixed reproductive strategies by females. Anim. Behav. 39:1179-1188.
11 WESTON (Roy F. Weston, Inc.). 2000a. Supplemental Investigation Work Plan for the Lower
12 Housatonic River. Prepared for U.S. Army Corps of Engineers and U.S. Environmental
13 Protection Agency. 22 February 2000. DCN GEP2-020900-AAME.
14 WESTON (Roy F. Weston, Inc.). 2001a. Quality Assurance Project Plan, Vol. I - Text, Vol. II -
15 Appendix A, Vol. IIA - Appendix A, cont'd., Vol. IV - Appendices E and F. Prepared for U.S.
16 Army Corps of Engineers and U.S. Environmental Protection Agency. DCN GE-021601-
17 AAHM.
18 Whitford, A.G., and A. Vinegar. 1966. Homing, survivorship, and overwintering of larvae in
19 Ambystoma maculatum. Copeia. 1966:515-519.
20 Whitaker, J.O. 1995. Food of the big brown bat Eptesicus fuscus from maternity colonies in
21 Indiana and Illinois. American Midland Naturalist 134:346-360.
22 Whitaker, J.O., Jr. and M.G. Ferraro. 1963. Summer food of 220 short-tailed shrews from Ithaca,
23 New York. Journal of Mammalogy 44:419.
24 Whitaker, J.O. and W.J. Hamilton. 1998. Mammals of the Eastern United States. Cornell
25 University Press. Ithaca, New York.
26 Woodlot Alternatives, Inc. 2003. Amphibian reproductive success within vernal pools associated
27 with the Housatonic River, Pittsfield to Lenoxdale, Massachusetts. DCN: 99-1275.
28 Wu, J., J.L. Vankat, and Y. Barlas. 1993. Effects of patch connectivity and arrangement on
29 animal metapopulation dynamics: a simulation study. Ecological Modeling 65:221-254.
30 Zabel, E.W., P.M. Cook, and R.E. Peterson. 1995. Toxic equivalency factors of polychlorinated
31 dibenzo-/>dioxin, dibenzofuran and biphenyl congeners based on early life stage mortality in
32 rainbow trout (iOncorhynchus mykiss). Aquat. Toxicol. 31:315-328.
MK01 |O:\20123001,096\ERA_PB\ERA_PB_12. DOC
12-87
------- |