HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
For
U.S. Environmental Protection Agency
Region 2
and
U.S. Army Corps of Engineers
Kansas City District
Book 1 of 1
TAMS Consultants, Inc.
and
Menzie-Cura & Associates, Inc.
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
REGION 2
290 BROADWAY
NEW YORK, NY 10007-1866
¦? \
April 27. 1999
To All Interested Parties:
The U.S. Environmental Protection Agency (USEPA) is pleased to release the Responsiveness
Summary for the Phase 2 Ecological Risk Assessment Scope of Work (ERASOW) for the
Hudson River PCBs Reassessment Remedial Investigation/Feasibility Study (Reassessment
RI/FS).
This Responsiveness Summary contains USEPA's responses to the public comments received
the September 1998 ERASOW. The ERASOW presented USEPA's general approach for
conducting the Ecological Risk Assessments for the Upper Hudson River and for the Lower
Hudson River. The Upper Hudson River Ecological Risk Assessment will be completed in
Summer 1999. The Lower Hudson River Ecological Risk Assessment will be completed
following USEPA's review of the revised Thomann-Farley model developed for the Hudson
River Foundation.
If you have any questions regarding this Responsiveness Summary or the Reassessment RI/FS
general, please contact Ann Rychlenski. the Community Relations Coordinator for the site, at
(212) 637-3672.
Sincerely yours,
Richard L. Caspe, Director
Emergency and Remedial Response Division
Internet Address (URL) http://wvyw.spa.gov
Recycled/RacyclabI* Printed with Vegetable 01 Based Inks on Recycled Paper (Minimum 25% Postconsumer)
-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
! sUzJ
PRO^
For
U.S. Environmental Protection Agency
Region 2
and
U.S. Army Corps of Engineers
Kansas City District
Book 1 of 1
TAMS Consultants, Inc.
and
Menzie-Cura & Associates, Inc.
-------
HUDSON RIVER PCBs REASSESSMENT Rl/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
TABLE OF CONTENTS
BOOK 1 OF 1 Page
ACRONYMS
I. INTRODUCTION AND COMMENT DIRECTORY
1. INTRODUCTION 1
1.1 Recent Developments 2
2. COMMENTING PROCESS 2
2.1 Distribution of ERASOW 2
2.2 Review Period and Public Availability Meetings 2
2.3 Receipt of Comments 6
2.4 Distribution of Responsiveness Summary 6
3. ORGANIZATION OF ERASOW COMMENTS AND RESPONSIVENESS
SUMMARY 6
3.1 Identification of Comments 6
3.2 Location of Responses to Comments 7
4. COMMENT DIRECTORY 8
4.1 Guide to Comment Directory 8
4.2 Comment Directory 9
II. RESPONSE TO COMMENTS ON THE ERASOW
General Comments 13
1. INTRODUCTION 13
1.1 Site History 13
1.2 Ecological Risk Assessment in the Superfund Process . 14
1.3 Results of Phase 1 Ecological Risk Assessment 14
1.4 Changes in EPA Risk Assessment Guidance Since the Phase 1 Assessment ... 14
1.5 Additional Toxicological Benchmarks Developed by ORNL Since the
Phase 1 Assessment 14
1.6 Organization of the Phase 2 ERA Based on USEPA 1997 Guidance 14
2. PROBLEM FORMULATION 15
i
T AMS/Menzie-Cura
-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
TABLE OF CONTENTS
BOOK 1 OF 1 Page
2.1 Site Characterization 16
2.1.1 Upper Hudson River 16
2.1.2 Thompson Island Pool 16
2.1.3 Lower Hudson River 16
2.2 Contaminants of Concern 17
2.3 Assessment Endpoints 17
2.4 Site Conceptual Model 18
2.5 Measurement Endpoints 18
2.6 Receptors of Concern 20
2.6.1 Macroinvertebrate Communities 20
2.6.2 Fish Receptors 20
2.6.3 Avian Receptors 20
2.6.4 Mammalian Receptors 21
2.6.5 Threatened and Endangered Species 21
2.6.6 Significant Habitats 21
2.7 Risk Questions 21
3. EXPOSURE ASSESSMENT 21
3.1 Exposure Pathways 22
3.1.1 PCBs in Sediments 22
3.1.2 PCBs in Water 22
3.1.3 Benthic Invertebrates 22
3.1.4 Fish Receptors 22
3.1.5 Avian Receptors 22
3.1.6 Mammalian Receptors 22
3.2 Quantification of PCB Fate and Transport 22
3.3 Observed Exposure Concentration 23
3.3.1 Sediment Concentrations 24
3.3.2 Water Column Concentrations 24
3.3.3 Benthic Invertebrate Concentrations 24
3.3.4 Fish Concentrations 25
3.3.5 Avian Concentrations 25
3.3.6 Mammalian Concentrations 25
3.4 Modeled Exposure Concentrations 25
3.4.1 Benthic Invertebrate Receptors 26
ii TAMS/Menzie-Cura
-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
TABLE OF CONTENTS
BOOK 1 OF 1 Page
3.4.2 Fish Receptors 26
3.4.3 Avian Receptors 26
3.4.4 Mammalian Receptors 26
4. EFFECTS ASSESSMENT 27
4.1 Estimating the Toxicity of PCBs 27
4.1.1 Total PCBs and Aroclor Toxicities 28
4.1.2 Congener-specific Toxicity and the Toxicity Equivalency Factors (TEF)
Approach 28
4.2 Measures of Effect 28
4.2.1 Benthic Invertebrate Communities 29
4.2.2 Fish Receptors 29
4.2.3 Avian Receptors 29
4.2.4 Mammalian Receptors 29
4.2.5 Threatened and Endangered Species 29
4.2.6 Significant Habitats 30
5. RISK CHARACTERIZATION 30
5.1 Surface Water Concentrations 30
5.2 Sediment Concentrations 30
5.3 Benthic Invertebrates 30
5.4 Fish Receptors 30
5.5 Avian Receptors 31
5.6 Mammalian Receptors 31
5.7 Threatened and Endangered Species 31
5.8 Significant Habitats 31
6. UNCERTAINTY ANALYSIS 31
6.1 Approaches to Assessing Uncertainty 32
REFERENCES 33
Table Number 1: Assessment and Measurement Endpoints 33
Table Number 2: Trophic Levels. Exposure Pathways, and Food Sources 34
iii TAMS/Menzie-Cura
-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY FOR
PHASE 2 - ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
TABLE OF CONTENTS
BOOK 1 OF 1
Page
APPENDIX A MODELING APPROACHES
34
ADDITIONAL REFERENCES
35
III. COMMENTS ON THE ECOLOGICAL RISK ASSESSMENT
Federal (EN)
State (ED)
Local (EA)
Community Interaction Program (EP)
Public Interest Groups and Individuals (ES)
General Electric (EG)
SCOPE OF WORK
35
T AMS/Menzie-Cura
-------
Acronyms
Ah
Aryl Hydrocarbon
ARAR
Applicable or Relevant and Appropriate Requirement
AWQC
Ambient Water Quality Criteria
BAF
Bioaccumulation Factor
BSAF
Biota:Sediment Accumulation Factors
CBR
Critical Body Residue
CERCLA
Comprehensive Environmental Response, Compensation, and Liability Act
COE
Corps of Engineers
DEIR
Data Evaluation and Interpretation Report
DNAPL
Dense Non-Aqueous Phase Liquid
DQO
Data Quality Objectives
ERA
Ecological Risk Assessment
ERL
Effects Range-Low
ERM
Effects Range-Median
FDA
Food and Drug Administration
FFBAF
Foraging Fish Bioaccumulation Factor
FS
Feasibility Study
GE
General Electric
GM
Geometric Mean
GSD
Geometric Standard Deviation
HROC
Hudson River PCBs Oversight Committee
JLG
Joint Liaison Group
LOAEL
Lowest-Observed-Adverse-Effect-Level
NCP
National Oil and Hazardous Substances Pollution Contingency Plan
NPL
National Priorities List
NOAA
National Oceanic and Atmospheric Administration
NOAEL
No-Observed-Adverse-Effect-Level
NYSDEC
New York State Department of Environmental Conservation
NYSDOH
New York State Department of Health
NYSDOS
New York State Department of Sanitation
ORNL
Oak Ridge National Laboratories
PBAF
Pelagic Invertebrate Bioaccumulation Factor
PCB
Polychlorinated Biphenyl
PEL
Probable Effect Level
PFBAF
Piscivorous Fish Bioaccumulation Factor
TAMS/Menzie-Cura
-------
Acronyms
RI
Remedial Investigation
RI/FS
Remedial Investigation/Feasibility Study
ROD
Record of Decision
RM
River Mile
RPI
Rensselaer Polytechnic Institute
RR1/FS
Reassessment Remedial Investigation/Feasibility Study
SARA
Superfund Amendments and Reauthorization Act of 1986
SMDP
Scientific/Management Decision Point
SOW
Scope of Work
STC
Science and Technical Committee
TAGM
Technical and Administrative Guidance Memorandum
TCDD
2,3,7,8-Tetrachlorodibenzo-p-dioxin
TEF
Toxicity Equivalency Factor
TIP
Thompson Island Pool
TRV
Toxicity Reference Value
TSCA
Toxic Substances Control Act
USEPA
United States Environmental Protection Agency
USFWS
US Fish and Wildlife Service
WHO
World Health Organization
WQC
Water Quality Criteria
TAMS/Menzie-Cura
-------
Introduction
-------
HUDSON RIVER PCBs REASSESSMENT RI/FS
RESPONSIVENESS SUMMARY
PHASE 2 - ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
APRIL 1999
I. INTRODUCTION AND COMMENT DIRECTORY
1. Introduction
The U.S. Environmental Protection Agency (USEPA) has prepared this Responsiveness
Summary to address comments received during the public comment period on the Phase 2
Ecological Risk Assessment Scope of Work (ERASOW) for the Hudson River PCBs Reassessment
Remedial Investigation/Feasibility Study (Reassessment RI/FS), dated September 1998.
For the Hudson River PCBs Reassessment RI/FS, USEPA has established a Community
Interaction Program (CIP) to elicit on-going feedback through regular meetings and discussion and
to facilitate review of and comment upon work plans and reports prepared during all phases of the
Reassessment RI/FS.
Because of the large number of CIP participants and associated costs of reproduction, the
ERASOW is incorporated by reference and is not reproduced herein. No revised ERASOW will be
published. The comment responses and revisions noted herein are considered to amend the
ERASOW. For complete coverage, the ERASOW and this Responsiveness Summary must be used
together.
The first part of this three-part Responsiveness Summary is entitled, "Introduction and
Comment Directory." It describes the ERASOW review and commenting process, explains the
organization and format of comments and responses, and contains a comment directory.
The second part, entitled, "Responses to Comments on the Ecological Risk Assessment
Scope of Work," contains USEPA's responses to all significant comments. Responses are grouped
according to the section number of the ERASOW to which they refer. For example, responses to
comments on Section 2.2 of the ERASOW are found in Section 2.2 of the Responsiveness Summary.
Additional information about how to locate responses to comments is contained in the Comment
Directory.
1
TAMS/Menzie-Cura
-------
The third part, entitled, "Comments on Ecological Risk Assessment Scope of Work,"
contains copies of the comments submitted to USEPA. The comments are identified by commenter
and comment number, as further explained in the Comment Directory.
1.1 Recent Developments
USEPA received the revised Thomann-Farley model, which was developed for the Hudson
River Foundation, on April 27, 1999. USEPA will review the model to determine its
appropriateness for use in performing the Mid-Hudson ERA. In the ERASOW (p. 1), USEPA noted
that the Upper Hudson and Mid-Hudson ERAs may be developed at different times.
2. Commenting Process
This section documents and explains the commenting process and the organization of
comments and responses in this document. Readers interested in finding responses to their
comments may skip this section and go directly to the tab labeled "Comment Directory."
2.1 Distribution of ERASOW
The ERASOW, issued in September 1998, was distributed to federal and state agencies and
officials, participants in the CIP and General Electric Company (GE), as shown in Table 1.
Distribution was made to approximately 100 agencies, groups, and individuals. Copies of the
ERASOW were also made available for public review in 17 Information Repositories, as shown in
Table 2 and on the USEPA Region 2 internet webpage, entitled "Hudson River PCBs Superfund Site
Reassessment," at www.epa.gov/hudson.
2.2 Review Period and Public Availability Meetings
Review of and comment on the ERASOW occurred from September 23, 1998 to November
2, 1998. On September 23, USEPA held a Joint Liaison Group meeting open to the public at the
Holiday Inn at Latham, New York. Subsequently, on October 20, USEPA sponsored an availability
session at the Marriott Hotel in Albany, New York to answer questions from the public regarding
the ERASOW. These meetings were conducted in accordance with USEPA's Community Relations
in Superfund: Handbook, Interim Version (1988). Minutes of the Joint Liaison Group meeting are
available for public review at the Information Repositories listed in Table 2.
2
TAMS/Menzic-Cura
-------
TABLE 1
DISTRIBUTION OF REPORTS
HUDSON RIVER PCBs OVERSIGHT COMMITTEE MEMBERS
USEPA ERRD Deputy Division Director (Chair)
US EPA Project Managers
USEPA Community Relations Coordinator, Chair of the Steering Committee
NYSDEC Division of Hazardous Waste Management representative
NYSDEC Division of Construction Management representative
National Oceanic and Atmospheric Administration (NOAA) representative
Agency for Toxic Substances and Disease Registry (ATSDR) representative
US Army Corps of Engineers representative
New York State Thruway Authority (Department of Canals) representative
USDOI (US Fish and Wildlife Service) representative
New York State Department of Health representative
GE representative
Liaison Group Chairpeople
Scientific and Technical Committee representative
SCIENTIFIC AND TECHNICAL COMMITTEE MEMBERS
The members of the Science and Technical Committee (STC) are scientists and technical researchers
who provide technical input by evaluating the scientific data collected on the Reassessment RI/FS,
identifying additional sources of information and on-going research relevant to the Reassessment RI/FS,
and commenting on USEPA documents. Members of the STC are familiar with the site, PCBs, modeling,
toxicology, and other relevant disciplines.
- Dr. Daniel Abramowicz
- Dr. Donald Aulenbach
- Dr. James Bonner, Texas A&M University
- Dr. Richard Bopp, Rensselaer Polytechnic Institute
- Dr. Brian Bush, New York State Dept. of Health
Dr. Lenore Clesceri, Rensselaer Polytechnic Institute
- Mr. Kenneth Darmer
- Mr. John Davis, New York State Dept. of Law
- Dr. Robert Dexter, EVS Consultants, Inc.
- Dr. Kevin Farley, Manhattan College
3
TAMS/Menzie-Cura
-------
TABLE 1
DISTRIBUTION OF REPORTS(Cont-)
Dr. Jay Field, National Oceanic and Atmospheric Administration
Dr. Ken Pearsall, U.S. Geological Survey
Dr. John Herbich, Texas A&M University
Dr. Behrus Jahan-Parwar, SUNY - Albany
Dr. Nancy Kim, New York State Dept. of Health
Dr. William Nicholson, Mt. Sinai Medical Center
Dr. George Putman, SUNY - Albany
Dr. G-Yull Rhee, New York State Dept. of Health
Dr. Francis Reilly, Jr., The Reilly Group
Dr. John Sanders
Ms. Anne Secord, U.S. Fish and Wildlife Service
Dr. Ronald Sloan, New York State Dept. of Environmental Conservation
USEPA Community Relations Coordinator (Chair)
Governmental Liaison Group Chair and two Co-chairs
Citizen Liaison Group Chair and two Co-chairs
Agricultural Liaison Group Chair and two Co-chairs
Environmental Liaison Group Chair and two Co-chairs
USEPA Project Managers
NYSDEC Technical representative
NYSDEC Community Affairs representative
STEERING COMMITTEE MEMBERS
FEDERAL AND STATE REPRESENTATIVES
Copies of the Reports were sent to relevant federal and state representatives who have been involved with this
project. These include, in part, the following:
The Hon. Daniel P. Moynihan
The Hon. Charles E. Schumer
The Hon. John E. Sweeny
The Hon. Michael McNulty
The Hon. Sue Kelly
The Hon. Nita Lowey
The Hon. Maurice Hinchey
The Hon. Ronald B. Stafford
The Hon. Benjamin Gilman
The Hon. Richard Brodsky
The Hon. Bobby D'Andrea
17 INFORMATION REPOSITORIES {see Table 2).
4
TAMS/Menzie-Cura
-------
TABLE 2
INFORMATION REPOSITORIES
Adriance Memorial Library
93 Market Street
Poughkeepsie, NY 12601
Catskill Public Library
1 Franklin Street
Catskill. NY 12414
A Cornell Cooperative Extension
Sea Grant Office
74 John Street
Kingston, NY 12401
Crandall Library
City Park
Glens Falls, NY 12801
County Clerk's Office
Washington County Office Building
Upper Broadway
Fort Edward, NY 12828
* A Marist College Library
Marist College
290 North Road
Poughkeepsie, NY 12601
* New York State Library
CEC Empire State Plaza
Albany, NY 12230
New York State Department
of Environmental Conservation
Division of Hazardous Waste Remediation
50 Wolf Road, Room 212
Albany, NY 12233
* A R. G. Folsom Library
Rensselaer Polytechnic Institute
Troy, NY 12180-3590
Saratoga County EMC
50 West High Street
Ballston Spa, NY 12020
* Saratoga Springs Public Library
49 Henry Street
Saratoga Springs, NY 12866
* A SUNY at Albany Library
1400 Washington Avenue
Albany, NY 12222
* A Sojourner Truth Library
SUNY at New Paltz
New Paltz, NY 12561
Troy Public Library
100 Second Street
Troy, NY 12180
U.S. Environmental Protection
Agency
290 Broadway
New York, NY 10007
* A U.S. Military Academy Library
Building 757
West Point, NY 10996
White Plains Public Library
100 Martine Avenue
White Plains, NY 12601
Repositories with Database Report
CD-ROM (as of 10/98)
Repositories without Project
Documents Binder (as of 10/98)
5
lAMS/Mcnzic-Cura
-------
As stated in USEPA's letter transmitting the ERASOW, all citizens were urged to
participate in the Reassessment process and to join one of the Liaison Groups formed as part of
the CIP.
2.3 Receipt of Comments
Comments on the ERASOW were received in two ways: letters and oral statements made
at the September 23, 1998 Joint Liaison Group meeting. USEPA's responses to comments raised
at the Joint Liaison Group meeting are provided in the meeting minutes.
All significant comments received on the ERASOW are addressed in this Responsiveness
Summary. Comments were received from 6 commenters. Total comments numbered
approximately 80.
2.4 Distribution of Responsiveness Summary
This Responsiveness Summary will be distributed to the Liaison Group Chairs and Co-
Chairs and interested public officials. This Responsiveness Summary will be placed in the 17
Information Repositories and is part of the Administrative Record.
3. Organization of ERASOW Comments and Responsiveness Summary
3.1 Identification of Comments
Each submission commenting on the ERASOW was assigned the letter "E" for
ERASOW and one of the following letter codes:
N -
Federal agencies and officials;
D -
State agencies and officials;
A -
Local agencies and officials;
P -
Community Interaction Program;
S -
Public Interest Groups and Individuals; and
G -
GE.
The letter codes were assigned for the convenience of readers and to assist in the
organization of this document. Priority or special treatment was neither intended nor
given in the responses to comments.
Once a letter code was assigned, each submission was then assigned a number, in
the order that it was received and processed, such as EN1, EN2 and so on. Each different
comment within a submission was assigned a separate subnumber. Thus, if a federal
agency submission contained three different comments, they would be designated as
EN 1-1, EN 1-2, and EN 1-3. Written comment letters are reprinted following the fourth
tab of this document.
The alphanumeric code associated with each reprinted written submission is
marked at the top right corner of the first page of the comment letter. The subnumbers
6
TAMS/Mcnzic-Cura
-------
designating individual comments are marked in the margin. Comment submissions are
reprinted in numerical order by letter code in the following order: EN, ED, EA, EP, ES
and EG.
3.2 Location of Responses to Comments
The Comment Directory, following this text, contains a complete listing of all
commenters and comments. This directory allows readers to find responses to comments
and provides several items of information.
The first column lists the names of commenters. Comments are grouped first by:
EN (Federal), ED (State), EA (Local), EP (Community Interaction Program), ES
(Public Interest Group or Individual), or EG (GE).
The second column identifies the alphanumeric comment code (e.g., EF1-1)
assigned to each comment.
The third column identifies the location of the response by the ERASOW Section
number. For example, comments raised in Section 2.1 of the ERASOW can be
found in the corresponding Section 2.1 of the Responses, following the third tab
of this document.
The fourth, fifth, and sixth columns list key words that describe the subject matter
of each comment. Readers will find these key works helpful as a means to
identify subjects of interest and related comments.
Responses are grouped and consolidated by section number in order that all responses to
related comments appear together for the convenience of the reader interested in
responses to related or similar comments.
7
T AMS/Menzie-Cura
-------
4. COMMENT DIRECTORY
4.1 Guide to Comment Directory
This section contains a diagram illustrating how to find responses to comments. The
Comment Directory follows. As stated in the Introduction, this document does not reproduce the
ERASOW. Readers are urged to use this Responsiveness Summary in conjunction with the
ERASOW.
Step 1
Step 2
Step 3
Find the commenter or the key
words of interest in the
Comment Directory.
Obtain the alphanumeric
comment codes and the
corresponding ERASOW section.
Find the responses following the
Responses tab. Use the Table of
Contents to locate the page of the
Responsiveness Summary for the
ERASOW section.
Key to Comment Codes:
Comment codes are in the format EX-a
E=ERASOW
X=Commenter Group (N=Federal, D=State, A=Local, P=Community Interaction Program, S=Public
Interest Group, G=GE)
a=Numbered comment
Example:
Comment Response Assignment for the ERASOW
AGENCY/
Comment
REPORT
KEY WORDS
Name
CODE
SECTION
1
2
3
NOAA /Rosman EN-1 1.2 Tissue Samples
8
T AMS/Menzie-Cura
-------
Comment Directory
-------
4.2 Comment Directory
AGENCY/
COMMENT
REPORT
KEYWORDS
NAME
CODE
SECTION
1
2
3
PCB Tissue
NOAA/Rosman
EN-1
1.2
Tissue Samples
Analysis
Dataset
Lower Hudson
NOAA/Rosman
EN-2
2,1.3
Endangered Species
Threatened Species
River
NOAA/Rosman
EN-3
2.5
Soil Clean up Levels
TAGMS
Sediments
Benthic
NOAA/Rosman
EN-4
4.2.1
Communities
Metrics
Taxa
NOAA/Rosman
EN-5
2.6.1
PCB Concentrations
Surface Water
AWQC
Mammalian
Exposure
NOAA/Rosman
FN-6
3.1.6
Receptors
Floodplain
Assessment
Exposure
Exposure Point
NOAA/Rosman
EN-7
3.3
Assessment
Body Burdens
Concentrations
NOAA/Rosman
EN-8
3.4
Ingestion Pathways
Dose Calculation
Surface Water
Benthic
NOAA/Rosman
EN-9
4.2.1
Communities
Metrics
-
Tissue
NOAA/Rosman
EN-10
4.1
TEQ
TEF
Concentrations
Measurement
Reproductive
NOAA/Rosman
EN-11
4.2.2
Effects
Effects
Fish
Benthic
NOAA/Rosman
EN-12
5.3
Communities
Uncertainty
Effects association
Population Level
Risk
NOAA/Rosman
EN-13
5.4
Effects
Characterization
Uncertainty
NOAA/Rosman
EN-14
References
Add References
-
-
Assessment
NOAA/Rosman
EN-15
Table I
Sediment Guidelines
Endpoints
Significant Habitats
NOAA/Rosman
EN-16
Table 2
Shortnose Sturgeon
Invertebrates
.
Exposure
NOAA/Rosman
EN-17
Appendix A
Models
Assessmnet
-
NOAA/Rosman
EN-18
Appendix A
Congeners
-
-
NOAA/Rosman
EN-19
Appendix A
River Segments
EPC Derivation
Sediments
NYSDEC/Ports
ED-1
1,1
Fort Edward Dam
Dam Removal
-
Dietary
NYSDKC/'Ports
ED-2
1.3
PCB Concentrations
Concentrations
-
NYSDEC/Ports
HD-3
2.1.1
Whorled Pogonia
-
-
Contaminants of
Ecological
NYSDEC/Ports
ED-4
2.2
Concern
PCBs
Assessment
Measurement
NYSDEC Ports
ED-5
2.5
Endpoints
PCB Body Burdens
Tree Swallow
NYSDEC/Ports
ED-6
2.3
Shortnose Sturgeon
Piscivorous Species
Omnivorous Species
NYSDEC/Ports
ED-7
2.6.4
Mink
New York State
Literature Cited
9
TAMS/Mcnzic-Cura
-------
NYSDEC/Ports
ED-8
3.3.5
Bald Eagles
NYSDEC Research
-
NYSDEC/Ports
ED-9
3
Linear Approach
Receptors
Peripheral habitats
NYSDEC/Ports
ED-10
6.1
Uncertainty
Clarity
-
SCEMC
EA-1
1.5
ORNL
Synopsis
Relevant Points
SCEMC
EA-2
1.6
NRDA Plan
Quantitative
Receptors
SCEMC
EA-3
2.1.3
PCB Sources
Upper Hudson
Lower Hudson
Benthic
SCEMC
EA-4
2.5
Disease
Deformaties
Communities
SCEMC
EA-5
2.6.5
Bald Eagle
Threatened
Endangered
SCEMC
EA-6
3.2
Striped Bass
Sturgeon
Lower Hudson
SCEMC
EA-7
3.2
Time varying model
_
SCEMC
EA-8
3.3.3
GE data
-
SCEMC
EA-9
3.4
Dose formula
PCB intake
Factor
SCEMC
EA-10
4.1
Level of Effort
Hudson River
Impacts
SCEMC
EA-11
4.1
Congener
Bzmi
-
SCEMC
EA-12
5.3
Qualitative
-
-
SCEMC
EA-13
5.4
Effects assessment
Extrapolation
-
SCEMC
EA-14
6.1
Exposure
Upper Bound
Lower Bound
Seasonal Time
SCEMC
EA-15
Appendix A
Food Chain Model
Bioenergetic Model
Scales
SUNY/Putman
EP-1
4.2.1
Exposure Levels
TRV
Uncertainty
Measurement
SUNY/Putman
EP-2
4.1.1
Chronic Exposure
Effects
Receptors
Exposure
SUNY/Putman
EP-3
General Comments
Concentrations
Great Lakes
St. Lawrence River
Scenic Hudson
ES-1
4.1
PCB Toxicity
Receptor Species
-
Scenic Hudson
ES-2
General Comments
Mink
Population Decline
Diversity
Scenic Hudson
ES-3
2.6
Avian Receptors
Knowledge
References
Scenic Hudson
ES-4
2.7
Impacts
Risk
NRDA
Scenic Hudson
ES-5
4.1
Endpoints
Reproduction
Development
Scenic Hudson
ES-6
4.1.1
PCB Toxicity
NOAF.L
Threshold
Toxic Equivalency
Scenic Hudson
ES-7
4.1
Factors
BZ#126
Congener
Measurement
Scenic Hudson
ES-8
4 2
Effects
Toxicological Data
Sources
Assessment
Measurement
Benthic
Scenic Hudson
ES-9
Table 1
Endpoint
Endpoint
Communities
GE
EG-1
1
Baseline ERA
Toxicity quotient
Uncertainty
Site conceptual
GE
EG-2
6.1
model
TRVs
Relevant endpoints
GE
EG-3
2.3 and 6.1
Endpoints
Translating
Population-level risk
GE
EG-4
2.2
Other stressors
PCBs
-
GE
EG-5
2
Objectives
Endpoints
-
GE
EG-6
2.3
Endpoints
Criteria
-
Measurement
Assessment
GE
EG-7
2.5
endpoints
endpoints
-
10
TAMS/Mcn/ic-Cura
-------
GE
EG-8
3.3
Site characterization
Definition of the
Receptors
Lower Hudson
-
GE
EG-9
2.1
"site"
Contaminants of
River
Herbicides,
GE
EG-10
2.2
Concern
Site Conceptual
Metals exposure
Pesticides
GE
EG-11
2.4
Model
Exposure pathways
Aquatic vegetation
GE
EG-12
2.6
Receptors
-
-
GE
EG-13
3.1
Exposure Pathways
-
-
GF.
EG-14
3.2
Fate & Transport
Exposure
Models
-
GE
EG-15
3.3
concentration
-
-
GE
EG-16
3.4
Models
-
-
GE
EG-17
4.1
Toxicity
Risk
TEF
TEQ
GE
EG-18
5.4
characterization
Models
TRV
GE
EG-19
6
Uncertainty
Sensitivity analysis
-
GE
EG-20
6.1
Uncertainty
Sensitivity analysis
-
TAMS/Miw it!-Cur;i
-------
THIS PAGE LEFT BLANK INTENTIONALLY
TAMS/Menzic-C'ura
-------
Responses
-------
II.
RESPONSE TO COMMENTS ON THE ERASOW
General Comments
Response to EP-3
Consistent with USEPA's risk assessment guidance and the NCP (§300.430(d)(4)). the
objective of the ERA is to assess risk on a site-specific basis. The ERA will not compare the Hudson
River PCBs site to other sites contaminated with PCBs.
Response to EG-1
Consistent with USEPA's risk guidance, the ERASOW begins with the contaminant of
concern and works up to assess risk to individuals, populations, and communities (a "bottom-up"
approach), rather than starting with field population and community information and working down
to identify the contaminant of concern (''top-down" approach), as proposed by the commenter. The
bottom-up approach is then combined with the independent results of a probabilistic analysis in a
weight-of-evidence approach to determine whether concentrations of PCBs present in the Hudson
River may cause adverse effects in individuals and populations of representative receptors. A
weight-of-evidence approach is preferred for the Hudson River PCBs site because examination of
several lines of evidence increases confidence in the results of the study of this complex Superfund
site.
Additional information regarding the contaminants of concern is provided in the responses
in Section II. 2.2, Contaminants of Concern. Additional information regarding risks to biological
communities and populations is provided in the responses in Section II. 2.3, Assessment Endpoints.
Additional information regarding uncertainty in the ERA is provided in the responses in Section II
6.1, Approaches to Assessing Uncertainty.
Response to ES-2
Qualitative information on the "health" of the Upper Hudson River ecosystem, limited to
any observations made during the ecological field sampling, will be presented in the problem
formulation (Step 3) discussion in the ERA.
1. INTRODUCTION
1.1 Site History
Response to ED-1
The comment on the relative importance of the removal of the Fort Edward Dam is
acknowledged and will be incorporated into future reports, where applicable.
13
TAMS'Menzie-Cura
-------
1.2 Ecological Risk Assessment in the Superfund Process
Response to EN-1
All relevant data received by USEPA during preparation of the ERA will be considered for
incorporation into the ERA, as time permits. Data received after release of the ERA will be
reviewed to determine whether they are directly applicable to the ERA, and if so, they will be
considered in any additional analyses or responses to comments.
1.3 Results of Phase 1 Ecological Risk Assessment
Response to ED-2
The comment on editing for clarity is acknowledged and will be incorporated into future
reports, where applicable.
1.4 Changes in EPA Risk Assessment Guidance Since the Phase 1 Assessment
No significant comments were received on Section 1.4.
1.5 Additional Toxicological Benchmarks Developed by ORNL Since the Phase 1
Assessment
Response to EA-1
USEPA will identify and explain the use of any Oak Ridge National Laboratories (ORNL)
report used in the ERA. The ORNL reports summarize scientific literature and may be used as a
starting point for a literature search on toxicological benchmarks.
1.6 Organization of the Phase 2 ERA Based on USEPA 1997 Guidance
Response to EA-2
The draft scope for the Hudson River Natural Resource Damage Assessment (NRDA) Plan
issued in 1998 correctly stated that much of the information to quantify injury to various receptors
is likely not available; however, it also states that the necessary information will be developed.
Similarly, the information necessary to assess risk in the ERA will be developed by USEPA based
on available water, sediment, aquatic invertebrate, and fish data.
Limited mink and otter data from the New York State Toxic Substances Monitoring Program
(1982) and Foley et al. (1988) will be provided in the ERA. The mink's exposure to PCBs will be
estimated using a food chain model that calculates PCB uptake from its food source (including
forage fish), from direct contact with PCB-contaminated sediments, and from ingestion of PCB-
14
TAMS/Menzie-Cura
-------
contaminated surface water. Heavy metals data from the Upper Hudson River (Phase 2 1993
Ecological Sampling Database) will be discussed in the benthic macroinvertebrate community
analysis section.
Available data are sufficient to perform a quantitative and qualitative assessment of
ecological risk posed by the Hudson River PCBs site.
2. PROBLEM FORMULATION
Response to EG-5
The commenter recommends focusing the ERA on three questions related to the adverse
effects of PCBs on biological community structure and population dynamics. The first question is
whether PCBs from the site currently are adversely affecting biological community structure or the
population dynamics of the key receptors. As explained in the response to EG-3. biological
community structure and population dynamics of vertebrate receptors are not emphasized in the
"bottom-up" approach of the Superfund baseline risk assessment. Instead, consistent with IJSEPA
guidance, assessment endpoints were selected to identify risk to receptors of concern that could be
adversely affected by contaminants from the site (ERASOW. p. 13).
The second question asks when will PCBs from the site no longer adversely affect biological
community structure or population dynamics under a "no action" remedy. Again, the bottom-up
approach of the Superfund risk assessment emphasizes risk to individual receptors of concern, rather
than on community structure or population dynamics. The ERA will model future fish body burdens
to determine if predicted concentrations of PCBs will adversely affect biological receptors over a
25-year time frame, beginning at the time of data collection (i.e., 1993 to 2018). Both mean and
95% upper confidence limit (UCL) PCB concentrations will be calculated, absent any remediation.
Because many receptors at various trophic levels are being evaluated, an estimate of when the PCBs
would no longer adversely affect the biological community is not an appropriate assessment
endpoint.
The third question is to what degree will remediation reduce the time to reach the point at
which PCBs are no longer adversely affecting biological community structure or population
dynamics. As stated above, the ERA focuses on key biological receptors rather than an entire
biological community or population. Moreover, evaluation of remedial alternatives that address risk
is part of risk management, which is the step after completion of the baseline risk assessment
(ERASOW. p. 9). Accordingly, this question is outside of the scope of the ERA.
15
TAMS/Menzie-Cura
-------
2.1 Site Characterization
Response to EG-9
USEPA has consistently defined the site to include the Lower Hudson River since at least
April 1984, when the Agency issued its FS for the site and before the site was listed on the National
Priorities List (codified at 40 CFR Part 300, App. B). In its September 25, 1984 Record of
Decision, USEPA defines the site by reference to three figures which, together, depict the site as the
entire 200-mile stretch of the River from Hudson Falls to the Battery in New York City, plus the
remnant deposits. In addition, during the Reassessment RI/FS, USEPA has consistently defined the
site as including the Upper and Lower River (e.g., the Scope of Work for Hudson River
Reassessment RI/FS (December 1990) and the Phase 1 Report for the Reassessment RI/FS (August
1991)). The comment regarding USEPA's use of the results of the ERA (including the Lower
Hudson ERA) in evaluating remedial alternatives is a risk management issue, and therefore beyond
the scope of the ERA. USEPA decision-makers will consider risk management following
completion of the ERA.
2.1.1 Upper Hudson River
Response to ED-3
The small whorled pogonia (Isotria medeloides), a member of the orchid family, is listed as
a federal threatened species. It is listed in New York State as a historic species, last seen in 1976.
However, because it is generally found in dry soils in mid-aged woodlands, it will not be discussed
in the ERA.
2.1.2 Thompson Island Pool
No significant comments were received on Section 2.1.2.
2.1.3 Lower Hudson River
Response to EN-2
As appropriate, endangered or threatened species in the Lower Hudson River will be
discussed in the ERA.
16
TAMS/Menzie-Cura
-------
Response to EA-3
The ERA for the Lower Hudson River is appropriate because, as noted in the ERASOW (p.
12), the Lower Hudson River has several ecologically sensitive areas that give it ecological
importance. Sufficient information exists to perform the ERA for the Lower Hudson River.
2.2 Contaminants of Concern
Response to EG-4. EG-10. and ED-4
The ERA is part of USEPA's reassessment of its 1984 no-action decision with respect to
PCB contaminated sediments in the Upper Hudson River (ERASOW, p. 12). As such, the
ERASOW focuses on PCBs in the Upper Hudson River sediments. Step 1, which includes screening
for preliminary contaminants of concern (COCs), was performed in the Phase I Report Review Copy
Interim Characterization and Evaluation Hudson River PCB Reassessment Rl/FS (August 1991) and
will be summarized in the ERA (see ERASOW, p. 7). PCBs were identified as the primary COCs.
given the quantity released, their persistence in the environment, their toxicity, and their known
bioaccumulative effects.
2.3 Assessment Endpoints
Response to EG-3 and EG-6
USEPA agrees that one of the assessment endpoints of the ERA is sustainability (i.e.,
survival and reproduction) of local Hudson River populations (ERASOW, p. 14). The ERA will
estimate the potential for risk under future conditions based on modeling results (Baseline Modeling
Report, due May 1999). Potential risk will be based on the probability that future concentrations will
result in body burdens or doses that exceed the selected toxicity reference values (TRVs) for each
receptor population.
Consistent with USEPA risk guidance (USEPA. 1997), the assessment endpoints identified
in the ERASOW are primarily individual-level risks, such as survival or growth, rather than the
sustainability of communities or populations. Direct measurement of population-level endpoints is
problematic because it is difficult to identify a discrete population and to determine the proportion
of individuals in a population that constitutes a population-level change, which could lead to an
underestimation of ecological risk. While there is uncertainty associated with extrapolating from
an individual-level risk to a community or population, such an approach is protective of the
environment.
USEPA agrees that it must select appropriate TRVs as measures of effect and that the
appropriate criteria are those provided in USEPA guidance. The TRVs that will be used are based
on peer-reviewed scientific studies. The uncertainty associated with them does not compromise the
17
TAMS/Menzie-Cura
-------
integrity of the ERA. It is unrealistic to expect the scientific literature to provide species- and
habitat-specific TRVs for all endpoints examined. Surrogate species are routinely used in ecological
risk assessments to calculate risks to the biological community. It should be noted that even TRVs
derived for specific areas and species have uncertainty and variability associated with them.
Response to ED-6
The shortnose sturgeon and brown bullhead are omnivorous, as indicated in the text
(ERASOW, p. 18-19). Table 1 of the ERASOW should have included a separate assessment
endpoint for omnivorous fish, as follows.
Assessment Endpoint
Survival, growth, and
reproduction of local
omnivorous fish populations
Specific Ecological Receptor
"Endpoint Species'"
brown bullhead
shortnose sturgeon
Measures
Exposure
Food Chain Modeling
PCB Cone, in Prey
Body Burdens
PCB in Water/Scd
Effect
Exceed of TRV
Exceed of Pop
Effects
Exceeds WQC
2.4 Site Conceptual Model
Response to EG-11
Consistent with USEPA guidance (USEPA, 1997). the site conceptual model in the
ERASOW (p. 14) identified the sources, media, pathways, and routes of exposure that will be
evaluated in the ERA. Environmental fate processes, such as burial and dechlorination, are not
included in the conceptual model, but will be presented in the Baseline Modeling Report (due May
1999), which will be used to estimate future concentrations of PCBs. The presence of other
contaminants and stressors will be considered during risk management, as appropriate (see response
to comments EG-4, EG-10, and ED-4). The diagram of the conceptual model (Figure 5 of the
ERASOW) could have shown aquatic vegetation as the food source of some lower trophic level
receptors (e.g., herbivores). However, aquatic vegetation was not included because it generally does
not bioaccumulate PCBs and therefore does not contribute a significant amount of PCBs to benthic
invertebrates.
2.5 Measurement Endpoints
Response to EG-7
Consistent with USEPA guidance (USEPA, 1997), the ERASOW identified individual-level
measurement endpoints, rather than population or community-level endpoints, for individual-level
assessment endpoints. As with the assessment endpoints. direct measurement of population-level
measurement endpoints is problematic because it is difficult to identify a discrete population and to
determine the proportion of individuals in a population that constitutes a population-level change.
18
TAMS.'Menzie-Cura
-------
which could lead to an underestimation of ecological risk The measurement endpoints will be used
to determine the potential for community or population-level effects based on individual risk. While
there is uncertainty associated with extrapolating from an individual-level risk to a community or
population, such an approach is protective of the environment.
The measurement endpoints identified in the ERASOW include ambient water quality criteria
AWQC), sediment quality values, and TRVs. A comparison of site data to these values is being
conducted as part of the ERA to update the comparison done earlier in the Phase 1 Report (USEPA,
1991). This is consistent with USEPA guidance, which states, "measurement endpoints can include
measures of exposure as well as measures of effect" (USEPA. 1997).
The TRVs, as well as other measurement endpoints, will be selected based on the criteria set
forth in the ERASOW (p. 15), which are strength of the association between the measurement
endpoint and the assessment endpoint, data quality, and study design and execution. These three
broad criteria encompass the eight criteria identified by the commenter.
USEPA acknowledges the commenters agreement with USEPA's use of a weight-of-
evidence approach. The quality of each measurement endpoint will be evaluated according to the
attributes identified by Menzie et al. (1996) and will be discussed in ERA. USEPA notes that Dr.
Menzie will be directly involved for the Hudson River PCBs Reassessment ERA.
Response to ED-5
Measured PCB body burdens are available for benthic invertebrates, forage fish, piscivorous
fish, and insectivorous birds, as listed in Table 1 of the ERASOW. PCB concentrations in receptors
that have no measured body burdens will be modeled based on food chain exposure models and
concentrations of PCBs measured in the Hudson River, and/or biomagnification factors from the
scientific literature (for example, for concentrations in eggs of piscivorous birds). USEPA will
contact agencies and organizations with information on fish and wildlife populations along the
Hudson to ensure that all receptor species feed in, at, or near the river. The potential receptor species
listed in the ERASOW were selected to be representative of different behaviors and feeding
strategies.
Response to EN-3
USEPA agrees that the correct citation for the New York State Department of Environmental
Conservation (NYSDEC) TAGM for screening contaminated sediments is NYSDEC (1998).
Response to EA-4
Disease and deformities observed during sampling would not necessarily be attributed to
PCBs. Rather, the observations would be included in a weight-of-evidence approach, which would
19
TAMS/Menzie-Cura
-------
consider the location of observations, the number of observations, the measured PCB concentrations
in the area, and other measurements.
2.6 Receptors of Concern
Response to EG-12
USEPA selected the largemouth bass, but not the smallmouth bass, as a potential fish
receptor (ERASOW, p. 19). The smallmouth bass was erroneously included in Table 1. As
discussed in the response to ED-3, the whorled pogonia will not be included in the ERA because its
habitat is not found along the Hudson River. The northern harrier is a State species of special
concern that is a potential receptor to be considered in the ERA because it feeds and nests in marshes
or wetlands that could receive PCBs during flood events. The striped bass and shortnose sturgeon
occur predominantly in the Lower Hudson River; however, shortnose sturgeon have been sighted
in the Upper Hudson River (Bain, personal communication). The striped bass was selected to
evaluate the ecological risk posed by contaminated sediment, water and fish in the Lower Hudson
(see also response to EA-6) and the shortnose sturgeon was selected because it is a federally-listed
endangered species.
Response to ES-3
Please provide the references so they can be evaluated for use in the ERA.
2.6.1 Macroinvertebrate Communities
Response to EN-5
USEPA agrees with the comment. As stated in the ERASOW (p. 44), the PCB
concentrations in Hudson River surface water will be compared to freshwater or marine AWQC,
depending on the salinity of the water. The Upper Hudson River is exclusively freshwater and only
the last three Lower Hudson River locations sampled in the Phase 2 ecological sampling program
(RM 58.7 and lower) are saline and can be compared to saltwater criteria.
2.6.2 Fish Receptors
2.6.3 Avian Receptors
No significant comments were received on Sections 2 6.2 and 2.6.3.
20
TAMS/Menzie-Cura
-------
2.6.4 Mammalian Receptors
Response to F.D-7
USEPA agrees with the comment. The citation for Foley et al. (1988) will be evaluated for
use in the ERA. The Palmer and Fowler (1975) reference will be included in the references section
of the ERA.
2.6.5 Threatened and Endangered Species
Response to F.A-5
The bald eagle is a federally-listed threatened and a New York State-listed endangered
species, and therefore is appropriately mentioned in Section 2.6.5. Threatened and Endangered
Species.
2.6.6 Significant Habitats
No significant comments were received on Section 2.6.6..
2.7 Risk Questions
Response to ES-4
USEPA will use the term '"adversely affecting"in the ERA rather than "impacting" to
distinguish it from a natural resources damage assessment
3. EXPOSURE ASSESSMENT
Response to ED-9
Consistent with USEPA guidance (USEPA, 1997), the ecological risk posed by the site will
be assessed both quantitatively and qualitatively. For the quantitative portion of the ERA, USEPA
first identified species that are most directly in contact with the PCBs in the Hudson River and were
therefore assumed to have the greatest potential for ecological risk. From this list of species, USEPA
identified receptors that occupy different positions within the food web and that have a variety of
behavioral and feeding strategies, such as the mink, the brown bat, and the great blue heron. The
overall risk to the ecosystem will be assessed by considering the quantitative risks to the various
individual receptor species and by measuring the benthic community structure, combined with a
discussion of the risk along exposure pathways that were not quantitatively assessed. Exposure via
pathways outside of the Hudson River, such as direct contact with PCBs in river bank and floodplain
sediments, will be mentioned as additional pathways of exposure.
21
T A M S/Menzie-Cura
-------
3.1 Exposure Pathways
Response to EG-13
Consistent with USEPA guidance (USEPA, 1997), the diets of receptor species will be
determined using the USEPA Wildlife Exposure Factors Handbook (USEPA, 1993) and, as
appropriate, site-specific information from the scientific literature. USEPA has copies of the
references cited by the commenter (Robinson, 1992; McCarty. 1995; Secord and McCarty, 1997;
Exponent. 1998; and Secord, 1998).
3.1.1 PCBs in Sediments
3.1.2 PCBs in Water
3.1.3 Benthic Invertebrates
3.1.4 Fish Receptors
3.1.5 Avian Receptor
No significant comments were received on Sections 3.1.1 to 3.1.5.
3.1.6 Mammalian Receptors
Response to EN-6
USEPA selected the mink, rather than the shrew or meadow vole, as one of the mammalian
receptors of concern. Although shrews, voles, and mink may all be exposed to PCBs from direct
contact with sediments on the river banks and floodplains, the mink was selected as a receptor of
concern because of its documented sensitivity to PCBs and its reliance on fish as a food source
(ERASOW, p. 20). USEPA selected the raccoon because it is a mammal with a different feeding
strategy (i.e., omnivore) than the mink. (With respect to avian receptors, the bald eagle was selected
as an avian receptor of concern, rather than other accipiters (hawks), because it is on both the federal
and New York State lists of threatened and endangered species and there have been recent sightings
of it along the Hudson River (ERASOW, p. 21)). The receptors of concern are representative of
different species that may be adversely affected by the site and is not intended to be all-inclusive.
3.2 Quantification of PCB Fate and Transport
Response to EG-14
USEPA will use fate and transport models to describe the distribution of PCBs in the
sediments and water of the Hudson River. These modeled concentrations of PCBs in sediments and
22
TAMS'Menzie-Cura
-------
water will be used in the ERA as initial concentrations in several bioaccumulation models. One of
the bioaccumulation models is a Gobas-type time-variable mechanistic model. The ERA will be
peer reviewed and USEPA will respond to the peer reviewers' comments in a responsiveness
summary.
Response to EA-6
The risks to striped bass and shortnose sturgeon will be assessed, though not explicitly
modeled, using sediment, water and fish data that are representative of their respective habitats,
including spawning locations and winter holdover locations (ERASOW, p. 28). USEPA intends to
assess the risk to the striped bass using the revised Thomann/Farley model.
RespQnse tQ EA-7
Consistent with USEPA guidance (USEPA, 1997), the ERA will assess future risk (1993 to
2018), assuming no remediation, by using modeled exposure concentrations. In modeling the future
concentrations at various time intervals for mammalian and avian receptors, the initial concentration
may be held constant for the first five or ten years, even if the model suggests a time-varying
decrease in concentration (ERASOW, p. 29), or may be recalculated annually, as suggested by the
commenter. Holding the concentrations constant is a conservative approach that may lead to an
overestimation of risk, but it is protective of the environment given the expected variations in
concentrations throughout the site. A sensitivity analysis performed for the Monte Carlo method will
provide some indication of the degree to which the risk may be overestimated, and the uncertainty
associated with using this approach would be discussed in the ERA.
3.3 Observed Exposure Concentration
Response to EN-7 and EG-15
USEPA will use appropriate statistics from the observed data to characterize exposures and
body burdens, including arithmetic averages and 95 percent upper confidence limits (UCLs) on these
averages (ERASOW, p. 30). Because the observed concentrations are best described by lognormal
distributions (Baseline Modeling Report, due May 1999), the formula to estimate 95% UCL for
lognormal distributions, also known as Land's method, will be used (Gilbert, 1987). These statistics
are appropriate for use in the ERA because the receptors of concern are expected to contact the
sediments and water column over a large area rather than remain in one localized spot. Appropriate
statistics may include a mean exposure level with appropriate quantification of uncertainty, where
the data are sufficient to be characterized as a distribution described by a mean and a standard
deviation (ERASOW. p. 30).
The field measurements of benthic invertebrate BSAFs will be compared to laboratory and
other field measurements, such as sediment data, in assessing the risk to benthic invertebrates. For
modeled exposure concentrations of PCBs in benthic invertebrates, the concentrations will be
23
TAMS'Menzie-Cura
-------
estimated from a distribution of the site-specific BSAFs derived in the Baseline Modeling Report
(due May 1999) (ERASOW, p. 34). The site-specific fish body burdens will be used to calibrate and
validate the models (Baseline Modeling Report, due May 1999). Insectivorous birds and mammals
living along the Hudson River are assumed to feed mainly on flying insects with partial aquatic life
histories based upon Hudson River and New York State studies (e.g., Buchler, 1976; Robertson et
al., 1992; Secord and McCarty, 1997). PCB concentrations in flying aquatic insect prey will be
based on the results of the field sampling program (i.e., macroinvertebrate and sediment data) with
a factor to account for metamorphic partitioning between shed exuviae and the emerging flying
insect.
Response to EG-8
The ERA will evaluate all relevant data sets that have been published or otherwise made
available to USEPA, where appropriate. Studies cited by the commenter will be reviewed, including
NYSDEC (1977), NMFS and USFWS (1985), ASMFC (1990), NYSDEC (1993), Kynard (1997),
NMFS (1997), Secord and McCarty (1997), Exponent (1998), and USFWS (1998). Mr. Nye of
NYSDEC will be contacted.
While USEPA will evaluate the data in the studies cited by the commenter, it does not agree
with all of the commenter's interpretations of those data with respect to benthic and phytophilous
macroinvertebrates, fish, and birds at the site. For example, studies on fish abundance do not
necessarily reflect the response of fish to PCBs as factors, because the fishing advisories may have
a greater overall effect on abundance. In addition, the commenter stated that USEPA should
eliminate the tree swallow as a receptor species due to the Secord and McCarty (1997) study,
claiming that the study "show(s) that even high body burdens of PCBs have not affected the
reproductive success of tree swallows." In fact, Secord and McCarty (1997) stated, "...tree swallows
breeding along the Hudson River had lower reproductive success than tree swallows from an
uncontaminated site" and "PCB concentrations and toxic equivalency quotients detected in tree
swallows have significant implications for migratory birds that breed or migrate along the Hudson
River." Therefore, it is appropriate that the tree swallow be retained as a receptor species.
3.3.1 Sediment Concentrations
3.3.2 Water Column Concentrations
No significant comments were received on Sections 3.3.1 and 3.3.2
3.3.3 Benthic Invertebrate Concentrations
Response to EA-8
GE has collected data on PCB concentrations in various media in the Hudson River, and
GE's data will be used to supplement USEPA's dataset, as appropriate (ERASOW, p. 30). GE's
24
TAMS/'Menzie-Cura
-------
sediment data will be evaluated for appropriateness to supplement USEPA's high-resolution and
low-resolution sediment sampling programs (ERASOW, p. 30). Data were collected in the 1993
Phase 2 field sampling program specifically for the ERA, and therefore these data will receive
priority in the ERA.
3.3.4 Fish Concentrations
No significant comments were received on Section 3.3.4.
3.3.5 Avian Concentrations
Response to ED-8
USEPA acknowledges the commenteds offer to provide additional data on bald eagles as it
becomes available and may incorporate the data into the ERA if time permits.
3.3.6 Mammalian Concentrations
No significant comments were received on Section 3.3.6.
3.4 Modeled Exposure Concentrations
Response to EG-16 and EA-9
In the ERA, USEPA will use modeled exposure concentrations in addition to observed
exposure concentrations to assess ecological risk. USEPA will use all appropriate data to calibrate
and validate both the probabilistic bioaccumulation food chain model and modified Gobas time-
varying mechanistic models (FISHPATH and F1SHRAND) for PCB concentrations in fish in the
Baseline Modeling Report (due May 1999). Even with such validation, there will be uncertainty
which, consistent with USEPA's risk guidance (USEPA, 1997), will be quantified to the extent
possible. There are insufficient avian and mammalian Hudson River PCB data to field-validate
vertebrate exposure models, and field collection of some vertebrate species could adversely affect
Hudson River populations (e.g., bald eagle, river otter).
The ERASOW (p. 32-33) presents general forms of the equations to be used for direct
ingestion of water and for dietary doses from ingestion of prey (i.e., food). For the dietary dose
equation, USEPA does not agree that the ingestion rate term is per day. As stated in the ERASOW
(p. 33). the dose may be expressed as either a critical body residue or as an average daily exposed
dose. USEPA does agree that the concentration of PCBs in food items is on a wet weight basis
The ERASOW does not detail how the ingested dose will be translated into body burdens.
The specific equations to describe the conversions require additional terms, such as assimilation and
metabolic efficiencies, and depend on the physiology of the given species and, to some extent, on
25
TAMSMenzie-Cura
-------
the structure of the PCB congener being ingested (ERASOW, p, 33). This level of detail is beyond
the scope of the ERASOW. However, where appropriate, USEPA will describe in the ERA the
method used to convert ingested dose to absorbed doses (i.e., body burdens) for each species and
each congener-based toxicity equivalency factors (TEFs). These methods may range from assuming
that 100% of ingested dose is available to the species, to using exposure models such as the
bioaccumulation model for fish. A literature-based biomagnification factor (USEPA, 1994) will
be used to predict concentration of PCBs in eggs of pisciverous birds.
As stated in the ERASOW (p. 33), the input values for these equations will be obtained from
the scientific literature. For the Hudson River ERA, the relevant literature includes site-specific
studies. USEPA's Wildlife Exposure Factors Handbook (USEPA, 1993), and other references. The
studies cited (Salyer and Langler, 1949; Davis, 1980, Landum et al 1993; and Kaufman, 1996) will
be evaluated for use in the ERA.
USEPA disagrees with the comment that it would be more appropriate to obtain exposure
point concentrations for fish by multiplying an average BSAF by an average surface sediment PCB
concentration. Rather, USEPA will obtain exposure point concentrations for fish by compiling PCB
concentrations by location (ERASOW, p. 31). The Upper Hudson River will be divided into three
reaches (RM 189, 168, and 154). The pooling of data from Upper Hudson River sampling locations
is done to account for forage fish obtaining benthic invertebrates from a large area rather than from
one isolated location. The locations in the Lower Hudson River are considered to be separate
habitats and therefore fish PCB concentrations are not combined.
Response to EN-8
Because limited data are available for nearshore water concentrations, whole water average
and 95% UCL concentrations will be used for all receptors. For species whose predominant habitat
is the shoreline, the 95% UCL may provide a more appropriate concentration than the average, as
concentrations may be higher along the shoreline.
3.4.1 Benthic Invertebrate Receptors
3.4.2 Fish Receptors
3.4.3 Avian Receptors
3.4.4 Mammalian Receptors
No significant comments were received on Sections 3.4.1 to 3.4.4.
26
TAMS'Menzie-Cura
-------
4. EFFECTS ASSESSMENT
4.1 Estimating the Toxicity of PCBs
Response to ES-1 and ES-5
For each receptor species, a toxicity reference value (TRV) will be selected based on studies
that examine the effects of PCBs on survival (lethality), growth, or reproduction (ERASOW, p. 37).
Reproductive effects are broadly defined to include larval/fetal development. Studies that examine
the effects of PCBs on other sublethal endpoints will be presented, but not used to select TRVs.
TRVs for fish will be expressed as critical body residues in adult fish and fry (e.g., nig/kg whole
body weight) and as lipid-normalized concentrations in eggs (e.g., mg/kg lipid in eggs). TRVs for
avian and mammalian receptors will be expressed as daily doses (e.g., mg/kg whole body weight).
TRVs for birds will also be expressed as concentrations in eggs (e.g., mg/kg wet wt egg). No
additional toxicity studies are planned. A summary of published studies in the scientific literature
will be provided in the ERA.
Response to EA-10
US EPA did not conduct site-specific toxicological studies at different locations along the
entire length of Hudson River because it would have required numerous toxicological studies
conducted over a period of several years. This would have delayed the Superfund process and added
significantly to the cost of the ERA. Rather, USEPA focused its efforts on obtaining site-specific
sediment, water, benthic invertebrate, and fish data, which will be used along with the toxicological
studies from the scientific literature to estimate PCB toxicity to Hudson River ecological receptors.
Toxicological studies used to establish TRVs will be chronic studies, because exposure of ecological
receptors to PCBs is expected to be long-term. In addition, reproductive effects of PCBs are
typically studied in long-term exposure scenarios. The uncertainty associated with using TRVs
derived from toxicological studies in scientific literature rather than site-specific toxicological
studies will be addressed in the ERA.
Response to EG-17. EN-10. ES-7. and EA-11
The ERASOW (p. 37-38) outlined two approaches for assessing PCB toxicity: the Total
PCBs and Arotior mixture toxicities approach, and the most recent PCB congener-specific toxicities
and Toxicity Equivalency Factors (TEFs) approach. USEPA agrees that use of the TEF approach
has significantly improved the understanding of the relative toxicities of different PCB congeners
and of the aggregated toxicity of PCB mixtures. However, USEPA disagrees with the suggestion
that the TEF approach should be eliminated from the ERA because TEF studies may yield
conservative benchmarks that indicate lower adverse effect levels than those derived using the Total
PCBs and Aroclor mixture toxicities approach. Rather, consistent with the recommendations
published following a USEPA-sponsored workshop (ERG, 1998), USEPA believes that both
27
TAMS/Menzie-Cura
-------
approaches are appropriate for use in the ERA. USEPA will use the TEF approach for all congeners
covered by World Health Organization (WHO) TEFs (i.e., BZ# 77, 81, 126, 169,105, 114, 118.123,
156. 157. 167, and 189; Van der Berg et al., 1998), if the site data are adequate to support its use
(ERASOW, p. 39). In the ERA, USEPA will present an evaluation of the data sets for each of the
12 congeners with a WHO TEF value.
4.1.1 Total PCBs and Aroclor Toxicities
Response to ES-6
To clarify the ERASOW (p. 38), one of the toxicity reference values (i.e., dose) for chronic
effects of PCBs will be based on the No-Observed-Adverse-Effect-Level (NOAEL). and will be
adjusted to reflect differences from test species to receptor species, and from sub-chronic to chronic
values.
Response to EP-2
The potential for adverse effects will be based on comparisons of measured and modeled
exposure concentrations to both appropriate regulatory standards and TRVs obtained from the
scientific literature, even if no adverse effects are observed in the receptors of concern. The 1993
Phase 2 benthic invertebrate data compares communities from similar habitats along different
reaches (i.e.. varying PCB concentrations) of the Thompson Island Pool in the Upper Hudson River
to determine community-level differences. As noted in the ERASOW (p. 39), areas with low PCB
concentrations will be considered more representative of reference areas than areas with elevated
PCB concentrations. TRVs for benthic invertebrates have been deleted as a measurement endpoint.
as discussed in the response to ES-9.
4.1.2 Congener-specific Toxicity and the Toxicity Equivalency Factors (TEF)
Approach
No significant comments were received on Section 4.1.2.
4.2 Measures of Effect
Response to ES-8
The full citations for all sources of toxicological data used in the ERA will be provided in
the references section of the ERA.
28
TAMS'Menzie-Cura
-------
4.2.1 Benthic Invertebrate Communities
Response to EN-4 and EN-9
A quantitative index of EPT species (Ephemeroptera, Plecoptera. and Tricoptera, such as
mayflies, stoneflies, and caddisflies) and an EPT to Chironomidae ratio are not appropriate for the
Hudson River, because large rivers generally do not have the shallow, fast-moving water and the
rocky bottom that is the preferred habitat of many EPT species. In addition, the EPT to
Chironomidae ratio may be skewed; the number of Chironomid species may be high for reasons
other than water quality. The diversity indices will be used as a measure of benthic
macroinvertebrate community structure to compare similar habitats along different reaches of the
river (ERASOW, p. 39). The USEPA (1989) reference cited by the commenter was revised in 1997
(Revision to Rapid Bioassessment Protocols For Use in Streams and Rivers: Periphyton. Benthic,
Macroinvertebrates. and Fish, USEPA 841-D-97-002).
Response to EP-1
The ERA will use predicted (1993 to 2018) sediment, water, and fish PCB concentrations
calculated for the Baseline Modeling Report (due May 1999), as well as measured concentrations
from the 1993 Phase 2 field sampling program. The predicted concentrations of PCBs in all media
modeled are lower than the initial (i.e., 1993) levels, in accordance with data collected since the 1993
Phase 2 field investigation.
4.2.2 Fish Receptors
Response to EN-11
As stated in the ERASOW (p. 40), the effects to fish will be calculated using measured (for
current) and modeled (for future) PCB body burdens, using a critical body residue approach.
Toxicity Reference Values (TRVs), which represent the lowest PCB concentrations that have been
shown to cause adverse effects in test species, will be obtained from the scientific literature. The
approach used to select TRVs will evaluate the sensitivity of the endpoint and quality of the study,
as recommended by the commenter. When selecting total PCB body burden TRVs for fish,
preference will be given to studies that measure actual tissue concentration. Studies examining the
effects of PCBs on fish eggs will be used to develop TRVs for fish eggs and will be provided on a
lipid-normalized basis.
4.2.3 Avian Receptors
4.2.4 Mammalian Receptors
4.2.5 Threatened and Endangered Species
29
TAMS/Menzie-Cura
-------
4.2.6 Significant Habitats
No significant comments were received on Sections 4.2.3 to 4.2.6.
5. RISK CHARACTERIZATION
5.1 Surface Water Concentrations
5.2 Sediment Concentrations
,\'o significant comments were received on Sections 5.1 and 5.2.
5.3 Benthic Invertebrates
Response to FN-12 and EA-12
Benthic community data will be analyzed qualitatively using a weight-of-evidence approach
due to the difficulty in attributing specific results or differences between stations to PCBs alone
(ERA SOW. p. 45). The presence of contaminants in sediments other than PCBs, natural variability
in ecosystems, and human disturbances of habitats and their potential effects on benthic community
structure will be discussed in the ERA.
5.4 Fish Receptors
Response to KG-18
USEPA agrees that the ERA should provide an estimate of population-level risks to fish
posed by PCBs at the site (ERASOW, p. 46) and that Suter (1993) presented a number of approaches
to characterize population-level risk (ERASOW, p. 45). However, USEPA disagrees with the
comments that the ERASOW misrepresents Suter's work and that the ERA will not address the
magnitude of population-level risk for fish. The equations presented in the ERASOW (p. 46-47) are
a general form of the model that may be used to relate individual-level to population-level risks;
however, the exact form would depend primarily on the availability of toxicological data and the
form in which the exposure data are expressed (ERASOW, p. 47). Other approaches that may be
used include logit or probit functions to describe dose-effect and a logistic model to express the
probability that a receptor of concern will exceed a particular effect level (ERASOW, p. 47). The
results of the logit or probit approach would be combined with mortality and reproductive rates in
a population growth model to estimate population-level effects. The approach ultimately selected
for use in the ERA will address the magnitude of population-level risk, if it is found to exist.
30
TAMS'Menzie-Cura
-------
Response to EN-13
USEPA acknowledges the statement that the commenter disagrees with the concept of
evaluating population-level effects as a component of an ecological risk assessment for remedial
decision-making. However, the evaluation of population-level effects is consistent with USEPA's
Ecological Risk Assessment Guidance for Superfund (USEPA, 1997). The term ''population" and
the approximate percent of affected individuals ascribed to population-level changes at an expected
exposure concentration will be discussed in the ERA. Consistent with USEPA guidance (USEPA.
1997), the ERA will be appropriately protective of the environment in estimating risk at the
population level.
Response to EA-13
As noted in the ERASOW (p. 46), the first step in characterizing population-level risk is to
"(djefine the effects assessment as an extrapolation of series or statistical extrapolations." To clarify,
this means statistically combining the individual dose effect curves, which are extrapolations of
measured observations and modeled data, into a single dose effect curve that describes the expected
response for the entire population, which is a further extrapolation.
5.5 Avian Receptors
5.6 Mammalian Receptors
5.7 Threatened and Endangered Species
5.8 Significant Habitats
No significant comments were received on Sections 5.5 to 5.8.
6. UNCERTAINTY ANALYSIS
Response to EG-19
USEPA disagrees with the comment that the absence of site-specific ecological data creates
such uncertainty regarding population and community-level risk that the ERA will be of little use
in the remedial analysis. First, the ERA will be based on site-specific water, sediment, and biota
(i.e.. benthic invertebrate and fish) data. Second, the assessment of effects to benthic
macroinvertebrate communities will be performed using the methods set forth in USEPA guidance
(USEPA, 1997). Third, consistent with USEPA guidance (USEPA, 1997), uncertainty will be
discussed qualitatively, and quantitatively when possible, in the ERA. Once completed, the results
of the ERA will be used in Step 8 - Risk Management, to evaluate remedial alternatives (ERASOW,
p. 9).
31
TAMS'Menzie-Cura
-------
6.1 Approaches to Assessing Uncertainty
Response to EG-2 and EG-3
The commenter expresses concern regarding the uncertainty in modeling ingested doses and
body burdens and uncertainty in TRVs. A variety of fish, avian, and mammalian receptors at various
trophic levels will be evaluated in ERA. Information on ingestion rates in wild populations will be
used when available. When these rates are not available, ingestion rates based on captive
populations or allometric equations (e.g., Nagy, 1987) will be used. The dosage of PCBs will be
calculated based on PCBs in sediment, water, invertebrates, and fish from both the measured
concentrations from the 1993 Phase 2 dataset and predicted concentrations from the bioaccumulation
modeling (Baseline Modeling Report, due May 1999). The calculation of two exposure estimates
provides a basis for comparison and a method of validating modeling results. The TRVs are based
on the results of scientific studies showing observable, repeatable effects, directly related to PCBs.
TRVs will be selected for each receptor based on long-term studies on taxonomically similar species.
TRVs are intended for individual-level effects. The extrapolation to population or community effects
is based on site-specific conditions.
Another part of the weight-of-evidence approach is comparing sediment and water column
studies to accepted standards, criteria, or guidelines. Although these comparisons do not provide
a quantitative estimate of risk, the magnitude of the risk can be approximated by examining the ratio
of observed and predicted PCB concentrations to guidelines. It should be noted that regulatory
criteria and guidelines are based on individual-level observations.
Conducting various river specific studies beyond what NYSDEC, the US Fish and Wildlife
Service, and others are already conducting would provide more elements to the weight-of-evidence,
but will introduce such broad uncertainties of their own that they are unlikely to reduce general
uncertainty in the assessment. Population numbers, age-class and annual reproductive success vary
so widely in nature that only large long-term studies can ever begin to address such uncertainty.
Further, the ERA will provide adequate information for decision makers when considered
in conjunction with other parts of the Reassessment Rl/FS, such as the Human Health Risk
Assessment, the Data Evaluation and Interpretation Report, and the results of the modeling. The
Hudson River PCBs Superfund site, stretching for nearly 200 river miles, has been contaminated by
PCBs for over 50 years. Consequently, the site is complex on both a spatial and temporal scale and
decision-makers will benefit from multiple lines of evidence regarding risk management.
Response to EA-14 and EG-20
To clarify, using a sensitivity analysis to assess uncertainty, one model parameter will be
varied by a fixed percentage while all other parameters are held constant (ERASOW, p. 53). Use
of the upper bound on the ingestion rate of a prey item for the avian and mammalian models was
presented as an example. The sensitivity analysis will use the lower bound as well.
CAMS Men/ie-Cura
-------
Response to ED-10
The ERASOW (p. 53) described an approach to assessing uncertainty, which is lack of
knowledge, and variability, which is the natural heterogeneity present in the environment.
Consistent with USEPA guidance (USEPA, 1996b), a Monte Carlo-type analysis can be used to
assess variability and uncertainty simultaneously by selecting one value to represent variability (i.e.,
the 75th percentile) and then systematically changing the value to consider the uncertainty around it
(i.e., a triangular distribution).
REFERENCES
Response to EN-14
USEPA will ensure that the missing references are included in the ERA, as appropriate.
Table Number 1: Assessment and Measurement Endpoints
Response to EN-15
Exceedance of sediment guidelines will be included as effects endpoints to the benthic
community structure; survival, growth, and reproduction of benthic invertebrates; and protection of
significant habitats endpoints in the ERA. The shortnose sturgeon and brown bullhead will be
identified as omnivores in relevant tables of the ERA (ERASOW, p. 19).
Response to ES-9
The ecological community indices (diversity, evenness, dominance) are considered to be
exposure measures because they provide an estimate of the overall health of the benthic invertebrate
community. In contrast, comparison of benthic community indices at a different time or location
(e.g., different PCB concentrations) is an effects measure. Estimated exceedance of toxicity
reference values (TRVs) for benthic community structure is deleted from the ERASOW. The TRVs
are not an appropriate measurement endpoint for benthic invertebrate community structure as a
source of food for local fish and wildlife because TRVs are species-specific and do not apply to risk
at the community level.
33
TAMS/'Menzie-Cura
-------
Table Number 2; Trophic Levels. Exposure Pathways, and Food Sources
Response to EN-16
USEPA agrees that the shortnose sturgeon feeds primarily on invertebrates. The feeding
habits of the shortnose sturgeon will be revised in the tables and text of the ERA, consistent with the
ERASOW (p. 27).
APPENDIX A MODELING APPROACHES
Response to EN-17
USEPA agrees with the comment that the modeling approaches in Appendix A are general.
The ERASOW provides an outline of the work to be performed in the ERA. The models that are
used to assess risk will be fully described in the ERA.
Response to EN-18
The congeners that may be modeled include the congeners proposed for evaluation in the
TEQ approach (see response to EG-17. EN-10, ES-7. and EA-11). Most of the TEF congeners
analyzed were not detected in the samples collected.
Response to EN-19
River segments will be determined based on the spatial scale of the fate and transport
(HUDTOX) modeling. Sediment concentrations in these reaches will be based on the results of the
HUDTOX modeling. For most fish species, the model segments are expected to encompass the
exposure zones for fish that may be caught in a particular segment of the river.
Response to EA-15
Bioaccumulation will be predicted using several models (each of which incorporates a
different methodology) (Baseline Modeling Report, due May 1999). Fish PCB body burdens will
be predicted over a 25-year time frame (1993-2018). The time-varying mechanistic model developed
by GE will not be used, but USEPA will use a modified Gobas time-varying mechanistic model (see
response to EG-14).
34
TAMS Menzie-Cura
-------
ADDITIONAL REFERENCES
Bain. M.B. March 1999. Cornell University. Personal communication by telephone.
Van den Berg, M., L. Bimbaum, A.T.C. Bosveld, B. Brunstrom, P. Cook, M. Feeley, J.P. Giesy, A.
Hanberg, R. Hasegawa, S.W. Kennedy, T. Kubiak, J. C. Larsen, F.X. Rolaf van Leeuwen, A.K. Jjien
Liem, C. Nolt, R.E. Peterson, L. Poellinger, S. Safe, D. Schrenk, d. Tillitt, M. Tyslind, M. Younges,
F. Waem, and T. Zacharewski. 1998. Toxic equivalency factors (TEFs) for PCBs, PCDDs, PCDFs
for humans and wildlife. Environmental Health Perspectives. Vol.106 (12):775-792.
III. COMMENTS ON ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
Copies of the comments received during the public comment period follow.
35
TAMS/Menzic-Cura
-------
THIS PAGE LEFT BLANK INTENTIONALLY
36
TAMS''Mcnzie-Cura
-------
Federal
-------
EN-
U.S. DEPARTMENT OF COMMERCE
National Oceanic and Atmospheric Administration
National Ocean Service
Office of Ocean Resources Conservation and Assessment
Hazardous Materials Response and Assessment Division
Coastal Resources Coordination Branch
290 Broadway, Rm 1831
New York, New York 10007
November 4,1998
Doug Tomchuk
U.S. EPA
Program Support Branch
290 Broadway
New York, NY 10007
Dear Doug:
Thank you for the opportunity to review the September, 1998 Hudson River PCBs Reassessment
RI/FS, Phase 2 Ecological Risk Assessment Scope of Work. The following comments are
submitted by the National Oceanic and Atmospheric Administration (NOAA).
Background
The primary objectives of the baseline ecological risk assessment (ERA) are to quantify risks to
selected biological receptors and communities exposed to releases of PCBs in the Hudson River.
Current risk evaluations will be derived primarily from the EPA 1993 Phase 2 ecological sampling
program and NYSDEC/NOAA 1993 and 1995 PCB congener-specific fish data. It will also
include other Phase 2 investigations. Future risk will be determined from the pending Baseline
Modeling Report Other data may include 1970's-1980's RI/FS data, NYSDEC annual fish tissue
data since 1971, GE fish data, NYSDOH 1972 benthic invertebrate data and USFWS/NYSDEC
avian PCB data.
The interim ERA of the Phase 1 report concluded that PCBs in surface water, sediment and fish
exceed state and federal guidelines. Specifically, surface water PCB concentrations exceeded
ambient water quality criteria for protection of aquatic life. Fish tissue in the Upper Hudson
exceed USFWS guidelines for trout Piscivorous wildlife are consuming fish with PCBs
potentially at higher concentrations than recommended by NYSDEC or USFWS. Sediment PCBs
also exceed sediment guidelines.
Summary
EPA issued new policies related to ecological risk assessment The Hudson River ERA will
incorporate new guidelines and criteria, utilize point and probabilistic estimates of effects,
qualitatively address endocrine effects, consider the applicability of toxic equivalency factors
(TEFs) and prioritize ecological entities for protection.
The Screening Level Problem Formulation and Ecological Effects Evaluation were completed as
part of the Phase 1 activities. Study Design, Data Quality Objectives and Field Verification of the
Sampling Design were performed as part of the 1993 Ecological Sampling Plan.
This Scope of Work (SOW) covers three of the eight steps of an EPA ERA. Baseline Risk
Assessment Problem Formulation will use new guidance and studies to select assessment
endpoints and to develop a conceptual model. Site Investigation and an Analysis of Exposure and
Effects will be based on site investigation results and model output Risk Characterization will
utilize a weight of evidence approach.
X
-------
NOAA comments on Hudson River Ecological Risk Assessment Scope of Work, September 1998 (11/4/98)
The SOW deals with three distinct sections of the Hudson Riven the Upper Hudson, the
Thompson Island Pool (TIP) and the Lower Hudson. It covers the 200 miles of the Hudson from
the Battery to Hudson Falls, encompassing freshwater, brackish and estuarine habitats. PCBs will
be examined as congeners, total PCBs and aroclors.
Macroinvertebrates were sampled for PCB-congener body burdens and for benthic community
ecological metrics. Structure will be measured by evenness, dominance and diversity. PCB
water-column concentrations will be compared to ambient water quality criteria (AWQQ and
sediments will be assessed using freshwater and estuarine screening level guidelines.
Eight species of fish representing various trophic levels will be evaluated. Measured PCB tissue
contaminant levels will be utilized for all species except the federally endangered shortnose
sturgeon for which body burdens will be modeled. Contaminant profiles will be developed for
each species.
Avian receptors include the tree swallow, mallard, belted kingfisher, great blue heron, and
federally threatened bald eagle. Mammalian species include the little brown bat, mink, and
raccoon.
Significant habitats to NOAA, NYSDEC, and USFWS will be evaluated.
Assessment endpoints include:
Benthic community structure as food source to fish and wildlife,
Survival, growth and reproduction of localized benthic macroinvertebrate community and local
forage and piscivorous fish populations,
Protection (i.e., survival and reproduction) of wildlife (piscivorous and insectivorous birds;
piscivorous, insectivorous and omnivorous mammals), and
Protection of significant habitats.
Measurement endpoints include
Benthic community indices relative to trophic transfer of PCBs,
Fish tissue residues to evaluate trophic transfer of PCBs,
Fish and wildlife tissue residues to determine exceedances of PCB effect-level thresholds,
Water PCB concentrations comparison to NYS AWQC for protection of piscivorous wildlife,
and
Sediment PCB concentrations comparison to applicable sediment benchmarks.
Risk estimates will be developed from measurement endpoints. Exposure concentrations will be
modeled far species few which data are unavailable.
Comments
The authors of the ERA SOW present a concise and clear description of the planned approach to
assess exposure and risk to receptors from Hudson River PCB-contaminated media. The approach
is well-thought out and is consistent with current ERA guidance. The ERA will rely entirely on
existing data and modeling in the exposure assessment and information in the published literature
to assess toxicity; while considerable information is available for the exposure assessment, the lack
of field measurements of toxicity endpoints may result in a high degree of uncertainty in the risk
characterization. Most of the comments that follow primarily identify points for clarification.
Page 4 Section 1.2: NYSDEC and USFWS are in the process of collecting tissue samples for
PCB tissue analysis (waterfowl, macroinvertebrates, mink, turtles, bald eagles). This pending
dataset should be considered during the ERA. We recognize that results are not currently available
2
-------
NOAA comments on Hudson River Ecological Risk Assessment Scope of Work, September 1998 (11/4/98)
but they should be considered upon availability to support the modeled concentrations and to
enhance the available data incorporated into the risk assessment
Page 12 Para 2: State and federally endangered and threatened species should have been listed in
the Lower Hudson River Section 2.1.3 to be consistent with the description of the Upper Hudson IL1
River.
Page 16 Last Bullet, Page 18 Para 1, Page 44 Para 2, Page 45 Para 3: The NYSDEC sediment
benchmark is incorrectly attributed to the NYSDEC soil cleanup levels (TAGMs) developed for EN-3
protection of human health. The cited document (NYSDEC 1993) correctly identifies the technical
guidance for screening contaminated sediments, but it should be noted that there is a March 1998
edition.
Page 16 First Bullet, Page 18 Para 1: Richness, abundance and biomass were assessed in the
benthic macroinvertebrate community study. Proposed ecological metrics include diversity,
evenness and dominance. Metrics selected should aid in the assessment of whether the biological
community is impaired and to what degree. Community similarity indices are appropriate but EN-4
reference sites are required for comparison and none were sampled. EPT index and the ratio of
EPT to Chironomidae are other metrics suggested by EPA (1989). Diversity indices should not be
used because they can provide misleading information. For example, it is well known that
diversity is low in unperturbed desert environments. Since diversity is based on total number of
taxa and abundance of each taxa, stations with low abundance of many taxa could have a higher
diversity index then stations with high abundance but fewer taxa. It is important to consider
whether species intolerant to the COC are absent or reduced and whether the contribution of
pollution tolerant species has increased.
Page 18 Section 2.6.1 PCB concentrations in the water column should be compared to freshwater EN-5
or marine AWQC depending on the salinity of the water.
Page 28 Para 1 and Page 31 Para 4: Mammalian receptors in the Upper Hudson should include
small mammals such as shrews and meadow voles, which utilize the river banks and floodplains
that are potentially contaminated with PCBs and represent an important feeding area for upper EN-6
trophic avian and mammalian carnivores. By not considering floodplains in the ERA, risk to
mink, raccoons and accipiters will most likely be underestimated. Excluding floodplains from the
risk assessment also ignores a potential source of PCBs that can recontaminate riverine habitats
during flood and scour events. The ERA should address all significant exposure pathways.
Page 30 Para 1: According to Section 3.3, appropriate statistics for characterizing exposures and
body burdens will include arithmetic averages and the upper 95% confidence interval (95% UCI). EN-7
Other references throughout the SOW refer to the use of averages but not the 95% UCI. Will the
95% UCI be employed for all receptors? In the case of benthic invertebrates (Section 3.4.1), PCB
concentrations will be estimated from a distribution of site-specific BSAFs. Can we assume that
the variable estimator will be the arithmetic mean and 95% UCI?
Page 32: Direct ingestion of water is considered an exposure pathway and the dose from water is
calculated from the average concentration. PCB concentrations may be higher along shorelines it N R
than in mid-channel transect sample concentrations, resulting in an higher exposure to receptors
predominantly utilizing the shoreline habitat It is not clear whether averaged nearshore water
column concentrations or averages over a broader spatial area will be used
Page 39 Para 1: Benthic macroinvertebrate effects measures should include the ecological metrics EN-9
specified above in the comment for page 18. See EPA (1989).
Pages 37-39: The ERA mentions the use of a toxicity equivalence (TEQ) approach with calculated
TEFs compared to modeled congener-specific PCB tissue concentrations. Appendix A does not tfN-lU
explain whether this will be done for each of the models.
3
-------
NOAA com men is on Hudson River Ecological Risk Assessment Scope of Work, September 1998 (11/4/98)
Page 40 Para 2: "Measurement effects will be based on test species most similar to die receptor..."
Limited data are available for PCB residue-based effects in fish, so it may be advisable to use an
approach that evaluates the sensitivity of die endpoint and the quality of the study. Measured body
burden concentrations will need to be adjusted for concentration in the target tissue. For example, EN-11
evaluation of reproductive effects, where the initial exposure occurs via maternal transfer to the
egg, may require estimation of PCB concentrations in the mature egg from measured whole-body
or muscle tissue concentrations. Exposure concentrations for developing larvae should also
include exposure to PCBs in water and food.
Page 43 Para 1 and Page 45 Para 2: Benthic community structure data will be analyzed
qualitatively because there are concerns that it will be difficult to attribute specific results or EN-12
differences between stations to PCBs alone. In that most sediments contain mixtures of
contaminants and that benthic community structure data are generally viewed quantitatively, we
recommend the latter analysis following a weight of evidence approach. Consideration can be
given to the presence of other contaminants in the assessment of uncertainty.
Pages 45-47 and Table 1: In general, NOAA does not agree with the concept of evaluating F N 1 ^
population-level effects as a component of an ecological risk assessment for remedial decision-
making. The vague termonology used to describe population-level risk characterization can lead to
nonconservative determinations. Neither "population" nor the percent of affected individuals
ascribed to population-level changes at an effective concencentration has been defined.
References: Add Bain (1997), Gilbert (1979) and Novak (198x). These are referenced in the text EN-14
but missing from the reference section. NYSDEC (1998) should be substituted for NYSDEC
(1993).
Table 1: See Page 18 comments above regarding the use of diversity indices. Add exceedance of
sediment guidelines in the "Effects" column for the following assessment endpoints: benthic r m i
community structure; survival, growth and reproduction of benthic invertebrates; and protection of kl>-l
significant habitats. Shortnose sturgeon and brown bullhead are incorrectly identified in Table 1 as
forage fish and piscivorous fish, respectively. The text correctly describes them as omnivores.
Tables 1 and 2: Shortnose sturgeon feed primarily on invertebrates (Bain 1997). Their feeding EN-16
habits should be revised in these tables, consistent with Section 3.1.4 on page 27.
Appendix A: Modeling Approaches. The descriptions of the models are very general. EN-17
Page A-l: It is stated that "individual congeners" will be modeled and estimates will be provided in
a form that can be used for ERA - which congeners will be modeled and how will data for those EN-18
congeners be used in the ERA?
Page A-2: How will river "segments" be determined? How will sediment concentrations for the EN -19
segmentation used in the modeling be derived?
Thank you for your continual efforts in keeping NOAA apprised of the progress at this site. Please
contact me at (212) 637-3259 or Jay Field at 206-526-6404 should you have any questions or
would like further assistance.
Sincpely,
| \
' C^($^9vPm"
Lisa Rosman
NOAA Coastal Resource Coordinator
4
-------
NOAA comments on Hudson River Ecological Risk Assessment Scope of Work, September 1998 (11/4/98)
References
Bain, MB 1997. Atlantic and shortnose sturgeons of the Hudson River common and divergent life history
attributes. Environ Biol Fish 48:347-358.
EPA 1989. Rapid Bioassessment Protocols for Use in Streams and Rivers, Benthk Macroinvertebrates and Fish,
EPA/444/4-89-001, Office of Water, May 1989.
NYSDEC 1998. Technical Guidance for Screening Contaminated Sediments, New York State Department of
Environmental Protection, Division of Fish and Wildlife, Division of Marine Resources, March 1998.
ee: Mindy Pensak, DESA/HWSB
Robert Hargrove, DEPP/SPMM
Doug Fischer, ORC/NYCSFB
William Ports, NYSDEC
Charles Merckel, USFWS
Anne Secord, USFWS
Anton P. Giedt, NOAA
5
-------
CZ3
to
ft
-------
New York State Department of Environmental Conservation
Division of Environmental Remediation
Bureau of Central Remedial Action, Room 228
50 Wolf Road, Albany, New York 12233-7010
Phone: (518) 457-1741 FAX: (518) 457-7925
November 9, 1998
Mr. Douglas Tomchuk
United States Environmental Protection Agency
Region II
290 Broadway - 20th Floor
New York, NY 10007-1866
Re: Hudson River PCBs Site Reassessment RI/FS
Site No.: 5-46-031
Dear Mr. Tomchuk:
The following comments are on the Hudson River PCBs Reassessment RI/FS, Phase 3 Feasibility Study
Scope of Work and Phase 2 Ecological Risk Assessment Scope of Work, dated September 1998.
Feasibility Study Scope of Work
Section 1, Page 3. The text states, "computer models will be employed to assist in the selection of
remedial objectives as well as to assess the likely success of any remedial action in attaining these goals."
Generally, the remedial action objectives are based on applicable or relevant and appropriate
requirements (ARARs) and/or risk assessment findings. Models are generally used to assist in predicting
whether specific remedial measures will enable the goals to be achieved. The scope of work should
explain how the computer models will be used to select remedial objectives.
Section 1.2, Page 5. The Scope of Work should clearly state that all of the exposure routes found to be of
concern in the Human Health Risk Assessment (HHRA) are addressed in the Feasibility Study.
Section 2.2 Page 12 and 13 and Section 4. We suspect that the depositional areas just south of Remnant
Site 3 and adjacent to Remnant Site 4 contain PCB contaminated sediments. Remedial alternatives
should include consideration of such areas above Rogers Island.
Section 2.2 Page 15 and 16. The DEC piscivorus wildlife criteria (0.1 ppm in whole fish) must be
included in the Final Selection of Remedial Action Objectives, particularly where the reference to the
'desired level in fish' occurs.
-------
Section 2.2 Page 15. The second bullet on the page cites cancer risk as the basis for determination of
acceptable PCB levels in near-shore sediments. This bullet should be clarified to indicate that the basis
for determination of acceptable PCB levels in near-shore sediments will be based on any appropriate
risks, cancer or non-cancer, as determined in the HHRA.
Section 2.3. The report between Section 2.3 (Development of Remedial Action Objectives and General
Response Actions, page 17) and Section 4 (Development, Screening and Detailed Analysis Remedial
Alternatives, page 26) is not consistent as to whether institutional controls are part of a no action
alternative. This apparent discrepancy needs to be clarified.
Section 2.3 Page 17. The term "on-site" as used in conjunction with identifying general response actions
should be more clearly defined. As it is currently defined in the text, one might conclude that the
superfund site includes a 2-mile corridor along either bank of the river.
Table 1 Potential Chemical -Specific ARARS and Criteria, Advisories and Guidance state that the
inclusion of the U.S. Food and Drug Administration tolerance limit for PCBs in fish (2 ppm) as an ARAR
is relevant and appropriate but that its consideration in the RI/FS is to be determined. This is confusing
and should be clarified. Regardless of what is done with the Table, the FDA tolerance limit should be
used as an ARAR.
Ecological Risk Assessment Scope of Work
Page 1, at bottom and top of Page 2 - The removal of the Fort Edward Dam is overstated as a defining
event for the impact on contaminating the Hudson River. The write-up on page 3 of the Feasibility Study ED-1
Scope of Work provides a better perspective on the dam removal and this should be reiterated in place of
the referenced passage in the Ecological Risk Assessment Scope of Work.
Page 5, fourth bullet - It is recommended that this passage be revised for clarity to read as follows: irn
Estimated PCB concentrations in the diets of fish eating birds and mammals at the site are similar to or
higher than dietary concentrations recommended by USFWS or NYSDEC (TAMS/Gradient, 1991).
Page 11, last line - Explain a little further about the 'whorled pogonia.' Is this a plant or animal? ED-3
Page 12 & 13, Section 2.2 Contaminants of Concern - There needs to be some rationale provided for ED-4
limiting the discussion to PCBs. If this is to be an 'ecological assessment,' recognition at least of the
existence of other contaminants in the system is in order.
ED-5
Page 16, 'Measurement endpoints', third bullet - 'PCB body burdens' are not included in the
measurement endpoints as listed in Table 1. For example, on the second page of the table the only species
mentioned that would have body burdens measured is the tree swallow. Where actual Hudson River
samples which provide body burden data are not available, literature values may be used. Also, please
note that other species of animals mentioned may not be feeding in the river and hence, may not be
accumulating high levels or may not be impacted.
Table 1 - Why is the short nose sturgeon listed as a forage species? If it is truly endangered, is it expected
to comprise a large part of the food base for piscivorus species? Under piscivorus fish at the bottom of
the first page, it would be more accurate to recognize many of these as omnivorous. Any piscivorus
habits of the species may be functions of life stage and size.
-------
Page 20, paragraph 4 - There are data available on PCBs in mink for New York State in Foley et al. The _
full citation is available if desired. In the next paragraph, Palmer and Fowler are not in the literature cited
section, whereas Hornshaw et al. from the paragraph above is in the literature cited On page 21, the
Gilbert 1989 citation is not in the literature cited section either.
ED-8
Page 31, section 3.3.5 - More data on bald eagles may soon be available based on ongoing NYSDEC
research. If this will be of use, please let us know.
Throughout Sections 3, 4 and 5, the assessment approach described is 'linear.' The focus is almost
exclusively on the river, not as an ecosystem, but as a north-south geographic feature. The species
outlined for the Ecological Risk Assessment are considered to derive their energy from the river itself and
there is little weight given to some of the most productive habitats or ecological zones. Except for some
forage species and benthic invertebrates the shallow, near shore littoral areas are not evaluated. There is
no mention of the transition zone from aquatic to terrestrial habitats and likewise the riparian, wetted ED-9
perimeter, and flood plain habitats are absent from discussion. Reptiles, amphibians, soil invertebrates
(e.g., earthworms, burrowing insect larvae), mammals (e.g., shrews and moles), birds such as woodcock
form a diverse complex array of organisms inhabiting these peripheral habitats which may be larger in
spatial extent than those directly associated with the river. In addition then to the direct exposure to
animals in these habitats, there is the re-exposure to the aquatic system of PCB running off the surface of
the flood-plain. Although the concentrations are relatively low, it represents a widespread surficial
phenomenon which should be taken into account in the risk assessment.
On page 53, paragraph 2 - This paragraph should be rewritten and expanded for clarity. Otherwise, ED-10
eliminate it since it does not impart useful information.
If you have any questions on the above, please call me at (518) 457-5637.
Sincerely,
William T. Ports, P.E.
Project Manager
Bureau of Central Remedial Action
Division of Environmental Remediation
Enclosure
cc: John Davis, NYSDOL
Robert Montione, NYSDOH
Jay Fields, NOAA
Lisa Rosman, NOAA
Anne Seacord, USF&WS
-------
E. Crotty
R. Tramontane
J. Lobby
R. Sloan
S. Sanford
I. Carcich
M, O'Toole
W, Daigle
W. Demick
K. Farrar
-------
e
n
to
-------
EA-1
HUDSON RIVER PCBs REASSESSMENT RI/FS
COMMENTS ON PHASE 2 ECOLOGICAL RISK ASSESSMENT
SCOPE OF WORK, DATED SEPTEMBER, 1998
Prepared by David D. Adams, Member-at-Large
Saratoga County Environmental Management Council
1. Section 1.5, P.6: It is recommended that a synopsis of the ORNL reports be EA-1
included in the Ecological Risk Assessment giving the relevant points used in the Risk
Assessment.
2. Section 1.6, P. 8 and Table 1: The draft scope for the Hudson River Natural Resource EA-2
Damages Assessment Plan, September 1988 states that information to quantify injury to
aquatic insects, birds, reptiles and amphibians, and mammels is "likely not available". In
view of this statement, why does EPA believe sufficient information to realistically
evaluate ecological risk in these areas is available? Wouldn't it make for a better
evaluation to focus the ecological risk assessment on those areas (water, sediment) and
species (fish, invertibrates) for which data are available? The ecological risk assessment
seems in many areas to be qualitative enough without resorting to further speculative
treatment of species for which data specific to the Hudson River are lacking. It is noted
that the Peer Review Committee reviewing the Preliminary Model Cabitration Report
commented on the desirability of getting data on mink and on concentration of heavy
metals.
3. Section 2.1,3, P. 12: Is there enough information on the sources of PCB's in the Lower EA-3
Hudson River to make a risk assessment of this area meaningful in terms of the effects of
remedial action in the Upper Hudson on species in the Lower Hudson? If not, it is not
evident what benefit a risk assessment of the Lower Hudson River has to this study.
4. Section 2.5, P. 17: How can observations on "disease & deformities during sampling" EA-4
be of use when toxic materials other than PCBs present in the Hudson River that could
cause any observed disease or deformity are not being evaluated? Again, this entire risk
assessment is so fraught with qualitative aspects that it should not be further clouded by
further unsubstantiated considerations. (For example, it is noted that in reply to a question
at the October 20th Availability Session regarding beuthic community indices, it was
stated that specific information relating these indices to PCB concentrations is not
available and a qualitative approach would be used. Similar responses occurred several
other times in response to question at the Availability Session. (See also the discussion on
P. 22 on measurement endpoints).
5. Section 2.6.5, P. 21: The bald eagle should be deleted from consideration as a
threatened or endangered species.
6. Section 3.2, P. 28: How can striped bass be properly evaluated using Upper Hudson
River data when they are most heavily exposed to PCB sources in the Lower Hudson
EA-5
EA-6
-------
-2-
independent of the Upper Hudson? This species, and the sturgeon, should be eliminated
from the risk assessment.
6. Section 3.2, P. 29: It makes no sense to run a model which shows a time-varying
decrease over time and then ignore the model results in the risk assessment. The risk EA-7
assessment procedure should be modified to allow proper consideration of the model
results.
EA-8
7. Section 3.3, P. 29: Available GE data should also be considered here and elsewhere.
8. Section 3.4, P. 32: The dose formula should include a factor to account for the fraction EA-9
of PCB intake absorbed
9. Section 4.1, P. 37: The size of the Hudson River and level of effort required to obtain ^ ^
site-specific data are not acceptable excuses when weighted against the potential impacts
resulting from the risk assessment. EPA owes the communities in the Hudson River area
a better evaluation than what is evident from this SOW.
10. Section 4.2.3, P. 39: The decision to use only congener BZ#77 should be reviewed to EA-11
see if the other congener data can also be used.
11. Section 5.3, P. 45: The second paragraph highlights another example of the highly EA-12
qualitative nature of this assessment.
12. Section 5.4, P. 46: The statement at the top of the page says the "effects assessment"
is an extrapolation of an extrapolation which seems highly questionable. This statement
should be clarified.
13. Section 6.1, P. 53: Lower bounds as well as upper bounds should be considered. EA-14
14. P.A.-2: The food chain model should be a bioenergetic model as recommended by
the Peer Review Committee for the Preliminary Model Calibration Report. Also, EA-15
consideration should be given to the model developed by GE and to seasonal time scales.
301016
-------
Community Interaction
Program
-------
Department of Earth and Atmospheric Sciences
College of Arts and Sciences
Earth Science 351
Albany, New York 12222
an a?
iS
University at Albany
STATE UNIVERSITY OF NEW YORK
518/442-4466 or 4556
Fax: 518/442-5825 or 4468
Chair@ atmos albany.edu
http://www.atmos.albany.edu
EP-1
October 29, 1998
Mr. Douglas Tomchuk
USEPA - Region 2
290 Broadway - 20th Floor
New York, NY 10007-1866
ATTN: ERA SOW Comments, Phase 2 Ecological Risk Assessment
Dear Mr. Tomchuck:
I have reviewed the above referenced document and submit the following comments:
1. The use of 1993 field sampling data for estimating body burdens will overestimate
exposure levels relative to toxicity reference values (p. 40). There will also be a bias
in comparisons to water quality criteria, because PCB concentrations in both the
water column and fish have since declined, and were also lower prior to 1990. The £p_J
relation of existing concentrations to body burdens is dynamic, not static, and body
burdens have a variable lag effect depending upon the receptor. To attempt to model
future body burden levels or response on the basis of one direct observation/organism
or group at a fixed point in time will produce unacceptable uncertainty in the results.
2. What will determine a chronic exposure level, or whether (p. 38) a toxicological
adverse reaction has occurred in organisms? Is this by comparison to a toxicity
threshold level from the literature or a regulatory standard (e.g. fish), even if no
adverse effect is observed, i.e. an assumed toxicity effect in the absence of any direct EP-2
observation? The relevance of threshold or reaction levels from the literature, or as
determined in other settings with different receptors or test species must be
established. It is not clear how a fixed time point (1993) benthic invertebrate
community assessment (p. 39) will accomplish this, given the dynamic variation
occurring in the other two "effect measures" (p. 39).
3. Comparison data (exposure concentrations; observed ecosystem reactions; type of
PCB and system recovery status, etc.) for other PCB contaminated sites, such as the
Great Lakes and St. Lawrence River, should be included; and the risk characterization
of same provided for perspective. What are the similarities and difference to the
Hudson, and how might these affect the risk assessment?
In retrospect, this type of comparison should be applied to the SOW previously prepared
for the assessment of human exposure risk. An estimate of population exposure risk derived via
extrapolation from laboratory exposure studies at high PCB doses or other indirect means is
-------
profoundly uncertain when the actual health statistics for a human population of documented
exposure history, and observable body burden, can be obtained and/or directly measured, as at
Ft. Edward, Hudson Falls, and the Mohawk Indian tract on the St. Lawrence River. This is no
less a matter of ecosystem effects evaluation, with the advantage of a much longer receptor
history. Data of this type is critical to a resolution of the current debate about risk assessment in
human.
Very truly yours,
George W. Putman,
Emeritus, Dept. of Earth and
and Atmospheric Sciences Faculty
-------
Public Interest Groups
-------
SCENIC
HUDSON
ES-1
INC
9 VASSAR STREET POUGHKEEPSIE, NY 12601 (914) 473-4440 FAX (914) 473-2648
Scenic Hudson's Comments on
USEPA's Hudson River PCBs Reassessment RI/FS
Phase 2 Ecological Risk Assessment
Scope of Work
As with the human health risk assessment, a full characterization of ecological effects associated
with exposure to Hudson River PCBs is a critical part of EPA's Reassessment of the Hudson
River. While we believe the overall approach to assessing ecological risks seems reasonable, as
does the discussion of PCBs, we have the following concerns:
General Comments:
Toxicity of PCBs - It appears that PCB toxicity is not sufficiently documented. We do not have
great confidence that the measurement of PCB toxicity in receptor species will be properly
conducted. We suggest that EPA review the approach used and the contractor's experience with
ecological risk assessment to ensure a useful ecological risk assessment is completed, v
The discussion of the toxicity of PCBs is- ladong-iri reference-material: A considerable amount
of literature exists: regarding' this-subjedt, atfd'"shbiiTd be included i&4he assessment.
Overall Ecological Health
From examining the range of PCB concentrations that exist in the Hudson River ecosystem and
the effects of those concentrations on various receptor species it would seem reasonable to
include a qualitative discussion of the health of the entire ecosystem and disappearance of
certain species such as the mink. Certain factors such as the population decline of the mink, is an
indicator as to the health and diversity of the ecosystem. Unquantified, but important, effects of
PCB contamination will fall by the wayside in a EPA number driven decision-making process.
From our experience, studying environmental impacts and assessments, issues relegated to
qualitative discussions are often dismissed and have little bearing on the outcome. However, a
qualitative discussion as to the overall health of the Upper Hudson River ecosystem is better than
no discussion at all.
Other concerns:
Section 2.6. Receptors of Concern, pg. 17 - Once again we are concerned that the risk
assessment is based on the proper knowledge :of receptor .species, particularly of avian receptors £S-3
(Section 2.6.3j pg. 19). Several critical references are not sited;for this section;:
more...
email: scenichu@mhv.net
-------
Section 2.7, Risk Questions, pg. 22 - In the last sentence, the use of the word "impacting" is not
appropriate for a risk assessment. The risk assessment should show whether there is a risk or ES-4
not, it does not have to show impact. The word impact would more appropriately be used in the
Natural Resources Damage Assessment.
ES-5
ES-6
Section 4.1, Estimating the Toxicity of PCBs, pg. 37 - Toxicological endpoints that will be
evaluated in the ERA should include Reproduction and 'Development
Section 4.1.1 Total PCBs and Aroclor Toxicities, pg. 38 - The last sentence in the first
paragraph incorrectly uses the word "threshold". A threshold is where effects start. This
sentence is confusing as, "No-Observed-Adverse-Effect-Level (NOAEL) toxicity values" will be
used as "thresholds" for estimating adverse ecological impacts. Please clarify.
Section 4.1.2 Congener-specific Toxicity and the Toxic Equivalency Factors (TEF) ES-7
Approach -pg. 39 - In discussion of congener specific toxicity assessment it is indicated that "of
the most toxic (coplanars) congeners, the Phase 2 database includes usable data for BZ#77 only.
If the data are adequate..." First, it is not clear as to how this data will be used. Secondly, we
question whether or not an assessment can be based on BZ#77 data only. BZ#126 is usually the
most important congener used in ecological risk assessment. The most meaningful data set
should be used to properly assess ecological risks
Section 4.2, Measures of Effect, pg. 39 - Sources of toxicological data, which "will include ES-8
refereed scientific literature, the USEPA AQUIRE database, and government publications", are
too vague. Please identify sources of toxicological data more specifically.
Table 1 Assessment and Measurement Endpoints, (page 1 of 3). Regarding the first ES-9
Assessment Endpoint identified, Benthic community structure as food source for local fish and
wildlife, Ecological community indices (diversity, evenness, dominance) placed under
Measures/Exposure is a measure of effect, not exposure, and should be moved to the Effect
column. Under the Effect column - Estimated Exceedance of Toxicity Reference Value (TR Vs) is
species-specific and is not generally used for a community and should be deleted.
11/10/98
-------
General Electric
-------
EG-1
GE Corporate
Environmental Programs
Metvln B. Schwaiger
Manager. Hudson River Project
New York Stata EHS Affairs
General Bectric Company
1 Computer Drive South
Albany. New York 1220S
Telephone (518) 458-6646
Fax:(518)458-1014
November 2, 1998
Mr. Douglas Tomchuk
USEPA- Region 2
290 Broadway - 20th Floor
New York, N.Y. 10007-1866
RE: ERA SOW Comments
Dear Mr. Tomchuk:
The process of preparing an ecological risk assessment to assess, with
reasonable certainty, the ecological effects of PCBs in the Upper Hudson River is
truly a daunting one. Extreme care must be given to several issues, including
adequately assessing communities or populations of benthic invertebrates, fish,
birds and mammals in the river and adjacent land, and differentiating between
effects of PCBs on these populations and effects of other contaminants, such as
mercury and lead, and other human disturbances.
Unfortunately, we are concerned that the process EPA has developed for this
assessment will not prove at all useful to its intended goal. The course which
EPA sets out in the Scope of Work is little more than a screening level analysis.
Apart from a limited study of benthic invertebrates, the Agency has not, and is
not planning to, collect data on the biota of the Hudson and its shoreline. No
information gathered during this process, other than the concentration of PCBs in
the water column, sediments or limited selection of biota, will be specific to the
Upper Hudson River. Such an analysis, while potentially useful for developing
future investigation and analysis, cannot illuminate whether PCBs are affecting
the sustainability of biological communities or populations present at this
particular site.
This failure to attain site-specific information on the wildlife populations of the
Upper Hudson River will result in significant uncertainty in the risk assessment -
we urge that the uncertainty be quantified and stated candidly so that the public
-------
may understand the likelihood of obtaining any particular ecological change as a
result of a recommended clean-up plan.
We look forward to the Agency's response to the enclosed detailed comments
regarding the Ecological Risk Assessment Scope of Work. If the Agency would
like to discuss these comments in greater detail, please do not hesitate to contact
me.
cc: Richard Caspe
William McCabe
Melvin Hauptman
John Cahill
Douglas Fischer
Albert DiBernardo
Charles Menzie
Sincerely,
Melvin B. Schweiger
301025
-------
COMMENTS OF GENERAL ELECTRIC COMPANY ON
HUDSON RIVER PCBS REASSESSMENT RI/FS
PHASE 2 ECOLOGICAL RISK ASSESSMENT SCOPE OF WORK
November 2, 1998
-------
TABLE OF CONTENTS Page
Introduction and Executive Summary 1
The Proposed Ecological Risk Assessment Will Not Provide Sufficient Information for
Remedial Decision-making 5
The SOW Lacks a Clear Statement of the Objectives and Description of Assessment and
Measurement Endpoints 10
A. The ERA Needs a Clear Statement of Objectives 10
B. The Assessment Endpoints Should Focus on the Sustainability of Biological
Communities or Populations 11
C. The Measurement Endpoints Should Provide Information Relevant to the
Assessment Endpoints and Should Go Beyond the Approaches Described in the
SOW 11
1. The Measurement Endpoints Described in the SOW Have Significant
Limitations 13
2. EPA Must Use Objective Criteria To Select TRVs 15
3. The ERA Must Consider Additional Lines of Evidence 16
Problem Formulation (Section 2) 18
A. Site Characterization (Section 2.1) 18
1. The SOW Does Not Acknowledge Significant and Relevant Data Sets . 18
a. Benthic Macroinvertebrates/Phytophilous Macroinvertebrates ..18
b. Fish 19
c. Birds 20
2. The Site Does Not Extend Below the Federal Dam at Troy 22
B. Contaminants of Concern (Section 2 2) 23
C. Site Conceptual Model (Section 2.4) 24
D. Receptors (Section 2.6) 25
Exposure Assessment (Section 3) 26
A. Exposure Pathways (Section 3.1) 26
1. Fish 26
2. Birds 26
3. Mammals 27
B. Quantification of PCB Fate and Transport (Section 3.2) 28
1. The ERA Should Rely Exclusively on Time-Variable, Mechanistic Models
Such as GE's 28
2. EPA Should Not Hold Initial Concentrations Steady Over an Extended
Period 30
C Observed Exposure Concentrations (Section 3 3) - 30
1. Benthic Invertebrates ~ 31
2. Fish 31
3. Birds 31
-------
D. Modeled Exposure Concentrations (Section 3.4) 32
VI. Effects Assessment (Section 4) 35
VII. Risk Characterization (Section 5) 35
VIII Uncertainty (Section 6) 38
IX. Conclusion 39
-------
I. Introduction and Executive Summary
The General Electric Company ("GE") is pleased to submit these comments on the
"Hudson River PCBs Reassessment RI/FS Phase 2 Ecological Risk Assessment Scope of Work"
("SOW").
The purpose of conducting an ecological risk assessment ("ERA") under
Superfund is to aid in making a decision on the appropriate remedial course of action. To be
useful in remedial decision-making, an ERA needs to be focused on the conditions at the
Superfund site in question and to have substantial certainty associated with its predictions. These
benchmarks must be met if the risk manager is to be able to use the risk assessment to project
with reasonable specificity the ecological changes that will be accomplished by various remedial
actions.
In the case of the Hudson River PCBs Superfund site ("Site"), the chemical of
concern to EPA entered the river beginning approximately 50 years ago. Direct discharges from
the manufacturing processes ceased approximately 20 years ago. For some period since, there
have been seeps of PCBs entering the river These have now been reduced to quite low levels.
Given the persistence of PCBs, often in dechlorinated form, these simple facts provide the
opportunity for an analysis of the ecological effects of PCBs on a large scale: Over 50 years and
40 miles of river, what impacts, if any, have PCBs had on the populations of benthic invertebrates,
fish, birds, and mammals in the River and the adjacent land?
Unfortunately, EPA has failed to take advantage of this opportunity. Apart from a
limited study of benthic invertebrates, the Agency elected not to collect data on the biota of the
Hudson and its shoreline in order to examine effects, if any, of PCBs on wild populations.
1
-------
Consequently, it will now be very difficult, if at all possible, for EPA to conduct an ERA that will
be useful in selecting a remedial action.
The course that EPA sets out in the SOW is largely a screening-level analysis.
Possible concentrations of PCBs in various receptor species will be deduced by "modeling" the
dose or body burden that the animals receive, and those values will be compared to PCB levels
reported in the literature as associated with some effect in individuals of the receptor species.
With the exception of benthic invertebrate community analysis, EPA has not proposed any
analysis of effects on populations or communities. The only site-specific information to be used is
the concentration of PCBs in the water column, the river sediments and a limited selection of
biota. This type of analysis can tell one whether there is a basis for further investigation and
analysis; it cannot be used to determine whether unacceptable impacts are occurring or are likely
to occur. In other words, while the proposed analysis may be useful for generating risk
hypotheses, it does not adequately test the hypotheses. Real data on ecological health are needed
for that.
The fundamental limitation of the proposed analyses is their inherent uncertainty.
First, there is little or no measurement cf the actual exposure of biota at the Site. Second, there is
no method to extrapolate from predicted effects in individuals to determine whether the
sustainability of the population is affected. Third, there is no critical analysis of the literature to
determine whether and how particular studies should be employed at the Site.
It is unfortunate that these obstacles were not overcome by data collection and
analysis at the Site over the more than 15 years that EPA has been assessing the Upper Hudson
River These unfortunate circumstances lead to two important resultsin the present ERA:
2
-------
Substantial effort must be made to limit and narrow the uncertainty in the risk
assessment, including the obvious task of testing the postulated risk; and
The degree of uncertainty at the end of the assessment will probably be large;
consequently, the final uncertainty of the results must be quantified where possible
and otherwise stated candidly so that the risk manager and the public understand
the likelihood of obtaining any particular ecological change as a result of a
particular remedial action.
At a Superfund site the size of the Hudson Site, EPA will presumably be
examining to what extent, if any, a remedial course of action can or will affect the sustainability of
the wild populations which are examined. At least five actions will be necessary to narrow and
limit the uncertainty of the risk assessment:
First, the exposure of animals to PCBs from the Hudson must be rigorously
analyzed, using site-specific information where possible, and the
uncertainty of estimated exposure must be quantified.
Second, the literature studies relied on in the assessment must be critically
examined. What is the uncertainty in extrapolating from one species to
another? What is known of the dose/response relationship? Was the study
conducted correctly? Where use of a study introduces uncertainty to the
analysis, it must be quantified.
Third, the relationship of effects on individuals to effects on the dynamics of the
population must be explicated, taking into account the density-dependent response
of the population. Once again, uncertainty must be recognized explicitly.
Fourth, the existing data on the biota of the Hudson and its shorelines must
be examined to assure that any projections of the risk assessment are
consistent with the facts. A similar comparison must be made to the peer-
reviewed literature.
Fifth, in order to determine whether the reduction of PCBs in the water
column or sediments will have an effect on the exposed biota, effects of
other organic chemicals and metals must be examined. Obviously,
chemicals other than PCBs as well as other stressors, affect the biota of the
Hudson. If those effects are not understood, it is not possible to tell
whether remediation of PCBs will have the desired beneficial effects.
3
-------
These five steps need to be applied to the assessment of the effects of PCBs not
only at the present levels found in the water column and the sediments but also at those future
levels which will occur under "no action" or other remedial courses. The element of time is
important; a central issue for the Agency is to determine how much a particular remedial action
will accelerate the time at which a particular population-level effect will cease.
At the conclusion of this exercise, the Agency must set out the uncertainty
associated with its estimates of how the sustainability of populations of the Hudson will be
changed by specific remedial actions, so that the risk manager and the public can determine the
level of confidence to place in the predictions of the ERA in deciding on a remedial course of
action.
4
-------
~. The Proposed Ecological Risk Assessment Will Not Provide Sufficient Information £q_j
for Remedial Decision-making
To be useful to a decision-maker, the baseline ERA at a Superfund site must
provide reliable information that is not subject to unreasonable uncertainty. EPA (1998a) states
that risks must be "characterized in terms of magnitude, severity, and spatial and temporal
distribution of effects." The SOWs proposal for the ERA at this Site fails this basic test. The
approach set out in the SOW is to use a very limited set of site-specific data in conjunction with
models to estimate both the ingested dose of PCBs by receptors in the Hudson River and the PCB
body burdens of these biota. These exposure estimates will then be compared to literature-
derived "toxicity reference values" ("TRVs") to develop a "toxicity quotient" ("TQ").
Alternatively PCB concentrations in water and sediment will be compared directly to generic
ambient water criteria or sediment values. Under this method, either concentrations in
environmental media that exceed regulatory criteria or values or a computed TQ which is greater
than one will be deemed to identify unacceptable risk to individual organisms. The substantial
uncertainty that this method incorporates, combined with its inability to provide quantitative
information about risks to biological communities or populations, makes it inadequate for a
baseline ERA for a site the size and complexity of the Hudson River PCBs Superfund site.
Consequently, the information generated by the planned ERA will be of little assistance to the
Agency's decision about appropriate remedies for the site. With no reasonably certain estimate of
the true risks that PCBs in the Hudson River pose to biological communities or populations, the
Agency will not be able to discern whether risk reduction measures are necessary and, if so, to
5
-------
what extent any remedial alternative would reduce those risks to acceptable levels faster than
would occur under the no-action scenario.
EPA could have planned to conduct a better and more useful ERA for the Site.
The Agency, other federal and state government agencies, and numerous independent scientists
have been studying the presence and effects of PCBs in the Hudson River for more than two
decades. In 1990, EPA announced that it would reassess its 1984 Record of Decision for the
Site. The initial step in this reassessment was to collect and summarize existing information about
the Site and determine what additional data were needed to complete the reassessment. The 1991
Phase 1 Report for the Site contained this initial summary, including the results of a screening-
level ERA for the Site, which concluded that a "comprehensive ecological risk assessment,
including population, community and ecosystem interactions in response to PCB exposure, is not
possible with available data" (EPA 1991). Despite this acknowledgment, the only new
information EPA proposed to collect to support the baseline ERA was some limited data on
macroinvertebrate communities in the upper Hudson (EPA 1992). EPA chose not to collect
additional data about the health and abundance of fish, birds, mammals or other species of
interest, and the SOW makes clear that the Agency does not intend to consider the substantial real
world data that can provide additional insight into these issues.
Instead, the Agency is proposing to perform a perfunctory ERA that is more
appropriate for a 5-acre landfill than for a 40-mile reach of river. This is wholly inadequate for
several reasons.
Several of the methods set out in the SOW are not intended for use in a baseline
risk assessment. Comparison of PCB concentrations in water and sediments to regulatory or
6
-------
guidance criteria, such as ambient water criteria or generic sediment guidelines, is a screening
technique intended to provide a quick indication of the potential ecological risk (EPA 1997). The
screening informs the risk assessor whether additional investigation is warranted but does not
provide a conclusion that ecological risk actually exists. Indeed, this was the method primarily
employed in the Phase 1 Report, and repeating it now will not provide additional knowledge of
actual ecological risk. The SOW, nevertheless, proposes to rely on such screening-level analyses.
Even where the Agency intends to go beyond this screening-level examination, the
information and analyses to be generated will be of little use in remedial decision-making because
they will be subject to substantial uncertainty. This uncertainty results from several aspects of the
Agency's approach. First, having collected little data on PCB concentrations in biota, the SOW
proposes to rely primarily on modeling both ingested doses and body burdens. There are little or
no data against which to validate and calibrate the results of such models, leaving the risk
assessment with an unknown level of uncertainty. The SOW, moreover, indicates that the Agency
does not intend to consider the majority of the available data on the health and abundance of
biota, including selected receptors, which provide a critical line of evidence and to which the
model results must be compared.
EG
Second, the SOW proposes to compare the uncertain results of modeling to
uncertain TRVs derived from the lab or other ecological settings which may bear little
resemblance to the Hudson. In many instances, TRVs may not exist for the endpoints or species
being examined, or the TRVs may be insufficient for conclusive risk analyses. This leads to
further uncertainty when these results must be extrapolated to relevant endpoints or species of
concern.
7
-------
EG-3
Third, the assessment endpoints set out in the SOW identify only individual-level
risks, such as survival or reproduction. As a result, the uncertainty of the models and TRVs is
compounded by the uncertainty of translating the analysis of individual-level risks to community
or population-level risks. What makes this problem worse, is that the full extent of this
uncertainty cannot be quantified. One can have little confidence in an analysis that is subject to
unquantifiable and unknown uncertainty. '
EG-4
An additional shortcoming of the proposed ERA is that it will to consider the risks
associated with other stressors, such as contaminants other than PCBs and human disturbances to
the ecosystem. An analysis that focuses solely on PCBs will not place those risks in the context of
the overall stresses on the ecosystem. Without an assessment of the relative risks associated with
PCBs, EPA will not be able to determine whether remedial actions focused on PCBs will actually
be beneficial to the ecosystem. In other words, the ERA will provide no information to allow the
decision-maker to assess whether the reduction of PCBs in the Hudson River will produce any
tangible benefits to the ecosystem.
Accordingly, the proposed ERA will be of little assistance to the remedial decision-
makers. The substantial uncertainty associated with the proposed approach, the failure to assess
population or community risks from PCBs, and the inability to place such risks in the context of
other stressors limit the utility of the information to be derived from the ERA. In fact, the ERA
will primarily be useful in demonstrating that-there is no risk to certain receptors from the present
concentrations and mass of PCBs. As a result, EPA has two choices if it decides to proceed with
an ERA of the type set out in the SOW. The Agency can examine the screening-level results, and
where the potential for risk is indicated, can set out to collect site-specific effects data.
8
-------
Alternatively, without sufficient data to test the risk hypotheses, the Agency must candidly
recognize the substantial uncertainty and very limited usefulness of the ERA in its remedial
decision-making.
9
-------
m. The SOW Lacks a Clear Statement of the Objectives and Description of A«<*smgnt
and Measurement Endpoints
A. The ERA Needs a Clear Statement of Objectives EG-5
Perhaps the most important element of the risk assessment is defining its
objectives a related set of questions which it is intended to answer. The structure, scope and
details of the ERA depend on and flow from clearly defined objectives. Without such goals, one
is unable to determine what endpoints should be assessed and what measurements should be
made. In other words, one needs first to develop a set of hypotheses in the context of what is
known about the Site before determining the analyses and data required to test those hypotheses.
As the SOW explains, the overall objective of the ecological risk assessment is to
aid in the reassessment of the 1984 Record of Decision (SOW at 12). This, in turn, depends on
an understanding of whether PCBs in the upper Hudson constrain "the ability of populations
and/or communities to sustain themselves at or near" the Site (EPA 1998a). The SOW, however,
does not present any questions whose answers will provide information relevant to this issue. We
recommend focusing the ERA to answer the following questions:
Are PCBs currently adversely affecting biological community structure or the
population dynamics of the key receptors?
If so, when will PCBs no longer adversely affect biological community structure or
population dynamics under No Action?
If so, to what degree will remediation reduce the time to reach the point at which
PCBs are no longer adversely affecting biological community structure or
population dynamics in comparison to No Action?
These questions provide a basis for deriving assessment endpoints that can be used to determine
whether biological communities or populations are at risk and, if so, whether PCBs are causing
10
-------
those risks. These objectives, moreover, permit the selection of measurement endpoints with
relevance to those assessment endpoints.
B. The Assessment Endpoints Should Focus on the Sustainabilitv of Biological EG-6
Communities or Populations
An assessment endpoint is a valued characteristic of an ecological receptor that
may be affected by exposure to a stressor (Suter 1990). Assessment endpoints are explicit
expressions of the actual environmental value to be protected (EPA 1997; 1998b). That is,
assessment endpoints are the specific ecological values or characteristics of the selected receptors,
the risks to which are quantified by the risk assessment. They usually consist of an entity
(population, community, or habitat) and a property (population number, rate of growth,
community structure) They are designed to provide answers to the defined objectives for the
ERA
Suter (1990) has suggested criteria for evaluating potential assessment endpoints:
unambiguous operational definition; accessibility to prediction and measurement; susceptibility;
biological relevance; and societal relevance. As EPA (1998a) states, the "goal of the Superfund
program is to maintain and ensure self-sustaining populations/communities." Therefore, in the
context of the Hudson River, the appropriate assessment endpoint is the sustainability of
communities or populations native to the Upper Hudson River. rr,
tli-7
C. The Measurement Endpoints Should Provide Information Relevant to the
Assessment Endpoints and Should Go Bevond the Approaches Described in the
SOW
Measurement endpoints are measurable responses to a stressor that are related to
the assessment endpoints (Suter 1990; 1993). Relationships between the measurement endpoints
11
-------
and assessment endpoints enable the risk assessor to use the results of field observations,
bioassays, and literature reviews to decide whether a risk of harm has resulted or is likely to result
from the stressors. While using a measurement endpoint to approximate or estimate the effects
on an assessment endpoint introduces uncertainty into the assessment, that uncertainty will be
minimized by selecting measurement endpoints that are closely related to the selected assessment
endpoints and by developing multiple lines of evidence.
There are several different types of measurement endpoints, including
benchmarks, literature-based TRVs, direct measures obtained through field surveys or toxicity
tests, and biomarkers. To match a measurement approach with a given receptor and assessment
endpoint, one must consider several aspects of the measurement endpoints, including: technical
feasibility; strength of association with the assessment endpoint; specificity of the measured effect
to the chemical, receptor and site of interest; and representativeness of the various measurement
approaches.
Unfortunately, the measurement endpoints proposed in the SOW are focused on
individual-level effects, such as survival, growth, and reproduction, and no method is suggested to
translate these results to community or population-level responses. Indeed, the SOW
acknowledges that, although the methods it intends to use characterize risk at the individual level,
"if risks are present at the individual level, they may or may not be important at the population
level." For example, the SOW identifies survival, growth and reproduction as assessment
endpoints for fish and invertebrates. While these may appear to relate to macroinvertebrate
communities and fish populations, they only reflect effects on individuals. Abundance,
persistence, diversity, and trophic structure are population or community level endpoints. The
-------
SOW does not include any method for linking changes in survival, growth or reproduction in
individuals to changes in population or community status. The ERA must establish measurement
endpoints aimed at ensuring the maintenance of sustainable populations and communities.
1. The Measurement Endpoints Described in the SOW Have Significant
Limitations
The measurement endpoints proposed in the SOW are primarily screening-level
endpoints in which PCB concentrations in water or sediments are compared to ambient water
quality criteria ("AWQC"), sediment quality values ("SQVs"), or TRVs. These approaches can
not constitute the primary measures of ecological risk in the baseline ERA for this Site.
Reliance on screening-level benchmark values, such as AWQCs and SQVs, in a
baseline ERA is both inconsistent with EPA guidance (EPA 1997) and redundant with the analysis
conducted in Phase 1. As the SOW notes, comparisons to AWQCs or SQVs merely indicate that
there is a "potential for risk" to aquatic organisms, including invertebrates, fish, and terrestrial
receptors, wetland community structure, and habitat value. The results do not provide a
quantitative estimate of risk to communities and populations. If the benchmarks are exceeded, it
simply indicates that further study is required. Moreover, EPA has already compared sediment
and water chemistry to defined benchmarks, including AWQC and generic sediment quality
guidelines, such as Long and Morgan (1991), in the Phase 1 investigation (EPA 1991). According
to EPA (1997), "requiring cleanup based solely on [a screening-level risk assessment] would not
be technically defensible."
The TQ technique set out in the SOW suffers from a number of similar problems.
Most important, because the proposed TQ approach typically focuses Qh an individual organism,
13
-------
it is appropriate only as a first step in the ERA. If this procedure predicts adverse effects on the
individual organism, further analysis is needed to assess effects at a higher level of organization.
Such additional work may involve obtaining more site-specific exposure information, probabilistic
TQ analyses, population-level modeling, or the collection of field data to attempt to verify the
predictions.
There are several additional weaknesses with TRVs and the TQ method:
Exceedance of a TRV (i.e., TQ>1) does not indicate the magnitude of the
potential impact or its biological significance.
There are few published TRVs which correlate body burdens in the relevant
species with the toxicological endpoints to be evaluated (survival, growth, and
reproduction). Further, in the few cases where body burdens and toxic effects are
reported both in a single study, the correlation between the two measurements is
either not reported or is weak.
PCB body burdens can be reported in a variety of ways, such as on a whole body
basis, on a lipid-normalized basis, or on an organ-specific basis. When body
burdens are reported in the scientific literature, they usually are reported as tissue
measurements for target organs (e.g., gonads, brain, liver), rather than for the
entire body. Because bioaccumulation factors are rarely reported for those same
target organs (bioaccumulation factors most often allow estimation of entire body
burden), the ecological risk assessor is often left comparing "apples to oranges"
when attempting to derive a TQ from a study which reports body burden rather
than dose.
TRVs are generally based upon No Observed Adverse Effects Levels
("NOAELs"). Appropriate use of TRVs based upon NOAELs depends on the
ability of the particular study to demonstrate a dose-response relationship. This, in
turn, depends on the dose intervals, which is a factor of the a priori
characterization of toxicity by the investigator. The variance of the observed
effects may inhibit the ability to distinguish a significant effect from a non-
significant effect.
14
-------
2.
EPA Must Use Objective Criteria To Select TRVs
Because of their limitations, TRVs must be selected with care using objective
criteria. The SOW does not contain any such criteria, proposing instead to select "appropriate"
TRVs While reliance on professional judgment is inevitable, there are more objective
considerations that should guide the selection of TRVs:
Relevance of the endpoint and species measured. The use of studies with the
same measurement endpoints as those described in the SOW (i.e., survival, growth
and reproduction) and the same species minimizes the need to extrapolate results
and will decrease the conservatism and the uncertainty of estimated risk.
Similarity of test species to receptor species, in terms of sensitivity. The SOW
states that TRVs will generally be derived from the most sensitive individual
(based on species and age class). This adds unnecessary conservatism to the ERA.
For example, much of the information on toxicity of PCBs to fish has been
generated using, and reflects the unique sensitivity of, salmonid species. Because
the proposed receptors are not salmonids, such studies can not be used to derive
TRVs for this Site.
Degree of chlorination of PCB mixture tested. The PCB mixture tested in the
TRV study must be comparable to that found of the Hudson River or in the prey
that are consumed by the receptor being evaluated.
Study duration. Chronic studies provide more relevant information for risk
assessment than acute studies. The appropriate duration of the TRV study should
be determined by the toxicity test endpoint and the toxicokinetics of PCBs. For
example, if the test endpoint is reproductive effects, the duration of the toxicity
study should be the entire reproductive cycle of the tested species.
Exposure pathway. Chronic studies with dietary exposures are the most
appropriate mimics of exposure for a receptor species in the field. Acute laboratory
studies with unnatural exposure scenarios, such as injection, cannot be used when
exposure is modeled on the basis of ingested doses.
Use of measured vs. estimated responses. Only TRVs related to actual, measured
responses by the test organism should be used.
Nature of dose-response relationship. Unbounded NOAEts should not be used as
TRVs An unbounded NOAEL (a NOAEL for a species for which the lowest
15
-------
observed adverse effect level has not been determined), provides little information
on the toxicity of the tested material.
Overall study quality. Demonstration of a relationship between a tested chemical
and a toxicity endpoint depends on statistical significance and power. There
should be an absence of confounding factors, such as inadequate nutrition in some
test groups, and the control population must exhibit normal survival and health.
Only high quality, controlled studies that are consistent with other such studies can be used to
provide measurement endpoints.
3. The ERA Must Consider Additional Lines of Evidence
Given the limitations of the TQ approach, the SOW correctly proposes to consider
different lines of evidence (page 15). Several lines of evidence not listed in the SOW that should
be considered include:
Benthic invertebrate community biomass, abundance, and diversity compared with
a range of these parameters measured in comparable aquatic habitats in the region.
Fish species abundance, diversity and richness compared with a range of these
parameters measured in comparable aquatic habitats in the region.
Data on the abundance and reproduction of Hudson River birds, including tree
swallows collected by the USFWS and nesting bald eagles on the Hudson River.
Data on the abundance and reproduction of Hudson River mammals.
Each line of evidence is characterized by its own individual strengths and weaknesses; the purpose
of using more than one is to reduce uncertainty associated with the overall assessment.
In order to reconcile the different lines of evidence, the Agency must develop at
the outset an objective method for identifying and weighing their relative strengths. This method
should define specific attributes that will be used to judge the quality of each measurement
endpoint Examples of such attributes (from Menzie et al., 1996) include: strength of association
16
-------
between the measurement endpoint and assessment endpoint; site-specificity; stressor-specificity;
quality of data and overall study; availability of an objective measure for assessing environmental
harm; sensitivity of the measurement endpoint for detecting changes; spatial representativeness;
temporal representativeness; quantitativeness; correlation of stressor to response; and use of a
standard method.
In short, a more sophisticated ERA than the approach set out in the SOW is
required for this Site. Such an assessment should not rely primarily upon comparisons with
benchmarks or TQs but should include site-specific informational and dose/response or
concentration/response relationships. Risk assessments based solely on chemistry, body burdens,
or model-derived doses are prone to excessive conservatism because the only readily identifiable
benchmarks are conservative ones. Consequently, modeled risk estimates must be verified by
field data.
17
-------
IV. Problem Formulation (Section 21
We discussed our concerns with the assessment and measurement endpoints in the
previous section. We identify below some additional problems with the SOWs problem
formulation discussion.
EG-
A. Site Characterization (Section 2.11
1. The SOW Does Not Acknowledge Significant and Relevant Data Sets
An important element of problem formulation is a detailed review and integration
of all available and relevant information concerning the Site. This is particularly important for the
type of ERA proposed in the SOW Where the primary analytical method is the use of exposure
and bioaccumulation models and highly simplified risk models, compiling the available data,
whether limited or substantial, provides an important means for constraining model results. At
this Site, although the Agency has not collected much data for use in the ERA, there is a
substantial body of data collected by others, including GE, concerning the general ecological
conditions in the Hudson River. The Site Characterization section of the SOW fails to present
such a discussion. The Agency should use these data to validate its desktop analyses. We identify
and discuss some of the data sources below.
a. Benthic Macroinvertebrates/Phvtophilous Macroinvertebrates
EPA's 1993 study of benthic macroinvertebrates (reported as part of its 1995
Database Report) shows that indices of macroinvertebrate community structure are not correlated
with PCB concentrations in sediments. NYSDEC's most recent report on macroinvertebrate
similarly concludes that there have been no community-level impacts in the Upper Hudson due to
PCBs. "High levels of PCBs have been found in the macroinvertebrate tissues downstream of Fort
18
-------
Edward since analysis began in 1976. ... No impact at the community level has been observed
at any site that can be attributed to high PCB levels" (NYSDEC 1993).
More recent macroinvertebrate community data are available for the Thompson
Island Pool ("TIP") and Stillwater regions of the Hudson River. GE sponsored surveys of benthic
and phytophilous macroinvertebrates in these areas in September 1997. The mean total
invertebrate abundances ranged from approximately 8,000 invertebrates/m2 to 35,000
invertebrates/m2, with a maximum of 87,000 invertebrates/m2 at a vegetated station in the
southern end of the TIP (Exponent 1998). Total taxa richness of benthic macroinvertebrates
ranged from 33 to 58 invertebrate taxa depending on whether the habitat was vegetated or not
and which area was sampled. These results suggest that PCBs have not impaired benthic and
phytophilous macroinvertebrate communities in the TIP where concentrations of PCBs in
sediment are likely to be highest in the River.
b. Fish
There is a vast database of fish abundance in the Lower Hudson River that
suggests that PCBs have not adversely affected relevant populations. While these data are not
directly applicable to the upper Hudson because of differences in species and the relatively lower
levels of PCBs in the lower river, they do support a conclusion that PCB levels comparable to
past and present levels in the lower Hudson are unlikely to affect fish abundance in the upper
river.1 Available information includes:
1 Although we explain in the next section that the ERA should not address risks from upper river
PCBs to lower river biota, these data provide useful and relevant information for the upper river
biota.
19
-------
Data showing that PCBs have not impacted the abundance of shortnose sturgeon
(Kynard 1997). Indeed, the most recent data demonstrate that the Hudson River
shortnose sturgeon population is booming (NMFS 1997).
* Numerous studies have showed that the decline in the coastal striped bass
population was not the result of Hudson River PCBs, but rather overfishing and
water quality conditions in the Chesapeake Bay (NMFS and USFWS 1985;
ASMFC 1990) ("given the very healthy status of the Hudson River stock, which is
well documented to have relatively high tissue concentrations of PCBs, it would
appear that such levels . .. may not pose a threat to striped bass from a population
biology perspective") Far from declining, the spawning stock of striped bass in
the Hudson River has shown remarkable growth.
NYSDEC and the utilities collect data each year on fish abundance in the lower
river which reflect generally healthy fish populations unaffected by PCBs.
c. Birds
The ERA does not need to conduct an additional analysis of tree swallows. The
data presented in Secord and McCarty (1997) show that even high body burdens of PCBs have
not affected the reproductive success of tree swallows. Despite some significant deficiencies in
the study design (e.g., lack of suitable reference area), the data from this study support the
following conclusions:
Swallows which are maximally exposed to PCBs in the Hudson River have normal
numbers of young, and these young grow normally.
Evaluation of reproductive parameters relative to tissue residue data shows that
tree swallows did not respond to PCBs in a dose-dependent manner. Hatchability
and reproductive success were highest, and nest abandonment lowest, at sites with
the highest PCB concentrations in eggs and hatchlings.
* Reproductive success, nest abandonment, hatch rates, nest quality and plumage
development observed in the experimental tree swallow populations nesting
adjacent to the Hudson River were within the range of reference or unexposed
populations.
20
-------
This information shows that PCBs in the Hudson do not affect reproduction in, and thus do not
pose population-level risks to, tree swallows.
EPA should not extrapolate Secord's and McCarty's tree swallow data to bird
species which have different diets or experience different exposures to PCBs. These species may
also be affected by other stressors which have a more demonstrable impact on population
dynamics {e.g., lost habitat, human disturbances, other chemicals).
Other relevant information concerning birds that should be considered in the ERA
includes:
Successful breeding by bald eagles in Hudson River habitats appears to be
increasing. There were two nesting pairs on the Hudson River in 1996. One pair
did not lay eggs; the other pair laid eggs which did not hatch. The cold, wet
weather of that year is suspected to have caused the failure of these eagles to
produce young (NYSDEC 1997). In 1997, two nests were also present. One
juvenile fledged from a nest in Greene County; the other nest was unsuccessful, as
it was in 1996 In 1998, three pairs of bald eagles attempted to breed. All three
pairs laid eggs. One attempt resulted in three fledged young, another pair
produced one fledgling and one unhatched egg, and the third nest had one
unhatched egg (Nye 1998).
Data showing that bald eagles use the Hudson River during winter migration and
are breeding successfully in Hudson River habitats. The mid-winter bald eagle
survey shows an increase from 29 bald eagles using the Hudson River in the winter
of 1979 - 1980 to 103 eagles in the winter of 1995 - 1996.
Data showing the increasing use of Hudson River habitats by peregrine falcons.
Peregrine falcons have established nests on bridges over the Lower Hudson River.
While these nests are producing live young, falcons have been injured or killed in
accidents with automobiles and from falling off nest sites before they are capable
of full flight (Nye 1998).
Data showing that mallards are "demonstrably secure" throughout the New York
Bight watershed and are "widespread, abundant and secure in the state of New
York" USFWS (1998). The NYSDEC (1997) reports that, on the basis of
breeding surveys, the mallard population using the Hudson River estuary is "stable
to increasing." Mid-winter counts of waterfowl show generally increasing
21
-------
numbers of mallards and other species with a peak in the 1995 survey of more than
16,000 birds. (NYSDEC 1997).
2. The Site Does Not Extend Below the Federal Dam at Trov EG-9
Another error in the site characterization is its definition of the Site as the 200-
miles of river from Hudson Falls to the Battery in New York Harbor. As GE has previously
informed the Agency,2 the administrative record supporting the addition of the Site to the
National Priorities List ("NPL") limits the Site to the 40 miles above the Federal Dam at Troy.
EPA's post-rulemaking statements to the contrary cannot change this limitation. SfiS U.S. v.
Asarcn Inc. No. C V96-0122-N-EIL (D. Idaho Sep. 30, 1998) (post-rulemaking statements
cannot change scope and size of site from the description provided in the NPL record). GE's
disagreement with EPA on the scope of the Site is particularly important in the context of the
ERA, which is purporting to assess risks from PCBs to biota and habitats in the Lower River.
EPA's Feasibility Study Scope of Work makes it clear that EPA is limiting its analysis to potential
remedial actions in the upper river. Assessing ecological risks in the lower river, however,
suggests that the Agency may in fact be attempting to justify a remedial action on the basis of
benefits to the lower river.
As we have previously explained, the limitation of the Site to the upper river
means that the Agency cannot seek to justify any remedial action in the upper river on the basis of
benefits to the lower river. If the Agency attempts to do so, it would be obligated to investigate
and evaluate remedial alternatives, such as source control, in the lower river; consider the
additional sources of PCBs (and other contaminants and stressors) to fish and other biota in the
2 See Nov. 6, 1997, letter from Angus Macbeth to Richard Caspe; May 5, 1998, letter from
Angus Macbeth to Douglas Fischer.
-------
lower Hudson; and identify the much wider group of parties who rightfully should be classified as
PRPs. The presence of other dischargers of PCBs in the lower river is well known to EPA; the
Agency has conducted recent studies of PCB discharges into New York Harbor, including
sampling outfalls. The Agency made the importance of other contaminants plain in its 1984 ROD,
concluding "that detectable levels of dioxin, dibenzofurans, mercury and chlordane (from known
and unknown sources) have also been identified in Hudson River fish, and that even if PCBs
decrease to an acceptable level, the fishing bans would continue on the basis of these other types
of contaminants."
EPA cannot have it both ways. The Agency cannot describe the Site as
encompassing the 150 miles from Troy to the Battery and then address only one contaminant and
one or two PRPs outside that 150 miles as the sole subjects for remedial consideration. The
scope of EPA's Superfund activity at the Site is circumscribed by the characterization and
definition of the site which EPA promulgated in its NPL rulemaking years ago.
B. Contaminants of Concern (Section 2 21 . EG-10
The SOW proposes to limit the ERA's consideration to PCBs. There are other
chemical and physical stressors in and along the Hudson River, which also must be considered.
For example, sediment concentrations of cadmium, copper, lead, mercury and zinc measured in
the 1993 EPA survey in the Upper Hudson exceed NYSDEC sediment quality values. Fish
sampling in the upper Hudson reveals that fish there have DDE and DDT concentrations that
have been shown to impair bald eagle reproduction elsewhere (Anthony et al. 1993). Similarly, it
is likely that habitat loss and alteration, modification to landscape for agriculture and
23
-------
urbanization, and recreational hunting and trapping have impacted wildlife populations and
communities in the Hudson River watershed.
The presence of other contaminants and stressors is important. EPA (1998b)
provides that ERAs should assess "an array of stressors that may be influencing the assessment
endpoints and describing the diversity of potential effects." In the limited instances where EPA
intends to consider site-specific data, such as benthic macroinvertebrates in the upper Hudson,
other contaminants and stressors may confound observations and/or have additive or antagonistic
interactions with PCBs Even where the SOW proposes to rely primarily on exposure modeling
for its assessment, other stressors are relevant. Although exposure modeling avoids issues
associated with confounding stressors, it still is necessary and appropriate to place the estimated
risks associated with PCBs against the backdrop of other stressors. Unless other stressors are
considered, the remedial decision-maker will be unable to determine whether a remedial action
directed at PCBs will have any beneficial effect. The ERA should interpret any stresses to
ecological receptors that may be due to PCBs in the context of-all anthropogenic stresses to the
Hudson River. In essence, EPA must determine what the conditions in the Hudson would be hul
for PCBs.
C. Site Conceptual Model (Section 2 4^ EG-11
The site conceptual model has several problems. First, it identifies exposure
pathways, but neglects environmental fate processes such as burial and dechlorination that act to
reduce exposure over time. Second, it fails to include other stresses on the Hudson River
ecosystem, including other chemicals, conventional water-quality degradation and other human
impacts. Third, the conceptual model diagram (Figure 5) ignores aquatic vegetation, which is an
24
-------
important component of the aquatic food web and which supports invertebrates that are important
for forage fish. This is important because the potential impacts of remediation activities on
aquatic vegetation and therefore on the ecosystem as a whole must be considered.
D. Receptors ("Section 2 61 EG-
EPA should take into account the following points on the receptors identified in
the SOW:
The feeding habits and habitat use of smallmouth bass, which are listed as a
receptor in Table 1, are similar to those of largemouth bass. Analysis of
smallmouth bass as a receptor is therefore redundant and unnecessary.
Some of the threatened and endangered species listed do not use the habitats of the
Hudson River For example, the whorled pogonia is an orchid which grows in
highly acidic, peat-like soils characteristic of bogs, not riparian wetlands; northern
harrier is not closely linked to aquatic habitats but forages for mammals over
meadows and older wetlands. Therefore, these species should not be considered in
the ERA.
Striped bass and shortnose sturgeon do not occur in the Upper Hudson River and
are therefore not appropriate as receptors for this ERA.
25
-------
V. Exposure Assessment (Section 3) EG-13
A. Exposure Pathways (Section 31)
The SOW contains several errors and inconsistencies in its description of exposure
pathways. In many instances, the SOW does not consider all the available site-specific and
species-specific information that can reduce uncertainty and improve exposure estimates.
1. Fish
The SOW is inconsistent and incorrect in its characterization of fish diets. Site-
specific, empirical information on fish diets is available (e.g., Exponent 1998) and should be used
as the basis for determining the relative proportions of different prey in the diets of fish.
2. Birds
As noted above, there is no need to conduct further assessments of tree swallows
because Secord and McCarty (1997) shows that PCBs in the Hudson are not affecting tree
swallow reproduction. In any event, the SOW incorrectly states that worms are consumed by tree
swallows Most studies of tree swallow diets report that insects make up the majority of the tree
swallow diet (e.g., McCarty 1995) and that tree swallows also rely on plant materials, particularly
during winter (Robinson 1992). Secord and McCarty (1997) also contains information on the
concentrations of PCBs in tree swallows for 1994, and the authors have similar data for 1995 and
1998. EPA should use this site-specific data.
Empirical, site specific information on the diets of bald eagles is also available.
Peter Nye of NYSDEC has found remains of grebes, eels, pickerel, bullhead, herring and carp in
eagle nests. Information on the PCB concentrations in some of these prey, including bullhead and
eels, are available (Secor 1998) Moreover, NYSDEC has collected samples of the blood of
26
-------
juvenile eagles, eagle eggs, and the remains of prey in nests from the eagles breeding on the
Hudson River. Samples are being stored for analysis of organochlorines. EPA should work with
NYSDEC to determine actual PCB concentrations in the samples of bird tissue and their prey,
3. Mammals
The SOW characterizes the diet of raccoons and mink incorrectly. Both species
consume a variety of foods. The diet of mink includes mammals, birds, and amphibians, and the
raccoon diet includes terrestrial insects and plant material. EPA should use the references cited in
the EPA Wildlife Exposure Factors Handbook (EPA 1993) to estimate the relative importance of
each prey type in the diets of mammalian receptors Mid develop exposure models accordingly.
EPA must also consider the specific use of Hudson River habitats by these receptors and weight
exposures accordingly. For example, mink or raccoons may favor small streams and wetlands for
foraging over the main stem Hudson River.
27
-------
EG-14
B. Quantification of PCB Fate and Transport f Section 3 21
1. The ERA Should Relv Exclusively on Time-Variable. Mechanistic Models
Such as GE's
The SOW describes EPA's plan to use the fate, transport and bioaccumulation
models and geochemical analyses presented in the DEIR/PMCR in the ERA. These models are
critical components of the reassessment, especially with regard to risk prediction under alternative
remedial scenarios. GE has previously submitted comments on these documents, identifying a
number of serious problems (GE 1996, 1997). The peer review group convened to review the
PMCR echoed many of GE's concerns. EPA must address the recommendations of GE and the
peer review group before using the models in the ERA.
GE has developed an Upper Hudson River sediment transport model that is
mechanistic and uses data-based formulations to describe resuspension and deposition processes
realistically. This model has been rigorously calibrated and validated and should be considered for
use in the fate and transport modeling.
Appendix A to the SOW describes how the Agency intends to use the different
bioaccumulation models for the Site: the Bivariate Statistical Model ("BSM," used to compute
average PCB levels), the Probabilistic Food Chain Model ("PFCM," used to compute average
levels and variability), the Gobas Model, and the time-variable, mechanistic bioaccumulation
model developed by QEA for GE. As GE has explained previously (and as the peer review panel
confirmed), EPA should not rely on the two steady-state statistical models in the reassessment
because they do not provide an appropriate basis for predicting average PCB levels in the Upper
Hudson River.
28
-------
While it generally is good scientific practice to check for consistency among
different approaches to a given problem, this is true only insofar as the approaches are developed
properly and (at least to some degree) independently and add useful information. As described in
GE's original comments concerning the PMCR, there are serious limitations to the BSM and
PFCM approaches that limit their utility:
The Hudson River has exhibited extreme variation in exposure levels as well as
lipid content of fish. PCB levels in the food web do not respond immediately to
changes in these parameters, so PCB levels measured in fish at any point in time
may be in the process of responding to changes in exposure levels. That is, the
steady state assumption in the BSM and PFCM is violated.
The BSM has no predictive power within the Thompson Island Pool, based on the
observation that there is no relationship between observed and computed
largemouth bass Aroclor 1254 levels (Figure 9-12 of the PMCR). The pattern of
observed vs. BSM predicted values differs among reaches. Without further study
into the mechanisms underlying bioaccumulation, it is not clear why the
bioaccumulation model should differ in its predictions among reaches. Finally, the
BSM overpredicts at low concentrations (Figure 9-12). PCB levels are now
declining in the Hudson River. Thus, it is anticipated that model predictions will
overestimate the PCB levels in biota in the future.
The PFCM confuses uncertainty with variability. For example, the computed
distribution of biota-sediment bioaccumulation factors ("BSAFs") should provide a
representation of the distribution of invertebrate PCB levels as seen by the fish.
The BSAF distribution used in the PFCM was determined directly from the data,
even though the variance in the data is due in large part not to the true variability
in the sediment/invertebrate relationship, but to uncertainty. This uncertainty is a
result of the sediment and invertebrate samples not being collected at exactly the
same location, because the species composition of the samples may not be the
same as the species composition of the diet of the forage fish, and because of
analytic uncertainty. Because of this confusion, the resulting distribution of BSAF
values has no physical meaning.
The PFCM requires the answer to solve the problem. The predator/forage fish
bioaccumulation factor ("PFBAF") was determined using the observed distribution
of predator PCB levels. However, the whole purpose of the model is to compute
the distribution of predator PCB levels. Similarly, the parameters that determine
the spread, or variance, of the computed PCB levels are constants that are
29
-------
validated by comparison with the distribution of the data. The computed spread
and shape of each distribution will not change in the future, even as average
sediment and water column PCB levels decline. The model provides no additional
information beyond what the data already provide.
In contrast, a mechanistic model avoids these problems and incorporates all of the
information that is in the steady state models (that is, measured PCB levels, dietary information),
as well as ancillary information (bioenergetics, toxicokinetics, and time-variable processes) in
computing average PCB levels. Therefore, EPA should use a time-variable mechanistic model to
compute average levels, such as the one developed by QEA for GE. The distributions of PCB
levels should be estimated directly from the data.
2. EPA Should Not Hold Initial Concentrations Steady Over an Extended
Period
The SOW states that the exposure model parameters will be assumed to remain
constant for extended periods of time (e.g., 5-10 years). No justification for this approach is
provided, and given the observed declines in PCB concentrations in sediment, water and biota,
this assumption runs counter to the available evidence. Initial .concentrations should not be
assumed to hold steady over five-to ten-year increments. Time-variable fate, transport and
bioaccumulation models, like GE's, should be used to generate this information.
C. Observed Exposure Concentrations (Section 3 31 EG-15
The SOW proposes to characterize exposures and body burdens based on 95
percent upper confidence limits on averages, an approach that will result in a conservative
estimate of actual Site conditions. Although this approach might be appropriate for a screening-
level assessment, the ERA is a baseline risk assessment. The ERA should use mean exposure
levels with appropriate quantification of uncertainty.
30
-------
The traditional measure of uncertainty (mean +/- x standard error) is not the
most appropriate for all data sets. The best measure depends on the shape of the distribution of
the data (e.g., normal, lognormal). The ERA should consider other techniques including Land's
method for lognormal populations and nonparametric bootstrap methods.
1. Benthic Invertebrates
EPA has collected data on benthic invertebrate communities and PCB body
burdens at a number of stations in the Upper Hudson River. According to the SOW, the ERA
will use these data to calculate site-specific biota-sediment accumulation factors for use in food-
chain modeling. BSAF values should be used along with sediment PCB data to estimate
invertebrate PCB levels. Because of the potentially large uncertainties associated with field
measurements of benthic invertebrate BSAFs, the ERA should compare the site-specific values
with other field and lab measurements for consistency.
2. Fish
It is not clear whether or to what extent the ERA will use the significant data on
PCB concentrations in fish collected by NYSDEC, GE, and others. To the extent they are
available and valid, the ERA should use all the site-specific data on fish body burdens to calibrate
and validate the models.
3. Birds
The approach to be used to assess exposures to insectivorous birds is not clearly
described. According to the SOW, data collected by the USFWS will be used to evaluate tree
swallow body burdens "for those locations at which data are available." What approach will the
ERA use for other locations? If the ERA uses BSAFs, what specific diet assumptions will the
31
-------
Agency employ? Employing unvalidated assumptions will create substantial uncertainty in the
ERA.
D. Modeled Exposure Concentrations ("Section 3 41 EG-16
Given the failure to collect data to support its ERA, the Agency is left to rely
primarily on modeling approaches to determine PCB concentrations in biota. Without sufficient
data for validation and calibration, modeling compounds uncertainty and often leads to
unnecessary conservatism. This is particularly true for the modeling proposed to support the
ERA. If models are used, the Agency must use the available data to develop and then calibrate
and validate them.
First, in the case of the Hudson River, outside of the substantial database on PCB
concentrations in fish, there are few data against which to calibrate and validate the proposed
models. Where data could readily be developed, as for PCB concentrations in the eggs of eagles,
EPA has chosen to ignore the relevant information. Without validation, the level of certainty in
the results is unknown.
Second, the SOW provides a general model to describe ingested dose,3 but does
not explain how the ERA will translate such doses to body burdens. For example, the SOW
refers to assimilation and metabolic efficiency terms, stating that dose (meaning exposure) will be
expressed as a body residue or as an ingested dose rate, depending on the TRV. The SOW does
not provide a model for determining bird and mammal bioaccumulation, examples of assimilation
3 The ERA presents two ingestion exposure algorithms for birds and mammals: one for water and
one for food. The units for the ingestion rate term (IRj in the equation on p. 33) should be kg wet
weight/day. The units for concentration of PCBs in food items (CJ should specify that
concentration is expressed on a wet weight basis.
32
-------
and metabolic efficiency terms, or data to support such terms. Similarly, the ERA states that
biomagnification factors ("BMFs") may be used to predict concentrations of PCBs in eggs of
piscivorous birds. With the exception of bald eagles, there are limited data to support the
development of BMFs for piscivorous birds. This is particularly inappropriate because, for most
of the bird and mammal receptors, the toxicity database provides TRVs in terms of ingested dose,
not body burden. Compounding uncertainty by extrapolation of a modeled ingestion rate to a
body burden, only to compare body burdens to the limited TRVs available for this endpoint,
cannot be justified, and will lessen the value of the risk assessment for decision-makers.
Turning to the specific inputs to the models, it is important to understand and
assess the relative proportions of different prey items in the diet in order to determine the proper
dose. For example, Salyer and Langler (1949) and Davis (1980) report that more than 75 percent
of the kingfisher's diet is comprised of cyprinid species, with crayfish comprising another thirteen
percent of the diet. Additional foods, such as lizards, frogs, small snakes, salamanders, insects,
small mammals, young birds and berries, have been reported in. the literature. Landum et al.
(1993) and Kaufman (1996) The ingestion models must also incorporate the correct size of
consumed prey species. Inclusion of larger prey, which may have higher PCB body burdens,
would overestimate exposure through the diet. This type of information and site-specific PCBs in
prey are needed to construct a food-web model that is truly representative for the Hudson River.
Similarly, the SOW states that PCB concentrations in benthic invertebrates will be
averaged using all samples to obtain exposure point concentrations ("EPCs") for fish. It is more
appropriate to estimate average BSAFs and then multiply this result by average surface sediment
PCB concentrations because surface sediment PCB levels vary considerably over long and short
-------
spatial scales. Therefore, PCB levels measured in a few locations are not necessarily
representative of pool-wide surface sediment concentrations.
34
-------
VI. Effects Assessment (Section 4)
The SOW proposes two different approaches for estimating the toxicity of PCBs:
(1) using total PCBs and Aroclor toxicities and (2) using congener-specific toxicities and toxic
equivalency factors ("TEF"). There are many factors which favor the use of total PCBs or
Aroclor-based concentrations at this Site. First, most of the data for the Hudson River are
presented in total or Aroclor-based concentrations. The available capillary column data do not
quantify many of the congeners relevant to a TEF analysis. Using Aroclor or total PCB-based
effects data are preferable to estimating congener compositions of aged Aroclor mixtures (with
attendant problems of selective depletion and enrichment) and comparing the resulting toxic
equivalent ("TEQ") values to TEQ-based effects data.
Second, there are significant problems in using the TEF approach. While the TEF
approach has provided significantly improved understanding of the relative toxicities of different
PCB congeners and of the aggregate toxicity of PCB mixtures, its value for ecological risk
assessment is limited because the endpoints of TEF studies ace often cellular or biochemical
indicators Benchmarks derived from these studies are even more conservative than screening
benchmarks derived from studies in which the endpoint is survival, growth, or reproduction.
As a result, GE recommends limiting the analysis to total or Aroclor-based
concentrations.
35
-------
VII. Risk Characterization (Section 51
As we have noted previously, use of the TQ approach as the only line of evidence
for certain receptors as described in the SOW is insufficient for a baseline ERA at a site the size
and scope of the Hudson River PCBs Superfund Site. One significant limitation of the TQ
method is that it only provides information about individual level effects. To be useful, these
results must be translated to effects at the community or population level.
The SOW presents a model which it claims achieves such a translation. Although
the model purports to be based on Suter's (1993) approach for characterizing population-level
risks, it, in fact, misrepresents Suter's work. As discussed in Chapter 8 of Suter (1993),
population-level risk assessments require, at a minimum, integration of toxicological information
with information on the survival rates and reproductive rates of the exposed population. The
model described in the SOW merely characterizes the distribution of exposures within a
population. No attempt is made to characterize actual changes in rates of survival, growth, or
reproduction. As long as the exposure distribution is being compared to a TRV, population-level
risk is not being addressed.
Moreover, the model will provide only the value of a normal cumulative
distribution function (the probability for a given value from a normal cumulative distribution) for
the standardized mean of the linear combination of two random variables. This is not the same as
estimating the probability that any given individual from one distribution is greater than any one
individual from the other distribution, as stated in the text. The model estimates the probability
that the average body burden is greater than the average TRV. This formula does not estimate
36
-------
the probability that the tissue chemical concentration of any given individual fish from a
population will exceed a TRV.
Chapter 8 of Suter (1993) discusses a range of methods for extrapolating
individual-level effects to population-level effects. The simplest approach is to (1) estimate dose
or exposure-response relationships for survival (by life stage) and reproduction, (2) calculate the
effects of exposure on the rates of survival and reproduction of the exposed species, and then (3)
translate changes in survival and reproduction into changes in the intrinsic rate of increase (r) of
the exposed population. This approach, however, is more suitable for comparative risk
assessments than for site-specific assessments because it does not consider the influence of
density-dependent processes that would be expected to partially offset PCB-related reductions in
survival or reproduction. Age-structured population models that incorporate density-dependence
are described in Chapter 8 of Suter (1998) and in Barnthouse et al. (1990). The time available to
EPA since the Preliminary Assessment would have been sufficient to collect site-specific data for
Hudson River fish populations for input into these types of models. Models developed using data
collected from comparable populations in other ecosystems could still be used, although the
uncertainty would be higher. Regardless, the results would be more meaningful than the TQ
approach proposed by EPA because they would address the actual magnitude of ecological
responses to PCB exposures, not just the absence or potential presence of a response.
37
-------
Vm. Uncertainty fSection 6^ EG-19
The SOWs description of uncertainty includes (1) general categories of
uncertainty as described in EPA guidance and (2) methods for Monte Carlo analysis of exposure
models. It Ms to mention the greatest single source of uncertainty in the assessment, which is
the virtual absence of site-specific ecological data. The principal consequence of this absence is
that the assessment might be able to suggest risk to individuals, but it will be unable to shed light
on the magnitude of actual risks to communities or populations related to past, present or future
PCB exposures. In fact, the results of the proposed ERA will be so uncertain that they will be of
little use in the remedial analysis
EG-20
While it is true that sensitivity analysis is one approach to assessing uncertainty,
the example provided in the first paragraph of page 53 is not a true sensitivity analysis. Rather,
running the avian and mammalian models using the upper bound on the ingestion rate of a prey
item represents a Maximum Exposed Individual (MEI) evaluation. To conduct a sensitivity
analysis, the model would have to be repeatedly executed, each time varying one parameter by the
same degree {e.g., +/- 10%) while holding all other parameters constant. Once all parameters
have been varied, the effects on the model output can be compared in order to draw conclusions
regarding the relative sensitivity of the model to the different parameters and the relative
uncertainty associated with each parameter.
38
-------
IX. Conclusion
The ERA proposed in the SOW has a number of significant deficiencies. The basic
approach is to conduct screening-type or TQ analyses that are subject to substantial uncertainty.
The result will be a highly uncertain estimate of risks to individuals; no information will be
provided to address the critical assessment endpoint of sustainability of biological communities or
populations. In these circumstances, EPA can not place substantial reliance on the ERA in its
remedial decision-making.
39
-------
REFERENCES
Anthony, et al. 1993. Environmental contaminants in Bald Eagles in the Columbia River Estuary.
J. Wildlife Management 57:10-19.
Atlantic States Marine Fisheries Commission. 1990. Source Document for the Supplement to the
Striped Bass FMP, Amendment #4. Appear. 1956.
Barnthouse, L.W., G.W. Suter II, and A.E. Rosen. 1990. Risks of toxic contaminants to
exploited fish populations: influence of life history, data uncertainty, and exploitation intensity.
Environmental Toxicology and Chemistry 9:297-312.
Davis, W.J. 1980. The belted kingisher, Megaceryle alcyon. Its ecology and territoriality. M.S.
Thesis, University of Cincinnati, Cincinnati, Ohio.
Environmental Protection Agency. 1991. Phase 1 Report - Review Copy Interim
Characterization and Evaluation, Hudson River PCBs Reassessment RI/FS. Volume 1. EPA No.
013-2N84. U.S. Environmental Protection Agency, Washington, DC.
Environmental Protection Agency. 1992. Final Phase 2 Work Plan and Sampling Plan, Hudson
River PCBs Reassessment RI/FS. EPA No. 013-2N84. U.S. Environmental Protection Agency,
Washington, DC.
Environmental Protection Agency. 1993. Wildlife Exposure Factors Handbook. Volume I of II.
EPA/600/R-93/187a. U.S. Environmental Protection Agency, Office of Research and
Development, Washington, DC.
Environmental Protection Agency. 1997. Ecological Risk Assessment Guidance for Superfund:
Process for Designing and Conducting Ecological Risk Assessments. EPA 540-R-97-0CS.
Environmental Protection Agency. 1998a. Ecological Risk Management Principles for
Superfund Sites. Draft. OSWER Directive 9285-7-28P.
Environmental Protection Agency. 1998b. Guidelines for Ecological Risk Assessment.
EPA/630/R-95/002F.
Exponent. 1998. Macroinvertebrate community and fish stomach content analysis for sampling
conducted in 1997. Data report. Exponent, Bellevue, WA.
General Electric Company, 1996. Comments of General Electric Company on Phase 2 Report -
Review Copy, Further Site Characterization and Analysis, Volume 2C. Data Evaluation and
Interpretation Report. Hudson River PCBs Reassessment RI/FS. October 1996.
40
-------
General Electric Company, 1997. Comments of General Electric Company on Phase 2 Report -
Review Copy, Further Site Characterization and Analysis, Volume 2C. Data Evaluation and
Interpretation Report. Hudson River PCBs Reassessment RI/FS. February 1997.
Kaufman, K. 1996. Lives of North American Birds. New York: Houghton Mifflin Company.
Kynard, B. 1997. Life history, latitudinal patterns, and status of the shortnose sturgeon,
Acipenser brevirostrum. Envir. Biol. Fishes 48:319-334.
Landrum,, C.L., T.L. Ashwood and D.K. Cox. 1993. Belted Kingfishers as Ecological Monitors
of Contamination: A Review. Oak Ridge National Laboratory, Oak Ridge, TN. ORNL/M-2533.
March.
Long, E.R., and L.G. Morgan. 1991. The potential for biological effects of sediment-sorbed
contaminants tested in the National Status and Trends Program. NOAA Tech. Memo. NOS
OMA 52. U.S. National Oceanic and Atmospheric Administration, Seattle, WA. 175 pp.
McCarty, J.P. 1995. Effects of short-term changes in environmental conditions on the foraging
ecology and reproductive success of tree swallows, Tachycinecta bicolor. Dissertation. Cornell
University. 305 pp.
Menzie, C., M.H. Henning, J. Cura, K. Finkelsten, J. Gentile, J. Maugham, D. Mitchell, S. Petron,
B. Potocki, S. Svirsky, P. Tyler. 1996. Special report of the Massachusetts weight-of-evidence
workgroup: A weight-of-evidence approach for evaluating ecological risk. Human Ecol. Risk
Assess. 2(2):277-304.
National Marine Fisheries Service and U.S. Fish and Wildlife Service. 1985. Emergency Striped
Bass Research Study, Report for 1984.
National Marine Fisheries Service and U.S. Fish and Wildlife Service. 1994. Emergency Striped
Bass Research Study, Report for 1993.
National Marine Fisheries Services. 1997. Recovery plan for the shortnose sturgeon (Acipenser
brevirostrum). Public review draft. National Marine Fisheries Service, Shortnose Sturgeon
Recovery Team.
Nye, P. 1998. Personal communication (telephone conversation with J. Sampson, Exponent,
Bellevue, WA, on June 23, 1998 regarding the status of Hudson River bald eagles and peregrine
falcons). New York State Department of Environmental Conservation, Delmar, NY.
New York State Department of Environmental Conservation. 1993a. 20 year tends in water
quality of rivers and streams in New York state based on macroinvertebrate data, 1972-1992.
New York State Department of Environmental Conservation, Albany, NY.
-------
New York State Department of Environmental Conservation. 1997. HREMP Annual Report and
State of the Hudson Report for Period 4/1/97-3/31/98. Albany, NY. 69 pp.
Robertson, R.J., B.J. Stutchbury, and R. Cohen. 1992. Tree swallow. The Birds of North
America. 11: 1-27.
Salyer, J.C. and K.F. Langler. 1949. The eastern belted kingfisher, Megacerylealcoyon alcyon
(Linnaeus), in relation to fish management. Trans. Am. Fish Soc. 76:97-117.
Secor, D. 1997. Personal communication (telephone conversation with J. Sampson, Exponent,
Bellevue, WA 1998 regarding PCB concentrations in eels), Chesapeake Biological Laboratory,
U. of Maryland for Environmental Science, Solomons, MD.
Secord, A.L., and McCarty, J.P 1997. Polychlorinated biphenyl contamination of tree swallows
in the Upper Hudson River Valley, New York: Effects on breeding biology and implications for
other bird species. U.S. Fish and Wildlife Service, New York Field Office, Corland, NY.
Suter, G.W. 1990. Endpoints for Regional Ecological Risk Assessments. Environmental
Management. 14(1): 9-23.
Suter, G.W, II. 1993. Ecological Risk Assessment. Lewis Publishers, Chelsea, MI.
U.S. Fish and Wildlife Service. 1997. Significant Habitats and Habitat Complexes of the New
York Bight Watershed. Southern New England - New York Bight Coastal Ecosystems Program.
Charlestown, Rhode Island. 1200 pp. (on CD)
42
------- |