FINAL SECOND FIVE-YEAR REVIEW COMMENT RESPONSE FOR THE
HUDSON RIVER PCBS SUPERFUND SITE
£
<
^£DSrX
&
%
% PRO^
z
LU
e?
-------
FIVE-YEAR REVIEW COMMENT RESPONSE FOR
HUDSON RIVER PCBs SITE
TABLE OF CONTENTS
I. INTRODUCTION 1
1.1 The Second Hudson River PCBs Superfund Site Five Year Review 1
1.2 FYR Public Outreach and Engagement 1
1.3 FYR Comment Review and Response 2
II. COMMENTS 01 I SI 1)1: SCOPE OF FYR 12
III. MASTER COMMENTS AND RESPONSES 12
3.1 Data Collection 12
3.1.1 Comment 1: Additional data are required to understand the effectiveness of the
remedy 13
3.1.2 Comment 6: Consumption survey required to assess new populations eating fish..16
3.1.3 Comment 7: Despite institutional controls people are still eating fish 17
3.1.4 Comment 14: EPA must reinstate suspended solids monitoring at Waterford to
improve evaluation of PCB load to the Lower Hudson River 18
3.1.5 Comment 31: EPA's species-weighted-average approach to estimating fish
recovery rates should be updated based on the current population's diet. EPA
must modify its homologue correction and use of data in developing temporal
trends 19
3.1.6 Comment 54: There is no basis in the record for the estimate of mass discharged
to the river by GE from the capacitor plants in Hudson Falls and Fort Edward
(1.3 million pounds) 21
3.1.7 Comment 61: Significant PCB deposits left behind are in excess of other
cleanup projects 21
3.2 Modeling Analysis 22
3.2.1 Comment 9: EPA models of recovery in fish, sediment, and water are
overestimated and should be revisited 22
3.2.2 Comment 15:EPA should update its models to reflect information obtained
during dredging 24
3.2.3 Comment 26: Conceptual site model - relationship of sediment, water, fish 25
3.2.4 Comment 27: EPA's model prediction that the Upper Hudson River PCB load
to the Lower Hudson River is the primary factor for recovery of Lower Hudson
River fish is proven incorrect by this Five-Year Review 27
3.2.5 Comment 34: Water quality improvements from dredging tend to decrease with
distance downriver from dredging 28
3.2.6 Comment 44: NOAA's models demonstrate that the EPA ROD models are
flawed and should be updated to correctly reflect the role of sediment
concentrations in evaluating protectiveness of the remedy 28
i
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.2.7 Comment 55: EPA needs to update the conceptual site model (CSM) and
recalibrate and update HUDTOX and FISHRAND models in order to properly
understand the impacts of the dredging on the resultant fish concentrations 29
3.3 Assessment 30
3.3.1 Comment 2: Adjust data treatment techniques for Aroclor data 31
3.3.2 Comment 3: Assess risks of PCBs based on changes in consumption 34
3.3.3 Comment 4: Assess risks of PCBs in air 35
3.3.4 Comment 8: EPA did not investigate the potential for links to autism in the first
five-year review 37
3.3.5 Comment 11: EPA must calculate the risks of dioxin contamination (or dioxin-
like congeners) 38
3.3.6 Comment 16: EPA should finalize the study done on black bass 40
3.3.7 Comment 18: EPA should look for updated information on the toxicity of PCBs...40
3.3.8 Comment 19: EPA should qualify the 2016 spring and fall data properly
according to the impacts expected by the dredging 45
3.3.9 Comment 20: EPA should recalculate human health risks 46
3.3.10 Comment21: EPA should require GE to conduct an RI/FS of the Lower Hudson
River 47
3.3.11 Comment 24: EPA should indicate the current state of testing and analysis of
human health impacts for users of the river 49
3.3.12 Comment 28: EPA will not reach the target levels as anticipated in the ROD 50
3.3.13 Comment 29: EPA's analysis of fish data is flawed 52
3.3.14 Comment 30: EPA's analysis of water PCB trends must consider changes in
both loading conditions and comparisons of monitoring data to model
predictions when developing and interpreting trends 53
3.3.15 Comment 35: Incorporate Hudson River Reference Material in future fish
analyses 60
3.3.16 Comment 36: Increase the use of congener PCB analysis and decrease use of
Aroclor analysis 61
3.3.17 Comment 40: The larger-than-expected mass of PCBs and higher surface
sediment PCB concentrations remaining in the sediment following remediation
will extend the recovery of the river 63
3.3.18 Comment 41: Reassess air risks 69
3.3.19 Comment 43: Resolve diverging views of data with other agencies 70
3.3.20 Comment 46: Use of the non-standard protocol (without rib-in vs. rib-out)
impacts how the data can be used 71
3.3.21 Comment 49: EPA's use of the data on fish body burdens to estimate the rates
of recovery is highly subjective. EPA's analysis of trends does not support their
conclusions about the rate of decline during the period 1995-2008 75
3.3.22 Comment 50: The impact of dredging on fish tissue PCB concentrations has
passed and concentrations have now reached equilibrium. Future declines in
concentration will be very gradual and prolong the time to achieve ROD targets ...80
3.3.23 Comment 51: Changes in fish sampling locations result in data that is not
suitable for long term PCB temporal trend analysis 81
3.3.24 Comment 53: Surface PCB Concentration of the Non-Dredge Areas in RS1 has
not declined 93
ii
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.3.25 Comment 56: Sediment concentrations remaining in the river are higher than
anticipated and sediment concentration rate of decline is overestimated 95
3.3.26 Comment 57: Analysis of sediment PCB data outside the dredge areas
miscalculated the concentration and mass located in these areas 96
3.3.27 Comment 60: Data incompatibilities Lead to Errors in Interpretations 97
3.4 Remedy 99
3.4.1 Comment 10: EPA must address whether the targets for improvements in water
quality have or will be met 99
3.4.2 Comment 22: EPA should track the attainment of the interim fish tissue targets
of 0.4 mg/kg and 0.2 mg/kg PCB as it assesses the success of the remedy 100
3.4.3 Comment 33: Habitat reconstruction did not achieve the project objectives 105
3.4.4 Comment 38: EPA should compare data to ROD forecast regardless of
implementation 107
3.4.5 Comment 42: The comprehensive sediment sampling data from the SSAP
should be treated as the baseline for evaluating recovery of PCB-contaminated
cohesive sediment in non-dredged areas 107
3.4.6 Comment 47: By leaving more PCBs than anticipated in portions of the Upper
Hudson River, the remedy as implemented may not achieve the targeted
reductions in water and fish PCB concentrations in the timeframes anticipated
by EPA 110
3.4.7 Comment 48: The Lower Hudson River (LHR) fish recovery is not responding
as expected 127
3.4.8 Comment 52: Adequacy of the OM&M sediment sampling program, especially
with respect to development of post-dredging baseline information 131
3.4.9 Comment 58: EPA recognized that more PCBs were present in the Upper
Hudson River sediments than originally estimated in the 2002 ROD but did not
alter remedial activities to account for this knowledge 134
3.5 Protectiveness Determination 137
3.5.1 Comment 12: EPA must consider protection of natural resources as fish
consumption advisories do not protect environmental receptors 138
3.5.2 Comment 13: EPA must include a site-wide protectiveness statement in
accordance with the guidance 139
3.5.3 Comment 32: "Will be protective" is not an appropriate determination for the
Hudson River PCBs Site. "Will be protective" is only appropriate when a
remedy is still "under construction." 140
3.5.4 Comment 37: Institutional controls should not be a part of the remedy 141
3.5.5 Comment 45: The remedy is not protective 142
3.5.6 Comment 59: Hudson River PCB concentrations will not reach the target levels
anticipated in the ROD and EPA is claiming a short-term impact to the fish
from recent dredging when such impacts should be negligible 145
3.6 FYR Process and Public Engagement 147
3.6.1 Comment 5: Consider the risks to Environmental Justice communities 147
3.6.2 Comment 17: EPA should ensure that there is adequate outreach to the diverse
communities in the Lower Hudson River 150
3.6.3 Comment 23: EPA should review all the data when developing the Five-Year
Review report in accordance with the guidance 151
iii
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.6.4 Comment 25: EPA should update the Community Involvement Plan 153
3.6.5 Comment 39: Public Involvement in the Five-Year Review Process 154
IV. REFERENCES 156
APPENDIX A - LIST OF COMMENTERS (INDEX)
- Government, Agencies, Organizations and Businesses/Corporations
- Individuals
APPENDIX B - DEFERRAL STATEMENT - Supporting Technical Information
APPENDIX C - TECHNICAL MEMORANDUM - Evaluation of Field, Kern and Rosman (2016)
Emulation Model
APPENDIX D - Special Study - Black Bass Fillet Tissue With and Without Ribs
iv
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
LIST OF TABLES
Table 1 FYR Master Comment Response List 4
Table Al-7 (as presented in Appendix 1 of the FYR report): Average Annual Water
Column Tri+ PCB ROD and Updated MNA Forecasts for 1998-2008.
Augmented by Pre-MNA Calibration Results for 1995-1998 55
Table Al-7a Average Annual Water Column Tri+ PCB ROD and Updated MNA Forecasts
for 1998-2008. Augmented by Pre-MNA Calibration Results for 1995-1998,
Assuming Constant Upstream Boundary PCB Loadings, 2000-2008 55
Table 49-1 Power to detect 8 percent or 5 percent annualized change with monitoring
data for 4 years and 8 years 80
Table 51-1 Impact of Sampling Locations on Temporal Fish Tissue Trends in RS 1 87
Table 22-la Comparison of Lower Hudson PCB Concentrations in Fish With Interim
Targets and Remedial Goals, Total PCB - Homologue Equivalent Basis 102
Table 22-lb Comparison of Lower Hudson PCB Concentrations in Fish With Interim
Targets and Remedial Goals, Total PCB - Aroclor Basis- 103
Table 22-2a Comparison of Upper Hudson PCB Concentrations in Fish With Interim
Targets and Remedial Goals, Total PCB - Homologue Equivalent Basis 104
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
v
April 2019
-------
LIST OF FIGURES
Figure 18-1 IRIS Assessment Development Process 44
Figure 30-1 Measured concentrations at Ft. Edward/Rogers Island and Thompson Island
Dam (TID) between 1991 and 2008 59
Figure 30-2 Monthly average Tri+ PCB load at Ft. Edward/Rogers Island and Thompson
Island Dam (TID) between 1991 and 2008 60
Figure 40-la Variability in surface sediment Tri+ PCB concentrations with distance from
dredging boundary in RS 1 67
Figure 40-lb Variability in surface sediment Tri+ PCB concentrations with distance from
dredging boundary in RS 2 68
Figure 40-lc Variability in surface sediment Tri+ PCB concentrations with distance from
dredging boundary in RS 3 69
Figure 51-la Fish sampling stations in River Section 1 87
Figure 51-lb Fish sampling stations in River Section 3 88
Figure 5 l-2a Variation of rates of decline in fish tissue concentration with station inclusion
for largemouth bass at River Section 1 89
Figure 51 -2b Variation of rates of decline in fish tissue concentration with station inclusion
for pumpkinseed at River Section 1 90
Figure 51-3 Variation of rates of decline in fish tissue concentration with station inclusion
for pumpkinseed at Albany/Troy 91
Figure 51-4 Variation of rates of decline in fish tissue concentration with restriction on
species length for pumpkinseed at River Section 3 92
Figure 51-5 Variation of rates of decline in fish tissue concentration with restriction on
season - Yellow Perch at Albany/Troy 93
Figure 53-1 Cumulative Distribution Function (CDF) Plot for 0-2 inch Sediment Samples
for RS 1 In Non-dredged Areas 95
Figure 47-la SSAP Surface Sediment Tri+ PCB Concentrations us. Distance from
Dredging Boundary, River Section 1, 2002-2005 (0-2 in. Samples) 115
Figure 47-lb OM&M Surface Sediment Tri+ PCB Concentrations us. Distance from
Dredging Boundary, River Section 1, 2016-2017 (0-2 in. Samples) 116
Figure 47-2a SSAP Surface Sediment Tri+ PCB Concentrations us. Distance from
Dredging Boundary, River Section 2, 2002-2005 (0-2 in. Samples) 117
Figure 47-2b OM&M Surface Sediment Tri+ PCB Concentrations vs. Distance from
Dredging Boundary, River Section 2, 2016-2017 (0-2 in. Samples) 118
Figure 47-3a SSAP Surface Sediment Tri+ PCB Concentrations us. Distance from
Dredging Boundary, River Section 3, 2002-2005 (0-2 in. Samples) 119
vi
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Figure 47-3b OM&M Surface Sediment Tri+ PCB Concentrations us. Distance from
Dredging Boundary, River Section 3, 2016-2017 (0-2 in. Samples) 120
Figure 47-5 SSAP Maximum Sediment Tri+ PCB Concentration us. Distance from
Dredging Boundary, River Section 1, 2002-2005 (Maximum value in 0-12
in. interval) 121
Figure 47-7 SSAP Maximum Sediment Tri+ PCB Concentration us. Distance from
Dredging Boundary, River Section 3, 2002-2005 (Maximum value in 0-12
in. interval) 123
Figure 47-8 SSAP Tri+ PCB MP A vs. Distance from Dredging Boundary, River Section
1, 2002-2005 124
Figure 47-9 SSAP Tri+ PCB MP A vs. Distance from Dredging Boundary, River Section
2, 2002-2005 125
Figure 47-10 S S AP Tri+ PCB MP A vs. Distance from Dredging Boundary, River Section
3, 2002-2005 126
vii
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
LIST OF ABBREVIATIONS AND ACRONYMS
A1016
Aroclor 1016
A1221
Aroclor 1221
A1242
Aroclor 1242
A1254
Aroclor 1254
ADD
Average Daily Dose
ARAR
Applicable or Relevant and Appropriate Requirement
AT
Albany-Troy
AT SDR
Agency for Toxic Substances and Disease Registry
BERA
Baseline Ecological Risk Assessment
BMP
Baseline Monitoring Program or best management practice
BSAF
Biota Sediment Accumulation Factor
BW
body weight
CAG
Community Advisory Group
CAM
Corrective Action Memo
CCC
Criteria Continuous Concentration
CCE
Cornell Cooperative Extension
CDC
Centers for Disease Control and Prevention
CDF
cumulative distribution function
CERCLA
Comprehensive Environmental Response, Compensation, and Liability
Act
CFR
Code of Federal Regulations
cfs
cubic feet per second
CIP
Community Involvement Plan
cm
centimeter
coc
chemical of concern
COPC
chemical of potential concern
CS
Catskill
CSF
Cancer Slope Factor
CSM
Conceptual Site Model
viii
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
CT central tendency
CTE Central Tendency Exposure (Exposed)
CU Certification Unit; dredging target area within which performance
metrics were applied
DAD Dredge Area Delineation
DDS Downstream Deposition Study
DDT di chl orodipheny ltri chl oroethane
DEC see NYSDEC
DoC depth of contamination
DOH see NYSDOH
DQO Data Quality Objective(s)
dw dry weight
EDI equal discharge increment
EPA see USEPA
EPC Exposure Point Concentration(s)
EPS Engineering Performance Standards
ERRD EPA Region 2's Emergency and Remedial Response Division
ERT Environmental Response Team
FCA Fish Consumption Advisory(ies)
FIR food ingestion rate
FS Feasibility Study
FISHRAND mechanistic, time-varying, fish tissue contaminant bioaccumulation
model
ft foot (or feet)
FWQC Federal Water Quality Criteria
FWS Fish and Wildlife Service
FYR Five-Year Review (unless otherwise indicated, the Second Five-Year
Review report initially released as "Proposed" in June 2017)
g/cm3 grams per cubic centimeter
g/day gram per day
g/m2 gram per square meter
ix
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
GAC
GC/ECD
GCL
GE
HDC
HHRA
HI
HQ
HQ-OSRTI
HUDTOX
I ARC
IC
IRIS
kg
Kg/day or kg/d
Kg/month
Kg/yr
Km
Kow
L/day
Lb
LCL
LHR
LOAEL
LPCB
M1668
M8082
granular activated carbon
Gas Chromatography/ Electron Capture Detection method
geosynthetic clay liner
General Electric Company
high-density core
Human Health Risk Assessment
Hazard Index
Hazard Quotient
EPA Headquarters' Office of Superfund Remediation and Technology
Innovation
Upper Hudson River Toxic Chemical Model; a mechanistic, numerical
chemical fate and transport model for water and sediment
International Agency for Research on Cancer
Institutional Control(s)
Integrated Risk Information System
kilogram
kilogram per day
kilogram per month
kilograms per year
kilometers
octanol/water partition coefficient
liters per day
Pound
Lower Confidence Limit
Lower Hudson River
Lowest Observed Adverse Effect Level
lipid normalized PCBs
EPA high-resolution gas chromatography / mass spectrometry
(HRGC/HRMS) congener-based PCB analysis method; version 1668c of
the method (Ml668c) has been used primarily since 2016
EPA gas chromatography (GC) Aroclor-based PCB analysis method
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
x
April 2019
-------
MADIS Multiple Aliquot Depth Integrated Sampler
MCA Monte Carlo Analysis
MCL maximum contaminant level
mGBM modified Green Bay Method; gas chromatography / electron capture
detector (GC/ECD) congener-based PCB analysis method adapted by GE
for the Hudson River from one originally developed for the Great Lakes
mg/kg milligram per kilogram
mg/kg-ww milligram per kilogram wet weight
MNA Monitored Natural Attenuation
MNA1 baseline MNA scenario
MNA2 "updated" MNA scenario used in Field et al (2016)
MNR Monitored Natural Recovery
MPA Mass Per Unit Area; typically expressed as grams per square meter
(g/m2)
MPUV mass per unit volume
NAPL non-aqueous phase liquid
NCP National Oil and Hazardous Substances Pollution Contingency Plan
ND Northumberland Dam
ng/L nanogram per Liter
ng/m3 nanograms per cubic meter
NHANES National Health and Nutrition Examination Survey
NIST National Institutes of Standards and Technology
NLOM non-lipid organic matter
NOAA National Oceanic and Atmospheric Administration
NOAEL No Observed Adverse Effect Level
NPL National Priorities List
NYC New York City
NYS New York State
NYSCC New York State Canals Corporation
NYSDEC New York State Department of Environmental Conservation
NYSDOH New York State Department of Health
xi
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
NYSDOT
O&M
OM&M
OSWER
OU
PAH
PCB
PCRDMP
PE
PKSD
ppb
ppm
ppt
PRA
PRG
PRP
PSCP
PWS
QA
QAPP
QoLPS
RA
RAM
RAMP
RAWP
RAO
REM 3/10/Select
RfC
RfD
New York State Department of Transportation
Operations and Maintenance
Operations, Maintenance, and Monitoring
Office of Solid Waste and Emergency Response
Operable Unit; an officially designated portion of a CERCLA site for
investigation and remediation purposes
polycyclic aromatic hydrocarbon
Polychlorinated Biphenyl
Post-Construction Remnant Deposit Monitoring Plan
Performance Evaluation
Pumpkinseed
parts per billion
parts per million
parts per trillion
probabilistic analysis
Preliminary Remediation Goal
Potentially Responsible Party
Performance Standards Compliance Plan
public water supplies
Quality Assurance
Quality Assurance Project Plan
Quality of Life Performance Standard
Remedial Action
Remedial Action Monitoring
Remedial Action Monitoring Program
Remedial Action Work Plan
Remedial Action Objective
Removal Criteria by respective River Sections as stated in the ROD
Reference Concentration
Reference Dose
xii
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
RI Remedial Investigation
RI/FS Remedial Investigation and Feasibility Study
RM River Mile
RME Reasonable Maximum Exposure (Exposed)
ROD Record of Decision
RPM Remedial Project Manager
RS River Section
SAV Submerged Aquatic Vegetation
SEDC Supplemental Engineering Data Collection
Site Hudson River PCBs Superfund Site
SMR standardized mortality ratio
SOP Standard Operating Procedure
SOW statement of work
SRM Standard Reference Material
SSAP Sediment Sampling and Analysis Program
TBC To Be Considered; criteria explored as potentially germane to remedial
decision-making in parallel with ARARs
TCDD 2,3,7,8-tetrachlorodibenzo-p-dioxin
TID or TD Thompson Island Dam
TIP Thompson Island Pool
TOC Total Organic Carbon
TPCB Total PCB
TPCBArocior PCB compounds measured as Aroclors
TPCBhe PCB compounds measured as homolog equivalents
Tri+ PCBs PCBs containing three or more chlorines
TRV toxicity reference values
TSCA Toxic Substances and Control Act
TSS Total Suspended Solids
UCL Upper Confidence Limit
UE Unrestricted Exposure
xiii
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
^g/L
microgram per liter
|ig/m3
micrograms per cubic meter
UHR
Upper Hudson River
USEPA
United States Environmental Protection Agency
USGS
Unites States Geological Survey
UU
Unlimited Use
WCS
Waste Control Specialists, LLC
WIR
water ingestion rates
WQ
Water Quality
WW
wet weight
xiv
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
I. INTRODUCTION
1.1 The Second Hudson River PCBs Superfund Site Five Year Review
The purpose of the Second Five-Year Review (FYR) for the Hudson River PCBs Superfund Site
is to determine if the Superfund cleanup remedy is working as intended and is protective of human
health and the environment. Superfund law requires that five-year reviews be performed when a
cleanup action leaves some hazardous substances on a site at levels that do not allow for unlimited
use and unrestricted exposure. These reviews are required every five years from the start of
construction of the cleanup action.
The second Hudson River PCBs Superfund Site FYR was led by United States Environmental
Protection Agency (EPA) Project Director, Gary Klawinski, and EPA Office of Superfund
Remediation and Technology Innovation (OSRTI) - Environmental Response Team (ERT)
manager Marc S. Greenberg, Ph.D. Participants also included other EPA staff within EPA Region
2's Emergency and Remedial Response Division (ERRD) and EPA Headquarters' Office of
Superfund Remediation and Technology Innovation (HQ-OSRTI) as appropriate.
EPA's Comprehensive Five-Year Review Guidance states that, for complex projects, a
multidisciplinary five-year review team of experts may be needed to adequately review the
protectiveness of the remedy. This five-year review included a rigorous and unprecedented
stakeholder and community engagement process. Because of the complexity of the Hudson River
Polychlorinated Biphenyls (PCBs) Operable Unit (OU) 2 remediation, EPA assembled a FYR
team that included representatives of state agencies, federal agencies, natural resource trustees,
Community Advisory Group members, and EPA subject matter experts. The team provided input
on remedy implementation and performance based on information that includes environmental
data and document review. Team members regularly and actively participated in meetings
throughout the review period.
Three public workshops were held during the FYR to provide information about the review and
the review process to the public. EPA also accepted public comments on the proposed FYR report.
Written correspondence was received during the FYR from the public, multiple State and Federal
agencies, environmental groups, and elected officials.
1.2 FYR Public Outreach and Engagement
Throughout the five-year review process, EPA provided various opportunities for public
participation. Before the initiation of the formal public comment period for the second FYR, the
public was notified of and invited to participate in the five-year review process via press releases,
public workshops, the Hudson River Listserv and EPA's Hudson River PCBs Superfund site
webpage: www.epa.gov/hudson. Additionally, EPA provided updates on the FYR process and
report to stakeholders represented by the project Community Advisory Group (CAG). These
meetings were and are open to the public. As mentioned in Section 1.1, three public workshops
were also held at varying locations in the proj ect area during the five-year review process to discuss
the purpose of the review and the timeline, and to provide status updates and an opportunity for
members of the public to provide input and ask questions.
1
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Although EPA does not typically seek public comment on a FYR report, EPA initiated a formal
public comment period in concert with the release of the proposed FYR report on June 1, 2017.
The comment period was originally set to end on June 30, 2017. Shortly after the release, on June
8, 2017, EPA extended the public comment period until September 1, 2017 in response to requests
from several stakeholders.
During the public comment period, as mentioned above EPA hosted three public information
meetings, in Upper Hudson River and Lower Hudson River communities, and in New York City.
EPA discussed the purpose, scope and findings of the five-year review and answered questions
from the public during those meetings:
• Lower River: June 28, 2017 6 p.m. - 8 p.m. at the Poughkeepsie Grand Hotel in
Poughkeepsie, New York
• Upper River: July 19, 2017 6 p.m. - 8 p.m. at the Saratoga Hilton in Saratoga, New York
• New York City: August 9, 2017 6 p.m. - 8 p.m. John Jay College of Criminal Justice in
New York City, New York
EPA reviewed and considered all written comments provided during the public information
meetings, as well as written comments received during the public comment period. By the close
of the comment period, EPA had received 1,968 discrete submissions of comments. Of the 1,968
submissions, 51 were from government (state and federal) agencies, other organizations and
businesses/corporations. 529 were unique submittals from individuals and 1,388 were additional
letters based on templates provided by organizations. An index listing the names of commenters
is attached to this report as Appendix A.
1.3 FYR Comment Review and Response
All comments submitted to EPA during the public comment period were carefully considered. To
ensure a complete and comprehensive evaluation and response to the FYR comments, all comment
documents were reviewed and catalogued within a database system. To manage the comments,
each comment letter was divided into segments that each captured a unique theme or topic with
respect to the FYR process or report. Each of the segments was also assigned representative
keywords (or key phrases) and entered into an electronic database for sorting and processing. The
segments were organized for content and then assigned to review by subject matter experts (SME).
All unique segments were identified and were individually adopted as a "master comment," or
were consolidated with other similarly themed segments (addressing similar issues) into a single
master comment. EPA prepared a response for each master comment.
A quality assurance program was implemented to verify that the full body of segments were
reviewed and categorized appropriately. All segments identified to be within the scope of the FYR
report and/or process were consolidated into master comments. The quality assurance program
was also used to verify that these topics were accurately represented in the master comments and
the responses are technically complete. EPA received some comments and opinions that were
outside the scope of the FYR report and process. These comments have been summarized briefly
in Section 2 and, as appropriate, are not addressed as part of the master comments and responses.
2
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Responses were informed by and drawn from:
• material presented in the second FYR report
• previous project reports and other literature
• the experience of other remedial projects and individuals
• EPA policy
• technical analyses that were performed specifically to address comments or questions
raised during the public comment period.
For ease of review and understanding, the comments and responses were grouped into 6 categories:
• Data Collection
• Modeling Analysis
• Assessment
• Remedy
• Protectiveness Determination
• FYR Process and Public Outreach and Engagement
These categories were selected to reflect the logical progression of the FYR process from data
collection to determination. Some master comments and responses pertain to two or more of these
categories. As an example, a comment pertaining to the evaluation of water samples, could apply
to both data collection (how that sample was collected and why) and assessment (how that
information was used to inform the FYR report). EPA placed the comments and responses into
the category where the comment and response are most focused and has provided content and
references to other categories as appropriate.
Table 1, below, provides a list of the master comments as organized into the categories, and
indicates which categories apply to the comment.
3
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Table 1 FYR Master Comment Response List
Topics Covered by Comment
Res
ponse
a
4)
o
U
15
=
<
o/j
=
S
to
¦o
=
>
Process and
ic Engagement
Comment
Number
Comment Title
Response Found
in Section
C8
C8
Q
¦a
©
s
4)
to
to
<
s
O
a.
FYR
Publi
1
Additional data are
3.1 Data
Collection
X
X
X
X
required to understand the
effectiveness of the remedy
2
Adjust data treatment
techniques for Aroclor data
3.3 Assessment
X
3
Assess risks of PCBs based
3.3 Assessment
X
on changes in consumption
4
Assess risks of PCBs in air
3.3 Assessment
X
5
Consider the risks to
Environmental Justice
communities
3.6 FYR Process
X
X
6
Consumption survey
required to assess new
populations eating fish
3.1 Data
Collection
X
X
7
Despite institutional controls
people are still eating fish
3.1 Data
Collection
X
X
8
EPA did not investigate the
potential for links to autism
3.3 Assessment
X
in the first five-year review.
9
EPA models of recovery in
fish, sediment, and water are
overestimated and should be
revisited
3.2 Modeling
Analysis
X
10
EPA must address whether
the targets for improvements
in water quality have or will
be met
3.4 Remedy
X
11
EPA must calculate the risks
of dioxin contamination (or
dioxin-like congeners)
3.3 Assessment
X
12
EPA must consider
protection of natural
resources as fish
consumption advisories do
3.5 Protectiveness
X
X
4
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Comment
Number
Comment Title
Response Found
in Section
Topics Covert
Res
;d by Comment
ponse
Data Collection
Modeling Analysis
Assessment
Remedy
Protectiveness
FYR Process and
Public Engagement
not protect environmental
receptors
13
EPA must include a site-
wide protectiveness
statement in accordance with
the guidance
3.5 Protectiveness
X
X
14
EPA must reinstate
suspended solids monitoring
at Waterford to improve
evaluation of PCB load to
the Lower Hudson River
3.1 Data
Collection
X
X
15
EPA should update its
models to reflect information
obtained during dredging
3.2 Modeling
Analysis
X
16
EPA should finalize the
study done on black bass
3.3 Assessment
X
X
17
EPA should ensure that there
is adequate outreach to the
diverse communities in the
Lower Hudson River
3 .6 FYR Process
X
18
EPA should look for updated
information on the toxicity
of PCBs
3.3 Assessment
X
19
EPA should qualify the 2016
spring and fall data properly
according to the impacts
expected by the dredging
3.3 Assessment
X
20
EPA should recalculate
human health risks
3.3 Assessment
X
X
21
EPA should require GE to
conduct an RI/FS of the
Lower Hudson River
3.3 Assessment
X
22
EPA should track the
attainment of the interim fish
tissue targets of 0.4 mg/kg
and 0.2 mg/kg PCB as it
3.4 Remedy
X
5
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Comment
Number
Comment Title
Response Found
in Section
Topics Covert
Res
;d by Comment
ponse
Data Collection
Modeling Analysis
Assessment
Remedy
Protectiveness
FYR Process and
Public Engagement
assesses the success of the
remedy.
23
EPA should review all the
data when developing the
Five-Year Review report in
accordance with the
guidance
3 .6 FYR Process
X
X
X
24
EPA should indicate the
current state of testing and
analysis of human health
impacts for users of the river
3.3 Assessment
X
25
EPA should update the
Community Involvement
Plan
3.6 FYR Process
X
26
Conceptual site model -
relationship of sediment,
water, fish
3.2 Modeling
Analysis
X
27
EPA's model prediction that
the Upper Hudson River
PCB load to the Lower
Hudson River is the primary
factor for recovery of Lower
Hudson River fish is proven
incorrect by this Five-Year
Review
3.2 Modeling
Analysis
X
28
EPA will not reach the target
levels as anticipated in the
ROD
3.3 Assessment
X
X
29
EPA's analysis of fish data is
flawed
3.3 Assessment
X
30
EPA's analysis of water PCB
trends must consider changes
in both loading conditions
and comparisons of
monitoring data to model
3.3 Assessment
X
X
X
6
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Comment
Number
Comment Title
Response Found
in Section
Topics Covert
Res
;d by Comment
ponse
Data Collection
Modeling Analysis
Assessment
Remedy
Protectiveness
FYR Process and
Public Engagement
predictions when developing
and interpreting trends
31
EPA's species-weighted-
average approach to
estimating fish recovery
rates should be updated
based on the current
population's diet. EPA must
modify its homologue
correction and use of data in
developing temporal trends.
3.1 Data
Collection
X
X
X
32
"Will be protective" is not an
appropriate determination for
the Hudson River PCBs Site.
"Will be protective" is only
appropriate when a remedy
is still "under construction."
3.5 Protectiveness
X
33
Habitat reconstruction did
not achieve the project
objectives
3.4 Remedy
X
X
34
Water quality improvements
from dredging tend to
decrease with distance
downriver from dredging
3.2 Modeling
X
35
Incorporate Hudson River
Reference Material in future
fish analyses
3.3 Assessment
X
X
36
Increase the use of congener
PCB analysis and decrease
use of Aroclor analysis
3.3 Assessment
X
37
Institutional controls should
not be a part of the remedy
3.5 Protectiveness
X
38
EPA should compare data to
ROD forecast regardless of
implementation
3.4 Remedy
X
X
X
7
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Comment
Number
Comment Title
Response Found
in Section
Topics Covert
Res
;d by Comment
ponse
Data Collection
Modeling Analysis
Assessment
Remedy
Protectiveness
FYR Process and
Public Engagement
39
Public Involvement in the
Five-Year Review Process
3 .6 FYR Process
X
40
The larger-than-expected
mass of PCBs and higher
surface sediment PCB
concentrations remaining in
the sediment following
remediation will extend the
recovery of the river
3.3 Assessment
X
X
41
Reassess air risks
3.3 Assessment
X
42
The comprehensive sediment
sampling data from the
SSAP should be treated as
the baseline for evaluating
recovery of PCB -
contaminated cohesive
sediment in non-dredged
areas
3.4 Remedy
X
X
X
X
43
Resolve diverging views of
data with other agencies
3.3 Assessment
X
X
44
NOAA's models
demonstrate that the EPA
ROD models are flawed and
should be updated to
correctly reflect the role of
sediment concentrations in
evaluating protectiveness of
the remedy
3.2 Modeling
Analysis
X
X
45
The remedy is not protective
3.5 Protectiveness
X
X
X
X
46
Use of the non-standard
protocol (without rib-in vs.
rib-out) impacts how the data
can be used
3.3 Assessment
X
47
By leaving more PCBs than
anticipated in portions of the
Upper Hudson River, the
3.4 Remedy
X
X
8
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Comment
Number
Comment Title
Response Found
in Section
Topics Covert
Res
;d by Comment
ponse
Data Collection
Modeling Analysis
Assessment
Remedy
Protectiveness
FYR Process and
Public Engagement
remedy as implemented may
not achieve the targeted
reductions in water and fish
PCB concentrations in the
timeframes anticipated by
EPA
48
The Lower Hudson River
(LHR) fish recovery is not
responding as expected
3.4 Remedy
X
49
EPA's use of the data on fish
body burdens to estimate the
rates of recovery is highly
subjective. EPA's analysis of
trends does not support their
conclusions about the rate of
decline during the period
1995-2008
3.3 Assessment
X
X
X
50
The impact of dredging on
fish tissue PCB
concentrations has passed
and concentrations have now
reached equilibrium. Future
declines in concentration
will be very gradual and
prolong the time to achieve
ROD targets
3.3 Assessment
X
X
X
51
Changes in fish sampling
locations result in data that is
not suitable for long term
PCB temporal trend analysis
3.3 Assessment
X
X
X
52
Adequacy of the OM&M
sediment sampling program,
especially with respect to
development of post-
dredging baseline
information
3.4 Remedy
X
X
X
9
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Comment
Number
Comment Title
Response Found
in Section
Topics Covert
Res
;d by Comment
ponse
Data Collection
Modeling Analysis
Assessment
Remedy
Protectiveness
FYR Process and
Public Engagement
53
Surface PCB Concentration
of the Non-Dredge Areas in
RSI has not declined
3.3 Assessment
X
X
54
There is no basis in the
record for the estimate of
mass discharged to the river
by GE from the capacitor
plants in Hudson Falls and
Fort Edward (1.3 million
pounds)
3.1 Data
Collection
X
X
55
EPA needs to update the
conceptual site model (CSM)
and recalibrate and update
HUDTOX and FISHRAND
models in order to properly
understand the impacts of the
dredging on the resultant fish
concentrations
3.2 Modeling
Analysis
X
X
56
Sediment concentrations
remaining in the river are
higher than anticipated and
sediment concentration rate
of decline is overestimated
3.3 Assessment
X
X
X
57
Analysis of sediment PCB
data outside the dredge areas
miscalculated the
concentration and mass
located in these areas
3.3 Assessment
X
X
58
EPA recognized that more
PCBs were present in the
Upper Hudson River
sediments than originally
estimated in the 2002 ROD
but did not alter remedial
activities to account for this
knowledge
3.4 Remedy
X
X
X
10
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Comment
Number
Comment Title
Response Found
in Section
Topics Covert
Res
;d by Comment
ponse
Data Collection
Modeling Analysis
Assessment
Remedy
Protectiveness
FYR Process and
Public Engagement
59
Hudson River PCB
concentrations will not reach
the target levels anticipated
in the ROD and EPA is
claiming a short-term impact
to the fish from recent
dredging when such impacts
should be negligible
3.5 Protectiveness
X
60
Data incompatibilities Lead
to Errors in Interpretations
3.3 Assessment
X
X
61
Significant PCB deposits left
behind are in excess of other
cleanup projects
3.1 Data
Collection
X
X
11
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
II. COMMENTS OUTSIDE SCOPE OF FYR
EPA received many comments on the FYR, with some being identified as outside the scope of the
FYR. General descriptions of these comments are provided below.
• Some commenters noted that General Electric (GE) should not have been allowed to
contaminate the river.
• Commenters wrote about past final decisions made by EPA on the project.
• Some commented about the Hudson River floodplain (including the Old Champlain
Canal). The floodplain project is a separate operable unit and is not part of the upper river
remedy or this FYR.
• Comments were received on the participation of other federal or state agencies in the
remedy review. Those comments are best addressed by those agencies.
• Some commenters asked EPA to give New York State lead agency status for the project,
while others discussed the disagreement on data interpretations between agencies, and
others wrote about the impact of the five-year review on the trustees' NRDA claims.
• Comments were received on other environmental issues that impact the Lower Hudson
River, such as transport of oil and gas on the river and the proposed closing of the Indian
Point nuclear facility.
• Some comments were received on other unrelated project reports written by EPA.
• Some comments were received on Operation, Maintenance, and Monitoring (OM&M)
work plans. Those comments will be considered by EPA as those plans are developed and
finalized. Responses have been provided for comments pertaining to the data collected as
part of the OM&M program during the FYR period.
• Some comments discussed the timing of the certification of completion of the remedial
action under the 2006 consent decree with GE with respect to the timing of the completion
of the five-year review. The determination of protectiveness as part of this FYR is
independent of the certification of completion of the remedial action.
III. MASTER COMMENTS AND RESPONSES
This section contains the 61 master comments and responses. Note that some comments and
responses may also be applicable to other categories, as identified in Table 1 of this document.
3.1 Data Collection
This section includes comments and responses concerning what data was collected (e.g., fish,
water, and sediment), how it was collected, and the need for additional data collection and
consideration.
12
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.1.1 Comment 1: Additional data are required to understand the effectiveness of the
remedy
Comment
Commenters raised concerns regarding whether the existing data are sufficient to determine if fish
will recover in the timeframes laid out in the ROD. Commenters requested that EPA require
additional studies and data collection within the Upper and Lower Hudson to provide the evidence
needed to determine future fish PCB concentrations and the subsequent risks to human health. A
more robust fish and sediment sampling program than that proposed by EPA was recommended,
with a focus on sample size and segmentation, and increased spatial resolution. Commenters point
out that the prior sampling program focused on determining PCB concentrations in fish by river
section but sampling each river pool would be more effective in determining accurate
contamination concentrations, because resident fish integrate their exposure within smaller areas
compared to larger river sections.
One commenter suggested that the post-remedial fish PCB concentrations are expected to be
higher than EPA anticipated at the time of remedy selection. The commenter further questioned
whether the deviation from the forecast trends were due to the higher-than-expected post-remedy
absolute sediment concentrations or the less-than-targeted relative reduction in sediment
concentration.
A commenter indicated that the lesser degree of improvement in surface sediment PCB
concentrations should be reflected in less improvement in fish PCB concentrations. The
commenter said that since surface sediment concentrations in RS 3 only improved by 4 percent
based on the dredging, then PCB levels in fish in RS 3 should only immediately improve ~ 4
percent as a direct result of the dredging, and fish in the Lower River, where no sediment
remediation was done, should show little additional improvement as a result of the remedy.
One commenter urged EPA to expressly include at least the following benchmarks as a way to
measure the success or failure of the remedy to protect human health and the environment both in
subsequent FYRs and as more data becomes available each year:
1. Species-weighted fish fillet Upper Hudson average PCB concentrations must be at or
below 0.4 mg/kg within five years of the completion of dredging (by 2020);
2. Species-weighted fish fillet Upper Hudson average PCB concentrations must be at or
below 0.2 mg/kg within sixteen years of the completion of dredging (by 2031);
3. Largemouth bass, whole body PCB concentrations must be within EPA's recalculated
forecast range of 0.2 mg/kg to 0.07 mg/kg for RS 1,2, and 3 within 23 years of the
completion of dredging (by 2038); and
4. Species-weighted fish fillet average PCB concentrations in RS 1 must be at or below 0.05
mg/kg within 43 years of the completion of dredging (by 2058).
13
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Response
EPA asserts that the existing data, representing limited post-dredging monitoring, are not sufficient
to assess the rate of post-dredging recovery. While the 2016 OM&M data characterize the fish
body burdens in the first year post-dredging, these fish may still be impacted by dredging activities.
Many of the sport fish included in the 2016 data set were obtained as part of the spring sampling
event (the normal time for their collection). These data represent adult fish, typically several years
old. Thus, the measured PCB levels in the spring 2016 fish were likely still influenced by the
dredging activities of 2015 and earlier, and therefore, are not exclusively part of the post-dredging
condition. The ROD anticipated at least a one-year equilibrium period of the system in response
to remedial activities. Sampling data from another remediated PCB site shows that this
equilibration period could be as many as 3 to 5 years following intrusive activities (AECOM,
2012). During the equilibration period, PCB concentrations in fish and the water column can
exhibit wide variation with little trend. After this equilibration period, the system is expected then
to follow a more predictable natural recovery. To determine how many years of data will be needed
to accurately identify post-dredging recovery rate and inform the OM&M sampling program
design, EPA performed statistical power calculations using the fish body burden data from 1998
to 2008, part of the MNA period that preceded the remediation. These calculations indicate that
approximately eight or more years of data will be required to accurately identify and confirm the
post-remedy rate of decline for each species in each river section.
Commenters suggested that EPA focus its future monitoring on a reach-by-reach basis rather than
for the whole river or by river section. While the ROD addresses the river on a section basis, EPA
agrees that there is value in also assessing reaches (pools) within the river. EPA notes that the
extensive surface sediment data sets collected by EPA/GE and NYSDEC do not indicate the
continued presence of "toxic hotspots." Nonetheless, EPA will include the reach consideration in
the continued assessment of the recovery of the river. All reaches will be monitored as part of the
surface sediment sampling program, thus also providing recovery information on a reach basis.
Depending on data availability and the degree of similarity of PCB concentrations, reaches may
be grouped as part of long-term trend analysis. Periodic reach-specific fish monitoring will also
be done as appropriate to confirm fish recovery is consistently occurring throughout the Upper
Hudson. Additionally, EPA will coordinate with NYSDEC and NYSDOH regarding location-
specific sampling modifications, as necessary to inform decisions regarding adjusting fishing
restrictions and advisories.
Individual fish species will respond to contaminant exposures in different ways depending on their
foraging strategies and life histories. It is important to note that any individual fish (and, more
broadly, any individual fish species) will achieve target levels at different times given: 1)
variability in actual exposures; 2) highly localized exposures; 3) the importance of sediment vs.
water exposure pathways, which can vary over time due to prey availability and natural variability
in exposure conditions; 4) variability in lipid content of fish and prey items; and 5) variability in
consumption of specific prey items and PCB concentrations in prey. Thus, while EPA will
continue to monitor the declines in fish PCB levels relative to the ROD's interim targets and the
remediation goals for fish fillet, the post-dredging model projections included in the ROD are not
precise dates that must be met in order for the remedy to be successful.
14
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Remedy protectiveness was evaluated in the ROD by comparing predicted fish tissue
concentration trajectories over time under different remedial alternatives. As noted in the ROD,
different target levels will be achieved at different times depending on the species and river section
(or river pool given species-specific foraging strategies and life histories). Evaluation of post-
dredging fish concentration trajectories is the best method for assessing recovery over time due to
uncertainty associated with long-term model predictions.
Regarding the amount of PCBs left behind and the reduction in surface sediment concentrations,
EPA notes in Table A4-5 of Appendix 4 of the FYR report that the removal activity alone achieved
substantially better reduction in surface PCB levels in RS 1, the targeted level in RS 3, and
somewhat less than expected reduction in RS 2. However, incorporating the surface sediment data
collected in 2016, the table shows greater reductions in all three river sections. It is likely that the
combination of the remedy itself, natural attenuation since the 2002 to 2005 sampling, and
additional attenuation since the remedy was important in achieving this reduction. These
observations are further supported and refined by the more extensive sampling conducted by
NYSDEC; see EPA's April 2019 Technical Memorandum entitled "Technical Memorandum
Evaluation of 2016 EPA/GE and 2017NYSDEC Surface Sediment Data" (www.epa.gov/hudson).
Given that the average surface sediment concentrations were estimated to decline more than 80
percent in all three river sections (based on comparison of the 2002 to 2005 SSAP and the 2016
OM&M sediment sampling datasets), EPA does not agree that an update to the Conceptual Site
Model (CSM) is needed. EPA will continue to monitor sediment concentrations closely.
EPA will continue to monitor PCB levels in fish and assess the corresponding declines in fish
tissue concentrations over time. EPA disagrees with a commenter's certainty that fish tissue
concentrations will be higher than what was anticipated in the ROD because the absolute
concentrations in the sediments were higher than expected. As documented in Appendix 4 of the
FYR report and further supported by the 2017 NYSDEC sediment survey, the sediment data would
suggest a substantial improvement in fish exposure to PCBs and, therefore, fish body burdens are
expected to decline.
As stated in the FYR report, it is EPA's assessment that fish tissue concentrations will decline in
response to the relative decline in surface sediment concentrations and the ensuing reduction in
water column concentrations and fish exposure. Absolute sediment concentrations do not dictate
specific PCB levels in fish, since factors such as organic carbon and sediment type influence the
bioavailability of the PCBs to fish. As documented in the FYR report, the surface sediments have
declined substantially since the 2002 to 2005 SSAP, in both dredged and non-dredged areas,
yielding reductions of 80 to 96 percent, depending on river section. Fish concentrations are
expected to decline similarly in future years.
In addition to the direct evidence on the recovery of the sediments themselves and the implications
for fish body burden declines in response, levels of PCBs in the water column are also expected
to continue their declines in response to the sediment improvements as well as the upstream source
control remedy implemented near the former Hudson Falls plant. The water column remains an
important means of exposure for fish. These reductions in PCB concentrations in sediment and
water are expected to lead to concomitant reductions in fish tissue concentrations. While EPA has
15
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
estimated these improvements through extensive modeling, additional long-term monitoring data
must be obtained before the success of the remedy can be more fully assessed.
Lastly, as discussed elsewhere in these responses, EPA is currently evaluating the sampling and
monitoring needs for the Lower Hudson.
3.1.2 Comment 6: Consumption survey required to assess new populations eating fish
Comment
Commenters question the effectiveness of NYSDOH's fish consumption advisories (FCAs) noting
that the public is still consuming fish from the Hudson River despite the advisories. Commenters
indicated that PCB concentrations currently found in fish continue to result in exposures to both
human and ecological receptors which are above EPA's acceptable risk range. Numerous
comments noted that comprehensive studies of FCA compliance or effectiveness have not been
performed in over 20 years. Some commenters were encouraged that EPA requested the assistance
of NYSDOH in evaluating the performance of the existing FCAs and the efficacy of the state's
outreach program. Some comments were concerned with environmental justice impacts, noting
that communities of color, low-income communities, and immigrants catch and eat fish from the
Lower Hudson River. Many comments requested that, since fish consumption patterns have
changed since the 1990s, when the last risk assessment was conducted, a detailed scientific, broad-
based fish consumption survey be conducted to quantify current and potential human exposure for
all contaminated river reaches in order to determine whether the advisories are sufficiently
protective over the short and long-term. If the survey finds that consumption patterns have
changed, the commenters request a review of risk assessment calculations to determine if updates
are needed. Additionally, comments request that EPA consider and evaluate the localized effects
of human exposure in more contaminated areas of the River.
Response
It is not possible or expected that the fish will recover immediately after the conclusion of
dredging. Rather, as EPA made clear in the 2002 ROD, such recovery will take many years. As is
the case at other contaminated sediment sites where the risk from fish consumption guides
remedial decisions. Natural attenuation is a necessary component of the remedy for the Hudson
River PCB Superfund Site. As a result, FCAs and/or fishing restrictions are a necessary component
of the remedy.
In the development of the ROD, various consumption surveys were taken into consideration when
identifying consumption patterns and quantities for the risk assessment. The 1991 New York
Angler survey (Connelly et. al., 1992) was used in the development of the exposure assessments
and in identifying species consumed. The Connelly survey was selected because the climate and
characteristics of the New York water bodies in that study were more likely to represent Hudson
River anglers than non-New York surveys. This survey was used in the development of the
exposure assessment and identification of species consumed. While it is understood that
consumption patterns will not remain the same over time, EPA was able to assess the need to
update the risk calculations based on the information currently in hand. Further, an update to the
16
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
consumption survey is not expected to increase the efficacy of the risk calculations because risks
are determined on an individual basis and the consumption rates as well as the species mix remain
appropriate and representative.
EPA reviewed the assumptions that were used as input to the risk assessment as part of this FYR,
including those associated with consumption. Further consideration of specific populations of fish
as a food source from the river does not affect the risk calculations. It does, however, highlight an
opportunity for additional/updated outreach efforts to inform the public about the advisories.
The FYR includes discussion of efforts New York State has taken to improve the effectiveness of
the advisories; Appendix 13 of the FYR report details New York State's efforts which include
more signage posting, updated graphics and informative materials, and the performance of angler
convenience surveys to target and expand its outreach. Given the iterative and ongoing nature of
outreach and recent NYSDOH efforts to enhance and focus efforts, the institutional controls
(fishing restriction and fish consumption advisories) appear to be functioning as expected. EPA
will continue to work closely with the NYSDEC and the NYSDOH to improve their fish advisory
outreach program. EPA strongly encourages the public to carefully review and adhere to the
advisories set by New York State.
3.1.3 Comment 7: Despite institutional controls people are still eating fish
Comment
Commenters questioned the effectiveness of NYSDOH's fish consumption advisories (FCAs),
noting that the public is still consuming fish from the Hudson River despite the advisories.
Commenters also indicated that various fish consumption convenience surveys taken over the last
few years indicate that fish are being consumed.
Response
EPA recognizes that some anglers do not comply with or may not be aware of the NYSDOH
Hudson River Fish advisories and NYSDEC fishing restrictions. EPA agrees that it is important
to continue and coordinate efforts with NYSDOH on its Outreach Program in an effort to optimize
awareness by the public. The Outreach Project includes ongoing efforts designed to more
effectively reach a broader and more diverse population of those who potentially catch and
consume fish from the Hudson River.
As discussed in Section 2.4 and Appendix 13 of the FYR report, Institutional Controls (IC) are an
integral part of Superfund site management, investigation, remediation, and post-remediation
monitoring. ICs have been effectively implemented by EPA and other government agencies at
contaminated sites for decades. As discussed in the 2002 ROD, site-specific ICs, including
continuation of fish consumption advisories and fishing restrictions, were anticipated to be
implemented as long-term control measures along with active remediation and a long-term
monitoring program. These site-specific ICs were designed to prevent or limit exposure to PCBs
through consumption of contaminated fish. EPA also acknowledged in the 2002 ROD that
consumption advisories are not fully effective by themselves in that they rely on voluntary
17
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
compliance in order to prevent or limit fish consumption. For this reason, they were implemented
as part of a broader remedy.
The 2002 ROD indicated that fishing restrictions and consumption advisories will remain in place
until the relevant remediation goals were met. EPA will not consider the OU2 remedy to be
complete until the MNA component also has been completed and project objectives are achieved,
including the fish consumption goal of 0.05 mg/kg PCBs in species-weighted fish fillet. As was
the case during the Baseline Monitoring Program (BMP) and Remedial Action Monitoring
Program (RAMP), EPA anticipates continued support and close coordination with New York State
to optimize the ongoing effectiveness of the consumption advisories and fishing restrictions during
the ongoing MNA phase of the Project.
As discussed in FYR report, EPA is encouraged by the post-dredging fish (see Appendix 3), water
(see Appendix 1), and sediment data (see Appendix 4), however given the limited amount of post-
dredge data a protectiveness determination cannot be made at this time. EPA also understands
that it is not possible for fish tissue PCB levels to recover immediately after the conclusion of
dredging - that recovery will take many years. As such, EPA will continue to monitor post-
dredging (natural recovery) results collected under OM&M, work with New York State to
continually improve IC effectiveness, and evaluate remedy protectiveness by comparing future
observations to expectations outlined in the ROD (which include various RAOs including fish
targets and goals). Therefore, as the river continues to recover, it is important for Hudson River-
area residents who fish to carefully review and adhere to the regulations and advisories set by New
York State.
3.1.4 Comment 14: EPA must reinstate suspended solids monitoring at Waterford to
improve evaluation of PCB load to the Lower Hudson River
Comment
Commenters suggested that the Hudson River Foundation recommendation to reinstate USGS
suspended sediment sampling at Waterford, along with additional high flow sampling, is needed
because higher than anticipated PCB mass remains in River Sections 2 and 3 after the completion
of dredging and because the current post-dredging data are insufficient to characterize the PCB
loading to the Lower Hudson River (LHR) that is associated with this remaining PCB mass.
Commenters noted that the updated HUDTOX model still underestimated measured 2004 to 2008
PCB loads by 8 to 41 percent during that pre-dredging period, so an improved data-based
characterization of the observed PCB loads to the Lower Hudson is important for evaluating the
effectiveness of EPA's implemented remedy.
Response
EPA agrees that the collection of additional suspended sediment data near Waterford would be
helpful to further characterize loads to the Lower River. TSS analysis has been done on all water
samples collected during the remedial action. Long term monitoring is expected to continue to
include TSS analysis. This monitoring provides comparable results to the historical data collected
18
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
by USGS near Waterford. Therefore, the current sampling program is sufficient to characterize the
PCB loading to the LHR.
EPA has and will continue to collect weekly water column data from Waterford to assess load to
the LHR. In addition to this weekly monitoring, high flow sampling will continue to be conducted.
This monitoring may help assess the effectiveness of the remedy with respect to impacts on the
LHR and the post remedy variability associated with high flow loads. EPA agrees this high flow
monitoring is important and that PCB mass transported downstream during these events can be
significant compared to the yearly load totals. The impact from high flow events is one
contributing factor to the differences in observed vs. predicted loads during the 2004 to 2008 MNA
period.
EPA will evaluate if an empirical relationship between suspended solids transport and PCB loads
can be derived from such post dredging monitoring including under high flow conditions. Ongoing
monitoring will continue to provide PCB loading information downstream of Waterford and its
effect on conditions in the LHR.
The reduction in the average surface sediment PCB concentration in RS 2 was lower than expected
by the ROD based on the SSAP data. As indicated in the FYR report, achievement of the various
remedial goals for RS 2 may lag those anticipated by the ROD by several years. However, based
on surface sediment data collected in 2016 and 2017,1 the inherent uncertainties in the model
forecasts, the long periods already anticipated to achieve the remedial goals in the Upper Hudson
and the better-than-anticipated improvements in RS 1 and RS 3, this delay in RS 2 is not deemed
a significant concern at this time.
3.1.5 Comment 31: EPA's species-weighted-average approach to estimating fish recovery
rates should be updated based on the current population's diet. EPA must modify its
homologue correction and use of data in developing temporal trends
Comment
EPA examined fish body burden decline on both an individual and a composite basis. However,
the composite basis only took into consideration three sportfish species (largemouth bass, yellow
perch, and brown bullhead). EPA noted that the pre-dredging PCB concentration decay rate for
the composite was 8 percent per year, consistent with its modeling analyses. EPA's choice of these
three sport fish species was intended to represent a typical fisherman's creel, based on studies done
during the 1990s. It is unclear whether these proportions are still representative of the exposed
population's diet, especially in light of population demographic changes. Different groups within
the population may consume different species or use different preparation techniques than the EPA
analyses assume. How will EPA address this possibility?
In an effort to test the effect of data transformation into homologue equivalent measurements on
the estimated decay rate, EPA calculated average decay rates by species and river section and
plotted these on a river-mile basis for both TPCBhe and TPCBArocior. However, about 50 percent
1 See Appendix 4 of the FYR report and EPA's April 2019 Technical Memorandum entitled "Technical
Memorandum Evaluation of 2016 EPA/GE and 2017 NYSDEC Surface Sediment Data" (www.epa.gov/hudson).
19
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
of the samples used in the TPCBhe trend analyses were eliminated for this step by selecting only
those species-river section combinations with at least 100 samples and 8+ years of data. This
elimination procedure censors out a large portion of the data and the effect of this has not been
statistically evaluated.
Response
EPA's use of three species to create a composite fisherman's creel for evaluating both risk and
remedy success represents a conservative approach, ensuring that estimates of exposure are based
on average conditions and not on a single species whose PCB concentrations might be high or low
relative to the risk-based thresholds. EPA agrees that the selection of fish consumed by local
populations can vary spatially and over time. EPA recognized this in the ROD, with different
species sets used in the Lower and Upper Hudson. Although the project uses a generic fisherman's
creel, EPA will look at all available data when evaluating the project. Additionally, NYSDOH sets
the fish advisories in NYS, and these advisories are specific to individual species and not based
on the fisherman's creel. EPA recognizes this and will require GE to collect additional fish species
as appropriate in the future to help inform NYSDOH on its advisories. EPA collects data on more
species than are included in the creel composite, recognizing the need to consider other species.
EPA modeled seven different species in the ROD and examined a total of eight different species
in the FYR. The analysis in the FYR determined the decay rates for each species at each station.
The decay rates for each species included in the composite form the upper and lower boundaries
of the rate that can be derived from a specific creel composite.
Regarding the calculation of decay rates throughout the Hudson as measured by both TPCBhe and
TPCBArocior, EPA's data presentations in the FYR report were intended to show that essentially all
permutations of the data revealed the same or similar relationships over time and across distance.
The commenter's assertion that half of the data were excluded did not take note that this was a
reference to the second of several figures prepared in support of EPA's FYR report. In Figure A3-
16A, of Appendix 3 of the FYR report, nearly all of the long-term data were used in both the
TPCBhe and TPCBArocior plots. In Figure A3-16B of Appendix 3 of the FYR report, EPA selected
only those fish trends with the most extensive records in an effort to eliminate less robust records
(with fewer measurements and over shorter periods of time), where the less robust trends might
confound the overall interpretation of trend with river mile. These data were eliminated from both
TPCBhe and TPCBArocior plots shown in Figure A3-16B of Appendix 3 of the FYR report, so there
is no difference in the amount of data used in either plot. The point of these analyses is to show
that the data describe the same trend, regardless of the length of time or the amount of data
available for individual species. That is, that the relatively rapid rates of decline in Upper Hudson
fish body burdens decline with distance downstream in the Lower Hudson. The exclusion of data
generated without ribs, which reduces the numbers of samples available for several species, does
not change the observed trend (see Figure A3-16C of Appendix 3 of the FYR report). There is no
need to conduct more rigorous statistical analysis to test the effect of the screening procedures
since all three approaches yield the same overall trend with distance downstream.
20
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.1.6 Comment 54: There is no basis in the record for the estimate of mass discharged to
the river by GE from the capacitor plants in Hudson Falls and Fort Edward (1.3
million pounds)
Comment
A commenter has concluded that there is no basis in the record for the estimate of the PCB mass
discharged to the river by GE from the capacitor plants in Hudson Falls and Fort Edward (1.3
million pounds) as presented in the FYR report. The commenter stated that the estimate is
uncertain, and believes that it is inaccurate and inappropriate to continue to cite this estimate. The
actual mass discharged to the river is unknown, and may be much more than 1.3 million pounds
(650 tons).
Response
The estimate of 1.3 million pounds was originally published by Professor John Sanders of
Columbia University in a peer-reviewed journal (Sanders, 1989; page 16). According to the report,
the estimate is based on historical use records of PCBs at the GE capacitor facilities between 1957
and 1975. The report stated that less than one percent of the PCB mass used during manufacturing
at these facilities (estimated at 133,100,000 pounds) was discharged into the Upper Hudson River
{i.e., about 1,331,000 pounds). Nonetheless, EPA acknowledges the possibility that the value may
underestimate PCB release from the GE facilities, and should not be interpreted as an upper bound
estimate.
3.1.7 Comment 61: Significant PCB deposits left behind are in excess of other cleanup
projects
Comment
Commenter stated that the remedy left behind significant deposits of PCB-bearing sediments
throughout the Upper Hudson that were not identified for removal by the ROD criteria. Those
deposits are in excess of standards used in other PCB cleanup projects and leave the river subject
to additional cleanup costs when other residential or public projects are attempted.
Response
EPA does not agree with the premise of this statement; the standards used in other PCB cleanup
projects are not relevant to the Upper Hudson River for two primary reasons.
First, EPA's 2002 ROD made the explicit decision to target Tri+ PCB and not Total PCBs, since
the Tri+ PCB fraction represented the main exposure to humans who consume fish and fish-eating
birds/mammals. The lighter PCB fraction does not significantly accumulate in fish tissue and thus
poses minimal risk. Therefore, the premise that the Hudson sediment contamination should be
addressed based on a Total PCB criterion is not supported.
21
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Second, the ROD explicitly specified removal criteria for surface sediment concentrations and
sediment inventory of Tri+ PCBs based upon a comparison of mathematically simulated remedial
alternatives. The selected remedial alternative provided a good balance between the degree of
reduction and the required extent of disruption to the river. The ROD recognized that active
remediation to the levels suggested by the commenter would have required bank-to-bank dredging
for the entire length of the Upper Hudson at much greater cost and disruption to the environment
and surrounding communities, but without correspondingly greater improvements. Rather than
require such an extensive remedy, EPA included a MNA2 component in the remedy as outlined in
the ROD. It was, and still is, EPA's contention that the Upper Hudson will attain Tri+ PCB levels
of approximately 0.25 mg/kg or less Tri+ PCB in the surface sediment as a result of the dredging
and natural recovery. EPA's ultimate remedial goal for the sediments is to achieve Tri+ PCB levels
at the surface that result in the anticipated reduction in fish.
3.2 Modeling Analysis
This section includes comments and responses concerning spatial scale, the relationship of fish,
sediment, and water data, and the assumptions on how the river is going to recover. Modeling
analysis includes justification of the models and assumptions used to assess the Remedial Action
Objectives (RAOs). EPA has completed additional technical analysis supporting comment
responses related to NOAA's emulation model. This analysis is included as Appendix C of this
document.
3.2.1 Comment 9: EPA models of recovery in fish, sediment, and water are overestimated
and should be revisited
Comment
Commenters indicated that the MNA recovery rates estimated by EPA for the MNA period prior
to dredging overestimate the rate of recovery, referring to NOAA's analyses and emulation of
EPA's ROD models (and subsequent application of those analyses and emulation with assumed
adjustments for decay and sediment PCB concentrations) to support that conclusion. Commenters
also indicated that based on NOAA estimates the remedy is not protective and further state that
post-dredging PCB concentrations in fish should be used to determine remedy
effectiveness/protectiveness as outlined in the ROD, rather than EPA relying on percent reduction
of PCBs in sediment and uncertain PCB decay rates. One commenter also stated that post-remedy
concentrations are driven by both the recovery rate and initial (post-dredging) concentrations. In
contrast, another commenter finds a lack of scientific credibility in the effort conducted by NOAA
to emulate, update, and forecast with the EPA ROD models (HUDTOX and FISHRAND) using
SSAP sediment monitoring in 2002-2005.
Response
EPA conducted an extensive independent review of NOAA's manuscript entitled Re-Visiting
Projections of PCBs in Lower Hudson River Fish Using Model Emulation (Field, et al., 2016),
2 The term MNA used in the 2002 ROD was consistent with then-current usage; subsequently EPA's 2005 Sediment
Remediation Guidance established Monitored Natural Recovery (MNR) as the consensus term of art for sediment
sites. The two terms are synonymous in the current context.
22
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
which uses model emulation to predict lengthy delays in fish recovery times relative to forecasts
made with EPA's models. EPA's detailed responses regarding NOAA's emulation model are
contained in EPA's white paper3 and in Appendix C of this document, and summarized here.
NOAA developed an "emulation" of EPA's models and subsequently "updated" the surface
sediment PCB concentrations to forecast fish tissue concentrations in a predictive scenario known
as MNA2 in their manuscript. This MNA2 model emulation is not valid because it ignores the
underlying mechanisms of the model and that the model was developed using actual water, fish
and sediment data. If the sediment concentrations were different than those used in the model (as
the MNA2 model emulation suggests) then the relationships between fish, water and sediment
would also need to be adjusted. Simply changing a variable such as sediment PCB concentration
without recalibrating the underlying model to maintain consistency with the calibration data
produces results that are flawed. EPA's evaluation of the NOAA manuscript shows that the MNA2
predictions are biased high for water column PCBs and fish tissue PCBs. This bias is the main
reason for NOAA's prediction of extended times to reach RAO targets. Appendix C of this
document provides additional supporting information and shows that most of the change in time
to recover claimed by NOAA is due to this upward bias caused by a failure to recalibrate. No
model-data comparisons are presented in the NOAA paper to support their water column PCB
predictions or wet-weight fish tissue predictions, even though extensive data on both have been
collected since the ROD. Also, NOAA did not consider all the available sediment data sets in their
analysis.
In contrast, the EPA models that supported the ROD successfully reproduced data from 1977-
1998, were peer-reviewed as part of the Superfund process and continued to match trends in water
and fish data through the extended 1998-2008 period of pre-dredging monitored natural recovery
reasonably well, as shown in the FYR report (Appendices 1 and 3).
Commenters indicated that post-dredging PCB concentrations in fish should be used to determine
remedy effectiveness/protectiveness as outlined in the ROD, rather than EPA relying on percent
reduction of PCBs in sediment and uncertain PCB decay rates. EPA agrees that fish tissue
concentrations and their recovery rates are the primary metric to be used when assessing the
effectiveness of the remedy. However, EPA must also use all available data when evaluating the
remedy effectiveness, including water, sediment, fish and any analysis of those data including
percent reduction. It is also important to consider that, for this five-year review period, limited
post-dredging fish data are available and these data are likely still impacted by dredging project
activities (including habitat reconstruction activity in 2016). Lastly, up to eight or more years of
post-dredging data are expected to be necessary to assess with statistical confidence when fish
tissue concentrations will achieve the goals set in the ROD.
EPA remains committed to collecting post-dredging data under the OM&M program to improve
its understanding of PCB concentration trends in fish, sediment and water over time.
3 See: White Paper: Responses to NOAA Manuscript Entitled: "Re-Visiting Projections of PCBs in Lower Hudson
River Fish Using Model Emulation" (Field, Kern and Rosman, 2015) (EPA, 2016)
23
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.2.2 Comment 15:EPA should update its models to reflect information obtained during
dredging
Comment
Commenters indicated that the conceptual site model (CSM) should be updated as part of the
ongoing management of the remedial program for this site, now in the monitored natural recovery
phase. The commenter asserts that data collected during dredging indicates a fundamental change
in the relative contribution of water-borne to sediment-borne PCB contamination to fish body
burdens. The commenter believes that the relatively limited increases in fish PCB body burdens
during dredging in response to the much larger increases in water column concentrations must
indicate that fish body burdens are controlled primarily by sediment with little water column input.
The commenter also indicated that the appropriate spatial scale {i.e., pool-by-pool, rather than
averaged over multiple pools) should be used in the design of sediment, water, and fish sampling
to be undertaken to better understand the performance of the remedy.
Response
Two years of post-dredging data are not sufficient to assess the post-remediation recovery. EPA
anticipates that it will take up to eight or more years of fish tissue data to identify trends with a
reasonable degree of scientific certainty. EPA does not believe that a fundamental change to the
CSM is needed. The main mechanisms for PCB transport, degradation, resuspension and fish
uptake in the CSM are still operative, although now driven by lower concentrations overall. While
EPA agrees that fish tissue body burdens did not increase as much as water column concentrations
during dredging, EPA does not believe those data are sufficient to invalidate the relative
contributions estimated in the ROD models. In particular, HUDTOX and FISHRAND models are
based on many years of calibration data, and forecasted trends and concentrations in fish body
burdens were well matched to those observed during the pre-dredging MNA period from 1998 to
2008. Thus, the models' mathematical representation of site conditions appears to be sound.
Moreover, although the models were used to roughly approximate dredging conditions, the models
were not designed to capture the highly variable and transient conditions associated with dredging.
EPA therefore does not consider the models' ability to represent the dredging period (including
the post-dredging equilibration period) to provide a test of the models' reliability. Thus, EPA does
not agree that the evidence requires a fundamental revision of EPA's models to change the relative
roles of sediment and water exposure to fish body burdens.
Commenters suggested that EPA focus its future monitoring on a reach-by-reach basis rather than
for the whole river or by river section. While the ROD's expectations are based on the river section
scale, EPA agrees that there is value in assessing reaches within the river, river sections, and the
whole river. EPA has and will continue to evaluate the river at these different scales. All reaches
will be monitored as part of the surface sediment sampling program, thus providing recovery
information on a reach basis. Depending on data availability and the degree of similarity of PCB
concentrations, reaches may be grouped as part of long-term trend analysis. Periodic reach-specific
fish monitoring will also be done as appropriate to confirm fish recovery is consistently occurring
throughout the Upper Hudson. Additionally, EPA will coordinate with NYSDEC and NYSDOH
24
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
regarding location-specific sampling modifications as necessary to inform evaluations of ongoing
recovery and decisions regarding adjusting fishing restrictions and fish consumption advisories.
3.2.3 Comment 26: Conceptual site model - relationship of sediment, water, fish
Comment
Commenters noted that fish tissue concentrations are closely tied to localized remedial activity and
sediment contamination in the Upper Hudson River (UHR). Specifically, commenters asserted
that UHR fish are not likely to travel between pools due to dams and locks and are exposed only
to the sediments of the pool in which they live and that data collected during the Remedial Action
Monitoring Program (RAMP) indicate that the local sediments play a larger role in influencing
fish PCB concentrations than was thought at the time of remedy selection. As a result, commenters
indicated that the scale of project management needs to change from river section to reach and the
assumed relationships between sediment, water, and fish, under which cleanup levels were
developed, need to be re-evaluated and re-quantified.
Response
EPA agrees that fish tissue concentrations are linked to sediment concentrations but does not agree
that the relationship between sediment, water, and fish tissue PCB concentrations was not
understood at the time of remedy selection. The HUDTOX and FISHRAND models were
calibrated, verified, and applied to the UHR to support EPA decision-making. The application of
the calibrated models to the UHR allowed direct comparisons of predicted water, sediment, and
fish tissue concentrations across proposed remedial alternatives. The FISHRAND model assumed
localized exposures on a reach-by-reach basis by relying on underlying sediment and water
exposure concentrations developed using the HUDTOX model at these localized scales. The
strength of the combined model framework lies in its ability to compare predicted concentration
trajectories over time using a consistent set of assumptions. The models successfully underwent
peer review in 2000.
The 2002 ROD reflects that relationships between key model components (sediment, water, and
fish) were well understood and contributed to EPA's designation of REM 3/10/Select as the
Selected Remedy. Consistent with the scale at which FISHRAND was calibrated and applied, fish
data collected during the Baseline Monitoring Program (BMP, 2004-2008), i.e., prior to the RAMP
(2009-2015), involved collecting multiple samples for each target species from multiple locations
within reaches. The sampling locations adopted for the BMP were based on, and represented an
expansion of, the number of long-term NYSDEC stations sampled for pre-dredge studies (Sloan,
et al. 2002) in UHR reaches 8 through 5. The sampling approach developed for the BMP and
carried forward as the RAMP reflected the BMP goals and DQOs. Specifically, the BMP (fish)
goal was to "provide data on PCB levels in fish and water to allow the evaluation of long-term
recovery trends" and the fish sampling DQO was to "establish baseline PCB levels in UHR
resident sport fish and resident forage fish to allow for documentation of the changes in PCB
concentration that result from remediation." As stated in BMP QAPP Section B. 1.2.3, a specific
objective of the fish sampling program was to provide "a reasonable estimate of reach average fish
25
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
PCB concentrations." Thus, the FISHRAND model and the BMP and RAMP fish data collection
approaches were each designed to evaluate fish tissue concentrations on a reach basis.
As discussed in Appendix 1 of the FYR report for the water column, Appendix 3 for fish, and
Appendix 4 for sediment, the model has performed within the range of expectations when
compared to observed data prior to dredging. In 2016, concentrations for individual species ranged
from 0.4 to 1.7 mg/kg depending on the species and location. Yellow perch, for example, had
already achieved the 0.4 mg/kg interim target at several locations. The 2016 species-weighted
average is about 1 mg/kg in all three river sections. These values are comparable to the model
results for the first year post-dredging as shown in Figure A3-19 of Appendix 3 of the FYR report.
Comments regarding potential challenges encountered in fish monitoring program implementation
are further addressed in Appendix 3 and Appendix 8 of the FYR report. There is no basis for
concern about the applicability of the CSM or model forecasts. The system underwent a "reset"
following dredging activities and established a new "baseline" from which post-dredging, or
MNA, trends will be evaluated. Accordingly, evaluating data-based trends into the future starting
with this new baseline will require additional data over multiple annual cycles to provide
statistically meaningful estimates of progress toward meeting the interim targets and final goal.
The ROD evaluated potential remedial alternatives at the river-section scale. Data collected in
support of remedy selection and modeling were calibrated, verified, and applied at the reach scale.
Results were then compiled as needed to the river-section scale as appropriate. EPA recognizes
that in some reaches minimal data were collected. EPA and GE are currently discussing fish and
sediment data collection scopes of work under the OM&M program, and have collected initial
baseline post-dredging sediment, water, and fish samples. Data collection for all three media
will continue under OM&M. Ongoing discussions include additional fish data collection in
reaches 1 through 4 (see Figure 2 of the FYR report), in part for the purpose of informing
NYSDOH fish consumption advisories and NYSDEC fishing restrictions. It is important to note
that reaches 1 to 4 were not included in baseline fish monitoring because it was expected that the
reach 5 would conservatively represent those reaches. Reaches 1 to 4 have significant sections of
bedrock bottom and in most areas have lower sediment concentrations than the upstream reaches.
Therefore, fish in those reaches will likely be lower in concentration than upstream reaches,
including reach 5.
As the water, sediment, and fish recover from dredging and the project transitions from the
remedial to the OM&M phase, the emphasis will be on comparing observed fish tissue levels to
ROD targets and RAOs rather than comparison to model forecasts. Therefore, it is unnecessary
to reevaluate and re-quantify the relationships between sediment, water, and fish PCB
compartments at this time. EPA will continue to monitor post-dredging (natural recovery) results
collected under OM&M and evaluate remedy protectiveness (as part of FYRs) by comparing
future data to project RAOs (including the fish targets and goal).
26
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.2.4 Comment 27: EPA's model prediction that the Upper Hudson River PCB load to the
Lower Hudson River is the primary factor for recovery of Lower Hudson River fish
is proven incorrect by this Five-Year Review
Comment
Commenters noted that post-dredging impacts on water column Tri+ PCBs (PCBs containing three
or more chlorines) in the Lower Hudson River average four times higher than predicted by the
2002 ROD models and that additional years of MNA will be required before PCB levels are
acceptable in water, sediment and fish because the EPA ROD model of the Lower Hudson River
failed to properly reflect the cyclic nature of sediment transport resuspension and deposition.
These commenters also argued that there are not enough data/evidence to support the assumption
that PCB loading from the Upper Hudson to Lower Hudson River plays a major role in Lower
Hudson River recovery because the FYR itself indicates that PCB loading from the Upper Hudson
River to Lower Hudson River is not a primary factor considering the slow recovery of Lower
Hudson River fish. The comments cite multiple studies and references (Thomann et al. 1989,
Farley et al. 1999, USEPA 2000a, Hydroqual 2007, Rodenburg and Ralston 2017) that suggest,
based on high-resolution core sampling, the primary source of PCBs to the Lower Hudson River
is the result of past and continued loading of PCBs originating from the Hudson Falls and Fort
Edward plant sites and sediments within the Upper Hudson River. Therefore, the commenters
recommend that a more in-depth analysis of Upper Hudson River effects on the Lower Hudson
River is needed.
Response
The Farley model that was used to extend EPA forecasts to the Lower Hudson River included
sediment settling and burial, resuspension, and diffusive exchange as processes contributing to the
complexity of PCB transport, as did EPA's model of the upper river. In contrast to development
of EPA's Upper Hudson River model, the water column data available for calibration of the Farley
model were very limited, and that model was not calibrated to water column data. As shown in
FYR report Appendix 1, Farley model water column forecasts for Albany through 2008, which
were driven primarily by EPA's HUDTOX Tri+ PCB forecasts for Troy Dam, the downstream
boundary of the Upper Hudson River, were accurate, but Farley model water column Tri+ PCB
forecasts for Poughkeepsie were systematically low, reflecting limitations in the Farley model
calibration. EPA agrees that past loadings of PCBs from the Upper Hudson River have been a
major source of PCBs to the lower river, including periods of uncontrolled historical release to the
Upper Hudson River and subsequent periods of declining loads. EPA has evaluated the extent to
which loadings from the Upper Hudson River currently contribute to Lower Hudson River
concentrations, and as discussed in Appendix 1 of the FYR report, the evidence from the dredging
period indicate that the water-column PCB response was greatly attenuated between Albany and
Poughkeepsie, and that local sources, including legacy deposits, likely account for elevated
Poughkeepsie water-column PCB concentrations. EPA agrees that it is important to collect
additional data/information about other sources and PCB fate and transport in the Lower Hudson
River. EPA is moving forward with supplemental studies of the Lower River.
27
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.2.5 Comment 34: Water quality improvements from dredging tend to decrease with
distance downriver from dredging
Comment
Commenters suggested that available post-dredging data show that the improvement in water
column PCB concentrations diminishes downstream of Thompson Island Dam (TID) and that it is
unclear whether ROD targets for PCB mass transport reductions will be achieved because of data
limitations and the complicating influence of year to year variations in flow.
Response
EPA agrees that annual variations in flow complicate PCB load evaluation and that trends over
brief time frames may not well represent long-term water quality improvements. EPA also agrees
that improvements to water quality decrease the farther one moves downstream from the dredging.
However, Appendix 1 of the FYR report shows that overall post-dredging reductions in surface
water PCB concentration were substantial at Waterford compared to the pre-dredging levels, as
well as at the other Upper Hudson River (UHR) locations. As the river continues to recover, EPA
(as part of OM&M) will continue to track water column PCB concentration and loading trends
downstream of the dredging (including to the Lower River).
3.2.6 Comment 44: NOAA's models demonstrate that the EPA ROD models are flawed
and should be updated to correctly reflect the role of sediment concentrations in
evaluating protectiveness of the remedy
Comment
Commenters cited NOAA's model emulation (Field, et al., 2016), including its update substituting
SSAP data for HUDTOX-simulated sediment concentrations, as a basis for determining that
EPA's models are no longer valid. One commenter asserted that EPA's models assume that only
sediments control fish exposures in RS 1 and 2, and that only the water column controls fish
exposures in RS 3. The commenter argued that local sediments control exposures everywhere.
Another commenter called for use of an updated model to assess the effect of post-dredging surface
sediment concentrations on the protectiveness of the remedy.
Response
EPA conducted an extensive review of NOAA's manuscript entitled Re-Visiting Projections of
PCBs in Lower Hudson River Fish Using Model Emulation (Field, et al., 2016). EPA's detailed
responses regarding NOAA's emulation model are contained in EPA's white paper4 and in
Appendix C of this document, and summarized in Master Comment 9 (see Section 3.2.1). As
discussed in the response to Master Comment 9 (see Section 3.2.1) and as detailed in Appendix C
of this document, the NOAA updated model emulation produced biased water column and fish
tissue simulations by failing to recalibrate after altering sediment concentrations.
4 See: White Paper: Responses to NOAA Manuscript Entitled: "Re-Visiting Projections of PCBs in Lower Hudson
River Fish Using Model Emulation" (Field, Kern and Rosman, 2015) (EPA, 2016)
28
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Contrary to the commenter's assertion, FISHRAND does, in fact, assume that sediments are a
critical element of PCB exposure for all fish in all River Sections, and simulates fish body burdens
as functions of local sediment exposures, taking into account local habitat and ranges and
recognizing dams as pool boundaries. In addition to sediment exposures, water column exposures
also matter in FISHRAND, consistent with the science on fish PCB uptake, and varying by species
according to their degree of benthic or pelagic exposure within the food web.
As shown in the FYR, EPA's models performed well in simulating water column concentrations
and fish tissue concentrations through 2008, just prior to dredging, and EPA does not see a need
at this time to develop an updated model. EPA notes that its models simulated water column and
fish tissue concentrations for 2004 to 2008 much more accurately than the NOAA emulation model
update cited by multiple commenters.
3.2.7 Comment 55: EPA needs to update the conceptual site model (CSM) and recalibrate
and update HUDTOX and FISHRAND models in order to properly understand the
impacts of the dredging on the resultant fish concentrations
Comment
Commenters stated that with fifteen years of data collected in the Upper Hudson River since the
ROD was issued, and the realization that more PCB mass was present than originally estimated in
the 2002 ROD, EPA needs to update the conceptual site model (CSM) and gather the data
necessary to determine if the amount of remedial work identified in the ROD will achieve the
targeted reductions in human health and environmental risk. EPA also needs to update the agency's
understanding of how the PCBs remaining in Hudson River sediments impact the water column
and fish in the river.
Commenters also stated that EPA must update, restructure and recalibrate the mathematical models
developed for the Site to properly take into account what has been learned since the ROD was
issued. These commenters asserted that EPA has never provided a valid scientific reason for not
updating its modeling and flawed predictions. The commenters also stated that, currently, EPA is
relying on overly optimistic model projections regarding the anticipated rate of natural recovery
in the river by underestimating the impacts of local sediments on fish and thus underestimating
the benefit of active remediation.
Response
Based on EPA's understanding of site conditions for the Upper Hudson River, the CSM
appropriately represents the interactions among PCBs in the sediment, water column and aquatic
organisms, and an update to the CSM is not warranted at this time. As presented in the 2002 ROD,
the CSM for the Upper Hudson River describes the source-to-receptor succession in simple terms
and identifies the major contamination sources, contaminant release mechanisms, secondary
sources, and pathways and receptors of concern (EPA, 2002). Data collected subsequent to the
release of the 2002 ROD have not altered EPA's understanding of how contaminant sources,
release mechanisms and pathways of contaminants impact receptors of concern, and no data have
29
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
been collected that would necessitate a substantive update to the CSM as presented in the 2002
ROD. More specifically, the identification of additional PCB mass in the Upper Hudson did not
require an alteration to the CSM, as PCBs in the sediment were accounted for in the original CSM
presented in the 2002 ROD. However, EPA continually reviews new data as they are collected and
will continue to assess whether updates to the CSM are necessary.
The HUDTOX and FISHRAND models were developed in a manner consistent with the CSM and
the spatial scales needed to inform the 2002 ROD, including linking fish tissue concentrations at
each sampling station to local sediment exposures. These models were also subject to a rigorous
peer review by a panel of international experts (ERG, 2000). After extensive document review and
a series of public meetings, the peer review panel determined that the models were acceptable and
adequately reproduced historical data. Model-data comparisons presented in Appendices 1 and 3
to the FYR reportdemonstrate that the models successfully represent the water column and fish
data collected during the 11-year MNA period from 1998 through 2008. Concerns over sediment
data do not indicate the need for the EPA models (HUDTOX and FISHRAND) to be modified
because their ability to accurately predict fish and water concentrations over the MNA period
demonstrates they were not overly optimistic in predicting future conditions and that they had
value as decision tools to inform the ROD at the time that it was issued in 2002.
Now that dredging activities have been completed, the development of empirical, data-based
trends for the recovery of fish tissue and water column concentrations will provide the strongest
evidence of whether the remedy is functioning as intended, rather than reliance on model-based
predictions of trends in PCB concentration. While model-based predictions were necessary before
the remedy was implemented, the project is entering a phase where the river bottom and PCB
inventory have been extensively modified, and where the data itself will determine rates of
recovery. It is unlikely that additional modeling work done at this time would add significant value
in predicting long-term recovery. It should also be noted that the time needed to develop an
updated suite of models, including necessary data collection, would be quite long. Updated models
would need to be initiated to represent post-dredging conditions and then calibrated to a dataset
adequate to support forecasting long-term trends. Accounting for potential model peer review, this
process would likely take many years, at the end of which EPA would likely have collected
sufficient post-dredging data to determine empirically-based MNA trends for PCBs in water,
surficial sediments and fish within an acceptable range of uncertainty.
3.3 Assessment
This section includes comments and responses on PCB Aroclor considerations, risk assumptions,
other EPA reports, fish data assessment, Lower Hudson River assessment as applicable, and other
similar comments. Most of the master comments and responses fall within this category. As
requested, EPA has finalized the Black Bass fillet tissue with and without ribs study. This study
is included as Appendix D of this document.
30
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.3.1 Comment 2: Adjust data treatment techniques for Aroclor data
Comment
Some comments recommended that the impact of using a single correction factor to adjust multiple
years of fish PCB data on the uncertainty of the temporal PCB trend analysis should be assessed.
They also commented to confirm that TPCBs in fish from mGBM are comparable to TPCBs from
M1668. The "homologue" adjustment of NYSDEC and GE fish data in the FYR uses a single
factor based on a geometric mean of the ratio of Aroclor PCBs to mGBM TPCBs. In the case of
the NYSDEC data, the adjustment factor from 1999-2000 is applied to all subsequent years
without any data to document applicability. The NYSDEC adjustment factor applies the factor
from wet weight analysis to the lipid-normalized concentrations, instead of more appropriately
using the factor from lipid-normalized analyses, which are substantially different for some years.
Also, for the 1997 NYSDEC fish data, the FYR relies on a model-estimated factor from Butcher
et al. (1998), ignoring the data from the split-sample approach used for subsequent years. Using a
single factor ignores the uncertainty/variability of the relationship for different subgroups (e.g.,
species, location, year) and may not represent the pattern in the underlying data. Some commenters
also stated that the transformation from Aroclor to TPCB homologue-equivalent introduced a very
large degree of uncertainty on the transformed data and therefore the fish tissue recovery rate of 8
percent is uncertain. Another commenter stated that EPA's conclusion that transforming the data
from Aroclor based to homologue equivalent measurements had virtually no effect on fish tissue
trends was based on an analysis that excluded about 50 percent of the total data.
Response
There are multiple assertions made in this comment, each of which is addressed below.
EPA's application of adjustment factors for Aroclor-based data is predicated on the theory that
homologue- and congener-based methods (e.g., capillary column-based methods, including GE's
mGBM and EPA's Ml 668) provide more accurate estimates of the true PCB concentration, since
these methods attempt to quantify the congeners themselves. Therefore, when matched pair data
were available,5 EPA developed relationships between Aroclor-based results and those of the
homologue- and congener-based methods so as to adjust the Aroclor-based results to a consistent
homologue-equivalent basis. This is discussed extensively in Appendix 5 of the FYR report. EPA
uses the most applicable matched pair data to derive the appropriate correction factor for a dataset.
EPA's primary estimates of the decay rates in fish tissue concentration are based on these
homologue equivalent values. EPA recognizes that these transformations introduce variation and
uncertainty, much of which can be difficult to quantify directly.
Because of the difficulties in assessing the magnitudes of the various individual uncertainties
involved (e.g., the use of a single conversion factor over multiple years of data), EPA did not
assess the individual sources of uncertainty. Rather, EPA examined the overall level of uncertainty
introduced by the transformation process. To assess the sensitivity of the decay rate estimates to
EPA's homologue-equivalent-based approach, EPA repeated the entire analysis of decay rates
5 Matched pair data refers to samples where both an Aroclor-based and a homologue- or congener-based method was
run on the same sample, providing two TPCB concentration values for the same sample.
31
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
using the Aroclor-based fish tissue PCB concentrations as originally reported, thereby avoiding
the uncertainties introduced by the transformations. In both approaches, the amount of data
available for the calculations was the same {i.e., there was no reduction in the amount of data used
for the calculations by the homologue-equivalent-based approach, contrary to the assertion by the
commenter.) These results are presented in Appendix 3 of the FYR report. In particular, Figures
A3-16A through C directly contrast the rates obtained for fish tissue across the Hudson by both
homologue-equivalent and Aroclor-based calculations. By using the Aroclor-based data as
reported, the second set of diagrams in each figure avoid any uncertainties introduced by the
transformation process.
To further reduce the uncertainty in the rate of decline evaluation, EPA relied on the integration
of multiple fish species at each station, even testing the sensitivity of the results across larger and
smaller data sets (compare Figure A3-16A to Figure A3-16B). EPA's analysis indicate that the
Aroclor-based rates of decline are much the same as the homologue-equivalent-based rates. This
conclusion was the same whether derived from all applicable fish tissue samples (Figure A3-16A)
or when derived by the subset limited to species with large numbers of samples and more extensive
temporal coverage (Figure A3-16B). Thus, the basis for the commenter's assessment that EPA's
conclusion was derived from only 50 percent of the total data is unknown and cannot be replicated
by EPA. By conducting this sensitivity analysis, EPA demonstrated that the concerns raised by
the commenter, such as the use of a single adjustment factor for the post-1999 NYSDEC data, do
not affect EPA's conclusions, since the decline rates show similar magnitude and spatial
relationships along the river, with and without adjustment.
EPA's analysis also shows that the lipid-based decay rates are consistently slower than wet weight-
based estimates. For both Aroclor and homologue-equivalent bases, lipid-based decay rates
average between 5 and 10 percent per year in the Upper Hudson. These rates consistently reduce
to much slower decay rates with distance downstream in the Lower Hudson by either PCB
measurement basis. Thus, EPA's main conclusions about the rates of decline in fish tissue PCB
levels over time are not sensitive to the treatment of the PCB data and are derived from all
applicable fish tissue samples.
One concern raised by a reviewer is that EPA applied the adjustment factor from 1999-2000
NYCDEC fish data to all subsequent years. It should be noted that matched pairs are not available
for NYSDEC data post-2000. Thus, there are no additional data from which to develop these
factors. Since NYSDEC used the same laboratory from 1999 through 2011 and the reported
Aroclor compositions are relatively similar during this period, the continued use of the 1999-2000
adjustment factor for the post-2000 period is the best approach based on available data. As
explained above, EPA's sensitivity analysis concludes that this data treatment approach does not
impact its conclusions regarding the magnitude or the spatial variation of the fish decay rates.
Regarding the use of lipid-normalized PCB adjustment factors vs. wet-weight PCB adjustment
factors, EPA also does not agree with the comment's assertion. The available matched pair fish
tissue data for Aroclor-based and the homologue-equivalent analyses are not consistently matched
with lipid analyses. In some years (1995, 1999 and 2000), independent lipid analyses are available
for both the Aroclor-based and the homologue-equivalent-based analyses, making lipid-
normalized correction factors possible. In other years (1997 and 1998), lipid analysis is only
32
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
available for the Aroclor-based analysis, eliminating the possibility of a lipid-normalized
correction factor between methods. For this reason, EPA approached this issue by developing
correction factors based only on the reported PCB data, which could be done consistently for all
years of data. As discussed previously in this response, EPA has compared the trends derived from
the homologue-equivalent data and those from the raw Aroclor data, on both a wet-weight basis
and a lipid-normalized basis. The comparison indicates that the average decay rates on a wet-
weight basis and the average decay rates on a lipid-normalized basis are not changed by the use of
homologue-equivalent data vs. the use of the original unmodified Aroclor data. These analyses
suggest that the spatial variation and mean values for the decay rates are not sensitive to the data
adjustment approaches.
The correction factor for the 1997 NYSDEC fish data was derived by regression analysis6 using
the matched pair data collected in 1997 (Butcher et al., 1998). Applying the correction factor from
subsequent years to the 1997 data will introduce more uncertainty due to the differences in Aroclor
and homologue analytical procedures applied by the different laboratories. These differences are
the reason EPA developed the sampling year-analytical laboratory-specific equations shown in
Table A5-20 in Appendix 5 of the FYR report.
Regarding the possible variations in these factors due to species differences, EPA examined the
possible effect of various subgroups in its analysis, as presented in Figures A5-11 and A5-15 of
Appendix 5 of the FYR report. These figures present the matched pair data on a species basis. It
is evident from the figures, as concluded in the appendix, that there is no discernable difference in
the relationship between the matched pairs of analytical results that can be explained by species
differences.
EPA does agree with the commenter's underlying assertion that it is important to establish an
accurate basis for PCB measurements in fish for the current and future monitoring efforts.
However, it is not necessary to establish the accuracy of the mGBM or its comparability to Ml 668
in this regard. Specifically, since the mGBM is no longer available, it will not be used in future
fish tissue monitoring. More importantly, the remediation changed in-river conditions
significantly. Thus, trends prior to dredging, which reflect pre-dredge conditions, do not impact
considerations going forward. Rather, it is the improvement in post-dredging conditions that will
be monitored and form the basis for any future consideration regarding river recovery. To provide
an accurate basis for long-term monitoring beginning in 2016-2017, EPA will be conducting
analyses of fish tissue by both Aroclor-based (M8082) and congener-based (specifically Ml668)
methods as part of the ongoing fish monitoring program. Similar to the sediment monitoring
program, samples will be homogenized and then split for analyses by the two methods. Also
similar to the sediment program, GE's laboratory will be required to run reference standards to
confirm analytical accuracy and provide a benchmark for future monitoring work.
6 It is noted that Butcher et al., 1998 did not include these relationships in their paper.
33
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.3.2 Comment 3: Assess risks of PCBs based on changes in consumption
Comment
Commenters state that since the risk assessment work was completed in the mid to late 1990s, it
appears that there has been a change in the species mix among sport fish in the Hudson River.
Walleye are now more prevalent than during the 1990s and are now commonly found throughout
the Lower Hudson and in the southern portion of the Upper Hudson. As a sought-after food fish,
walleye may represent a portion of the overall take of fish for human consumption, particularly in
the Lower Hudson. Available data indicate that the PCB concentrations in walleye are 1.5 to 2
times higher than in bass, another commonly sought after game fish, which was the species used
in EPA's risk assessment. EPA needs to update the current understanding of risks posed by fish
consumption given the change in fish species available for consumption. Surveys of people taking
fish from the Hudson would help inform this issue. A comment stated EPA should consider crab
consumption by both humans and other species (birds, fish) in the analysis of human health risks.
Commenters additionally stated that the FYR failed to carefully examine the magnitude or extent
of existing and previously ignored exposure pathways, such as the prevalence of the consumption
of Hudson River fish.
Response
The Revised HHRA completed in 2000 included Walleye as a consumed species. As stated in the
HHRA, the six species from the Connelly et al. (1992) survey that are potentially caught and eaten
in the Upper Hudson River (bass, walleye, bullhead, carp, eel, and perch), were grouped in order
to develop the fish ingestion weights from which the weighted concentration term was developed.
Carp and eel, which are bottom feeders, were grouped with brown bullhead as Group 1. Walleye,
which is similar to bass based on its large size and piscivorous diet, was grouped with the bass as
Group 2. Group 3 is perch, for which yellow perch modeled concentrations were used. Using this
approach, the concentrations of PCBs in fish species that were not modeled (i.e., carp and eel,
walleye and some bass) were approximated based on the two species consumed that were modeled
(brown bullhead and largemouth bass), so that consumption of the non-modeled species could be
included in the species weighted exposure point concentrations (EPCs) which are the
concentrations of PCBs in a given environmental medium at the point of human contact. Table 3-
4 in the HHRA summarizes species-group intake percentages by summing the frequency
percentage of the individual species in each group.
The point estimate EPCs in fish were derived using the species ingestion fractions shown in Table
3-4 multiplied by the PCB concentrations in each of the three modeled fish species. Thus, the point
estimate of the weighted EPC is:
EPC = EPC Group 1 X 0.44 + EPCoroup2 X 0.47 + EPCoroup3 X 0.09
The fish species used to evaluate the EPCs in the HHRA are representative of the species to which
people may be exposed at the Site, and that it is not necessary to perform risk calculations for
additional species. EPA did not calculate risks and hazards from exposure to crabs in the Hudson
River since crab data was not collected at the time of the RI/FS and considering that crabs are only
34
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
found in the New York Bay. In the 1996 and 1991-1992 Hudson Angler Surveys (NYSDOH,
1999b; Barclay, 1993), the NYSDOH conducted a creel survey of Hudson River anglers in 1996
(NYSDOH, 1999b). The Surveys found that Blue crabs were caught only south of Catskill, not in
the Upper Hudson River (NYSDOH, 1999b). In addition, work conducted by EPA at another
Superfund Site in the Newark Bay Complex found that crab consumption was lower than for fish
consumption. For example fish ingestion rates for the Adult in the Hudson River was 31.9 g/day
while the Crab Ingestion Rate for the Newark Bay Complex is 20.9 g/day. The Newark Bay
Complex analysis looked at both fish and crab consumption finding the risks from fish
consumption for the adult was 3 x 10-4 and 7 x 10-5 for crabs. The non-cancer Hazard for fish
consumption was an HQ = 24 for fish consumption and the HQ was 5 for the crab
consumption. EPA found that typically individuals consume either crab or fish so these values are
not combined. In the case that an individual consumes both crab and fish the ingestion rate for
each species is less than those who consume only fish or crab.
The NYSDOH has issued advisories which prohibit consumption of fish from the Upper Hudson
River and recommend strict limitations on consumption of fish from the Lower Hudson River.
EPA is working with NYSDOH to continue to improve the outreach efforts to inform anglers about
the importance of following the advisories and regulations. While the studies show that fish
consumption is occurring, EPA calculates risks from PCB exposure to the reasonably maximally
exposed individual. Additional information about the number of people who consume fish
therefore does not directly affect the risk calculations. Consistent with USEPA's Superfund
guidance, this risk assessment does not estimate the number of anglers that consume their catch or
the number of women of child-bearing age exposed through consumption of fish because
CERCLA requires consideration of risk to an individual with a reasonable maximum exposure. It
would be difficult to identify the number of anglers who are consuming fish in part because of the
presence of fishing bans and fish consumption advisories and because of the potential for
underreporting and the threat of fines for anglers keeping fish from the Upper Hudson River. It is
also not possible to project with any certainty the number of potential anglers within various
stretches of the river who would consume fish if there were no health advisories in the Upper
Hudson River.
EPA included a detailed explanation of the calculations performed for human health and ecological
risks in Appendix 11 of the FYR report. The calculations include an evaluation of the toxicity of
PCBs, the assumed ingestion rate of PCBs from a number of pathways including from eating the
fish, and described the Monte Carlo analysis from the original risk assessment. The appendix also
describes the exposure assumption for eating fish based on the angler surveys used in the risk
assessment.
3.3.3 Comment 4: Assess risks of PCBs in air
Comment
Commenters state that EPA's risk analysis should consider aerosolized PCBs and their cancer and
non-cancer risks. Exposure to aerosolized PCBs has been shown to increase risk of cancer,
hypertension, heart disease, and diabetes. Commenters further state that recent science indicates
that exposure to PCBs through inhalation is a more significant risk than previously believed. The
35
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
risk characterization of the ROD and the intention of the RAOs are primarily intended to control
unacceptable PCB exposures through consumption of contaminated food (i.e. fish). However,
since 2002, the scientific community has documented that exposures to PCBs can occur through
contaminated water, direct skin contact, or breathing contaminated air. A comment also stated that
PCB contamination has migrated south to threaten New York City.
Response
EPA evaluated in the HHRA potential exposures to residents and recreators from exposure to
PCBs in the Hudson River through a range of exposure pathways. The HHRA found that PCBs
that volatilize from the river water may be inhaled by both recreators and residents living near the
river and that the risks were de minimus (U.S. EPA, 2000), i.e., they were significantly lower than
the unacceptable cancer risks and non-cancer health hazards from ingestion of fish. In addition,
EPA conducted extensive sampling of PCBs in air during the dredging, as discussed further in the
FYR report. A discussion of the air data during dredging is provided in Appendix 6 of the FYR
report, along with information regarding the derivation of the toxicity values used in the risk
assessment for air exposures.
At the current time, EPA's Integrated Risk Information System (IRIS) program is updating the
Chemical Assessment for PCB non-cancer toxicity. A component of this assessment will be an
evaluation of the existing scientific studies e.g., animal and human epidemiological studies, to
determine if an inhalation toxicity value can be derived. The evaluation will include systematic
review, evaluation of dose-response based on report exposures, public comment, and external
peer-review consistent with the IRIS process.
This on-going reassessment of non-cancer health effects of PCBs included an October 2014
meeting of independent experts to provide input on the science underlying the development of
IRIS reassessment. The experts discussed key science topics related to the non-cancer toxicity of
PCBs including inhalation. IRIS is currently evaluating the extensive database of non-cancer
toxicity information including inhalation studies.
Upon completion of this evaluation, the IRIS program will continue with the various steps in the
IRIS process including internal Agency review, intra-agency review, external peer-review with a
response to comments and updates to the report, and finally release of the document. Any updates
to the IRIS chemical file for non-cancer PCBs will be evaluated as part of answering FYR -
Question B (are the exposure assumptions, toxicity data, cleanup levels, and RAOs used at the
time of the remedy still valid?).
EPA evaluated exposures to volatilized PCBs in the 2000 HHRA and found the risks were
significantly lower than the unacceptable risks and hazards from ingestion of fish. EPA is currently
evaluating available studies on PCB exposures through inhalation as part of the IRIS reassessment
for non-cancer health effects. The IRIS assessment will include evaluation of the dose-response
and exposures to determine whether existing toxicity values need to be updated. EPA provided
copies of the studies identified by commenters to the IRIS staff for consideration during the current
reassessment of non-cancer PCB toxicity. EPA will re-evaluate the impacts of this reassessment
in future FYRs when the IRIS assessment is completed.
36
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Regarding areas of the Lower River, the objective of the HHRA for the mid-Hudson river was to
quantitatively evaluate current and potential cancer risks and non-cancer health hazards from river
water, sediment, and fish in the Mid-Hudson River. This HHRA provides estimates of risks both
to the Reasonable Maximum Exposure (RME) individual, or high-end risk (>90th to 99th
percentiles), and to the Average Exposed Individual, or central tendency cancer risks and non-
cancer health hazards (50th percentile). Since the Phase 1 Risk Assessment, USEPA has conducted
extensive modeling efforts in order to forecast PCB concentration trends in environmental media
in the Mid-Hudson River region (USEPA, 2000a; 2000f; 2000g). The results from these model
forecasts were incorporated into this Phase 2 assessment. EPA also plans additional data collection
and supplemental studies of the Lower River which will further inform our understanding of the
extent of PCB contamination in that portion of the river.
3.3.4 Comment 8: EPA did not investigate the potential for links to autism in the first five-
year review
Comment
CoAl016mmenters state that the EPA did not investigate potential links between PCB exposure
and autism within the numerous Hudson River communities. They indicated that given the high
and increasing prevalence of autism and its seriousness and apparent linkage to environmental
agents that may include maternal exposure to PCBs during pregnancy, that the project success
should be evaluated with this consideration.
Response
EPA's response to Question B in the FYR report evaluated whether existing data would change
the overall outcome of the HHRA. The data evaluated included peer-reviewed documents on
exposure assessment and plans for updating the Integrated Risk Information System database
(IRIS). As explained in the FYR report, updates to the exposure assumptions in the ROD do not
change the conclusions of the HHRA or the protectiveness of the remedy. EPA is re-evaluating
the non-cancer toxicity information and any updates will be evaluated in the next FYR.
One commenter states that there is an "emerging link between PCBs and possible causation of
autism." The commenter also states that EPA has "neither addressed this issue substantively, nor
alluded to it." EPA disagrees with the commenter's suggestion that EPA did not address potential
links between autism and PCBs. The response to Question B in the FYR highlights the upcoming
updates to the non-cancer toxicity values, it is expected that these updates will include
consideration of autism.
The following text provides additional information on this topic and EPA's approach to evaluating
the toxicity of PCBs.
• EPA relies on the IRIS as the primary source of toxicity information in the Superfund
program to evaluate cancer risks and non-cancer toxicity. The IRIS program represents
that Agency's consensus toxicity information database for over 500 chemicals.
37
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
• Currently, the IRIS program has non-cancer Reference Doses (RfDs) for Aroclors 1016
and 1254 (A1016 and A1254) and a cancer assessment, including a Weight of Evidence
that PCBs are a probable human carcinogen, for total PCBs. Information on IRIS is
available at www.epa.gov/iris.
• The response to Question B in the FYR discusses the on-going reassessment of non-
cancer toxicity for PCBs by toxicologists in EPA's IRIS program. The IRIS reassessment
is evaluating thousands of studies on PCB toxicity using the Systematic Review process.
The systemic review process will evaluate a large number of health endpoints that
include autism, based on the available literature.
• The information from the systemic review will be used by the IRIS program to evaluate
dose-responses for a number of diseases. Based on the evaluation, IRIS will determine if
there is adequate information to update the current oral RfD or develop a new inhalation
Reference Concentration (RfC).
This on-going assessment process is discussed on the IRIS webpage (www.epa.gov/iris) which
includes documentation of the March 2014 draft literature searches and associated search
strategies, evidence tables, and exposure response arrays for PCBs as a means to obtain input from
stakeholders and the public. The literature search strategy, which describes the processes for
identifying scientific literature, contains the studies that EPA considered and selected to include
in the evidence tables. The preliminary evidence tables and exposure-response arrays present the
key study data in a standardized format. The evidence tables summarize the available critical
scientific literature. The exposure-response figures provide a graphical representation of the
responses at different levels of exposure for each study in the evidence table.
EPA also held a meeting of scientific experts in the field of PCB toxicity on June 17-18, 2015 to
discuss information on PCB toxicity (see https://www.epa.gov/iris/iris-public-meeting-jun-2015).
As the process progresses, EPA will make information available on the webpage
www.epa.gov/iris.
EPA's response to Question B in the five-year review acknowledges these efforts by the IRIS
process and indicates that once the IRIS process is completed, EPA will re-evaluate the non-cancer
toxicity risks at the Site with new reference dose numbers that are applicable. At this point in time,
it is premature to prejudge the outcome of the assessment and the study and health endpoint that
will be selected. Any changes to the toxicity values will also be evaluated in future Five Year
Reviews based on the completion of the IRIS reassessment for non-cancer toxicity.
3.3.5 Comment 11: EPA must calculate the risks of dioxin contamination (or dioxin-like
congeners)
Comment
EPA did not address dioxin or heavy metal contamination within the Hudson River. EPA should
calculate these risks and until then, the Hudson River cannot be considered remediated.
38
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Response
The commenter raised concerns regarding the evaluation of dioxins at the site. As part of the
Revised HHRA EPA 2000), EPA evaluated cancer risks from exposure to dioxin-like PCBs
following the guidance provided in the 1996 PCB Cancer Reassessment. The assessment of
dioxin-like PCBs did not find an enhancement of risk associated with the dioxin-like PCBs. EPA
as part of this FYR, evaluated the new toxicity information for dioxins using the approach in the
1996 PCB Cancer Reassessment and did not find enhancement of risk from dioxin-like PCBs. The
results of this analysis did not change the conclusions from the HHRA.
In addition, the HHRA describes the process used to identify the chemicals of potential concern
for the site and indicated that dioxins and other chemicals were not included as contaminants-of-
concern because fish collected by NYSDEC found those other chemicals to be present either at
very low levels or below detection limits. As discussed in the Revised HHRA:
A typical baseline Superfund risk assessment includes an evaluation of those chemicals at a
contaminated site that pose a potential health concern, or chemicals of potential concern (COPCs).
In the HHRA, PCBs were identified as the COPCs and later as chemicals of concern (COC),
because the HHRA was being conducted as part of EPA's Reassessment of its 1984 interim No
Action decision for the PCB-contaminated sediments in the Upper Hudson River, and because
PCBs in fish tissues were detected at greater concentrations than other contaminants. As discussed
in the Revised Baseline Modeling Report, in addition to monitoring for PCBs, fish collected by
NYSDEC at the site were analyzed for total dichlorodiphenyltrichloroethane (DDT), total
chlordane, total endrin, total endosulfan, dieldrin, aldrin, mirex, total heptachlor, total
hexachlorobenzene, toxaphene, methoxychlor, individual polycyclic aromatic hydrocarbons
(PAHs), cadmium, mercury, dioxins, and dibenzofurans. These compounds were found to be
present at relatively low levels or below detection limits (Sloan, 1999), confirming that PCBs are
the primary COCs in the Hudson River. Consequently, no screening of COPCs was performed
during the Revised HHRA for this assessment.
Contamination in the Lower Hudson
The commenter states that there is a need to consider exposures from PCBs in the Lower Hudson
River. The HHRA for the Mid-Hudson River (included in the HHRA) quantifies both
carcinogenic and non-carcinogenic health effects from exposure to PCBs following EPA risk
assessment policies and guidance. The Mid-Hudson was identified as the area between the Federal
Dam in Troy, New York (River Mile 153.5) going south to the salt-water front at approximately
River Mile 64. Both current and future cancer risks and non-cancer health hazards to young
children, adolescents, and adults were evaluated based on the assumption of no remediation or
institutional controls, in accordance with the National Contingency Plan, 40 CFR Part 300. The
HHRA for the Mid-Hudson found the Reasonable Maximum Exposure (RME) cancer risks and
non-cancer health hazards from ingestion of fish in the Mid-Hudson are about one-half the cancer
risks and non-cancer Hazard Indices determined for ingestion of fish in the Upper Hudson.
As indicated in the FYR report, EPA agrees that is important to collect additional data and conduct
supplemental studies to better understand the PCB contamination in the Lower Hudson River,
which includes the Mid-Hudson.
39
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.3.6 Comment 16: EPA should finalize the study done on black bass
Comment
Two commenters recommended that EPA finalize the 2014 black bass DEC standard fillet vs rib-
out study, so it can potentially be peer reviewed and the paper can then be cited in documents by
the Natural Resource Damage Trustees.
Response
EPA determined that fish collected and processed between 2007 and 2013 by GE did not have the
ribcage included as part of the fillet as required by project documents. In response to this deviation
from the QAPP, EPA required that GE perform a special study that would facilitate evaluation of
whether or not inclusion of the rib cage (ribs) had a significant impact on fish tissue PCB
concentrations and lipid levels. Black bass (smallmouth bass and largemouth bass) were the focus
of the 2014 study because they are large enough to produce fillets of sufficient size for comparison,
were generally considered to be representative of other species and are collected from monitoring
stations in the Upper and Lower Hudson River.
EPA completed its evaluation of GE's 2014 special study on black bass and provided the report as
Appendix D of this document. The preliminary results of the analyses were shared with
commenters, project stakeholders, and the Hudson River Community Advisory Group in October
2015. EPA's analyses found that on a wet-weight basis, the difference between fillets prepared
with and without ribs was variable and could be greater than a factor of two. For lipid normalized
data, the difference between the two fillet approaches averages less than 20 percent. As a result,
EPA determined that comparison of lipid normalized results from fillets prepared with and without
ribs could be conducted, but that results should be evaluated cautiously. It should be noted that the
period of data collection when the rib was not included (2007 to 2013) was just prior to and during
dredging activities. Fish tissue PCB concentrations observed during dredging were impacted by
dredging-related PCB resuspension and are not useful for establishing post-dredging fish recovery
trends. No significant project decisions were made or altered based on fish data from the years
when the ribcage was not included in the fillet samples. Also, no adjustments to fish advisories or
regulations were made by New York State based on those data.
3.3.7 Comment 18: EPA should look for updated information on the toxicity of PCBs
Comment
Commenters state that EPA has failed to acknowledge any new information related to exposure
assumptions or toxicity data that could impact the human health risk assessment, siting that recent
science indicates that PCBs are more toxic than previously thought. Commenters note concern that
EPA is still classifying PCBs as probable human carcinogens (in the Integrated Risk Information
System [IRIS] listing) with a cancer weight-of-evidence classification B2, whereas the
International Agency for Research on Cancer, of the World Health Organization, has now listed
PCBs as a known human carcinogen.
40
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
In addition, dioxin-like PCBs can now be evaluated via EPA's listing of non-cancer endpoints for
dioxin via the reference dose in EPA's IRIS as well as several additional toxicological endpoints
which have been updated in terms of health effects. All of this information adds to the growing
body of research which demonstrates that PCBs are more toxic to humans than previously believed
when the human health risk assessment was being developed for the ROD. As a result, the FYR
report needs to address the greater toxicity as a change in assumptions and new information that
was not available at the time the ROD was developed.
Commenters also indicated that there is not sufficient data available to evaluate if the cleanup
levels in the ROD are still valid; to determine if the exposure pathways used in the risk assessments
are still valid (due to changes in fish species distribution, and in population demographics among
human fish consumers); and to determine if the toxicity assumptions are still valid, as EPA has not
yet completed the anticipated update to the IRIS database for PCBs.
Commenters noted that PCBs pose a significant risk to public health including cancer,
cardiovascular disease, and cognitive and development disorders in children while some also noted
that the link from PCBs to health impacts has not been proven. When considering remedies to
address PCB contamination in the Hudson, the EPA determined that cancer and non-cancer health
risks were well above the acceptable risk range for people who ate fish from both the Upper and
Lower Hudson. The Superfund cleanup remedy was intended to address the risks, and the FYR is
intended to ensure the risks have been adequately addressed.
Response
Commenters stated that PCBs pose a significant risk to public health. EPA uses risk assessment
to inform decisions under CERCLA, commonly referred to as the Superfund law. The goal of the
risk assessment is not to predict specific diseases but rather to assess risks to support risk
management decisions. The risk assessment provides a methodology for evaluating current and
future risks under specific exposure assumptions for different age ranges (e.g., young child through
adult) and different activities (e.g., outdoor workers, construction workers, recreational users, and
residents). Both human health and ecological risk assessments were developed for the Hudson
River. The documents were externally peer-reviewed and were updated to reflect the comments
from the peer-reviewers before the final HHRA was issued in November 2000. The HHRA was
a component of the decision to take remedial action at the site. Currently, EPA does not plan to
conduct additional risk assessments as discussed in the response to the FYR Question B: Are the
Exposure Assumptions, Toxicity Data, Cleanup Levels, and RAOs Used at the Time of the
Remedy Still Valid? At the next FYR and subsequent FYRs, EPA will review updates to toxicity
values and exposure assumptions to determine the need to update the risk assessment.
EPA conducts evaluations of the toxicity of chemicals, such as PCBs, through IRIS. The webpage
www.epa.gov/iris provides IRIS assessments for carcinogenicity (for total PCBs) and non-cancer
toxicity (Aroclors 1016 and 1254 [A1016 and A1254]). The IRIS process involves the evaluation
of a large number of studies of the toxicity of the chemical including evaluation of the chemicals
potential to cause cancer and non-cancer health effects.
41
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
At the current time, EPA is re-evaluating the non-cancer toxicity of PCBs following the IRIS
process outlined in the graphic below. The update includes an evaluation of the available
published scientific literature, development of dose-response information where data is adequate,
and the development of toxicity values. When the process is completed, any updates to the non-
cancer toxicity values will be evaluated in the next FYR in Question B.
In 2000, EPA conducted an HHRA to support the decision to take action at the Site. The HHRA
found that the cancer risks exceeded the risk range of 1 x 10"4 to 1 x 10"6 (risk of one in ten thousand
to one in a million) established under the NCP. The HHRA also found that the non-cancer hazards
exceeded the goal of protection of a Hazard Quotient (HQ) or a Hazard Index (HI) = 1. Fish
consumption was the risk driver; other exposure pathways posed risks within the risk range and
below a HI = 1.
EPA is working with NYSDOH to inform anglers fishing in the Hudson River of the NYSDOH
Fish Consumption Advisories. EPA will continue these efforts and share fish sampling results
with NYSDOH and NYSDEC to inform the on-going need to maintain or modify the fish
consumption advisories. (See Appendices 3 and 11).
Future FYRs will review the on-going NYSDOH outreach program for the fish consumption
advisories and the concentrations of PCBs in fish necessary to inform the fish advisories.
Commenters noted that the link from PCBs to health impacts has not been proven. EPA studied
the effects of PCBs on humans, including evaluation of cancer and non-cancer health effects. IRIS
provides the Agency's consensus database for toxicity information used in assessments of risks
and hazards at Superfund sites. The IRIS assessments for PCBs classify PCBs as a probable human
carcinogen based on limited human evidence and adequate animal evidence. The IRIS document
summary of the cancer hazards to humans is described below.
• A cohort study by Bertazzi et al. (1987) analyzed cancer mortality among workers at a
capacitor manufacturing plant in Italy. PCB mixtures with 54%, then 42% chlorine were
used through 1980. The cohort included 2,100 workers (544 males and 1556 females)
employed at least 1 week. At the end of follow-up in 1982, there were 64 deaths reported,
26 from cancer. In males, a statistically significant increase in death from gastrointestinal
tract cancer was reported, compared with national and local rates (6 observed, 1.7
expected using national rates, standardized mortality ratio [SMR]=346, confidence
interval [CI]=141 -721; 2.2 expected using local rates, SMR=274, CI=112-572). In
females, a statistically significant excess risk of death from hematologic cancer was
reported, compared with local, but not national, rates (4 observed, 1.1 expected,
SMR=377, CI=115- 877). Analyses by exposure duration, latency, and year of first
exposure revealed no trend; however, the numbers are small.
• A cohort study by Brown (1987) analyzed cancer mortality among workers at two
capacitor manufacturing plants in New York and Massachusetts. At both plants, the
Aroclor mixture being used changed twice, from 1254 to 1242 to 1016. The cohort
included 2,588 workers (1,270 males and 1,318 females) employed at least 3 months in
areas of the plants considered to have potential for heavy exposure to PCBs. At the end of
follow-up in 1982, there were 295 deaths reported, 62 from cancer. Compared with
42
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
national rates, a statistically significant increase in death from cancer of the liver, gall
bladder, and biliary tract was reported (5 observed, 1.9 expected, SMR=263, p<0.05).
Four of these five occurred among females employed at the Massachusetts plant.
Analyses by time since first employment or length of employment revealed no trend;
however, the numbers are small.
• A cohort study by Sinks et al. (1992) analyzed cancer mortality among workers at a
capacitor manufacturing plant in Indiana. A1242, then A1016, had been used. The cohort
included 3,588 workers (2,742 white males and 846 white females) employed at least 1
day. At the end of follow-up in 1986, there were 192 deaths reported, 54 from cancer.
Workers were classified into five exposure zones based on distance from the
impregnation ovens. Compared with national rates, a statistically significant excess risk
of death from skin cancer was reported (8 observed, 2.0 expected, SMR=410, CI=180-
800); all were malignant melanomas. A proportional hazards analysis revealed no pattern
of association with exposure zone; however, the numbers are small, looked for an
association between occupational PCB exposure and cancer mortality. Because of small
sample sizes, brief follow-up periods, and confounding exposures to other potential
carcinogens, these studies are inconclusive.
• Accidental ingestion: Serious adverse health effects, including liver cancer and skin
disorders, have been observed in humans who consumed rice oil contaminated with PCBs
in the " Yusho" incident in Japan or the " Yu-Cheng" incident in Taiwan. These effects
have been attributed, at least in part, to heating of the PCBs and rice oil, causing
formation of chlorinated dibenzofurans, which have the same mode of action as some
PCB congeners (Safe, 1994).
Since the IRIS assessment was finalized, there have been additional studies of worker exposures to
PCBs published in the scientific literature, documenting the risks.
Some commenters noted that the change in designation by the International Agency for Research
on Cancer (IARC) as a "known carcinogen" is justification for new calculations on risk, with some
commenters stating that this should be considered "new information". EPA is re-evaluating the
non-cancer toxicity from exposure to PCBs. The toxicity values in the original HHRA have not
changed and these values are listed in the IRIS chemical files available on the database. Upon
completion of the reassessment, EPA will evaluate the impacts of the reassessment in a future
FYR.
The Reassessment of PCB non-cancer toxicity is a multi-level evaluation involving a number of
steps listed below including internal review and external peer-review. In addition, the toxicity of
PCBs for non-cancer health effects includes thousands of studies that will be reviewed using
Systematic Review by the IRIS program. As such, it is anticipated that the review process and
development of the next toxicity value will take significant time to complete. The progress on
developing this toxicity value will be evaluated in future FYRs (see figure 18-1 below).
43
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Review Finalize
ft
Scoping and
Problem Formation
• Scoping: Identify needs
of EPA's program and
regional offices
• Problem formulation:
Frame scientific
questions specific to the
assessment
Draft Development
Apply principles of
systematic review to:
• Identify pertinent studies
• Evaluate study methods
and quality
• Integrate evidence for
each health outcome
• Select studies for
deriving toxicity values
• Derive toxicity values
n
Agency Review
Review by health
scientists in EPA's
program and regional
offices
©
Interagency Science
Consultation
Review by other federal
agencies and Executive
Office of the President
Public Comment
Release for public review
and comment
External Peer
Review
Release for independent
external peer review
Revise Assessment
Address peer review and
public comments
I
(^Final Agency Review
and Interagency
Science Discussion
Discuss with EPA health
scientists and with other
federal agencies and
Executive Office of the
President
t
I
Post Final
Assessment
Post to IRIS website
IRIS Assessment Development Process
The 7-step process has not changed. This figure refines earlierversions and includes the2013 IRIS enhancements and the in corporation of systematic review approaches.
Figure 18-1 IRIS Assessment Development Process
There is no new toxicity information that would change the calculated cancer risks and non-cancer
hazards. As such, updates to the calculated risks/hazards are not necessary. EPA will monitor the
progress of the updates to the IRIS non-cancer toxicity assessment for PCBs.
The FYR guidance calls for updating the risk assessment if new information is available that will
change the results of the human health risk assessment. The cancer risks and non-cancer hazards
are representative of the exposures to the Reasonably Maximally Exposed (RME) indivi dual. The
information identified in the comment does not provide specific information and references to
peer-reviewed studies necessary to evaluate if there are any changes in the exposure assumptions
that would change the conclusions of the HHRA. As discussed in previous responses, EPA
evaluated risks/hazards from exposure to Walleye as part of the HHRA.
The Superfund program relies on IRIS and the 2011 Exposure Factors Handbook and Superfund
Standard Default Exposure Assumptions. IRIS provides the Agency's consensus toxicity database
for use in assessing cancer risks and non-cancer toxicity. Another component of the risk
assessment is the exposure assessment that evaluates the routes of exposure for various receptors
(e.g., resident, recreator, outdoor worker, construction worker, etc.) and age groups. The combined
information on exposure and toxicity are used to calculate risks. At the current time, IRIS is
updating the non-cancer toxicity values for PCBs including evaluation of inhalation toxicity.
44
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Based on the extensive number of studies on PCB toxicity in the scientific literature this review
will include a number of scientists at EPA with expertise in this area. In addition, the draft
document will go through internal Agency review, public comment, and external peer-review
before the final document is available for application in HHRAs. The development of the exposure
assessment for this FYR relied on information from EPA's 2011 Exposure Factors Handbook that
is updated on an on-going basis, and the Superfund Standard Default Exposure Assumptions. Both
documents were evaluated in responding to Question B to determine if any new information in the
published scientific literature would require changes in the 2000 Hudson River Risk Assessment
and its conclusions. This review did not identify any new information that would require updating
of the HHRA.
3.3.8 Comment 19: EPA should qualify the 2016 spring and fall data properly according to
the impacts expected by the dredging
Comment
The 2016 spring sport fish in the Upper Hudson (black bass, bullhead, perch) should be assessed
as being impacted by the dredging work which ended in 2015, as the trend in fish PCB data
indicates that the spring fish represent the previous years' conditions. The fall 2016 forage fish,
however, should indicate the first year of post-dredging conditions, as they went through an entire
growth season in 2016 without dredging impacts.
Response
The spring 2016 sport fish are the first fish collected after completion of the dredging activities.
Nonetheless, EPA agrees that the spring sport fish obtained in 2016 likely included fish that were
exposed to dredging impacts. Those collected in the fall are further removed from direct dredging
activities, although in both cases it is not fully known to what degree the fish were exposed to
dredging-related conditions prior to when they were caught. EPA agrees the fall 2016
pumpkinseed and forage fish, including young-of-the-year fish, represent the first sampling and
analysis of fish that were not directly exposed to conditions during dredging (assuming those fish
were born after dredging completed). However, both spring and fall 2016 data sets may have been
influenced by dredging-related impacts. While the spring fish, which are largely adults, were
present in the river during dredging, the young-of-the-year fish from the fall of 2016 could still
have been exposed to dredging-related disturbances, such as transport of unconsolidated surface
sediments that remained after dredging, or from sediment disturbances related to habitat
replacement and reconstruction activities, which continued throughout 2016. EPA maintains that
further monitoring of fish levels will be important to assess post-dredging recovery. This condition
supports the need and importance of fish monitoring for the foreseeable future.
With limited fish, water, and sediment data post-dredging, it is unclear exactly what riverine
conditions the 2016 data represent. EPA considers 2016 to be a transitional year within a re-
equilibration period for the system, which was anticipated by the ROD. The comment serves to
further emphasize the need for additional data before trends in recovery of water, sediment and
fish can be further evaluated.
45
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.3.9 Comment 20: EPA should recalculate human health risks
Comment
Commenters state that the EPA should recalculate risks to specific populations — specifically,
populations in New York City, and communities along the Hudson River that may still face the
same health threats as they did prior to dredging. Commenters indicate that EPA should determine
if the changes in fish species availability for consumption, and changes in community population
demographics, result in a significant change to the risk assessment inputs and results.
Commenters state that the EPA should recalculate the risks to human health based on the fact that
two or more times more contamination than previously estimated was found in the Hudson River.
Specifically, the EPA should consider the links between PCB exposure and impacts to the health
of children, and whether the dredging project should be extended to remediate remaining PCBs.
EPA should be conservative, not only in protecting the scientific knowledge base, but in protecting
public health.
Response
The externally peer reviewed revised HHRA describes the risk assessment process and how it was
applied to the Hudson River. The report is available at:
https://www3.epa.gov/hudson/revisedhhra-text.pdf. The goal of the risk assessment is to
determine the need to take action at the Superfund site. The FYR considered a broad range of data,
including any changes in toxicity or exposure information that may impact the protectiveness
determination.
EPA's response to Question B in the FYR evaluated whether existing data would change the
overall outcome of the Revised HHRA. The data evaluated included peer-reviewed documents on
exposure assessment and plans for updating the Integrated Risk Information System (IRIS)
database, discussed below. As explained in the FYR report, updates to the exposure assumptions
in the ROD do not change the conclusions of the HHRA or the protectiveness of the remedy. EPA
is re-evaluating the non-cancer toxicity information for PCBs and any updates will be evaluated
in the next FYR.
The following is additional information regarding EPA's approach to evaluating the toxicity of
PCBs:
• EPA relies on IRIS as the primary source of toxicity information in the Superfund
program to evaluate cancer risks and non-cancer toxicity. The IRIS program represents
the Agency's consensus toxicity information database for over 500 chemicals.
• Currently, the IRIS program has non-cancer Reference Doses (RfDs) for Aroclors 1016
and 1254 (A1016 and A1254) and a cancer assessment, including a Weight of Evidence
that total PCBs are a probable human carcinogen. Information on IRIS is available at
www.epa.gov/iris.
46
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
• The response to Question B in the FYR discusses the ongoing reassessment of non-
cancer toxicity for PCBs by toxicologists in EPA's IRIS program. The reassessment will
evaluate thousands of studies on PCB toxicity using the systematic review process, which
will evaluate a large number of published studies on the effects of PCBs. The report
"Scoping and Problem Formulation for the Toxicological Review of Poly chlorinated
Biphenyls (PCBs): Effects Other Than Cancer" (EPA/635/R-14/198) provides a
preliminary survey of the literature conducted in 2015. A preliminary list of broad health
effect categories in which effects were observed and for which there may be enough data
to further evaluate specific health endpoints includes: cardiovascular, dermal and ocular,
developmental effects on growth and maturation, endocrine, gastrointestinal,
hematological, hepatic, immunological, metabolic, neurological, and reproductive effects.
As the reassessment moves forward this evaluation of the literature may be updated to
incorporate newer studies
This information from the systematic review will be used by the IRIS program to evaluate
dose-responses for a number of diseases. Based on the evaluation, IRIS will determine if
there is adequate information to update the current oral RfD or develop a new inhalation
Reference Concentration (RfC).
The ongoing assessment process is discussed on the IRIS webpage (www.epa.gov/iris). including
documentation of the March 2014 draft literature searches and associated search strategies,
evidence tables, and exposure response arrays for PCBs as a means to obtain input from
stakeholders and the public prior to developing the draft IRIS assessments for PCBs. The literature
search strategy, which describes the processes for identifying scientific literature, contains the
studies that EPA considered and selected to include in the evidence tables. The preliminary
evidence tables and exposure-response arrays present the key study data in a standardized format.
The evidence tables summarize the available critical scientific literature. The exposure-response
figures provide a graphical representation of the responses at different levels of exposure for each
study in the evidence table. EPA also held a meeting of scientific experts in the field of PCB
toxicity on June 17-18, 2015 to discuss information on PCB toxicity (see
https://www.epa.gov/iris/iris-public-meeting-jun-2015). As the process progresses, EPA will
make information available on the webpage www.epa.gov/iris. The FYR, Question B,
acknowledges these efforts by the IRIS process indicating that EPA will re-evaluate the non-cancer
toxicity information as part of the reassessment for non-cancer that is anticipated to include
updates to the oral RfD. It is premature to prejudge the outcome of the assessment and the study
and health endpoints that will be selected. Any changes to the toxicity values will be evaluated in
future FYRs based on the completion of the IRIS reassessment for non-cancer toxicity.
3.3.10 Comment 21: EPA should require GE to conduct an RI/FS of the Lower Hudson
River
Comment
A number of commenters asserted that EPA should require GE to complete a RI/FS of the Lower
Hudson River (LHR). The commenters cited the significant magnitude and long duration of GE's
PCB releases as evidence that there is a significant need to investigate contamination to the Lower
47
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Hudson. Commenters asserted that while other PCB sources do exist in the Lower Hudson, EPA
has stated in public meetings that GE was the primary contributor. Commenters stated that the
remedy has had little to no beneficial impact on the Lower Hudson to date. This is demonstrated
by decay rates of PCB concentrations in fish tissue that are not statistically different from zero.
Additionally, a number of government agencies have published findings that substantial PCB
contamination remains within the river, necessitating study and remediation.
Commenters also indicated a study of the LHR must include a historical study, establishing the
extent and elevation of the river throughout its course prior to and during the PCB dumping, as
well as the evolution of the navigational channel within the Lower River. This historical study will
reveal areas that are not currently part of the Hudson River and were not included in the initial risk
assessment and monitoring, such as deposits behind dykes and above the current high tide line.
Other commenters expressed concern with waterways/access points to the River besides the main
river channel. Specifically, individuals who own marinas in the LHR have to pay a large cost to
dredge out their marinas due to contaminated sediment. Some commenters suggested that GE
should pay for this type of cleanup. It was stated that these marinas in the LHR, which cannot
dispose of contaminated dredge spoils economically, will be impacted for a much longer
timeframe than indicated in the FYR report.
Response
As stated in the FYR report, data collected from the LHR indicate that the LHR is not recovering
as quickly as the Upper Hudson River (UHR). This suggests that the declining PCB concentrations
in the water of the UHR may have less of an impact downstream of the project area than
anticipated. EPA agrees that much of the PCB contamination of sediments in the Lower Hudson
originated from GE releases from the Upper Hudson. As part of the remedial investigation that led
up to the OU2 ROD, EPA collected a series of high-resolution cores in both the Upper and Lower
Hudson. The analysis of these cores demonstrated that at the time of the coring study (1992),
Lower Hudson sediment PCB patterns and, therefore PCB inventory, could be attributed to Upper
Hudson GE releases as far south as RM 50. Below RM 50 there were other notable sources of
PCBs to the river. It should also be noted that there are other known ongoing and historic PCB
releases from several PCB contaminated sites above RM 50 in the LHR. The furthest upstream
source is near Albany. Understanding and resolution of all of these sources of PCBs creates
challenges and uncertainty for fish recovery in the LHR. EPA has met with NYSDEC and
discussed other sources the state is aware of. To better understand how PCBs in the UHR affect
water, sediment and fish recovery in the LHR, more information/data will need to be collected.
EPA has informed the public that it is important to collect additional data and conduct
supplemental studies in order to better understand the PCB contamination in the LHR. The specific
obligations of GE and any other parties with respect to the Lower Hudson will be defined as that
process proceeds.
EPA expects that the supplemental studies of the LHR will start in 2019 and will take several years
to complete. These studies will supplement information collected during EPA's investigation of
the LHR in the 1990s, the routine monitoring of LHR fish and water by GE under EPA oversight
since 2004, and the periodic monitoring of LHR fish by New York State. The supplement studies
48
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
will inform the need for a RI/FS. Such an RI/FS, if undertaken, would likely be extensive and
complex and could take a number of years. It is too early in the process to determine if a cleanup
is needed in the LHR. Based on the studies completed, EPA would decide whether remedial work
is called for; such a decision would be made after an opportunity for public review and comment.
EPA will continue to work closely with NYSDEC, the Hudson River Natural Resources Trustees
(NYSDEC, FWS, NOAA) and other stakeholders to determine what additional supplemental
studies are necessary to further evaluate how the LHR and UHR are linked and how sediment
contamination in the Lower River will affect water column, sediment and fish tissue PCB recovery
over time. PCB loads from the Upper Hudson to the Lower Hudson are expected to continue to
decrease and natural attenuation recovery will continue for the entire Hudson River system.
3.3.11 Comment 24: EPA should indicate the current state of testing and analysis of human
health impacts for users of the river
Comment
Commenters state that the EPA should test people for the presence of PCBs who work, live, or
play on or near the Hudson River. Testing should be conducted over a period of years (decades).
All test results and associated reports should be made available to the public.
Response
The CERCLA (i.e. the Superfund law), does not provide authority for EPA to conduct human
studies such as evaluation of blood PCB levels in populations. The Superfund law established the
ATSDR which conducts such population studies at specific sites where ATSDR determines a need
for such an analysis. ATSDR has not determined that population testing for PCBs is needed at the
site.
Other federal Agencies such as the CDC also conduct ongoing research on blood PCBs levels
across the U.S. population through the National Health and Nutrition Examination Survey
(NHANES) (https://wwwn.cdc.gov/Nchs/Nhanes/2009-2010/PCBPOL_F.htm). State agencies
may also conduct such studies, through grants from the National Institute of Environmental Health
Sciences, EPA, or National Institute of Health.
EPA's evaluation of current and future risks supporting the decision to take action and the remedial
goals for PCBs in fish were based on the peer-reviewed HHRAs, which are available at
https://www3.epa.gov/hudson/reports.htm. EPA uses risk assessment as a tool to evaluate the
likelihood and degree of chemical exposure and the possible adverse health effects associated with
such exposure. The basic steps of the Superfund HHRA process are the following: 1) Data
Collection and Analysis to determine the nature and extent of chemical contamination in
environmental media, such as sediment, water, and fish; 2) Exposure Assessment, which is an
identification of possible exposed populations and an estimation of human chemical intake through
exposure routes such as ingestion, inhalation, or skin contact; 3) Toxicity Assessment, which is an
evaluation of chemical toxicity including cancer and non-cancer health effects from exposure to
49
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
chemicals; and 4) Risk Characterization, which describes the likelihood and degree of chemical
exposure at a site and the possible adverse health effects associated with such exposure.
A component of the HHRA is the evaluation of the toxicity of the chemical. EPA's Integrated
Risk Information System (IRIS), a consensus database of toxicity information used to support
decisions at Superfund sites, and across the Agency, provides information on both cancer and non-
cancer toxicity information. A component of the assessment is studies in animals exposed under
laboratory conditions to PCBs, including information on the blood PCB levels in the animal used
in the studies (e.g., two year cancer bioassays in rats) for the cancer assessment and Rhesus
monkeys for the non-cancer toxicity assessment. Both studies showed direct linkages between the
blood PCB levels and health effects. The IRIS chemical files for PCBs (cancer assessment) and
Aroclor 1016 (A1016) and Aroclor 1254 (A1254) are available at:
https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=294.
https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=462. and
https://cfpub.epa.gov/ncea/iris2/chemicalLanding.cfm7substance nmbr=389. respectively.
3.3.12 Comment 28: EPA will not reach the target levels as anticipated in the ROD
Comment
Several reviewers commented on the time for fish tissue concentration to reach anticipated levels
specified in the ROD. The ROD states that the time to reach target PCB concentrations in fish was
a primary factor in the comparison of remedial alternatives. Comments state that EPA predicted
that the dredging remedy would result in rapid reductions in PCB levels in fish so that might allow
for fish consumption restrictions to be relaxed in five to ten years, as opposed to many decades as
is now predicted. Commenters also believe that the recovery rate of fish tissue is lower than the 8
percent presented in the FYR report. Given the recovery rates anticipated and derived from the
data and the recent 2016 fish tissue data, several have commented that it will take several
additional decades for ROD targets to be achieved.
Response
The first year of post-dredging data (2016) provides the baseline for the post-dredging monitoring
period, and additional data and time are required after the remediation to assess when fish tissue
concentrations will achieve the goals set in the ROD. The ROD anticipated at least a year of
equilibration in the system in response to remedial activities. Another site, as discussed below, has
taken several years to reach post-dredging equilibrium. It is expected that after the post-dredging
equilibration period, the system will then follow natural recovery trends. An accurate
determination of the time to reach specific ROD targets and goals requires information on the
equilibrated PCB concentration following dredging. EPA therefore disagrees with commenters
who used the 2016 PCB fish tissue as the post-dredging equilibration concentration and applied
various recovery rates to determine the time frame to reach certain fish tissue targets.
Post-dredging equilibration over several years has also been observed in other remedial sites. For
example, as described in Section 2.7 of Appendix 8 of the FYR report, at Cumberland Bay (Lake
Champlain, New York), fish tissue PCB levels were observed to require several years to recover
50
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
in the wake of a removal action (NYSDEC 2012). At the Cumberland Bay (Wilcox Dock) Site,
the Wilcox Dock remediation was implemented by NYSDEC in 1999 and 2000, and fish tissue
concentrations for two species, including rock bass and yellow perch, indicated that several post-
dredging years passed before concentrations began to stabilize. Following stabilization, fish tissue
concentrations recovered at a rate of about 25 percent per year.7 While there is limited pre-
dredging data available for the Wilcox Dock Site, it is likely that the pre-dredging recovery rate
must have been significantly lower than the estimated post-dredging recovery rate. Overall, these
observations suggest that some time is required for remedial sites to undergo equilibration before
it is reasonable to expect the ultimate post-dredging trend to decline towards remedial goals and
target PCB levels.
The observed pre-dredge Hudson River MNA recovery rate in fish was approximately 8 percent
per year. The post-remedial recovery rate is expected to incorporate a significant adjustment to the
new post-dredge river conditions followed by the period of continued MNA. Although the
recovery was forecast by the HUDTOX and FISHRAND models, there are few examples of
remedial actions at the magnitude of the Hudson River project where sufficient time has passed
since remediation to develop a robust understanding of how sediment contaminant concentrations
recover after such an action. Because of uncertainty in the post-remedial recovery rate, it is
difficult to make definitive predictions of the time to reach specific recovery goals. The remedy
was designed and constructed with the expectation that both interim targets and ultimate risk-based
goals would be reached over a period of time. The interim targets, in particular, provide useful
milestones to help assess the actual rate of recovery relative to expectations. As the long-term
monitoring data are amassed, the understanding of remedial effectiveness will be refined. With
this refinement, EPA will be able to determine whether additional data collection and investigation
are needed.
EPA recognizes that because individual fish species will respond to contaminant exposures in
different ways depending on their foraging strategies and life histories, the ROD utilized a
calculated "average" or "composite" fish to represent the variety of fish likely to be consumed by
anglers. It is important to note that any individual fish (and any individual fish species more
broadly) will achieve "target levels" at different times given a number of factors including: 1)
variability in actual exposures; 2) highly localized exposures; 3) the importance of sediment vs.
water exposure pathways, which can vary over time due to prey availability; 4) uncertainty and
variability in lipid content of fish and prey items; 5) uncertainty and variability in consumption of
specific prey items and PCB concentrations in those prey; and 6) measurement uncertainty
(including allowing for differences in sampling programs and analytical methods).
It is important to note that the Hudson River is a large, diverse and dynamic natural system and it
is unrealistic to expect that an average, species-weighted concentration (measured in mg/kg wet
weight) will be achieved consistently within a precise time-frame. The variability inherent in large,
dynamic systems such as the Hudson River may well lead to a situation in which an average
concentration (based on data) for any specific species and sampling location might achieve a target
7 From Figures A8-5.1 and A8-5.2in Appendix 8 of the FYR report, fish PCB concentrations in Cumberland Bay
declined from 4-5 mg/kg in 2005 to about 1 mg/kg in 2009, corresponding to an average rate of decline of about
25 percent per year.
51
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
threshold in a given year, but again rise just above it the following year due to any one of the
processes listed above, particularly when averaging across species and sampling locations.
Based on 2016 fish tissue monitoring data (presented in Figure A3-19 in Appendix 3 of the FYR
report), fish tissue concentrations for individual species range from 0.4 to 1.7 mg/kg depending on
the species and location. Even though the system has not equilibrated, yellow perch, for example,
has already achieved the 0.4 mg/kg interim target at several locations. The Upper Hudson species-
river section-weighted average based on 2016 data is about 1 mg/kg, this value is comparable to
the model results from the first year post dredging (2010) , as shown in Figure A3-19 of Appendix
3 to the FYR report.
The HUDTOX and FISHRAND models were calibrated, verified and applied to the Upper Hudson
River, and designed to support decision-making by allowing direct comparisons of predicted
water, sediment, and fish tissue concentrations across proposed remedial alternatives. The strength
of the models lies in their ability to compare predicted concentration trajectories in sediment,
water, and fish overtime based on a consistent set of assumptions. Absolute model predictions are
likely to differ from actual observations due to the same six factors noted above. In addition,
differences in environmental conditions (e.g., flow rates, upstream boundary conditions, etc.) also
contribute to potential differences between predicted versus modeled tissue concentrations,
particularly given that the models were primarily designed to predict relative tissue concentrations
across remedial alternatives rather than absolute concentrations over time. Nonetheless, model-
data comparisons, as presented in Appendix 3 of the FYR report for the pre-dredging MNA period,
show that the model performed well, and continues to perform well based on 2016 data (Figure
A3-19). The pre-dredge data comparisons in this FYR were not intended to be a predictor of future
recovery trends as some commenters indicated. Those analyses simply indicate that the modeling
tools used for EPA decision-making performed well, thereby supporting decisions made in the
ROD.
In conclusion, the Upper Hudson River system underwent a "reset" with dredging, which
established a new baseline from which post-dredging MNA trends must be evaluated.
Accordingly, evaluating data-based trends into the future starting with this new baseline will
require additional data over multiple annual cycles to provide statistically meaningful estimates of
progress toward meeting the interim targets and remedial goals. EPA estimates that as many as
eight or more years of fish data will be needed to confidently determine recovery trends.
3.3.13 Comment 29: EPA's analysis of fish data is flawed
Comment
A commenter indicated that EPA compares observations of fish tissue concentrations using median
values against the fish tissue concentration goals listed in the ROD, which are based on average
concentrations. Additionally, EPA compares average individual fish PCB concentrations against
the remedial goals although achievement of the goals will be evaluated based on species-composite
averages across river sections and the entire Upper Hudson. Such comparisons to individual
species at individual locations are not particularly meaningful when comparing to the metrics EPA
chose in the ROD. EPA should compare fish data using the same consistent basis of measurement
52
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
{i.e., average to average). However, comparisons at specific locations are very important in
understanding trends in site media over time, and the commenter encouraged EPA to gather fish,
sediment, and water data on a pool-by-pool basis rather than river section basis.
Response
EPA agrees that clarity is needed when comparing various metrics concerning PCB contamination.
In Section 5.1.1.3 and Appendix 3 of the FYR report, EPA compares median PCB concentrations
observed in 2016 fish with the interim target levels for fish tissue. EPA agrees that it is the average
value that ultimately determines achievement of an interim target or a remedial goal. Nonetheless,
the observation that a median value has fallen below a target or goal indicates that an important
milestone has been reached. Specifically, this indicates that more than half the fish caught showed
tissue concentrations less than the target or goal.
EPA does not agree that comparisons of individual species to the target levels or remedial goals
are meaningless. Since the species composite consists of a weighted average of individual species,
it is important to identify which species may be contributing to an exceedance of a goal or target,
if that is observed. EPA recognizes that a species composite target or goal can only be met when
individual species begin to meet that level.
EPA agrees that data collected on a pool-by-pool basis is useful in interpreting the recovery of the
system. For this purpose, EPA has analyzed data collected at the pool-by pool (i.e., reach) scale
and will continue to do so. EPA recognizes that there is limited fish data from Reaches 1 through
4 {i.e., the lower half of RS 3) and that additional fish collection from these reaches is necessary.
The 2019 fish sampling program will include fish collected from each of these reaches. Note that
while EPA will continue to evaluate the data on a reach basis, the success of the remedy is assessed
primarily on a river section basis.
3.3.14 Comment 30: EPA's analysis of water PCB trends must consider changes in both
loading conditions and comparisons of monitoring data to model predictions when
developing and interpreting trends
Comment
Commenters stated that the GE facility source control assumption reflected in the upstream
boundary condition for EPA's model (HUDTOX) MNA forecasts {i.e., PCBs containing three or
more chlorines [Tri+ PCBs] decreasing from 0.16 kg/day to 0.0256 kg/day starting in 2005) is a
significant factor resulting in high model-based decay rates presented in Table Al-7 of Appendix
1 of the FYR report, as opposed to natural recovery processes unrelated to source control.
Additionally, commenters stated that EPA did not account for the effects of the Allen Mill gate
structure and bedrock seeps of PCB oil prior to GE's completion of upstream source control
measures when interpreting data-based estimates of water column PCB declines presented in the
FYR report. Commenters also suggested that major changes in water column PCB sampling
locations and methods were not accounted for and that these changes make determination of
temporal changes in water column PCBs unreliable. Citing these reasons, the commenters
53
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
recommended that the post-source control period from 2005 to 2008 be used as a baseline when
calculating both data and model (HUDTOX)-based MNA decay rates for water column PCB
concentrations.
Another commenter suggested that the Farley model's under-prediction of Lower Hudson River
water column PCB concentrations during the pre-dredging period was a result of the Farley model
only being calibrated to sediment and fish data. The commenter suggested that an increase in
observed Tri+ PCB concentrations between Albany and Poughkeepsie that is not reflected in
TPCB water column data indicates the presence of a local PCB source, but that 2016 PCB
concentrations at Poughkeepsie were lower than Baseline Monitoring Program (BMP)
observations and, therefore, still show a response to dredging.
Response
EPA does not concur with commenters' suggestions that 2005 to 2008 is an appropriate time frame
over which to characterize water column PCB decay rates under MNA from data or from
HUDTOX simulations, because variability in annual flows dominates the temporal decline in
water column PCB decline over such a short period. This flow dominance produces trend
estimates that are highly uncertain and have no applicability to longer periods that are relevant to
assessing MNA performance as an aspect of the remedy.
The assumed trend in upstream boundary loads in the ROD MNA forecast simulations reflected
control of the Hudson Falls and Fort Edward GE plant sites as external sources to the river. To
better understand the influence of upstream source controls on MNA simulations, EPA has
conducted an alternate diagnostic HUDTOX MNA forecast using the lower of the two-constant
upstream boundary PCB loads assumed in the ROD (0.0256 kg/day), and starting this load in 2000
instead of2005. This alternate diagnostic forecast controls for variability in the upstream boundary
loads and also negates the potential influence of the data-based 1997-1999 upstream boundary
conditions as well as the 1991 Allen Mill gate failure.
Table Al-7 of Appendix 1 of the FYR report presented simulated decay rates, 1995 to 2008, in the
ROD MNA forecast and the updated MNA forecast incorporating actual flows. Table Al-7 is
reproduced below, and Table Al-7a presents the alternative simulations that assume a constant
upstream boundary load starting in 2000, and computing decay rates for the period 2000 to 2008.
54
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Table Al-7 (as presented in Appendix 1 of the FYR report): Average Annual Water
Column Tri+ PCB ROD and Updated MNA Forecasts for 1998-2008.
Augmented by Pre-MNA Calibration Results for 1995-1998
ROD MNA (step-down upstream PCB load)
MNA Update (step-down upstream PCB load)
Year
Tl Dam
Schuyler-
ville
Stillwater
Waterford
Tl Dam
Schuyler-
ville
Stillwater
Waterford
Ul g
a: o
1995
55.8
63.1
50.4
42.7
55.8
63.1
50.4
42.7
1996
30.2
38.3
37.0
34.3
30.2
38.3
37.0
34.3
1997
29.0
35.9
36.6
34.7
29.0
35.9
36.6
34.7
1998
38.3
44.2
38.7
35.8
38.2
43.6
41.4
39.4
1999
32.7
38.4
34.2
29.8
34.0
39.2
40.0
35.0
H
2000
24.7
29.0
26.5
25.0
24.8
29.6
28.0
25.7
<
2001
25.1
30.4
26.6
24.6
32.8
35.8
33.1
29.9
O
111
2002
27.6
30.3
23.7
21.1
27.8
30.5
28.0
24.7
£
o
2003
26.6
28.8
23.0
19.9
23.0
26.0
23.6
21.4
LL
2004
29.3
31.0
23.7
19.7
20.9
23.1
21.0
18.7
<
z
2005
13.5
17.0
15.5
14.6
12.0
15.5
16.0
15.6
m
2006
11.6
14.9
13.5
12.4
00
bo
12.2
13.0
12.5
2007
12.1
15.3
13.2
12.1
13.0
15.2
14.7
13.1
2008
13.8
15.9
12.3
10.4
10.0
12.0
15.7
12.9
1995-
2008
Decay
Rate
9.7%
9.6%
10.4%
10.6%
11.7%
11.4%
9.9%
10.0%
Table Al-7a Average Annual Water Column Tri+ PCB ROD and Updated MNA Forecasts
for 1998-2008. Augmented by Pre-MNA Calibration Results for 1995-1998,
Assuming Constant Upstream Boundary PCB Loadings, 2000-2008
ROD MNA (constant upstream PCB load)
MNA Update (constant upstream PCB load)
Year
Tl Dam
Schuyler-
ville
Stillwater
Waterford
Tl Dam
Schuyler-
ville
Stillwater
Waterford
PRE-
ROD
1995
55.8
63.1
50.4
42.7
55.8
63.1
50.4
42.7
1996
1997
30.2
29.0
38.3
35.9
37.0
36.6
34.3
34.7
30.2
29.0
38.3
35.9
37.0
36.6
34.3
34.7
1998
38.3
44.2
38.7
35.8
38.2
43.6
41.4
39.4
1999
32.7
38.4
34.2
29.8
34.0
39.2
40.0
35.0
h
2000
15.4
20.8
21.6
21.8
15.6
21.5
22.8
22.3
<
2001
16.0
22.4
21.6
21.3
19.3
24.3
25.4
24.7
o
111
2002
16.3
20.5
18.0
17.2
16.0
20.1
20.9
20.0
a:
o
LL
2003
14.9
18.6
16.7
15.6
13.2
17.2
17.4
17.0
2004
16.4
19.7
16.7
15.0
11.4
14.5
14.8
14.1
<
z
2005
13.1
16.3
14.2
13.2
11.7
15.0
14.9
14.3
E
2006
11.3
14.4
12.6
11.4
8.6
11.7
12.3
11.6
2007
11.8
14.8
12.4
11.1
12.7
14.7
13.9
12.3
2008
13.5
15.4
11.6
9.6
9.8
11.6
15.2
12.3
2000 -
2008
Decay
Rate
3.8%
5.5%
8.4%
10.3%
7.5%
8.7%
7.7%
9.5%
The right panel of Table Al-7a presents the alternative MNA Update forecast, holding upstream
loadings constant starting in 2000, and computing decay rates starting in that year. The decay rates
from this alternate HUDTOX MNA forecast ranged from 7.5 percent to 9.5 percent per year and
are lower at each station than the corresponding rates presented in Table Al-7, but are generally
consistent with the MNA-based rates of water column PCB decline reported. Note that these rates
of decline apply to the 2000 to 2008 period and are expected to become smaller over time, as PCB
attenuation in the Upper Hudson River becomes increasingly controlled by upstream and other
sources rather than by PCB mass transfer from the sediment bed.
55
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
The HUDTOX-simulated 1995 to 2008 water column PCB trends presented in Table Al-7 are
representative of actual observed conditions over that period. EPA does not assert that post-remedy
attenuation rates will match those trends, only that continued declines are anticipated. Post-remedy
MNA trends can, and will, be addressed through the OM&M program that is an integral part of
the ongoing MNA component of the remedy.
Table Al-7a also shows an alternative version of the ROD MNA forecast, which used synthetic
future flows assumed at the time of the ROD, and holding upstream loads constant starting in
2000. As with the alternative MNA update, the assumption of constant loads and the shift to the
2000 to 2008 timeframe reduces the estimated recovery rates, in this case to a range of 3.8 percent
to 10.3 percent for the four stations.
With respect to FISHRAND predicted trends, some portion of the 1995 to 2008 predicted recovery
simulated by that model would similarly be due to source control assumptions because the
HUDTOX simulations provided exposures for FISHRAND in the Upper Hudson River and PCB
loads for the Lower Hudson River.
Using data only from Rogers Island, a commenter concluded that the impact of the Allen Mill gate
failure continued well past 1995 (for Waterford and Stillwater) and 1997 (for Thompson Island
Dam [TID] and Schuylerville), rendering them inappropriate starting years for MNA trend
analysis. In order to assess the impact of the Allen Mill Gate failure on PCB loads to the Upper
Hudson River, EPA reviewed Tri+ PCB concentration data and estimated loads at Fort
Edward/Rogers Island and Thompson Island Pool (TIP) between the period of 1991 (when the
gate failure occurred) and 2008 (the last year of the BMP). Figure 30-1 (below) plots the Tri+
PCB concentration (ng/L) for samples collected at Fort Edward/Rogers Island and TIP between
1991 and 2008. Figure 30-1 indicates that for the years 1991-1992, during and immediately after
the Allen Mill gate failure, concentrations were the highest at both stations and the stations
exhibited similar concentrations. By 1993, peak summertime concentrations at both stations had
decreased by a factor of almost 10 and concentrations at Fort Edward/Rogers Island had typically
fallen below concentrations at TIP. By the beginning of 1996, concentrations at Fort
Edward/Rogers Island were almost a factor of 10 lower than in TIP, and concentrations at Fort
Edward/Rogers Island remained consistently lower than TIP throughout the year. Note also that
beginning in 1996 the number of samples collected at Fort Edward/Rogers Island that were below
the analytical detection limit began to increase, while the vast majority of samples collected at TIP
were still above the detection limit. Between 1996 and 2003 (the last year of pre-BMP data),
concentrations at Fort Edward/Rogers Island were consistently lower than in TIP throughout the
year and were dominated by non-detect results. Starting in 2004, samples at both stations were
collected as part of the BMP. The analytical method used to quantify PCB data during the BMP
included a lower detection limit, such that all samples contained detectable concentrations of
PCBs. The comparison of Tri+ PCB concentrations and the number of non-detect samples at the
two stations indicates that beginning in 1996, concentrations at TIP were driven more by localized
sources of PCBs (e.g., release of PCBs from sediments and PCB-contaminated in-river debris)
than by upstream sources such as the Allen Mill gate failure.
Figure 30-2 (below) plots the monthly average Tri+ PCB load (kg/month) at Fort Edward/Rogers
Island and TID between 1991 and 2008. The monthly Tri+ PCB load at each site was calculated
56
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
by averaging the measured Tri+ PCB concentration by month including only days where
measurements occurred at both stations and multiplying the average daily concentration by the
number of days in the month. As with the concentration plots, the highest loads occurred between
1991 and 1992. However, the load calculations indicate that by 1993 to 1994 the loads at both
stations had substantially declined, and that by 1995 to 1996 the loads at TIP were substantially
higher than Fort Edward/Rogers Island. As the loads at TIP were substantially greater than at Fort
Edward/Rogers Island by approximately 1996, this provides further evidence that upstream
sources of PCB load (including PCB releases from the Allen Mill gate failure) no longer
contributed a substantial loading of PCBs into the upstream boundary of the Site {i.e., Ft.
Edward/Rogers Island), and that loads measured at TIP could largely be attributed to localized
sources present between Fort Edward/Rogers Island and TIP. Thus, declines in water column PCB
loads beginning in 1996 can be largely attributed to natural attenuation/natural recovery of the
Site, as opposed to reduction in upstream inputs of PCBs. In conclusion, it is EPA's position that
starting the MNA period in 1996 is justified and appropriate.
With respect to water column monitoring, EPA recognized that changes in the location and method
of sample collection between various datasets could impact the ability to analyze long-term
changes in water column PCB concentrations. Therefore, pilot studies were initiated at TID and
Schuylerville that involved concurrent sample collection at both the pre-BMP {i.e., Post-
Construction Remnant Deposit Monitoring Plan (PCRDMP) monitoring program) and BMP
stations to assess whether the change in station location and method of collection produced a bias
in PCB water column concentration. At the Schuylerville station, concurrent samples were
collected between June 2004 and May of 2006. At TID, concurrent samples were collected
between June 2004 and August 2008. These studies were summarized in documents reviewed by
EPA prior to allowing the station location and method collection to be altered (Corrective Action
Memo (CAM) 6 (GE 2006) and CAM 14 (GE 2008)). Results of these pilot studies indicated that
alteration of the sample collection location and/or method of collection at the long-term monitoring
stations did not produce a significant bias in the water column PCB concentration. As such, it is
EPA's position that it is appropriate to use multiple datasets to calculate long-term trends in water
column PCB concentrations at these stations. Unlike the Thompson Island Dam and Schuylerville
monitoring stations, the Stillwater and Waterford monitoring stations were not monitored during
the PCRDMP monitoring program. The trend analysis included in Appendix 1 of the FYR report
did not include data at the Stillwater monitoring station between 1998 and 2004 (the beginning of
the BMP) and between 2001 and 2004 for the Waterford monitoring station. Prior to 2001 (for
Waterford) and 1998 (for Stillwater), USGS-collected water column PCB data were used. While
a comprehensive comparison of datasets could not be carried out for the Stillwater and Waterford
monitoring stations, sample collection methodology used by the USGS was similar to methods
used during the PCRDMP monitoring program at TID and Schuylerville. Further, sample
collection at all four sites during the BMP period was based on a common approach {i.e., all used
a multiple aliquot depth integrated sampler (MADIS)). Thus, results from TID and Schuylerville
provide evidence that sampling differences between the USGS and BMP datasets do not produce
a significant bias in reported water column PCB concentrations.
EPA recognizes that the analytical methods used by the USGS differed from the methods used
during the BMP period, including the USGS method having a higher detection limit (11 ng/L).
However, of the 61 USGS samples included in the trend analysis at Stillwater station from 1995
57
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
through 1997, only 8 samples were identified as non-detect for Tri+ PCBs, and the PCB
concentrations for these samples were set at 5 ng/L, approximately one-half the detection limit of
the method. Similarly, only 8 out of 68 USGS samples collected at Waterford station from 1995
through 1997 were identified as non-detect for Tri+ PCBs, and these samples were also assigned
a concentration of 5 ng/L. Beginning in 1999, USGS updated their method for measuring PCBs.
However, issues related to the new USGS method produced PCB concentrations that were biased
high between 1999 and 2000. By 2001, the issues related to the updated methodology were
resolved and USGS data from 2001 were included in the trend analysis at the Waterford station.
The updated USGS methodology had a lower detection limit than the previous method and was
able to detect concentrations below 11 ng/L. In conclusion, while EPA recognizes that sampling
and analytical methods did change during the time period used in the trend analysis at the four
long-term monitoring stations, these differences were deemed not to cause significant bias in the
water column PCB concentration so that it is appropriate to combine datasets.
EPA acknowledges that concentrations shown in FYR report Appendix 1 Figure Al-3b show
generally higher concentrations at Poughkeepsie than at Albany. Data reviewed for the FYR did
not reveal an explanation for the difference in concentrations between these two locations. EPA
does not agree that the decline in concentrations at Poughkeepsie between the BMP and 2016
necessarily shows a response to dredging. The decline over this time period could reflect an
ongoing decay in concentrations due to processes in the Lower Hudson River, a response to
dredging, or a combination. In particular, the absence of increases in water column concentrations
at Poughkeepsie as a response to dredging suggests that the lower concentrations in 2016 do not
reflect a response to the decrease in water column loads from the Upper Hudson at the end of the
dredging period. EPA will continue to carefully evaluate data collected from the Lower River
including from Poughkeepsie and Albany as part of future supplemental studies.
58
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
10000
1000
D)
o
c
o
O
CG
O
CL
+
100
10
0.1
\i .
•• •*
\ V • £ •
• Fort Ed./Roger's Island Detect
x Fort Ed./Roger's Island Non-Detect
• Thompson Island Pool Detect
x Thompson Island Pool Non-Detect
v WA&A
• -' '
Cva
FE/RI NDs
- 3K
X X
X >K X XK X
TIP NDs
- X
1 1
X
1 I
1 1 1 1
Ill
a>
LO
CM
O)
CO
cn
in
O)
CD
CD
Is*-
Gi
CO
cn
CT)
cn
o
o
LO
o
ID
CN
O
LO
CO
o
p
LO
O
CD
O
LO
o
LO
00
o
in
CD
o
Figure 30-1 Measured concentrations at Ft. Edward/Rogers Island and Thompson Island
Dam (TID) between 1991 and 2008. Top panel shows samples with and
without Tri+ PCB detections. Non-detect samples are assigned a
concentration of 5.5 ng/L, which is one-half the detection limit. Bottom panel
indicates when non-detect samples were collected. Note the Allen Mill gate
failure occurred in 1991 to 1992 and the Baseline Monitoring Program
(BMP) began in 2004.
59
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
450
Date
Figure 30-2
Monthly average Tri+ PCB load at Ft. Edward/Rogers Island and Thompson
Island Dain (TID) between 1991 and 2008. Note the Allen Mill gate failure
occurred in 1991 to 1992 and the Baseline Monitoring Program (BMP) began
in 2004.
3.3.15 Comment 35: Incorporate Hudson River Reference Material in future fish analyses
Comment
Commenters requested the incorporation of Hudson River Reference Material in future fish
analyses.
Response
EPA agrees that some form of performance evaluation (PE) material should be incorporated into
future fish analyses under OM&M. Currently, EPA is evaluating the use of multiple potential
materials, including National Institutes of Standards and Technology (NIST) Standard Reference
Material (SRM) 1946 (Lake Superior Lake Trout) and NIST SRM 1947 (Lake Michigan Lake
Trout). EPA understands that NYSDEC has used SRM 1947 along with Hudson River Reference
Material for QA in its fish analyses. EPA is currently discussing with GE the future use of PE and
reference materials as part of laboratory QA approaches for both fish tissue (NIST SRMs 1946
and 1947) and sediment (e.g., NIST SRM 1944) samples.
60
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
3.3.16 Comment 36: Increase the use of congener PCB analysis and decrease use of Aroclor
analysis
Comment
Underestimation of Total and Tri+ PCBs in sediment based on EPA Method 8082 (M8082) relative
to recent EPA Method 1668 (Ml668) split-sample analysis is not addressed in the FYR report.
The 2016 split-sample analysis for sediment suggests that both the Aroclor and modified Green
Bay Method (mGBM) PCB analyses may significantly underestimate the Total PCB concentration
as compared to full congener analysis (M1668).
The analytical method for measuring the PCBs in the samples is outdated and must be updated.
The 2016 Sediment Work Plan indicates that PCBs will be measured via M8082 (modified via the
Green Bay procedure) and EPA indicates that 4% of the samples will also use Ml 668 to measure
PCBs. Given the greater accuracy of M1668, the justification for relying on the older and less
accurate M8082 is unclear. As EPA moves away from using M8082, and adopts M1668, there will
be a problem unless a much larger percentage of samples use both methods to establish a rigorous
conversion basis. For these reasons, EPA should increase the number of samples analyzed by both
methods in every reach of the river. This will ensure enough data are available for substantive and
statistically significant comparisons between the methods to facilitate accurate conversion before
EPA switches to only Ml668 for OM&M sediment samples in the future. A commenter questioned
whether the exposure assumptions, toxicity, data, cleanup levels, and remedial actions objectives
used at the time of the remedy selection are still valid, stating that the variability of testing methods
has tainted the results to date.
Response
EPA agrees with the comment concerning the importance of comparability between the two
analytical methods. It is EPA's intention to establish a representative, precise and accurate estimate
of PCB concentrations in the sediments, water and fish as part of the baseline monitoring program.
Each of these attributes is addressed by different aspects of the OM&M program. However, the
issue raised in this comment concerns accuracy most directly. That is, how can EPA be sure that
the concentrations for sediments obtained in 2016 and 2017 are accurate while using M8082, an
Aroclor-based method which approximates the Total PCB concentration as the sum of the reported
Aroclors? Accuracy is particularly important for long-term monitoring. Analytical variation
through time (essentially a "drift" in the report values) must be minimized so that changes
observed in reported average concentrations over time can be attributed to real changes in the river
over time, and not to variations in the analytical methods. Maintaining accuracy helps to reduce
the uncertainty in the long-term monitoring data.
To this end, EPA began an initial program in 2016 to compare M8082 results with those obtained
by M1668. The 2016 program consisted of the analysis of a limited number of samples by both
methods. However, due to the sequencing of sample splits during collection, as well as differences
in the processing of the samples at the respective laboratories, resolution of the differences
between M8082 and Ml668 results is challenging since the data are confounded by these
additional technical considerations. While it is expected that Ml 668 is more precise given its direct
61
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
quantitation of individual PCB congeners, this does not mean that the M8082 results are
inaccurate. Historically, EPA has recognized differences between M8082 and other analytical
methods and addressed these differences by means of an adjustment factor to reconcile the
differences (e.g., EPA 1997; Butcher etal 1998; EPA 2017). In some instances, the M8082 results
fall below the second method results, whereas in other instances, the M8082 results exceed the
second method results. Recently, EPA has also evaluated sediment data matched pairs of M8082
and M1668 analyses collected by NYSDEC in 2017 to better determine the relationship between
results associated with two methods.
As a result of this concern, and in recognition of the need to maintain both accuracy and precision
throughout the OM&M program, EPA has undertaken a more rigorous approach in 2017 to
compare the two methods. Specifically, EPA has required GE to standardize its collection and
processing of surface sediment samples in a more rigorous fashion, based on the lessons learned
from 2016. The sample processing steps used by GE to produce samples for both M8082 and
M1668 analysis have been revised to use standard techniques for removal of larger particles (i.e.,
sieving), as well as to incorporate standardized procedures for sample homogenization. In this
manner, differences in the absolute values obtained by the two methods can be reconciled with a
known and acceptable level of uncertainty. That is, this approach will quantify the precision
between the two methods.
In addition, EPA is also requiring the use of reference materials by both the GE and EPA labs.
These standards will include both NIST and other reference materials that are designed for long-
term stability, so that future sampling programs will also be referenced to the same known
standards. In this manner, the EPA's approach will establish accuracy for the 2016-2017 program
while also tying future sediment monitoring to the same reference values. Finally, based on the
findings of the methods study using 2016 and 2017 samples as well as the results of the NYSDEC
work described below, EPA will identify a frequency of analysis by Ml668 for future sampling
programs to confirm and maintain comparability over time.
As discussed briefly above, EPA has also evaluated the recently available 2017 NYSDEC
sediment dataset of matched pairs of M8082 and M1668 analyses. The 2017 NYSDEC dataset
contains 117 matched pair samples, obtained as part of NYSDEC's 2017 Upper Hudson surface
sediment investigation. The analysis of these matched pair samples found that both Total PCB and
Tri+ PCB concentrations derived from M1668 measurements were higher than those obtained by
M8082. Based on the analyses, Total PCB by M1668 was approximately 55 percent higher than
the sum of Aroclors based on M8082. Similarly, Tri+ PCBs by M1668 was approximately 44
percent higher than those predicted from Aroclor data (M8082) using GE's equation [Tri+ =
0.13*A1221 +0.89*(A1242+A1254)]. These differences between M8082 and M1668 are not as
large as those suggested by the GE 2016 samples, which may be the result of consistent split
sample preparation between the two methods. EPA will continue to evaluate the difference
between two methods once more data are available.
The exposure assumptions, toxicity, data, cleanup levels, and remedial actions objectives used at
the time of the remedy selection were based on Aroclor data from M8082. If M1668 had been used
in the time of the remedy selection, it is likely that the cleanup levels would have been adjusted as
appropriate. Thus, the continuous use of M8082 for the analysis of fish, water and sediment is
62
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
required to maintain long tern internal consistency and compatibility with the remedial objectives.
For environmental media (fish, water and sediment), EPA developed and applied its own
congener-based dual column Gas Chromatography/ Electron Capture Detection (GC/ECD)
method throughout the RI/FS process to provide accurate estimates of PCB levels, which was used
in conjunction with M8082, providing some of the best data available anywhere at the time. PCB
analytical methods have continued to evolve, and EPA is applying them as appropriate. Notably,
the EPA and the NYSDEC relied and continue to rely on M8082 for fish characterization, although
confirmation via more sophisticated methods is also being developed.
It is important to note that EPA has not simply relied on the various methods themselves to provide
accurate and precise results. Throughout the remedial design sampling (SSAP), GE was required
to run performance evaluation (PE) samples (essentially laboratory-certified samples of known
concentration) to demonstrate accuracy in their analyses. Similarly, EPA required PE samples be
included in the post-dredging residual sampling program conducted by GE. Thus, while there
remain analytical accuracy and precision challenges to be considered for the 2016 data and for
subsequent OM&M monitoring, EPA has required GE to monitor the accuracy of its results
throughout the remedial design and remediation periods. In this regard, EPA is confident it has
based its decisions on reliable data throughout the study and remediation of the Upper Hudson
River.
3.3.17 Comment 40: The larger-than-expected mass of PCBs and higher surface sediment
PCB concentrations remaining in the sediment following remediation will extend the
recovery of the river
Comment
Commenters stated that actual PCB sediment concentrations should be the primary measure of
remedy success as defined by the ROD rather than decay rates or percent reduction. The
commenters assert that the success of the remedy does not depend on the percentage or amount of
PCBs removed, but the magnitude and spatial extent of PCBs left behind, which greatly exceeded
expectations in the 2002 ROD, and that the FYR incorrectly emphasizes the percent reduction in
PCB mass in the river. Using actual values of the residual PCB concentrations rather than
percentages, commenters state that the remedy as implemented does not conform to the 2002 ROD
expectations or meet remediation goals judged necessary to achieve protection of human health
and the environment in RS 2 and RS 3.
Commenters also stated that EPA compared PCB residual concentrations with the "less stringent
interim expectations" described in the 2012 FYR without any justification of why this is correct.
Furthermore, commenters indicate that there are insufficient post-remedial data available to
evaluate if the remedy is functioning as intended by the decision documents, and that the FYR
should acknowledge that the highly contaminated areas adjacent to the dredged areas identified
during remedial design as part of the SSAP have not been re-sampled sufficiently to determine
post-dredging PCB concentrations, percent reduction, or decay rates.
63
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Response
EPA acknowledges that the pre-design SSAP sediment samples collected from 2002 to 2005
contained PCBs at concentrations that were higher than ROD-based modeled concentrations. The
practical significance of this apparent difference was carefully and extensively considered by EPA.
It was determined that the change in fish tissue concentrations could be predicted without re-
calibrating the model to account for the higher sediment concentrations identified in the SSAP.
This is because the change in fish tissue concentrations is known to be proportional to the change
in sediment concentrations irrespective of the absolute sediment concentrations prior to
implementation of the remedy.
The remedy was developed by considering the proportional change in fish tissue PCB
concentrations that was required to meet risk-based thresholds over a period of time. Even if the
actual surface sediment PCB concentrations are different from those that were expected at the time
of the ROD, reducing the absolute PCB surface sediment concentrations by the same percentage
as anticipated by the ROD is expected to achieve the same percentage reduction of fish PCB
concentrations projected in the ROD. This analysis is explained below.
The physical premise underlying sediment to fish accumulation is that fish tissue PCBs are
approximately proportional to sediment concentrations to which water and prey items are exposed.
This is expressed mathematically as:
PCBfiSh = k x PCBsed
where, k is defined as the fish to sediment accumulation factor, which is a relative constant value
for a specific species from a specific portion of the site that have similar ecological and chemical
conditions (Burkhard, 2009).
With this assumption of proportionality, the ratio of post-dredging to pre-dredging fish tissue PCB
concentration can be related to sediment concentrations as follows:
PCBfish—post k X PCBsgd—pgst PCBsgd—pQsf
PCR ~ k v PCR ~ PCR
i pre a A 1 ^ ^ sed—pre ' X-,1Jseel—pre
This formulation indicates that the desired proportional change in fish tissue concentrations can be
obtained via an equal proportional change in surface sediment concentrations. Therefore, it is only
necessary to achieve the proportional change in sediment concentrations, as opposed to achieving
some absolute sediment concentration.
Remedy Outcome Compared to Anticipated Percent Reduction in 2012 FYR
As described above, the use of proportional change in PCB concentration in sediment as an
indicator of proportional change in fish tissue concentrations is consistent with the conceptual site
model and physics underlying the development of the ROD targets. Appendix 4 of the FYR report
acknowledges that estimates of proportional change in sediment concentration could be uncertain
due to differences in sampling designs and sediment collection equipment in 2002-2005 and 2016.
EPA specifically referred to these changes as "apparent" change, identifying that they embodied
64
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
actual change due to the remedy, actual change due to natural recovery, and potential artifacts of
differing sampling methods. In an effort to understand and fully explore these issues, EPA also
developed a hind-cast method for estimating changes in concentration that helped to reduce these
effects. See Section 5 of Appendix 4 of the FYR report for details.
As presented in Table A4-5 of Appendix 4 of the FYR report, based on comparison of the 2002 to
2005 SSAP dataset and the 2016 OM&M sediment sampling dataset, the percentage declines in
average Tri+ PCB (PCBs containing three or more chlorines) concentrations in surface sediments
(0-2 inch interval) as a result of dredging and MNA were 96, 88 and 80 percent in RS 1, RS 2 and
RS 3, respectively. These reductions exceed estimates presented in the 2012 FYR (87, 36, and 5
percent reductions as a result of dredging in RS 1, RS 2 and RS 3, respectively) and the 2002 ROD
(79, 64 and 4 percent reductions as a result of dredging alone, in RS 1, RS 2 and RS 3,
respectively). Thus, the actual percentage reductions achieved by dredging and MNA are
substantially greater than those anticipated by the ROD or the 2012 FYR.
Residual Concentrations at the Edges of Certification Units
Surface sediment data (0-2 inch) from the 2016 OM&M program were used to evaluate whether
there are any "highly contaminated areas adjacent to the dredged areas." The OM&M sediment
sampling program is based on a probability-based sample selection procedure that supports
unbiased estimation of mean PCB concentrations per river section. The sample selection process
is spatially balanced and includes samples from edges of certification units in proportion to the
size of these areas. This sampling procedure is referred to as a self-weighting design because any
underlying stratification of the population is represented proportionally to stratum size. For
example, if CU edges represent 10 percent of a given river section, then approximately 10 percent
of samples will be from these areas. This approach provides the data necessary for unbiased
estimates of river-section averages which are expected to be proportional to fish tissue
concentrations averaged over the same river section. These data are appropriate and unbiased for
judging the average effect of the remedy at the river section scale.
Figures 40-la, lb and lc below show how the surface sediment Tri+ PCB concentrations in non-
dredged areas vary with distance from the closest dredged area boundary. The figures also include
a weighted average line, which represents a running average through the data as a function of
distance from the dredging boundary. The weighted curves for the areas outside dredging
boundaries show no significant positive increase in surface sediment Tri+ PCB concentrations
from dredging boundary out to the maximum distance values on the plot (150 ft in RS 1 and 300
ft in RS 2 and RS 3). These plots indicate that sediments close to the dredging boundaries are not
particularly more contaminated than those located far away. Furthermore, the concentrations of
Tri+ PCB in surface sediments are generally low. The average surface sediment concentration of
Tri+ PCB in non-dredged areas from the 2016 OM&M program was 1.7 mg/kg, 1.5 mg/kg and
0.8 mg/kg in RS 1, RS 2 and RS 3, respectively. When comparing the individual sample results to
their respective dredging criteria (i.e., 10 mg/kg for RS 1, 30 mg/kg for RS 2 and RS 3), only one
sample (out of 215 samples) exceeded the criteria. If the most stringent RS 1 criterion was applied
to the entire Upper River, there were only two exceedances. Future OM&M sampling will provide
more data to determine the post-dredging percent reduction rate8.
8 The 2017 NYSDEC surface sediment sampling program further confirmed that the surface sediment concentrations
outside the dredging boundary are generally low. The average surface sediment concentration of Tri+ PCB in non-
65
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
EPA acknowledges that the effect of the remedy may vary by reach. EPA has and will continue to
evaluate the remedy on smaller spatial scales, including by river reach (i.e., stretches of the river
that are separated by dams or locks), for future assessment of the recovery of the river. As an
example, EPA's evaluation of the combined 2016 EPA/GE and 2017 NYSDEC sediment data
presented in the EPA's April 2019 "Technical Memorandum Evaluation of 2016 EPA/GE and
2017 NYSDEC Surface Sediment Data" (www.epa. gov/hudson) indicates that there are three very
localized areas where PCB levels are statistically elevated compared with surrounding areas. EPA
and NYSDEC will undertake analyses to jointly define "Areas of Interest" to be tracked in greater
detail (e.g., with increased local sampling density) in the future.
dredged areas from the 2017 sediment survey was 2 mg/kg, 3 mg/kg and 0.76 mg/kg in RS 1, RS 2 and RS 3,
respectively. When using the combined 2016 OM&M and 2017 NYSDEC surface sediment datasets, there are only
4 sample locations out of 1,304 (or 0.3 percent) where Tri+ PCB concentrations exceed their respective criteria in
recoverable sediments across both dredged and non-dredged areas (3 in RS 1 and 1 in RS 2). Further, there are only
8 locations (inRS 2 and RS 3 combined, or 0.74 percent of 1,078 locations) where Tri+ PCB concentrations exceed
the lower RS 1 removal criterion of 10 mg/kg. Overall, if the RS 1 criterion were applied to the entire Upper River
(which is not what the ROD required, but what some have suggested), just 11 sample locations (3 in RS 1,2 in RS
2 and 6 in RS 3, or 0.84 percent overall) exceed that most stringent threshold.
66
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
2016 Data(0-2 in)
Distance from Dredge Boundary
1000-
100 -
O) 10"T
"3)
E
CO
o
CL
+ 1-
0.1-
0.01 ¦
River Section
¦ i i i i i i i i i i i i i i i i i i i i i i i
I I I I | I I I I | I I I I | I I I I | I I I I
1
I I I I I I I I I I I I I I
I I I I | I I I I | I I I I
Legend
— Weighted Average
Dredging Threshold
-250 -200 -150 -100 -50 0 50 100 150
Dredge Nondredge
Distance From Dredge Boundary (ft)
Figure 40-la Variability in surface sediment Tri+ PCB concentrations with distance from
dredging boundary in RS 1. Samples were collected under the 2016 OM&M
program. Samples with distance greater than 150 ft are not shown.
67
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
2016 Data(0-2 in)
Distance from Dredge Boundary
1000-
100 -
O) 1 0 -
"3)
E
CO
o
CL
+ 1-
0.1-
0.01
_l I I I I I I I L-
River Section 2
_l I I I I I I I I I I I I L-
i—i—i—i—|—i—i—i—r
200 -100
Legend
— Weighted Average
Dredging Threshold
0
t—i—i—i—|—i—i—i—i—|—i—i—i—r
100 200 300
Dredge Nondredge
Distance From Dredge Boundary (ft)
Figure 40-lb Variability in surface sediment Tri+ PCB concentrations with distance from
dredging boundary in RS 2. Samples were collected under the 2016 OM&M
program. Samples with distance greater than 300 ft are not shown.
68
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
1000 ¦
100 -
10-
"3)
E
m
O
CL
+
1-
0.1-
0.01 ¦
2016 Data(0-2 in)
Distance from Dredge Boundary
River Section 3
J I I I I I I I I I I I I I I I I I I I L-
Legend
— Weighted Average
- - Dredging Threshold
t—i—|—i—i—i—i——i—i—i—i—|—i—i—i—i—|—i—i—i—r
-100 0 100 200 300
Dredge Nondredge
Distance From Dredge Boundary (ft)
Figure 40-lc Variability in surface sediment Tri+ PCB concentrations with distance from
dredging boundary in RS 3. Samples were collected under the 2016 OM&M
program. Samples with distance greater than 300 ft are not shown.
3.3.18 Comment 41: Reassess air risks
Comment
Commenters state that PCBs from the Hudson River will volatilize and be inhaled, and these air-
borne PCBs become a significant exposure to anyone living or spending significant time near the
river. They say that people who live along the Hudson River have a significantly increased risk of
hospitalization for heart disease and diabetes, which is because of their proximity to PCBs from
the Hudson River or sediment that volatilize and pollute the atmosphere near the river. This
exposure is not voluntary for anyone living near the river. Commenters noted that simply living
near a PCB contaminated site poses a risk of exposure and to disease. The very large amounts of
PCBs that GE now plans to leave behind greatly exacerbate this problem. EPA should verify that
the air route of exposure is not a significant route of exposure requiring remedial action,
particularly in the Lower Hudson.
69
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Response
The HHRA evaluated a number of exposure pathways including air exposures. The results of this
analysis indicated that the cancer risks from inhalation of PCBs in air were 1 x 10"6 based on
modeling of air concentrations. These risks are at the lower bound of EPA's generally acceptable
cancer risk range for exposures at Superfund sites, and 100 times lower than the upper bound of
the risk range of 10"4 (one in ten thousand). In addition, the inhalation risks are significantly lower
than all other pathways, including ingestion of fish.
Prior to and during the remediation, air data was collected in areas around the dredging (both
upwind and downwind) and in the areas near the river. This data was collected by multiple
agencies over a period of years from 2005 to 2015. This data was analyzed in Appendix 6 of the
FYR report. The data indicate that PCB air concentrations before the dredging (GE and NYSDEC
data) and during dredging (GE data) are both below estimates in the HHRA. Also, estimates of
post-dredging PCB air data indicate that concentrations are lower than those estimated in the
HHRA. As PCB concentrations in water are likely to decrease over time due to monitored natural
attenuation, it is expected that the PCB emissions from the river will also continue to decrease
over time based on the mass of PCBs removed from the River.
3.3.19 Comment 43: Resolve diverging views of data with other agencies
Comment
Commenters stated that EPA should reconcile divergent views on the timing required to meet the
goals of the cleanup and on the protectiveness determination by taking credible data and analyses
from studies conducted by other federal and state agencies into consideration, notably NOAA and
the NYSDEC.
Response
EPA has considered the data and input from the other agencies. EPA believes that reconciliation
of diverging government agency views about cleanup and protectiveness determinations has been
complicated by the NOAA emulation model (Field, et al., 2016) which attempts to "update" the
surface sediment PCB concentrations to forecast fish tissue concentrations. Simply changing a
variable, such as sediment concentrations, as NOAA did, without recalibrating the underlying
model to maintain consistency with the calibration data, produces results that are flawed. EPA's
detailed responses regarding NOAA's emulation model are contained in EPA's white paper9 and
in Appendix C of this document, and summarized in MasterComment 9 (see Section 3.2.1).
Additionally, EPA believes other agencies may have misinterpreted the significance of the SSAP
data with regard to the projected fish PCB recovery rates as discussed in the responses to Master
Comments 47 (see Section 3.5.5) and 58 (see Section 3.4.9). EPA will continue to take into
consideration other federal and state agencies' views regarding the ongoing OM&M phase of the
9 See: White Paper: Responses to NOAA Manuscript Entitled: "Re-Visiting Projections of PCBs in Lower Hudson
River Fish Using Model Emulation" (Field, Kern and Rosman, 2015) (EPA, 2016)
70
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
remedy. EPA has extensively evaluated surface sediment data collected by NYSDEC in the
summer of 2017 and has prepared a detailed technical memorandum, which is being published
concurrently with the Final Second FYR (see: Technical Memorandum Evaluation of 2016
EPA/GE and 2017 NYSDEC Surface Sediment Data, April 2019, [www.epa.gov/hudson]).
Notable findings of that evaluation are that: (1) the 2017 NYSDEC data and the 2016 EPA/GE
data collected outside dredged areas yielded similar estimates for surface sediment PCB
concentrations; (2) the remedy significantly reduced PCB concentrations in dredged areas and
there has not been substantive recontamination of those dredged areas; and (3) no hot spots (i.e.,
areas exceeding the ROD removal criteria) were identified.10 EPA and NYSDEC have agreed that
additional data are needed to determine if the remedy is effective and if any additional remedial
work would be necessary or beneficial. EPA will continue to consider information and data
provided by NYSDEC and other agencies and use those data as appropriate to inform future
evaluations of the progress of the remedy.
3.3.20 Comment 46: Use of the non-standard protocol (without rib-in vs. rib-out) impacts
how the data can be used
Comment
Comments from multiple reviewers focused on the findings of EPA's preliminary evaluation of
the differences in PCB concentrations between black bass fillets with and without rib cages
included in the sample. Reviewers pointed out that EPA found differences in PCB concentrations
in fish tissue samples on both wet-weight and on a per-lipid basis which could influence
interpretation of these data generated by GE's contracted analytical laboratory from 2007 through
2013. Some reviewers noted that actual wet weight concentrations in fish tissue samples could be
on the order of two times higher than measurements without ribs would suggest, which could
influence NYSDOH fish consumption advisories. Several reviewers expressed that combining
these samples with historical samples that included the rib cage and surrounding tissue could bias
estimates of natural recovery rates, bio-accumulation rates and wet weight concentrations needed
to inform fish consumption advisories.
Reviewers pointed out that efforts to correct fish tissue PCB concentrations could be unreliable
due to high variability in the ratios of with-rib to without-rib PCB concentrations for individual
pairs. Reviewers also identified that the EPA study was restricted to black bass species and
suggested that EPA should embark on a similar study of other species comprising important
components of the monitoring program. It was suggested that these additional studies should be
aimed at understanding root causes of sample variability. The primary concerns focused on
potential inability to: 1) accurately estimate temporal trends in PCB concentrations, 2) make fair
comparisons to modeled predictions, and 3) forecast future concentrations and the time to reach
remedial objectives.
Several reviewers expressed a desire for EPA and or GE to conduct additional comparative studies
looking at differences between standard fillet and without-rib processing protocols. One reviewer
pointed out that understanding the effects of the change in protocol was necessary for trustees to
10 Of the nearly 1,900 sediment locations occupied and sampled by NYSDEC and GE, there were only four sampling
results at scattered locations that exceeded the ROD surface sediment removal criteria.
71
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
quantify injuries as part of the natural resource damage assessment. Finally, several reviewers
stated that this study was necessary to understand the anticipated time for fish tissue PCBs to reach
remedial goals.
Response
Based on information provided by New York State, EPA's understanding is that between 2007
and 2013 GE's contract laboratory did not follow the NYSDEC standard fillet approach by not
including the rib cage material in the analyzed fish fillets. The 2004 BMP QAPP indicated that,
"All fish will be prepared for contaminant analyses following collection according to the SOP for
Annual Fish Sampling (Appendix 21; adapted from NYSDEC procedures)." The NYSDEC
standard fillet approach and Appendix 21 of the 2004 BMP QAPP require inclusion of the rib-
bones and belly flap with the fillet that is removed from the fish and subsequently analyzed for
PCBs and lipids. In 2013, GE indicated to EPA that "the ribcage was not included with the fillet
in either the BMP or the Remedial Action Monitoring Program (RAMP) for samples collected
since 2007." In response, EPA requested that GE perform a special study that would facilitate
evaluation of whether or not inclusion of the rib cage (ribs) had a significant impact on fish tissue
PCB concentrations and lipid levels. Black bass (small mouth bass and largemouth bass) were the
focus of the resulting 2014 study because they are large enough to produce fillets of sufficient size
for comparison, are processed with skin on, and are collected from RAMP stations on the Upper
Hudson River (UHR) and Lower Hudson River (LHR).
The 2014 study indicated that on a wet weight basis, the difference between fillets prepared with
ribs vs. without ribs was variable and could be greater than a factor of two. For lipid-normalized
data, the difference between the two fillet approaches averages less than 20 percent. As a result,
EPA determined that the difference in fillet methods does not affect lipid-normalized fillet trend
data. However, EPA recognizes that the results of the 2014 special study found differences in PCB
concentrations in fish tissue samples with and without ribs on both wet-weight and per-lipid bases,
which could influence interpretation of project data. However, the majority of the data influenced
by the change in fillet protocol were collected prior to, or during remediation, and currently
collected (post-dredging) data are not routinely compared to that period. There are also PCB data
for samples processed as whole-body for pumpkinseed and forage fish from the UHR and LHR
and for samples processed following the NYSDEC standard fillet method from LHR stations
during the BMP and RAMP that span the period in question. Whole body processed fish are not
affected by the change in fillet protocol. In addition, post-dredging filleted fish samples are being
processed in a manner that is consistent with NYSDEC filleting protocols. EPA has been
conducting robust oversight of fillet protocol for post-dredging fish. Post-remedial evaluations of
fillet data will focus on how PCB concentrations relate to the interim targets (i.e., 0.4, 0.2 mg/kg)
and project goals (e.g., 0.05 mg/kg PCBs), and the time required to reach those targets and goals.
The Natural Resource Trustees requested that the study entitled "Special Study of Black Bass Fillet
Tissue With and Without Ribs" be finalized as part of the FYR. The special study is not directly
part of the FYR but the findings have been considered in the process of reviewing and presenting
the data in the FYR report. The study which was conducted in 2014 has been finalized and
provided to the public and trustees as requested (Louis Berger & Kern, 2019; see Appendix D of
this document).
72
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
EPA agrees that the change in processing methodology is an important issue and that temporal
evaluations involving standard fillet and non-standard fillet data may be biased (low) due to
variations in sample preparation methods. Because this issue has been discovered and corrected
for current and future monitoring efforts, EPA disagrees with reviewers with respect to the value
of additional study of other species. EPA acknowledges suggestions that the disposition of sample
preparation should be accounted for when estimating past natural recovery rates.
As part of the five-year review, EPA conducted exhaustive evaluations, considering alternative
ways to handle varying fish tissue lipid levels, standard fillet vs rib-out methods, and variable
starting and ending dates in efforts to identify the most robust methods for estimating PCB
recovery rates in fish tissues. These efforts are reflected in the variety of methods reported in
Appendix 3 of the FYR report. EPA anticipates further evaluation of PCB and lipid levels in fish
tissue in efforts to develop the most effective approaches for the design of OM&M plans for the
Hudson River remedy. Temporal variation in lipid concentrations in fish tissue is not unique to the
Hudson River, but has also been observed in monitoring data from both the Fox River and
Kalamazoo River Superfund sites. Although the causes are not well understood, there is general
agreement that empirical estimates of PCB trends should adjust statistically for these temporal
trends in lipid content.
Pre-dredging recovery rates estimated from various media and analysis methods and time intervals
are generally consistent with the approximately 8 percent recovery rate projected by EPA with the
HUDTOX model in the upper river. Although some estimated rates are lower than the expected 8
percent, other estimates exceed the 8 percent expectation and no estimates are sufficiently
definitive to suggest that future trends in fish PCB recovery will fall below the recovery rates
anticipated in the ROD.
As reviewers have pointed out, because of the large change in the Upper Hudson in-river
environment as a result of the remedy, pre-dredging natural recovery rates may or may not be
predictive of future rates, particularly as tissue PCB levels approach regional background levels
when recovery rates are expected to decline. Nonetheless, no other empirical data are available for
estimating recovery rates at this time. Through development of Appendix 3 of the FYR report,
EPA found that recovery rate estimates were sensitive to how lipid content was handled either as
a variable in multiple regression or as a normalizing factor. It was noted that in many species-by-
location combinations, lipid content varied substantially through time; it was also found through
examination of paired lipid measurements that there was some measurement error in lipid content.
Understanding that lipid normalization has strong parallels to regression methods and is therefore
subject to similar sensitivities, EPA plans to further evaluate: 1) how temporal trends in lipid
content may influence reliability of PCB trend estimates, 2) if temporal trends in lipid content may
vary by species or tissue type, and 3) whether diagnostics can be developed to identify situations
for which PCB trend estimates are most likely to be accurate.
As stated previously, EPA disagrees with the need for additional retrospective studies of
differences in PCB measurements associated with the change in processing protocols because
remaining technical questions related to fish tissue PCB concentrations are prospective and do not
73
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
require linkages to pre-dredging monitoring data. Resources would be better allocated toward
prospective evaluations that will improve future data quality to provide more reliable estimates of
key metrics in the OM&M period. Post-remedial evaluation of fillet data will focus on how PCB
concentrations relate to the interim fish targets and remediation goal (i.e., 0.4, 0.2 and 0.05 mg/kg
PCBs) and the time required to meet those milestones. Comparisons with pre-dredging tissue
levels are not useful for these evaluations.
One could argue that better understanding of pre-dredging recovery rates would inform
understanding of times to recovery, but this requires the assumption that pre- and post-remedial
rates will be similar. While this may be true, the only way to test that hypothesis is to estimate
post-dredge rates from data and compare them. However, once post-dredging rates can be
estimated for comparison, there would no longer be a need to apply the pre-dredging rates.
At this time, EPA's focus is on understanding current tissue PCB concentrations, post-dredging
natural recovery rates and the time necessary to reach remedial objectives. For these prospective
objectives, the utility of a paired comparison of the effects of the protocol deviation would be of
little value because:
1. The non-standard fillet data will not really influence future decisions on the
protectiveness of the remedy. These decisions will be based on standard fillet data
collected post-dredging.
2. The rib-out data span a relatively short portion (2007 to 2008) of the overall baseline
monitoring period (to ), and therefore, do not have substantial influence considering the
extent of the NYSDEC baseline data (back to 1997 and earlier). In addition, NYSDEC
collected some samples processed according to the protocol during the 2007 to 2008
period, so we could identify large discrepancies if they occurred.
3. The special study conducted by GE and EPA shows that when the non-standard fillet data
are used with lipid normalization, the data can be combined with standard fillet data for
determining trends, but that results should be interpreted judiciously.
4. Samples analyzed without ribs were mostly obtained from the UHR during dredging. A
correction to the "true" value for the dredging period is largely academic since we cannot
hope to recreate the actual exposure conditions during dredging due to their highly
transient nature. As a result, there is very limited application to model improvement or
even Biota-Sediment Accumulation Factor (BSAF) refinement to be gained from such a
correction.
5. There are uncertainties associated with the differences between Aroclor-based, historical
capillary column (homologue-equivalent)-based, and congener-specific isotope dilution-
based analytical methods. EPA is working on procedures to minimize and understand any
differences between laboratories and methods as part of the OM&M program. However,
accounting for these uncertainties, by implementing a fillet-processing driven correction,
would yield highly uncertain values of little technical value, particularly with respect to
fish data from the remedial action period (2009 to 2015).
74
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.3.21 Comment 49: EPA's use of the data on fish body burdens to estimate the rates of
recovery is highly subjective. EPA's analysis of trends does not support their
conclusions about the rate of decline during the period 1995-2008
Comment
A number of comments were provided regarding the use of fish tissue PCB concentrations as a
means to calculate the rate of decline and support the viability of EPA's original modeling analysis.
These comments include:
a) The fish data show shorter term variations that yield very different decay rates than what
EPA calculated for the entire 1995 to 2008 period, thus EPA's choice of period is
arbitrary. In general, other selected averaging intervals yielded slower rates than those
obtained by EPA.
b) The rates of recovery across the individual species-location pairs vary drastically. The
use of an average rate is deceptive in supporting EPA's protectiveness statement for the
Site, because those fish populations with slow recovery rates or slightly increasing trends
have half-lives several decades longer than the 8 years suggested by the 8 percent rate.
These populations will continue to be an exposure risk for human health beyond the
timeframe suggested by the FYR. The use of average recovery rates does not consider the
variability in individual recovery rates by species.
c) Exclusion of the fillet samples generated without ribs from 2007 to 2008 yields
dramatically slower decay rates for Hudson River fish. If inclusion of the fillet samples
generated without ribs produced a trend line truly representative of fish tissue MNA
recovery, then the rate of recovery would not be consistently slower across species and
River Sections once those data are removed.
d) Fish tissue concentration decay rates are extremely variable such that the 8 percent
average decay rate is a highly uncertain, biased high, and oversimplified representation of
this variation. EPA's claimed 8 percent rate of recovery exaggerates the estimate of the
rate of natural recovery in the Hudson River. At present, it cannot be concluded from any
of the analyses performed that rates of recovery are on track with the ROD model output.
The data does not support EPA's conclusion that the goals of the ROD will be achieved.
e) EPA overestimated the average rate of decline for adult sport fish in the Upper Hudson
River (UHR).
f) Pumpkinseed (PKSD) samples did not show the variation in rate estimates since they
were not subject to sample processing differences with respect to inclusion or exclusion
of the rib cages.
Response
In the FYR report, EPA's examination of the historical record from 1995 to 2008 was intended to
characterize fish body burden trends when external inputs to the river were largely controlled and
significantly reduced relative to previous periods. The period 1998 to 2008 (a subset of the 1995
to 2008 period) also represents a forecast period for EPA's models (HUDTOX and FISHRAND),
75
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
thereby a means to test their accuracy by comparing the forecasts with observations. For the
models to generally agree with the data over such a long period of time, they would need to
represent the internal exchanges of PCBs between fish, sediment and water in a manner that
reflected actual conditions and the rates at which those exchanges took place. From these
comparisons, EPA could examine how well the models represented the actual environmental
conditions and processes that occur in the river, as the water flows downstream from Ft. Edward.
The agreement between model and data, if it could be demonstrated, would justify EPA's use of
the models as decision tools to evaluate the relative benefits of several remedial scenarios. In
addition, EPA's goal in this FYR was not to establish the decay rate for MNA for future conditions
for the river. EPA does not expect the decay rates observed during 1995 to 2008 to fully represent
post-remedy recovery rates. EPA disagrees with the commenter's assertion that the models
overestimated the rates of recovery during the MNA period. As shown in the multiple figures in
Appendix 3 of the FYR report and as discussed further below, there is good agreement between
model forecasts and the data during the MNA period. Thus, the results justify EPA's use of the
models as decision tools to evaluate the relative benefits of several remedial scenarios. As to future
conditions, reliable estimates of the actual post-remedy recovery rates are best derived from post-
dredging data.
EPA has characterized UHR decay rates with a single average based on the results of several
monitoring stations, each of which yielded similar average decay rates across species. However,
EPA did not characterize the rates in the Lower Hudson River (LHR) in this manner. EPA has
already indicated that LHR decay rates below Albany are not strongly linked to UHR conditions
since the rates of change are slower there and decline with distance downstream of Albany. Thus,
the decay rates in this region are not well-represented by a single LHR recovery rate average (and
EPA does not calculate one). EPA notes this observation is not fully consistent with EPA's original
model expectations. This was an important finding concerning the LHR, as discussed elsewhere
in the FYR.
The observation that decay rates vary by species and location across the Hudson but yield similar
averages in the UHR speaks to the robust nature of the fish monitoring program and EPA's
approach in analyzing the data. Local conditions ultimately control fish body burdens but these
conditions can vary widely even within a single river reach. Thus, EPA's approach, by considering
the larger data sets for individual fish at each station, effectively averages across the various local
conditions at each station. As shown in Appendix 3 of the FYR report, the model forecasts agree
well with many fish species across the UHR, although not all species in all instances. EPA's use
of an average rate of decline integrates across the various sources of uncertainty and incorporates
the fish species/river mile pairs that did not match well. Nonetheless, if EPA's analyses were as
uncertain as maintained by the commenters, it is highly unlikely that each UHR section would
yield approximately the same average rate of decline. The examination of many fish at each station
yields a robust basis on which to determine the average decay rates.
As noted above, fish body burdens of PCB are determined by local conditions. However, these
local conditions can be significantly impacted by external variables such as river flow, water
temperature, and nutrient loads, which impact fish growth, fish feeding preferences, food
availability and other factors. Thus, short-term variations in the fish tissue concentrations are
anticipated. EPA agrees that the estimated decay rates for shorter time periods would be very
76
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
different than, for example, the nine percent decay rate for the 1995 to 2008 period based on lipid-
normalized concentrations for largemouth bass in RS 1, with some slower periods and some faster
periods than those estimated by EPA. For example, if only the last nine years are considered (the
period 2000 to 2008), the decay rate more than doubles to 19 percent per year for largemouth bass
lipid-normalized concentrations in this river section. Figures 49-la and 49-le further demonstrate
the range of the actual decay rates based on varying time windows, specifically the decay rates for
3, 5, 8, 9 and 10-year intervals for lipid-normalized PCB concentrations for five different fish in
RS 1 based on the 1998 to 2008 data. (See Appendix B of this document for further explanation
of these figures). From the figures it is evident that short-term rates can vary substantially (more
than 600 percent) from the long-term rates. The results also indicate that the variability of the
decay rate decreases as the length of the window approaches the period of available data.
EPA also conducted a power analysis to determine the ability to detect a 5 percent or an 8 percent
annualized decline over 8 years and 4 years. The analysis was based on lipid-normalized data for
largemouth bass and PKSD at RS 1-TD 5 station (RM 189). The sample size was assumed to be
15 to 20 samples per year. A power of 0.8 means that there is 80 percent probability to detect a
true trend. The results (Table 49-1) indicate that with 8 years of monitoring data, the probability
to detect an 8 percent annualized decline is 99 percent for PKSD and 90 percent for largemouth
bass. If the annualized rate is 5 percent, the probability to detect the trend is reduced to 85 percent
for PKSD and 53 percent for largemouth mass. When the analysis is conducted over a 4-year
period, the probability to detect a 5 percent or 8 percent decline for either species is less than 50
percent. A power of 0.8 or greater is the required power to identify a true trend. The analysis
supports the premise that a trend derived from short-term interval is highly uncertain, and that at
least 8 years of monitoring data are needed to detect an 8 percent annualized decline. Additional
years of data may be required if the rate is slower than 8 percent. Therefore, in order to examine
long-term recovery rates, it is important to examine the longest possible record available, so that
these shorter-term fluctuations are averaged out. Thus, the numerical model simulated conditions
from 1977/1978 to the present while also forecasting future conditions; the overall goal of such a
long simulation period was to capture the long-term average trends in PCB concentrations in the
various media. In this regard, the EPA's choice to examine 1995 to 2008 was not arbitrary, but
rather designed to examine the longest period where external loads to the UHR were relatively
small and well-defined. The observation that fish tissue concentrations increase and decrease over
short periods of time does not detract from the model's ability to capture the long-term trends
across the entire UHR for the 1995 to 2008 period.
Further to this point, EPA is not basing its evaluation on the observations of the MNA rate itself
prior to dredging. This period simply provided the EPA the opportunity to test the models'
accuracy while waiting for actual measurements of post-dredging conditions to become available.
EPA's evaluation indicated the following:
• The remedy removed more PCB mass than anticipated;
• The remedy reduced surface concentrations better than anticipated in RS 1, as expected in
RS 3 and only somewhat less than planned in RS 2 [post-dredging sampling indicates that
surface concentrations in all three sections have declined even further than originally
targeted by remediation since completion of the remedy];
77
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
• The models on which the EPA based its decision were able to forecast UHR conditions
accurately over a lengthy period (1998 to 2008), and thus, they should be useful
indicators of the anticipated degree of recovery post-dredging; and
• PCB levels in fish continued to decline during the period 1998 to 2008, and have returned
to levels at or below 2008 conditions in most areas of the river as of 2016.
With regard to individual species decay rates, EPA has pointedly displayed the relationships
between model forecasts and available data for all of the main monitoring stations throughout the
Hudson, on both wet weight and lipid-normalized bases (see Figures A3-2 to A3-15 of Appendix
3 of the FYR report). In presenting these data and model forecasts together, EPA has shown where
the model and data agree and where they do not. EPA does not agree with the assertion that the
observed rates of recovery during 1998 to 2008 are not in agreement with those predicted by the
model for the UHR and Albany area.
In Figures A3-16A to A3-16C, EPA shows the individual decay rates for all species with sufficient
data to support a decay rate estimate. The variability among species and across stations is directly
shown and considered in these figures. It is clear from these figures that the averages are good
representations of the estimated decay rates. More to the point, although these graphs depict pre-
dredging conditions, and confirm the accuracy of the model forecasts of MNA, the graphs do not
show post-dredging behavior. For that, EPA is awaiting data to be collected in the coming years.
While the model and data comparisons during the pre-dredging period confirm that the model's
use in analysis and decision-making in the feasibility study and the ROD is justified, the data-
based rates of recovery observed prior to dredging are not a basis to estimate post-remedy recovery
rates. Furthermore, although the models did forecast post-remedy rates of recovery, reliable
estimates of the actual post-remedy recovery rates are best derived from post-dredging data.
The commenters also raise the concern that the departure from the standard fish sample processing
protocol {i.e., without ribs) has so impacted the 2007 to 2008 data that no data generated without
ribs should be considered. As EPA shows in Figure A3-16C, exclusion of the data generated
without ribs does add variance to the estimates of decay rates but still leads to the same major
conclusions; that is, that lipid-normalized decay rates are about eight percent per year in the UHR,
and that these rates decline with distance downstream in the LHR below Albany. However, the
assertion that the data generated without ribs are invalid is simply incorrect. As discussed in
Appendix 3 of the FYR report and detailed in Appendix D of this document, EPA directed GE to
complete a special study of black bass fillet tissue with and without ribs. The results of the analysis
show that data generated without ribs were largely comparable to those generated with ribs on a
lipid-normalized basis, with a difference of less than 20 percent. Thus, it is appropriate to include
the data generated without ribs in the lipid-normalized data trend analysis.
EPA further explored the commenter's assertion that UHR fish body burden decay rates decrease
markedly if the 2007 and 2008 data generated without ribs are excluded. EPA agrees that these
rates do decrease without the additional 2 years of data, but this change is not due to the data
generated without ribs. Rather, it is the result of the particularly low PCB concentrations in fish
tissue in most species for years 2007 and 2008. EPA notes that the exclusion of the 2007 and 2008
data generated without ribs has a significant impact on the decay rate for some species such as
largemouth bass in RS 1, roughly about a sixty percent reduction in rate. However, for a similar
78
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
species, specifically smallmouth bass, the decay rate was actually increased by excluding the 2007
to 2008 data generated without ribs, by about 35 percent (indicating faster recovery). This suggests
that the exclusion of the rib during sample processing does not always yield a less contaminated
sample.
EPA further explored this possibility by examining the change in decay rate for PKSD, a species
analyzed on a whole-body basis, and not subject to the rib processing issue. Lipid-normalized
PKSD PCB data are plotted in the attached figure (Figure 49-2), replicating the PKSD graphs from
Figures A3-9A and A3-10A in Appendix 3 of the FYR report. In both instances, the original
regression line and equation are shown based on a fit to the data from 1995 to 2008. The decay
rates are the same for both RS 1 and RS 2, as it turns out, -4.9 percent per year. Also shown on
each graph is an additional regression, fit to the data from 1995 to 2006, excluding the 2007 and
2008 data. This parallels the analysis done by the commenter for species processed on a fillet basis,
excluding those samples processed without rib cages. For PKSD, the decay rate in RS 1 drops
from -4.9 percent per year to -0.8 percent per year by excluding those two years of data. In RS 2,
the data actually show a positive trend, changing from to -4.9 percent per year to +0.8 percent per
year by excluding those two years of data. These changes yield an 80 to 115 percent decline in the
decay rate simply by arbitrarily excluding the last two years of data. This analysis shows that
PKSD show a similar change in decay rate if the last two years of data are excluded as that seen
for fillet-based species trends when the rib-excluded samples (from 2007 and 2008) are omitted
from the analysis. This result is contrary to the assertion by the commenter who ascribes the change
in trend to the effect of the rib-excluded samples. Rather, EPA's analysis indicates that the
reduction in decay rate calculated by the commenter is largely due to the omission of the data for
2007 and 2008.
While EPA agrees that the exclusion of the rib cage from fillet samples does result in lower PCB
levels on a wet weight basis, this observation regarding lipid-normalized PCB levels in PKSD
sample indicates the observation of slower decay rates with the exclusion of 2007 and 2008 data
for all sample types is more likely related to the exclusion of these years of data. On this basis, the
exclusion of the last two years of data is not justified. The relatively small effect of rib removal on
lipid-normalized concentrations (estimated to be less than 20 percent) does not justify the
exclusion of the 2007 and 2008 data from the trend analysis. As EPA has already noted, the goal
of EPA's analysis is to estimate the long-term rate of decline over the period 1995 to 2008 as
compared with the model-estimated rate of decline for the same period. This analysis shows that
the 2007 to 2008 data should be a part of that analysis.
Lastly, EPA agrees with the commenter that there are not enough data available since the
completion of dredging and related project activities in 2015 to determine if the remedy will be
protective within the time frame anticipated by the ROD. While sediment and water both have
sufficient data to identify the reductions due to dredging, EPA estimates that as many as eight or
more years of post-dredging fish tissue data are needed to establish a statistically relevant trend
for fish.
79
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Table 49-1 Power to detect 8 percent or 5 percent annualized change with monitoring data
for
years and 8 years
5th
95th
Years of
Percentile
Percentile
Monitoring
True Rate of
Estimate
Estimate
Data
Decline (%/year)
Species
Power
(%/year)
(%/year)
4
-5%
Pumpkinseed
0.20
-12%
2%
Largemouth Bass
0.13
-15%
4%
-8%
Pumpkinseed
0.46
-14%
0%
Largemouth Bass
0.28
-18%
1%
8
-5%
Pumpkinseed
0.85
-8%
-2%
Largemouth Bass
0.53
-9%
-1%
-8%
Pumpkinseed
0.99
-10%
-5%
Largemouth Bass
0.90
-12%
-4%
Notes:
• Analysis was performed based on lipid-normalized Tri+ PCB concentrations in largemouth bass and
pumpkinseed at RS1-TD5 station (RM 189).
• The sample size was set at 17 per year. Similar results were obtained when sample size varied from 15 to 20.
• A power of 0.8 means that there is 80% probability to detect a declining trend that is there. A power of 0.8 or
higher is the desired power for trend analysis.
• Positive value for the 95th percentile indicates a reasonable probability for incorrectly detecting a positive trend
when the true trend is negative.
3.3.22 Comment 50: The impact of dredging on fish tissue PCB concentrations has passed
and concentrations have now reached equilibrium. Future declines in concentration
will be very gradual and prolong the time to achieve ROD targets
Comment
Commenters asserted that fish tissue in all river sections experienced a transient increase in PCB
concentration in the one to two years following dredging upstream and then a subsequent stepdown
in concentration. They stated that the data do not support EPA's contention that fish tissue
concentrations are still being significantly impacted by the dredging activity. Commenters argued
that fish tissue concentrations have returned to pre-dredging concentrations and have reached
equilibrium concentrations, with additional declines in fish tissue PCB concentrations occurring
only gradually, over a very long time. Commenters stated EPA is now left with fish tissue
concentrations that are more elevated than expected at the time of the 2002 ROD and it is very
unlikely that these concentrations will decline at the rate EPA predicted.
Response
EPA agrees with commenters that fish tissue at many stations experienced short-term and transient
increases in PCB concentrations during proximal dredging or dredging-related activities. In
addition, barge traffic around previously dredged areas may have had an impact on equilibration
of conditions. However, EPA disagrees that it can be concluded that fish tissue PCB concentrations
have now reached a post-dredging equilibrium. For many species, (e.g., brown bullhead and black
bass at Thompson Island Pool [TIP], pumpkinseed [PKSD] and yellow perch at Stillwater, and
80
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
other species discussed in Appendix 8 of the FYR report), concentrations are only now returning
to pre-dredging conditions and 2016 data exhibit continuing downward trends in fish tissue
concentrations following the most recent upstream dredging activities. With only one year of post-
dredging fish tissue data collected, there currently isn't enough data to conclude that post-dredging
equilibrium has or has not been established; additional data collection over a number of years will
be required to fully establish the trajectories of the fish tissue concentration recovery and conclude
when equilibrium has been reached.
A review of data at the Cumberland Bay - Wilcox Dock Superfund Site (Lake Champlain) where
remedial dredging was completed in 2000 may provide context on approximately how long it may
take to establish equilibrium concentrations. Appendix 8 of the FYR report, Figure A8-5.1 and
A8-5.2 show yearly wet weight TPCB concentrations in fish tissue of rock bass and yellow perch,
respectively, collected at that site. In both cases, PCB concentrations continued to decline for at
least 8 years following dredging activities. In the case of yellow perch, concentrations returned to
pre-dredging concentrations following dredging activities and remained there for upwards of 5
years before declining to PCB concentrations that were significantly below pre-dredging
concentrations. Thus, based on pre- and post-dredging data collected at the Cumberland Bay -
Wilcox Dock Site, 7 to 9 years may be required to develop a complete picture of rates of decline
of fish tissue concentrations.
Therefore, it is important that additional data over multiple annual cycles (likely 8 years or more)
be collected to more fully understand how fish are responding to dredging and provide statistically
meaningful estimates of progress toward meeting the interim targets and final goals. While EPA
finds the 2016 fish data results encouraging, one year of data does not suggest trends and cannot
not be used to conclude that fish tissue concentrations have reached an equilibrium. EPA will
continue to monitor post-dredging (natural recovery) results collected under OM&M and to
evaluate remedy protectiveness through the FYR process which includes comparing future
observations to the ROD targets and remedial goals.
3.3.23 Comment 51: Changes in fish sampling locations result in data that is not suitable for
long term PCB temporal trend analysis
Comment
Several commenters noted that fish PCB tissue concentration data exhibit variability across the
Upper Lower River (ULR) and Lower Hudson River (LHR) and through time, specifically when
viewed from the perspective of pre-dredge compared to post-dredge. Reviewers asserted that some
of the observed variability may be attributed to changes in the locations from which some fish
species were collected, while others suggested that the observations warrant further investigation,
with commenters mentioning pumpkinseed (PKSD) in particular. A commenter further suggested
that EPA only "use only lipid-normalized data to evaluate temporal trends and for comparison to
food web model projections use wet weight values adjusted to the standard lipid content for each
fish species used in the modeling."
Furthermore, commenters concluded that the decay rates in the UHR were generally low when
only using the long-term monitoring species (or species groups) and stations established by
81
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
NYSDEC and when restricting the size range and time of year to be consistent with NYSDEC
monitoring. Specifically, their analysis shows that "only black bass and yellow perch from the
Thompson Island Pool monitoring station show PCB decay rates greater than 8 percent. Bullhead
and pumpkinseed from that same location have PCB decay rates of less than 5 percent and 0
percent, respectively. At the other UHR long-term monitoring locations in the Stillwater Pool, all
species had PCB decay rates less than 5 percent."
In addition, a commenter stated that the MNA period for fish should begin in 1997 rather than
1995, which is consistent with prior practice {i.e., use consistent data).
Response
EPA agrees that there is considerable variability in long-term fish tissue data trends (see
Appendices 3 and 5 of the FYR report), even when data associated with non-NYSDEC fillet
processing protocols are excluded from the analyses. EPA also agrees that the affected lipid data
(i.e. using lipid-normalized data) should be included in temporal trends analyses. The 2002 ROD
anticipated that fish tissue data would continue to exhibit variability by species and across time
and stations, both during dredging and following the conclusion of dredging (See Appendix 8 of
the FYR report).
EPA does not agree that changes in sampling locations over time make fish tissue data unreliable
or inappropriate for assessing temporal trends. As described in the 2004 BMP QAPP (Section B1.2
"Upper Hudson Fish Monitoring"), the locations, sampling frequency, target numbers and species
for the Baseline Monitoring Program (BMP) and Remedial Action Monitoring Program (RAMP)
were based on NYSDEC long-term monitoring approaches and represented species associated
with a range of sediments and human and ecological uses. As indicated in the NYSDEC 2005
Report on PCBs in the Hudson River, "With the adoption of the Baseline Monitoring Program
(BMP), as part of the PCB Remediation effort, GE took over a portion of the fish monitoring
beginning in 2004, but they have adopted the basic DEC plan for the Upper River" (Sloan et al
2005). These long-term data, in conjunction with data generated during the BMP, were intended
to be used in spatial and temporal trends analyses of PCB concentrations in UHR fish and are still
used, in concert with RAMP and post-dredging data, for this purpose. The 2004 BMP QAPP
(along with the Phase 1 and Phase 2 RAMP QAPPs) indicates that "reasonable attempts will be
made to maintain sample location integrity throughout the program," but in the event that fish
cannot be collected at each location in every year, at least two locations within each pool or reach
and two locations within each river section {e.g., stations ND-1, ND-2, ND-3 and ND-5 in the
Northumberland pools in RS 2, consisting of Reaches 7 and 6) would be sampled. Note that the
number of fish stations within Reaches 8 through 5 increased under the BMP and these stations
represent the current (RAMP) collection stations. EPA, GE, and NYSDEC are discussing potential
fish sampling stations in Reaches 4 through 1 under the OM&M program.
An examination of the locations of the fish stations actually sampled each year at the UHR and
Albany-Troy monitoring locations indicates that, during both the BMP and RAMP periods, the
location of fish collected during the spring and fall sampling windows changed over time. These
changes were sometimes necessary to collect target species each year in the vicinity (plus or minus
approximately one river mile) of a historical station {e.g., Station TD-1 2004 to 2016), loss of
82
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
habitat (e.g., the abandonment of Station ND-4 in Reach 6 in 2004/2005), or a combination of
habitat loss and operational considerations (e.g., the transition from the south Albany turning basin
at RM 143 to transects along the east and west shores of the Hudson River between RM 145 and
147 between 2004 and 2013). Figure 51-1 presents the change of fish stations over time in RS 1
and RS 3. Note that long-term monitoring was not conducted by NYSDEC in RS 2. Because the
sampling of fish is performed along a transect and the actual sampling location depends on the
availability of fish, the NYSDEC "stations" historically occupied represent relatively large areas
(usually within a mile radius) generally centered on the designated station location (as described
below consistent with the 2004 BMP QAPP). In consideration of this, GE BMP locations that are
within a 1-mile radius of the assigned "NYSDEC stations" can be considered as equivalent to the
NYSDEC stations. These equivalent locations are marked within the red rectangle boxes in Figure
51-1. These figures show that brown bullhead, largemouth bass, PKSD, and yellow perch were all
consistently monitored at the "NYSDEC stations" from 1995 to 2008.
As described in the 2004 BMP QAPP (Table B-3) and subsequent RAMP QAPP, the intent of the
selected post-2003 stations was (and is) to sample from available habitats "approximately evenly
distributed (depending on habitat availability) within the pool [or pools]" for the purpose of
"establishing] a baseline for comparison to construction and post-construction conditions." Also,
as discussed in the BMP Data Summary Reports, changes in sampling locations were made in
consultation with NYSDEC and EPA field oversight. For these reasons, EPA does not agree that
changes in sampling locations over time significantly impact EPA's ability to use these data in
long-term trend analyses. Of particular note, an insistence on limiting data to a single "station"
effectively limits the temporal coverage of the data, reducing the ability to detect long-term trends
in the data.
To illustrate EPA's perspective, EPA evaluated the impact of station inclusion on the decay rates
of brown bullhead, large-mouth bass, PKSD, and yellow perch in RS 1. Specifically, EPA
compared the rates that were derived from samples at the established NYSDEC station (RM 189),
within a 1-mile radius of the NYSDEC station (RM 188.5-190), and from all RS 1 locations (RM
189-194). The decay rates were calculated on a lipid-normalized concentration basis using samples
collected between 1997 and 2008. Note that based on the reduction of PCB loads originating above
OU2, MNA began sometime between 1995 and 1996. On this basis, 1995 was used as starting
year for the MNA period in the FYR report. However, to address the commenter's concern that
MNA did not begin until 1997, the decay rate analysis presented in this response begins in 1997.
EPA obtains the same or similar decay rates whether the analysis begins in 1995 or 1997.
As shown in Table 51-1, for three of the four species, the rates are not impacted by the stations
included when the restriction on the location does not impact the available length of time (number
of years) for deriving the trend. For brown bullhead, large-mouth bass and yellow perch, their rates
of decline based on data within a 1-mile radius of the NYSDEC station (RM 188.5-190) were
similar to the rates obtained using all RS 1 stations. This is attributed to the similar temporal
coverage of the data (i.e., 1997 to 2008) available for both data sets. In contrast, the rates for these
species were much slower when only using the single NYSDEC station (RM 189) because they
were derived from a shorter time period, (i.e., 1997 to 2005). By restricting the samples to the
NYSDEC station, the number of years for the trend analysis is significantly reduced and the
downward trend observed across the entire period being examined here (1997 to 2008) is not
83
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
strongly evident from 1997 to 2005. Conversely, all stations from 2004 to 2008 exhibit a strong
downward trend. From these observations, as illustrated in detail in Appendix 3 of the FYR report,
it is likely that all stations exhibit both slow and rapid periods of decline, providing further support
for the longer-term rate calculation applied by EPA to minimize the effect of short-term trends.
PKSD does not show the same behavior as the other species. Its rate of decline based on data
within a 1-mile radius of the NYSDEC station (RM 188.5-190) was lower than that obtained using
all RS 1 stations. However, when looking at the most recent data (2004 to 2008), the rate of decline
of PKSD is high and is comparable to the other species. This is further illustrated in Figures 51-
2a and 51-2b, which present lipid-normalized PCB concentrations for largemouth bass and PKSD
in RS 1. Note that a single station covers the period up to about 2003 for PKSD and to 2005 for
largemouth bass, with subsequent conditions documented at other locations within the same river
section. Note that all stations post-2003 indicate a downward trend. Additionally, within a given
year, the variability of PCB levels across stations is comparable to the variation within a single
station, indicating that most stations are tracking similar conditions. These observations support
the combination of multiple stations within a river section in the calculation of rates of decline.
These results suggest that it is not the sampling location, but the temporal span of the data that
primarily impacts the calculated rate of decline. Longer-term trends provide the best estimates of
the actual rate of fish tissue recovery.
A commenter also mentioned that the high rate of decline for PKSD at the Albany/Troy location
was highly unreliable because of the change in sampling location. The conclusion was made based
on the commenter's findings that "at the Albany/Troy location all species except pumpkinseed had
PCB decay rates of 4% or less" and "the decay rate of pumpkinseed was low at other locations in
the Lower Hudson River (LHR) (Catskill and Poughkeepsie)". EPA does not agree with the
commenter's assertion that at the Albany/Troy location all species except PKSD had PCB decay
rates of 4 percent or less. EPA's analysis, based on lipid-normalized data, shows that the decay
rates of smallmouth bass, largemouth bass, brown bullhead, yellow perch, PKSD, spottail shiner
and striped bass were all greater than 8 percent for this river section (Table A3-3 of Appendix 3
in the FYR report). Considering that PKSD is known to show high site fidelity compared to
largemouth bass, smallmouth bass and bullhead, the similar decay rate of these species implies
that the change of locations is not the cause of the high decay rate of PKSD observed in this river
section. The lipid-normalized PCB concentrations for PKSD at the Albany/Troy location are
shown in Figure 51-3. This figure shows that PKSD PCB levels are declining no matter which
time interval or monitoring location is selected. The period prior to 2004 is characterized by a rate
of decline of -7 percent per year while the period 2004 to 2008 is characterized by a rate of -20
percent per year. The overall rate of decline for 1997 to 2008 is -17 percent per year. EPA notes
that in Table A3-3 of FYR report Appendix 3, EPA obtains a rate of-13 percent per year, based
on a slightly long period, 1995 to 2008.
The lower rates of decline for PKSD observed at other locations in the LHR as claimed by the
reviewer do not support the conclusion that "the high decay rate of pumpkinseed at the
Albany/Troy location is highly unreliable". This is because the rate of decline across all species is
shown to decrease with distance downstream (i.e., downstream locations recover more slowly than
the upstream locations under MNA). The spatial pattern of decay rates is illustrated in Figure A3-
16 of Appendix 3 of the FYR report.
84
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
EPA agrees that based on one year (2016) of post-dredging data, differences between BMP and
post-dredge PCB fish tissue levels can be observed at the river section and station scale. However,
such variability was anticipated by the ROD as "short-term temporary impacts" to aquatic species
and habitats resulting from dredging. As discussed in Appendix 8 of the FYR report, some of this
variability may be due to the proximity of dredging and dredging-related activities to fish habitat
and fish sampling stations. Other differences between pre- and post-dredging fish tissue
concentrations could be attributed to stresses resulting from habitat changes, variations in lipid
levels and changes in other uncharacterized environmental conditions.
A commenter accurately noted that RS 1 "post-dredging TPCB concentrations in pumpkinseed
and small forage fish were three to six times lower than observed pre-dredging
levels...[while]...further downstream, the results are mixed." However, RS 1 (including PKSD
and forage species) fish tissue concentrations started out at higher pre-dredging concentrations
than either the Albany-Troy or LHR Stations (See Appendix 3 of the FYR report, Figures A3-2
through A3-5). And, prior to dredging, target fish species in RS 1-3 and RM 152 at Albany-Troy
also exhibited similar average recovery rates (Appendix 3 of the FYR report, Figure A3-16). In
contrast, stations below RM 152 started out at lower pre-dredge fish tissue levels and exhibited
recovery rates prior to dredging that were slower than or even positive (meaning increasing over
time) when compared with those observed at UHR stations or at RM 152. In fact, striped bass,
yellow perch, and white perch at LHR stations below RM 90 (NYSDEC data) have exhibited
tissue PCB concentrations approaching the 0.4 mg/kg target level since the BMP. PCB levels in
Lower Hudson fish below RM 140 are primarily governed by local sediment conditions and are
not closely linked to conditions in the Upper Hudson, nor the Upper Hudson PCB loads to the
Lower Hudson. Large reductions in fish tissue levels are anticipated for the Upper Hudson fish in
response to the remedy, while less change is expected in response in Lower Hudson fish.
The extent of variability in fish tissue concentrations observed at the various fish sampling stations
during the BMP and RAMP is neither inconsistent with levels anticipated in the ROD nor
unexpected considering the range of species observed and the distance over which they were
collected for the project. For the reasons listed above, while EPA acknowledges the variability in
results across fish monitoring stations, it does not agree that variation among species and locations
requires additional investigation at this time in the Upper Hudson. Given the lack of correlation
between Upper Hudson and Lower Hudson fish tissue responses, EPA has identified the LHR as
an area requiring further investigation.
Regarding the use of lipid-normalization, EPA does not agree that field-collected wet-weight data
should be adjusted to (modeled) lipid values such as those reflected in pre-dredging forecasts from
the ROD. For the ROD forecasts, lipid values were assigned using random values from a tri-modal
distribution of lipid concentrations based on historical data. As a result, forecast lipid values reflect
a potential range from within an estimated population. This approach was used because it was
understood that because species' lipid values vary over time, it was not possible to predict lipid
levels precisely into the future, but an accurate estimate was still required to construct forecasts.
In contrast to modeled lipids, field-collected fish data reflect the actual (observed) lipid content of
harvested sample species and at the level of an individual fish. Additionally, EPA notes that while
lipid levels and PCB levels in fish tissue are correlated, the relationship between the two
85
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
parameters is not always linear. That is, for two fish of the same species, age and size with
comparable PCB exposure but one with twice the lipid content, the fish with the higher lipid level
will not necessarily have twice the PCB level. As a result, the utility of adjusting observed
(individual fish) lipid contents to reflect a population-level model distribution is not clear.
A commenter raised the concern that the rate of decline for PCB levels in fish tissue was slower
than expected when only using the long-term monitoring species (or species groups) and stations
established by NYSDEC and when restricting the size range and time of year to be consistent with
NYSDEC monitoring. Without these restrictions, EPA's analysis using lipid-normalized data
shows a much higher decay rate than the commenter noted. In the UHR, only brown bullhead in
RS 2 and RS 3 shows a decay rate less than 5 percent (See Table A3-3 of Appendix 3 in the FYR
report). The low rates of decline reported by the commenter are the result of a specific sample
selection process. The commenter limited his/her selection to tissue samples exclusively from
NYSDEC-established stations, within a certain specimen size range and a specific time of year
from 1997 to 2006, thereby eliminating any of the GE data. As discussed above, the restriction to
NYSDEC-established stations basically excludes samples collected after 2004 and also restricts
data to a single station. Therefore, the commenter's analysis does not reflect long-term (temporal)
trends or fish tissues representative of the reach-scale. EPA further evaluated the impact of
specimen size on the rate of decline. The analysis indicates that the exclusion of fish samples that
do not meet the NYSDEC monitoring size criteria do not substantively affect the overall rates of
decline. It was noted that only a small fraction of samples do not meet NYSDEC's selection criteria
for brown bullhead (0.3 percent with length less than 175 mm), largemouth bass (3.6 percent with
length less than 250 mm), and yellow perch (9.3 percent with length less than 150 mm). However,
for PKSD, about 25 percent of samples do not meet NYSDEC's selection criteria. In Figure 51-4,
EPA presents the trends of PKSD in RS 3 to show that removal of data points by length is
inconsequential for the rate of decline. The restriction to the time of year also does not impact the
rate of decline as fish samples were largely collected from the same season from 1997 to 2008. As
an example, Figure 51-5 compares the trend of yellow perch at Albany/Troy using only spring
samples against that using all samples. The rate of decline for the spring samples is 9.7 percent per
year, which is similar to 10.4 percent per year derived from all samples.
The commenter also excluded the tissue samples from 2007 to 2008 in the trend analysis because
the samples were analyzed using a non-NYSDEC-standard fillet approach by not including the rib
cage material in the fillet harvested for analyses. GE has conducted a specific study to evaluate
whether or not inclusion of the rib cage (ribs) had a significant impact on fish tissue PCB
concentrations and lipid levels. For lipid-normalized data, the difference between the two fillet
approaches averages less than 20 percent (see Section 3.3 in Appendix 3 of the FYR report). EPA
also conducted a sensitivity analysis to compare the fish recovery rates from data generated with
and without ribs (compare Figure A3-16A with Figure A3-16C in Appendix 3 of the FYR report).
The results of this sensitivity analysis indicate a similar distribution of the estimated rates of
decline, with or without the non-NYSDEC standard fillet data. Therefore, EPA's conclusion on
the rates of decline and their distributions across the Hudson River, on a lipid-normalized basis, is
consistent regardless of whether the non-NYSDEC-standard fillet data are used or not.
The low rates of decline reported by the commenter are mainly attributable to the restriction to the
"NYSDEC station" which excludes the samples after 2004 and does not reflect the long-term
86
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
trend. EPA's goal is to derive trends that represent the entire species population in a wider area,
such as those 011 River Section basis, and over as long a period as possible so as best to capture
the long-term trends. The trend figures provided in Appendix 3 of the FYR report (Figures A3-9
through A3-11) clearly show a consistent downward trend of tissue PCB concentrations for various
species in three river sections. These figures also suggest that the rates of decline on average are
consistent throughout the UHR and RM 152.
Table 51-1 Impact of Sampling Locations on Temporal Fish Tissue Trends in RS 1
The decay rates were derived from lipid normalized Tri+ PCB data from 1997 to 2008
Species
RM 189 to RM 189.4
RM 188.5 to RM 190
All Locations
Count
Rate of
Decline
(%/yr)
Actual years
of data
available
Count
Rate of
Decline
(%/yr)
Actual years
of data
available
Count
Rate of
Decline
(%/yr)
Actual years
of data
available
Brown Bullhead
139
-3%
1997-2005
166
-6%
1997-2008
260
-8%
1997-2008
Larqe-mouth Bass
165
-3%
1997-2005
195
-11%
1997-2008
214
-11%
1997-2008
Pumpkinseed
104
5.4%
1997-2003™
156
-3%
1997-2008
262
-6%
1997-2008
Yellow Perch
152
-8%
1997-2005
182
-13%
1997-2008
316
-14%
1997-2008
Note: (1) Pumpkinseed: one sample from 2004 and one sample from 2005 were excluded from the analysis since the
data are too limited for these years.
Brown Bullhead
1994 1996 1998 2000 2002 2004 2006 2008
Year
Largemouth Bass
Legend
• NYSDEC sampling event
#GE sampling event
• EPA samples from NYSDEC
NYSDEC Historic Monitoring Station
TD5
1994 1996 1998 2000 2002 2004 2006 2008
Year
195
194
193
JBJ
2 192
&
» 191
ct
190
189
188
Pumpkinseed
TD5
Yellow Perch
TD5
1994 1996 1998 2000 2002 2004 2006 2008
Year
1994 1996 1998 2000 2002 2004 2006 2008
Year
Figure 51 -1 a Fish sampling stations in River Section 1
87
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
Brown Bullhead
SW3? 176
Largemouth Bass
1994 1996 1998 2000 2002 2004 2006 2008
Year
SW3
1994 1996 1998 2000 2002 2004 2006 2008
Year
Legend
# NYSDEC sampling event
• GE sampling event
• EPA samples from NYSDEC
NYSDEC Historic Monitoring Station
Pumpkinseed
SW5
1994 1996 1998 2000 2002 2004 2006 2008
Year
Yellow Perch
182
180
178
5 176
1 174
oc
172
170
168
SW3
1994 1996 1998 2000 2002 2004 2006 2008
Year
Figure 51-1 b Fish sampling stations in River Section 3
88
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
Large-mouth Bass in RSI
10000
1000
E
• • i I
—. ; I « f «
ii.ji • i - *
100
• RM189
¦ RM190
A RM191-193
¦ Expon. (All Data)
¦ Expon. (RM189-190)
Expon. (RM189)
All Data: y = 2l*99e^uu
RM 189-190: y = lE+95e-° *¦«
RM189: y - 2E.12e-°tft"
10 1 i 1 1
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Year
Figure 51-2a Variation of rates of decline in fish tissue concentration with station inclusion
for largeinouth bass at River Section 1
89
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
Pumpkinseed in RSI
1000
E
100
i !
• I. . ;
-M . ; . —+.
i
All Data: y = lE+53e0059*
RM189-190: y = 6E+25e-°D:i'*
RM189: y = 2E-4Se0-cl541,
I t~
- i I
> i
• RM189
B RM190
* RM191-194
Expon. (RM189)
Expon. (RM 189-190)
Expon. (All Data)
10
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Year
Figure 51-2b Variation of rates of decline in fish tissue concentration with station inclusion
for pumpkinseed at River Section 1
90
Final Second Five-Year Review Comment Response for the Hudson River PCBs Snperfimd Site April 2019
-------
Pumpkinseed at Albany/Troy
100
£
- •
S N|
• * .
i
:-i
• * • : X
• \ •
• ¦>.
All Data: y = 2E+145e-°-165x
RM142: y = lE+62e-°l)69*
RM150-152: y = 1E+I80e 0J05*
|N-
:i
s i
.a
• RM142( 1997-2003)
B RM150-152
Expon. (All Data)
Expori. (RM 142(1997-2003))
Expon. (RM 150-152)
10
0 a
1
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Year
Figure 51-3 Variation of rates of decline in fish tissue concentration with station inclusion
for pumpkinseed at Albany/Troy
91
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
Pumpkinseed in RS3
100
~Q0
£
CD
U
I ¦ a ¦ a
B . •
x a •
' "I I ! • " '
1 ¦ • ! ¦ B
i 1**1 fl " I 8
~ « | * *J o I ¦
: ! ¦ ¦ s..s !
:: : Mi s
¦ a x
¦ D 8
Q ¦ * ,
i ! a § | :
Size<120 mm: y = 4E+82e^093s
All Sizes: y - 5E+82e~°Q33"
¦ PKSD{<120 mm)
• PKSD{>120mm)
Expon. (All Size)
Expon. (PKSD{<120 mm))
10
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Year
Figure 51-4 Variation of rates of decline in fish tissue concentration with restriction on
species length for pumpkinseed at River Section 3
92
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
Yellow Perch at Albany/Troy
100
BO
E,
CO
u
• •
J •
" N ^ ^
S *
1
¦
¦
~
I
s
1
¦ Spring
• Fail
Expon. (Spring)
- - - Expon. (All Data)
All Samples: y = 2E+92e"°-104*
Spring Fish only: y = IE +86e^097x
10
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Year
Figure 51-5 Variation of rates of decline in fish tissue concentration with restriction on
season - Yellow Perch at Albany/Troy
3.3.24 Comment 53: Surface PCB Concentration of the Non-Dredge Areas in RSI has not
declined
Comment
A commenter stated that when the SSAP dataset is separated into dredged and non-dredged area
sample sets, cumulative distribution plots show lesser degrees of improvement in non-dredged
areas than the improvement shown by plotting all SSAP samples. Non-dredged areas in RS 1 show
very little or no improvement.
Response
EPA disagrees with the commenter's conclusion that non-dredged areas in RS 1 show very little
or no improvement, because the commenter's analysis was based on comparison of incompatible
data. The commenter compared the 2016 OM&M data surface sediment data representing the 0-2
inch interval, to the 2002 to 2005 SSAP sediment data representing sediment intervals from 0-12
inches. In general, deeper sediments outside of the dredged areas tend to be less contaminated, so
93
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
juxtaposition of the two differing sampling intervals does not provide a comparison of equivalent
metrics.
EPA has repeated the analysis by directly comparing the 0-2 inch SSAP data to the 2016 OM&M
data in RS 1, using a cumulative distribution function (CDF) plot shown in the Figure 53-1 below,
and the results show that there is significant improvement in the non-dredged areas. The y-value
on the CDF plot is interpreted as the proportion of population or probability with value less than
the corresponding x-value. For an environmental dataset, the y-values can be considered as
percentiles of a dataset. Figure 53-1 compares the distributions of concentrations (represented by
cumulative probability or percentiles) between 2002 to 2005 SSAP samples and 2016 OM&M
samples. At any given percentile (y-value), the concentration (x-value) from the 2002 to 2005
SSAP dataset is always greater than that from the 2016 OM&M samples. The data indicate that
spatially comparable concentrations in non-dredged areas of RS 1 have declined between the 2002-
2005 period and 2016, with a median (50th percentile) value decreased from 5.4 mg/kg to 2.3
mg/kg. The geometric mean has decreased from 4.3 mg/kg to 1.7 mg/kg, representing a two and a
half-fold decrease in TPCB concentration. The arithmetic mean has decreased from 8.4 mg/kg to
4.1 mg/kg. These results clearly indicate a reduction in TPCB concentration in RS 1 surficial
sediments within the past thirteen years. Based on the geometric mean results, over the course of
thirteen years, there has been a yearly 7 percent decrease in the concentration of TPCBs in the 0-
2 inch interval of the non-dredged areas.
Unlike RS 1, which was extensively sampled outside the dredged areas, RS 2 and RS 3 cannot be
effectively compared using the method applied by the commenter. The RS 2 and RS 3 sample
locations in the 2003 data set are focused primarily on the areas surrounding the CUs and do not
provide a spatially representative sample of the non-dredged river areas.
Using the side-scan sonar surveys of sediment texture across all three river sections, EPA
integrated TPCB concentrations based on cohesive and non-cohesive sediment textures to estimate
the change in TPCB concentration in each river section. This more rigorous assessment of the data
can be found in Appendix 4, Table A4-5 of the FYR report. The table shows the area-weighted
average concentration in RS 1 has decreased from 4.15 mg/kg to 1.7 mg/kg. This is equivalent to
the two and a half-fold reduction in TPCB concentration seen in the geometric mean.
94
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
1.03
2002-2005 SSAP
2016 OM&M
T 1—I I I I 11
0.01
a£B 0.1" " 1 2 ' 10 M 100a
230 AB7M
Total PCB [mg/kg)
Figure 53-1 Cumulative Distribution Function (CDF) Plot for 0-2 inch Sediment Samples
for RS 1 In Non-dredged Areas
3.3.25 Comment 56: Sediment concentrations remaining in the river are higher than
anticipated and sediment concentration rate of decline is overestimated
Commenters state that data collected after 2002 show higher levels of surface sediment
contamination than anticipated in portions of RS 2 and 3 that were not targeted for dredging and
that estimated post-dredging surface PCBs are ~5X higher than expected in RS 2 and RS 3 and
~3X higher than expected in RS 1. Commenters argue this increases the uncertainty as to whether
all remedial action objectives, including target PCB levels in fish, will be fully achieved.
Commenters also state that the ROD expected that the target cleanup levels for RS 2 and RS 3
would result in those river sections having post-dredging surface sediment PCB concentrations
comparable to those in RS 1.
Commenters indicated that it was an error to try to anticipate or estimate the rate of post-dredging
recovery in surface sediment concentrations based upon the rate of improvement before the
remedy, as there has been fundamental changes in the system due to source control before dredging
and sediment removal/backfilling as part of the dredging.
Commenters indicate surface sediment TPCB concentrations show a general improvement
between SSAP (2002 to 2005) and OM&M (2016) datasets. However, when the SSAP dataset is
separated into dredging and non-dredging area sample sets, lesser degrees of improvement in non-
dredging areas are indicated, with non-dredging areas in RS 1 showing very little or no
improvement.
Comment
95
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Several commenters stated that other studies indisputably show PCB concentrations in river
sediment are two to three times higher than estimated at the time the cleanup remedy was selected
and draw conclusions that additional work in the upper Hudson is essential.
Response
EPA acknowledges that sediment PCB data from the 2003 pre-design sampling program had
higher concentrations than average concentrations observed in the previous sampling program in
1998 and that the average concentrations were higher than modeled predictions. EPA is also aware
that these results inject a certain level of uncertainty into the remedial process, but it is important
to note that a very large percentage of the surface area in RS 1 was remediated and the change in
surface concentrations in dredged and un-dredged areas combined for RS 1 and RS 3 met
expectations. The change in surface concentrations in RS 2 was less than expected. EPA continues
to monitor all three river sections, including any areas that have higher than expected surface
concentrations. Relative change as opposed to absolute concentrations is the primary consideration
related to predicted change in PCB concentrations in biota. This relationship is fully embedded in
the basic physics and site conceptual model regarding the Hudson River and other contaminated
sediment sites nationally. EPA also recognizes that the time for PCBs in fish tissue to reach target
concentrations is a function of both the absolute post-dredging concentrations and the natural
recovery rate. However, because the negative short-term effects of the remedy may remain in
effect for some period of time, it will take some time for the near-term post-dredging PCB
concentrations in fish tissue to stabilize.
3.3.26 Comment 57: Analysis of sediment PCB data outside the dredge areas miscalculated
the concentration and mass located in these areas
Comment
Commenters indicated EPA has not yet made a quantitative assessment of the PCB mass remaining
in non-dredged areas as compared to previous estimates presented in the 2002 ROD. Commenters
stated this assessment was important in understanding long-term performance of the remedy.
Commenters presented reservations regarding how estimates of the mass of PCBs in non-dredged
areas of the Upper Hudson River were calculated. Specifically, commenters took issue with the
methodology used to estimate the mass in "unclassified" areas of RS 3, how core MPA values
were aggregated spatially to derive an areal estimate of MPA for each River Section or each
sediment type within RS 3, and the overall estimate of mass remaining. An alternative estimate of
the mass remaining outside the dredged areas was presented by GE.
Response
In consideration of the comments and alternative estimate received, EPA developed a modified
method for estimating the mass of PCBs in the "unclassified" sediment of RS 3, which utilizes the
cores collected in the "unclassified" sediment and improves upon the original methodology
presented in the FYR report.
96
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
EPA compared its methodology for estimating mass outside dredged areas with GE's
methodology, and has identified two important factors that result in a substantive difference
between GE's and EPA's estimates. Depending on whether a core recovery correction is applied
to sediment core data, EPA's estimates of total PCB inventory remaining can be between
approximately 10 percent and 50 percent more than GE's estimates {i.e., 50 percent more when
the core recovery correction is applied, 10 percent more when it is not). The issue of including or
omitting recovery correction of core segment lengths highlights some of the uncertainty in
estimating the mass of PCBs remaining in the Upper Hudson River. Second, the use of different
spatial aggregation techniques for MPA values also introduces some uncertainty and results in
differences between estimates of mass remaining. EPA's analysis indicated that the remaining 10
percent difference between EPA's and GE's estimates {i.e., when a core recovery correction is not
applied to EPA's estimate) largely arose from the differences in spatial aggregation. It may not be
possible to definitively conclude that one spatial aggregation method is preferable to another.
Thus, the Final FYR has been revised from the Proposed FYR to present the likely range of PCB
inventory remaining. Based on EPA's analysis, 40,000 kg of PCB inventory is likely a best
estimate while 60,000 kg of PCB inventory is likely an upper bound estimate on PCB inventory
remaining in the Upper Hudson River.
3.3.27 Comment 60: Data incompatibilities Lead to Errors in Interpretations
Comment
Various commenters identified several challenges with data, including conflation of natural
recovery and source control efforts prior to about 1995, variation in sampling equipment and
analytical methods, changes in sample preparation and handling techniques for fish, and temporal
changes in lipid content that may be conflated with temporal changes in PCBs in several species
and location combinations. One commenter argued that EPA is making a fundamental error -
assuming that all of the changes in sediment PCB concentrations are the result of natural recovery.
The commenter also pointed out that without taking the impact of source control into account, all
of EPA's estimates of rates of natural recovery represent overestimations and upper bounds;
recovery rates could be no higher, but the recovery due to natural processes are very likely much
less. Another commenter reinforced this idea stating that the MNA period includes major source
control. It was also noted that the rate of post-remedial recovery could not be estimated from the
recovery rate estimated prior to the remedy. One commenter further suggested that the natural
recovery rate prior to the remedy is known to be 1.3 percent in sediment based apparently on a
partial reading of Field et al. (2016). Other commenters suggested that apparently larger amounts
of PCB mass in the river than anticipated will necessarily delay fish tissue concentrations reaching
targets within time frames anticipated in the ROD.
Response
EPA agrees that the variety of historical sediment PCB data present challenges for developing
empirically-based estimates of natural recovery rates prior to and after completion of active
remediation in the Upper Hudson River. Many of the challenges identified by commenters were
also pointed out by EPA in the FYR report and its appendices. In particular, EPA went to great
lengths to analyze data in multiple ways to reduce adverse effects of several of the factors
97
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
identified by commenters, and also discussed uncertainties in the recovery rate estimates
presented. Some of the challenges are further discussed below; however, it is important to
emphasize that these challenges are not unique to the Hudson River. In, fact some may be the
inadvertent, but inevitable, consequence of the development of data quality objectives at different
stages of the remedial investigation process. In the early phases of an investigation, analytical
methods are selected and sampling programs are largely designed to identify worst case situations
with limited data collection. These efforts are largely forensic and are not designed for comparison
with future data sets to understand mechanisms of recovery. As sites evolve from forensics to site
characterization and feasibility analysis, objectives evolve and existing data are invariably added
to in efforts to fill spatial and temporal gaps. As risk evaluations evolve, some early actions for
source control are initiated, and investigatory data begins to be replaced with more representative
sampling efforts as time progresses toward remedial design. At this point in time the need for
understanding of current and future recovery rates increases and mechanistic models and empirical
evaluations are embarked upon. Unfortunately, when interest turns to understanding recovery
rates, the available data have been developed for a variety of differing objectives presenting the
kinds of difficulties identified by commenters and discussed at length by EPA.
The commenter' s statement that the recovery rate in Upper Hudson River sediment was 1.3 percent
during the MNA period appears to be a reference to Field, et al. (2016), who reported such a rate.
EPA disagrees with the reviewer that the recovery rate is known. In the FYR report, EPA discusses
several issues that complicate efforts to reliably estimate recovery rates including differing spatial
layouts of samples and sediment collection equipment. Field et al. also pointed out that lack of
unbiased estimates of mean surface PCB concentration at multiple points in time limited potential
to reliably estimate the natural recovery rate. This is consistent with how EPA has discussed
empirical estimates of the recovery rate in sediment, which support the conclusion that further
monitoring is needed to understand post-dredging recovery rates and expected time for
contaminated media to reach targets. It should also be noted that this issue is not unique to the
Hudson River PCBs Superfund site. At many sites nationally, sampling efforts are focused more
on characterization of nature and extent of contamination in sediment as opposed to estimation of
recovery rates. Generally, monitoring of fish tissue and water are relied upon more heavily for
understanding recovery rates.
The anticipated recovery rate in fish water and sediment was expected to be approximately 8
percent per year, although little is known about post-dredging recovery rates because the scale of
the remedial action is nearly unique. There are few examples of remedial actions of the magnitude
of the Hudson River project where sufficient time has passed since remediation to develop a robust
understanding of how sediment contaminant concentrations recover after such an action. Because
the post-dredging recovery is not fully understood, it is difficult to make definitive predictions of
the time to reach specific recovery goals. To rectify this, the remedy was designed and constructed
with the understanding that the ultimate risk-based goals would be reached over a period of time
and that, after completion of the dredging, remedy effectiveness (i.e., risk reduction) would be
evaluated through long term monitoring. Through long term monitoring, additional data will be
generated which will allow rigorous estimation of recovery rates and, as these data are developed,
understanding of remedial effectiveness will be refined. With this refinement, EPA will be able to
determine if and when additional investigation may be needed.
98
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
EPA undertook an in-depth evaluation to understand recovery rates based on fish, water and
sediment data prior to implementation of the remedy. With these rigorous efforts to account for
the factors identified by commenters, recovery rate estimates generally span a relatively wide
range of values, but the anticipated rate of 8 percent per year is within the range of uncertainties
of these estimates. EPA disagrees with commenters' assertion that existing data are adequate to
conclude that the remedy has failed and believes that it would be premature to change course with
the level of uncertainty around post-remedial recovery rates. EPA agrees with the commenter's
suggestions that monitoring methods and operating procedures need to be decided and
standardized throughout the monitoring period in order to minimize the uncertainties that have
complicated estimating rates in the pre-remedial period.
Regarding unanticipated amounts of PCB mass in the dredge areas, it cannot be assumed that
additional PCB mass is necessarily an indicator of higher exposures for biota in the post-dredging
period. Generally fish tissue concentrations are proportional to surface sediment PCB
concentrations as opposed to PCB mass as the reviewer suggests. EPA disagrees with the
commenter because the dredging portion of the remedy achieved the approximate percentage
reductions in average surface sediment PCB concentrations, the primary driver of percentage
change in water and tissue PCBs.
3.4 Remedy
This section includes comments and responses on topics such as requests for more dredging,
modifications to the remedy, attainment of the targets and goals, and the time for recovery of the
river.
3.4.1 Comment 10: EPA must address whether the targets for improvements in water
quality have or will be met
Comment
One commenter states that available post-dredging data show that the improvement in water
column PCB concentrations diminishes downstream of Thompson Island Dam (TID) and that,
because of limited data, it is unclear whether the ROD targets for PCB mass transport reductions
will be achieved within anticipated timeframes. Another commenter indicated that the 2016 water
column PCB data and EPA's estimated 10 percent per year recovery rate suggest that the
freshwater aquatic life criterion of 14 ng/L will be met sooner than originally estimated by EPA.
Response
Appendix 1 of the FYR report states that EPA expects the water quality criterion for aquatic life
to be met consistently within several decades. Given the short time period since the end of
dredging, there is limited water column data available to determine water column concentration
trends at this time. Also, the effects from future flows and upstream loads provide some uncertainty
in terms of future water column concentrations. Therefore, it is not possible to project an expected
date of compliance with this criterion with a high degree of precision using currently available
post-dredging trends. Uncertainty in the time to meet the aquatic life standard will be reduced by
the continuing collection of post-dredging water column data to support the development of post-
99
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
dredging water column decay rate estimates. It is worth noting that during 2016 water column
concentrations were generally below the 14 ng/L threshold. However, meeting this criterion is also
challenging due to higher water column concentrations during high flow events in the river.
EPA will continue to evaluate post dredging PCB transport reductions from the upper to the lower
river. This evaluation will include consideration of PCB transport during high flow events in the
river. There have been minimal high flow events since dredging ended, so additional data
collection is needed to inform estimation of the impact of these events on future PCB loading
trends.
3.4.2 Comment 22: EPA should track the attainment of the interim fish tissue targets of 0.4
mg/kg and 0.2 mg/kg PCB as it assesses the success of the remedy
Comment
EPA appears to be abandoning the ROD's interim remedial targets for fish PCB concentrations
that formed the basis for justifying the dredging remedy, and in doing so is arbitrarily ignoring
critical questions A & B in its own FYR guidance. In the FYR report, EPA is now stating that the
remedy will not be protective until the ultimate remedial goal of 0.05 parts per million PCB in fish
is reached. EPA should take the actions necessary to ensure that the remedy rapidly achieves the
interim targets identified in the ROD - specifically, achieving the first interim target (0.4 mg/kg
PCB in average fish concentrations) within five years after dredging, and the second interim target
(0.2 mg/kg) in sixteen years.
Response
Remedy protectiveness was evaluated in the ROD by comparing predicted fish tissue
concentration trajectories over time under different remedial alternatives. As noted in the ROD,
different target levels will be achieved at different times depending on the species and river section
(or river pool) given species-specific foraging strategies and life histories. Interim target levels
and the final remedial goal were developed for the ROD and continue to be evaluated, as
mentioned in the FYR report and its appendices. The first interim target level (0.4 mg/kg) has been
achieved in about 30 percent of the long-term Lower Hudson species-location combinations (the
species collected at a particular monitoring stations e.g., black bass collected at Catskill), based on
the average TPCBhe levels from 2009 to 2016 (Table 22-la Summary column). In addition,
average TPCBhe levels in yellow perch in the Lower Hudson were near or below the 0.2 mg/kg
interim target from 2009 to 2016 and, with a TPCBhe of 0.053 mg/kg, essentially achieved the
0.05 mg/kg remediation goal in the Albany area in 2016 (Table 22-la). In the Upper Hudson,
primarily in RS 3, average TPCBhe levels in yellow perch were near or below the 0.4 mg/kg
threshold in 2016 (Table 22-2a). Concentrations for other species are farther away from the interim
targets, as noted in Tables 22-la and 22-2a attached to this response. The above results are based
on TPCBhe concentrations. The TPCBArocior results are essentially the same and are provided in
Tables 22-lb and 22-2b.
EPA did not focus its FYR on attainment of the interim target levels. As noted in the FYR report
(Section 5.1), the post-dredging data are too limited to confirm attainment, and too little time has
100
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
elapsed since the dredging was completed. EPA anticipates that as many as eight or more years of
post-dredging fish tissue data will be needed to establish statistically relevant rates of decline in
post-dredging fish tissue PCB levels. Therefore, the FYR primarily focused on the documented
achievements of the remedy, such as PCB mass removed, reduction in surface sediment
concentration and control of the PCB loads to the Lower Hudson. Given that the FYR occurred so
close to the completion of dredging, it is not yet possible to accurately assess the long-term
improvements in fish tissue.
In the ROD, EPA stated that the remediation goal for protection of human health with regard to
fish consumption was attainment of 0.05 mg/kg in fish fillet (species-weighted average
concentration) throughout the Hudson, but primarily in fish of the Upper Hudson. As indicated in
the ROD, the interim target levels of 0.4 and 0.2 mg/kg are not remediation goals but interim
targets to be achieved along the way to the final remediation goal, the achievement of which could
be used by the State as a basis to reevaluate the fish advisories and potentially relax some fishing
restrictions. It is true that modeling conducted by EPA and discussed in the ROD projected that
the 0.4 and 0.2 mg/kg interim targets would be achieved within 5 and 16 years, respectively.
However, actual conditions during dredging did not, and were not expected to, match up in every
way with conditions as understood when the ROD modeling was conducted. Therefore, direct
comparisons of observed fish tissue concentrations to ROD forecasts need to be carefully
considered. It should also be noted that dredging started later than the model considered. Also
short-term and localized increases and subsequent decreases in fish tissue PCB concentrations
were anticipated in the FS and ROD (and observed between 2009 and 2016) were not directly
reflected in the long-term fish tissue forecasts presented in support of remedy selection. For these
reasons, direct comparisons of observed data to ROD forecasts need to be done carefully with the
various factors taken into consideration.
EPA will continue to use the interim targets to track progress toward the remediation goal. At this
time, EPA does not have sufficient data to determine if the interim targets will be achieved within
EPA's expectations. As stated above, as EPA obtains more years of fish data, the Agency will be
better able to assess progress toward the interim targets and the final remediation goal.
101
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Table 22-la
Comparison of Lower Hudson PCB Concentrations in Fish With Interim Targets and Remedial Goals
Total PCB - Homologue Equivalent Basis
2009-2015
Post-dredging 2016
Summary*
No. of
Average
Median (mg/kg-
No. of
Average Median (mg/kg
Average Median (mg/kg-
RM Species
samples
(mg/kg-ww)
ww)
samples
(mg/kg-ww)
ww)
(mg/kg-ww) ww)
152 Pumpkinseed
125
0.8
0.8
20
0.5
0.4
Smallmouth Bass
26
1.8
2.1
19
1.4
1.5
Spottail Shiner
46
1.0
0.8
9
0.2
0.3
Striped Bass
279
2.0
1.4
20
0.3
0.2
White Perch
26
0.9
0.9
12
1.2
1.1
Yellow Perch
14
0.3
0.2
8
0.05
0.03
White Catfish
6
2.5
1.9
Channel Catfish
134
2.7
2.6
20
2.0
1.9
113 Brown Bullhead
84
0.3
0.3
19
0.4
0.4
Largemouth Bass
14
0.9
0.7
Pumpkinseed
49
0.6
0.6
Smallmouth Bass
64
0.7
0.6
Striped Bass
145
0.5
0.3
20
0.61
0.71
White Perch
30
0.6
0.5
Yellow Perch
30
0.3
0.2
Channel Catfish
20
2.5
1.7
90 Brown Bullhead
22
0.6
0.5
Largemouth Bass
5
0.3
0.3
Pumpkinseed
39
1.1
1.0
Smallmouth Bass
15
0.3
0.3
Striped Bass
124
0.6
0.4
White Perch
20
1.2
1.0
Yellow Perch
20
0.2
0.2
Channel Catfish
25
3.0
2.2
50 Striped Bass
123
0.5
0.4
No. of RM-species pairs
25
25
9
9
25 25
Between 0.4 and 0.2 mg/kg-ww
6
9
3
2
7 8
Between 0.2 and 0.05 mg/kg-ww
0
0
1
1
1 1
Below 0.05 mg/kg-ww
0
0
0
1
0 1
Notes: Value less than 0.4 mg/kg interim target but greater than 0.2 mg/kg interim target
Value less than 0.2 mg/kg interim target but greater than 0.05 mg/kg remedial goal
Value less than 0.05 mg/kg remedial goal
Sportfish shown in bold.
* Column integrates the most recent data for each RM-species pair.
White catfish and channel catfish were not presented in the Appendix 3.
102
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
Table 22-lb
Comparison of Lower Hudson PCB Concentrations in Fish With Interim Targets and Remedial Goals
Total PCB - Aroclor Basis
2009-2015
Post-dredging 2016
Summary*
No. of
Average
Median (mg/kg
No. of
Average Median (mg/kg
Average Median (mg/kg-
RM Species
samples
(mg/kg-ww)
ww)
samples
(mg/kg-ww) ww)
(mg/kg-ww) ww)
152 Pumpkinseed
125
1.0
1.0
20
0.5
0.5
Smallmouth Bass
26
2.3
2.6
19
1.7
1.7
Spottail Shiner
46
1.3
1.1
9
0.3
0.3
Striped Bass
279
1.8
1.3
20
0.3
0.2
White Perch
26
1.2
1.1
12
1.5
1.3
Yellow Perch
14
0.4
0.3
8
0.1
0.0
White Catfish
6
3.1
2.4
Channel Catfish
134
3.4
3.3
20
2.3
2.2
113 Brown Bullhead
84
0.3
0.3
19
0.4
0.4
Largemouth Bass
14
1.0
0.7
Pumpkinseed
49
0.5
0.5
Smallmouth Bass
64
0.7
0.6
Striped Bass
145
0.4
0.3
20
0.7
0.8
White Perch
30
0.5
0.4
Yellow Perch
30
0.3
0.2
Channel Catfish
20
2.1
1.5
90 Brown Bullhead
22
0.5
0.4
Largemouth Bass
5
0.3
0.2
Pumpkinseed
39
0.9
0.9
Smallmouth Bass
15
0.3
0.3
Striped Bass
124
0.5
0.3
White Perch
20
1.0
0.8
Yellow Perch
20
0.2
0.2
Channel Catfish
25
2.6
1.9
50 Striped Bass
123
0.5
0.3
No. of RM-species pairs
25
25
9
9
25 25
Between 0.4 and 0.2 mg/kg-ww
5
8
2
2
5 7
Between 0.2 and 0.05 mg/kg-ww
1
2
1
0
2 2
Below 0.05 mg/kg-ww
0
0
0
1
0 1
Notes: Value less than 0.4 mg/kg interim target but greater than 0.2 mg/kg interim target
Value less than 0.2 mg/kg interim target but greater than 0.05 mg/kg remedial goal
Value less than 0.05 mg/kg remedial goal
Sportfish shown in bold.
* Column integrates the most recent data for each RM-species pair.
White catfish and channel catfish were not presented in the Appendix 3.
103
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
Table 22-2a
Comparison of Upper Hudson PCB Concentrations in Fish With
Interim Targets and Remedial Goals
Total PCB - Homologue Equivalent Basis
Post-dredging 2016
No. of
Average
Median
Area
Species
samples
(mg/kg-ww)
(mg/kg-ww)
RSI
Brown Bullhead
30
1.4
0.9
Largemouth Bass
16
1.0
0.5
Pumpkinseed
30
1.0
0.8
Smallmouth Bass
14
1.7
1.0
Spottail Shiner
7
1.2
1.0
Yellow Perch
30
0.4
0.3
RS2
Brown Bullhead
24
1.7
1.6
Largemouth Bass
10
0.6
0.6
Pumpkinseed
25
2.6
2.0
Smallmouth Bass
15
2.5
2.2
Spottail Shiner
5
3.1
3.3
Yellow Perch
25
0.5
0.4
RS3
Brown Bullhead
29
1.0
0.9
Largemouth Bass
26
1.2
1.0
Pumpkinseed
30
1.0
1.0
Smallmouth Bass
4
2.0
1.7
Spottail Shiner
5
2.4
2.4
Yellow Perch
30
0.3
0.3
No. of Area-Species pairs
18
18
Between 0.4 and 0.2 mg/kg-ww
1
2
Between 0.2 and 0.05 mg/kg-ww
0
0
Below 0.05 mg/kg-ww
0
0
Notes:
Value less than 0.4 mg/kg interim target but greater than 0.2 mg/kg interim target
Value less than 0.2 mg/kg interim target but greater than 0.05 mg/kg remedial goal
Value less than 0.05 mg/kg remedial goal
Sportfish shown in bold.
104
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfimd Site April 2019
-------
3.4.3 Comment 33: Habitat reconstruction did not achieve the project objectives.
Comment
Several commenters, including the NYSDEC, indicated that EPA has not required GE to perform
enough habitat reconstruction to allow for the work to achieve the habitat reconstruction goals.
Comments state that habitat reconstruction has not resulted in repopulation of species within the
parameters that the ROD anticipated. Other commenters stated that habitat reconstruction work
is not relevant to the protectiveness determination in the FYR.
Response
The success of the habitat reconstruction work is relevant to Question A of the FYR (i.e., whether
the remedy is functioning as intended by the ROD). The backfill and cap materials play dual roles
in isolating residual contamination (and therefore reducing exposures of fish to PCBs) and as an
integral habitat reconstruction component. The 2002 ROD indicates that the habitat reconstruction
program was anticipated to include the following dimensions:
• Backfill of dredged areas with approximately one foot of clean material to isolate
residual PCB contamination and to expedite habitat recovery, where appropriate;
• Various measures to address the anticipated short-term impacts to floodplains,
wetlands, and SAV communities (including minimizing impacts to wetlands,
controlling resuspension, stabilizing shorelines, and reconstructing habitats impacted
by implementation in an adaptive management context); and
• Monitoring the restoration of aquatic vegetation until benchmark followed by success
criteria have been achieved.
Remedial activities were anticipated to result in short-term temporary impacts to aquatic and
wildlife habitat of the Upper Hudson River (UHR). As discussed in Appendix A to the ROD
(Statement of Findings on Floodplains and Wetlands) implementation was anticipated to "remove
considerably more material from the river bottom than it [would] place as fill." For these reasons
and where appropriate, habitat replacement/backfilling measures were implemented, and
monitoring programs have been established to verify the attainment of the habitat replacement
objectives.
As discussed in Section 5.1.1.2.3 (Habitat Reconstruction) of the FYR report, and consistent with
the 2002 ROD, project habitat reconstruction activities began and were implemented in an
adaptive management framework to replace SAV communities, wetlands, and to stabilize river
bank habitat and shorelines. These activities have included:
• Backfilling of dredged areas with approximately 1.4 million cubic yards of backfill
and cap materials, including approximately 1 foot of clean backfill material to isolate
residual PCB contamination and support re-establishment of designated habitats;
105
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
• The installation of approximately 1.5 million individual riverine fringing wetland and
SAV plants (of which approximately 65 percent were locally harvested) and
approximately 1,700 pounds of seed mixes over approximately 29 acres of wetland
and 39 acres of submerged aquatic habitat reconstruction areas to help expedite
habitat recovery;
• The installation of approximately 13.5 miles of various shoreline stabilization
measures; and
• A monitoring program to facilitate implementation of habitat reconstruction and
shoreline stabilization during and following construction.
EPA does not agree that habitat reconstruction is not achieving project objectives. While
remediation goals specific to vegetation replacement requirements may not have been detailed in
the ROD, the placement of fill and backfill materials to isolate residual contamination and/or serve
as an attenuating layer (including layers to prevent bioturbation or inhibit other disturbance to cap
materials and to serve as a clean habitat for the benthic organism repopulation) was anticipated
and discussed. Such measures were designed, adapted to accommodate river bottom and operation
considerations, and implemented during construction. Where appropriate during the remedial
action, backfill and caps (including habitat backfill and caps topped with additional backfill layers)
were installed in accordance with project requirements and to performance standards. Details
regarding specific backfill and cap installations or habitat reconstruction areas can be found in the
CU Form 2 and Form 3 packages that were submitted by GE and reviewed by EPA. Monitoring
of these caps, backfill surfaces, and shorelines continues under the OM&M program.
Furthermore, EPA disagrees with other reviewers who commented that habitat reconstruction has
not resulted in repopulation of species within the parameters that the ROD anticipated. As
discussed in the ROD, habitat reconstruction to reduce impacts to wetlands and SAV communities
was designed and implemented to reflect pre-dredge and existing wetland and submerged aquatic
vegetation communities. Specifically, and as outlined in the 2003 Habitat Delineation and
Assessment Work Plan, the primary goal of the habitat reconstruction program is to replace the
functions of the habitats of the UHR to within the range of functions found in similar physical
settings in the UHR. Plant species installed as live plants and seed mixes were based on extensive
pre-dredge vegetation monitoring data collected between 2003 and 2008. Monitoring of
reconstructed habitats, as described in the Phase 1 and Phase 2 Adaptive Management Plans and
annual Operation, Maintenance, and Monitoring Plans and implemented through the OM&M
program is on-going and is currently in the benchmark monitoring phase. This benchmark phase
of monitoring includes evaluating individual reconstruction areas using quantitative but non-
destructive (not harvesting) measures. The purpose of benchmark monitoring is to help areas get
on trajectory to success by measuring their progress and evaluating the need for potential response
actions. Benchmark monitoring can last for up to 6 or more years. Response activities such as
replanting, reseeding, removal of loose coir fabric, and invasive species control have been
implemented in past years and are also planned for 2019.
The next phase of habitat monitoring is the Success Criteria phase in which reconstructed SAV
and wetlands are grouped by reach or an alternate spatial scale. These groups of reconstructed
SAV and wetlands are then assessed using quantitative comparisons to reference areas to
106
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
determine if the habitat reconstruction areas have been successfully re-established. This Success
Criteria phase of monitoring is anticipated to last for approximately another 2 to 5 years beyond
the benchmark phase. Because reconstruction areas are grouped together for this phase of the
evaluation, many individual areas may be under observation for an extended period of time (7 to
10 years). The initial habitat reconstruction effort has resulted in the installation of the species
and quantities called for in the designs and in the areas dredged. Habitat survey results regarding
species composition and overall coverage are encouraging. However, it is too early in the
monitoring process to determine whether or not the overall project habitat reconstruction goals
have been met. Monitoring and adaptive management will continue under the OM&M program.
EPA will continue to coordinate with NYSDEC regarding restoration activities.
3.4.4 Comment 38: EPA should compare data to ROD forecast regardless of
implementation
Comment
Commenters stated that EPA has rejected attempts to compare post-dredging data to 2002 ROD
forecasts because of the operational changes during dredging.
Response
EPA does not reject such comparisons and has made explicit comparisons in Appendix 1 and
Appendix 3 of the FYR report between ROD expectations based on EPA's models and post-
dredging data for water-column and fish tissue PCBs. EPA does cite differences between
anticipated and actual dredging operations in interpreting those comparisons: fish-tissue data from
the period immediately after dredging reflect a transition from conditions experienced during
dredging, which differed from anticipated conditions as described in Appendix 8 of the FYR
report. Model-data comparisons for fish tissue are presented in Appendix 3 of the FYR report
(Figure A3-19). Figure A3-19 shows a comparison between species-weighted model results
(Model Mean) for the selected remedy for the year 2010 (projected in the ROD to be the first post-
dredging year) and observed monitoring data for the actual first post-dredging year (2016) (Data
Mean). These results show that the model-anticipated concentrations in the first year post-dredging
are similar to those observed. As stated in Appendix 3 of the FYR report, ongoing post-dredging
monitoring over as many as eight or more years is needed to draw a scientifically reliable
conclusion. Nonetheless, these early data are encouraging and, when compared to model
predictions, indicate that the model has performed as expected.
3.4.5 Comment 42: The comprehensive sediment sampling data from the SSAP should be
treated as the baseline for evaluating recovery of PCB-contaminated cohesive
sediment in non-dredged areas
Comment
A number of comments were provided regarding the comparison of various surface sediment
datasets in Appendix 4 of the FYR report. Specifically, commenters stated that comparing the
2002 to 2005 SSAP data with the 2011 to 2013 Downstream Deposition Study (DDS) and 2016
107
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
OM&M dataset was not appropriate, as the DDS sampling did not target highly contaminated areas
and thus, had very limited data collected from the highly contaminated cohesive sediments
surrounding the dredge areas, and DDS sampling only sampled the top 2 inches and not the top 12
inches of surface sediment used to define dredge areas. Commenters also stated that the surface
sediment PCB concentrations for cohesive sediment in RS 2 and RS 3 estimated from the DDS
sediment survey and 2016 sediment monitoring survey should be considered to be biased low.
Further, commenters did not agree with how surface sediment texture was classified in RS 3. A
commenter stated that the predictive model used to classify sediment texture in 2016 OM&M
samples in RS 3 incorrectly categorized cohesive sediments. They commented that of the cohesive
sediments in RS 3 that were identified by the model, only approximately 1/3 of samples contained
at least 25 percent fine-grained sediments and more than 20 percent of the samples were described
by field samplers as "coarse" or "rock". Finally, they commented that the congener-based M1668
produced significantly higher PCB concentrations than Aroclor-based PCB analysis method
(M8082).
Response
EPA disagrees with the commenter's assertion that the SSAP dataset should not be compared to
the DDS sampling program. First, the evaluation carried out in Appendix 4 of the FYR report
with SSAP and DDS (and OM&M) samples only included the 0 to 2-inch sample depth interval,
to avoid comparing PCB concentrations at different depths. Second, while the DDS program did
not specifically target highly contaminated cohesive sediments surrounding dredging target areas,
in RS 2 the median PCB concentration of the SSAP samples targeted for re-occupation by the
DDS cores was significantly higher than the overall median PCB concentration of the SSAP
samples. In RS 3, the median PCB concentration of the SSAP locations targeted for re-occupation
by the DDS samples was not statistically different than the median PCB concentration of all SSAP
cores collected within RS 3. Thus, the commenter's claim that DDS samples did not target highly
contaminated areas, and thus are not comparable to SSAP samples, is unfounded.
The 2016 OM&M sampling program was designed in an unbiased fashion in order to detect long-
term changes in surface sediment concentrations. Unlike the SSAP sampling program, the OM&M
does not target specific areas, such as locations in close vicinity to dredging target areas. EPA
acknowledges that the biased nature11 of the SSAP program, particularly in RS 3 where a large
number of samples were located in the vicinity of dredged areas, versus the unbiased design of the
OM&M sampling program, creates a challenge when comparing PCB concentrations between the
two datasets. However, the OM&M program is specifically designed to alleviate issues that arose
when attempting to compare historical sediment datasets collected within the Upper Hudson River.
Thus, in the future, the OM&M data will provide a comprehensive, "apples to apples" dataset that
will allow EPA to detect changes in surface sediment concentrations through time.
EPA disagrees that the 2011 to 2013 DDS estimates of RS 2 and RS 3 surface sediment PCB
concentrations in cohesive sediments are biased low. In RS 2 outside the dredged area, the 95%
lower confidence limit (LCL) of SSAP locations targeted for DDS re-occupation was greater than
11 EPA notes that the sampling design bias was purposefully implemented by GE under EPA's direction to identify
and delineate areas of contaminated sediment. As such the SSAP sampling design was not intended to provide
average sediment concentrations, as was the 2016 sampling program.
108
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
the 95% upper confidence limit (UCL) for the full set of RS 2 SSAP locations (Field, Kern,
Rosman, 2016, Figure 13). This indicates that, in fact, the targeted SSAP locations for DDS were
representative of relatively high SSAP concentrations. As such, it would be expected that
resampling of the targeted SSAP locations during the DDS program would produce lower PCB
concentrations as a result of the Central Tendency Theorem (commonly referred to as "regression
to the mean," or median in the case of a log-normally distributed dataset). However, the median
of the DDS locations did not simply regress back toward the median of the entire SSAP data set,
Rather, the DDS locations exhibited a 95% UCL that was lower than the 95% LCL for the entire
population of SSAP cores in RS 2. This indicates that the median of the DDS locations was less
than the median of the corresponding targeted SSAP locations (as would be expected), but, more
to the point, that the median PCB concentration of the DDS locations was significantly lower than
the median of the entire SSAP core dataset in RS 2. This clearly indicates that cohesive surface
sediment outside CUs exhibited improved conditions (i.e., reduction in PCB concentration)
between collection of the SSAP and DDS data.
In RS 3 outside the dredged area, the 95% UCL and LCL of PCB concentrations of the SSAP
locations that were targeted for re-occupation by DDS cores bracketed the 95% UCL and LCL of
PCB concentrations of all SSAP locations in RS 3, indicating that the targeted SSAP locations
were representative of the distribution of all SSAP locations in RS 3, implying a fair comparison
between the datasets (Field, Kern, Rosman, 2016, Figure 14). As with RS 2, the observation that
the 95% UCL of the DDS cores was lower than the 95% LCL of both the targeted SSAP locations
and all SSAP locations in RS 3 indicates that RS 3 cohesive surface sediments outside CUs also
exhibited improved conditions (i.e., reductions in PCB concentrations) between the times of
collection of the SSAP and DDS data.
Thus, EPA's analysis indicates that PCB concentrations of DDS samples collected in RS 2 and RS
3 should not be considered biased low. Instead, the lower median PCB concentration of DDS
samples compared with both the re-occupied SSAP cores and all SSAP cores in RS 2 and RS 3
indicates recovery of surface sediment.
With regard to classification of sediment type in RS 3 based on a predictive model, as was done
for Appendix 4 of the FYR report, we disagree that the use of the model adds uncertainty to the
classification of sediment texture. EPA was not able to reproduce the commenter's results
regarding the performance of the predictive model in RS 3. In RS 3, EPA identified 21 samples
from the 2016 OM&M surface sediment sampling program that were identified as cohesive by the
model. Of these 21 surface samples, 16 (73 percent) had greater than or equal to 25 percent fines
(defined as the sum of percent fine clay and percent fine silt). EPA's percentage (73 percent) is
substantially larger than the value of "about 1/3," as presented in the comment. Similarly, the
commenter asserted that more than 20 percent of the cohesive sediments identified by the model
were described by field samplers as "coarse" or "rock." EPA did not identify any cohesive
sediment samples collected in RS 3 during the 2016 OM&M sampling program that were
described by field samplers as "coarse" or "rock." Furthermore, during the 2016 OM&M sampling
program, a steel probing rod was extensively used by samplers to assess sediment texture prior to
sample collection. Sediment probing is a standard, acceptable technique that involves physically
penetrating the river bottom with a metal rod to assess the sediment texture. Of the 21 samples
identified as cohesive in RS 3, 19 were classified as "fine grained," while the remaining two
109
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
samples were classified as "transitional." Thus, EPA's analysis of the predictive model used in RS
3 indicates that it performed well when classifying sediment as either cohesive or non-cohesive,
and there is no indication that the model increased the uncertainty in sediment classification.
Finally, EPA is currently investigating differences between sediment PCB concentrations using
M1668 and M8082. EPA has evaluated the difference in PCB concentrations from the two
methods using the matched pairs of sediment samples analyzed by both methods in the 2017
NYSDEC sediment dataset. As the commenter indicated, the comparison indicates that Total PCB
concentrations derived from Ml668 measurements are approximately 55 percent higher than those
derived from M8082. Similarly, the Tri+ PCB concentrations from the sum of congeners (M1668)
are approximately 44 percent greater than those predicted from Aroclor data (M8082) using GE's
equation [Tri+ = 0.13*A1221 +0.89*(A1242+A1254)].12 M1668 provides a more robust basis
than M8082 and modified Green Bay Method (mGBM) to determine both Total PCB and Tri+
PCB concentrations. However, the spatially extensive records of sediment PCB concentrations
collected as part of the Remedial Design (SSAP data) were based on M8082. Therefore, to track
changes in surface sediment concentrations relative to the SSAP data, sediment samples will
continue to be analyzed via M8082. To provide an accurate basis for long-term monitoring and
future evaluations regarding river recovery, EPA will be conducting analyses of sediment by both
M8082 and specifically Ml668 as part of the ongoing monitoring program. The laboratory will
also be required to run reference standards to confirm analytical accuracy and provide a benchmark
for future monitoring work.
3.4.6 Comment 47: By leaving more PCBs than anticipated in portions of the Upper
Hudson River, the remedy as implemented may not achieve the targeted reductions
in water and fish PCB concentrations in the timeframes anticipated by EPA
Comment
Commenters indicated that the dredging left behind high levels of PCB contamination in the
sediment and that the remaining PCB inventory was larger than originally estimated in the 2002
ROD with the result that fish will recover at a slower rate than originally estimated. Concern was
expressed that without additional removal of "toxic hotspots" in the Upper Hudson River, there
will be a substantial delay in the recovery of the resource and a delay in reaching remedial action
objectives. Commenters also stated that the selected remedy as applied in RS 2 and RS 3 left
behind substantial PCB mass in the vicinity of dredged areas, creating a "donut" of PCB inventory
around dredged areas.
Response
The following important information is provided to address commenters concerns: 1) how are fish
exposed to PCBs in the Upper Hudson River; 2) distinguish between surface sediment PCB
concentrations (measured in mg/kg) and PCB inventory in the sediment (measured in kg of PCBs
in river sediments or in g/m2 of river bottom), and 3) distinguish between the absolute amount of
PCBs removed (measured in kg of PCBs removed from the river) and relative amount of PCBs
12 GE's equation (and similar forms of the equation) have been used to estimate Tri+ PCB concentrations from the
M8082 (Aroclor-based) results throughout the SSAP and DDS sampling programs.
110
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
removed (measured as a percentage of total PCBs in the river sediment). As explained below, it
is the PCBs in the surface sediments and water column of the Upper Hudson River that directly
drive the PCB concentrations found in fish tissue, and it is these PCB sources - and not the total
PCB inventory remaining - that are the best indicators of fish exposure to PCBs.
In the Upper Hudson River, the PCBs in fish tissue are driven by PCBs in the water column (both
dissolved PCBs and PCBs bound to suspended solids) and PCBs in the upper few inches of the
sediment bed. It is these two compartments of PCBs that directly affect long-term trends in fish
tissue PCB concentrations. In the 2002 ROD, these two PCB compartments were forecast using
the HUDTOX model, and were also used as inputs to the FISHRAND model, which produced
projections of PCB concentrations in fish tissue into the future. Analyses presented in Appendices
1 and 3 of the FYR report indicate that water column and fish tissue concentrations declined at
similar rates during the baseline monitoring period (1995 to 2008).13 Long-term measurements of
surface sediment data also indicate declining PCB concentrations. Post-dredging data collected in
2016 show further decline, although data will need to be collected over more OM&M program
cycles to establish long-term trends. The similarity in rates of decline between the three different
media (fish tissue, water and surface sediment) highlight the close linkage between them and
support the use of water and surface sediment measurements as direct indicators of the reduction
in fish exposure to PCBs over time.
In the Upper Hudson River, the surface sediment PCB concentration, for purposes of long-term
monitoring and direct fish exposure, is defined as the concentration of PCBs in the upper 2 inches
of the sediment, closest to the sediment-water interface. PCB inventory refers to the mass of PCBs
throughout the sediment bed and does not distinguish between PCBs in the surface sediment or
PCBs greater than 2 inches below the sediment-water interface. Thus, the inventory of PCBs in
the Upper Hudson River includes PCBs in the surface layer that fish are regularly exposed to as
well as PCBs that have little interaction with fish {i.e., PCBs that are below the surface sediment
layer). Because of the limited access to the deeper layers, it is not appropriate to link rates of
decline in fish tissue PCB concentrations with PCB inventory. PCB inventory does not directly
characterize the concentrations of PCBs to which fish are exposed. In particular, tracking PCB
inventory through time does not account for reductions in surface sediment concentrations due to
burial by cleaner sediments produced upstream of the Site. In this instance, while PCB
concentrations in the surface sediments would decline, there would be little impact on the
undisturbed PCB inventory at depth.
Instead, as discussed above, surface sediment concentrations and water column concentrations
should be the direct metrics used to assess the degree of PCB exposure to fish. For surface
sediment concentrations, the recent 2016 OM&M data show that the overall percent reductions
(remediation plus natural recovery) of PCBs containing three or more chlorines (Tri+ PCBs)
concentrations were estimated to be 96 percent, 88 percent and 80 percent for RS 1, RS 2 and RS
3, respectively. These percentage reductions are greater than were anticipated by the ROD. The
data also indicate that post-dredging average surface sediment concentrations of Tri+ PCBs are
near or below 1 mg/kg (see Tables A4-5 and A4-6 in Appendix 4 of the FYR report).14 Since the
13 Similar rates of decline are observed for later starting dates, e.g., 1996 to 2008 or 1998 to 2008.
14 The average concentration and level of reduction obtained from the 2016 EPA/GE data were also confirmed by the
2017 NYSDEC surface sediment survey.
Ill
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
surface sediment concentration is a direct driver of fish tissue concentration, the above evidence
{i.e., the large relative reduction in surface sediment PCB concentration as well as the current low
surface sediment concentration) suggests that it is likely that fish will also show a corresponding
decline in the near future.
EPA recognized that the PCB inventory may impact fish tissue PCB levels indirectly by supplying
PCBs to the surface sediment and water column PCB concentrations through sediment
resuspension during high flow events {i.e., flows in excess of 15,000 cfs), and set removal criteria
for sediment PCB inventory as well as for surface concentration. However, predictions from the
HUDTOX model indicated that the very high flow conditions associated with a 100-year peak
flow event would remove, on average, less than a 1-cm layer of sediment. The results from model
simulation are supported by the measurements taken during the spring 2011 high flow event when
actual flow rates exceeded the 100-year peak flow conditions used in the HUDTOX model
simulations. Data collected during and after the 2011 event did not find evidence of widespread
scour and transport of highly-contaminated PCB-bearing sediment within the Site.
When assessing the success of the remedy with regard to the removal of PCB inventory from the
Upper Hudson River, it is important to base the assessment on a comparison of 2002 ROD
estimates of the amount of PCB inventory that would be removed relative to the amount of PCB
inventory that EPA estimated in the ROD was present (measured as a percent reduction in PCB
inventory) vs. the actual percent reduction in PCB inventory as a result of dredging. The reason
for basing the effectiveness of PCB removal in part on the percent reduction in PCB inventory is
that any changes to the estimate of absolute inventory present prior to dredging {e.g., as a result of
sampling challenges and characterization of PCB concentrations in the sediment subsequent to the
release of the 2002 ROD) do not change the relative reductions in fish exposure that are needed to
achieve the remedial objectives for fish tissue.
In the 2002 ROD, the specific goals for the relative reduction in PCB inventory, measured as a
percentage of total PCBs in the river sediment in each river section prior to dredging, were: 80
percent of PCB mass removed in RS 1; 86 percent of PCB mass removed in RS 2; and 28 percent
of PCB removed in RS 3.15 As presented in Appendix 2 (Table A2-6b) of the FYR report, the
targeted percent reduction in PCB inventory was exceeded in RS 1 and RS 3, but not for RS 2 (the
2002 ROD stated 86 percent of PCBs were to be removed, but calculations in Appendix 2 indicate
that approximately 82 percent of PCB inventory was removed). However, the overall target
reduction for the Upper Hudson of 65 percent was exceeded by the actual removal, which achieved
76 percent. Given that PCB inventory is not the direct driver of fish tissue concentration and given
the overall reduction of 76 percent of the Upper Hudson inventory, it is unlikely that the additional
4 percent of the original inventory remaining in RS 2 will substantively impact the rate of recovery
in the river section. RS 2 will be regularly monitored under OM&M to quantify the actual rate of
recovery.
With regard to the assertion that the selected remedy left behind contaminated sediment around
dredged areas, EPA analyzed the 2002-2005 SSAP and 2016 GE/2017 NYSDEC sediment data in
15 From EPA, 2002. Responsiveness Summary to the Hudson River PCBs Site Record of Decision, Table 363334-1.
These percentages for inventory reduction should not be confused with anticipated reductions for surface sediment
and fish tissue, discussed elsewhere in these responses.
112
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
each river section to investigate whether "donuts" (rings) of elevated PCB inventory or surface
sediment concentrations were left around dredged areas. In the remedial design, EPA specifically
looked at two criteria for determining whether a location needed to be dredged: 1) a maximum top
12-inch interval Tri+ PCB concentration in excess of the dredging criteria of 10 mg/kg, 30 mg/kg
and 30 mg/kg Tri+ PCB for RS 1, RS 2 and RS 3 respectively, and 2) a Tri+ PCB MPA value in
excess of 3 g/m2, 10 g/m2 and 10 g/m2 for RS 1, RS 2 and RS3 respectively. As part of applying
these criteria, an adjustment was made where selected areas were allowed to remain undisturbed
when the PCB inventory was buried below 12 inches or more of low-concentration sediments (less
than 1 mg/kg).
EPA explored this concern by examining the variation of concentrations of Tri+ PCB in surface
sediment (0-2 inch) as a function of distance from the sampling location to the nearest dredged
area boundaries. The results were represented in Figures 47-la, 2a, and 3a for samples from the
SSAP program, and in Figures 47-lb, 2b, and 3b for samples from the 2016 GE and 2017
NYSDEC programs, respectively. Figure 47-4 illustrates how the distances were assigned to each
location using the 2016 and 2017 surface sediment data as an example.
In each of the figures, a blue line has been added, which represents a weighted least square fit to
the data. This line approximates the variation of the median of the data with distance from the
dredged area boundary. The curves were fit separately for the data inside the dredging boundary
and outside the dredging boundary. Hence the curves do not meet at the boundary itself (0 on the
horizontal axis). In reviewing these curves, it is apparent that there is little variation in the surface
sediment concentration as a function of distance from the dredged area boundary. That is, the
median concentration as approximated by the weighted curve is nearly flat in both halves of each
figure.
EPA further examined the variations of PCB inventory as a function of distance from the sampling
location to the nearest dredged area boundaries. The PCB inventory was represented by the
maximum Tri+ PCB concentration in the top 12 inches of sediment16 and MPA data from the
SSAP program. Figures 47-5, 47-6 and 47-7 show the results for maximum Tri+ PCB
concentration, and Figures 47-8, 47-9 and 47-10 show the results for MPA data. These figures also
display the threshold for removal for each river section as described above.
The figures illustrate that for cores outside dredged areas in all three river sections, a very limited
number of points were above the threshold for removal, confirming the successful selection of
locations according to the criteria. Nearly all above-threshold locations, as well as a large number
of below-threshold locations were included in the dredged areas. GE was not required to "chase"
isolated cores above the threshold, as this would likely have caused more sediment disturbance (a
negative impact on the river ecosystem) with little positive gain from the removal of the isolated
contaminated sediments. The orange dots shown in Figure 47-10 represent spatially isolated
locations that met the "Select" criterion of an inventory greater than 10 g/m2 underlying a
minimum of 12 inches of surface sediment less than 1 mg/kg Total PCB, which were permitted to
remain according to the ROD criteria.
16 As defined in the 2007 Dredge Area Delineation (DAD) Report (GE, 2007)
113
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
The weighted curves on these figures (blue lines) indicate that there is little variation in the PCB
inventory from the dredging boundary out to the maximum distance values on the plot (150 ft in
RS 1 and 300 ft in RS 2 and RS 3), indicating little gradient. These plots indicate that sediments
close to the dredging boundaries are not particularly more contaminated than those located further
away. Thus, these graphs indicate that finer-grained sediments close to the dredging boundaries
are similar in average concentration to finer-grained sediments elsewhere in the river section, and
a "donut" feature of high concentrations immediately proximate to the dredging boundaries as
argued by the commenter is not apparent. Fine-grained sediments have similar Tri+ PCB MPA
value and maximum top 12-inch interval Tri+ PCB concentrations throughout each river section.
In conclusion, the dredging activities achieved their stated goals in the 2002 ROD, when evaluated
using an appropriate metric (the amount of PCBs removed relative to the total amount of PCBs
present in the Upper Hudson River sediment prior to dredging). Further, the remaining PCB
inventory in sediments of the Upper Hudson River is not an appropriate metric to project future
rates of decline in fish tissue PCB concentrations, as PCB inventory does not quantify the amount
of PCBs that fish are exposed to, largely by failing to account for sediment burial. Instead, surface
sediment concentrations and water column concentrations provide more informative measures of
the amount of PCBs that fish are exposed to. These concentrations along with fish tissue
concentrations should form a basis for assessing rates of decline in fish tissue concentrations
moving forward (and ultimately the success of the remedial activities). Lastly, concerns that high
levels of contamination are found in the immediate vicinity of dredged areas is not borne out by
the SSAP data or by the 2016 OM&M and 2017 NYSDEC surface sediment data. (See also the
Response to Master Comment 40 (see Section 3.3.17), regarding the absence of "hot spots" post-
dredging.)
114
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
1000 I i i » i I i »» » I » i » i i i i i i I i i i i | i i i i I i i i i I i i i i 1
100
- ¦ ¦ ¦ «¦¦¦¦{
- 1 \ % ¦ " .'¦"•# "' ¦
10 mg/kg
i i i i i i i i i i i i i i i i i i i i i i I i i i i i i i i i i i i i i
-250 -200 -150 -100 -50 0 50 100 150
Inside Dredge Boundary Outside Dredge Boundary
Distance From Dredge Boundary (ft)
Legend
• 2002-2005 SSAP Data
Weighted-Average
— Dredging Threshold
L : Uiulstnx Mi l'!-
SSAP Surface Sediment Tri+ PCB Concentrations vs. Distance from Dredging Boundary
River Section 1
2002-2005 (0-2 in. Samples)
Figure 47-la
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
115
April 2019
-------
I I I I I I I I I I I I I I I I
I I I I I I I I
10 mg/kg
0.01 ! I I I | I I I I | I I I I | I I I I | I I I I I I I I I I I I I | I I I I
-250 -200 -150 -100 -50 0 50 100 150
Inside Dredge Boundary Outside Dredge Boundary
Distance From Dredge Boundary (ft)
Legend
• 2016 EPA/GE Data
• 2017 NYSDEC Data
Weighted-Average
Dredging Threshold
OM&M Surface Sediment Tri+ PCB Concentrations vs. Distance from Dredging Boundary
River Section 1
2016-2017(0-2 in. Samples)
; nrnCSm: Ritfr
Figure 47-lb
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
116
April 2019
-------
iooo 4—1—1—'—>—11—1—1—1—*-
100 -
0.1-=
0.01
_1—I—I—I—I—I—I—I—I—I—I—I—I—1_
¦ - - -4- - -
¦TWV' - »*'
1 W . t 1 » ¦ ¦
m\i • . » * . " ¦ ¦¦
¦ % ¦ : ¦¦ . •
30 mg/kg
Legend
• 2002-2005 SSAP Data
Weighted-Average
— Dredging Threshold
T 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ! 1 1 1 1 1 1 1-
-200 -100 0 1 00 200 300
Inside Dredge Boundary Outside Dredge Boundary
Distance From Dredge Boundary (ft)
Q Knrr
SSAP Surface Sediment Tri+ PCB Concentrations v.v. Distance from Dredging Boundary
River Section 2
2002-2005 (0-2 in. Samples)
Figure 47-2a
April 2019
Final Second Fix e-Year Review Comment Response for the Hudson River PCBs Superfund Site
117
April 2019
-------
1000-
100-
•3 10-1
M
~S)
E
CQ
O
Q-
+ 1-
0.1-=
0.01
J I l_l I l_l I I I I
_J I I I
- 30 mg/kg
/ .-J* •'
. ¦ ¦*¦
T 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1"
-200 -100 0 100 200 300
Inside Dredge Boundary Outside Dredge Boundary
Legend
• 2016 EPA/GE Data
• 2017 NYSDEC Data
Weighted-Average
Dredging Threshold
Distance From Dredge Boundary (ft)
OM&M Surface Sediment Tri+ PCB Concentrations v.v. Distance from Dredging Boundary
River Section 2
2016-2017 (0-2 in. Samples)
: i lliutim i Hlti-r
Figure 47-2b
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
118
April 2019
-------
-j goo -a-J 1 ' 1 1 1 1 1 1 1 1 ' 1 1 1 1 ' 1 1 1 1
30 mg/kg
Legend
• 2002-2005 SSAP Data
Weighted-Average
— Dredging Threshold
~i—i—|—i—i—i—i——i—i—i—i—|—i—i—i—i—|—i—i—i—r~
-100 0 100 200 300
Inside Dredge Boundary Outside Dredge Boundary
Distance From Dredge Boundary (ft)
; IhutsmTilKiivr
SSAP Surface Sediment Tri+ PCB Concentrations vs. Distance from Dredging Boundary
River Section 3
2002-2005 (0-2 in. Samples)
Figure 47-3a
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
119
April 2019
-------
100-
10^
D)
e
CO
O
Q.
1-
0.1-
0.01-
, J 1- 1 —1 1 1 1
1 J 1 1 L 1 I 1 1 I 1 1 1 L
:
" -
" " B "
- ¦ ¦ 1
-i-l 1 1 1 1 1 1
¦
¦ _ ¦"
"¦¦¦" • ¦ ¦ ¦" ¦ ,
¦ ¦¦
' V :
¦ " : 11
• _
• ¦ a
1 1 1 1 1 1 1 1 1 1 1 1 1 1
30 mg/kg
Legend
• 2016 EPA/GE Data
• 2017 NYSDEC Data
Weighted-Average
— Dredging Threshold
-100 o
Inside Dredge Boundary
100 200 300
Outside Dredge Boundary
Distance From Dredge Boundary (ft)
OM&M Surface Sediment Tri- PCB Concentrations v.s. Distance from Dredging Boundary
River Section 3
2016-2017 (0-2 in. Samples)
Ihuhnill Hi'l-r
Figure 47-3b
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
120
April 2019
-------
1000 - ' ' ' ' I ' ' ' ' I ' ' ' ' I ' ' ' ^ I ' ' ' ' i i i i I i i i i I i i i i
100-
. ¦ ¦ ¦ ¦ -¦ ^ •
I-• . . .
" ¦ ¦ ¦
¦ «v I . > ¦ ¦ ¦¦ *
O) 10-= J
CQ
o
CL
c 1-
I-
r. ¦ • • >v,
' . r ¦¦ • •. 5
10 mg/kg
¦:VT'¦: 'J
**P" ¦ ¦"
0.1-
¦ ¦
Legend
• 2002-2005 SSAP Data*
Weighted-Average
Dredging Threshold
i i i i l i i i i I i i i i
0.01 i i i i | i i i i | i i i i | i i i i | i i i i
-250 -200 -150 -100 -50 0 50 1 00 1 50
Inside Dredge Boundary Outside Dredge Boundary
Distance From Dredge Boundary (ft)
•Majority of points from the SSAP data set were
from 2002 to 2005, a small portion of the points
were obtained in 2007 as a part of the SEDC
sampling program.
7 IhutimrtM*
SSAP Maximum Sediment Tri+ PCB Concentration vs. Distance from Dredging Boundary
River Section 1
2002-2005 (Maximum value in 0-12 in. interval)
Figure 47-5
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
121
April 2019
-------
100 ^
10-
CQ
o
CL
1-.
0.1-
0.01
V'-ii
#w. . . ¦ \"V_ -hH
¦» . '%»»5 .*
. - v i - .w: - -r. i
" - at -
¦ ."a: yiflrV' ..
- 30 mg/kg
?v., /¦
5" ¦ .= • ' ¦ :
. ?Y
Legend
• 2002-2005 SSAP Data*
Weighted-Average
— Dredging Threshold
t—i—i—i—[—i—i—i—r-
200 -100 0
Inside Dredge Boundary
-i—i—i—[—i—i—i—i—]—i—i—i—i-
100 200 300
Outside Dredge Boundary
Distance From Dredge Boundary (ft)
*Majority of points from the SSAP data set were
from 2002 to 2005, a small portion of the points
were obtained in 2007 as a part of the SEDC
sampling program.
if) llttitsimi Km;
SSAP Maximum Sediment Tri+ PCB Concentration vs. Distance from Dredging Boundary
River Section 2
2002-2005 (Maximum value in 0-12 in. interval)
Figure 47-6
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
122
April 2019
-------
- 30 mg/kg
Legend
• 2002-2005 SSAP Data*
Weighted-Average
— Dredging Threshold
~]—i—i—i—r
-100 0
Inside Dredge Boundary
1—i—i—i—i—|—i—i—i—r-
100 200 300
Outside Dredge Boundary
Distance From Dredge Boundary (ft)
*Majority of points from the SSAP data set were
from 2002 to 2005, a small portion of the points
were obtained in 2007 as a part of the SEDC
sampling program.
i lliut'snfrillttrer
SSAP Maximum Sediment Tri+ PCB Concentration vs. Distance from Dredging Boundary
River Section 3
2002-2005 (Maximum value in 0-12 in. interval)
Figure 47-7
April 2019
Final Second Five-Year Review Comment Response for the Fludson River PCBs Superfund Site
123
April 2019
-------
1000
100 ^
¦6 io-
0.01
1 1 1 1 J_1 J-1 1 ,1 1 i 1 1 1 1 1 1 1 1 1 1 1 j
¦ s
- ¦ " - ¦ .
: >U § .S s.. .
I I l l 1 1 l I i 1 i 1 l_l
r r
k ¦ %
t_^i—^~ - -1-1.
mob ¦
if' ¦ ^ ^ ¦TV&t fF&g&'l
m m ¦ 1 |" i ¦ fV* n
- . i j " . V.S". *
. ¦ ¦ ¦ " t . ¦ "
¦' . •
• -V:
¦ ¦¦
~ ¦
i ¦ ¦ ¦
T I T 1 I 1 1 1 T 1 T 1 1 1 1 1 1 1 1 I rrf T
'¦:fc. " ¦ -
.«r : .
¦ ¦
i i i i i • *1 i i i i i i i
- 3 g/m2
Legend
• 2002-2005 SSAP Data*
Weighted-Average
— Dredging Threshold
-250 -200 -150 -100 -50
Inside Dredge Boundary
0 50 100 150
Outside Dredge Boundary
Distance From Dredge Boundary (ft)
~Majority of points from the SSAP data set were
from 2002 to 2005, a small portion of the points
were obtained in 2007 as a part of the SEDC
sampling program.
L) Hudson! Hilf:
SSAP Tri+ PCB MPA vs. Distance from Dredging Boundary
River Section 1
2002-2005
Figure 47-8
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
124
April 2019
-------
1000 -
100
_J I I 1 I 1 I I I I I I I I I I I I 1 I I I L_
¦¦¦
, i iff
¦ • ¦ W<4
0.-H
¦*. *•' ¦
r_." _ 10 g/m2
* \ i.
cAv-.
J,Ht L V _¦ ? mm
- . ¦.. _ ¦ • ¦
•aFo-*-:- J;— ¦*
¦ ¦ rn m * ¦ ~m ¦ _¦ .
i-S*. ¦ r, ¦
¦Mil ¦ ¦ . ¦
Legend
• 2002-2005 SSAP Data*
Weighted-Average
— Dredging Threshold
0.01 -j—i—i—i—i—|—i—i—i—i——i—i—i—i—|—i—i—i—i—|—i—i—f—i—
-200 -100 0 1 00 200 300
Inside Dredge Boundary Outside Dredge Boundary
Distance From Dredge Boundary (ft)
*Majority of points from the SSAP data set were
from 2002 to 2005, a small portion of the points
were obtained in 2007 as a part of the SEDC
sampling program.
Hlu/soFTilltll*
SSAP Tri+ PCB MPA vs. Distance from Dredging Boundary
River Section 2
2002-2005
Figure 47-9
April 2019
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
125
April 2019
-------
1000-q-1—1—1—'—I—'—<-
100 T
CM
E
3
<
CL
CO
o
CL
0.1-
0.01
J I I I I I
r 10 g/m2
Legend
• 2002-2007 SSAP Data1
• Meets Select Criteria2
Weighted-Average
Dredging Threshold
t—i—|—i—i—i—r
-100 0
Inside Dredge Boundary
-i—i ¦ r—i—|-
100 200 300
Outside Dredge Boundary
Distance From Dredge Boundary (ft)
1. Majority of points from the SSAP data set were from 2002 to 2005, a small portion of the points were obtained in 2007 as a part of the SEDC
sampling program.
2. The orange dots shown present spatially isolated locations that met the "Select" criterion of an inventory greater than 10 g/m2 underlying a
minimum of 12 inches of surface sediment less than 1 mg/kg Total PCB, which were permitted to remain according to the ROD criteria.
IhuisoirtliHit'r
SSAP Tri+ PCB MPA vs. Distance from Dredging Boundary
River Section 3
2002-2005
Figure 47-10
April 2019
126
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.4.7 Comment 48: The Lower Hudson River (LHR) fish recovery is not responding as
expected
Comment
Commenters raised a number of related issues concerning the rate of fish recovery in the LHR,
and EPA's conclusion that fish PCB levels in LHR fish are not strongly linked to Upper Hudson
River (UHR) loads and conditions. There were seven major points that are outlined and addressed
below.
1. A commenter stated that the FYR appears to disregard prior conclusions and modeling
results in the ROD (USEPA, 2002) that the UHR PCB load to the LHR is the primary
factor in the recovery of LHR fish. The FYR report cites slower recovery of LHR fish as
evidence that the UHR does not play an important role in the LHR and speculates about
"other sources." Based on high-resolution core sampling data and modeling (Thomann et
al. 1989, Farley et al. 1999, USEPA 2000a, Hydroqual 2007, Rodenburg and Ralston
2017), the primary source of PCBs to the LHR is the result of past and continued loading
of PCBs originating from the Hudson Falls and Fort Edward plant sites and sediments
within the UHR.
2. A commenter stated that the FYR Report indicates that the remedial work in the UHR
will have little or no beneficial impact in the LHR. The commenter notes that this is in
contrast to the ROD assumption that PCB loading from UHR to the LHR plays a major
role in recovery of the LHR. EPA appears to have rejected this major ROD assumption
with little technical basis provided in the FYR.
3. A commenter stated that the remedy in the UHR is not likely to have a significant impact
on fish in the LHR and says that EPA should not state that PCB sources other than GE's
discharges in the UHR are controlling LHR fish PCB concentrations unless the agency
has data to support such a conclusion.
4. A commenter stated that the identification of both fish tissue and sediments in the LHR
with significantly elevated PCB concentrations suggests that the remedial work in the
UHR is less likely to achieve the targeted reductions in PCB concentrations in the
estuarine portion of the river than anticipated by EPA in the ROD.
5. A commenter noted that the increase in water column PCB concentration due to dredging
was not reflected in a commensurate impact on the fish in the Hudson River, and that,
typically, only those fish in the immediate vicinity of the dredging work, or immediately
downstream, showed a significant reaction to the dredging. This indicated to the
commenter that the local sediments are much more important in controlling fish PCB
concentrations than impacts from upstream sources, which in the Hudson River primarily
means upstream sediments. This is most important for the LHR, where the fish showed
little to no response to the dredging work upstream, so that it can no longer be expected
that the remedial program in the UHR will result in significant improvement in fish PCB
concentrations south of Albany.
6. It is stated by a commenter that PCB levels in the 160-mile portion of the LHR have not
benefited much, if at all, from upriver dredging, and that contamination in fish at
127
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Poughkeepsie remains as high as it was before the dredging project. Below the Troy Dam
all the way to New York City, EPA's own studies show PCB concentrations in fish
haven't declined as expected as a result of the upriver dredging.
7. A commenter stated that EPA should recognize that there is much more work to be
accomplished to address the human health and ecological risk posed by the disposal of
PCBs in the Hudson River. EPA should do the work necessary to ensure that the remedy
in the UHR is protective, and to implement a full investigation and remedial program in
the LHR south of Troy.
Response
EPA recognized many of the concerns raised by the commenters, noting in the FYR report that
fish body burden decay rates declined with distance downstream of Albany, and that fish tissue
concentrations below RM 113 in the LHR did not respond to the increased loads at Waterford
during dredging. These observations, along with others listed below, are sufficient to support
EPA's assertion that PCB concentrations in LHR fish are not strongly linked to current loads
originating from the UHR. Specifically, while fish in the UHR and LHR fish at RM 152
(Albany/Troy), clearly responded to the increased water column loads and concentrations due to
dredging, below RM 152 fish tissue levels increased little, if at all, during dredging. Water column
concentrations at Poughkeepsie (RM 90) also did not respond to dredging as was observed
upstream, and instead gradually decreased slightly during the dredging period. In addition, water
column concentrations at Poughkeepsie during 2004-2008 were slightly higher than those
observed at Albany (See Figure Al-2 of Appendix 1 of the FYR report), a condition that is
incompatible with LHR conditions caused by UHR loadings alone. These multiple lines of
evidence indicate that LHR conditions, at least those at RM 113 and downstream, are not strongly
linked to current PCB loads and conditions of the UHR.
It is important to distinguish between past (pre-dredging and dredging) and current (post-dredging)
loadings from the UHR to the LHR. While the link between UHR loadings and LHR impacts has
reduced over time, GE sources in the UHR have been the primary source of PCBs in the LHR.
EPA notes that further studies are needed to better understand the extent of PCB contamination
in the sediments of the LHR. As discussed below, EPA is relying on the investigative sediment
work for the LHR that was conducted prior to the ROD
EPA agrees with the commenter's assertion that the weakness of the current link does not
necessarily mean that external downstream loads to the LHR have suddenly grown to greater
importance. Rather, it is most likely that an extensive inventory of PCB contamination in the
sediments of the LHR is primarily responsible for LHR PCB levels in both fish and water. This
inventory is derived primarily from historical GE discharges and UHR loads to the LHR. As noted
by the commenter, the magnitude of the historical GE loads was examined as part of the
investigations that led to the ROD, primarily using dated sediment cores, and more recent work
continues to support the ROD conclusions in this regard. While LHR external sources may have
increased in relative importance, EPA believes that the majority of the LHR PCB inventory (and
by inference, the majority of PCB exposure) can be attributed to GE-related PCBs originating from
the UHR, extending to approximately RM 50 or further downstream. EPA is currently evaluating
what additional studies need to be completed for the LHR.
128
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
The ROD states as follows with regard to the impacts of UHR remediation on the LHR:
...the reduced PCB load over the Federal Dam projected by the selected remedy will
ultimately result in reduced concentrations of PCBs in fish, sediment and water. This in
turn will result in reduced risks to humans and ecological receptors living in and near the
Lower Hudson River from PCB contamination originating in the Upper Hudson River.
(EPA, 2002; p. 2)
While this statement is still true, the strength of the link between Upper and LHR PCB levels has
diminished since the collection of data which formed the basis for the 2002 ROD. Reduced loads
at Waterford translate to reduced fish body burdens in the LHR, but the reduction may now be
relatively minor likely due to the current dominance of LHR legacy sediment contamination in
exposure.
The FYR recognized the variation in decay rates in fish tissue along the LHR. However, EPA does
not agree with the assertion that increases in UHR fish levels were constrained to the areas in the
immediate vicinity of the dredging. While the effects may have been greatest in these locations,
all UHR fish monitoring stations downstream of dredging showed some increases. EPA agrees
with the commenter's assertion that PCB levels in fish did not respond proportionately to the
increase in water column concentrations (e.g., compare the relative rise in water column
concentrations in Figure Al-1 of Appendix 1 of the FYR report with Figures A3-2 to A3-5 of
Appendix 3 of the FYR report, which could reflect the roles played by sediments in fish exposure,
as well as the localized increases in dissolved water column PCB that were observed in near field
monitoring. However, even fish in the LHR at RM 152 and RM 113 were affected by the increased
loads to the LHR, as evidenced by the change in slope, or actual increase in body burden, for most
fish at these stations during the dredging period. See Figures A3-5 and A3-6 in Appendix 3 of the
FYR report for the trends in absolute body burden. Lipid-normalized trends show less of an impact,
particularly at RM 113 but still are suggestive of a weak dredging-related impact for some species.
See Figures A3-12 and A3-13 of Appendix 3 of the FYR report.
Downstream of RM 113, the data do not suggest a dredging-related impact. This gradual
attenuation with distance in dredging-related increases in fish tissue PCB concentrations parallels
the decline in PCB level decay rates in fish. Because LHR fish body burdens are declining more
slowly than UHR fish body burdens and are less responsive to UHR dredging-related loads, it can
be concluded that the factors driving LHR fish body burdens are not now strongly linked to those
driving UHR fish body burdens. As noted by the commenter and stated in the FYR, given the lack
of strong correlation, these lines of evidence indicate that further remediation of the UHR would
be unlikely to result in substantial improvements in LHR fish PCB levels, particularly for areas
downstream of Albany.
This conclusion does not mean, however, that all dredging-related benefits to the LHR have been
realized. As also discussed for the UHR, EPA anticipates it will take several years before the
improvements directly resulting from the dredging are evident in the UHR, and this also applies
to locations in the upper portion of the LHR that clearly responded to UHR loadings, including
Albany/Troy (RM 152). EPA has determined that at least 8 years of monitoring data on PCB
129
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
levels in fish, water, and sediment are needed before the magnitude of the dredging-related
improvements can be determined.
EPA does not agree with the commenter's assertion that all areas of the LHR were declining more
slowly than anticipated prior to dredging. As shown in Appendix 3 of the FYR report, Figures A3-
5 to A3-6, body burdens at RM 152 and RM 113 were declining from 1997 to 2008 at rates
comparable to those predicted by the ROD. Note the close agreement in line slope as well as
absolute magnitude for the actual observations (dashed blue line) with the model forecasts (red
and purple lines) for most fish species at each of these stations. Lipid-normalized results show
similar agreement at the RM 152 station (see Figure A3-12) but less agreement at RM 113 (see
Figure A3-13). The greatest deviations between anticipated recovery and actual observations in
fish occur for the RM 90 andRM 50 stations (see Figures A3-7, A3-8, A3-13 and A3-14). In these
instances, the fish levels are clearly declining more slowly than anticipated by the ROD. As
mentioned previously, EPA is evaluating the needed monitoring requirements and additional study
needs for the LHR in order to understand the observed trends.
Despite the much more gradual decline evident in the most-downstream LHR fish tissue levels,
EPA notes that LHR fish tissue levels are approaching, and in some cases have fallen below, the
interim remedial target levels. In examining these data, EPA has developed Tables 22-1 a and 22-
lb, which present mean and median PCB concentrations in fish tissue for long-term fish
monitoring stations in the LHR, based on homologue-equivalent TPCBhe and TPCBArocior results,
respectively. In each table, data are provided for two periods, 2009 to 2015, representing mean
and median fish tissue concentrations in the LHR during the dredging period, and for 2016,
representing the first year of post-dredging data. Note that 2016 data are not available for all
species and stations. Additionally, data available for RM 50 are limited to a single species for the
post-2009 period. While data from a single year post-dredging cannot provide an estimate of the
long-term decay rate, the table does note which species have achieved one or more of the interim
targets or final remedial goals.
For 2016, mean and median concentrations for yellow perch, spot tail shiner and striped bass at
RM 152 all fall below the interim target of 0.4 mg/kg-ww {i.e., on a wet weight basis), using either
measurement basis {i.e., TPCBhe or TPCBArocior). This represents three of the seven species with
long-term data studied at this station. Below RM 152,2016 fish data are sparse but dredging period
data are available for many species and locations. At RM 113, mean fish tissue concentrations for
two species, brown bullhead and yellow perch, were at or just below the interim remedial target
of 0.4 mg/kg-ww during the dredging period, using either measurement basis. At RM 90, three
species fell below the interim target of 0.4 mg/kg-ww during 2009-2015. Median concentrations
fell below the interim value of 0.4 mg/kg-ww for even more species. Note that by definition, when
median concentrations fall below a threshold, this indicates that more than half of the observations
fall below this threshold. Overall, based on the most recent data available (see the last column in
each table), about thirty percent of the species-station pairs {e.g., yellow perch at RM 152) fell
below the first interim remedial target of 0.4 mg/kg-ww using either measurement basis. Forty
percent of the median values (10 of 25 pairs) fell below this target as well. In the case of yellow
perch, all three stations with data were near or below the second interim remedial target of 0.2
mg/kg-ww. These observations do not mean that the fish consumption advisories can be modified
130
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
for yellow perch or any other species because that is a decision for the State of New York.
Nevertheless, these data do indicate that some recovery is occurring in the LHR.
3.4.8 Comment 52: Adequacy of the OM&M sediment sampling program, especially with
respect to development of post-dredging baseline information
Comment
Sediment
Commenters stated that the current surface sediment OM&M sampling program is not adequate to
provide an appropriate baseline conditions of post-dredging concentration. Available surface
sediment data are not sufficient to evaluate the percent reductions in surface sediment
concentrations that are achieved by the remedy. The data collected pursuant to the 2016 work plan
for the FYR are not in compliance with the 2010 decision documents and are inadequate to track
effectiveness and protectiveness of the remedy and therefore, cannot appropriately be used in the
FYR. The Downstream Depositional Study (DDS) does not provide post-remediation baseline
concentrations and is not suitable for evaluating sediment recovery rates.
Comm enters asserted that an estimate of the rate of change developed over a 10-year interval and
on a River Section basis is not sufficient to evaluate the performance of the remedy in a time frame
commensurate with the remedial targets. EPA should direct that an increased number of samples
be collected such that there is a statistical power to determine sediment concentration trends in 5
rather than 10 years. Sediment sampling should be performed on pool-by-pool basis, and should
occur at smaller time intervals (i.e., more frequently). EPA should reveal the fundamental basis
for the sample design analysis. EPA should develop a robust and data-driven monitoring program
for surface sediment.
A commenter recommended using the existing framework of the SSAP (80-foot sampling grid) to
quantify the concentrations in each reach (pool). Transects also should extend beyond the SSAP
sampling area to extend coverage to the entire area of each reach, including previously unsampled
areas as well as remediated CUs. EPA should use probability-based statistical design for selection
of sample locations within dredged and non-dredged areas. Sample size should be determined
using variability of existing data to quantify temporal decay rates with adequate precision. A
commenter recommends that an additional 1,800 sediment samples be collected in each sampling
event.
Fish
A commenter stated that it should not be necessary to wait eight years to determine the rate of
decline for fish PCB concentrations. EPA needs to perform a statistical power analysis to
determine the number of fish samples to collect.
Overall
Commenters requested that EPA increase the number of sediment and fish tissue samples to the
scale and frequency necessary to optimize the remedy through further remedial work as necessary
to achieve the targeted fish PCB reductions identified in the ROD. EPA should ensure the
131
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
collection of sufficient water, sediment, and fish data to fully assess whether the remedy will meet
the targets in the ROD, starting with the initial target of 0.4 ppm PCBs in fish by 2020.
If the targets are not likely to be met, EPA must direct that sufficient additional remedial work be
done. To date, EPA's persistent refusal to collect and analyze a full array of data has run counter
to EPA's original commitment to clean up the site. EPA has thus far refused to do so, and as a
result NYSDEC has begun gathering the needed sediment data starting in summer 2017.
Response
The commenters request EPA to increase the number of sediment and fish tissue samples to the
scale and frequency that they believe is necessary to achieve the targeted fish PCB reductions
identified in the ROD. Inherent in the comments is an underlying inference that available sampling
results indicate that the recovery of sediment and fish is known to be on a trajectory that will miss
the targets stated in the ROD. However, EPA's analysis conducted for the FYR report shows that
prior to dredging, the actual rates of decline in surface sediment and fish tissue PCB concentrations
in the Upper Hudson River remained within reasonable bounds of uncertainty as compared to those
anticipated at the time of the ROD, and therefore EPA does not have reason to believe that the
targets identified in the ROD will not be met within the general timeframes identified. Although
pre-dredging data does not provide reason to believe the targets will be missed, EPA recognizes it
does not at this time have sufficient data to determine the post-dredging rates of decline and
therefore cannot determine if the ROD targets will be achieved within the expectations of the
ROD. If further sampling data indicate that remedial goals are unlikely to be met within the
timeframes contemplated by the ROD, EPA will evaluate whether further action should be taken.
Further action could include additional sampling and analysis.
The current OM&M sampling program was statistically designed to estimate the spatial average
within each of three river sections with relative error of 50 percent in RS 1, 40 percent in RS 2 and
25 percent in RS 3. To meet this objective, the 2010 OM&M Scope of Work (USEPA, 2010)
estimated that 350 sample locations from the non-dredged areas and a minimum of 50 locations
from backfilled areas in each river section would have to be sampled during each sampling event.
With this design, it was anticipated that in each river section a 5 percent annualized decline over
10 years would be detectable with at least 80 percent power. However, the estimate of the number
of samples relies on the variance of available data. Using estimates of variance developed from
results for surface sediment samples collected during the 2011-2013 DDS program, EPA
subsequently reduced the number of samples required to 226 in non-dredged areas. It should be
noted that although EPA's sampling design is on a river section basis, the sampling locations were
selected independently within dredged and non-dredged areas and were allocated proportionally
to the size of each stratum within each designated river-mile segment. This stratified random
sampling design yields samples that are spatially balanced along the entire length of the Upper
Hudson River in proportion to the area to be sampled. Therefore, this sampling approach provides
data for each river reach. EPA will continue to update the number of samples required for future
OM&M sampling events as new data become available. Based on EPA's power analysis, the
frequency and scale of data collection established were appropriate for meeting EPA's DQOs
based on the ROD requirements. EPA's monitoring program is also consistent with or exceeds the
132
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
level of post-remedial monitoring effort at other Superfund sediment sites, including the Fox
River, Wisconsin (USEPA and WNDR, 2009) and Portland Harbor, Oregon Superfund Sites.
NYSDEC asserted that EPA's OM&M sediment sampling plan was not sufficiently rigorous to
determine whether the remedy was performing in a manner needed to reach the fish tissue
concentration targets outlined in the ROD. As a result, NYSDEC undertook a more intensive
sediment program in 2017 with DQOs established to detect an 8 percent annual change in total
PCBs in surface sediment over a 5-year timeframe with statistical power of 80 percent for each of
eight river reaches. The results of NYSDEC's surface sediment sampling program became
available in early 2018, a few months after the end of the public comment period established by
EPA for the FYR. In consideration of concerns raised by commenters, and in collaboration with
NYSDEC, EPA undertook an extensive technical review of the results from the approximately
1,200 sediment samples collected by NYSDEC in 2017, together with the sediment samples taken
by EPA/GE under the OM&M sampling program in 2016. EPA has documented the findings in a
technical memorandum (Louis Berger & Kern Statistical Services, 2019), which is available on
EPA's Hudson River web page (https://www3.epa.gov/hudson/).
EPA's evaluation shows that the 2017 NYSDEC samples yielded comparable results to the 2016
EPA/GE OM&M samples regarding the mean concentrations on both the river reach and river
section basis. Both datasets (individually and collectively) suggest that the remedy achieved the
required percent reductions in surface sediment concentrations. 17 The surface sediment
concentrations in non-dredged areas in 2016/2017 were at or below concentrations forecast by the
empirical trends derived from historical data. The data also show that the rate of decline in surface
sediment concentrations from non-dredged areas in RS 1 are consistent with ROD model forecasts
(8 percent), with a best estimate rate of 6 percent per year. The combined 2016 EPA/GE OM&M
and 2017 NYSDEC data will be used as the baseline conditions of post-dredging concentration.18
Estimates of recovery rates will be based on comparison of these baseline concentrations with
future rounds of sediment data collected within the OM&M program. Despite the larger number
of samples collected by NYSDEC in 2017 compared to the samples collected by EPA/GE in 2016,
the similar results from these two datasets suggest that the scale of data collection under EPA's
OM&M program is appropriate to meet the requirements of the ROD and the 2010 Statement of
Work. Prior to the next round of sampling, EPA will evaluate the number of samples needed to
confirm that sufficient samples will be collected to allow for a meaningful evalution on both a
river reach and river section basis.
With respect to commenters' assertions that sampling frequency should be increased so that the
rate of decline could be evaluated in 5 years rather than 10 years, and that sediment sampling
should be performed on pool-by-pool basis. EPA agrees that sampling and evaluation of sediment
17 On an area-weighted average concentration basis, the 2016 OM&M data show that the overall percent reductions
(remediation plus natural recovery) of Tri+ PCB were estimated to be 96 percent, 88 percent and 80 percent for RS
1, RS 2 and RS 3, respectively (see Table A4-5 in Appendix 4 of the FYR report). When using the combined 2016
OM&M and 2017 NYSDEC datasets, the overall percent reductions were 93 percent, 89 percent and 87 percentfor
RS 1, RS 2 andRS 3, respectively (compare Table A4-5 in Appendix 4 of the FYR report and Table 3.2-1 in EPA's
technical memorandum [Louis Berger & Kern Statistical Services, 2019] These percentage reductions are
substantially greater than what were anticipated by the ROD.
18 EPA agrees that the DDS program was not designed to evaluate the average concentrations in a specified area and
should not be relied on as the only basis for evaluating sediment recovery rates.
133
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
data by river reach and river section are both useful. EPA proposes to add the reach consideration
to the future assessment of the recovery of the river. However, EPA reasoned that large areas of
undiscovered contamination that would cause one reach to be markedly higher than others in the
same river section would not have been missed by the remedial design sampling. Therefore, it
was, and is, reasonable to design the sampling program on a river section basis. This is further
supported by the agreement between the 2016 EPA/GE and 2017 NYSDEC data as discussed
above, even though GE's sampling program consisted of significantly fewer samples than
NYSDEC's. EPA also believes that a 5-year window is too short to detect a recovery rate of 5
percent with acceptable statistical power. Commenters appear to underappreciate the data
requirements necessary to accurately estimate first order recovery rates from empirical data. For a
given population of PCB data, the precision of recovery rate estimates varies with the sample size,
the frequency of monitoring, the duration of the monitoring program, and the variability of the
sample data. Of these parameters, the precision of the recovery estimates is the most sensitive to
the duration of the monitoring period, followed by the number of samples collected in the first and
last monitoring time step. Although counterintuitive, increasing the frequency of monitoring has
much less influence on the accuracy and power of the monitoring program. EPA's program
optimizes these design parameters recognizing that little is to be gained by frequent monitoring
over a short period of time relative to a 10-year program that will estimate recovery rates much
more accurately. Based on EPA analyses, statistical power to detect recovery rates in this 3 to 10
percent range would be dramatically reduced by restricting attention to a 5-year period, and
increased temporal monitoring frequency would not mitigate the problem.
EPA agrees that the OM&M program should be used to optimize next steps in evaluating the Upper
Hudson River remedy. EPA agrees that the post-dredging fish and sediment data results available
for this FYR are inconclusive indicators of remedy "protectiveness." More monitoring is needed.
EPA will continue to review fish tissue data from semi-annual sampling, and fish will in the future
be collected from additional sampling locations beyond those that have been used for many years.
EPA will also carefully consider each round of sediment data collected within the OM&M
program and consider a range of adaptive responses as those data become available.
3.4.9 Comment 58: EPA recognized that more PCBs were present in the Upper Hudson
River sediments than originally estimated in the 2002 ROD but did not alter remedial
activities to account for this knowledge
Comment
Several commenters concluded that a change was needed but not made in the remedial strategy for
PCBs. Specifically, the commenters noted that the remedial design investigation identified
substantially more PCBs in the sediments of the Hudson than originally anticipated in the ROD.
Commenters indicated that EPA has not provided a satisfactory scientific rationale for not
expanding the remedial work to take into account the increase in PCB mass that was identified
prior to dredging activities. As a result, commenters state that there will be more PCB mass left
behind than originally anticipated.
Commenters were concerned that PCB-contaminated sediment in the shallow portions of the river
was missed during dredging, and these areas are where exposure to children and wildlife occur. It
134
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
is the commenters' understanding that the floodplain project will extend only to the edge of the
water, not into the shallows. Thus, these contaminated areas will be left for future generations.
Commenters state that two to three times as much PCBs remain in the river as originally expected,
and without additional removal, a "series of Superfund-caliber sites will be left behind" in the
Hudson River.
Response
EPA agrees that more PCBs were identified and dredged than originally anticipated but disagrees
with the contention that an adjustment was needed in the remedial design to address the additional
PCB inventory. The 2002 ROD estimates of PCB inventory in sediments of the Upper Hudson
River were based on data collected prior to the publication of the ROD, including sediment data
from various sampling programs conducted between 1977 and 1998. These datasets were collected
for different purposes using different sampling methodologies but were the best sediment PCB
data available at the time of the 2002 ROD. Additional information on the differences between
sediment sampling programs is described in more detail in the 2002 Responsiveness Summary for
the ROD (EPA, 2002) and EPA's white paper19 (EPA, 2016).
Subsequent to the release of the 2002 ROD, from 2002 to 2005 an extensive sediment coring
program (SSAP) collected data at a higher resolution both spatially and vertically than the pre-
ROD sampling. The purpose of the SSAP sampling was to refine the areal and vertical extent of
dredging within the Upper Hudson River as part of the Remedial Design (GE, 2002; GE, 2007).
The higher spatial and vertical resolution of the SSAP dataset identified more PCB inventory and
higher surface concentrations than originally anticipated in the ROD. However, the purpose of the
sediment removal program in the ROD was to reduce surface sediment concentrations and
sediment inventory so as to reduce fish tissue concentrations. In other words, the primary goal was
not to reach a pre-determined sediment concentration or PCB mass, but rather to achieve a
sufficient reduction in sediment concentrations to yield a proportional reduction in fish tissue
concentration (see also Master Comment 40 [see Section 3.3.17], regarding the relationship
between surface sediment concentration and fish tissues concentrations).
The 2002 ROD design was based on ultimately reducing fish concentrations in the Upper Hudson
by approximately 99 percent or more, basically by reducing their exposure to contaminated surface
sediments by 79 percent in RS 1, 64 percent in RS 2, and by 4 percent in RS 3 via dredging20 and
allowing natural attenuation to achieve the further reduction in exposure (additional information
regarding calculation of these reduction percentages is provided in Table 1 of Appendix A of the
2012 FYR report (EPA, 2012)). As can be seen in these numbers, the majority of the reduction in
RS 1 and RS 2 was expected to be achieved by dredging. If the SSAP data (which was not available
at the time of the ROD) are taken into account, the anticipated levels of reduction via dredging
alone based on the ROD criteria (3/10/Select) were 87 percent, 36 percent, and 5 percent,
19 See: White Paper: Responses to NOAA Manuscript Entitled: "Re-Visiting Projections of PCBs in Lower Hudson
River Fish Using Model Emulation" (Field, Kern and Rosman, 2015) (EPA, 2016)
20 Note that for an initial average concentration of 10 mg/kg, the planned 79 percent reduction would yield 2.1 mg/kg
after dredging. Similarly, a 64 percent reduction yields 3.6 mg/kg and a 4 percent reduction yields 9.5 mg/kg. In a
parallel manner, the achieved reductions based only on S SAP data would yield 1.3 mg/kg for a 87 percent reduction,
6.3 mg/kg for a 37 percent reduction and 9.5 mg/kg for a 5 percent reduction.
135
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
respectively, for the three river sections (EPA, 2012). EPA did not adjust the ROD criteria after
the SSAP data became available since it was anticipated that the thresholds set by the ROD would,
when accounting for the SSAP dataset, result in greater proportional reductions in surface
sediment concentrations than estimated in the ROD for RS 1 while RS 3 would achieve the planned
proportional reduction. Only in RS 2 the remedial action appears not to achieve the proportional
reduction when accounting for the SSAP dataset. However, as discussed in more detail below,
when more recent and more representative data from 2016 are examined for RS 2, it is apparent
that this section also achieved a reduction similar to that anticipated by the ROD.21
As part of the FYR report, EPA compared the SSAP data with the 2016 surface samples and the
CU backfill sampling results to estimate the overall change in surface sediment concentrations.
This comparison accounts for both the active remedy {i.e., dredging) and natural recovery. On an
area-weighted average basis, the overall percent reductions (the active remedy plus natural
recovery) in surface sediment concentration (0-2 inch interval) of Tri+ PCB were estimated to be
96 percent, 88 percent and 80 percent for RS 1, RS 2 and RS 3, respectively. These percentage
reductions are substantially greater than anticipated in the ROD for the active remedy alone, i.e.,
79, 64, and 4 percent for RS 1, RS 2 and RS 3, respectively. The calculations also indicate that
post-dredging average surface sediment concentrations of Tri+ PCBs are near or below 1 mg/kg.
Additional details on the calculation of these reductions can be found in Tables A4-5 and A4-6
and accompanying text in Appendix 4 of the FYR report (EPA, 2017). By implementing the active
remedy as specified in the ROD and accounting for natural recovery, EPA achieved a better than
planned reduction in surface sediment Tri+ PCB concentrations in all river sections based on the
2016 sampling data.
It should be noted that the discussion above is predicated on the representativeness of the SSAP
data for the non-dredged areas. That is, the percentage reductions estimated above assume that the
SSAP data obtained for the non-dredged areas can be used to accurately estimate the average PCB
concentrations in all non-dredged areas. In RS 1, the SSAP data can probably be considered
representative since the sampling grid extended across the entire river section, and few areas were
left un-sampled. However, in RS 2 and RS 3, the non-dredged area sampling was focused on the
areas closest to the dredging zones to establish the boundary between areas for removal and those
that could be left in place. As such, it is likely that these data are not representative of the entirety
of non-dredged areas. This systematic bias in the SSAP data for RS 2 and RS 3 was appropriate,
given the goals of the program, but EPA's use of the SSAP data to estimate inventory remaining
and average surface concentration in non-dredged areas has yielded values that are almost certainly
biased high in RS 2 and RS 3.
This observation on the bias in the SSAP data is supported by the more recent surveys of surface
sediment concentrations outside the dredging prisms. For both the Downstream Deposition Study
(DDS) program directed by EPA in 2011 to 2013 and for the 2016 non-dredged area study
designed by EPA with a statistically unbiased sampling layout, the surface concentrations of PCBs
in non-dredged areas are substantially lower than those would be suggested by the SSAP data,
even after allowing for natural recovery. As shown in Table A4-5 in Appendix 4 of the FYR report,
current estimates of Tri+ PCB concentrations in all non-dredged areas in all three river sections
21 Levels of reduction greater than 87 percent, 36 percent, and 5 percent for RS 1, RS 2 and RS 3, respectively, were
also confirmed by the 2017 NYSDEC surface sediment survey.
136
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
are less than 2 mg/kg. These values should be contrasted with EPA's estimates of pre-dredging
PCB concentrations in non-dredged areas (Pre-dredging SSAP Survey column in Table A4-5) also
shown in the table. Given the statistically rigorous unbiased sampling design used in 2016, these
results either suggest a very rapid rate of recovery in non-dredged areas (the rate would be
equivalent to a 2.5-year half-life) or, more likely, that the earlier data sampling design was biased
and did not provide an accurate estimate of the average PCB concentration or, most likely, some
combination of the two factors. In any case, the most recent data indicate that surface
concentrations of Tri+ PCB in non-dredged areas are less than 2 mg/kg in all river sections, and
that after combining these data on an area-weighted basis with the low concentrations of the
dredged areas, overall average concentrations are not statistically different from 1 mg/kg on a river
section basis.22
In conclusion, based on the 2016 EPA/GE and 2017 NYSDEC data, surface sediment
concentrations of Tri+ PCB are between 80 and 95 percent lower than those observed in the SSAP
survey. While the decrease is undoubtedly due to some combination of active remediation, natural
recovery, and artifacts of sampling design, the most recent data clearly show low average surface
sediment concentrations throughout the Upper Hudson.
The risk assessment completed as part of EPA's investigation of the Upper Hudson identified fish
consumption as the major pathway of exposure yielding risks to human health. No cleanup criteria
were developed for sediments based on direct human exposure since this pathway does not yield
unacceptable risks to humans. However, in the investigation of shoreline areas during the RI and
the remedial design, the shoreline and shallow areas were extensively sampled, identifying those
areas that exceeded EPA's removal thresholds. These areas were not missed but, if they were not
dredged, they were either demonstrated to be low in PCB level or likely to be low based on indirect
lines of evidence (sediment texture), or the shoreline area was too unstable to permit remediation.
Based on measured PCB levels, swimming in the Upper Hudson River does not pose an
unacceptable risk to human health. Hence, no further remediation is needed in the Upper Hudson
River to reduce exposure via in-river activities such as swimming and wading. As part of the
floodplain comprehensive study, EPA is investigating shoreline/floodplain areas up to the edge of
the water at normal river elevation. Additionally, areas of human use along the shoreline that
become exposed as water levels recede are being investigated.
3.5 Protectiveness Determination
This section includes comments and responses with respect to whether the determination is
consistent with the EPA FYR guidance and policy, RAOs, and considerations related to
institutional controls. Since the proposed Second Five-Year Review was issued in 2017, EPA has
completed additional technical analysis supporting comment responses related to the deferral
protectiveness statement discussed in this section (See Appendix B of this document).
22 The current surface sediment concentration in non-dredged area was confirmed by the 2017 NYSDEC surface
sediment survey
137
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.5.1 Comment 12: EPA must consider protection of natural resources as fish consumption
advisories do not protect environmental receptors
Comment
Commenters noted that fish consumption advisories (FCAs), which address human consumption
of impacted wildlife, do not protect all environmental receptors such as fish, birds, small
mammals, and benthic organisms that could be exposed to PCBs left behind by the remedy.
Commenters indicated that EPA, as the environmental agency charged with implementing the
remedy and ensuring its protectiveness to human health and the environment, should quantify the
impacts to these receptors in the Second Five-Year Review.
Response
The RAO established the ROD for protection of ecological receptors is to "reduce the risks to
ecological receptors by reducing the concentration of PCBs in fish," since consumption of fish
contaminated with PCBs remains the primary route of exposure for most upper trophic level
wildlife species. The results of the Baseline Ecological Risk Assessment (BERA) supported EPA's
decision that remedial action was necessary to reduce unacceptable risks to ecological receptors.
The FYR report contains a summary of the BERA conducted for the ROD and, consistent with
Question B of the FYR process, an evaluation of ecological risk exposure assumptions and toxicity
values (see Section 5.2.3.2 Ecological Toxicity and Appendix 11 of the FYR report). Based on
current information available in the scientific literature, EPA concluded that updates to ecological
exposure assumptions and refinement of the toxicity values do not affect the protectiveness
determination of the selected remedy with respect to ecological receptors.
The risk-based goal for the ecological exposure pathway is a range from 0.3 to 0.03 mg/kg PCBs
in fish (largemouth bass, whole body) and for consumption of fish by the river otter. This
ecological goal is considered protective of all the ecological receptors evaluated because it was
developed for the river otter, determined to be at greatest risk from PCBs at the Site. In addition,
a range from 0.7 to 0.07 mg/kg PCBs in spottail shiner (whole fish) was developed for the mink,
which is a species known to be sensitive to PCBs. Other species, such as the bald eagle, were
considered but are at less risk than the river otter.
The dredging remedy has reduced PCB inventory in the sediment, thereby reducing exposures to
wildlife. More data will need to be collected before a determination can be made as to the longer-
term effect the dredging has had on reducing fish tissue concentrations relative to the risk ranges
given above. The number of years needed to reach a conclusion will be based in part on the
variability of the data. However, EPA anticipates that it will take as many as eight or more years
post-dredging fish tissue data to identify trends with a reasonable degree of scientific certainty.
138
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.5.2 Comment 13: EPA must include a site-wide protectiveness statement in accordance
with the guidance
Comment
Commenters indicated that EPA must make a site-wide protectiveness determination since
remedial construction is complete at the Hudson River Superfund Site. The protectiveness
determination should generally be the same protectiveness determination as the one for the least
protective OU at the site. In addition, because the OU2 remedy here includes the use of institutional
controls by way of the NYSDOH fish consumption advisories, EPA must also evaluate the current
and long-term effectiveness of the fish consumption advisories and include relevant information
about the advisories as part of the protectiveness determination.
Commenters also indicated that EPA admits that the cleanup is not protective of human health and
the environment in the Lower Hudson River (LHR) by omitting a protectiveness determination for
the 150-mile stretch below the Federal Dam. In the First FYR, EPA issued a site-wide
protectiveness determination for the entire 197-mile Superfund site. However, the Proposed
Second FYR did not contain a site-wide determination. While EPA claims that the cleanup "will
be protective" in the Upper Hudson River (UHR), EPA makes no determination about the cleanup
for the 150-mile stretch of the Hudson River below the Federal Dam. Omitting a protectiveness
determination for this portion of the Site is concerning and has caused confusion among the people
who live, work, and play along the LHR.
Response
EPA developed the protectiveness statements for OU1 and OU2 using the Agency's
comprehensive five-year review guidance and supplemental memoranda on the use of
protectiveness statements. EPA considered adding a site-wide protectiveness statement, as in the
2012 FYR report. However, in accordance with the guidance EPA did not include a site-wide
statement in this FYR because the Agency is still at the RI/FS stage regarding OU4 (Upper Hudson
River Floodplain) and is just beginning to conduct supplemental studies of the LHR.
As per the guidance, a site-wide protectiveness statement is typically issued when a site that has
multiple operable units (OUs) and has reached construction completion. The guidance discourages
issuing a site-wide statement prior to this because all remedies at the site may not have been
selected and constructed. Therefore, to minimize any confusion and in accordance with the
guidance, EPA chose not to issue a site-wide statement in this FYR.
Limited data collection from the LHR indicates that recovery rates are slower than in the UHR and
may no longer be strongly associated with PCB loading from the UHR. The rate of decline of fish
tissue PCB concentrations generally decreases with distance downstream. As a result, there is a
decrease in the correlation between fish PCB concentrations in the UHR and LHR with distance
downstream. This indicates that PCB sources in the UHR have less of an impact on LHR fish than
on fish in the UHR. PCB removal by dredging in the UHR has reduced PCB transport to the LHR.
This beneficial reduction, along with continued natural recovery, is expected to continue to reduce
PCBs in the LHR.
139
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Water column PCB concentrations at Albany/Troy were consistent with modeling predictions
during the MNA period and, as expected, increased during the dredging. By contrast, results at
Poughkeepsie were generally higher than model predictions and were not impacted by the
dredging, indicating that the strength of the relationship between UHR and LHR water column
concentrations weakens with distance downstream. It should be noted that there are other sources
of PCBs in the LHR, including legacy sediment contamination and possible local sources.
Although the local sources have been less significant than the GE sources of PCBs originating in
the UHR, both these LHR sources and legacy sediment contamination should continue to be
further investigated.
EPA agrees that it is important to carry out supplemental studies of the LHR and will begin that
work in 2019. These studies will supplement information collected during the Reassessment
process in the 1990s that led to the 2002 Record of Decision, along with the results of periodic
monitoring of LHR fish and water by GE under EPA oversight since 2004, and periodic
monitoring of LHR fish by New York State. The supplemental studies will also help inform the
need for a remedial investigation and feasibility study. It is too early in the process to determine if
a cleanup is needed in the LHR.
Regarding the effectiveness of the fish consumption advisories, the State of New York has in place
fishing restrictions and advisories against consumption of fish to control human exposure
pathways that could result in unacceptable risks. EPA acknowledged in the ROD that the
consumption advisories are not fully effective in that they rely on voluntary compliance in order
to prevent or limit fish consumption. EPA will continue to work with New York State to ensure
the ongoing maximum effectiveness of the advisories. See FYR report Section 2.4.2 (Institutional
Controls for OU2) and Appendix 13 for additional details regarding the fish consumption
advisories.
3.5.3 Comment 32: "Will be protective" is not an appropriate determination for the
Hudson River PCBs Site. "Will be protective" is only appropriate when a remedy is
still "under construction."
Comment
Commenters state that for the purposes of developing a protectiveness statement, construction of
the remedial action is complete. According to the Protectiveness Determination Guidance, a "will
be protective" determination is only appropriate when remedial construction activities are
ongoing, but the remedy is anticipated to be protective upon completion and no remedy
implementation or performance issues have been identified. Therefore, "will be protective" is not
an available option for the OU2 remedy because construction of the remedy is complete. The
physical (dredging) and engineering components of the remedial action were completed in 2015
and 2016, respectively.
140
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Response
A protectiveness determination of "will be protective" is an appropriate option for remedies at
which construction activities are ongoing. Construction was not complete at the end of the time
for the FYR (December 2016). A brief discussion of this point is included below. However,
because there is limited post-dredging data available and EPA has determined that as many as 8
or more years of fish tissue data are necessary to establish statistically reliable trends, EPA has
differed making a protectiveness determination at this time. Additional supporting technical
information regarding EPA's deferral determination is included in Appendix B of this document.
EPA appropriately considered data and information collected through December 2016 for the
Second FYR and evaluated OU2 protectiveness as of the end of that year. Data and other
information obtained after December 2016 will be considered in the next FYR for OU2.
Although demobilization of the sediment processing facility was largely completed in December
2016, certain demobilization activities, including removal of filter presses and subsequent
sampling in the filter press building, were not completed until April 2017. EPA project staff also
coordinated with EPA Headquarters FYR staff on interpretation of EPA's five-year review
guidance and it was agreed that construction was not complete at the end of 201623.
3.5.4 Comment 37: Institutional controls should not be a part of the remedy
Comment
One reviewer commented that EPA should take note of the effectiveness of institutional controls
(ICs) as stated in the ROD and indicated that understanding that fish advisories rely on voluntary
compliance and therefore are not completely effective in preventing fish consumption is a primary
basis for the need, identified in the ROD, for rapid reductions in human health risk in the years
immediately following remediation.
Response
ICs are an integral part of Superfund site management, investigation, remediation, and post-
remediation monitoring. Specifically, ICs are non-engineered measures such as administrative
and legal controls that help minimize the potential for human exposure to contamination and/or
protect the integrity of the remedy. ICs are routinely employed at remedial sites and are routinely
used by EPA and other government agencies at Superfund sites. As discussed in the 2002 ROD,
ICs, including continuation of fish consumption advisories and fishing restrictions, were
anticipated to be implemented as long-term control measures, along with active remediation and
a long-term monitoring program. These controls are designed to prevent or limit exposure to PCBs
through consumption of contaminated fish. Hudson River ICs and their role in the overall
remediation approach are discussed in Section 2.4 and Appendix 13 of the FYR report.
23 EPA guidance included: (Comprehensive Five-Year Review Guidance (OSWER 9355.7-03B-P) and
Memorandum: Clarifying the Use of Protectiveness Statements for Comprehensive Environmental Response,
Compensation, and Liability Act Five-Year Reviews (OSWER 9200.2-111).
141
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
ICs are an important component of the remedy for the Hudson River project and EPA continues
to work closely with NYS to implement them. EPA understands that while ICs rely on voluntary
human compliance and are not by themselves protective of the environment, the remedy is
significantly more protective with the outreach conducted and information disseminated through
the ICs than without it. Because ICs are not "stand alone" remedial components, EPA selected
from a range of remedial alternatives and selected the alternative Removal Criteria by respective
River Sections as stated in the ROD (REM 3/10/Select), which includes upstream source control,
fish consumption advisories/fishing restrictions, and long-term monitoring of post-construction
natural attenuation. EPA anticipates that ICs will need to remain in place for the foreseeable future
as the long-term monitoring component of the remedy continues.
3.5.5 Comment 45: The remedy is not protective
Comment
Commenters have asserted that current site human health and ecological risk levels are in excess
of EPA's acceptable range, that the remedy is thus not protective, and that EPA's protectiveness
statements contradict the fundamental goals of the 2002 ROD. These assertions are based on
comparisons of available post-dredging fish data to the modeling projections contained in the ROD
regarding the estimated time of achievement of the interim fish target of 0.4 mg/kg of PCBs in the
species-weighted Upper Hudson River (UHR) average (Table 11-2 of the ROD). Based on these
assertions, commenters conclude that the remedy is not protective and should indicate to the
agency that further active remediation is necessary.
Response
EPA is deferring its determination of protectiveness because there is not enough data available
since the completion of dredging and related project activities in 2015 to evaluate whether the
remedy is functioning as intended as described in the ROD and the underlying FS. The following
are several key relevant points from the FYR:
• The dredging portion of the remedy was implemented as designed and within
expectations described in the ROD;
• Prior to dredging, MNA was occurring at rates of decline that are generally in agreement
with the modeling done for the ROD;
• Early post-dredging results are within expectations of the modeling analyses presented in
the FS and ROD;
• Fish, sediment and water data are not sufficient to evaluate post-dredging trends and
likely reflect continued impacts from dredging operations [as noted in the ROD (e.g., pp.
68-69), EPA's expectation was that following dredging, the river system would require at
least a year or more to equilibrate to post-dredging conditions and exposures];
• The 2002 ROD exposure assumptions are still valid and appropriate for the Site;
• No other information has come to light that could call into question the protectiveness of
the remedy; and
142
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
• EPA continues to work with New York State to control human exposure pathways that
could result in unacceptable risks from the consumption of fish.
EPA recognizes the remedy for OU2 is not yet protective of human health and the environment.
However, as the ROD makes clear, the remedy includes an extensive post-dredging period of
natural recovery (termed "monitored natural attenuation," or MNA, in the ROD). EPA will
continue to monitor the progress of MNA in the OM&M phase of the remedy for the foreseeable
future.
EPA agrees that, in general, fish tissue concentrations are currently above the ROD's interim fish
target concentration of 0.4 mg/kg. However, EPA does not agree that fish tissue concentrations
are significantly different from model projections. For the 2002 ROD, remedy protectiveness was
evaluated by comparing predicted fish tissue concentration trajectories over time under different
remedial alternatives. The HUDTOX and FISHRAND models were calibrated, verified and
applied for the UHR and designed to support decision-making by allowing direct comparisons of
predicted water, sediment, and fish tissue concentrations across proposed remedial alternatives.
The strength of the models lies in their ability to predict concentration trajectories in sediment,
water, and fish over time for multiple scenarios representing remedial alternatives which could
then be compared based on a consistent set of assumptions. However, model predictions were
likely to differ from actual observations due to: 1) variability in actual exposures; 2) highly
localized exposures; 3) the importance of sediment vs. water exposure pathways, which can vary
over time due to prey availability and natural variability in exposure conditions; 4) uncertainty and
variability in lipid content of fish and prey items; 5) uncertainty and variability in consumption of
specific prey items and PCB concentrations in prey; and 6) measurement uncertainty (including
allowing for differences in sampling programs and analytical methods).
In addition, model forecasting involves population-level assumptions regarding key components
such as lipid content and average exposure. However, post-dredging data collection involves
collection, processing and analyses of data regarding individual fish. Individual fish and species
will respond to contaminant exposures in different ways depending on their foraging strategies
and life histories. As a result, individual fish (and any individual fish species, more broadly) will
achieve "target levels" at different times and may not completely match "absolute" model forecasts
because of varying (real) exposures, diet/available prey, uncertainty in lipids contents based on
diet and available prey, and potential data collection and measurement uncertainty.
Differences between assumptions underlying remedy design and actual implementation are
discussed in Appendix 8 of the FYR report. Although there were differences in the implementation
compared to the underlying assumptions in the analyses presented in the ROD, in general the
implementation was not significantly different than those underlying assumptions. In addition to
uncertainty resulting from differences in design and implementation, the modeling analysis
presented in the ROD assumed a post-dredging "equilibration" period depending on remedy-
specific implementation and construction schedule assumptions that would not (as anticipated) be
tested until design details were worked out. Furthermore, even if implementation had exactly
matched design, some uncertainty was always expected related to water, sediment, and fish in
terms of observed (actual) data exactly matching forecasts. For example, differences in
environmental conditions (e.g., flow rates, upstream boundary conditions) may contribute to
143
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
potential differences between forecasts and observed (actual) fish tissue concentrations,
particularly given that the models were designed primarily to predict relative tissue concentration
trajectories across remedial alternatives rather than absolute concentrations over time.
The FYR report presents comparisons in Appendix 1 for the water column, Appendix 3 for fish,
and Appendix 4 for sediment. Appendix 3 demonstrates that for the pre-dredging MNA period the
model performed well and continues to perform well based on 2016 data (Figure A3-19). This
figure shows that the mean fish tissue PCB concentrations in 2016 for individual species range
from 0.3 to 1.7 mg/kg depending on the species and location. Yellow perch, for example, has
already achieved the 0.4 mg/kg interim target at several locations. The species-river section-
weighted average based on 2016 data is 1.0 mg/kg, which compares well to model predictions as
shown in Figure A3-19. Early post-dredging results therefore are consistent with modeling
analyses and expectations presented in the FS and ROD and do not suggest that there are flaws in
the model forecasts. Fish PCB data will continue to be collected and evaluated to determine
whether subsequent observation cycles demonstrate consistency with ROD-anticipated trends.
EPA acknowledges there are some challenges with certain aspects of data collection and analyses.
Specifically, it has been challenging to reconcile the surface sediment data from the 2002 to 2005
SSAP dataset with trends based on the 1977, 1991, and 1998 datasets to which HUDTOX was
calibrated. Appendix 4 of the FYR report addresses these challenges EPA has also received
comments regarding potential challenges to fish monitoring program implementation. These are
addressed in Appendix 3 and Appendix 8 of the FYR report and in responses to Master Comments
51 (see Section 3.3.23 regarding changes in fish monitoring locations), and 46 (see Section 3.3.20
regarding changes in sample processing procedures). Based on the available information, at this
time there is no reason to question underlying model assumptions or to evaluate the significance
of post-dredging fish tissue results from the perspective of time to attain target levels.
Additional data will be required to evaluate the long-term trend. Power calculations conducted to
support the OM&M sampling program design indicate that as many as eight or more years of data
will be required to evaluate fish tissue trends in a statistically significant and robust manner. The
UHR underwent a "reset" with the implementation of dredging, and post-dredging data will
establish a new "baseline" from which trends must now be evaluated. Post-dredging fish tissue
data reported in the FYR pertain to a single year and were collected in a ROD-anticipated "year
of equilibration." Accordingly, evaluating data-based trends into the future, starting with this new
baseline, will require additional data over multiple annual cycles to provide statistically
meaningful estimates of progress toward meeting the interim and final targets. EPA finds the 2016
fish data results encouraging, but one or two year of data does not establish a "trend" (toward or
away from target levels) and a single year of post-dredging data is not sufficient to conclude that
the remedy is not protective or that further active remediation is warranted. As such, EPA will
continue to monitor post-dredging (natural recovery) results collected under OM&M and to
evaluate remedy protectiveness by comparing future observations to ROD targets and remedial
action objectives.
144
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.5.6 Comment 59: Hudson River PCB concentrations will not reach the target levels
anticipated in the ROD and EPA is claiming a short-term impact to the fish from
recent dredging when such impacts should be negligible
Comment
Commenters asserted that PCB concentrations in the river will not reach the target levels
anticipated in the ROD and that EPA is claiming a short-term impact to the fish from changes in
the dredging and construction schedule when in fact the impact should be negligible. Work in RS
1 in the last year of dredging should not have had a significant effect on fish tissue concentrations
observed during fall 2015 or in 2016. The impact of the dredging work on the fish clearly shows
that the increase in water column PCB concentration did not have a commensurate impact on the
fish in the Hudson River. Typically, only those fish in the immediate vicinity of the dredging work,
or immediately downstream, showed a significant reaction to the dredging.
Response
Although post-dredging MNA recovery rates are not impacted by the construction schedule, it is
not reasonable to assume that construction activities during dredging could have been predicted
exactly as anticipated in the ROD. Direct comparisons between ROD calendar year forecasts and
observed fish tissue concentrations are not necessarily "apples-to-apples" comparisons.
Additionally, while short-term and localized increases and subsequent rapid decreases in fish
tissue PCB concentrations were anticipated in the FS and ROD, and were observed between 2009
and 2016, they were not directly reflected in the long-term fish tissue forecasts presented in support
of remedy selection. For these reasons, direct comparisons of observed data to ROD forecasts
during dredging are not appropriate.
Individual fish species respond to contaminant exposures in different ways depending on their
foraging strategies and life histories. It is important to note that any individual fish (and any
individual fish species more broadly) will achieve "target levels" at different times given: 1)
variability in actual exposures; 2) highly localized exposures; 3) the importance of sediment vs.
water exposure pathways, which can vary over time due to prey availability and natural variability
in exposure conditions; 4) uncertainty and variability in lipid content of fish and prey items; 5)
uncertainty and variability in consumption of specific prey items and PCB concentrations in prey;
and 6) measurement uncertainty (including allowing for differences in sampling programs and
analytical methods). As a result, while the ROD anticipated perturbations to post-dredging fish
tissue recoveries, the full range of specific impacts and the timing of such delays on each fish
species or population could not reasonably have been predicted.
EPA agrees that fish tissue PCB concentrations may be influenced by dredging and related support
work but does not agree that "only those fish in the immediate vicinity of the dredging work, or
immediately downstream, showed a significant reaction to the dredging." Data show (Figures A8-
4.1 through A8-4.12 of Appendix 8 of the FYR report) that fish tissue concentrations may or may
not have varied significantly (statistically) from Baseline Monitoring Period (BMP) levels.
However, species at most stations exhibited elevated tissue concentrations as dredging approached
a sampling location, or in the year of dredging or after dredging, that were significantly
145
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
(statistically) different from either station BMP levels or levels observed in the years immediately
preceding dredging.
Specifically, fish collected in both spring and fall, including black bass, yellow perch, and
pumpkinseed (PKSD), as anticipated by the ROD, exhibited localized and transient increases in
response to dredging at 4 out of 5 Thompson Island Pool (TIP) fish stations during remediation.
Figures A8-4.1 through A8-4.4 of Appendix 8 of the FYR report, also suggest that fish tissue PCB
concentrations for all species continue to drop from the elevated levels observed during dredging
(which concluded in 2015) and do not appear to have leveled off or stabilized yet (data through
2016). This observation is consistent with results from another remedial site, the NYSDEC
Cumberland Bay Site in Lake Champlain. As discussed in Appendix 8 of the FYR report, while
limited pre-dredging data are available for the Cumberland Bay Site, Figures A8-5.1 and A8-5.2
indicate that for both fall-collected species (i.e., rock bass and yellow perch), several post-dredging
years passed before fish tissue PCB levels began to stabilize.
This pattern is also observed for other species at stations located in RS 2 and RS 3 during dredging
despite differences in the duration of dredging immediately upstream (other than RS 1/Reach 8)
of or within a reach. Upper Hudson River (UHR) fish tissue levels not returning to BMP or pre-
dredging levels immediately after dredging may be a product of dredging and support vessel traffic
following dredging but is certainly associated with the approach and implementation of local (i.e.:
at the scale of the reach or station) dredging. As indicated in Tables A8-5 and A8-6, of Appendix
8 of the FYR report, dredging platforms were accompanied by a fleet of support vessels and
sediment barges that also had to transit upstream reaches (e.g., moving dredged materials to the
processing facility) for several years after sediment removal at a given location. As a result, the
end of dredging and backfill operations in a CU may not have resulted in the immediate end of
project activities (and consequent environmental disturbances, such as near-shore wave action) in
the vicinity of individual fish data collection stations or within a reach or river section.
Overall, the data in hand are consistent with ROD expectations regarding localized and transient
increases in fish tissue concentrations. In addition, available data do not conclusively indicate that
fish tissue concentrations have leveled off or stabilized since dredging concluded. In fact, fish
PCB tissue concentration data collected during and since dredging suggest a general downward
trend for several species at multiple sampling stations from within all three river sections. This
pattern is reasonable to expect given that dredging within reaches or sections ended in different
years, but vessel traffic in these reaches may not have. Upper Hudson fish tissue levels not
immediately (or by the spring or fall of 2015 or 2016) returning to BMP or pre-dredging levels
may be a product of site-specific dredging and support activities. However, and as is suggested
by the NYSDEC Cumberland Bay data, it may also reflect that sediments and fish tissue levels
simply require time to stabilize from short-term, transient impacts associated with active
remediation. Appendix 8 of the FYR report (See Figure A8-5.2) indicates that while downward
trends in fish tissue PCB concentrations can be observed in the first few years after dredging, it
will take as many as 8 or more years of data collection before trends in the data can be determined
with statistical confidence. Taken together, these observations suggest that while recent UHR
results are generally consistent with the ROD's expectations, post-dredging data may still be
exhibiting "localized and transient" impacts; and that it is still too early to determine the full extent
146
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
of dredging impacts on local fish tissue concentrations {i.e., more data collection cycles are
needed).
3.6 FYR Process and Public Engagement
This section includes comments and responses on the FYR process, the FYR team formation,
public engagement, the Community Involvement Plan (CIP), and interactions between EPA and
the trustees and other stakeholders.
3.6.1 Comment 5: Consider the risks to Environmental Justice communities
Comment
Commenters noted that the original HHRA did not consider newer subpopulations of anglers, such
as minority or immigrant populations, who rely on subsistence fishing, use different species of
fish, and consume small forage fish in different ways. Additionally, commenters noted that more
people are relying on fish for subsistence than when the ROD was issued, pointing to significant
changes in demographics and fish consumption patterns on the Hudson River, particularly in the
Lower Hudson. Commenters also noted that EPA's Community Involvement Plan's (CIP) goal
with regard to environmental justice is to increase awareness and information about the project,
especially in communities that may not know how to access information or that may not have
many opportunities or methods to do so and that the EPA should consider developing specific
strategies for reaching out to underrepresented communities, as it has done in other locations. As
such, commenters requested that the risk assessment be revisited to take into account all
consumption patterns in order to accurately capture human health risks. They requested that this
information be included in the FYR. Commenters further requested that EPA should ensure that
the communities that are most interested in using the Hudson for subsistence fishing are adequately
informed and have a meaningful opportunity to participate in the public comment process for the
FYR report.
Response
Under CERCLA, cancer risks and non-cancer hazards were evaluated based on potential exposures
to the Reasonably Maximally Exposed (RME) individual. RME is defined as the maximum
exposure that is reasonably expected to occur in the Upper Hudson River under baseline conditions
{i.e., assuming no remediation and no other measures to control exposure, such as fishing
advisories and restrictions) and is not a worst-case exposure scenario. The risk assessment
considers exposures currently and in the future. As described below, the HHRA describes the
process used to evaluate exposures in the Hudson River including evaluation of consumption of
fish by subsistence anglers. Based on the available information EPA considers the original
ingestion rate representative of the RME individual.
Fish Ingestion Rate. The fish ingestion rate used in the HHRA was based upon an estimate of
the long term average consumption of self-caught fish in the angler population, expressed as an
annualized daily average rate in units of grams of fish per day (g/day).
147
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
The HHRA evaluated a number of fish consumption survey studies as described in the assessment.
Based on this assessment, EPA selected the fish ingestion rate based upon a survey of over 1,000
New York anglers (Connelly et al., 1992) who caught and consumed fish. For the adult exposure,
the Central Tendency Exposed (CTE) fish ingestion rate (for the average exposed individual) is
the 50th percentile of the empirical distribution (4.0 g/day) and the RME ingestion rate is the 90th
percentile (31.9 g/day). For a one-half pound serving, these ingestion rates represent approximately
6.4 and 51 fish meals per year, respectively. The process used to develop these ingestion rates are
outlined in the HHRA and externally peer-reviewed.
Subsistence Subpopulations. Subpopulations of highly exposed or less exposed anglers have not
been explicitly characterized, but instead are assumed to be represented in the fish ingestion rate
distribution. For example, the 99th percentile fish ingestion rate from the 1991 New York Angler
survey is 393 meals per year, or more than one fish meal per day. Furthermore, even those
responses claiming a consumption rate of up to 1,000 meals per year were included from the 1991
New York Angler survey. Although it is possible that there are subsistence or highly exposed
individuals who do not obtain fishing licenses, and therefore would not have been captured in the
1991 New York Angler survey or included in the generated distribution of ingestion rates, there
are no known, distinct subpopulations that may be highly exposed in the Upper Hudson River area.
Review of the limited literature on subsistence or highly exposed angler populations supports the
assumption that these subpopulations are likely to be adequately represented in the total
distribution of fish ingestion rates developed for Upper Hudson River anglers. As presented in a
thesis by Wendt entitled "Low Income Families' Fish Consumption of Freshwater Fish Caught
From New York State Waters," low-income families in 12 counties throughout New York,
including Albany and Rensselaer counties were interviewed (Wendt, 1986). Wendt reported that
between 9% and 49% of the low-income families in each county ate freshwater fish from New
York State waters. Wendt then conducted a more in-depth survey of low-income families in
Wayne County, New York, bordering Lake Ontario and determined fish consumption rates. The
average consumption rate was 17.5 meals per year, or 10.9 g/day. In comparison, the arithmetic
average consumption rate from the distribution selected to represent Upper Hudson River anglers
is 27.8 meals per year, or 17.3 g/day.
Some commenters indicated that EPA should take additional steps to ensure that there is sufficient
outreach to the diverse communities in the Lower Hudson River, including low-income
communities, communities of color, and subsistence fishing communities.
Based on public input received when the cleanup decision was made, the EPA committed to
developing a comprehensive public involvement program to be employed throughout the design
and construction phases of the project. As a comm enter accurately noted, according to the EPA's
2009 CIP, EPA's community involvement efforts over the last several years have largely focused
on the upriver communities. This is the area where dredging took place and where the impacts and
effects of the dredging were most directly felt. Environmental justice considerations not only
recognize the burden of industrial pollution from historical practices, but the potential impacts of
cleanups themselves.
148
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
While the dredging component of the cleanup remedy is now complete, the EPA remains
committed to keeping the public informed about future work, including the long-term monitoring
that will be conducted to track the recovery of the river over time, and any efforts that are initiated
in the future to collect additional information/data in the Lower Hudson River. The EPA will
continue to coordinate with New York State and the site's CAG, which includes Lower Hudson
River interest groups, to evaluate outreach needs.
Several commenters also stressed the importance of continued and ongoing outreach to subsistence
fishing communities in the Lower Hudson River to ensure that they are adequately informed about
the PCB contamination in the river and the existing New York State fish consumption advisories.
NYSDOH has primary responsibility for educating and informing people who fish in the Hudson
River about the current New York State restrictions and advisories. As discussed in Appendix 13
of the FYR report, pursuant to the Consent Decree between GE and EPA, GE has contributed $4
million to Health Research, Inc., of Rensselaer, New York, in order to support the State's
implementation of appropriate fish consumption advisories and fishing restrictions. The NYSDOH
has a Hudson River Fish Advisory Outreach program which specifically targets its communication
to high-risk populations, such as women, children and low-income citizens and works to develop
specific strategies for reaching out to underrepresented communities in both the upper and Lower
Hudson River.
The Hudson River Fish Advisory Outreach Project uses various outreach strategies that include
distribution of written and electronic materials, partnerships, and a presence at community events
and public venues to achieve its objectives. NYSDOH fish advisory outreach work has been
conducted in partnership with other state and local agencies. NYSDOH has established
partnerships with commercial fishermen, recreational anglers, boating community representatives,
environmental justice advocates, immigrant rights advocates, local health officials, environmental
conservation officials, parks and recreations officials, health care provider representatives,
community group leaders, and food pantry and community food networks.
To improve its outreach, NYSDOH has also been making educational materials more accessible
to lower-literacy and non-English speaking individuals. NYSDOH is also working with partners,
such as the Latinos Unidos of the Hudson Valley, the U.S. Committee for Refugees and
Immigrants, and the Chinese America Planning Council, to learn about different cultures and
communities to more effectively communicate information to a more diverse audience via both
existing and new venues. These efforts include making presentations to faith-based groups and
establishing "youth ambassadors" to help communicate health advice to their communities. The
addition of part-time project staff who attend public outreach events and possess Spanish and
Chinese language skills also enables the project to reach a broader audience more effectively.
The EPA will continue to coordinate closely with NYSDEC and NYSDOH on the implementation
of the outreach program.
149
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
3.6.2 Comment 17: EPA should ensure that there is adequate outreach to the diverse
communities in the Lower Hudson River
Comment
Commenters state that EPA's community involvement goals include providing understandable
information to the public, ensuring that the public has a meaningful opportunity to engage with
EPA, and helping the public understand the Superfund decision-making process. Commenters
questioned the number of EPA's community involvement activities in the downriver communities.
Commenters provided examples of outreach in the Upper Hudson communities that Lower
Hudson residents do not benefit from such as EPA's enhanced physical presence in the Upper
Hudson through field offices, public meetings, community events, and media appearances.
Commenters question whether EPA has made specific efforts to ensure that its outreach materials,
like fact sheets, technical documents, and e-mails, are widely available to various audiences.
Some commenters indicated that EPA should take additional steps to ensure that there is sufficient
outreach to the diverse communities in the Lower Hudson River, including low-income
communities, communities of color, and subsistence fishing communities.
Response
Based on public input received when the remedy was selected for OU2, EPA committed to
developing a comprehensive public involvement program to be used throughout the design and
construction phases of the project (which included the dredging work). As a commenter accurately
noted, according to the EPA's 2009 Community Involvement Plan (CIP), EPA's community
involvement efforts over the last several years have largely focused on the upriver communities.
This is the area where dredging took place and where the impacts and effects of the dredging were
most directly felt. Environmental justice considerations not only recognize the burden of industrial
pollution from historical practices, but the potential impacts of cleanups themselves.
While the dredging component of the OU2 remedy is now complete, EPA remains committed to
keeping the public informed about future work, including the long-term monitoring that will be
conducted to track the recovery of the river over time, and efforts that are initiated in the future to
collect additional information/data in the Lower Hudson River. EPA will continue to coordinate
with New York State and the site's CAG, which includes Lower Hudson River interest groups, to
evaluate outreach needs. EPA expects that the supplemental studies of the Lower Hudson River
will start in 2019 and will take several years to complete.
The primary risk to people from Hudson River PCBs is the consumption of PCB-contaminated
fish. As natural recovery of the river continues, human exposure to PCB-contaminated fish will
continue to be controlled through fishing restrictions and fish consumption advisories issued by
New York State. As discussed in Appendix 13 of the FYR report, NYSDOH has a Hudson River
Fish Advisory Outreach Program, which specifically targets its communication to high-risk
populations, such as women, children, and low-income citizens and works to developed specific
strategies for reaching out to potentially underrepresented or informed communities in both the
Upper and Lower Hudson River. Informing people about the New York State fishing restrictions
150
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
and advisories is primarily the responsibility of NYSDOH. EPA continues to coordinate with
NYSDOH to ensure that the state's fish advisory information is integrated into project
informational materials, discussed during public meetings and presented on the EPA's project
webpage. Updates on the status and progress of the NYSDOH's Hudson River Fish Advisory
Outreach Program are also presented to the site's CAG periodically.
Pursuant to the Consent Decree between GE and EPA, GE has contributed $4 million to Health
Research, Inc., of Rensselaer, New York, in order to support the State's implementation of
appropriate fish consumption advisories and fishing restrictions. Much of the outreach conducted
as part of the Hudson River Fish Advisory Outreach Program focuses on informing the community
about the risks from high PCB concentrations in fish, strategies to reduce exposure to PCBs during
fish consumption, and the recommended frequency of consumption of Hudson River fish.
NYSDOH staff who conduct outreach also provide advice to anglers on alternate waters near the
Hudson River that are safer in terms of fish consumption.
The Hudson River Fish Advisory Outreach Program uses various outreach strategies that include
distribution of written and electronic materials, partnerships, and a presence at community events
and public venues to achieve its objectives. NYSDOH fish advisory outreach work has been
conducted in partnership with other state and local agencies.
EPA understands the challenges faced by NYSDOH regarding informing the public about fish
consumption and the importance of the Outreach Program to reducing human exposure to
contaminated fish. EPA will continue to coordinate with NYSDEC and NYSDOH on the
implementation of the outreach program and to identify potential additional and/or more effective
outreach techniques into the future.
3.6.3 Comment 23: EPA should review all the data when developing the Five-Year Review
report in accordance with the guidance
Comment
Commenters state that EPA should follow its own guidance and include credible data and analyses
that are independently verified and peer reviewed, including those conducted by NYS and federal
agencies, in its FYR. They state it is imperative that the FYR process be conducted in the most
expeditious manner possible, and that the study include a comprehensive, independent, and
objective analysis of all available data, including the NOAA analysis, and an opportunity for full
participation by the NYSDEC, NYSDOH, the federal natural resource trustees, and other
interested stakeholders.
Commenters claim that EPA's draft FYR report of the Site lacks clear metrics to evaluate the
success or failure of the cleanup, and without clear metrics, the public is left in the dark as to how
EPA compared current conditions with the 2002 ROD expectations to reach its conclusion that the
remedy will be protective. Therefore, EPA should identify and list the criteria that it used to
evaluate the performance of the remedy in the FYR, as well as the criteria that the agency will use
for subsequent reviews. This should lead to a fair consideration of all relevant targets, not a
selective view of only the targets that are being met.
151
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Response
EPA's 2002 remedy selection for the Upper Hudson River (UHR) explicitly relied on two separate
elements: first, the very extensive dredging project, covering almost 500 acres and involving
removal of a large volume of PCB-contaminated sediment; and second, natural recovery with
extensive monitoring, predicted to take more than five decades.
The ROD also identifies objectives for the cleanup, including the reduction in fish tissue PCB
concentrations, because the main threat to people's health (and the health of other animals) from
PCB contamination in the Hudson River is through fish consumption. EPA's overarching approach
was to significantly improve the rate of fish recovery by removing sediment (with limited capping)
so that the river could recover quicker than by natural recovery alone.
Computerized models were used to compare dredging options and estimate how long it would take
under each option to achieve the interim fish recovery targets and the long-term remediation goal.
The model runs extended for 55 years after the end of dredging. No dredging alternative, even the
most aggressive, was predicted to achieve EPA's goal for fish recovery (0.05 mg/kg of PCBs in
fish) within this time period, in the UHR as a whole. The EPA therefore laid out two interim targets
for the cleanup remedy. The first of these (0.4 mg/kg in fish) would allow people to consume one
fish meal every two months. The second (0.2 mg/kg) would allow people to consume one fish
meal every month.
EPA will measure success for the UHR dredging remedy by comparing the goals set in the ROD
with data gathered through an extensive program of water, sediment and fish monitoring. Fish are
collected twice each year, in spring and fall, from a specified series of locations throughout the
UHR and Lower Hudson River (LHR). Water quality data are collected weekly or monthly
depending on location. Sediment data will be collected every five years. It will take up to eight or
more years of fish tissue data to identify trends with a reasonable degree of scientific certainty.
EPA will continue to carry out FYRs into the future, which will consider all data including the
new data gathered since the previous review.
As mentioned above, EPA has and plans to collect an extensive amount of fish, water quality and
sediment data from the Hudson River. This FYR considered all available project data (e.g., fish,
water, sediment, air) through 2016. Data collected in 2016 reflects conditions less than a year after
completion of dredging and are still influenced by dredging-related impacts. EPA has considered
in the FYR the analysis provided by other agencies including NOAA and NYSDEC. Members of
these agencies were on the FYR review team and contributed to the meetings held to discuss the
project and progress of the FYR. EPA has evaluated the NOAA analyses mentioned by the
commenters and presented its findings at a FYR team meeting. EPA response to the NOAA
analysis is in EPA's White Paper titled Re-Visiting Projections of PCBs in Lower Hudson River
Fish Using Model Emulation -March 2016 (https://www3.epa.gov/hudson/pdf/
EPA%20White%20Paper%20-%20Responses%20to%20NQAA%20Manuscript.pdf), and further
supplemented in Appendix C of this document. EPA concluded that NOAA's claim that fish tissue
concentrations will not meet remedial goals until many decades longer than anticipated by EPA's
152
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
model forecasts is not supported by the data. NOAA's analysis did not reflect the breadth of proj ect
sediment and fish data. NOAA also did not complete an appropriate emulation model calibration.
EPA established the metrics to be used in the FYR at the beginning of the process and presented
them to the FYR team. During early FYR team meetings, team members asked specific questions
about the data to be used and approach for evaluation. EPA with its technical experts discussed
the data and approach to be used for the FYR. Following the issuance of the draft FYR report,
EPA held a follow up meeting with the FYR team where the data and analysis were further
explained and discussed. At each team meeting, EPA allowed time for full discussion of questions
and concerns of team members. The criteria being used in assessing the data are based on the EPA
guidance on conducting FYRs and are presented in Section 5 of the FYR report.
EPA disagrees that the FYR lacks clear metrics against which to evaluate the effectiveness of the
remedy. The remedial action objectives, remedial goals and fish PCB target concentrations all
serve as metrics for evaluating the remedy, and the FYR includes discussion of the available data
in relation to those metrics. Data collected to date (primarily fish tissue data) may still be impacted
by dredging related activities and more data is needed. However, actual conditions during dredging
did not and were not expected to match up in every way with conditions as understood when the
ROD modeling was conducted. Therefore, direct comparisons of observed fish tissue
concentrations to ROD forecasts need to be carefully considered. It should also be noted that
dredging started later than the model considered. Also short-term and localized increases and
subsequent decreases in fish tissue PCB concentrations were anticipated in the FS and ROD (and
observed between 2009 and 2016) were not directly reflected in the long-term fish tissue forecasts
presented in support of remedy selection. For these reasons, direct comparisons of observed data
to ROD forecasts need to be done carefully with the various factors taken into consideration.
In the Final FYR report EPA is deferring a final protectiveness determination because it has
determined that there are not yet sufficient years of post-dredging data available on which to
support making a protectiveness determination.
3.6.4 Comment 25: EPA should update the Community Involvement Plan
Comment
Commenters state that EPA is not performing adequate outreach to communities along the Hudson
River. While EPA has a Community Involvement Plan (CIP), it has not been updated since 2009
and was intended to guide activities through the completion of dredging. Now that dredging is
complete, EPA should revise the CIP to better address the ongoing risks associated with PCB
contamination that will continue for decades along the entire Hudson River Superfund Site.
Response
As commenters accurately noted, the most recent update to the CIP for the in-river dredging
portion of the cleanup was in 2009 and was intended to guide activities through the completion of
dredging. Under the Superfund program, the CIP lays out the approach and rationale for
community involvement efforts and activities throughout the Superfund cleanup process and is
153
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
typically prepared early in that process. The original CIP for the Hudson River PCBs Superfund
Site was developed in 2003, during the design portion of the cleanup, and was subsequently
updated in 2009, prior to the start of dredging. EPA is currently in the process of preparing the
CIP for the Upper Hudson River (UHR) floodplain component of the Superfund site, for which
GE currently is performing a remedial investigation under an administrative consent order with
EPA. While CIPs are not developed specifically for Five-Year Reviews (FYR), or to guide post-
cleanup outreach efforts, the CIP is a valuable resource when planning community involvement
activities during the FYR and for continued community engagement after cleanups are completed.
Based on public input received when the cleanup decision was made, EPA committed to
developing a comprehensive public involvement program to be employed throughout the design
and construction phases of the project. While the dredging component of the cleanup remedy is
now complete, EPA remains committed to keeping the public informed about future work,
including the long-term monitoring that will be conducted to track the recovery of the river over
time and any efforts that are initiated in the future to collect additional data in the Lower Hudson
River (LHR). Information will be available on the Hudson River PCBs site webpage and EPA will
continue to develop fact sheets and news releases related to elements of the work that are of
greatest interest to the community. EPA also plans to continue to participate in meetings of the
site's CAG, as requested, to provide project updates. CAG meetings are open to the public.
Some commenters noted that although the dredging has ended, information should continue to be
provided and available to all Hudson River communities, and particularly down river subsistence
fishing communities, regarding the risks associated with PCB contamination in the Hudson River.
The NYSDOH has primary responsibility for informing people about current New York State
fishing advisories. More information about New York State's Hudson River Fish Advisory
Outreach Project is discussed in Appendix 13 of the FYR. EPA will continue to coordinate closely
with New York State on the Hudson River Project including matters related to the fishing
restrictions and advisories.
3.6.5 Comment 39: Public Involvement in the Five-Year Review Process
Comment
Commenters had various concerns pertaining to stakeholder and public involvement in the FYR
process. Commenters requested a defined scope for FYR team members to provide input on topics
such as identifying objectives, a timetable for completing tasks, a scoping process that solicits
input from agencies and the public, and criteria for transparency in the process. A request to
extend the public comment period was also provided along with a request for EPA to respond to
any comments received in writing.
In addition, commenters were concerned that EPA only held two public information meetings
along the entire "197-mile stretch of the Hudson River Superfund Site, neither of which are located
in or near New York City." Commenters stated that the EPA should hold a public information
meeting in NYC regarding the Proposed Second Five-Year Review Report for the Hudson River
Superfund Site. Commenters also said that it is crucial that the local community, including those
154
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
along the Lower Hudson River, have the opportunity to hear directly from the EPA on this
proposed report and to have their own voices be heard.
Commenters also stated that EPA is conducting vastly more extensive community outreach at
similar Superfund sites. Commenters provided the following example: EPA Region 10 has held
more than eighty community outreach and engagement activities since 2012 regarding the Portland
Harbor Superfund Site, which is also contaminated with PCBs. There, EPA identified strategies
for reaching out to underrepresented communities in the region, had translators present at
meetings, and attended cultural events to promote greater community engagement.
Response
While the five-year review was underway, the EPA consistently indicated the Agency's
commitment to a transparent and inclusive five-year review process. While not required by law,
or the usual Superfund procedures, the EPA took the nearly unprecedented step of offering an
opportunity for the public to comment on the draft five-year review report.
Per the EPA guidance on conducting five-year reviews, EPA is expected to obtain input on the
review from multiple groups and agencies, including state-level agencies, community groups, and
other federal partners. For the Second Five-Year Review, EPA established a team that included
representatives from state and federal agencies, the Hudson River Natural Resource Trustees, and
representatives from the site's Community Advisory Group. The scope for the team members was
established at the first team meeting and team member responsibilities were identified and
discussed. Over the course of the review period, the five-year review team met 13 times to discuss
data and other relevant project information. In these meetings, EPA presented data being used in
the analyses for the five-year review and explained EPA's understanding of the data to date. EPA
also dedicated multiple meetings to receive input from team members on the analysis of the data,
the concerns and questions on the analyses being conducted, clarification on the protectiveness
determination, and to discuss the draft report with team members to assist in their development of
comments. EPA has shared all the data and, to the extent possible, the technical assessment
documents and related materials that were part of its decision-making on this five-year review.
In addition, EPA held workshops, open to the public, to discuss important aspects of the review
process and to discuss EPA's progress on the analyses conducted to date. EPA also reported to
the community advisory group at multiple meetings throughout the review process to update the
group as well.
The Proposed Five-Year Review report, including all the technical appendices, and a brief fact
sheet were made publicly available on the Hudson River PCBs site webpage
(www.epa.gov/hudson) along with information about how to submit written comments during the
public comment period.
On June 1, 2017, EPA issued the Proposed Second Five-Year Review Report and initiated a 30-
day public comment period. The public comment period was subsequently extended as requested
by the public to 90 days and concluded on September 1, 2017. Three public information meetings
were held at various locations in the project area during the comment period. One of the public
155
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
meetings was held in NYC as requested by the public. Approximately 2,000 comments were
received from the public, as well as State and Federal agencies, environmental groups, and elected
officials. All comments received were carefully considered in development of the Final Second
Five-Year Review Report.
IV. REFERENCES
Barclay, B. 1993. "Hudson River Angler Survey." Hudson River Sloop Clearwater, Inc.
Poughkeepsie, New York.
Bertazzi, PA, L. Riboldi, A. Pesatori, L. Radice, and C. Zocchetti. 1987. "Cancer mortality of
capacitor manufacturing workers." In Am J Ind Med. 1987; 11(2): 165-76.
Brown, DP. 1987. "Mortality of workers exposed to polychlorinated biphenyls-an update". In
Arch Environ Health. 1987;42(6):333-9.
Burkhard, L. 2009. Estimation of Biota Sediment Accumulation Factor (BSAF) from Paired
Observations of Chemical Concentrations in Biota and Sediment. U.S. Environmental Protection
Agency, Ecological Risk Assessment Support Center, Cincinnati, OH. EPA/600/R-06/047.
Butcher, J.B., T.D. Gauthier, and E.A. Garvey, 1998. Use of Historical PCB Aroclor
Measurements: Hudson River Fish Data. In Environmental Toxicology and Chemistry, Vol. 16,
No. 8, pp. 1618-1623.
Connelly, N.A., B.A. Knuth, and C.A. Bisogni. 1992. Effect of the Health Advisory changes on
Fishing Habits and Fish Consumption in New York Sport Fisheries. September. Cornell
University: Ithaca, NY.
Eastern Research Group, Inc. 2000. Report on the Peer Review of the Hudson River PCBs Human
Health Risk Assessment. Prepared for U.S. Environmental Protection Agency, Region II.
September, 2000. Final Report.
Eastern Research Group, Inc.. 2000. Report on the Peer Review of the Revised Baseline Modeling
Report for the Hudson River PCBs Superfund Site. Final. Prepared for U.S. Environmental
Protection Agency, Region II, Emergency and Remedial Response Division. EPA Contract No.
68-W6-0022, Work Assignment No. 4-12, May 10.
EPA. 1997. Phase 2 Report, Further Site Characterization and Analysis. Volume 2C - Data
Evaluation and Interpretation Report (DEIR), Hudson River PCBs Reassessment RI/FS. Prepared
for USEPA Region 2 and USACE, Kansas City District by TAMS Consultants, Inc., The Cadmus
Group, Inc., and Gradient Corporation. February 1997.
EPA. 2000a. Revised Baseline Modeling Report (RBMR), Hudson River PCBs Reassessment
RI/FS Volume 2D. Prepared for USEPA Region 2 and USACE, Kansas City District by TAMS
Consultants, Inc., Limno-Tech, Inc., Menzie-Cura & Associates, Inc., and Tetra Tech, Inc. January
2000, http://www.epa.gov/hudson/reports.htm (Last accessed: 02/10/2015)
156
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
EPA. 2000b. Phase 2 Report Further Site Characterization and Analysis Volume 2F - Revised
Human Health Risk Assessment Hudson River PCBs Reassessment RI/FS - Upper Hudson River
and Mid-Hudson River. Vol. 2F. November 2000.
EPA. 2000c. Phase 2 Report - Review Copy Further Site Characterization and Analysis Revised
Baseline Modeling Report, Hudson River PCBs Reassessment RI/FS, Volume 2D/ Book 1 of 4,
Fate and Transport Models. Prepared for EPA and USACE by TAMS Consultants, Inc. and
Menzie-Cura & Associates, Inc. January 2000.
EPA. 2000d. Hudson River PCBs Reassessment RI/FS Phase 3 Report: Feasibility Study. Book 1
of 6: Report Text, Book 2 of 6: Figures and Tables, Book 5 of 6: Appendix D through Appendix H.
Prepared by TAMS Consultants, Inc., for EPA and USACE, December 2000.
EPA. 2001. Comprehensive Five-Year Review Guidance. EPA 540-R-01-007. OSWER No.
9355.7-03B-P.
EPA. 2002a. Record of Decision for the Hudson River PCBs Superfund Site. U.S. Environmental
Protection Agency. February 2002.
EPA. 2002b. Responsiveness Summary Hudson River PCBs Site Record of Decision. January
2002. Prepared for EPA Region 2 and USACE Kansas City District by TAMS Consultants, Inc.
EPA. 2010. Attachment E to Statement of Work, Operations, Maintenance and Monitoring Scope
for Phase 2 of the Remedial Action. December 2010.
EPA. 2012a. First Five-Year Review Report for Hudson River PCBs Superfund Site. June 1, 2012.
EPA. 2012b. Memorandum: Clarifying the Use of Protectiveness Statements for Comprehensive
Environmental Response, Compensation, and Liability Act Five-Year Reviews (OSWER 9200.2-
111). From James E. Woolfort, Director, Office of Superfund Remediation and Technology
Innovation and Reggie Cheatham, Director, Federal Facilities Restoration and Reuse Office, to
National Superfund Program Officers, Region 1-10. September 13, 2012.
https:// semspub. epa. gov/work/HQ/174829.pdf
EPA. 2014. Scoping and Problem Formulation for the Identification of Potential Health Hazards
for the Integrated Risk Information System (IRIS) Toxicological Review of Ethylbenzene
[CASRN 100-41-4], https://nepis.epa.gov/Exe/ZvNET.exe/P100L2BQ.TXT?ZvActionD=
ZyDocument&Client=EPA&Index=201 l+Thru+2015&Docs=&Querv=&Time=&EndTime=&S
earchMethod= 1 & T ocRestri ct=n& T oc=& T ocEntry=& QFi el d=& OFi el d Year=& OF i eldMonth=&
OFieldDav=&IntOFieldOp=0&ExtOFieldOp=0&XmlOuerv=&File=D%3A%5Czyfiles%5CInde
x%20Data%5C 1 lthrul 5%5CTxt%5C00000013%5CP 1 OOL2BO.txt&User=ANONYMOUS&Pa
ssword=anonvmous&SortMethod=h%7C-&MaximumDocuments=l&FuzzyDegree=0&Image
Oualitv=r75g8/r75g8/xl50vl50gl6/i425&Displav=hpfr&DefSeekPage=x&SearchBack
=ZvActionL&Back=ZvActionS&BackDesc=Results%20page&MaximumPages=l&ZyEntrv=l
& SeekPage=x&ZyPURL
157
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
EPA. 2016a. Responses to NOAA Manuscript Entitled: "Re-Visiting Projections of PCBs in
Lower Hudson River Fish Using Model Emulation" (Field, Kern, Rosman, 2015).
https ://www3. epa. gov/hudson/pdf/EP A%20White%20Paper%20-
%20Responses%20to%20NQAA%20Manuscript.pdf.
EPA. 2017a. PCB Data Collection and Analyses with Sediment Grain Size Distribution: Results
of Analyses of Sediment Samples Collected During EPA Oversight of General Electric
Contractors Performing 2016 Surface Sediment Data Collection. Data Collected Fall 2016 and
Analyzed January 2017. Prepared by Axys Analytical Services and Louis Berger for USEPA
(Filename: 2017-01-30 FINAL Hudson HRFO 20-16 - Master Database [validated].xlsx). January
31, 2017.
EPA. 2017b. Proposed Second Five-Year Review Report for Hudson River PCBs Superfund Site.
May 31, 2017.
EPA. 2019a. Technical Memorandum - Evaluation of 2016 EPA/GE and 2017 NYSDEC Surface
Sediment Data, Hudson River PCBs Superfund Site. Prepared for USEPA Region 2 and USACE
Kansas City District by Louis Berger, U.S., Inc. and Kern Statistical Services. January 2019.
(https://www3.epa.gov/hudson/).
EPA. 2019b. Special Study of Black Bass Fillet Tissue With and Without Ribs. Prepared for
USEPA Region 2 and USACE, Kansas City District by Louis Berger, U.S., Inc. and Kern
Statistical Services, Inc. January 2019. (https://www3.epa.gov/hudson/).
EPA and Wisconsin Department of Natural Resources (WDNR). 2009. Lower Fox River
Remedial Design Percent Design Report for 2010 and Beyond Remedial Actions. Volume 2 of 2.
Appendix I: Long-Term Monitoring Plan, November 2009.
Farley, K. J., R. V. Thomann, T. F. I. Cooney, D. R. Damiani and J. R. Wands. 1999. An Integrated
Model of Organic Chemical Fate and Bioaccumulation in the Hudson River Estuary.
Riverdale, NY, Report to The Hudson River Foundation: 170.
Field, L.J., J.W. Kern, and L.B. Rosman. 2016. "Re-visiting projections of PCBs in Lower Hudson
River fish using model emulation." Science of Total Environment 557-558 (2016) pp. 489-501.
(http: //dx. doi. or g/10.1016/i. scitotenv .2016.02.072).
GE. 2002. Design Support Sediment Sampling and Analysis Program, Field Sampling Plan.
Prepared for General Electric Company by Quantitative Environmental Analysis, LLC. (QEA).
April 2002.
GE. 2006. Corrective Action Memorandum (CAM) No. 6 - Modification to the Daily Far-Field
Monitoring Schedule. Memorandum from John Roche, Anchor QEA, to Bob Gibson, General
Electric. August 12, 2013.
158
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
GE. 2007. Hudson River PCBs Site Phase 2 Dredge Area Delineation Report. Prepared for General
Electric Company by QEA, LLC. December 17, 2007.
GE. 2008. CAM No. 14 - Hudson River Phase 2 Remedial Action Monitoring Program -
Analytical Procedures for Water Samples Collecting at Beginning of 2017 through Start of
OM&M. Memorandum from Meg Michell, Environmental Standards, Inc., to Bob Gibson,
General Electric. May 17, 2017.
GE. 2011. Hudson River PCB's Site 2011 Remedial Action Monitoring Quality Assurance Project
Plan (RAMQAPP). Prepared by Anchor QEA (Glens Falls, NY) in Conjunction with
Environmental Standards, Inc. (Valley Forge, PA). Appendix 3.8-4 (SOP for the Tissue
Reduction/Grinding for Whole Body and Filleted Fish, August 2011).
Huber, P.J. 1973. Robust Regression: Asymptotics, Conjectures and Monte Carlo. Annals of
Statistics, 1, 799-821. http://dx.doi.org/10.1214/aos/1176342503
Huber, P. J. 1981. Robust Statistics. John Wiley and Sons, New York.
Hudson River Phase 2 Remedial Action Monitoring Program - Analytical Procedures for Water
Samples Collecting at Beginning of 2017 through Start of OM&M. Memorandum from Meg
Michell, Environmental Standards, Inc., to Bob Gibson, General Electric. May 17, 2017.
HydroquaL. 2007. A Model for the Evaluation and Management of Contaminants of Concern in
Water, Sediment, and Biota in the NY/NJ Harbor Estuary, Contaminant Fate & Transport &
Bioaccumulatoin Sub-models, July 2007 http://carpweb.org/modeling/Modeling reports?
CFTB.pdf
Lenth, R. V. 2009. Java Applets for Power and Sample Size [Computer
software]. Retrieved January 29, 2014, from http://www.stat.uiowa.edu/~rlenth/Power.
Lilliefors, H. W. "On the Kolmogorov-Smirnov test for normality with mean and variance
unknown." Journal of the American Statistical Association. Vol. 62, 1967, pp. 399-402.
Neter, J., Kutner, M. H., Nachtsheim, C. J., and Wasserman, W., 1996. Applied Linear Statistical
Models, 4thEdition, Richard D. Irwin, Inc., Burr Ridge, Illinois. NYSDEC 2005. Of Time, PCB's
and the Fish of the Hudson River. NY State Department of Environmental Conservation, Division
of Fish, Wildlife, and Marine Resources. Albany, NY, July 2005, Appendix 1 (Fish Collection,
Sample Preparation, and Analytical Procedures, page 259).
NYSDEC. 2012. Cumberland Bay Sludge Bed Wilcox Dock Site # 5-10-017 Removal and
Disposal Project Pre-to Post-Dredging Monitoring (Volumes I ofII and II ofII) Ten Year Review.
Prepared by AECOM, Latham, NY. June 2012.
NYSDOH. 2000. 1996 Survey of Hudson River Anglers - Hudson Falls to Tappan Zee Bridge at
Tarrytown, New York. Final Report, February 10, 2000.
Rodenburg, L.A. and D.K. Ralston. 2017 Historical sources of poly chlorinated biphenyls to the
sediment of the New York/New Jersey Harbor.
159
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Safe. S.H. 1994. "Polychlorinated Biphenyls (PCBs) Environmental Impact, Biochemical and
Toxic Responses, and Implications for Risk Assessment". Critical Reviews in Toxicology, 24, 87-
149. http://dx.doi.org/10.3109/104084494090493Q8
Sanders, J.E. 1989. PCB-pollution problem in the Upper Hudson River: From environmental
disaster to environmental gridlock.' Northeastern Env. Sci. 8(1): 1-86.
Sinks, T., G. Steele, AB Smith, K. Watkins, RA Shults. 1992. "Mortality among workers exposed
to polychlorinated biphenyls." In American Journal of Epidemiology. 1992;136(4):389-98.
Sloan, R.J., M.W. Kane and L.C. Skinner. 2002. 1999 as a Special Spatial Year for PCBs in
Hudson River Fish. Bureau of Habitat, Division of Fish, Wildlife and Marine Resources, New
York State Department of Environmental Conservation, Albany, New York. 111 p.
Sloan, Ron. 1999. Briefing on 1997 Striped Bass PCB Results. Memorandum from Ron Sloan to
James Colquhoun. February 11, 1999. https://semspub.epa.gov/work/02/68407.pdf
Sloan et al. 2005. "Of Time, PCBs and the Hudson River." Ronald J. Sloan, Michael W. Kane,
and Lawrence C. Skinner. Bureau of Habitat, Division of Fish, Wildlife and Marine Resources,
NYSDEC, Albany, New York. July 2005.
Thomann, R. F., J. A. Mueller, R. P. Winfield and C. R. Huang. 1989. Mathematical model on the
long-term behavior of PCBs in the Hudson River Estuary. New York, NY, Hudson River
Foundation.
Wendt, M.E. 1986. Low Income Families' Consumption of Freshwater Fish Caught From New
York State Waters. Thesis, Cornell University. August.
White. H. 1980. "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test
for Heteroskedasticity". Econometrica. 48 (4): 817-838
160
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Final Second Five-Year Review Comment
Response for the
Hudson River PCBs Superfund Site
APPENDIX A
LIST OF COMMENTERS (INDEX)
-------
Government, Agencies, Organizations and
Businesses/Corporations
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
April 2019
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agencv/Organi/alion
First .Name
l.asl .Name
Division
Title/Uole
Dale
submitted
Federal and Stale Government
1
United States Senate
Kirsten
Gillibrand
United States Senator
6/7/2017
2
United States Senate
Charles
Schumer
United States Senator
7/18/2017
Kristen
Gillibrand
United States Senator
3
New York State Assembly
Didi
Barrett
106th District
Assemblymember
8/28/2017
4
New York State Assembly
Didi
Barrett
106th District
Assemblymember
6/28/2017
5
New York State Assembly
Ellen
Jaffee
97th District
Assemblymember
6/7/2017
6
New York State Senate
David
Carlucci
38 th District
State Senator
8/30/2017
Terrance
Murphy
40th District
State Senator
Martin J.
Golden
22nd District
State Senator
Marisol
Alcantara
31st District
State Senator
Jesse
Hamilton
20th District
State Senator
7
New York State Senate
Brad
Hoylman
27th District
State Senator
8/9/2017
8
New York State Senate
Liz
Krueger
28th District
State Senator
7/19/2017
Carrie
Woerner
113th District
Assemblymember
Joseph P.
Addabbo, Jr.
15th District
State Senator
Jamaal
Bailey
36th District
State Senator
Brian
Benjamin
30th District
State Senator
John E.
Brooks
8 th District
State Senator
Leroy
Comrie
14th District
State Senator
Martin
Malave
Dilan
18 th District
State Senator
George
Latimer
37th District
State Senator
Kevin S.
Parke
21st District
State Senator
Jose
Peralta
13 th District
State Senator
Gustavo
Rivera
33rd District
State Senator
James
Sanders, Jr.
10th District
State Senator
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agencv/Organizalion
I'irsl .Name
I.asl Name
Division
Tillc/Kolc
Dale
submitted
Jose M.
Serrano
29th District
State Senator
Thomas J.
Abinanti
92nd District
Assemblymember
Didi
Barrett
106th District
Assemblymember
Kevin A.
Cahill
103rd District
Assemblymember
Jeffrey
Dinowitz
81st District
Assemblymember
Anthony
D'Urso
16th District
Assemblymember
Patricia
Fahy
109th District
Assemblymember
Sandra R.
Galef
95th District
Assemblymember
Deborah J.
Glick
66th District
Assemblymember
Richard N.
Gottfried
75th District
Assemblymember
Pamela J.
Hunter
128th District
Assemblymember
Ellen
Jaffee
97th District
Assemblymember
Brian P.
Kavanagh
74th District
Assemblymember
William
Magee
121st District
Assemblymember
Shelley
Mayer
90th District
Assemblymember
John T.
McDonald, III
108th District
Assemblymember
Yuh-Line
Niou
65th District
Assemblymember
Daniel
O'Donnell
69th District
Assemblymember
J. Gary
Pretlow
89th District
Assemblymember
Linda B.
Rosenthal
67th District
Assemblymember
Nily
Rozic
25th District
Assemblymember
Rebecca A
Seawright
76th District
Assemblymember
Jo Anne
Simon
52nd District
Assemblymember
Dan
Stec
114th District
Assemblymember
Fred W.
Thiele
1st District
Assemblymember
Mary Beth
Walsh
112th District
Assemblymember
Jaime R.
Williams
59th District
Assemblymember
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agency/Organization
I'irsl .Name
I.asl Name
Division
Titlc/Uolc
Dale
submitted
Kenneth P
Sel">row ski
District
AssemMx member
i)
New York State Senate
.lose
Peralta
1 .illi District
State Senator
S 31 2() 17
Agencies
10
National Oceanic and
Atmospheric Administration
Thomas
Brosnan
Hudson River Case
Manager
9/1/2017
11
National Oceanic and
Atmospheric Administration
Jay
Field
9/1/2017
Lisa
Rosman
12
New York State Bridge
Authority
Joseph
Ruggiero
Executive Director
6/28/2017
13
New York State Department
of Environmental
Conservation
Kevin
Farrar
9/1/2017
14
New York State Department
of Environmental
Conservation
Basil
Seggos
Commissioner
6/7/2017
15
New York State Department
of Environmental
Conservation
Basil
Seggos
Commissioner
8/30/2017
16
New York State Office of the
Attorney General
Maureen
Leary
Assistant Attorney
General
9/1/2017
James
Wood
Assistant Attorney
General
Brittany
Haner
Assistant Attorney
General
John D.
Davis
Environmental Scientist
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agencv/Organi/alion
I'irsl .Name
l.asl Name
Division
Title/Uole
Dale
submitted
l.ocal Government
17
Albany County
Daniel
McCoy
County Executive
9/1/2017
Rockland County
Edwin J.
Day
County Executive
Dutchess County
Marcus J.
Molinaro
County Executive
Ulster County
Michael P.
Hein
County Executive
Orange County
Steven M.
Neuhaus
County Executive
Westchester County
Robert P.
Astorino
County Executive
18
Columbia County
Environmental Management
Council
Edwin
Simonsen
Chair
8/31/2017
19
Dutchess County
Marcus
Molinaro
Dutchess County
Executive
6/28/2017
20
Dutchess County Regional
Chamber of Commerce
Frank
Castella Jr.
President and CEO
8/16/2017
21
Kingston Conservation
Advisory Council
Julie
Noble
Chair
8/31/2017
Elizabeth
Broad
Lorraine
Farina
Emilie
Hauser
Lynn
Johnson
Kevin
McEvoy
Casey
Schwarz
22
Town of Saratoga
Thomas N.
Wood III
Supervisor
8/21/2017
23
Town of Saugerties
Greg
Helsmoortel
Supervisor
8/29/2017
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agency/Organization
First .Name
l.asl Name
Division
Title/Uole
Dale
submitted
24
Town of Sluyvesant Town
Board
Melissa
jNaegeli
Town Clerk
8/10/2017
Ed
Scott
Councilman
Tom
Burrall
Councilman
Brian
Chittenden
Councilman
Kelley
Williams
Councilwoman
Ron
Knott
Supervisor
25
Ulster County Environmental
Management Council
Dave
Haldeman
Chair
8/31/2017
26
Village of Schuylerville
Dan
Carpenter
Mayor
9/1/2017
27
Village of Schuylerville
Dan
Carpenter
Mayor
9/1/2017
28
Westchester County
Robert
Astorino
County Executive
8/28/2017
Organizations
29
Catskill Mountainkeeper
Kathleen
Nolan
Senior Research
Director
9/1/2017
30
The Chamber of Southern
Saratoga County *Signatories
include organizations and
businesses
9/1/2017
Mechanicville-Stillwater
Chamber of Commerce
Barbara A.
Corsale
President
NYS Building and
Construction Trades Council
James
Cahill
President
Dutchess County Regional
Chamber of Commerce
Frank M.
Castella, Jr.
President & CEO
R. L. Baxter Building
Corporation
Robert
Baxter
Owner
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agency/Organization
First .Name
l.asl Name
Division
Tillc/Kolc
Dale
submitted
Schuylerville Area Chamber
of Commerce
Maria
Hodge
President
McDonald's REAAL, Inc.
Roger E.
Grout
President
The Chamber of Southern
Saratoga
Pete
Bardunias
President & CEO
Elyse Harney Real Estate
Elyse D.
Harney
Principal Broker/Owner
Local Union 21
Ron
Diaz
Business Agent
Plumbers and Steamfitters
HVACR
Thomas
Carey
Business Agent
Walkway Over the Hudson
Elizabeth
Waldstein-Hart
Executive Director
Poughkeepsie Alliance
Paul
Calogerakis
Chairman
Hudson Development
Corporation
Sheena
Salvino
Executive Director
American Towns
Ted
Buerger
Chairman
Dutchess Community College
Pamela
Edington, Ed.D
President
The Business of Your
Business
Wiley
Harrison
Owner
Rbeach & Bartolo Realtors
Victor
Mendolia
Associate Real Estate
Broker
Finance& Corporate
Development Omnicom
Group
John
Hamilton
Vice President
Bryant Rabbino LLP
Kim
Taylor
Of Counsel
Saugerties Lighthouse
Patrick
Landewe
Keeper
Bonura Hospitality Group
Joe
Bonura
Principal
IKOR - Life Care
Management Solutions
James
Sullivan
President & Managing
Director
Consigli Construction, NY
Gregory
Burns
President
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agency/Organization
I'irsl .Name
l.asl .Name
Division
Tillc/Kolc
Dale
submitted
Meyer Contracting
Corporation
Christian W.
Meyer
President
Putnam Market
Catherine
Hamilton
President
Obercreek Farm LLC
Alex
Reese
Owner
Ugly Rooster Cafe
Ariel
Pagan
Owner
Northshire Bookstore
Chris
Morrow
Co-Owner
Five Porch Farms
Dan
Lundtquist
Owner
Healthy Living
Eli
Lesser-Goldsmith
Co-owner and General
Manager
H H Hill Realty Services, Inc.
Harry
Hill
Principal Broker
National Resources, Inc.
Joseph
Cotter
CEO
Green Conscience Home &
Garden
Karen
Totino
Licensed Real Estate
Salesperson
Peak Magazine
Kellie
McGuire
Owner
Hudson River Cruises
Kevin
Buckel
General Manager
Spath Counseling Services
Kevin
Spath
Owner
Landscape Architects, P.C.
Kim
Mathews, RLA,
FASLA
Principal
Kit Burke-Smith Jewelry
Kit
Burke-Smith
Owner
Storm King Adventure Tour
Kris
Seiz
Owner
Mohawk Maiden Cruises,
LLC
Mara Hodge
&
Maria Saavedra
Owners
Dutchess Tourism Inc.
Mary Kay
Verba
President & CEO
Bellefield Development
Partners, LLC
Michael
Oates
Managing Partner
Fusion Lab, Inc.
Alon
Koppel
Partner
Jeffrey Russell Werner, LLC
Jeffrey
Russel
Werner, Esq
Attorney
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agency/Organization
I'irsl .Name
l.asl .Name
Division
Tifle/Uole
Dale
submitted
Spatial Dynamics
Jaime
McMillian
Founder
Arts Center on Hudson
Jaime
McMillian
Founder
Growler and Grill
Mike
Fitzgerald
Owner
Saratoga Apple, Inc.
Nathan
Darrow
Owner
Gardening Angels
Peggy
Fusco
Owner
Alisson Spears AIA
Alison
Spears
Chip
Lowenson
Daniel
Kramer
Mary W. Harriman
Foundation
David H.
Mortimer
President
David Redden, LLC.
David
Redden
Director
Deco Works Ltd.
Evan Mason
and
Garrard Beeney
Principals
Gary
Glynn
Hoke
Slaughter
Jay
Saunders
James
Goodfellow
Julia
Widowson
Kristin
Flood
Leigh
Seippel
United Catalyst, LLC.
Maijorie
Hart
Acting CEO
Land Trust Alliance
Michael P.
Dowling
Immediate Past Chair
Dillion, Ready & Co, Inc.
Ned
Whitney
Retired Managing
Director
Richard
Klapper
Pierpoint Capital
Richard
Krupp
Managing Partner
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agency/Organization
I'irsl .Name
l.asl .Name
Division
Tifle/Uole
Dale
submitted
Debevoise & Plimpton LLP
Sara A. Q.
Fitts
31
Hudson River Fishermen's
Association
Gil
Hawkins
Vice President
6/15/2017
32
Riverkeeper, Inc.
Jeremy
Cherson
Campaign Advocacy
Coordinator
7/7/17
33
Riverkeeper, Inc.
Jeremy
Cherson
Campaign Advocacy
Coordinator
7/19/17
34
Riverkeeper, Inc.
Richard
Webster, Esq
6/5/2017
Hudson Fishermen's
Association
Gil
Hawkins
Natural Resources Defense
Council
Daniel
Raichel, Esq
Scenic Hudson, Inc.
Althea
Mullarkey
Hudson River Sloop
Clearwater, Inc.
Manna Jo
Greene
35
Riverkeeper, Inc.
Richard
Webster
Legal Director
6/16/2017
36
Saratoga Unites
Environmental Action
Committee
Julie
Wash
8/31/2017
37
Scenic Hudson, Inc.
Hay ley
Carlock
Director of
Environmental
Advocacy
9/1/2017
Hudson River Fishermen's
Association
Gil
Hawkins
Riverkeeper, Inc.
Richard
Webster, Esq
Hudson River Sloop
Clearwater, Inc.
Manna Jo
Green
Sierra Club, Atlantic Chapter
Roger
Downs
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agency/Organization
First .Name
l.asl Name
Division
Title/Uole
Dale
submitted
Natural Resources Defense
Council
Mark
Izeman
38
Scenic Hudson, Inc.
Hay ley
Carlock
Director of
Environmental
Advocacy
9/1/2017
Hudson River Fishermen's
Association
Gil
Hawkins
Riverkeeper, Inc.
Richard
Webster, Esq
Hudson River Sloop
Clearwater, Inc.
Manna Jo
Green
Sierra Club, Atlantic Chapter
Roger
Downs
Natural Resources Defense
Council
Mark
Izeman
39
Society of Saint Ursula
Kathleen
Donnelly
8/15/2017
40
The Historic Hudson - Hoosic
Rivers Partnership
Tom
Richardson
Partnership Chairperson
8/31/2017
41
Walkway Over the Hudson
Elizabeth
Waldstein-Hart
Executive Director
8/31/2017
42
Hudson River Sloop
Clearwater, Inc. Petition
* Petition with 503 signatures
8/22/2017
43
Hudson River Sloop
Clearwater, Inc. Petition
*Petition with 150 signatures
8/28/2017
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Government, Agencies, Organizations and Businesses/Corporations
EPA
Index
N il in her
Agencv/Organi/alion
I'irsl .Name
l.asl Name
Division
Tillc/Kolc
Dale
submitted
Businesses/Corporations
44
Bonura Hospitality Group
Joseph
Bonura Jr.
Owner
8/1/2017
45
ecoSPEARS
Ian
Doromal
Vice President
9/1/2017
46
General Electric
John
Haggard
Global
Remediation;
Global
Operations,
Environmental,
Health & Safety
Leader
9/1/2017
47
Hudson Development
Corporation
Sheena
Salvino
Executive Director
8/31/2017
48
Mohawk Maiden Cruises
Maria
Hodge
Master Captain, Owner
8/30/2017
49
Seaweed Yacht Club; Hudson
River Boat & Yacht Club
Association
Janice
Anderson
Commodore; Director
8/28/2017
50
The Business of your
Business
Wiley
Harrison
Owner
8/7/2017
51
United Campus Holdings
Company, LLC
Wayne
Senecal
President and CEO
Emeritus
7/19/2017
-------
Individuals
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site April 2019
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irst Name
Last Name
Dale Submitted
I nique Submittals
52
Patricia
Aakre
7/24/17
53
Emm
Ache
8/30/17
54
Claudia
Ackerman
8/21/17
55
Jeff
Adams
9/1/17
56
Sam
Adels
8/14/17
57
Deborah
Adler
8/21/17
58
Joanna
Albertson
8/28/17
59
Tomara
Aldrich
9/1/17
60
Elizabeth
Allee
6/5/17
61
Richard
Allen
8/21/17
62
Suzanne
Allen
8/21/17
63
Roland
Alley
8/21/17
64
Thomas
Amisson
8/28/17
65
Mary
Andrews
9/1/17
66
Anonymous
Anonymous
7/25/17
67
Anonymous
Anonymous
8/21/17
68
Emi
Araki
8/29/17
69
Patricia
Arcuri
8/21/17
70
A1
Arioli
8/21/17
71
Dwight
Arthur
6/6/17
72
Tom
Artin
7/24/17
73
Judith
Asphar
8/24/17
74
Doris
Bachmann
9/1/17
75
Talya
Baharal-Gnida
8/21/17
76
Patrick
Bailey
8/21/17
77
Eric
Baker
8/21/17
78
Marni
Bakst
8/21/17
79
Kathryn
Barry
7/7/17
80
Scott
Basal
8/21/17
81
Susan
Basu
9/1/17
82
Bill
Bates
8/10/17
83
Cari
Bates
8/21/17
84
Alex
Beauchamp
8/29/17
85
Laurel
Becker
8/28/17
86
Andrew
Bell
8/29/17
87
Ros
Bell
8/21/17
88
Sandra
Bensalah
8/21/17
89
Lisa
Berry
8/21/17
90
Ryan
Blum
8/29/17
91
Cora
Bodkin
8/4/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
92
Betty
Boomer
7/26/17
93
Jon
Bowermaster
8/29/17
94
Danielle
Brecker
8/28/17
95
Nancy
Breen
8/21/17
96
Claire
Briguglio
8/9/17
97
Kristin
Brown
8/21/17
98
Helene
Browning
9/1/17
99
Ronda
Brunsting
7/28/17
100
John
Buckley
8/31/17
101
Tom
Buckner
6/16/17
102
Tom
Buckner
8/21/17
103
David
Budd
8/31/17
104
Ted
Buerger
8/25/17
105
Jack
Burke
7/28/17
106
Linda
Burke
8/21/17
107
Sanford
Bush
8/31/17
108
Brenda
Campbell
8/21/17
109
Alyssa
Carbone
8/21/17
110
Valerie
Carlisle
7/5/17
111
Arthur
Carlucci
8/29/17
112
Miani
Carnevale
8/29/17
113
Jeremy
Carpenter
8/21/17
114
Jay
Cartagena
9/1/17
115
Brian
Caserto
8/21/17
116
Thomas
Cathcart
8/21/17
117
Dana
Chaifetz
5/30/17
118
Gwendolyn
Chambers
8/2/17
119
Martha
Cheo
6/17/17
120
Jeremy
Cherson
8/2/17
121
Jean
Chung
8/30/17
122
CD.
Clark
7/19/17
123
Lawrence
Clarke
8/21/17
124
Blythe
CI ark-McKitri ck
8/31/17
125
Stephen
Cluskey
6/5/17
126
Nora
Cofresi
8/25/17
127
Nancy
Colas
8/29/17
128
Jon
Cole
8/25/17
129
Kelly
Collins
7/28/17
130
Daniel
Convissor
8/25/17
131
Jennifer
Convissor
8/28/17
132
James
Corcoran
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
133
Isabel
Cotarelo
8/21/17
134
Kyle
Cottier
8/29/17
135
Linda
Coupart
7/9/17
136
Michael and Reva
Cowan
9/1/17
137
Caroline
Craig
8/29/17
138
Patrick
Cunningham
7/28/17
139
Lawrence
Curtin
8/22/17
140
Nancy
Cutler
8/29/17
141
Caroline
Cutroneo
6/6/17
142
Peter
Cutul
9/1/17
143
Tara
DAndrea
8/29/17
144
Roya
Darling
8/21/17
145
D
Darvie
7/28/17
146
George
Dashnaw
8/30/17
147
Eileen
de Munck
9/1/17
148
Margaret
Dean
8/21/17
149
Susan
Deane-Miller
8/21/17
150
Eva
Deitch
8/29/17
151
Darin
DeKoskie
6/28/17
152
Victoria
Delgado
7/25/17
153
OA
Dell
8/21/17
154
Alex
DeRosa
6/21/17
155
Jim
Desmond
8/24/17
156
Yvonne
Devlin
8/21/17
157
Frank & Joan
DiChiaro
8/22/17
158
Joanna
Dickey
9/1/17
159
Rita
Dixit-Bubiak
7/28/17
160
Jennifer
Dob son
8/22/17
161
Ron
Dombroski
6/10/17
162
Judy
Dong
9/1/17
163
Elke
D'Onofrio
9/1/17
164
Colleen
Dougherty
8/30/17
165
Ryan
Doyle
8/29/17
166
Jacquelyn
Drechsler
9/1/17
167
Jill
Dunay
8/21/17
168
Jake
Dunn
8/21/17
169
Rebecca
Dwyer
8/28/17
170
Jeff
Economy
8/30/17
171
Seth
Edelman
8/21/17
172
Jane
Ehrlich
8/29/17
173
Sarita
Eisenstark
8/21/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
174
Wallace
Elton
9/1/17
175
Katherine
Enberg
8/21/17
176
Cory
Ethridge
8/10/17
177
Mary
Evans
8/30/17
178
Russell
Faller
6/21/17
179
Russell
Faller
6/27/17
180
Russell
Faller
8/29/17
181
Armanda
Famiglietti
6/4/17
182
Peter
Farrell
8/21/17
183
Nina
Faver
8/21/17
184
Nancy
Felcetto
8/29/17
185
Roy
Felcetto
8/29/17
186
Deborah
Felder
8/29/17
187
Ricardo
Fernandez
8/22/17
188
Linda
Fernberg
8/25/17
189
Elvira
Ferrario
8/16/17
190
Mary
Fetherolf
8/21/17
191
Joe
Finan
8/30/17
192
Margaret
Finch
8/29/17
193
Rebecca
Finnell
8/30/17
194
John
Fisher
8/21/17
195
Lynn
Flanagan
7/19/17
196
Peter
Flanagan
9/1/17
197
Kristin
Flood
8/19/17
198
Patricia
Flood
8/21/17
199
Craig
Fogel
8/29/17
200
Bob, Marie
Foster
8/28/17
201
Marion
Foster
7/29/17
202
Tiffani
Francisco
8/29/17
203
Marcus
Frank
9/1/17
204
Florence Joan
Freeman
8/21/17
205
Linda; Chester
Freeman
8/18/17
206
Kate
Frizzell
8/25/17
207
Sharon
Gagne
8/21/17
208
Gail
Galitzine
8/29/17
209
Nancy
Gardner
8/21/17
210
Linda
Geary
8/21/17
211
Sheila
Geist
8/29/17
212
Sheila
Geist
8/30/17
213
Linda
Gerena
8/22/17
214
Ira
Gershenhorn
8/9/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
215
Jacquelyn
Gier
7/25/17
216
Steve
Gilman
6/2/17
217
Mary
Goddard
8/23/17
218
Nadine
Godwin
8/30/17
219
Steve
Gold
8/29/17
220
Patricia
Goldberg
8/22/17
221
Allan
Goldhammer
8/21/17
222
Freya
Goldstein
8/21/17
223
Karen
Goodman
6/6/17
224
Leslie
Gordon
7/25/17
225
Cindy
Gould
8/29/17
226
Nicole
Graf-Javery
8/23/17
227
Meryl
Greenblatt
8/22/17
228
Hannah
Greene
8/28/17
229
Rosalie
Griffith
8/22/17
230
Joan
Grishman
8/21/17
231
Daley
Gruen
8/29/17
232
Carol
Grunkemeyer
7/6/17
233
Robert
Grunkemeyer
7/6/17
234
Christine
Guarino
8/21/17
235
Michael
Gunderson
9/1/17
236
Mary
Gunter
8/21/17
237
Anne
Hager
8/28/17
238
Nancy
Hager
8/29/17
239
Christine
Hague
8/16/17
240
Emily
Hague
8/29/17
241
Paul
Hague
8/16/17
242
Brandon
Hakulin
8/21/17
243
Karen
Hall
8/21/17
244
Rhonni
Hallman
8/21/17
245
Mary
Hammett Stevenson
8/17/17
246
Martin
Hangarter
8/26/17
247
Terence
Hannigan
7/21/17
248
Beth
Hanson
8/31/17
249
Marc
Happet
9/1/17
250
Anne
Heaney
8/7/17
251
Anne
Heaney Johnson
8/17/17
252
Patricia
Heller
8/21/17
253
Irene
Herz
8/22/17
254
Jonathan
Herzog
9/1/17
255
Deborah
Highley
8/21/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
256
Annie
Hillay
7/25/17
257
Barbara
Hobens
8/21/17
258
Dana
Hoey
8/21/17
259
Miriam
Hoffman
8/4/17
260
Karin
Holloway
8/21/17
261
Timothy
Holmes
8/29/17
262
Wendy
Holtzman
8/18/17
263
Arlene
Holzman
7/19/17
264
Patrick
Hono
8/21/17
265
Joseph
Hope Jr.
8/21/17
266
Robin
Horowitz
8/29/17
267
Pat
Hughes
8/21/17
268
Carole
Hunt
8/22/17
269
David
Hupert
8/29/17
270
Ryan
Jafri
8/21/17
271
Ed
Jahn
8/26/17
272
Lee
Jamison
8/29/17
273
Lois
Janove
6/6/17
274
Susan
Johnson
8/22/17
275
Abigail
Jones
8/30/17
276
Justin
Jordak
8/21/17
277
Ellen
Jouret-Epstein
5/30/17
278
Christopher
Joy
9/1/17
279
Peter
Jung
8/4/17
280
Elissa
Jury
8/30/17
281
F. Michael
Kadish
7/11/17
282
Gloria
Kadish
8/7/17
283
Robert
Kalman
8/21/17
284
Sara
Kaminker
6/6/17
285
Carole
Kane
8/20/17
286
Edith
Kantrowitz
8/29/17
287
Edith
Kantrowitz
8/31/17
288
Nancy
Kaplan
8/29/17
289
Michelle
Karell
5/30/17
290
George
Katopis
8/21/17
291
Deb Peck
Kelleher
9/1/17
292
William
Kelleher
7/10/17
293
Laird
Kelly
8/16/17
294
Quinn
Kelly
8/21/17
295
Marci
Kenneda
8/24/17
296
John
King
8/21/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
297
Laurence
Kirby
8/21/17
298
Rachel
Kish
8/22/17
299
Cary
Kittner
8/21/17
300
Caroline
Klapproth
8/21/17
301
Amy
Kletter
8/29/17
302
Vladimir
Klimenko
8/30/17
303
Pete
Klosterman
8/29/17
304
J.
Knott
7/20/17
305
Wayne
Kocher
8/8/17
306
Susan
Koff
7/6/17
307
Laura
Kohlmann
8/22/17
308
Phil
Kovacs
8/21/17
309
Patricia
Kram
9/1/17
310
Pamela
Krimsky
8/28/17
311
Thomas
Kryzak
8/29/17
312
Peggy
Kurtz
8/22/17
313
A. Norman
Kvam
8/21/17
314
Marc
Lallanilla
8/31/17
315
Frank
Lancellotti
8/31/17
316
Barbara
Landa
7/25/17
317
Sasha
Langesfeld
7/25/17
318
Julie
Lappano
8/28/17
319
Michael
Laser
8/23/17
320
Judy
Lass
8/29/17
321
J. Eva
Lau
9/1/17
322
Robin
Laurita
8/22/17
323
Margaret
Leather
9/1/17
324
Patti
Lenseth
8/2/17
325
Jean
Leo
9/1/17
326
Esther
Light
9/1/17
327
David
Limburg
8/21/17
328
Hedvig
Lockwood
8/21/17
329
Elizabeth
LoGiudice
8/21/17
330
Skyler
Long
8/21/17
331
Albert and Doris
Lowenfels
8/29/17
332
Barbara
Lubell
8/21/17
333
David
Macaluso
8/25/17
334
Andrew
Maclnnes
8/31/17
335
Edward
Mack
8/21/17
336
Cathy
Mackey
8/22/17
337
Molly
MacQueen
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
338
Sarah
MacWright
8/21/17
339
Kevin
Magee
8/21/17
340
Tom
Mahoney
8/30/17
341
Tom
Mahoney
8/30/17
342
Barry
Maisel
7/17/17
343
Pamela
Malcolm
8/21/17
344
Lucy
Manning
7/21/17
345
Mickey
Marcella
6/9/17
346
Jeffrey
Marino
8/31/17
347
Jeffrey
Marino
9/1/17
348
Kate
Marriott
8/21/17
349
Daniel
Marshall III
8/21/17
350
Matthew
Martini
8/29/17
351
Kara
Masciangelo
8/22/17
352
Kara
Masciangelo
8/22/17
353
Kara
Masciangelo
8/29/17
354
Kara
Masciangelo
8/30/17
355
Janice
Mastromarchi
8/31/17
356
David
Mathis
8/28/17
357
Debra
Mathis
8/29/17
358
Anne
McCabe
8/21/17
359
Christa
McCauley
8/21/17
360
Nora
McDowell
9/1/17
361
Willis
McEckron
6/14/17
362
Susan
McGrath
8/21/17
363
Virginia
McGreevy
7/31/17
364
Grant
McKeown
8/30/17
365
Merry
McLoryd
9/1/17
366
Jaime
McMillan
8/29/17
367
Patrick
McMullan
8/29/17
368
Christopher
McNally
8/24/17
369
David
McNally
8/21/17
370
Kathryn
McNamara
8/22/17
371
Francis
Metelski
8/21/17
372
Julie
Metz
8/21/17
373
Carol
Meyer
8/21/17
374
Robert
Michaels
8/30/17
375
Checko
Miller
8/21/17
376
Checko
Miller
9/1/17
377
Patricia
Miller
8/21/17
378
Scott
Miller
7/17/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
I.asl Name
Dale Submitted
379
Katharine
Millonzi
8/29/17
380
Giles
Mitchell
8/25/17
381
Deidre
Moderacki
8/29/17
382
Julian
Moll-Rocek
7/25/17
383
Carol
Monteleoni
7/26/17
384
Philip and Carol
Monteleoni
7/26/17
385
Kimberly
Mooers
8/31/17
386
Kimberly
Mooers
8/31/17
387
Sol
Mora
7/26/17
388
Teresa
Morelle
8/18/17
389
David
Mortimer
8/28/17
390
Eric
Munson
8/21/17
391
Maria
Muro
8/29/17
392
Jay
Murphy
8/31/17
393
Sean
Murray
8/31/17
394
Judy Gelman
Myers
8/16/17
395
Ani
Nappa
8/21/17
396
Jonathan
Nedbor
9/1/17
397
Patrick
Nelson
9/1/17
398
Mike
Newman
7/6/17
399
Grace
Nichols
8/21/17
400
Bob
Nirkind
8/25/17
401
William
Nixon
8/31/17
402
Jean
Noack
8/29/17
403
Wendy
Nodop
8/21/17
404
Erika
Nonken
8/29/17
405
Brian
Nowitski
8/29/17
406
Alexis
O'Brien
8/29/17
407
Kathryn
O'Brien
8/21/17
408
Annemarie
O'Connor
8/22/17
409
Mary Anna
O'Donnell
8/21/17
410
Rick
Oestrike
7/6/17
411
Margot
Olavarria
8/24/17
412
Victoria
Oltarsh
8/22/17
413
Victoria
Oltarsh
8/29/17
414
Kathryn
Ornstein
8/29/17
415
Eric
Ortner
8/22/17
416
Lauree
Ostrofsky
9/1/17
417
Margaret
Othrow
6/9/17
418
Carl
Otto
8/29/17
419
Craig D.
Palmer
8/25/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
420
John
Palmer
8/21/17
421
Julie
Pari si
8/21/17
422
Greg
Patch
8/21/17
423
Barbara
Paterson
8/21/17
424
Joy
Pell
9/1/17
425
Valerie
Percy
8/22/17
426
Katherine
Perino
8/29/17
427
Robert
Perretti
5/30/17
428
Robert
Perretti
8/16/17
429
Robert
Perretti
8/16/17
430
Allison
Philpott
6/14/17
431
Kate
Phipps
8/29/17
432
Steven
Plotnick
7/13/17
433
Philip
Podmore
9/1/17
434
Rhonda
Pomerantz
8/22/17
435
Gail
Porter
5/30/17
436
Nicole
Porto
8/29/17
437
Sarah
Posner
8/29/17
438
Beth
Propper
8/29/17
439
Teri
Ptacek
9/1/17
440
Carmen
Pujols
6/27/17
441
Carmen
Pujols
6/28/17
442
Merrilyn
Pulver-Moulthrop
8/31/17
443
Patrick
Purcell
8/21/17
444
Ann
Quota
8/30/17
445
B
R
6/5/17
446
Amparo
Rally
8/30/17
447
Donald
Rally
8/30/17
448
Dorrit
Ram
8/16/17
449
Michael
Reed
7/25/17
450
James
Renner
8/31/17
451
Ryan
Reutershan
9/1/17
452
Heidi
Reyes
8/15/17
453
Michelle
Riddell
8/21/17
454
Michael
Riggio
8/29/17
455
Dennis
Riley
8/22/17
456
Andres
Rivera
8/29/17
457
David and Mary
Roberts
8/26/17
458
Timothy
Roberts
8/26/17
459
Clinton
Robinson
8/29/17
460
Matthew
Robinson
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
461
Jennifer
Roeder
8/29/17
462
Jessica
Roman
8/29/17
463
Christine
Root
9/1/17
464
Edith
Root
8/21/17
465
Bruce
Rosen
8/25/17
466
Martha
Roth
8/29/17
467
Matt
Rowan
7/20/17
468
Ann
Royston
9/1/17
469
Leah
Rubenstein
8/21/17
470
Franz
Safford
8/30/17
471
Donald
Sagar
9/1/17
472
Patricia
Santiago
8/21/17
473
Jeffrey
Scales
8/29/17
474
Lisa
Scerbo
8/31/17
475
Karin
Scheele
7/25/17
476
Marilyn
Schiller
7/24/17
477
Marian
Schoettle
8/22/17
478
Roni
Schotter
8/30/17
479
Penny
Schoutn
8/21/17
480
Greg
Schultz
7/25/17
481
Phillip
Schwartz
8/21/17
482
Annie
Scibienski
8/21/17
483
Nancy
Sconza
8/21/17
484
Pat
Sexton
8/21/17
485
Eric
Shelfin
8/22/17
486
Laurel
Shute
8/31/17
487
Laurel
Shute
9/1/17
488
Claire
Siegel
7/28/17
489
Bena
Silber
9/1/17
490
Sherrill
Silver
7/26/17
491
Donna
Simms
8/21/17
492
Marianne
Siniopkin
8/25/17
493
Joanne
Sinovoi
8/29/17
494
Donald
Smith
8/22/17
495
Mark
Smith
8/21/17
496
Marie
Snyder
7/25/17
497
Sara
Sogut
8/21/17
498
Sara
Sogut
8/29/17
499
Jessica
Soloman
6/2/17
500
Leola
Specht
8/7/17
501
Leola
Specht
8/10/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
502
Paula
Speer
8/21/17
503
Judith
Stahl
8/31/17
504
Colin
Stair
8/29/17
505
Judy
Stanley
8/21/17
506
Alex
Stavis
8/21/17
507
Alex
Stavis
8/21/17
508
Maxina
Stearn
8/9/17
509
Stephanie
Stefanski
8/29/17
510
Joe
Stefko
8/16/17
511
Evelyn
Stein
8/29/17
512
Barbara
Stemke
6/28/17
513
Fred
Stern
9/1/17
514
Marylou
Stern
8/22/17
515
Eric
Stiller
8/25/17
516
Julia
Stokes
8/31/17
517
Barbara
Sugin
8/29/17
518
Leonard
Sugin
8/29/17
519
Eileen
Sullivan
6/18/17
520
James
Sullivan
8/29/17
521
Marilyn
Sullivan
8/21/17
522
Christian
Sweningson
8/29/17
523
Nava
Tabak
8/30/17
524
Linda
Tafapolsky
8/21/17
525
Constance
Taft
8/21/17
526
Silvana
Tagliaferri
7/2/17
527
Jeff
Tanenbaum
8/9/17
528
Maria-Luisa
Tasayco
8/29/17
529
Annabel
Taylor
8/29/17
530
Marie
Taylor
9/1/17
531
Jaden
Thompson
7/25/17
532
Jack
Thorpe
8/21/17
533
Judith
Timke
7/26/17
534
Sarah
Todd
7/27/17
535
Nancy
Torchia
9/1/17
536
Vito
Trasmonte
9/1/17
537
Diane
Trieste
8/30/17
538
Barbara
Ungar
8/25/17
539
Michael
Vagnetti
8/25/17
540
Peter
Van Aken
8/21/17
541
Mark
Varian
8/29/17
542
Jessica
Vaughan
8/22/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
543
Jason
Velez
8/22/17
544
Harry
Vincent
8/25/17
545
Connie
Vixon
8/29/17
546
Tico
Vogtt
8/21/17
547
Leslie
Von Pless
8/23/17
548
Dorothy
Wadsworth
8/21/17
549
Jennifer
Walford
8/25/17
550
Alison
Waller
7/21/17
551
Emily
Waller
7/27/17
552
Bella
Wang
8/28/17
553
Kathleen
Wanser
8/29/17
554
Laura
Ward
8/22/17
555
Robyn
Waters
8/29/17
556
Noah
Watts
7/25/17
557
Russell
Wege
7/25/17
558
Laura
Weiland
7/25/17
559
Gerald
Wein
9/1/17
560
Mark
Weinstein
8/21/17
561
Harvey
Weiss
9/1/17
562
Tierney
Weymueller
8/21/17
563
Cindy
Wian
8/28/17
564
Jared
Widjeskog
8/21/17
565
Trisha
Wild
8/23/17
566
Courtney M.
Williams
8/25/17
567
Jason
Williams
8/21/17
568
Autumn
Williams-Wussow
8/21/17
569
Geniene
Wilson
8/21/17
570
Sally
Wilson
7/19/17
571
Sarah
Wilson
7/20/17
572
Tania
Wolf
8/30/17
573
Bill
Wolfsthal
8/31/17
574
Doug
Wygal
8/29/17
575
Elizabeth
Yalkut
6/12/17
576
Erin
Yarrobino
8/23/17
577
Kathleen
Young
8/21/17
578
Brook
Zelcer
8/30/17
579
John
Zimmerman
7/20/17
580
Juliette
7/25/17
Komi Letters
581
Patricia
Aakre
8/25/17
582
Betty
AbajianSeaman
8/21/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
583
Gabriel
Abate
8/29/17
584
August
Abel
8/19/17
585
Katherine
Abel
8/29/17
586
Steven
Abel
8/25/17
587
Olya
Abezgauz
8/21/17
588
Olya
Abezguaz
8/22/17
589
Doug
Abramson
8/21/17
590
Mary
Abrey
8/22/17
591
Bobbie
Adams
8/29/17
592
Sean
Adams
8/18/17
593
Jana
Adler
8/26/17
594
Joan
Agro
8/24/17
595
Grace
Aiello
8/29/17
596
Son] a
Aiken
8/22/17
597
Pascal
Akesson
8/29/17
598
Donald
Albrecht
8/30/17
599
Diane
Alden
8/24/17
600
Rick
Alfandre
8/21/17
601
Jill
Alibrandi
8/26/17
602
Gail
Allan
8/29/17
603
Jeannette
Allan
8/24/17
604
David
Allen
8/30/17
605
Kendra
Allenby
9/1/17
606
Ivanya
Alpert
8/29/17
607
Steven
Altarescu
9/1/17
608
Karen
Ambrosetti
8/21/17
609
Martin
Amsel
8/24/17
610
Amy
Anderson
8/29/17
611
Emily
Anderson
8/30/17
612
Katherine
Anderson
8/29/17
613
Tracy
Anderson
8/29/17
614
Nancy
Andreas si
8/29/17
615
Audrey
Ang
8/28/17
616
Paul
Annetts
8/24/17
617
Lisa
Arbisser
9/1/17
618
Mercedes
Armillas
8/29/17
619
Lindsey
Arnell
8/30/17
620
K
Arnone
8/7/17
621
Barbara
Aronowitz
8/24/17
622
Eric
Arroyo
8/29/17
623
Karen
Asher
8/21/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
624
Jude
Asphar
8/29/17
625
Bianca
Assim-Kon
8/18/17
626
Alexis
Audette
8/24/17
627
Carol
Auer
8/22/17
628
Melisa
Auf der Maur
8/31/17
629
Brian
Austin
8/29/17
630
Sharon
AvRutick
8/22/17
631
S
B
8/24/17
632
Katherine
Babiak
8/30/17
633
Jesse
Bachir
8/29/17
634
Frances
Backofen
8/21/17
635
Marta
Baez
8/29/17
636
Cari
Bailey
8/21/17
637
Melissa
Bailey
8/22/17
638
Jeffrey
Bains
8/29/17
639
P
Baker
8/16/17
640
Candace
Balmer
8/30/17
641
Janice
Banks
8/29/17
642
Peter
Bannon
8/29/17
643
Daniel
Barclay
8/29/17
644
Alan
Bare
8/24/17
645
John
Barone
8/21/17
646
Enzo
Barrios
8/30/17
647
Marina
Barry
8/29/17
648
Carolyn
Bartholomew
8/24/17
649
Olga
Bartnicki
8/29/17
650
Cat
Basciano
8/16/17
651
Mark
Bastian
9/1/17
652
William
Battaglia
8/30/17
653
Pamela
Battle
8/30/17
654
Deborah
Bauer
8/30/17
655
Joan-Marie
Bauman
8/24/17
656
Deborah
Baumann
8/29/17
657
John
Bauza
8/21/17
658
Susan
Baxter
8/24/17
659
Bonnie
Bayardi
8/25/17
660
Linda
Beach
8/24/17
661
Carol
Bean
8/22/17
662
Elisabeth
Bechmann
8/29/17
663
Juan
Bedoya
8/22/17
664
Stephan
Beffre
8/26/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
665
Bertram
Beissel
8/29/17
666
Stephen
Bellomo
8/30/17
667
David
Bennett
8/29/17
668
Frances
Berger
8/22/17
669
Stephanie
Berger
9/1/17
670
Deborah
Bergman
8/28/17
671
Jill
Berliner
8/7/17
672
Janice
Bernard
8/29/17
673
Jean
Bernard
8/22/17
674
Bonnie
Bernstein
8/29/17
675
Lesley
Bernstein
8/22/17
676
Lisa
Berrol
8/22/17
677
Lisa
Berry
8/30/17
678
Joseph
Bertolozzi
8/22/17
679
Karyn
Bevet
8/22/17
680
Bob
Bickford
9/1/17
681
Annie
Bien
8/18/17
682
Alex
Billig
8/21/17
683
Gene
Binder
8/21/17
684
Janet
Binion
8/21/17
685
Janet
Binion
8/29/17
686
Richard
Binkele
8/22/17
687
Beth
Birnbaum
8/24/17
688
Jacqueline
Birnbaum
8/7/17
689
Maureen
Black
8/25/17
690
Sandy
Black-McDonough
8/29/17
691
Jeremiah
Blatz
8/25/17
692
Ashley
Blazer
8/29/17
693
Brandon
Block
8/17/17
694
Corliss
Block
8/25/17
695
Josephine
Bloodgood
8/21/17
696
Donald
Bluestone
9/1/17
697
Richard
Bodane
8/24/17
698
Dwight
Bodycott
8/18/17
699
Pauline
Boehm
8/10/17
700
Hollis
Bogdanffy
8/21/17
701
Gusti
Bogok
8/19/17
702
David
Bogoslaw
8/29/17
703
Gabrielle
Bordwin
8/29/17
704
Jim
Botta
8/24/17
705
Garrison
Botts
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
706
KJ
Bo wen
8/30/17
707
Grace
Bowne
8/24/17
708
Mary Alice
Boyle
8/22/17
709
Mary Alice
Boyle
8/22/17
710
Diane E.
Bradley
8/25/17
711
Kathleen
Brady
8/30/17
712
Ira
Brandenburg
8/23/17
713
Peter
Brandt
8/7/17
714
Nancy
Breen
8/22/17
715
Sophie
Breitbart
8/22/17
716
Lise
Brenner
8/29/17
717
Patricia
Brescia-Cantine
8/29/17
718
Frank
Brice
8/21/17
719
John
Brinkman
8/24/17
720
Anna
Bristow
8/30/17
721
Undine
Brod
8/30/17
722
Kathleen
Brodbeck
9/1/17
723
Marinus
Broekman
8/24/17
724
Alan
Brown
8/29/17
725
Babette
Brown
8/7/17
726
Denise
Brown
8/29/17
727
Janelle
Brown
8/25/17
728
Elizabeth
Bruen
8/22/17
729
Deborah
Brunner
8/22/17
730
Nancy
Bruno
8/29/17
731
Jan
Buchalter
8/8/17
732
Anne Marie
Bucher
8/24/17
733
Joseph
Buchheit
8/11/17
734
Teresa
Buchholz
8/29/17
735
Karin
Bucklin
8/29/17
736
Catherine
Budd
8/22/17
737
Katie
Bull
8/29/17
738
Diane
Burke
8/29/17
739
Sue
Burke
8/22/17
740
Kit
Burke-Smith
8/22/17
741
Margaret
Burton
8/31/17
742
Elena
Busani
8/24/17
743
Edward
Butler
8/29/17
744
Susan
Butterfass
8/22/17
745
Joyce
Byrne
8/21/17
746
Suzanne
Cachon
8/30/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
747
Peter
Callaway
8/29/17
748
R
Cammisa
8/25/17
749
Dac
Campbell
8/30/17
750
Patti
Candelari
8/29/17
751
Irwin
Cantos
8/22/17
752
Michelle
Capuano
8/22/17
753
Patricia
Cardello
8/30/17
754
Patricia
Cardoso
8/24/17
755
Rachel
Careau
9/1/17
756
Elisa
Caref
8/21/17
757
Kathy
Carey
9/1/17
758
Patsy
Carl
8/30/17
759
Nancy
Carmichael
8/22/17
760
Christy
Carosella
8/29/17
761
Katelyn
Carroll
8/22/17
762
Matthew
Carroll
8/21/17
763
Teri-Ann
Carryl
8/30/17
764
Matthew
Carson
8/22/17
765
Carmen
Casado
8/30/17
766
Jose Chicaiza
Casado
8/30/17
767
Lynn
Cascio
8/29/17
768
Allan
Casement
8/29/17
769
Leslie
Cassidy
8/29/17
770
Elizabeth
Castaldo
8/29/17
771
Dorinda
Cataldo
8/24/17
772
Armanda
Catenaro
8/25/17
773
Mikki
Chalker
8/24/17
774
Michael
Chameides
9/1/17
775
Henry
Charles
8/29/17
776
Phylicia
Chartier
8/3/17
111
Lisa
Chason
8/31/17
778
Myrel
Chernick
8/30/17
779
Elaine
Cherry
8/30/17
780
Russell
Chiappa
8/29/17
781
Evelyn
Chiarito
8/29/17
782
Evonne
Cho
8/22/17
783
Kelly
Choi
8/30/17
784
Doris
Chorny
8/31/17
785
Peggy
Christian
8/30/17
786
Bob
Christianson
8/24/17
787
Stephanie
Christoff
8/20/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
788
Lauren
Ciborski
8/31/17
789
Monique
CI ague
8/29/17
790
Lawrence
Clarke
8/29/17
791
Meryl
Classen
8/29/17
792
Anne Katherine
Cleary
8/22/17
793
Susan
Clelland
8/29/17
794
Geralyn
Clemens
8/31/17
795
Jesse
Clinton
8/29/17
796
Joseph
Cloidt
8/29/17
797
Laura-Christina
Cobb
8/17/17
798
Claudia
Cockerill
8/22/17
799
Florence
Cohen
8/29/17
800
Wendi
Cohen
8/29/17
801
Herbert
Coles
8/29/17
802
Bonnie
Collins
8/25/17
803
Thomas
Comiskey
8/7/17
804
David
Condon
8/29/17
805
Patricia
Connolly
8/24/17
806
Douglas
Cooke
8/29/17
807
James
Cooper
8/29/17
808
Adam
Cooperstock
8/24/17
809
Ryan
Coraldi
8/22/17
810
Marion
Corbin
8/22/17
811
Marion
Corbin
8/22/17
812
Marion
Corbin
8/29/17
813
Phyllis
Corcacas
8/29/17
814
Jared
Cornelia
8/29/17
815
Sean
Cortright
8/22/17
816
Victoria
Costello
8/22/17
817
Fiona
Cousins
8/17/17
818
Sherrill
Cox
8/25/17
819
Susan
Cox
8/7/17
820
Laurrie
Cozza
8/29/17
821
Marcelle
Crago
8/30/17
822
Joy
Cranker
8/22/17
823
Fran
Crilley
8/25/17
824
A1
Cruz
9/1/17
825
Helen
Cu
8/29/17
826
Ann Marie
Cunningham
8/29/17
827
Benjamin
Curran
8/23/17
828
Annalise
Curtin
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
829
Whitefeather
Curtiss
8/22/17
830
Caroline
Cutroneo
8/21/17
831
Clarissa
Cylich
8/21/17
832
Jane
Cyphers
8/16/17
833
Julie
Dahl
8/21/17
834
Marge
Dakouzlian
8/25/17
835
Jordan
Dale
8/30/17
836
Susan
Damato
8/19/17
837
Donna
Dangelo
8/22/17
838
Beth
Darlington
8/7/17
839
Kate
Darringo
8/18/17
840
Nina
David
8/24/17
841
Davis
8/28/17
842
Juanita
Dawson-Rhodes
8/29/17
843
Carol
De Angelo
8/24/17
844
C
de Ben
8/18/17
845
Noel
De La Cruz
8/25/17
846
Gerald
Dean
8/23/17
847
Nita
DeBono
8/19/17
848
Diane
DeChillo
9/1/17
849
Theresa
DeGraw
8/25/17
850
Julia
Dehn
9/1/17
851
Charles
Del Regno
8/23/17
852
Charlie
Del Regno
9/1/17
853
Arthur
Delaney
8/20/17
854
Robert
DeLay
8/30/17
855
Peter
DeLorenzo
8/29/17
856
Sheila
Dempsey
8/7/17
857
Laura
deNey
8/29/17
858
Daryl
Denning
8/24/17
859
Donna
Denny
8/30/17
860
Margaret
DeRose
8/30/17
861
Mark
Dery
8/21/17
862
Roberta
Desalle
8/29/17
863
Claudia
Devinney
8/7/17
864
Sterling
DeWeese
8/22/17
865
Harris
Diamant
9/1/17
866
Josh
Diamond
9/1/17
867
Rosalind
Dickinson
8/29/17
868
Tara
DiDonna
8/22/17
869
David
Dienes
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
870
James
DiMunno
8/18/17
871
Jacalyn
Dinhofer
8/29/17
872
NoA©
Dinnerstein
8/30/17
873
Doreen
Diorio
8/30/17
874
Vincent
DiTizio
8/30/17
875
Barbara
DiTommaso
8/19/17
876
James
Doherty
8/25/17
877
Adam
Domini ak
8/18/17
878
Ann
Donohue
8/29/17
879
Elaine
Donovan
8/7/17
880
Chris
Doolittle
8/22/17
881
David
Douglas
8/29/17
882
Susan
Downes
8/21/17
883
Taylor
Doyle
8/21/17
884
Muriel
Doyne
8/18/17
885
Christine
Drosky
8/22/17
886
Bette
Druck
8/16/17
887
Chris
Drumright
8/29/17
888
Brian
Duea
8/29/17
889
Diane
Duffus
8/25/17
890
Brian
Duffy
8/23/17
891
John
Dugan
8/22/17
892
John
Dugan
9/1/17
893
Timothy
Dunn
8/29/17
894
Bernadette
Duquette
8/22/17
895
Janet
Duran
8/30/17
896
Gregory
Durniak
8/29/17
897
Virginia
Dwyer
8/29/17
898
Emily
Eckart
8/21/17
899
Choral
Eddie
8/21/17
900
Alisa
Eilenberg
8/7/17
901
Esmee
Einerson
8/29/17
902
Josh
Eisenstark
8/24/17
903
Liz
Elkin
8/29/17
904
Jan
Emerson
8/29/17
905
Anne
Endler
8/7/17
906
Anna
Engdahl
8/29/17
907
D.
E-Platt
8/21/17
908
Lori
Epstein
9/1/17
909
Susan
Epstein
9/1/17
910
Alessia
Eramo
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
911
Jessica
Ettinger
8/30/17
912
Alicia
Everett
8/30/17
913
Jennifer
Fahey
8/25/17
914
Judy
Fairless
8/29/17
915
Eugene
Falik
8/12/17
916
Russell
Faller
8/8/17
917
Dan
Famer
8/21/17
918
Stacey
Farber
8/21/17
919
Raymond
Farrington
8/29/17
920
Tami Lin
Farrow
8/29/17
921
Mary
Fasano
8/22/17
922
Wendy
Fast
8/30/17
923
Mary Ann
Fastook
8/29/17
924
Pat
Faye
8/21/17
925
Kristina
Fedorov
8/25/17
926
Arnold
Feinsilber
8/30/17
927
Dianne
Felix
8/25/17
928
Ellen
Fenton
8/31/17
929
Roxanne
Ferber
8/25/17
930
Yvette
Fernandez
8/30/17
931
Andrew
Fetherolf
8/25/17
932
Ariel
Feuz
8/25/17
933
Jon
Fields
8/29/17
934
Francisco
Figueirido
8/30/17
935
Cristina
Fiorillo
8/29/17
936
Chrissy
Fischetti
8/22/17
937
Mel
Fish
8/22/17
938
Norman
Fisher
8/22/17
939
Kaitlin
Fitch
8/7/17
940
Julia
Fitzgerald
8/29/17
941
Mike
Fitzgerald
8/21/17
942
Barbara
Fitzhugh
8/31/17
943
Barbara
Fitzhugh
9/1/17
944
Ellen
Fleishman
8/24/17
945
Diana
Flood
8/22/17
946
Patricia
Flood
8/22/17
947
Patricia
Flood
8/25/17
948
Patricia
Flood
8/29/17
949
Patricia
Flood
8/29/17
950
Patricia
Flood
8/30/17
951
Bobbie
Flowers
8/24/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
952
Jillian
Flynn
8/17/17
953
Thomas
Folkl
8/25/17
954
JR.
Fontaine-Serra
8/29/17
955
Maureen
Ford
8/29/17
956
Tanya
Foret
8/21/17
957
Janet
Forman
8/18/17
958
Laura
Forman
8/21/17
959
Devlin
Foster
8/30/17
960
Ian
Fountain
8/21/17
961
Ian
Fountain
9/1/17
962
Steven
Fowler
8/21/17
963
Andrea
Frank
8/29/17
964
Elaine
Fr ankle
8/30/17
965
Brian
Frederick
8/24/17
966
Misha
Fredericks
9/1/17
967
Heather
Free
8/6/17
968
Ava
Freeman
8/30/17
969
Ronald
Friedman
8/24/17
970
Justin
Fromm
8/16/17
971
L.
Fron
8/29/17
972
Romain
Fruge
8/28/17
973
Mark
Frusciante
8/22/17
974
Carrie
Fudge
8/30/17
975
Jane
Fuller
9/1/17
976
Roy
Fuller
8/24/17
977
Dorian
Fulvio
8/29/17
978
Lee
Furbeck
9/1/17
979
Victoria
Furio
8/29/17
980
Rob
Fursich
8/7/17
981
Deborah
Fusco, RMT
8/22/17
982
Maria
Gagliardi
8/30/17
983
Bernard
Galiley
8/29/17
984
Barbara
Galli
8/22/17
985
Dianne
Galliher
8/29/17
986
Angel
Garcia
8/18/17
987
Cari and Donald
Gardner
8/7/17
988
Joy
Garland
8/18/17
989
Ktie
Garton
8/29/17
990
Nathan
Gauthier
8/29/17
991
John
Gebhards
8/24/17
992
Sharon
Gelfand
8/22/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
993
Sharon
Gelfand
8/22/17
994
Michael
Gelfer
8/7/17
995
Derek
Gendvil
8/29/17
996
Donna
George
8/29/17
997
Thomas
George
8/29/17
998
Paul
Ghenoiu
8/22/17
999
Helen
Ghiradella
8/24/17
1000
Mary
Gianetto
8/22/17
1001
Mary
Gianetto
8/22/17
1002
Anthony
Giannantonio
8/22/17
1003
Laurette
Giardino
8/22/17
1004
Thomas
Giblin
8/18/17
1005
Ward
Giblin
8/18/17
1006
David
Gilbert
8/22/17
1007
Nina
Gimmel
8/30/17
1008
Mark
Ginsburg
8/30/17
1009
Clarice
Glandon
8/29/17
1010
Toni
Glikes
8/21/17
1011
Matthew
Glock
8/22/17
1012
Matthew
Glock
8/30/17
1013
Rise
Gluck
8/29/17
1014
Alexander
Goasdoue
8/7/17
1015
Susan
Goldfarb
8/21/17
1016
Allan
Goldstein
8/21/17
1017
Howard
Goldstein
8/29/17
1018
Mary
Goldstein
8/22/17
1019
Louise
Golub
8/29/17
1020
Ronaldo
Gonzalez
8/22/17
1021
Mike
Good
8/30/17
1022
Karine
Gordineer
8/25/17
1023
David
Gordon
8/27/17
1024
Emily
Gordon
8/28/17
1025
Nancy
Gordon
9/1/17
1026
Richard
Gordon
8/29/17
1027
Sarah
Gordon
8/30/17
1028
Cyd
Gorman
9/1/17
1029
Deborah
Gorman
8/29/17
1030
Mark
Gorsetman
8/19/17
1031
Laura
Grady
8/25/17
1032
Jacqueline
Grand Pre
8/30/17
1033
George
Graney
8/21/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1034
D
Green
8/31/17
1035
Jeff
Greenberg
8/29/17
1036
Karen
Greenspan
8/29/17
1037
Dana
Gregg
8/8/17
1038
Sophie
Greller
8/29/17
1039
Homer Ellis
Griffin
8/29/17
1040
Lucy
Grimes
8/29/17
1041
Tracy
Griswold
8/7/17
1042
Andrew
Grod
8/21/17
1043
John
Gromada
8/31/17
1044
Martin
Gromulat
8/7/17
1045
Sabina
Gross
8/18/17
1046
Yonni
Groza
8/23/17
1047
Gina
Guarino
8/22/17
1048
Richard
Guier
8/29/17
1049
James
Guilianelli
8/22/17
1050
James
Guilianelli
8/29/17
1051
Paula
Gullo
8/23/17
1052
Rachel
Gumina
8/24/17
1053
Karlene
Gunter
8/9/17
1054
Marina
Gutierrez
8/21/17
1055
Zinnia
Gutowski
8/29/17
1056
Dominique
ha
8/17/17
1057
Connie
Haack
8/21/17
1058
Jeffrey
Haas
8/23/17
1059
Renee
Hack
8/24/17
1060
Renee
Hack
8/30/17
1061
Heather
Haggerty
8/22/17
1062
Brandon
Hakulin
8/21/17
1063
Peter
Halewood
8/28/17
1064
Brett
Hall
8/22/17
1065
Margaret
Halliday
8/25/17
1066
Hagit
Halperin
8/29/17
1067
Jane
Halsey
8/29/17
1068
Colleen
Hamilton
8/18/17
1069
John
Hamilton
8/25/17
1070
Michele
Hamilton
9/1/17
1071
Sarah
Hamilton
8/7/17
1072
Susan
Hamilton
8/2/17
1073
Mary Lynn
Hanley
8/29/17
1074
Terence and Norma
Hannigan
8/22/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1075
Rosalie
Harman
8/16/17
1076
Elizabeth
Harrington
8/23/17
1077
Emmalia
Harrington
8/16/17
1078
Elaine
Hartel
8/29/17
1079
Joyce
Hartsfield
8/22/17
1080
Christine
Harvey
8/18/17
1081
David
Harvey
8/22/17
1082
Bjorn
Harvold
8/17/17
1083
Tracey
Hastings-Ward
9/1/17
1084
Martin
Hauser
8/30/17
1085
Jill
Hausman
8/29/17
1086
Kathy
Haverkamp
8/29/17
1087
Gerry
Hawkins
8/22/17
1088
Sheryl & Don
Haynie/Samuel
8/24/17
1089
Mary
Hays
8/28/17
1090
Chris
Hazynski
8/24/17
1091
William
Healey
8/7/17
1092
Thomas
Hearty
8/24/17
1093
Josh
Heffron
8/24/17
1094
Eli
Hegeman
8/19/17
1095
Adriana
Heguy
8/16/17
1096
Michael
Heimbinder
8/29/17
1097
Jenny
Heinz
8/24/17
1098
Mary
Heller
8/29/17
1099
Laurie
Henderson
8/22/17
1100
-
Hera
8/29/17
1101
Jan
Herndon
8/18/17
1102
Carol
Herring
9/1/17
1103
Marianne
Herrmann
8/22/17
1104
Nava
Herzog
8/25/17
1105
Brenda
Hewett
8/31/17
1106
Pat
Hickey
8/25/17
1107
Brian
Higbie
8/25/17
1108
Jeanne
Hobert
9/1/17
1109
Mark
Hockman
8/18/17
1110
Matthew
Hoff
9/1/17
1111
Deborah
Hoffman
8/25/17
1112
Randi
Hoffmann
8/29/17
1113
Paul
Hofheins
8/18/17
1114
Constance
Hoguet Neel
8/24/17
1115
Hussein
Hollan
8/25/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
1116
Susan
Holland
9/1/17
1117
Tamsin
Hollo
8/22/17
1118
John
Holodak
8/29/17
1119
F
Holz
8/29/17
1120
J
Holz
8/29/17
1121
Teresa
Hommel
8/29/17
1122
Natalia
Hook
8/21/17
1123
Stephen
Hopkins
8/17/17
1124
Jennifer
Horowitz
8/19/17
1125
Lily
Hou
8/29/17
1126
Jennifer
Houston
9/1/17
1127
Paticia
Houston
8/24/17
1128
Claire
Howard
8/24/17
1129
Nina
Howes
8/21/17
1130
Vicki
Huber
8/29/17
1131
Christina
Hubrt
8/22/17
1132
Jerold
Huebner
8/23/17
1133
Marc
Humphrey
8/30/17
1134
Obie
Hunt
8/16/17
1135
Heather
Hurley
8/30/17
1136
June
Hurst
8/29/17
1137
Noelene
Hutchinson
8/25/17
1138
A
I
8/29/17
1139
Hatti
lies
8/29/17
1140
Cora
Impenna
8/22/17
1141
Daniel
Incristo
8/3/17
1142
Margaret
Innerfoher
8/7/17
1143
Adam
Isler
8/29/17
1144
Susan
Italia
8/31/17
1145
Lisa
Izes
8/30/17
1146
Sandy
J
8/29/17
1147
B.L.
Jacobi
8/22/17
1148
Carol
Jagiello
8/29/17
1149
Chip
James
8/21/17
1150
Chip
James
8/30/17
1151
Jared
Jamesson
8/29/17
1152
Shahla
Jannetta
8/31/17
1153
Alan
Jasper
8/29/17
1154
Payont
Jatasanont
8/29/17
1155
Lynne
Jeanette
8/30/17
1156
Barbara
Jesrani
8/30/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1157
Angela
Johnsom
8/23/17
1158
Carla Rae
Johnson
8/28/17
1159
Kathy
Johnson
8/21/17
1160
Margaret
Johnson
9/1/17
1161
Theresa
Johnson
8/24/17
1162
David
Johnston
9/1/17
1163
Nathaniel
Johnston
8/22/17
1164
Blanche
Jones
8/22/17
1165
Marjorie
Jones
8/22/17
1166
Robert
Jones
8/19/17
1167
Walter
Jones
9/1/17
1168
Barbara
Joslyn
8/29/17
1169
Adrian
Juarez
8/30/17
1170
Carol
Jurczewski
8/29/17
1171
Elaine
Jurumbo
8/29/17
1172
Deedra
Kaake
8/22/17
1173
Marilyn
Kaggen
8/24/17
1174
Lyle
Kahn
8/29/17
1175
Sabrina
Kahn
8/12/17
1176
Paul
Kalka
9/1/17
1177
Jean
Kallina
8/22/17
1178
Edith
Kantrowitz
8/31/17
1179
Sandra
Kaplan
8/29/17
1180
Sylvia
Kaplan
8/29/17
1181
Joe
Karr
8/24/17
1182
Beth
Kashmann
8/25/17
1183
Sheri
Kastner
9/1/17
1184
Lora
Katen
8/29/17
1185
Nikki
Katsikas
8/28/17
1186
Alayne
Katz
8/30/17
1187
Stacy
Katz
8/21/17
1188
Annie
Katzman
8/29/17
1189
Andreas
Kaubish
8/7/17
1190
Alix
Keast
8/24/17
1191
John
Keiser
8/24/17
1192
Peter
Keiser
8/19/17
1193
Charles
Keller
8/24/17
1194
Matthew
Kelly
8/29/17
1195
Vincent
Kelly-Brownell
8/29/17
1196
Jane
Kendall
8/30/17
1197
Meredith
Kent-Berman
8/19/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
1198
Maria
Keramari
8/22/17
1199
David
Kern
8/24/17
1200
Ethan
Kerr
8/23/17
1201
Lisa
Ketchum
8/25/17
1202
JK
Kibler
8/29/17
1203
Johanna
Kiernan
8/30/17
1204
Joh
Killen
8/22/17
1205
Kevin
Kilner
8/29/17
1206
Donald
Kimmel
8/25/17
1207
D.
King
9/1/17
1208
David
King
8/26/17
1209
Julie Parisi
Kirby
8/7/17
1210
Lori
Kirsch
9/1/17
1211
Leonard
Kirsch III
8/21/17
1212
Leonard
Kirsch III
8/22/17
1213
Leonard
Kirsch III
8/22/17
1214
Leonard
Kirsch III
8/25/17
1215
Leonard
Kirsch, III
9/1/17
1216
Sandra
Kissam
8/24/17
1217
Eresha
Kissoon-Fareed
8/22/17
1218
Timothy
Kleeger
8/30/17
1219
Amy
Kletter
8/29/17
1220
David
Klinke
8/7/17
1221
Ulrike
Klopfer
8/24/17
1222
Claudine
Klose
9/1/17
1223
Nina
Knanishu
8/19/17
1224
Brian
Knowles
8/31/17
1225
Michael
Kodransky
8/30/17
1226
Laura
Koestler
8/29/17
1227
Laura
Kohlmann
8/22/17
1228
Alon
Koppel
9/1/17
1229
Ray
Koretsky
8/30/17
1230
George
Kormendi
8/29/17
1231
Ellen
Korz
8/27/17
1232
Ellen
Kozak
8/30/17
1233
JAmes
Kozlik
8/22/17
1234
Lori
Krane
8/29/17
1235
Steven
Krauss
8/21/17
1236
Jennifer
Krawitz
8/11/17
1237
Pam
Kray Gallivan
8/18/17
1238
Elena
Krumova
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
1239
Richard
Krupp
8/25/17
1240
Walter
Kuciej
8/29/17
1241
William
Kuehnling
8/18/17
1242
Elyse
Kunz
8/30/17
1243
Pat
Kush
8/24/17
1244
Tor en
Kutnick
8/18/17
1245
Katie
Kynast
8/29/17
1246
John
Lacey
8/21/17
1247
Dimitri
Laddis
8/28/17
1248
Dennis
Ladner
8/31/17
1249
Annik
LaFarge
8/30/17
1250
Terri
Laidman
8/22/17
1251
Andrew
Laiosa
8/29/17
1252
Marion
Lakatos
8/29/17
1253
Catherine
Lai a
8/22/17
1254
Katrina
Lalonde
8/22/17
1255
Tara
Lambert
8/28/17
1256
Wendy
Lambert
8/22/17
1257
William
Landau
8/22/17
1258
Hilary
Lander
8/22/17
1259
Michelle
Lange
8/30/17
1260
Norbert
Langer
8/29/17
1261
Hatti
Langsford
8/30/17
1262
Bianca
Lanza
8/30/17
1263
Bianca
Lanza
9/1/17
1264
Ricky
Lark
8/22/17
1265
Nancy
Larsen
8/22/17
1266
Carol
Latourette
9/1/17
1267
Lynn
Lauber
8/21/17
1268
Julianna
Lavin
9/1/17
1269
Linda
Lavin
8/22/17
1270
Susan
Lawrence
8/22/17
1271
Michael
Lebron
8/22/17
1272
Jo-Ann
Lechner
8/19/17
1273
Benjamin
Lee
8/29/17
1274
Deborah K.
Lee
8/29/17
1275
Diane
Lee
8/30/17
1276
Michel
Lee
8/29/17
1277
Steven
Lee
8/31/17
1278
Steven
Lee
9/1/17
1279
Arthur
Leibowitz
8/7/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1280
Hannah
Leider
8/29/17
1281
Doug
Leihbacher
8/22/17
1282
Jill
Lein
8/30/17
1283
B. R.
Lemonik
8/24/17
1284
Bernice
Lenahan
8/4/17
1285
Eileen
Lennon
8/21/17
1286
Wayne
Lensu
8/7/17
1287
Gale
Leonard
8/25/17
1288
Gerson
Lesser
8/29/17
1289
Kathleen
Letchford
8/29/17
1290
Rhonda
Levine
8/7/17
1291
Ellen
Levinson
8/21/17
1292
Jeffrey
Levitt
8/18/17
1293
David
Levy
8/21/17
1294
Erma
Lewis
8/29/17
1295
Erma
Lewis
8/29/17
1296
Mike
Lieber
8/22/17
1297
D. M.
Linkie
8/25/17
1298
Matthew
Liponis
8/31/17
1299
Danette
Lipten
8/22/17
1300
Jennifer
Lischak
8/25/17
1301
Jim
Littlefield
8/29/17
1302
Elaine
Livingston
8/24/17
1303
Patricia
Livingston
8/30/17
1304
Patricia
Livingston
9/1/17
1305
Rich
Locicero
8/22/17
1306
Diane
Lombardi
8/22/17
1307
Diane
Lombardi
8/22/17
1308
Catherine
Lombardo
8/30/17
1309
Robert
Long
8/22/17
1310
Scott
Longstreet
8/21/17
1311
Mary
Loomba
8/29/17
1312
Michael
Loos
8/29/17
1313
Nancy
Lopez
8/24/17
1314
Christopher
Lord
8/19/17
1315
Mark
Lotito
8/27/17
1316
Evan
Loughran
8/10/17
1317
Hilarie
Louis
8/24/17
1318
Joe
Lowenbraun
8/23/17
1319
Alison
Lucek
8/30/17
1320
Nicole
Luciani
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1321
Rachel
Lugo
8/23/17
1322
Brian
Luman
8/24/17
1323
Martin
Lupowitz
8/25/17
1324
Susan
Lupul
8/22/17
1325
Barbara
Lynch
8/24/17
1326
Lois
Lynn
8/18/17
1327
Clarinda
Mac Low
8/29/17
1328
Stephen
Mac Nish
8/29/17
1329
Marissa
Macagnone
8/22/17
1330
Michael
Macelhiney
8/29/17
1331
Christine
Maciel
8/22/17
1332
Robert
Mackey
8/29/17
1333
Michael
Madden
8/7/17
1334
Robert
Madorran
8/30/17
1335
Laraine
Mai
8/21/17
1336
Karyn
Maier
8/30/17
1337
Linda
Maldonado
8/24/17
1338
Matthew
Malina
8/29/17
1339
Kenneth
Malkin
8/21/17
1340
Athena
Malloy
8/18/17
1341
Mitch
Mai oof
8/24/17
1342
Danielle
Maltby
8/22/17
1343
Lindsay
Mandel
8/28/17
1344
Michael
Mangino
8/22/17
1345
Alexandra
Manning
8/29/17
1346
Clint
Marallo
8/24/17
1347
Marlena
Marallo
8/2/17
1348
Jack David
Marcus
8/17/17
1349
Jack David
Marcus
8/22/17
1350
Kimberly
Marcus
8/29/17
1351
Karlene
Maresco
8/22/17
1352
Jordan
Margolis
8/23/17
1353
Kathy
Margulis
8/29/17
1354
Phillip
Marinelli
8/29/17
1355
Jane
Marinsky
8/21/17
1356
Darian
Mark
8/29/17
1357
Emily
Maroney
8/29/17
1358
Debbie
Marotta
8/22/17
1359
Jim
Marrinan
8/30/17
1360
Laurence
Martin
8/22/17
1361
Rea
Martin
8/30/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1362
Tina
Martin
8/29/17
1363
Isabel
Martins
8/18/17
1364
Joan
Martorano
8/22/17
1365
Toby
Marxuach-Gusciora
8/29/17
1366
Kara
Masciangelo
8/28/17
1367
Ben
Mastaitis
8/24/17
1368
Angela
Mastracchio
8/21/17
1369
Frances
Mastrota
8/7/17
1370
Dennis
Mathews
8/29/17
1371
Larissa
Matthews
8/18/17
1372
Elizabeth
Maucher
8/29/17
1373
Hope
Mauran
8/29/17
1374
Kurt
Mausert
8/21/17
1375
George Louis
Mayer
8/29/17
1376
Francis
Mayle
8/29/17
1377
Kathleen
Mazza
8/21/17
1378
Linda
McArdle
8/30/17
1379
Diane
McAteer
8/29/17
1380
Paul
McCarthy
8/28/17
1381
Richard
McCauley
8/24/17
1382
Flannery
McDermott
8/25/17
1383
John
McDonald
8/29/17
1384
Roland
McDonald
8/24/17
1385
Mary
McGeary
8/7/17
1386
Chris
Mcginn
8/29/17
1387
Emma
McGregor-Mento
8/16/17
1388
Steven
Mclntyre
8/30/17
1389
Grant
McKeown
8/28/17
1390
Mary
Mckeown
8/22/17
1391
Alan
McKnight
8/7/17
1392
Brian
McLaughlin
8/29/17
1393
Kathleen
McLaughlin
8/24/17
1394
Elizabeth
McMahon
8/7/17
1395
Jennifer
McMorrow
8/25/17
1396
Jennifer
McMorrow
8/29/17
1397
Susan
McNamara
8/17/17
1398
William
McNamara
9/1/17
1399
Monica
McQuade
8/25/17
1400
Robert
McQuilkiin Jr.
8/22/17
1401
Joanna
Meakin
8/25/17
1402
Tatiana
Mejia
8/19/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1403
Dominic
Melita
8/29/17
1404
Donna
Menconeri
9/1/17
1405
Rik
Mercaldi
8/21/17
1406
Jonathan
Mernit
8/21/17
1407
Andrew
Meyer
8/22/17
1408
Laurie
Miccio
8/25/17
1409
Bonnie
Michaels
8/22/17
1410
Sharon
Michales
8/24/17
1411
Ragnar
Midtskogen
8/21/17
1412
Lyndsey
Milcarek
8/20/17
1413
Joanne
Miller
8/29/17
1414
John
Miller
8/5/17
1415
Jonathan
Miller
8/16/17
1416
Marjorie
Miller
8/24/17
1417
Matthew
Miller
8/22/17
1418
Alvin
Miller Jr
8/25/17
1419
Alvin
Miller Jr
8/30/17
1420
Alvin
Miller Jr
9/1/17
1421
Alvin
Miller Jr.
8/22/17
1422
Judy
Miller-Lyons
9/1/17
1423
Jackie
Mills
8/29/17
1424
Laura
Milsom
8/22/17
1425
Harut
Minasian
8/31/17
1426
Hayley
Mink
8/29/17
1427
Ellen
Miret
8/22/17
1428
Lily
Mleczko
8/29/17
1429
Alexis
Mohr
8/30/17
1430
Phyllis
Mollen
8/24/17
1431
Barbara
Moloney
8/21/17
1432
Barbara
Moloney
8/22/17
1433
Jesse
Monahan
8/21/17
1434
Joanne
Moncada
8/29/17
1435
Gail
Moore
8/29/17
1436
Robert
Moore
8/24/17
1437
Thomas
Moore
8/29/17
1438
Anne
Mor
8/22/17
1439
Sylvia
Morais
8/22/17
1440
Will
Morel
8/29/17
1441
Teresa
Morelle
8/18/17
1442
Dennis
Morley
8/29/17
1443
Lewis
Morrison
8/19/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1444
Janet
Moser
8/24/17
1445
Chelsea
Mozen
8/7/17
1446
Norine
Muhfeld
8/10/17
1447
James
Mulder
8/29/17
1448
Ellen
Mulkerin
8/22/17
1449
Mary
Mullaney
8/22/17
1450
Monuca
Mulligen
8/29/17
1451
Dory
Munder
8/30/17
1452
Laura
Munisteri
8/22/17
1453
Eric
Munkelt
8/30/17
1454
Eric
Munkelt
9/1/17
1455
Maki
Murakami
8/29/17
1456
Lizzie
Murchison
8/29/17
1457
Susan
Murphy
8/21/17
1458
Susan
Murphy
8/29/17
1459
Dara
Murray
8/29/17
1460
William
Murtha
8/29/17
1461
Michael
Musante
8/23/17
1462
Roger
Muzii
8/29/17
1463
Lindsey
Muzzio
8/29/17
1464
Carol
Myers
8/24/17
1465
Emma
Myers
8/31/17
1466
Laura
Myerson
8/24/17
1467
Sandra
Naidich
8/18/17
1468
S.
Nam
8/18/17
1469
Courtney
Nandagiri
8/24/17
1470
Jean
Naples
8/7/17
1471
P.
Naprstek
8/31/17
1472
Gretchen
Nau
8/22/17
1473
Rosemary
Neer
8/21/17
1474
Lisa
Neste
8/29/17
1475
Eric
Neuman
8/21/17
1476
Lynn
Neuman
8/29/17
1477
Ted
Neumann
9/1/17
1478
John
Neumeister
8/21/17
1479
John
Neumeister
9/1/17
1480
Bob
Nevelus
8/30/17
1481
Roxie
Newberry
8/30/17
1482
Antonella
Nielsen
8/29/17
1483
Anthony
Nigro
8/29/17
1484
Carla
Ninos
8/28/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1485
Sajendra
Nithiananthan
8/29/17
1486
Joseph
Nitzberg
9/1/17
1487
Mary
Noll
8/24/17
1488
Lauren
Noonan
8/21/17
1489
Terry
Nord
8/22/17
1490
Mary Ann
Nordheimer
8/29/17
1491
liana
Novick
8/29/17
1492
Laura
Nowack
8/28/17
1493
Natalie
Nussbaum
8/29/17
1494
Kathy
Oconnor
8/28/17
1495
Mary Beth
OConnor
8/29/17
1496
Patricia
Odell
8/29/17
1497
Cynthia
Ofer
8/29/17
1498
Kerry
O'Flynn
9/1/17
1499
Barb
OFriel
9/1/17
1500
Elizabeth
O'Hara
8/29/17
1501
William
O'Hearn
8/29/17
1502
Luis
Olavarria
9/1/17
1503
Margot
Olavarria
8/16/17
1504
Kevin
Oldham
8/19/17
1505
Joseph
Olejak
8/23/17
1506
Victoria
Oltarsh
8/25/17
1507
Carole
Osterink
8/30/17
1508
Linde
Ostro
8/25/17
1509
Joseph
O'Sullivan
8/21/17
1510
Tara
O'Sullivan
9/1/17
1511
Jane
Osuna
9/1/17
1512
Marge
Othrow
8/24/17
1513
Maxwell
Owen
8/30/17
1514
Michael
Owen
8/29/17
1515
Roseanne
Pacheco
8/22/17
1516
Linda
Pachter
8/29/17
1517
Sarah
Page
9/1/17
1518
Harela
Paglia
8/21/17
1519
Vic
Paglia
8/7/17
1520
Carol
Painter
8/21/17
1521
Laura
Pakaln
8/22/17
1522
Tami
Palacky
8/29/17
1523
Anne
Palagano
8/22/17
1524
Craig
Palmer
8/29/17
1525
Julie
Palmeri
8/22/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1526
Charlie
Pane
9/1/17
1527
Drew
Panko
8/24/17
1528
Laura
Pantazis
8/29/17
1529
John
Papandrea
8/24/17
1530
Joan
Paris
8/21/17
1531
Pat
Pascual
8/18/17
1532
Michael
Pastore
9/1/17
1533
Jacob
Patenaude
8/22/17
1534
Randolph
Patrick
9/1/17
1535
Ernest
Paviour
8/18/17
1536
Anrea
Payne
9/1/17
1537
Gail
Payne
8/24/17
1538
Jennifer
Paynter
8/21/17
1539
Barbara
Pearson
8/7/17
1540
Pippa
Pearthree
8/29/17
1541
Robert
Pease
8/21/17
1542
Mary
Peck
8/30/17
1543
Melanie
Pedicini
8/7/17
1544
Annadora
Pedro
8/22/17
1545
Susan
Pelosi
8/30/17
1546
Vickiana
Pena
8/28/17
1547
Eliane
Pereira
8/24/17
1548
Martha
Perlmutter
8/18/17
1549
Richard
Perras
8/18/17
1550
Robert
Perretti
8/7/17
1551
Tony
Perrottet
8/17/17
1552
Debbie
Peters
8/29/17
1553
Laura
Petit
8/22/17
1554
Joe
Pfister
8/18/17
1555
Gaelene
Phelps
8/29/17
1556
Gaelene
Phelps
9/1/17
1557
Trent
Philipp
8/24/17
1558
Brother Robert
Pierson OHC
8/22/17
1559
Jon
Pike
8/30/17
1560
Thomas
Pintagro
8/29/17
1561
Debra
Plishka
8/29/17
1562
Jane
Podell
8/22/17
1563
Albert
Poland
8/25/17
1564
Jack
Polonka
8/18/17
1565
Marian
Pompa
8/31/17
1566
Charles
Pompey
8/22/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1567
Tyler
Poniatowski
8/29/17
1568
Bernadette
Powis
8/29/17
1569
Diane
Praus
8/19/17
1570
Ralph
Preiss
8/21/17
1571
Spencer
Prevallet
8/13/17
1572
Elysee
Price
8/29/17
1573
Lou
Priem
8/19/17
1574
Richard
Procida
8/24/17
1575
Camala
Projansky
8/25/17
1576
Clifford
Provost
8/8/17
1577
Lise
Prown
8/24/17
1578
Nicholas
Prychodko
8/24/17
1579
David
Prystal
8/29/17
1580
Laurie
Puca
8/27/17
1581
Katy
Purtee
9/1/17
1582
Katheryn
Quick
8/21/17
1583
Diane
Quinn
8/21/17
1584
Edythe Ann
Quinn
8/29/17
1585
Mary
Quinn
8/29/17
1586
Joseph
Quirk
8/28/17
1587
Laura
Rabinow
8/23/17
1588
Tracy
Raczek
8/8/17
1589
Mary
Rader
8/31/17
1590
Joann
Ramos
8/7/17
1591
Hale
Randers-Pehrson
8/20/17
1592
Edward
Rashba
8/22/17
1593
Andrew
RatZin
9/1/17
1594
Marie
Rayho
8/30/17
1595
Jeff
Reagan
8/22/17
1596
Lobi
RedHaw
8/29/17
1597
Joyce
Reeves
8/29/17
1598
Lenore
Reeves
8/29/17
1599
Pam
Rehm
8/29/17
1600
Cynthia
Reichman
8/29/17
1601
Michael
Reichman
8/29/17
1602
Mary
Reilly
8/21/17
1603
John
Reimnitz
8/31/17
1604
Josephine
Reina
8/22/17
1605
Edward
Rengers
8/29/17
1606
Beth
Renner
8/29/17
1607
Beth
Renner
8/30/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
1608
Beth
Rennig
8/30/17
1609
Athena
Resch
8/23/17
1610
Haleigh
Reutershan
8/22/17
1611
Cathy
Revis
8/7/17
1612
Cathy
Revis
8/22/17
1613
Annia
Reyes
8/23/17
1614
Adelaide
Reynolds
8/29/17
1615
Thomas
Reynolds
8/24/17
1616
Robert
Rice
8/16/17
1617
Frederich
Rich
8/25/17
1618
Amanda
Richards
8/24/17
1619
Kathleen
Richardson
8/7/17
1620
Diana
Riddle
8/29/17
1621
George
Riggs
8/24/17
1622
James
Riley
8/29/17
1623
Kelly
Riley
8/29/17
1624
Dianne
Rinaldi
8/31/17
1625
Melissa
Rinzler
8/29/17
1626
Diane
Rios
8/30/17
1627
Elaine
Risch
8/22/17
1628
Barbara
Riso
9/1/17
1629
Javier
Rivera
8/24/17
1630
Renee
Rizzo
8/18/17
1631
Krystal
Roach
8/27/17
1632
Chuck
Roberts
8/29/17
1633
Cynthia
Roberts
8/22/17
1634
Marcia
Robinson
8/18/17
1635
Robert
Robinson
8/30/17
1636
Iris
Rochkind
8/19/17
1637
Zachary
Rodgers
8/22/17
1638
Heriberto
Rodriguez
9/1/17
1639
Sylvia
Rodriguez
8/16/17
1640
Lily
Rodulfo
8/22/17
1641
Robert
Rogers
8/24/17
1642
Johanna
Rose
9/1/17
1643
Stephen
Rose
8/24/17
1644
Chris
Rosen
8/29/17
1645
Jenny
Rosenthal
8/23/17
1646
Robert
Rosenthal
8/18/17
1647
Suzie
Ross
8/21/17
1648
Timothy
Rosser
8/7/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1649
Janice
Rossi
8/24/17
1650
Jodie
Rossi
8/22/17
1651
Li via
Rossi
8/30/17
1652
Janice
Rost
8/29/17
1653
Janice Arlene
Rost
8/22/17
1654
Rochelle
Rothbaum
8/23/17
1655
Margery
Rothenberg
8/22/17
1656
Christina
Rousseau
8/29/17
1657
Wileen
Rowley
9/1/17
1658
Rebecca
Roy
8/30/17
1659
Jonathan
Rubin
8/19/17
1660
Paul
Rubin
8/16/17
1661
Karen
Rubino
8/29/17
1662
Helena
Rudd
8/16/17
1663
Rosalee
Ruediger
8/21/17
1664
Vincent
Rusch
8/29/17
1665
Mike, Pat
Ruscigno, Hilliard
8/31/17
1666
Paul
Russell
8/21/17
1667
Samantha
Russo
8/20/17
1668
Seth
Rutman
8/19/17
1669
Megan
Ryan
8/29/17
1670
Elaine
Sacco
9/1/17
1671
Marysa
Sacerdote
8/30/17
1672
Emma Lou
Sailors
8/24/17
1673
Diana
Salsberg
8/28/17
1674
Laurie
Salzberg
8/31/17
1675
Ahide
Sanchez
8/31/17
1676
Dominick
Santise
8/29/17
1677
Mary
Sari
8/29/17
1678
Carolyn
Sas
9/1/17
1679
Daniel
Savatteri
8/30/17
1680
Jason Douglas
Saville
8/22/17
1681
Marietta
Scaltrito
8/24/17
1682
Chris
Scanga
8/28/17
1683
Christopher
Scanga
9/1/17
1684
Kelley
Scanlon
8/24/17
1685
Martin
Schabu
8/20/17
1686
Wendy
Scheir
8/29/17
1687
Joan
Schildwachter
8/29/17
1688
Elaine
Schindler
8/21/17
1689
Pierre
Schlemel
8/24/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1690
Erica
Schmidt
8/29/17
1691
Naomi
Schmidt
8/30/17
1692
Chris
Schneebeli
8/29/17
1693
Shirley
Schue
8/29/17
1694
Marthe
Schulwolf
8/22/17
1695
Phillip
Schwartz
8/30/17
1696
Sybil
Schwartzbach
9/1/17
1697
Sabine
Schwarz
8/29/17
1698
Thomas
Scialo
8/7/17
1699
Carina
Scorcia
8/29/17
1700
Amanda
Scott
8/22/17
1701
P.
Scoville
8/7/17
1702
Margaret
Scripp
8/29/17
1703
Shelley
Seccombe
8/31/17
1704
Michael
Seckendorf
8/29/17
1705
Laura
Seitz
8/31/17
1706
Kim
Sellon
8/14/17
1707
Richard
Sena
8/29/17
1708
Yoshihiro
Sergei
8/29/17
1709
Donna
Serpentini
8/30/17
1710
Linda
Sewell
8/7/17
1711
Susan
Shaak
8/21/17
1712
Karen
Shalom
8/22/17
1713
Barbara
Shapiro (Raskopf)
8/29/17
1714
William
Sharfman
8/7/17
1715
William
Sharfman
8/25/17
1716
Janis
Sharkey
9/1/17
1717
Gary
Shaw
8/23/17
1718
Clare
Sheridan
8/21/17
1719
Ian
Sheridan
8/29/17
1720
Samantha
Sherry
8/28/17
1721
Kate
Sherwood
8/24/17
1722
Alice
Shields
8/7/17
1723
Susan
Shockett
8/23/17
1724
Beth
Shortsleeves
8/29/17
1725
Lisa
Shumate
8/19/17
1726
Elizabeth
Shundi
8/22/17
1727
Susan
Sie
8/30/17
1728
Ana
Sierra
8/18/17
1729
Ethan
Signer
8/18/17
1730
Jeffrey
Silman
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1731
Jill
Silverman
8/30/17
1732
Laura
Silverman
8/24/17
1733
Sasha
Silverstein
8/29/17
1734
Virginia
Simek
8/22/17
1735
Beatrice
Simmonds
8/3/17
1736
Eileen
Simon
8/22/17
1737
Norman
Sissman
8/29/17
1738
John
Skelly
8/23/17
1739
Caren
Skibell
8/29/17
1740
Darren
Skotnes
8/29/17
1741
Katherine
Slawinski
8/27/17
1742
Jessica
Smith
8/23/17
1743
Kevin
Smith
8/21/17
1744
Mary
Smith
8/7/17
1745
Melinda
Smith
8/20/17
1746
Vanessa
Smith
8/27/17
1747
Addie
Smock
8/7/17
1748
Virginia
Snider
8/29/17
1749
Elena
Snyder
9/1/17
1750
Sandy
Sobanski
8/24/17
1751
Gillian
Sobocinski
8/27/17
1752
Sabrina
Solomon
8/29/17
1753
David
Sorensen
8/7/17
1754
Nicolai
Soriano
9/1/17
1755
Cynthia
Soroka-Dunn
8/30/17
1756
Deniseadenise
Sossa
8/30/17
1757
Rebecca
Soule
8/29/17
1758
Trevor
Southlea
8/31/17
1759
Harvey
Spears
8/7/17
1760
Leola
Specht
8/7/17
1761
Elaine
Sperbeck
8/29/17
1762
Vanessa
Spiegel
8/30/17
1763
Barry
Spielvogel
8/24/17
1764
Abby
Spitzer
8/21/17
1765
Abby
Spitzer
8/28/17
1766
Stuart
Spolin
8/21/17
1767
Rebekkah
Sprague
8/30/17
1768
Ann
Sprayregan
8/29/17
1769
Judy
St. Hedley
8/24/17
1770
Jane
Stabile
8/21/17
1771
Shannon
Stagman
8/28/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
1772
Anna
Stahlie
8/30/17
1773
Carol
Stamets
8/29/17
1774
Judyth
Stavans
8/30/17
1775
Alex
Stavis
8/16/17
1776
Jean
StClair
8/22/17
1777
Jean
StClair
8/22/17
1778
Fern
Stearney
8/22/17
1779
Doug
Steckler
8/19/17
1780
Deborah
Stedge
9/1/17
1781
Joanne
Steele
8/29/17
1782
Dylan
Stein
8/21/17
1783
Herbert
Stein
8/19/17
1784
Herbert
Stein
8/24/17
1785
Jane
Stein
8/24/17
1786
Lorenz
Steininger
8/29/17
1787
Richard
Stern
8/18/17
1788
Susan
Stevens
9/1/17
1789
Paige
Stevenson
8/22/17
1790
Heather
Stewart
8/28/17
1791
Michael
Stocker
8/7/17
1792
Jill
Stolt
8/22/17
1793
Claudia
Stoltman
9/1/17
1794
Marcia
Stone
8/29/17
1795
Peggy
Stork
8/22/17
1796
Laurie
Storm
8/29/17
1797
James
Strickler
8/23/17
1798
Caroline
Stupple
8/30/17
1799
Moraima
Suarez
8/29/17
1800
Josh
Subin
8/30/17
1801
Anna
Sullivan
8/21/17
1802
Terry
Sullivan
8/29/17
1803
Karen
Sussan
8/30/17
1804
Judith
Swallow
8/22/17
1805
Tami
Swartz
8/29/17
1806
Kathleen
Sweeney
8/28/17
1807
Leslie
Sweeney
8/29/17
1808
Glynis
Sweeny
9/1/17
1809
Alexandra
Sweeton
8/28/17
1810
Michael
Szeto
8/29/17
1811
Sandy
Tabin
8/30/17
1812
Susan
Tabor
8/31/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
1813
Christen
Tallas
9/1/17
1814
Gail
Tauber
8/28/17
1815
Abigail
Taylor
8/30/17
1816
Jason
Taylor
9/1/17
1817
Nancy
Taylor
8/30/17
1818
Margaret
Teahan
8/7/17
1819
Gary
Telfer
8/21/17
1820
Michele
Temple
8/7/17
1821
Edith
Tempi eton
8/29/17
1822
Hannah
Tennant-Moore
8/21/17
1823
Lynne
Teplin
8/18/17
1824
Ron
Tergesen
8/29/17
1825
Rashida
Tewarson
8/22/17
1826
Deborah
Thackrey
8/22/17
1827
Robert
Thibault
8/24/17
1828
Irene
Thiel
8/24/17
1829
Tracy
Thomas
8/22/17
1830
Lorraine
Thompson
9/1/17
1831
James
Thoubboron III
8/25/17
1832
Robert
Tipp
8/30/17
1833
Jo
Toland
8/29/17
1834
Elizabeth
Tolliver
8/22/17
1835
Lynn
Tondrick
8/29/17
1836
Susan
Torres
8/18/17
1837
Joan
Traber
8/22/17
1838
Joanne
Trapanese
8/22/17
1839
Nancy
Traverse
8/22/17
1840
Thomas
Trengove
8/29/17
1841
Adam
Trese
8/30/17
1842
Mary
Troland
8/29/17
1843
Ryan
Trow
8/30/17
1844
Ann
Troxler
8/29/17
1845
Barbara
Trypaluk
8/29/17
1846
Ling
Tsou
8/16/17
1847
Ling
Tsou
8/21/17
1848
Leigh Ann
Tulleson
8/30/17
1849
Alexander
Turkenich
8/29/17
1850
Charity
Turner
8/22/17
1851
Deborah
Turner
8/30/17
1852
Sean A.
Twohig
8/29/17
1853
Francine
Tyler
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
I'irsl Name
l.asl Name
Dale Submitted
1854
Joel
Tyner
9/1/17
1855
Kathy
Upham
8/22/17
1856
Chris
Usami
8/20/17
1857
Nick
Vailakis
9/1/17
1858
Fernando
Valentin
8/14/17
1859
Matthew
Van Brocklin
8/21/17
1860
Brent
Van Dyke
8/30/17
1861
Marcsha
Vander Hey den
8/16/17
1862
Theresa
Vanyo
8/22/17
1863
Patrick
Varekamp
8/24/17
1864
Alexandra
Vargo
9/1/17
1865
Anna
Varney
8/20/17
1866
Joseph M
Varon
8/29/17
1867
Francisco J.
Velez
8/25/17
1868
Joanna
Venditto
8/21/17
1869
Maria
Venidis
8/18/17
1870
Robert
Veralli
8/24/17
1871
David
Verhoff
8/25/17
1872
David
Verhoff
8/28/17
1873
Margaret
Vernon
8/24/17
1874
Paolo
Vidali
8/19/17
1875
Nicole
Vidor
8/30/17
1876
Lauren
Vigna
8/29/17
1877
Harry
Vincent
8/22/17
1878
Richard
Vincent
9/1/17
1879
Jerald
Vinikoff
8/18/17
1880
Andy
Von Salis
8/18/17
1881
Helen
Vose
8/29/17
1882
Carla
Waldron
8/19/17
1883
Ruth
Walker
8/22/17
1884
Steven
Walker
8/29/17
1885
Robert
Waller
8/29/17
1886
Brad
Walrod
8/29/17
1887
Gerald
Walsh
8/18/17
1888
Ruth
Walter
8/21/17
1889
Wendy
Walters
8/24/17
1890
Jonathan
Wang
8/31/17
1891
Eddie
Ward
8/9/17
1892
Ken
Ward
8/19/17
1893
Marc
Ward
8/24/17
1894
Paula
Ward
8/30/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
1895
Bob
Warren
8/22/17
1896
Carol
Warren
8/29/17
1897
Edward
Warren
8/21/17
1898
Dina
Wasserman
8/21/17
1899
Marc
Waters
8/19/17
1900
Eli
Watts
8/18/17
1901
Michael
Watts
9/1/17
1902
Noah
Watts
8/23/17
1903
Clifford
Weathers
8/23/17
1904
Esther
Weaver
8/24/17
1905
Melissa
Weaver
9/1/17
1906
Marie
Webster
8/25/17
1907
Marie
Webster
8/28/17
1908
Annie
Wei
8/30/17
1909
Carmen
Wei
8/18/17
1910
Penelope
Weinberg
8/22/17
1911
Adam
Weinert
8/22/17
1912
Adam
Weinert
8/22/17
1913
Florence
Weintraub
8/15/17
1914
Elaine
Weir
8/18/17
1915
Stana
Weisburd
8/30/17
1916
Marcia
Weiss
8/22/17
1917
Alicia
Weissman
8/21/17
1918
Shaye
Wei
8/21/17
1919
William
Welkowitz
8/29/17
1920
Heather
Wells
8/23/17
1921
Molly
Westbrook
8/30/17
1922
Patrick
Whalen
8/29/17
1923
Ian
Wheeler
8/29/17
1924
Mona
White
8/29/17
1925
Penny
White
8/24/17
1926
Edward B.
Whitney
8/25/17
1927
Wheel ock
Whitney
8/25/17
1928
Teena
Wildman
8/29/17
1929
Kimberly
Wiley
8/24/17
1930
Michael
Wiley
8/31/17
1931
Seth
Wiley
8/30/17
1932
Andrea
Williams
8/29/17
1933
Andrew
Williams
8/24/17
1934
Suzanne
Williams
8/22/17
1935
Nathanel
Williams Jr.
8/29/17
-------
Appendix A - List of Commenters on the Proposed Second Five-Year Review Report:
Individuals
EPA Index
Nil in her
First Name
l.asl Name
Dale Submitted
1936
Thomas
Windberg
8/29/17
1937
Dana
Winkler
8/30/17
1938
Amy
Winter
8/24/17
1939
Marsha
Wiseltier
8/18/17
1940
Ron
Wish
8/24/17
1941
Frederick
Wishner
8/29/17
1942
Andrew and Kathleen
Wittenborn
8/29/17
1943
Ellen
Wolfe
8/7/17
1944
Peter
Wood
8/29/17
1945
Rick
Wood
8/29/17
1946
Veronica
Wood
8/29/17
1947
Sarah
Woodard
8/18/17
1948
Richard
Wright
8/22/17
1949
Richard
Wright
8/29/17
1950
Richard
Wrobel
8/21/17
1951
Yishin
Yang
9/1/17
1952
Donna
Yannazzone
8/22/17
1953
Emma
Young
8/21/17
1954
Jean
Young
8/29/17
1955
Kathy
Young
8/30/17
1956
Kristina
Younger
8/29/17
1957
J
Yuzawa
8/24/17
1958
Phyllis
Zahnd
8/7/17
1959
Susan
Zeiger
8/29/17
1960
Brook
Zelcer
9/1/17
1961
Janet
Zies
9/1/17
1962
Andrea
Zinn
8/29/17
1963
Pamela
Zino
8/30/17
1964
James
Zorn
8/22/17
1965
Carlo
Zucchi
8/29/17
1966
Cordelia
Zukerman
8/21/17
1967
Anonymous
8/18/17
1968
Anonymous
8/18/17
-------
APPENDIX B
DEFERRAL STATEMENT - Supporting Technical Information
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
April 2019
-------
Final Second Five-Year Review Comment
Response for the
Hudson River PCBs Superfund Site
APPENDIX B
DEFERRAL STATEMENT - SUPPORTING TECHNICAL
INFORMATION
-------
Appendix B - Deferral Statement - Supporting Technical Information
Given the limited temporal coverage of post-dredging data currently available, EPA has decided to defer
the Second Five-Year review (FYR) protectiveness determination at this time. Specifically, there are not
enough data available since the completion of dredging in Fall 2015 for EPA to determine at this time
whether the remedy was sufficiently successful in accelerating the reduction of human health and
ecological risks to meet the remedial action objectives of the 2002 Record of Decision (ROD) for
Operable Unit 2 (OU2) of the Site. The ROD anticipated a robust remedial action followed by
"monitored natural attenuation" (MNA) to meet the remedial action objectives. To evaluate and estimate
the remedy's long-term reduction of risk, EPA needs a number of years of post-dredging data that are not
influenced by the dredging activities. While EPA and others have made extensive analyses of the rates of
decline of PCBs in fish tissue during the period prior to dredging, the in-river conditions were extensively
modified by the remedy, and these historical rates are therefore not expected to reflect future conditions.
EPA's decision to defer its determination of protectiveness for the Upper Hudson River (UHR) remedy
recognizes the challenges in determining the post-dredging long-term rates of recovery for PCB levels in
fish throughout the UHR so soon after completion of the dredging. The dredging portion of the remedy
removed more than 70 percent of the PCB inventory, and more than 500 acres of river bottom were
backfilled or capped, dramatically reducing PCB concentrations in the most contaminated areas.
However, only two years of post-dredging fish data are available for review, and these data are still
expected to be impacted by dredging-related sediment conditions and dredging-related disturbances1 (as
opposed to exposure to only post-dredging conditions) since fish concentrations in adult sportfish are
known to reflect uptake over multiple years of exposure.
Fish body burdens of PCBs are the result of several processes involving PCB exposure through prey,
water and sediment; PCB metabolism and depuration; environmental conditions that affect prey
availability; lipid storage in fish; PCB levels in exposure media; and duration of exposure to PCBs. This
last factor means that PCB body burdens in larger (older) fish, in particular, integrate across multiple
years of exposure. Thus, adult sport fish (Bass, Bullhead and Perch) collected in the first and second year
after dredging will have derived a portion of their body burdens from exposure during the period of
dredging. While PCBs continue to decline in the water and sediment as a result of the dredging and
associated capping and backfilling, as documented in the monitoring data, the rates of decline in fish
tissue are confounded in the short-term by the processes mentioned above. Additionally, EPA believes it
is likely that the dredging-related disturbances further increased the year-to-year variability in fish tissue
concentrations, at least in the short-term, making identification of the overall rate of decline more difficult
to discern at this time.
Evidence for the variable nature of fish tissue PCB levels in the Hudson River, and hence the importance
of evaluating trends of decline over a longer period of time, can be observed in the fish monitoring data
obtained prior to the remedy (1998 to 2008). This period represents the MNA period immediately
following completion of EPA's modeling effort for the ROD which was characterized by a typical range
of flow events and declining PCB transport from above Rogers Island. Despite the relatively consistent
environmental conditions, the fish tissue data for PCBs fluctuates from year-to-year, with increases for
short periods, to be followed by periods of more regular decline. An example is provided in Figure B-l
for brown bullhead in River Section (RS) 1. This figure illustrates the variation in annual mean PCB
levels as described. Both wet weight and lipid-normalized concentrations are presented in the figure. Both
1 These would include contaminated sediments exposed while dredging, increased water column concentrations due to
dredging or habitat reconstruction activities, and temporary deposits of resuspended sediment, among others.
Appendix B Deferral Statement - Supporting Technical Information
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site B-l
-------
metrics show occasional substantive deviations (e.g., year 2004) from the overall declining trend
exhibited for the decade prior to dredging. This declining trend was evident at the time of the ROD,
indicating that natural recovery was occurring; the remedy was selected to accelerate this trend.
EPA's statistical analyses of past fish tissue data, as discussed below, indicated that to accurately
represent the actual rate of decline in fish tissue PCB levels, it is necessary to examine the annual record
over extended periods of time that are generally eight or more years. As noted previously, only two years
of post-dredging data (2016 and 2017) are available at this time. The importance of the longer perspective
is further illustrated in Figures B-2 through B-4. Figures B-2 and B-3 show the observed rates of decline
in lipid-normalized PCB concentrations in fish when calculated over successive 5-year windows (e.g.,
1998-2002, 1999-2003), as opposed to a longer 11-year integration of the data (i.e., 1998-2008). As
shown in these figures, the rates of decline based on the five-year windows vary substantially, often
deviating far from the longer-term trend (i.e., up to 3 times faster and as much as 10 times slower, even
indicating net rates of increase in some instances). Note that the 11-year long-term rates reasonably agree
across species (8 to 15 percent per year), suggesting that fish tissue concentrations decline across all five
species at similar rates when viewed over longer periods.
To further support EPA's decision to defer a protectiveness determination for the remedy, EPA conducted
two separate analyses. First, the 1998 to 2008 lipid-normalized concentrations of PCBs in five fish
species from RS 1 were used to develop rates of decline over progressive time windows, specifically 3, 5,
8, 9 and 10-year intervals. The results are shown in a series of plots (Figures B-4a through B-4e). In each
instance, the apparent rate of decline is calculated for each species for each time window (e.g., 1998 to
2000, 1999 to 2001, etc. for 3-year windows; 1998 to 2002, 1999 to 2003, etc. for 5-year windows; and
1998 to 2005, 1999 to 2006, etc. for 8-year windows). These rates are then plotted against the length of
the data window to illustrate the reduction in variability of the rate estimate as the window is extended.
The number of calculable windows becomes small as the length of the window approaches the length of
the data period (11 years). In each diagram, a set of empirical curves has been added to approximately
bound the range of values.
It is evident for each species, that increasing the length of the data window (i.e., the period of available
data) greatly reduces the variability of the estimates, converging on the 11-year average. For all but the
Pumpkinseed, the variability of the estimates for the 8-year window is within +/- 50 percent of the 11-
year rate of decline, indicating that 8 years is likely the minimum period of data needed to assess the fish
trends.
In the second analysis, EPA conducted a power analysis for the ability to detect a trend in fish tissue
concentrations. The analysis, summarized in Master Comment 49, found that approximately of 8 years of
data are needed to detect a declining trend of 8 percent per year with an 80 percent level of confidence. A
longer period is needed if the rate is less than 8 percent per year.
The discussion above focuses on lipid-normalized data, since these data generally show less variation
from year to year. A similar analysis based on wet weight data would yield similar or greater variation in
year-to-year rates of decline, since the rates of decline of wet weight concentrations incorporate variations
in lipid content as well as PCB exposure overtime.
These results have important implications pertaining to drawing conclusions on the remedy. A
determination of the remedy's protectiveness will need to rely heavily on the observed rate of decline of
PCB levels in fish; however, this rate can easily be misrepresented or would at least be highly uncertain if
a short-term assessment of the data formed its basis. Given the history of year-to-year variation in fish
tissue levels during a relatively undisturbed period (1998 to 2008), there is a high likelihood that the
Appendix B Deferral Statement - Supporting Technical Information
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site B-2
-------
apparent rate of decline, influenced by dredging (and the accompanying disturbance to the system), as
measured after only 3 to 5 years of data would not adequately represent the actual rate of decline. As is
evident from Figures B-2, B-3 and B-4, the short-term estimates can both underestimate and overestimate
the actual rate of decline.
Although EPA does not believe it has sufficient post-dredging fish tissue data to support a protectiveness
determination, the information available for sediments and water are not inconsistent with a remedy that
will be protective. The dredging portion of the remedial action was implemented successfully and within
the expectations described in the ROD, substantially reducing PCB inventory and surface sediment
concentrations in the UHR. Source control actions at the former GE plant and the reductions in sediment
PCBs from the dredging have also led to declines in surface water concentrations in the Upper Hudson.
EPA is anticipating a similar reduction in PCB levels in fish in the early portion of the post-dredging
MNA period, followed by continued but more gradual declines in fish tissue concentrations during the
later post-dredging MNA period.
EPA carefully considered over 2,000 comments provided by the public on the Proposed Second FYR
Report. Many of the comments focused on the rates of recovery in fish, sediment and water. By adjusting
to deferring a protectiveness determination, EPA acknowledges that there are limitations and some
uncertainty in the existing data. It is also important that EPA have more data before a protectiveness
determination is made. Therefore, EPA has decided to defer its decision on the protectiveness of the
remedy until the agency is able to obtain sufficiently reliable, longer term estimates of the rates of decline
of PCB levels in fish tissue in the UHR.
EPA will not consider the OU2 remedy to be complete until the natural attenuation component also has
been completed and the remedial action objectives are met.2
2 Note: There is no inconsistency between this statement and EPA's decision to issue a Certification of Completion
of the Remedial Action to GE under the 2006 Consent Decree. The term "Remedial Action" has a specific
meaning in the Consent Decree. Importantly, the term does not include Operation, Maintenance and Monitoring
(OM&M). While the post-dredging monitored natural attenuation period is a key explicit part of the remedy, it is
part of the OM&M rather than the "Remedial Action" activities under the Consent Decree. In the Consent
Decree, the term "Remedial Action" refers to the dredging itself and the associated construction work by GE
(principally, the capping, backfilling, habitat reconstruction and later decommissioning of the sediment processing
facility). GE remains responsible for carrying out all of the OM&M under the Consent Decree.
Appendix B Deferral Statement - Supporting Technical Information
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site B-3
-------
Brown Bullhead at RS 1
Wet Weight
Brown Bullhead at RS 1
Lipid-Normalized
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Year
Year
Notes:
• Data shown as annual mean and 95th percentile confidence interval on the mean.
£^2 - HudsfmT'River
Trend of Tri+ PCB Concentrations for Brown Bullhead in RS 1
wet weight and lipid-normalized basis
Figure B-l
April 2019
-------
10
Rates of Decline for Fish Tissue
1998 to 2008
Lipid-normalized Fish Tissue Concentrations in RS 1
J I L
•99 *"03
£ £
104 *'08
CD
0)
CI
c
1*0
-C
u
o
CD
+->
ro
C£L
IB
3
c
c
<
-10 -
-20 "
•98*102
'00 *w
'01 *105
"99 * X)3
'02 * "06
¦03 - f)7
*"04 - '08
"98 * '02
m * '07
'01 -TOtTO-TO
"04* r08
*W*'02
'02 - "06 * "03 - '07
F99 * "03
*01 • '05
"02 * ,06
'00*-04
'01 * *05
•38 - '02 S
'03 -'07
¦04*X)8
"96 * '02
"99*'03
r0'D * '04
'03 - '07* '04 - "08
Ci
c
"in "C
ro C
CD OJ
V-
U (-
0)
Q
-30 -
-40
Legend
Average Rate 1998-2008
¦M+D3 Average for 5 year interval
~T
Brown
Bullhead
T"
Large-mouth
Bass
WW
T
Small-mouth
Bass
Pumpkinseed
¦onus
~T
Yellow
Perch
Variation in Rate of Decline using 5-year Averaging Interval:
Lipid-normalized Concentrations in RS 1
Figure B-2
April 2019
-------
Rates of Decline for Fish Tissue
10
-10 -
fO
CD
s
o^w
O)
cn
c
(T3
-C
u
o
a>
i—
u (—
0)
a
V
Variation in Rate of Decline using 5-year Averaging Interval:
Lipid-normalized Concentrations in RS 2
Figure B-3
April 2019
-------
ai
+¦>
a)
»
(0
U
E
"7
cn
£
o
400%
300%
200%
100%
0%
-100%
-200%
re
>
q -300%
-400%
ro
cc
c
ro
(D
c
CD
+-1
00
ro
ro
cr
c
ro
(D
c
o
Rates of Decline vs. Years of Data Available
Brown Bullhead, lipid-normalized data, RS 1
x •.
x
X
X
X
X
XX
X
X
Legend
x Rate of decline
based on 3,5, 8, 9 or
10 year data interval
¦x*
x
11-year Rate
= -9.1%/yr
8
10
12
Notes: Years of Data Available for Averaging
Deviation from long-term average rate (y) was calculated as the relative change of the short-term average rate to the 11-year average rate.
As an example, the symbols at the five-year interval on the X- axis represent the rates calculated for the following intervals: 1998 to 2002, 1999 to 2003, 2000 to 2004, 2001 to 2005, 2002 to
2006, 2003 to 2007 and 2004 to 2008, resulting in seven separate estimates of the decay rate, represented by the seven x's on the graph at 5 years.
Dotted lines are empirical lines to show the approximate decline in variance with increasing number of years for averaging.
Note that a positive deviation of 100% is equal to a decay rate that is twice as fast as the 11 year rate, whereas a negative deviation of 100% is equal to a decay rate of 0%/year (a flat line
trend).
The data used in this figure represents the 11-year period, 1998 to 2008.
£^2 - HudsfmT'River
Deviation from Long-Term Average Rate vs. Years of Data Available
Brown Bullhead, lipid-normalized data, RS 1
Figure B-4a
April 2019
-------
200%
^ 150%
»
(0
U
E
"7
cn
£
o
100%
50%
0%
-50%
-100%
re
>
q -150%
-200%
ro
cc
c
ro
(D
c
ro
cr
c
ro
CD
c
o
Rates of Decline vs. Years of Data Available
Large-mouth Bass, lipid-normalized data, RS 1
x \
X
Legend
x Rate of decline
based on 3,5, 8, 9 or
10 year data interval
x
x
x
X
X
X
X
X
X
X
11-year Rate
= -14.9 %/yr
8
10
12
Notes: Years of Data Available for Averaging
Deviation from long-term average rate (y) was calculated as the relative change of the short-term average rate to the 11-year average rate.
As an example, the symbols at the five-year interval on the X- axis represent the rates calculated for the following intervals: 1998 to 2002, 1999 to 2003, 2000 to 2004, 2001 to 2005, 2002 to
2006, 2003 to 2007 and 2004 to 2008, resulting in seven separate estimates of the decay rate, represented by the seven x's on the graph at 5 years.
Dotted lines are empirical lines to show the approximate decline in variance with increasing number of years for averaging.
Note that a positive deviation of 100% is equal to a decay rate that is twice as fast as the 11 year rate, whereas a negative deviation of 100% is equal to a decay rate of 0%/year (a flat line
trend).
The data used in this figure represents the 11-year period, 1998 to 2008.
£^2 - HudsfmT'River
Deviation from Long-Term Average Rate vs. Years of Data Available
Lar^e-mouth bass, lipid-normalized data, RS 1
Figure B-4b
April 2019
-------
700%
600%
5 500%
"S 400%
cc
6 300%
re
| 200%
<
E
k.
"7
cn
c
o
100%
0%
-100%
-700%
ro
cc
c
ro
(D
c
(D
+-1
00
ro
E -200%
o
^ -300%
c
•+3 -400%
(0
a -500%
O
-600%
ro
cc
c
ro
(D
c
_o
LO
0
Rates of Decline vs. Years of Data Available
Small-mouth Bass, lipid-normalized data, RS 1
Legend
x Rate of decline
based on 3,5, 8, 9 or
10 year data interval
x
x
X
X
*
X
11-year Rate
= -13.2 %/yr
8
10
12
Notes Years of Data Available for Averaging
Deviation from long-term average rate (y) was calculated as the relative change of the short-term average rate to the 11-year average rate.
As an example, the symbols at the five-year interval on the X- axis represent the rates calculated for the following intervals: 1998 to 2002, 1999 to 2003, 2000 to 2004, 2001 to 2005, 2002 to
2006, 2003 to 2007 and 2004 to 2008, resulting in seven separate estimates of the decay rate, represented by the seven x's on the graph at 5 years.
Dotted lines are empirical lines to show the approximate decline in variance with increasing number of years for averaging.
Note that a positive deviation of 100% is equal to a decay rate that is twice as fast as the 11 year rate, whereas a negative deviation of 100% is equal to a decay rate of 0%/year (a flat line
trend).
The data used in this figure represents the 11-year period, 1998 to 2008.
£^2 - HudsfmT'River
Deviation from Long-Term Average Rate vs. Years of Data Available
Small-mouth bass, lipid-normalized data, RS 1
Figure B-4c
April 2019
-------
700%
600%
§£ 500%
400%
300%
(0
I 200%
<
E
k.
"7
cn
c
o
_l
E
o
100%
0%
-100%
-200%
"¦ -300%
c
+3 -400%
(0
v -500%
O
-600%
-700%
ro
cc
c
05
CD
C
o
Rates of Decline vs. Years of Data Available
Pumpkinseed, lipid-normalized data, RS 1
x*.
x
x
Legend
x Rate of decline
based on 3,5, 8, 9 or
10 year data interval
JL.
X
-------
600%
500%
400%
Cp
0s
+¦>
* 300%
u
£ 200%
>
^ 100%
¦?
cn
o -100%
0%
£
o
re
• ¦
>
U
O
-600%
-200%
-300%
-400%
-500%
ro
cc
c
ro
(D
c
ro
cr
c
ro
(D
c
(D
£
o
0
Rates of Decline vs. Years of Data Available
Yellow Perch, lipid-normalized data, RS 1
Legend
x Rate of decline
based on 3,5, 8, 9 or
10 year data interval
x
x
X
X
X
X
•x*
4-
"*•
11-year Rate
=-14 %/yr
8
10
12
Notes: Years of Data Available for Averaging
Deviation from long-term average rate (y) was calculated as the relative change of the short-term average rate to the 11-year average rate.
As an example, the symbols at the five-year interval on the X- axis represent the rates calculated for the following intervals: 1998 to 2002, 1999 to 2003, 2000 to 2004, 2001 to 2005, 2002 to
2006, 2003 to 2007 and 2004 to 2008, resulting in seven separate estimates of the decay rate, represented by the seven x's on the graph at 5 years.
Dotted lines are empirical lines to show the approximate decline in variance with increasing number of years for averaging.
Note that a positive deviation of 100% is equal to a decay rate that is twice as fast as the 11 year rate, whereas a negative deviation of 100% is equal to a decay rate of 0%/year (a flat line
trend).
The data used in this figure represents the 11-year period, 1998 to 2008.
£^2 - HudsfmT'River
Deviation from Long-Term Average Rate vs. Years of Data Available
Yellow Perch, lipid-normalized data, RS 1
Figure B-4e
April 2019
-------
Final Second Five-Year Review Comment
Response for the
Hudson River PCBs Superfund Site
APPENDIX C
TECHNICAL MEMORANDUM
EVALUATION OF FIELD, KERN AND ROSMAN (2016)
Prepared by:
Louis Berger US, Inc.
&
LimnoTech
April 2019
-------
Some comments on the draft Five-Year Review (FYR) report asserted that EPA's models of the
Hudson River are no longer considered to be scientifically valid, citing as evidence a paper
(Field, Kern, and Rosman 2016) senior authored by L. Jay Field, NOAA Office of Response and
Restoration. In this paper, the authors developed a regression-based replica of EPA's models (an
"emulation"), updated levels and trends in surface sediment concentrations (an "updated
emulation scenario"), and predicted lengthy delays in Lower Hudson River (LHR) fish recovery
times relative to forecasts made with EPA's models. EPA disagrees with the assertion that its
models are not scientifically valid, and with the commenters' citation of the NOAA paper as the
basis for the comments. Specifically, EPA has reviewed NOAA's use of an "updated emulation
scenario" to estimate recovery times of LHR fish, as presented in Field et al., and finds it to be
unreliable for the following reasons:
• The authors changed the value of a key input to their baseline emulation (surface
sediment PCB concentration) without recalibrating its relationship to closely linked
model outputs (water column and fish tissue PCB concentrations);
• This resulted in a substantial upward bias in their simulations of water column and fish
tissue PCBs, which can be readily seen in comparisons of model to data; and
• That upward bias was the main factor accounting for the lengthy recovery times that they
forecasted, as opposed to their additional assumption of a slower sediment recovery
trend.
The NOAA baseline emulation is not a new model, but is an approximate replication of results
from EPA's models, and consists of a set of simple statistical correlations between EPA's
predictions of Upper Hudson River (UHR) sediment and water column PCBs and Lower Hudson
River fish tissue PCBs. The statistical correlations that constitute NOAA's baseline emulation
model are closely fit to EPA's Monitored Natural Attenuation (MNA) and Selected Remedy
forecasts, as the authors discuss in their Appendix A1. In describing their "updated emulation
scenario," however, they stated in their Abstract that their study "applied model emulation to
evaluate the impact of updated sediment concentrations in the original mechanistic model
projections of time to reach risk-based thresholds in fish in the LHR." This statement implies
that the authors show the output that one would obtain by updating sediment concentrations in
EPA's models, but that is not true: substituting updated sediment concentrations for baseline
sediment inputs in NOAA's statistical emulation produces a shift in the outputs of the statistical
equations away from the water and fish data to which the underlying models were calibrated, but
without recalibrating. Without recalibration, the NOAA "updated emulation scenario" cannot be
accepted as reliable and, in fact, its results are inconsistent with observed data. This failure to
recalibrate should have been pointed out during the peer review process, and remedied before the
paper was accepted for publication.
These points are briefly summarized below and demonstrated in more detail in the body of this
Technical Memorandum.
1 The outputs taken from the Selected Remedy simulation to fit the emulation equations were limited to the post-
dredging period.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-l
-------
The EPA models (HUDTOX and FISHRAND) were calibrated to all of the historical datasets for
water column, sediment and fish tissue PCB concentrations for 1977-1998, and then used to
conduct 70-year forecast simulations for MNA for 1998-2067. Field et al. (2016) developed a
baseline emulation of EPA's models using regression equations and "updated" the inputs to
those regressions by modifying the inputs to their set of regressions. The particular "updated
emulation scenario" that they say provides the best estimates of long-term recovery times for
Lower Hudson River fish assumes a one-time increase in Upper Hudson River surface sediment
concentration in 2003, based on the 2002 to 2005 Sediment Sampling and Analysis Program
(SSAP) dataset (averaged to 2003), and also assumes a decrease (to 3 percent) in the assumed
rate of recovery of sediment. The discussion of NOAA results that follows focuses on this
"updated emulation scenario," but the overriding conceptual criticism of updating inputs to their
baseline emulation without recalibration applies to each updated scenario presented in the paper.
With regard to surface sediment PCB concentrations, temporal trends are complicated by the fact
that none of the past sediment sampling programs were designed and implemented in a manner
which allows for a consistent analysis of temporal trends. Appendix 4 of the Second FYR
addresses the complications and limitations in the long-term sediment data as part of the analysis
of temporal trends in sediment PCB concentrations in the UHR.
NOAA's "updated emulation scenario" errs conceptually by re-assigning the values of a key
model state variable (surface sediment PCB concentration) in 2003, after five years of the 70-
year MNA forecast simulation, and re-starting the simulation at 2003 without re-calibrating the
original underlying EPA models. This re-assignment interjects an artificial discontinuity in 2003
that fractures the connection between the original EPA model calibrations and the results of
NOAA's "updated" projections from 2003 onward. Because the re-assignment of surface
sediment PCB concentrations in 2003 represents substantial up-scaling (by factors of 2 to 6,
depending on location), NOAA's "updated" projections for water column and fish tissue PCBs
are biased substantially high relative to the observed fish tissue PCB data in 2003, when this
upscaling factor is applied. This substantial upward bias in predicted wet weight fish tissue PCB
concentration is the main reason that Field et al. (2016) predicted longer times to reach specific
risk thresholds (such as 0.2 mg/kg wet weight) than the EPA models.
In contrast, the EPA models that supported the Record of Decision (ROD) for the Upper Hudson
River Superfund Site successfully reproduced observed data for UHR water and fish PCB
concentrations from 1977 to 1998, were successfully peer-reviewed as part of the Superfund
process, and those models continued to closely match observed trends in UHR water and fish
PCB concentration data through the extended 1998 to 2008 period of MNA, as shown in in
Appendices 1 and 3 of the FYR. Field et al. (2016) did not present an "updated" emulation of
UHR fish tissue PCB concentrations, so their paper does not offer an alternative to EPA's
projections of UHR fish tissue recoveries; the paper is focused exclusively on Lower Hudson
River fish recovery times.
EPA shares the concern about slower-than-expected fish tissue recoveries in the LHR that was
expressed by some commenters. The decline in recovery rates with distance from the UHR was
shown in Figure A-16B of Appendix 3 and discussed there. For the LHR, EPA employed
HUDTOX to project PCB loadings from the UHR to the LHR. Those loadings were input to the
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-2
-------
Farley Model, which projected LHR water and sediment PCB concentrations, and FISHRAND
was then used to project LHR fish tissue concentrations. As shown in Appendix 3, EPA's models
tended to under-predict fish tissue PCB concentrations at River Mile (RM) 90 (Kingston) and
RM 50 (West Point) for the period 2004 to 2008, whether model and data are compared in wet
weights or after lipid normalization, and also showed a faster rate of decline for this period than
indicated by data. Note, however, that FISHRAND was not explicitly calibrated to LHR fish
data; instead, the model as calibrated for the UHR was directly applied to the LHR.
Specification of inputs for sediment and water exposure concentrations is central to the
application of the FISHRAND model. Insufficient historical LHR water column PCB data
existed at the time of the ROD to support a calibration of Farley Model water column PCB
predictions, and Appendix 1 shows that the Farley model under-predicted water column PCBs at
Poughkeepsie (RM 75) for the period 2004-2008. Consequently, water column PCBs may have
been under-predicted for earlier time periods as well, contributing to potential mismatches
between predicted and observed fish tissue concentrations in the FISHRAND model. In addition,
while PCB loadings from the UHR to the LHR are well characterized, their contribution to
sediment and water exposure concentrations experienced by LHR fish are less well understood.
EPA is committed to additional studies of fish tissue recovery trends in the LHR and the factors
that affect those trends.
Notwithstanding EPA's concern about LHR fish tissue recovery rates, EPA does not accept
NOAA's "updated emulation scenario" as a reliable predictive tool for the LHR because, and as
shown below, the PCB exposure concentrations in this model are substantially biased and
inconsistent with observed data.
I lated Emulation Scenario is Not Calibrated and Does Not Match Observed Data
Field et al. (2016) relied on an "updated emulation scenario," which included up-scaled sediment
concentrations in the UHR, to reach conclusions about time to meet fish tissue recovery targets
in the LHR. Their baseline emulation model has a simplified structure that links key HUDTOX
and FISHRAND outputs using regressions, without explicitly representing the complex
interactions between sediment, water, and the food chain. They developed their baseline
emulation model as follows:
• They first fitted an exponentially declining time trend to the simulated surface sediment
PCB concentrations in EPA's ROD MNA forecast. The time series that they developed
had a recovery rate (i.e. annual rate of decline) of about 8 percent per year;
• They then used that emulation of EPA's surface sediment forecast series as an
independent variable, in regressions, to predict EPA's ROD MNA forecast values of
water column PCBs.
• Finally, they used those emulated water column concentration forecasts for one Upper
Hudson River location (Waterford) as an independent variable, in regressions, to predict
EPA's ROD MNA forecasts of wet weight fish tissue PCBs at four Lower Hudson River
locations.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-3
-------
This completed their baseline emulation model, a linearization which closely tracked the
HUDTOX and FISHRAND MNA forecasts and, indeed, could not exist without these underlying
mechanistic forecasts. The baseline emulation model links EPA's key outputs to each other using
a statistical correlation, as opposed to a representation of the underlying physical, chemical, and
biological processes. FISHRAND predicts median wet-weight concentrations in each year for
each species and location, based on assumed a priori population distributions of percent lipid,
and the NOAA emulation also predicts median wet weight concentrations for each species and
location for each year, for the same assumed percent lipid distribution.
NOAA then developed their "updated emulation scenario" as follows:
• They altered the sediment forecast time series in 2003 by up-scaling sediment PCB
concentrations by factors of 2 to 6 (depending on location) to match their own estimates,
which were based on the 2002 to 2005 SSAP sediment data;
• They then projected their re-assigned sediment PCB concentrations forward from 2003 at
a declining 3 percent per year rate of recovery, based on their fit to a trend relating their
interpretation of 1991 and 2002 to 2005 data for sediment PCB concentrations;
• Finally, they plugged the "updated" sediment values into the regressions described above
to produce "updated" predictions of water column and wet weight fish tissue PCB
concentrations, thus completing their "updated emulation scenario." Their Table A.2
provides coefficients for the regressions relating sediment to water column PCB and their
Table S-l provides the coefficients for their equations that predict wet weight fish tissue
concentrations.
The NOAA "updated emulation scenario" differs from their baseline emulation scenario in two
important ways. First, sediment PCB concentrations in 2003 were up-scaled by factors of 2 to 6;
and second, assumed recovery rates from 2003 forward were down-scaled from 8 percent to 3
percent per year. This "updated scenario" results in new predicted water column and fish tissue
PCB concentration predictions throughout the entire system from 2003 onward, and shifting
away from contemporaneous water and fish data that without a mechanistic basis or
understanding.
Extensive water column and fish PCB data are available for multiple UHR and LHR monitoring
stations for 2003 to 2008 to evaluate and test the accuracy of the NOAA "updated emulation
scenario" and assess the effect of these assumptions on previously calibrated relationships. The
NOAA authors presented no such tests for water column concentrations, or for the wet weight
fish tissue concentrations predicted by their "updated scenario." These tests are presented in the
next section of this Technical Memo in the form of comprehensive model-data comparisons for
the outputs of the NOAA "updated emulation scenario," including four Upper Hudson River
water column sampling locations and for the four fish species that NOAA simulated at three
Lower Hudson River locations.
The schematics below illustrate the difference between EPA's mechanistic models and the
NOAA emulation of those models, before and after "updating" inputs. Figure C-l represents
EPA's models. Multiple complex physical, chemical and biological processes that affect
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-4
-------
sediment, water, and fish tissue PCBs in the Hudson River are represented and linked. PCBs in
sediment and water, and their interactions, are represented in HUDTOX, and processes affecting
fish tissue PCBs are represented in FISHRAND, which relies on predicted sediment and water
column exposures at appropriate spatial and temporal scales. In the development of EPA's
models, all simulations were constrained by the available sediment, water column, and fish tissue
data for 1977 to 1998. Parameters governing the processes represented in HUDTOX and
FISHRAND were calibrated to achieve consistency of historical simulations with data.
Modeled Processes Simulations Data
•
•
—*—»——
•
Hp
•
•
•
•
Hp
^^
•
•
• • .
Figure C-l: Schematic of Linkages between EPA Model Representations of Sediment,
Water, and Fish, and Sediment, Water, and Fish Data
Figure C-2 illustrates the structure of the baseline emulation model implemented by Field et al.
They predicted HUDTOX UHR water column PCB forecast outputs using HUDTOX surface
sediment PCB forecast outputs. Similarly, they predicted FISHRAND fish tissue PCB forecast
outputs in the LHR using only HUDTOX water column predictions of PCB concentrations at
Waterford. These predictions for water column and fish tissue PCBs were based on regression
equations that do not attempt to represent important elements of the UHR conceptual site model.
Unlike EPA's models, the NOAA model does not relate fish tissue concentration to sediment
exposure, but uses only Upper Hudson River water column PCB concentration as a simplified
predictor of Lower Hudson River fish PCB concentration. Further, the NOAA model does not
represent home ranges or include seasonal fluctuations in predicting fish PCB exposure, as does
EPA's FISHRAND model. The NOAA model only predicts fish tissue PCB levels for LHR
stations beginning at RM 152 (Albany/Troy). It uses annual average water column PCB
concentrations at Waterford as the driving exposure concentrations to predict fish tissue
concentrations as far south as RM 50 (West Point). Thus, a water column station in the Upper
Hudson River was used to represent Lower Hudson River fish exposures over a range of more
than 100 miles.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-5
-------
Modeled Processes
Sediment
Fish Tissue
Field et al. Emulation
Data
Figure C-2: Schematic of NOAA Baseline Emulation of EPA's MNA Model (MNA1)
The "update" that Field et al. (2016) introduced to their emulation model involved up-scaling
UHR sediment PCB concentrations in 2003 by factors of 2 to 6 and down-scaling assumed UHR
sediment recovery rates from 2003 onward. This is illustrated conceptually in Figure 3 which
depicts an altered surface sediment PCB concentration time series in response to the insertion of
a new surface sediment data point that was not part of the original HUDTOX calibration dataset.
Sediment
Modeled Processes
Water
Fish Tissue
Field et al. Emulation
Data
•
•
—
Regression
•
—-
•
* • •
•
Regression
•
_ -
•
* • •
•
Figure C-3: Schematic of NOAA "Updated Emulation Scenario" of
EPA's MNA Model (MNA2)
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-6
-------
By "updating" their emulation model to scale all their predictions relative to the 2002 to 2005
SSAP dataset, Field et al. (2016) severed the connection that existed between their baseline
emulation and the temporal trends in the calibration datasets for the original EPA models. The
same would be true if the HUDTOX MNA forecast were interrupted in 2003 and its predicted
sediment concentrations replaced by 2002 to 2005 averages. If this were done, HUDTOX would
no longer be consistent with the prior sediment and water column data to which it was calibrated.
As discussed below, not only does the NOAA "update" sever the connection with the calibration
datasets for the original EPA models, it results in forecasts for water column and fish tissue
PCBs that are biased substantially higher than the observed data.
I dated emulation scenario computes water column PCB concentrations that
are biased substantially high, relative to observed data.
Field et al. (2016) used their regression equations to predict average annual values of the
following:
• Water column PCB concentration in four HUDTOX model subsections in the UHR
- Their equations predicted annual PCB loadings from sediment to the water
column, assuming an "updated" sediment trend for each UHR subsection and
fitting their equations to the HUDTOX water column forecast; and
• Wet weight fish tissue PCB concentrations for four fish species at four stations in the
LHR below Federal Dam
- Their equations predicted annual average LHR wet weight fish tissue PCB
concentrations using emulated water column PCB concentrations at Waterford as
their sole independent variable for each combination of species and LHR location.
Thus, the NOAA emulation assumes that fish tissue concentrations in Lower
Hudson River fish can be adequately predicted based on an Upper Hudson River
water concentration.
EPA replicated the NOAA emulation model, using the documentation of regression equations
and coefficients provided in Field et al. (2016). Initial 2003 sediment PCB conditions for the
emulation were obtained from their Table 2, for both the baseline and "updated" MNA scenarios,
which they denoted as MNA1 and MNA2. Model equations were obtained from their Appendix
A, and inputs and coefficients from their Tables A.l and A.2. EPA's replication of the NOAA
procedures assumed an upstream source concentration for Tri+ PCB of 2 ng/L for every year.2
Comparisons between EPA's reconstruction of the NOAA "updated" MNA scenario (MNA2)
with observed data show systematic upward bias for UHR water column Tri+ PCB
concentrations. This is shown in Figure C-4 for the period 2003 to 2008. This interval begins
with the first year of the MNA2 simulation and ends with 2008, the last year of MNA before
dredging. Figure C-4 shows model-data comparisons for the four water column sampling
2 This value was assumed because Field et al. (2016) wrote that their paper focuses on scenarios with upstream
concentrations decaying to 2 ng/L by 2005, for consistency with recent monitoring data, and held constant at this
value for the remainder of their long-term simulations. A sediment decay rate of 3 percent per year was also
assumed in EPA's replication of MNA2.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-7
-------
locations in the MNA2 scenario from upstream to downstream: Thompson Island Dam (TID),
Schuylerville, Stillwater, and Waterford.
100
90
80
70
60 -
CO
50
2 40
30
20
10
Tl Dam
100
90
80
70
at 60
t 50
5 40
30
20
10
Schuylerville
•
.
¦ -
tt r •?'
ir, : v •••>_•
•.v •••-
•-••• _• • S L. '»• *• . <
• ;• ; % • ••
\ 7. r * «
. .i . %••... v- / • ;
2003 2004 2005 2006 2007 2008 2009
• Monitoring Data — MNA2
2003 2004 2005 2006 2007 2008 2009
• Monitoring Data — MNA2
100
90
80
70
"Si 60
CQ
t 50
2 40
30
20
10
Stillwater
90
80
70
"SBO
*50
240
30
20
10
0
20
Waterford
.
• *
V •»
'\ . -*•. A
i • ; . .
\ 't• •" »°!" i -• 5-
.* <3 .* . •••£.
;7" ' .V . *• : ,'•£
. v »y •
•v. .. "«?.• J" ^ J- ?•••
2003 2004 2005 2006 2007 2008 2009
• Monitoring Data — MNA2
03 2004 2005 2006 2007 2008 2009
• Monitoring Data - MIMA 2
Figure C-4: Comparisons between NOAA "Updated Emulation Scenario" (MNA2) and
2004-2008 GE Data for Water Column PCBs at Four UHR Sampling Stations.
At TID, predicted Tri+ PCB concentrations in the MNA2 scenario exceed all but the most
extreme values of 2004 to 2008 water column data, and the same is true for the Schuylerville and
Waterford locations.3 As is clear from Figure C-4, MNA2 scenario concentrations at Waterford
are higher than at the other locations throughout the 2003-2008 period, and the simulated
increase in water column concentration between Stillwater and Waterford is wholly inconsistent
with site-specific data for these locations. MNA2 scenario concentrations at Waterford exceed
those at Schuylerville by more than 15 ng/L in every year, and exceed observations in the other
two reaches by even more. Accuracy of prediction at Waterford is critical to the NOAA "updated
emulation scenario," because the computed MNA2 water column concentration at Waterford is
the single independent variable that represents exposures for fish tissue PCB concentrations at all
of the LHR locations in their model.
Predicted water column Tri+ PCB concentrations at Waterford in the MNA2 scenario are biased
especially high, relative to the other three stations, for three reasons:
• NOAA "updated" 2003 surface sediment concentrations in each reach, and this inflated
the water-column load gain (i.e., the upstream-to-downstream increase in water column
3 Note that there are no water column data at these stations for 2003 to compare to the emulation predictions.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-8
-------
concentration) within each reach. These successive load gains are additive from TID to
Schuylerville, to Stillwater, and to Waterford in their "updated emulation model;"
• The NOAA sediment update for the Waterford reach is also proportionally much greater
(increasing by a factor of 6) than the adjustments that were applied for the other reaches
(factors of 2 to 3); and
• The NOAA regression coefficient that they used to compute water column PCB load gain
from their assumed surface sediment concentrations (reported as "Sed to water" in their
Table A.2) is much greater for the Waterford reach than for the other three reaches.
The substantial bias of the MNA2 scenario for water column Tri+ PCB concentrations, in
general and on a reach-by-reach basis, demonstrates the pitfalls of re-assigning surface sediment
concentrations in a baseline emulation model and then re-starting the "updated" model without
testing the results against observed data.
I dated emulation scenario also computes wet weight fish tissue PCB
concentrations that are biased substantially higher than observed data.
The NOAA "updated emulation scenario" predicts wet weight fish tissue PCB concentrations, as
does EPA's FISHRAND model, but while FISHRAND relies on sediment and water exposure
concentrations at appropriate spatial and temporal scales, the NOAA "updated scenario"
characterizes exposure to Lower Hudson River fish using only Upper Hudson River water
column concentrations. Because water column Tri+ PCB concentrations in the "updated"
MNA2 scenario show substantial upward bias, this same bias is propagated to the fish tissue
concentrations in this scenario. All of the predicted fish tissue concentrations in MNA2, from
Albany to West Point, are functions of the MNA2 scenario water column concentrations at
Waterford, and all of the bias in the Waterford water column concentrations is transferred to fish
tissue concentrations in the "updated emulation scenario" via its regression equations.
This bias is demonstrated below with time series plots that show comparisons between the
NOAA "updated emulation scenario" and observed data for LHR fish tissue wet weight PCB
concentrations for 2003 to 20084, for the species and locations in the MNA2 scenario, at RM 152
(Albany/Troy), RM 113 (Catskill), and RM 90 (Kingston). EPA has not constructed comparable
model-data comparisons for RM 50 (West Point), because there were insufficient data for 2003
to 2008 at this location for an informative comparison.
Figure C-5 shows wet weight PCB concentration data for white perch, brown bullhead,
largemouth bass, and yellow perch at RM 152 versus results from the MNA2 scenario. Each fish
sample is shown as an individual data point. Although data for brown bullhead and largemouth
4 Model data comparisons are provided for 2003 to 2008 because NOAA provided sediment updates for 2003 in
its Table 1 to initialize a scenario, and did not provide a time series of upstream boundary conditions sufficient to
produce a spreadsheet replication for the full historical period. The period 2003 to 2008 provides the best test of
the "updated emulation scenario," because this is the period for which NOAA up-scaled sediment concentrations
by the greatest amount. NOAA relied on sediment data from one of the calibration datasets (1991) in developing
its "updated" sediment trend, so the fit to data of its "updated emulation scenario" should be more and more
similar to the fit of the original calibration if extended backward from 2003 toward 1991.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-9
-------
bass are quite sparse for this period, in general the NOAA MNA2 scenario results are
substantially higher than the data, with few data points serving as exceptions.
2 3
%
<1
White Perch
•
•
i
•
s
•
•
•
1
• •
5 i
2005
Data
2006
¦ NOAA MNA2
3
%
Brown Bullhead
2003 2004 2005 2006 2007 2008
• Data —•— NOAA, MNA2
Largemouth Bass
& 3
^3
Yellow Perch
2005 2006
Data - N0AA.MNA2
2005 2006
Data - NQAA.MNA2
Figure C-5: Comparisons between NOAA "Updated Emulation Scenario" (MNA2) and
Observed Wet Weight Fish Tissue PCBs at RM 152.
Figure C-6 shows wet weight PCB concentration data for white perch, brown bullhead,
largemouth bass, and yellow perch at RM 113 versus results from the MNA2 scenario.
Numerous observations are available for all four species at this location, and again, the NOAA
MNA2 scenario results are, in general, substantially higher than the data, with few data points
serving as exceptions.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-10
-------
2 2
"5
White Perch
2 2
5
Brown Bullhead
•
l
•
!
•
1
•
S
j
•
•
! i
2005 2006
Data NOAA MNA2
2005 2006
Data - - N OAA, M NA2
jP 2
^3
Largemouth Bass
•
• 1
•
1
1
•
• •
: 1
g 2
2005
Data
2006
• - NOAA, MNA2
2008 2003
2005 2006
Data NOAA, MNA2
Yellow Perch
Figure C-6: Comparisons between NOAA "Updated Emulation Scenario" (MNA2) and
Observed Wet Weight Fish Tissue PCBs at RM 113.
Figure C-7 shows wet weight PCB concentration data for white perch, brown bullhead,
largemouth bass, and yellow perch at RM 90 versus results from the MNA2 scenario. At RM 90,
as at RM 152 and RM 113, the NOAA MNA2 scenario results are generally higher than the data,
although the deviations from data are smaller.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-l 1
-------
White Perch
2005
Data
2006
- NQAA MNA2
Largemouth Bass
2005
Data
2006
¦ NQAA, MNA2
•
Brown Bullhead
•
f
•
•
•
•
I
2004 2005 2006
• Data NQAA, MNA2
2007 2008
2005 2006
Data - - N QAA, MNA2
Yellow Perch
Figure C-7: Comparisons between NOAA "Updated Emulation Scenario" (MNA2) and
Observed Wet Weight Fish Tissue PCBs at RM 90.
In their paper, the authors contend that their "updated emulation scenario" provides a better fit
than EPA's models to LHR fish tissue PCB data, after those PCB data have been transformed to
a common lipid content5. For example, while their "updated scenario" performs poorly in
matching 2004 to 2008 white perch wet weight data at RM 152 (see Figure C-5), it appears to
show a satisfactory fit to lipid-adjusted data (see their Figure 10). It is questionable whether a
simple lipid normalization is an appropriate assumption at the very low lipid contents reported
for many of these fish, as EPA more fully discusses in Appendix 3 of the FYR. Also using a
simple lipid normalization, EPA estimated a 3 percent recovery rate for white perch at RM 152,
so it is not surprising that NOAA shows in its Figure 10 a close overlay between their "updated
scenario" (assuming 3 percent decay from 1991 sediment concentrations) and lipid-adjusted data
for this species and location. In general, EPA agrees that lipid-normalization is a useful tool in
estimating data-based rates of recovery, and has shown in Appendix 3 of the FYR that lipid
normalization tends to reduce estimates of fish tissue recovery rates in both the UHR and the
LHR, based on available historical data.6 Moreover, lipid adjustment cannot offset the mismatch
between predicted wet weight concentrations and data for the other species simulated in
NOAA's "updated scenario:" adjusting 2003 to 2008 largemouth bass or yellow perch data to
FISHRAND's assumed median lipid would actually shift the data downward, increasing the gap
between the "updated scenario" and the data at each station. For brown bullhead, normalization
5 They transform to the median value in FISHRAND 's probabilistic distribution of lipid inputs.
6 However, lipid normalization for the purpose of evaluating temporal trends is different from lipid adjustment for
the purpose of comparing model predictions of wet weight tissue concentrations to observed data, as explained in
Appendix 3.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-12
-------
of the 2003 to 2008 data using FISHRAND's median lipid value would cause an upward shift by
a factor of about 2, which is not enough to match the NOAA "updated scenario" to data at RM
152 or 113. Thus, the model-data comparison at RM 152 that the authors show in their Figure 10
appears to be a special case, and is not in itself sufficient to confirm the accuracy of their
method.
Regardless of the role of lipid adjustment in computing rates of recovery, the authors constructed
their specific predictions of LHR wet weight fish tissue PCBs, and thus their predicted times to
reach wet weight PCB targets in the LHR, based on their predicted water column exposure
concentrations in the UHR at Waterford using their "updated emulation scenario." As shown
above, these computed exposure concentrations are biased substantially high and are completely
inconsistent with observed data. The authors also assumed that the water concentrations at this
single location in the Upper Hudson River can be used to reliably predict Lower Hudson River
fish concentrations over a range of more than 100 miles. Consequently, EPA does not accept
the NOAA "updated emulation scenario" as an accurate or reliable tool to replace EPA's models
in projecting LHR fish tissue concentrations.
~-scaling of 2003 sediment PCB concentrations accounts for most of the
difference between their predictions of recovery times and those in The
down-scaling of their assumed sediment recovery rate lias a much smaller effect on their
predicted recovery times.
As discussed above, the Field et al. "updated emulation scenario" involved (i) up-scaling
sediment PCB concentrations in 2003 by factors of 2 to 6 and (ii) down-scaling assumed
sediment recovery rates from 2003 onward from 8 percent to 3 percent per year. As shown
above, the "update" to 2003 surface sediment concentrations imparted substantial upward bias to
the water column and fish tissue PCB concentrations in the "updated scenario" (MNA2). The
additional downscaling of the trend in sediment concentrations also affects the subsequent rate of
change in simulated water column and fish tissue concentrations, slowing the recovery rate of
each via the emulation model regression equations. Of the two changes in the Field et al. (2016)
"update," the up-scaling of sediment PCB concentrations in 2003 has the dominant effect on
estimates of time to achieve ROD goals, with the down-scaling in recovery rate playing a more
minor role.
This is shown in Figure C-8 below, which plots the Field et al. (2016) estimates of time to reach
the 0.2 mg/kg Tri+ PCB (wet weight) human health risk threshold, by species and location, after
remediation, from their supplemental Table S-3. They present the following three cases:
• Their baseline case (REM1), using their original emulation of EPA's ROD model and an
8 percent post-dredging sediment recovery rate;
• A variant of REM1 assuming 3 percent post-dredging decline in surface sediment
concentrations, and
• The "updated emulation scenario" case (REM2), with up-scaled surface sediment
concentrations in 2003, as in their MNA2 simulation, and a 3 percent post-dredging
sediment recovery rate.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-13
-------
A comparison of the two variants of REM1 in Figure C-8 shows that the imposition of the 3
percent sediment recovery rate by itself has a relatively minor effect on times to recover. The
largest predicted increase is for largemouth bass at RM 152, from two to 16 years. Scenario
REM2 retains the 3 percent sediment recovery rate and adds the up-scaling of 2003 sediment
concentrations. The results for REM2 in Figure C-8 show that imposition of up-scaled surface
sediment concentrations lengthens the predicted time to recover by many decades for most
species and locations. Thus, the long times to recover shown by Field et al. (2016) are due
primarily to the up-scaling of sediment concentrations, which imparted an upward bias in
predicted water column and wet weight fish tissue PCB concentrations, an issue that should have
been addressed through recalibration.
80
70
2
o
-C
4)
£
E
a.
60
50
40
£ 30
o
w
IS 20
10
Ed
i Original Sediment (REM1), 8%
I Original Sediment (REM1), 3%
l Updated Sediment (REM2), 3%
(NfOOO(NfOOO(NrOO
u^THcrtLOLOTHcrii-nLnTHcri
tHtHO-O-tHtHCQCQtHtHcQ
CQ CO 5 5 CO CO CQ
5 5 J J Q3 CD
5 3
5 5
Fish Species and Sampling Location (RM)
Figure C-8: Estimated Years to Reach 0.2 mg/kg Human Health Risk Threshold by Species
and Location for Selected Scenarios (Source: Field et al. Supplemental Table S-3)
The original mechanistic models (HUDTOX-FISHRAND) used by EPA to inform the ROD
for the UHR successfully reproduced the available historical data for 1977-1998 and
remain scientifically valid management tools.
The mechanistic models HUDTOX and FISHRAND were constrained through calibration to
UHR data for the period 1977 to 1998 (EPA, 2000a). Those long-term historical calibrations of
HUDTOX and FISHRAND to all the available data provided the foundation for use of those
models in conducting forecast simulations to estimate long-term responses to remedial
alternatives in the RI/FS.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-14
-------
HUDTOX, the PCB transport and fate model, was calibrated to data for the following variables:
• Tri+ PCB surface sediment concentration trends;
• Solids burial rates;
• In-river solids and Tri+ PCB mass transport at high and low flows; and
• Solids and Tri+ PCB water column concentrations.
The historical calibration of HUDTOX was also tested through model-data comparisons for total
PCBs and five individual congeners from 1991 to 1997.
FISHRAND, the mechanistic bioaccumulation model, was calibrated to data for five species of
fish: largemouth bass, brown bullhead, yellow perch, spottail shiner and pumpkinseed using
Bayesian updating. Model calibration was conducted for Tri+ PCB concentration in fish tissue
on a wet weight basis by optimizing a priori input distributions for lipid (empirical) and Log of
the octanol/water partition coefficient (K0W) (congener-specific based on literature data) as
discussed in greater detail in EPA (2000a).
HUDTOX and FISHRAND were subject to a rigorous peer review by a panel of international
experts (Eastern Research Group, Inc, 2000). After extensive document review and a series of
public meetings, the peer review panel determined that the models are acceptable and adequately
reproduce historical data. The panel noted that the models do not reflect a fully mechanistic
understanding of all chemical, physical, and biological processes, and expressed concern about
increasing temporal uncertainty over time in the model forecasts. In its Response to Peer Review
Comments, EPA acknowledged uncertainties in the models, but reiterated that the models
provide a sufficient understanding of the system on which to base a decision for the site.
The original mechanii ulels, when extended to include the
1998-2008 pre-dredging period, continue to show good agreement with the observed data.
After the last round of sediment sampling used to test EPA's models, an eleven-year period of
sampling and pre-remedial design followed, providing a test of the accuracy of the models in
predicting water column and fish tissue PCB concentrations under conditions of monitored natural
recovery. In Appendices 1 and 3 of the FYR, HUDTOX and FISHRAND forecasts were compared
to monitoring data for that period (1998 to 2008), which ended with the commencement of
remedial dredging. Appendix 1 shows a 1998 to 2008 HUDTOX simulation of water column
PCBs, updated to include actual river flows, to be generally faithful to both seasonal and long-
term trends in water column PCBs for the full period, including the intensive data collection period
of 2004-2008. Appendix 3 shows that FISHRAND showed good agreement between data and
predicted wet-weight and lipid-adjusted fish tissue concentrations for the UHR (where the model
was calibrated) and RM 152 using the 1998-2008 HUDTOX simulations to provide PCB exposure
concentrations.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-15
-------
References
Eastern Research Group, Inc. 2000. Report on the Peer Review of the Revised Baseline
Modeling Report for the Hudson River PCBs Superfund Site. Final. Prepared for U.S.
Environmental Protection Agency, Region II, Emergency and Remedial Response Division. EPA
Contract No. 68-W6-0022, Work Assignment No. 4-12, May 10.
Field, L.J., J.W. Kern, and L.B. Rosman. 2016. "Re-visiting projections of PCBs in Lower
Hudson River fish using model emulation." Science of Total Environment 557-558 (2016) pp.
489-501. (http://dx.doi.Org/10.1016/i.scitotenv.2016.02.072).
EPA. 2000a. Revised Baseline Modeling Report (RBMR), Hudson River PCBs Reassessment
RI/FS Volume 2D. Prepared for USEPA Region 2 and USACE, Kansas City District by TAMS
Consultants, Inc., Limno-Tech, Inc., Menzie-Cura & Associates, Inc., and Tetra Tech, Inc.
January 2000, http://www.epa.gov/hudson/reports.htm (Last accessed: 02/10/2015)
EPA. 2016a. Responses to NOAA Manuscript Entitled: "Re-Visiting Projections of PCBs in
Lower Hudson River Fish Using Model Emulation" (Field, Kern, Rosman, 2016).
https://www3.epa. gov/hudson/pdf/EP A%20White%20Paper%20-
%20Responses%20to%20NQAA%20Manuscript.pdf.
Appendix C Technical Memorandum - Evaluation of Field, Kern and Rosman (2016) Emulation Model
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site C-16
-------
Final Second Five-Year Review Comment
Response for the
Hudson River PCBs Superfund Site
APPENDIX D
SPECIAL STUDY - BLACK BASS FILLET TISSUE
WITH AND WITHOUT RIBS
Prepared by
Kern Statistical Services, Inc
April 2019
-------
Appendix D
Special Study - Black Bass Fillet
Tissue With and Without Ribs
1 Introduction and Background:
In 2004 GE began conducting analyses for PCB on fish samples collected as part of the Baseline
Monitoring Program (BMP). Protocols established for and approved for use under the BMP Quality
Assurance Project Plan (QAPP) included procedures used by NYSDEC for fish collection and sample
preparation. Specifically, the 2004 BMP QAPP indicated that: "All fish will be prepared for contaminant
analyses following collection according to the SOP for Annual Fish Sampling (Appendix 21; adapted
from NYSDEC procedures)." Both the NYSDEC standard fillet approach and the standard operating
procedure (SOP) detailed in BMP QAPP Appendix 21 involve inclusion of the rib-bones and belly flap
with the fillet that is removed from the fish and subsequently analyzed for PCB's and lipids. EPA
determined, based in part on information provided by NYSDEC oversight staff, that the adult sportfish
fillets processed by GE from 2007 to 2013 did not include the ribcage as required. In response to this
deviation from the approved project procedures, EPA required that GE perform a special study in 2014
evaluating the degree to which exclusion of the rib cage (ribs) impacted measurement of fish tissue PCB
concentrations and lipid levels during the six years in question. Black bass (smallmouth bass and
largemouth bass) were the focus of the special study because they are large enough to produce fillets of
sufficient size for comparison and are collected from project monitoring stations in both the Upper and
Lower Hudson River.
The primary objective of this special study was to evaluate the comparability of lipid-normalized PCBs in
black bass fillets with and without ribs. This study compared Aroclor PCB (TPCB\mcinr). lipid, and lipid
normalized PCBs (LPCB) in black bass fillets processed with and without the ribs. Consistent processing
of fillet samples for chemical analysis is important because organic chemicals such as PCBs are
preferentially found within lipid rich tissues, such as those surrounding the gut and rib cage. Comparison
of chemical concentrations in groups of samples derived from differing sample preparation procedures
can result in systematic biases, which could influence their interpretation. In this study, the influence of
differential processing of fillets with respect to inclusion of the ribs was investigated through a paired
samples analysis comparing chemical parameters in left and right hand fillets from each fish.
2 Methods:
The difference between fillets with and without ribs was evaluated to determine if the two-sided 95
percent confidence interval contained zero (i.e., zero difference), and the half width is less than 20 percent
of the mean [LPCB] (all samples). Aroclor PCB concentrations from the two processing methods were
considered comparable within the margin of error and would be considered adequate for subsequent
analysis and interpretation. If both of these conclusions could not be drawn, the data would be evaluated
to determine the potential value of a larger sample sizes which would be generated in a subsequent year of
study.
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site D-l
-------
2.1 Hypothesis and Data Quality Objective
This study tests the null hypotheses that after adjustment for lipid content, the LPCB concentrations of
samples with ribs are consistent with those without ribs. This hypothesis was predicated on an ability to
achieve a detectable difference of 20 percent change in the concentrations of LPCB with 80 percent
power (|3 = 0.2). The number of paired samples was determined under the assumption that LPCB
concentrations would be compared using a paired Student's t-test at 5 percent level of significance (a =
0.05) and that a 20 percent difference in mean LPCB concentration would be detected with 80 percent
power. Sample data from black bass collected in 2013 were used to determine the minimum number of
fillet pairs necessary to meet this Data Quality Objective (DQO) (Table 1). Power and sample size
calculations were performed using methods for paired Student's T-tests (Lenth, 2009). The number of
fish required to meet the special study objective of detecting a 20 percent change in LPCB at 80 percent
power (|3 = 0.2) with 5 percent level of significance is n = 130.
Table 1. Lipid-Normalized PCBs in Hudson River Black Bass Fillets 2013.
STN
TISS
Sample fish
per Station
Average [LPCB]
(mg PCB/kg-
Lipid)
Standard
Deviation
Coefficient of
Variation (percent)
ATI
SF
20
153
41
27
CS1
SF
20
68
26
39
FD1
SF
30
4
7
153
ND1
SF
5
781
140
18
ND2
SF
5
780
477
61
ND3
SF
5
468
178
38
ND5 (BMP)
SF
10 *
576
491
85
SW1
SF
5
525
226
43
SW2
SF
5
535
200
37
SW3 (BMP)
SF
10
584.
321
55
SW4
SF
5
261
105
40
SW5
SF
5
303
146
48
TD1
SF
5
377
205
54
TD2
SF
5
473
326
69
TD3
SF
5
1304
814
62
TD4
SF
5
1008
717
71
TD5 (BMP)
SF
10
450
219
49
* Note: At n=130 the sampling design for this special study calls for 5 more fish from within ND pool (RS2) to
attain n=130 and make numbers for each RS equal.
2.2 Approach:
1) Fish collected under the Remedial Action Monitoring Program (RAMP) were used to supply sets
of paired (left and right) black bass fillets.
2) The left and right fillets of each fish were processed in exactly the same way with the exception
that the ribs were included (taken with) one fillet and excluded (not taken) from the other.
3) The method used to process fillets including ribs is the NYSDEC standard fillet method
(Attachment, 1: Standard Operating Procedure (SOP) PrepLab3, dated 2011). The approach used
to process fillets that do not include ribs is Appendix 3.8-4 of the 2011 Hudson River PCBs Site
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site D-2
-------
RAMP Quality Assurance Project Plan (QAPP), Attachment 2: SOP for the Tissue
Reduction/Grinding for Whole Body and Filleted Fish, August 2011).
4) In order to avoid bias and provide representative sample populations, equal numbers of left and
right rib samples were processed from the fish collected within each River Section and from
historical NYSDEC monitoring locations.
5) EPA provided robust oversight at the GE contractor laboratory throughout the study. NYSDEC
also participated in observing the work.
6) The 130 required Black bass were collected from the usual RAMP fish sampling locations from
RS 1 through RS 3 and the Albany-Troy (AT) and Catskill (CS) locations. An additional five
fish were also collected from RS 2 yielding 130 paired fillets.
7) Data generated by the study, for both fillets, was prepared and analyzed consistently with the
RAMP and current post-dredging program, including relevant QA/QC and laboratory controls
using a modification of the EPA M8082A Aroclor Sum Method (Pace SOP S-NY-O-314-rev.OO;
Appendix A3-1 of Revised Attachment A to the Phase 2 RAMP QAPP).
2.3 Statistical Analysis Methods
Each fish resulted in a pair of PCB measurements and a pair of lipid measurements which were compared
statistically. To test the null hypothesis of the study, the distribution of differences in paired LPCBs were
evaluated relative to the mean difference in LPCB concentrations. To further examine the results of the
special study, EPA also subjected the data to the following additional statistical evaluations:
1) robust regression between paired lipid-normalized PCBs,
2) robust regression between paired lipid content measurements,
3) robust regression between paired wet weight PCB measurements, and
4) average ratios of rib to without-rib wet weight PCB measurements.
Statistical methods for each analysis are described in the next sections.
2.3.1 Paired Differences
The distribution of differences in paired LPCBs were evaluated relative to the mean difference in LPCBs,
li0. and the null hypothesis was:
H0: IMrw ~ V-No-Rib I — 0.2 X fi0
versus the alternative hypothesis:
Ha- \H-Rib ~ H-No-Rib I ^ 0.2 X [Xq.
The null hypothesis, or assumed condition, is that the difference between preparation methods is greater
than 20 percent. Therefore, rejection of this null hypothesis at the 5 percent level of significance indicates
that one can be 95 percent confident that the difference in preparation methods is less than 20 percent.
This null hypothesis can be tested with the paired-samples Student's t-test, or equivalently by
constructing a 95 percent interval for the mean difference in LPCBs and comparing the absolute value of
the confidence limits with 0.2 X /10. The null hypothesis is rejected when the upper and lower confidence
limits are less in magnitude than 0.2 x ;U0. Rejection of the null hypothesis indicates satisfaction of the
data quality objective stated above.
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
D-3
-------
2.3.2 Robust Regression
Traditional regression analyses based on least squares fitting (Neter et al., 1996) require the assumption
that regression residuals are normally distributed with constant variance, across the range of the predictor
variables. These assumptions were tested by fitting a least squares regression line and testing the residuals
for normality with the Lilliefors test (Lilliefors, 1967), and for constant variance using the White test
(White, 1980). When tests of either or both assumptions failed, then robust regression was used to
estimate the relationships between rib and without-rib LPCB, PCB and lipid levels. The M-estimation
procedure (Huber 1973) was used to fit the robust regression line to data using a bi-square weighting
function (Huber, 1981). The idea behind robust regression is to minimize the influence of extremes by
weighting samples close to the fitted line more heavily than those more distant from the line. In this way
the influence of potential outliers is reduced without the need to remove them from the data set.
2.3.3 Average of Ratios
It is not uncommon for environmental scientists to work with ratios in efforts to explore bivariate
relationships, although there are some disadvantages relative to regression analysis including: 1) ratios
can be shown to be equivalent to a line through the origin which forces the relationship to be a direct
proportionality; and 2) distributions of ratios tend to be highly right skewed and therefore estimates based
on ratios tend to be highly uncertain. Despite these general concerns the mean ratio of without-rib to rib
PCBs (wet-weight concentrations) was estimated for completeness and to acknowledge traditional use of
ratios in environmental data analysis.
3 Results
3.1 Difference in Lipid-Normalized PCBs
The average of LPCBs in fillets with and without ribs were 439 mg PCBs/kg lipid and 473 mg PCBs/kg
lipid respectively, and the overall average was 456 mg PCBs/kg lipid (Figure 1). The zone of compliance
(for data quality) of 20 percent of the mean was (0.2 x /xo) = 91 mg PCBs/kg lipid. The upper and lower
confidence limits were -0.69 mg PCBs/kg lipid and 69 mg PCBs/kg lipid respectively showing that the
average of differences between LPCBs in fillets with and without ribs was less than the 20 percent zone
of compliance with 95 percent level of confidence.
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
D-4
-------
Difference in Paired Lipid Normalized Fillet PCBs
(Fillet With Rib) - (Fillet Without Rib)
No Difference Line
— 95% T-lnterva[ (-0.69 to 69 )
~ Mean Difference (34)
•300 -200 -100 0 100 200 300
Difference (Rib - No-rib) Total PCB (mg-PCB/kg-Lipid)
vce T-raeooseivaeioris (-11C6 -78S arc 1327 mg-pcMg-Uplf) oje&oene-530 » 500 irwervQiarePBtsioar.
Figure 1. Differences in LPCBs with 20 percent zone of compliance for data quality (gray band) and 95
percent confidence interval for the mean difference between LPCBs in fillets with and without ribs.
This bias at approximately 8 percent with samples with ribs being higher is sufficiently small as to not
present a major issue for data interpretation. Because the two-sided 95 percent confidence interval
contains zero (i.e., zero difference), and the half width is less than 20 percent of the mean [LPCB], EPA
determined that a second year of comparisons was not needed.
3.2 Robust Regression:
3.2.1 LPCB
The nature of the relationship between rib and without-rib LPCBs was investigated further through
regression analysis. A traditional least squares line was first fit to the data and residuals were found to
vary with concentration (White Test, p<0.001) and were non-normal (Lilliefors Test, p<0.001). To
compensate for these deviations from standard regression assumptions, the data were subjected to a robust
regression through the origin. The paired data and fitted line are plotted in Figure 2, and parameter
estimates are provided in Table 2. This analysis confirms that LPCB in samples including the rib average
were approximately 8 percent% (range of 6 percent to 10 percent) higher than those without the rib.
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
D-5
-------
3000
Q.
txo
CO
u
Q.
ao
i/i
-Q
>/)
-4-»
co
u
Cl
2000
1000
Observations 130
Parameters 1
Error DF 129
MSE 40993
R-Square 0.9091
AdjR-Square 0.9084
500 1000 1500 2000 2500
LPCB in fillets without ribs (mg-PCB/kg-Lipid)
— Fit ~ 95% Confidence Limits 95% Prediction Limits
Figure 2. Fitted robust regression line for LPCB concentrations in black bass fillets with ribs vs. those
without ribs, Hudson River, NY.
Table 2. Parameter estimates for robust regression between LPCB concentrations hi black bass
fillets with and without ribs, Upper Hudson River.
Parameter
DF
Estimate
Standard
Error
95 Percent Confidence
Limits
Chi-Square
Pr > ChiSq
LPCB
1
1.08
0.0109
1.06
1.10
9797.
<.0001
Scale
1
67
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
D-6
-------
3.2.2 Percent Lipid
Robust regression was used to estimate the relationship for percent lipid in fillets with and without ribs.
The percent lipid in rib fillets was approximately 16 percent (CI: 10 percent, 22 percent) greater than in
without-lib samples (Figure 3. Percent lipid in fillets with ribs vs. those without ribs with robust
regression fit., Table 3).
Figure 3. Percent lipid in fillets with ribs vs. those without ribs with robust regression fit.
Table 3. Parameter estimates for robust regression between percent lipid in black bass fillets
with and without ribs, Upper Hudson River.
Parameter
DF
Estimate
Standard
Error
95 Percent
Confidence
Limits
Chi-
Square
Pr > ChiSq
Percent Lipid
1
1.16
0.03
1.10
1.22
1314
<.0001
Scale
1
0.22
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
D-7
-------
3.2.3 Wet Weight Total PCBs
A robust regression line was also fit to describe the relationship between rib and without-rib wet weight
total PCBs in black bass fillets. Wet weight TPCBatogIot in rib samples were approximately 16 percent (CI:
11 percent to 21 percent) higher than in without-rib samples (Figure 4 and Table 4) and there is
substantially more variability in wet-weight concentrations reflecting differences in the amount of lipid in
rib and without-rib samples.
12
GO
10
Qfl
E
0
(/)
O
o
JJ
JO.
8
O
I
i/i
%
0 O
jy
6
c
o ° °
-t—'
-E
00
'ai
5
QJ
§
CD
U
CL
h-
t
o °
o©
0 °° ° °
© °
0
°o«o o>V
jjjfeoo o o
5^ v
0
%
o
o
0 2
4 6 8
TPCB (wet weight) in fillets without ribs (mg/kg)
10
Figure 4. Wet weight total PCBs in fillets with ribs vs. those without ribs with robust regression fit.
Table 4. Parameter estimates for robust regression between wet weight total PCB in black bass
fillets with and without ribs, Upper Hudson River.
Parameter
DF
Estimate
Standard
Error
95% Confidence
Limits
Chi-
Square
Pr > ChiSq
TPCB
1
1.16
0.024
1.11
1.21
2315
<.0001
Scale
1
0.61
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
D-8
-------
3.3 Mean Ratios Wet Weight TPCB
The arithmetic averages of ratios of rib to without-rib wet weight TPCBArodor ranged from approximately
1.5 to 1.75 with 95% confidence intervals ranging from a high of 2.25 to below 1.25 (Figure 5). Average
ratios were similar across all pools, suggesting that relationships are unrelated to specific environmental
conditions, as opposed to differences in PCB content in muscle tissue as compared with that in and
around the rib cage and belly flap. These estimated ratios are relatively imprecise, reflecting smaller
sample sizes within pools, as well as high variation in the ratios themselves. Ratios of random variables
for right skewed distributions are characteristically highly variable group comparisons based on averages
of ratios are generally discouraged relative to other statistical approaches. These ratios were compiled
here for comparison because statistical practitioners calculate such averages, however, we discourage
their use.
2.5
-------
data recognizing that PCB concentrations generally co-vary with lipid content in fish tissue samples.
LPCB tissue concentrations in this study were noticeably less variable than wet weight PCBs indicating
that lipid normalization reduces bias between rib on and without rib PCB concentrations and improves
power and precision of statistical analyses relative to analyses based on wet weight PCBs.
Some temporal comparisons of LPCB concentrations of interest may span periods when both fillets with
and without ribs would be compared. This analysis indicates that rib and without-rib LPCB measurements
are on average within the 20 percent zone of compliance for data quality with 95 percent level of
confidence. This suggests that there is compatibility of the LPCB concentrations in rib and without-rib
samples collected under the approved monitoring plans since 2004 but for transparency, the results from
2007 to 2013 should be identified as years where the fillet samples were prepared without the ribs.
Robust regression analysis suggests that differences in wet weight concentrations between rib and
without-rib samples are substantially larger and more variable than lipid-normalized concentrations,
suggesting that more care should be taken when interpreting temporal patterns in wet weight
concentrations, or comparing wet weight PCB absolute thresholds where the data from years when
without-rib fillets were analyzed (2007-2013) are included. The robust regressions can potentially be used
to convert from rib to without-rib PCB concentrations to facilitate such careful evaluations of both wet
weight and lipid-normalized PCB concentrations; although analyses based on combined data should be
evaluated with and without conversions to understand the sensitivity of any particular analysis
EPA has determined that for the years in question, a 34 mg PCB/kg lipid difference (about 8 percent)
would not appreciably effect data interpretation. It should be noted that the period of data collection
when the rib was not included (2007 to 2013) was just prior to and during dredging activities. Fish data
during dredging was impacted by dredging-related PCB resuspension. Therefore, fish monitoring data
collected during that period would not be used to establish pre- or post-dredging fish recovery trends. No
significant project decisions were made or altered based on fish data from those years; the data served its
primary purpose to monitor changes during dredging and the results were within EPA's expectations.
Also, no adjustments to fish advisories or regulations were made by New York State based on those data.
Therefore, based on the results of the special study, EPA does not intend to complete any follow up
studies associated with this matter. All post-dredging fish processing procedures have and will follow
NYSDEC standard fish processing procedures as required by project documents. EPA will continue to
oversee the fish processing at the laboratory as it monitors post-dredging recovery in fish.
5 References
General Electric Corporation 2011. Hudson River PCB's Site 2011 Remedial Action Monitoring Quality
Assurance Project Plan (RAMQAPP). Prepared by Anchor QEA (Glens Falls, NY) in
Conjunction with Environmental Standards, Inc. (Valley Forge, PA). Appendix 3.8-4 (SOP for
the Tissue Reduction/Grinding for Whole Body and Filleted Fish, August 2011).
Huber, P. J. 1981. Robust Statistics. John Wiley and Sons, New York.
Huber, P.J. (1973) Robust Regression: Asymptotics, Conjectures and Monte Carlo. Annals of
Statistics, 1, 799-821. http://dx.doi.org/10.1214/aos/1176342503
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
D-10
-------
Lenth, R. V. (2009). Java Applets for Power and Sample Size [Computer software]. Retrieved January
29, 2014, from http://www.stat.uiowa.edu/~rlenth/Power.
Lilliefors, H. W. "On the Kolmogorov-Smirnov test for normality with mean and variance
unknown." Journal of the American Statistical Association. Vol. 62, 1967, pp. 399-402.
Neter, J., Kutner, M. H., Nachtsheim, C. J., and Wasserman, W., 1996. Applied Linear Statistical Models,
4thEdition, Richard D. Irwin, Inc., Burr Ridge, Illinois. NYSDEC 2005. Of Time, PCB's and the
Fish of the Hudson River. NY State Department of Environmental Conservation, Division of
Fish, Wildlife, and Marine Resources. Albany, NY, July 2005, Appendix 1 (Fish Collection,
Sample Preparation, and Analytical Procedures, page 259).
White, H. (1980). "A Heteroskedasticity-Consistent Covariance Matrix Estimator and a Direct Test for
Heteroskedasticity". Econometrica. 48 (4): 817-838
Attachments to Appendix D
Attachment D-l: NYSDEC 2011. Prep Lab Standard Operating Procedure (SOP PrepLab3). NY State
Department of Environmental Conservation (Hale Creek Field Station). 3/16/2011.
Attachment D-2: Anchor QEA 2011. Standard Operating Procedure: Tissue and Preparation &
Homogenization for Biota and Plant Matrices. Appendix 3.8-4 SOP for the Tissue Reduction/Grinding
for Whole Body and Filleted Fish (NE132 07) to the Hudson River PCBs Superfund Site Phase 2
Remedial Action Monitoring Program Quality Assurance Project Plan. Anchor QEA and Environmental
Standards Inc. August 2011.
Appendix D Special Study - Black Bass Fillet Tissue With and Without Ribs
Final Second Five-Year Review Comment Response for the Hudson River PCBs Superfund Site
D-ll
-------
Attachment D-l
-------
SOP - PrepLab3 (3-16-2011)
PREP LAB STANDARD OPERATING PROCEDURE
NYS DEPARTMENT OF ENVIRONMENTAL CONSERVATION
Hale Creek Field Station
Name of document: SOP - PrepLab3 (3-16-2011)
Revision date: 3/16/2011
Previous revision: Preplab2
Reasons for this revision:
• Delete references to dBase IV for entering data into databases.
• Add details regarding type of grinder used and instructions to mix the tissue and repeat the
grinding step at least two more times and until the sample appears to be homogeneous.
• Add Section VI. - Minimizing sample contamination during sample preparation.
• Add reference, summary and background sections to the SOP.
Reference: Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories, Volume 1,
3rd edition (USEPA Office of Water, November 2000)
Summary: Samples are received at Hale Creek Field Station and dissected, ground and homogenized for
future chemical analysis. In addition, samples for organochlorine analysis are freeze-dried to remove
moisture.
Background:
New York State Department of Environmental Conservation conducts studies requiring chemical
analysis on fish or other biological tissues. Routine monitoring and surveillance studies develop data on
contaminants in fish for several reasons:
1. To identify sources of environmental contamination;
2. To identify the geographic extent of environmental contamination;
3. To identify temporal trends of contaminants in fish and wildlife;
4. To identify potential impacts to fish and their consumers; and
5. To provide information regarding human consumption advisories.
Chemical analyses of edible fish flesh have been determined to be the most appropriate analyses for
satisfying all of these objectives. The following methodology has been developed in order to standardize
the tissues under analysis and to adequately represent the contaminant levels of fish flesh. The portion of
edible flesh analyzed will be referred to as the standard fillet unless otherwise noted. For some species,
the procedure is modified as indicated below.
I. SAMPLE RECEIPT
A) All samples received by the lab are to be accompanied by a Collection Record and
Continuity of Evidence form.
B) After comparison of samples received with the Collection Record, the Continuity of
Evidence form is signed and dated.
C) The original forms are to be retained by the lab. Copies may be returned to the delivery
person.
D) Depending upon sample type, the samples are to be stored locked in either the cooler or
freezer.
II. SAMPLE LOGIN
A) All samples are assigned a unique serial Lab # which corresponds to a specific Tag # or
ID# on the sample or sample container.
B) The Lab #s are to be indicated on the Continuity of Evidence form and the Collection
Record.
Page 1 of 3
-------
C) From the Collection Record the Lab #, Tag #, Species, Location, Program, Length,
Weight, Sex, and Age are entered into the computer Log file.
III. SAMPLE DISSECTION
A) Samples are removed from the freezer and allowed to partially thaw (large samples may
be removed the previous night).
B) FISH: The portion of edible flesh analyzed will be referred to as the standard fillet unless
otherwise noted. For some species, the procedure is modified as indicated below.
1) Standard Fillet
a) Remove scales from fish. Do not remove the skin.
b) Make a cut along the ventral midline of the fish from the vent to the base
ofthe jaw.
c) Make a diagonal cut from the base of the cranium following just behind
the gill to the ventral side just behind the pectoral fin.
d) Remove the flesh and ribcage from one-half of the fish by cutting from
the cranium along the spine and dorsal rays to the caudal fin. The ribs
should remain on the fillet.
e) Score the skin and homogenize the entire fillet.
2) Modifications to the Standard Fillet
a) Four modifications of the standard fillet procedure (see b,c,d,e) are
designed to account for variations in fish size or known preferred
preparation methods of the fish for human consumption.
b) Some fish are too small to fillet by the above procedure. Fish less than
approximately 6 inches long and rainbow smelt are analyzed by cutting
the head off from behind the pectoral fin and eviscerating the fish.
Ensure that the belly flap is retained on the carcass to be analyzed.
c) Some species are generally eaten by skinning the fish. The skin from
these species is also relatively difficult to homogenize in the sample.
Hence, for the following list of species, the fish is first skinned prior to
homogenization:
Brown Bullhead White Catfish
Yellow Bullhead Channel Catfish
Black Bullhead Lake Sturgeon
Atlantic Sturgeon
d) American eel are analyzed by removing the head, skin, and viscera;
filleting is not attempted.
e) Forage fish and young-of-year fish are analyzed whole.
C) Wildlife/Other: Generally non-fish samples that are to be prepared have already been
dissected. See supervisor for appropriate instructions.
D) All dissection tools are to be rinsed, washed with soap, rinsed, rinsed with DI water and
dried between each sample dissection.
IV. HOMOGENIZATION
A) Thoroughly grind and homogenize fish fillets using a Waring commercial
chopper/grinder model WCG75. Alternatively, a comparable food chopper, food
processor, grinder, blender or homogenizer may be used.
B) Mix the tissue and repeat the grinding step at least two more times and until the sample
appears to be homogeneous.
Page 2 of 3
-------
C) The homogenized sample is then subsampled into appropriate glass bottles. Generally 2-
10 g is needed for metals analysis and 20g for organochlorine analysis. For the OC
sample label and weigh an empty sample bottle. Add ca 20g of sample into the bottle and
weigh again.
D) The bottles are capped and stored in the freezer.
E) All homogenization tools are to be rinsed, washed with soap, rinsed, rinsed with DI water
and dried between each sample.
V. FREEZE DRYING
A) Generally samples for organochlorine analysis are freeze dried.
B) Make sure that the unit has been drained, all valves closed and the vacuum pump oil is
clear and within the acceptable markings on the site vial.
C) Turn the refrigeration unit on. After the temperature OK light comes on (less than -40 C)
turn the vacuum pump on. After the vacuum OK light comes on (less than 100 millitorr)
the samples may be placed on the freeze dryer (make sure samples are frozen).
D) The samples are freeze dried ca 16 hours or until the samples reach a constant weight.
E) When freeze dried, the sample bottle is weighed again.
F) The sample is stored in the freezer until analysis is started.
VI. MINIMIZING SAMPLE CONTAMINATION DURING SAMPLE PREPARATION
A) Conduct all work in a clean environment, preferably a laboratory setting. All work
surfaces, utensils and grinder work bowls and covers should be cleaned with soap and
water, then rinsed with clean water, prior to working with samples, between each
sample, and upon completion of sample preparation for the day. Alternatively, between
samples aluminum foil may be placed on the work surface for the succeeding fish
sample; discard foil after one use. DO NOT use aluminum foil if metals analyses are to
be conducted on the sample.
B) Wear a clean laboratory coat for protection of clothing. Wear nitrile or latex gloves at all
times while preparing samples. Clean gloves with soap and water between each sample,
or discard gloves between samples and place new gloves on hands. If a glove is torn or
punctured, immediately discard the glove and replace with a new glove. Discard gloves
at the end of the day, or earlier if they become unsuitable for clean preparation of
samples..
C) Rinse fish or other biological samples in clean water if soil, debris or other matter are
evident on the exterior surfaces. Allow water to run off and dry exterior surface.
D) Following preparation of sample portions, place sample in clean containers of suitable
size
for the sample. For example, place small samples in chemically clean glass jars, cover
and label immediately. Jars should have PTFE-lined caps and be precleaned and certified
to meet EPA standards for metals, pesticides and semi-volatiles. For large samples (e.g.,
a fish fillet), wrap in hexane-rinsed aluminum foil and label externally. Place foil
wrapped sample in a labeled food-grade plastic bag for subsequent storage and transport.
If hexane-rinsed aluminum foil is unavailable, and samples are not to be analyzed for
phthalates, the excised sample may be placed in a food grade plastic bag, labeled
externally and placed in frozen storage. DO NOT use aluminum foil if metals analyses
are to be conducted on the sample.
SOP-PrepLab3 .doc
3-16-2011
Anthony J. Gudlewski
Page 3 of 3
-------
Attachment D-2
-------
APPENDIX 3.8-4
SOP FOR THE TISSUE
REDUCTION/GRINDING FOR WHOLE
BODY AND FILLETED FISH
(NE132_07)
-------
m
Pace Analytical*
STANDARD OPERATING PROCEDURE
Tissue and Preparation & homogenization
FOR BIOTA AND PLANT MATRICES
Reference Methods: US EPA SW-846 Test Methods for Evaluating Solid Waste
LOCAL SOP NUMBER: NE132_07
EFFECTIVE DATE: 03/29/2011
SUPERSEDES: NE132_06
SOP TEMPLATE NUMBER: SOT-ALL-Q-006-rev.03
Approvals
~TA Da -
v*- y-y'
03/29/2011
Dan Pfalzcr
Assistant General Manager Date
03/29/2011
Christina L. Braidwood
Quality Manager Date
Periodic Review
Signatures below indicate no changes have been made since previous approval.
Signature Title Date
Signature Title Date
Signature Title Date
© 2002 - 2010 Pace Analytical Services, Inc. This Standard Operating Procedure may not be reproduced, in part or in full, without written
consent of Pace Analytical Services, Inc. Whether distributed internally or as a "courtesy copy" to clients or regulatory agencies, this document
is considered confidential and proprietary information.
Any printed documents in use within a Pace Analytical Services, Inc. laboratory have been reviewed and approved by the persons listed on the
cover page. They can only be deemed official if proper signatures are present.
This is COPY# distributed on by and is CONTROLLED or X UNCONTROLLED.
-------
PACE ANALYTICAL SERVICES, INC
2190 TECHNOLOGY DRIVE
SCHENECTADY, NY 12308
(518) 346-4592
STANDARD OPERATING PROCEDURE
LABORATORY PROCEDURE NE132_07.DOC
REVISION 7 (03/29/11)
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
2 of 32
-------
1.0
IDENTIFICATION OF TEST METHOD
1.1 Standard Operating Procedure for tissue preparation, processing and
homogenization prior to extraction/digestion and analysis.
2.0 APPLICABLE MATRIX OR MATRICES
2.1 This method is applicable to the preparation and homogenization of animal and
plant matrixes; including but not limited to: fish (whole body and fillets), mollusks
(mussels, clams, etc.), crustaceans (lobster or shrimp, etc.), mammals (mice,
mink, muskrat, shrew etc.), reptiles and amphibians (frogs or turtles, etc.), macro
invertebrates (benthic worms, eels, insects and other biota), and vegetation
(coastal and wetland grasses/plants).
3.0 DETECTION LIMIT
3.1 Not applicable
4.0 SCOPE AND APPLICATION
4.1 This method is intended to describe the preparation and homogenization
procedures prior to the extraction, digestion and/or clean up of sample extracts.
This procedure uses a variety of cutting, grinding and scaling equipment for size
reduction, composting, and homogenization. Client and/or project may dictate
additional specific requirements than stated below. Samples are best processed
when partially frozen. Samples may be re-frozen after processing pending
extraction or digestion.
5.0 SUMMARY OF TEST METHOD
5.1 Fish
5.1.1 Samples are weighed, measured, and gender determined if possible.
The fish may be processed whole body or as fillets, and with the skin on
or off. If fillets are to be removed and processed separately, this is
generally done after the removal of the skin. If compositing is required,
the identified samples for composite are filleted or skinned prior to
homogenization. The carcass of the fish (after removal of the fillet) may
be maintained for separate homogenization and analysis if requested.
5.2 Mollusks, crustaceans and other like invertebrates
5.2.1 Samples are measured and weighed prior to processing. Mollusks must
be removed from their shells before processing. Due to the low weight of
a single mollusk, crustacean, or invertebrate, these sample types are
generally composited with others of the same species and/or sampling
area prior to homogenization. Gender determination may need to be
performed, i.e. lobsters. This is done prior to any processing and
recorded. Additionally, lobsters are usually dissected, and the edible
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 3 of 32
-------
meat (tail and claw) is removed for homogenization. Certain internal
organs such as the hepatopancreas may need to be processed
separately. If crabs are being processed, the legs, claws and body
cavity are generally homogenized together.
5.3 Mammals
5.3.1 Mammals such as mink, mice, shrew or other rodents, must be prepared
in a glove box or bio-hazard hood with the use of a HEPA biological
respirator due to the potential health hazards associated with mammal
tissue. All project specific sample preparation (weighing, skinning,
compositing and homogenization) is performed in the glove box. Waste
from the process must be treated with bleach before disposal. The
outside surfaces of the sample containers must be disinfected before
removal from the glove box.
5.4 Reptiles and Amphibians
5.4.1 Samples are generally processed as whole body samples. Depending
upon the size, the specimen may need to be cut into small pieces and
processed in part, then re-combined as a single sample. Due to the
thickness of the skin of most reptiles, such as frogs, it is recommended
that these be processed without the skin. If the skin must be processed,
ensure that the grinder or processor blades are sharpened before use.
The blades may need to be re-sharpened between every few samples as
needed. Turtles must be removed from the shell prior to processing by
digging out the head and legs, and as much of the body as feasible.
5.5 Macro invertebrates
5.5.1 Macro invertebrates such as worms, eels, insects or benthic biota are
generally processed as whole body samples. Depending upon the size,
the specimen may need to be cut into small pieces and processed in
part, then recombined as a single sample. Due to the low weight of a
single invertebrate, these sample types are generally composited with
others of the same species and/or sampling area prior to
homogenization.
5.6.1 Samples are rinsed prior to processing to remove soil, silt, small insects
or other debris. Depending upon the size of the plant and the leaves, the
sample may be processed mechanically, or may have to be cut into
small pieces by hand. Plants can be processed either wet or dry,
depending upon project specifications
6.0 DEFINITIONS
6.1 Abdomen- the posterior section of the body behind the thorax in an arthropod.
5.6 Plants
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
4 of 32
-------
6.2
6.3
6.4
6.5
6.6
6.7
6.8
6.9
6.10
6.11
6.12
6.13
6.14
6.15
6.16
6.17
Abductor- to draw or spread away (as a limb or the fingers) from a position near
or parallel the median axis of the body or from the axis of a limb.
Arthropod- any of a phylum (Arthropoda) of invertebrate animals (as insects,
arachnids, and crustaceans) that have a segmented body and jointed
appendages, a usually chitinous exoskeleton molted at intervals, and a dorsal
anterior brain connected to a ventral chain of ganglia.
Biota- the flora or fauna of a region.
Bivalve- being or having a shell composed of two valves (shells).
Caudal- directed toward or situated in or near the tail or posterior part of the
body.
Carapace- bony or chitinous case or shield covering the back or part of the back
of an animal (as a turtle or crab).
Composite- combining the typical or essential characteristics of individuals
making up a group.
Crustacean- any of a large class (Crustacea) of mostly aquatic mandibular
arthropods that have a chitinous or calcareous and chitinous exoskeleton, a pair
of often much modified appendages on each segment, and two pairs of antennae
and that include the lobsters, shrimps, crabs, wood lice, water fleas, and
barnacles.
Digestate- product of digesting.
Fillet- to cut, a boneless cut of fish.
Head- the upper or anterior division of the animal body that contains the brain,
the chief sense organs, and the mouth.
Hepatopancreas- a glandular structure (as of a crustacean) that combines the
digestive functions of the vertebrate liver and pancreas.
Homogenize- to reduce the particles of so that they are uniformly small and
evenly distributed.
Mantle- a fold or lobe or pair of lobes of the body wall of a mollusk or brachiopod
that in shell-bearing forms, lines the shell and bears shell-secreting glands.
Pectoral muscle- any of the muscles which connect the ventral walls of the chest
with the bones of the upper arm and shoulder and of which there are two on each
side of the human body.
Swimmerets- one of a series of small unspecialized appendages under the
abdomen of many crustaceans that are best developed in some decapods (as a
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
5 of 32
-------
lobster) and usually function in locomotion or reproduction
6.18 Telson-the terminal segment of the body of an arthropod or segmented worm.
6.19 Thorax-1) the middle of the three chief divisions of the body of an insect also,
the corresponding part of a crustacean or an arachnid. 2) the part of the
mammalian body between the neck and the abdomen also, its cavity in which the
heart and lungs lie.
7.0 INTERFERENCES
7.1 Samples being tested for metals must be processed with a ceramic knife and/or
ground with a plastic blade to prevent contamination from metals such as steel or
tin.
7.2 Samples being tested for organics must be processed with metal, Teflon, PTFE
and or glass utensils. The use of plastics may cause interferences with the
analysis of samples.
8.0 SAFETY
8.1 The use of laboratory equipment and chemicals exposes the analyst to several
potential hazards. Good laboratory techniques and safety practices shall be
followed at all times. Approved PRE, which includes safety glasses, gloves, must
be worn at all times in the lab. Lab coats are provided and may be worn. All
Personal Protective Equipment (PPE) must be removed before leaving the
laboratory area and before entering the employee lounge or eating area. Always
wash your hands before leaving the laboratory.
8.2 All standards, reagents and solvents shall be handled under a hood using the
proper PPE. All flammable solvents must be kept in the flammable storage
cabinet, and returned to the cabinet immediately after use. When transporting
chemicals, make sure to use a secure transporting devise and/or secondary
outer container.
8.3 The chemist should have received in-house safety training and should know the
location of first aid equipment and the emergency spill/clean-up equipment
before handling any apparatus or equipment.
8.4 Extreme caution must be taken when using or handling knives, descalers, and
grinders to homogenize the biota samples.
8.5 Re-useable cotton mesh glove liners may be worn under latex or PVC gloves as
an additional measure when using sharp tools or knives, or when dealing with
samples that have sharp teeth, spines, fins, or thorns. The mesh lining can help
prevent piercing of the skin in case a tool or sample slips, during dissection or
other preparation steps.
8.6 Polychlorinated biphenyls should be treated with extreme caution; as a class of
chemical compounds they possess both toxic and suspected carcinogenic
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 6 of 32
-------
properties.
8.7 All additional company safety practices shall be followed at all times as written in
the Pace Analytical Chemical Hygiene Plan.
9.0 EQUIPMENT AND SUPPLIES
9.1 Cutting board-made of either glass or polyethylene.
9.2 Food processor with titanium cutting blade (small), or blender with stainless steel
blades (large).
9.2.1 2- Retsch Grindomix (model GM200) with glass and or plastic mixing
bowls
9.2.2 1-Kitchen Aid Little Ultra Power
9.2.3 1-Tor Rey (model M22) Large Food Processor
9.3 Knives: ceramic stainless steel, or titanium. (See Section 7.0 for interferences
and/or contamination associated with different material knives and blades).
9.3.1 Gerber Stainless Steel Boning knives
9.3.2 Dexter Russel Chopping knives
9.3.3 Oneida Stainless Steel fillet knives
9.3.4 URI Eagle Ceramic Knife
9.4 Necropsy dissection kits
9.5 Analytical balance with precision to 0.01g.
9.6 Labconco multi-hazard glove box.
9.7 Advantage 200 LS Respirator Facepiece
9.8 Bench liner material (Lab Mat) and scissors.
9.9 Aluminum foil.
9.10 Plastic wrap or wax paper.
9.11 Titanium fork.
9.12 Teflon-coated spatula.
9.13 Teflon or stainless steel tweezers and dissection scissors.
9.14 PVC or Latex gloves.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 7 of 32
-------
9.15
Ruler.
9.16 Mallet.
9.17 Stainless steel or plastic strainer.
9.18 Salad spinner.
9.19 Pre-cleaned glass sample jars with Teflon or PTFE-lined caps.
9.20 Kim wipes.
9.21 Nylon bristled brushes for cleaning.
10.0 REAGENTS AND STANDARDS
10.1 Deionized (PI) water- Deionized (Dl) water or reagent water is ASTM Type II
laboratory reagent grade water or better (Type I).The Millpore NANO-pure
system provides Type I water used in the metals laboratory for rinsing lab glass
and plastic ware. Other grades may be used, provided it is first ascertained that
the reagent is of sufficiently high purity to permit its use without lessening the
accuracy of the determination. If the purity of a reagent is in question, analyze for
contamination.
10.2 Hexane - Pesticide grade
10.3 Acetone - HPLC grade
10.4 Nitric acid 25% - Add 250ml_ concentrated HN03 to 400ml_ of reagent water
and dilute to 1L in an appropriate flask. (See metals lab for this prepared
solution).
10.5 10% Bleach solution - Add 100mL of commercial bleach to 500mL of reagent
water and dilute to 1 liter in an appropriate beaker or flask.
10.6 Alconox - cleaning solution.
11.0 SAMPLE COLLECTION, PRESERVATION, SHIPMENT and STORAGE
11.1 Sample collection is not applicable to the Pace laboratory operation.
11.2 Please see the Pace SOP (NE227) that describes the responsibilities of sample
custody including all proper documentation, verification, and tracking procedures
following Chain of Custody (COC) protocols, sample receipt procedures, and
Internal COC procedures for sample tracking include the use of sample tracking
logbooks.
11.3 All samples should remain frozen at all times unless being tested. Fish usually
arrive whole bodied or already filleted. Once received the sample must be ground
and homogenized so that it may be analyzed. The homogenized fish tissue can be
held for 6 to 12 months. The fish solvent extracts can be held for 3 months. Some
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 8 of 32
-------
clients may request that the body and/or head offish be saved once the fillets are
cut out. Other biota material may have other specifications stated specifically for
that project.
11.4 If samples are not shipped frozen, they will be stored in freezers at Pace
Analytical upon arrival, and until processing. The samples must remain frozen
and maintained at < -20 C. Sample processing and extraction/digestion hold
times are suspended by freezing the sample. Hold time monitoring is resumed
when samples are removed from freezers for processing and then returned to
freezers pending extraction or digestion. The organic hold time is 14 days from
sample collection to extraction, and 40 days from extraction to analysis. The
metals hold time is six months from sample collection to digestion and analysis. If
mercury is to be determined, the hold time is 28 days from sample collection to
digestion and analysis.
11.5 Tissue samples: As guidance, a minimum of 50 grams of sample must be
collected for organic analyses, and 5 grams for metals analyses, in a glass jar
with a Teflon or PTFE lined screw cap. The amount of sample needed, will
depend upon the project management plan such as reporting limits and the need
for MS/MSD and/or duplicate analyses. Extra sample must be collected, if
possible, to allow the laboratory adequate sample volume in case of re-extract
and reanalysis is needed. Large whole individual fillets or vegetation may be
wrapped in plastic or aluminum foil depending upon the requested analyses.
Large crustaceans, reptiles or amphibians may be individually packed in well-
labeled Styrofoam coolers.
12.0 QUALITY CONTROL
12.1 Contamination Prevention
12.1.1 If the purity of a reagent is in question, analyze for contamination.
12.1.2 Blades for dissection may need to be re-sharpened between every few
samples as needed.
12.1.3 Certain project specific sample preparation (weighing, skinning,
compositing and homogenization) is performed in the glove box. Waste
from the process must be treated with bleach before disposal. The
outside surfaces of the sample containers being processed must be
containerized, treated and disinfected before removal from the glove box.
12.2 The procedures described below are general cleaning and pre-processing
procedures that are to be followed regardless of the type of tissue being
processed. Samples are prioritized by the Laboratory Supervisor or Lab Manager
based on hold time and client due date. All weights, measurements and other
project required observations are recorded in LIMS.
12.2.1 Wash all utensils, sample processors (blades, blade post, cup and lid)
and cutting boards with an Alconox solution and a sponge. Rinse
thoroughly with tap water, then with Dl water.
12.2.2 If the samples are going to be processed for organic analyses only, rinse
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 9 of 32
-------
all washed utensils, processor parts and surfaces with hexane followed
by rinsing with acetone.
12.2.3 If samples are going to be processed for metal analyses only, rinse all
plastic and ceramic utensils with Dl water and then Nitric acid 25%
solution and then Dl water again.
12.2.4 If requested by the client, the equipment or processing blank should be
collected at this time by pouring Dl water into and out of the processor,
over the surfaces of the utensils and over the cutting board. The blank is
collected in the appropriate container, at the project specification
frequency, for the determinative analysis.
12.2.5 Gloves must be worn when handling tissue samples. Latex gloves may
be worn. All gloves must be talc or dust free.
12.2.6 Tissue samples should be partially thawed before starting, to the point
where it becomes possible to make an incision in, or cut through, the
flesh. When samples are completely thawed they become soft and
difficult to cut or fillet. NOTE: If whole bodies are not being processed,
and the tissue is partially frozen during dissection, there is less of a
chance of puncturing the gut cavity and any internal organs. Inadvertent
puncture of the internal organs may contaminate the part(s) of the animal
that have been selected for analysis. Also, internal organs may rupture
during freezing. If this is observed during dissection, it must be noted in
the processing records. Note any morphological abnormalities on the
processing records.
12.3 Hold times: The homogenized fish tissue can be held for 6 to 12 months. The fish
solvent extracts can be held for 3 months.
13.0 CALIBRATION AND STANDARDIZATION
13.1 Not Applicable
14.0 PROCEDURES
14.1 Fish Tissue Preparation:
14.1.1 Determine the wet weight for each individual fish using a calibrated
balance and record in LIMS. The balance should be covered with
aluminum foil if aluminum is not a metal of concern. If aluminum is a
metal of concern and the sample will not be analyzed for organic
compounds the balance should be covered with plastic wrap. If the
sample is for both metal and organic compounds, wax paper may be
used. Catch any excess fluid coming from the thawing specimen into the
wax paper, foil or plastic wrap. All liquid from thawed whole fish must be
kept as part of the sample. The technician must remember to zero the
balance with the aluminum foil, plastic wrap, or wax paper on it before
weighing the specimen. The foil, plastic wrap, or wax paper must be
changed after each weighing.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
10 of 32
-------
14.1.2
Determine the length of each fish using a ruler, and record in LIMS.
Some measurements may, or may not be, a part of the project
specifications.
14.1.3 If gender identification is needed this must be done prior to the scaling
and filleting processes.
14.1.4 Removal of Scales or Skin: If required by project specifications, the
scales and/or skin of the fish will be removed prior to filleting.
14.1.5 Lay the fish on the cleaned, and/or lined, cutting board.
14.1.6 Scrape the fish from tail to head using the electric, automated descaler
with ceramic claws to remove the scales. Note: If performing metals
analysis, titanium or ceramic must be used.
14.1.7 Rinse the cutting board between fish with Dl water and Alconox. If
plastic, wax paper, or foil is used, change between fish.
14.1.8 Rinse the outside of the fish with Dl water and pat dry with paper towel
Place the fish on its side, on a clean cutting board, for filleting or
skinning.
14.1.9 To skin the fish, loosen the skin behind the gill cover and pull the skin off
toward the tail with a Catfish skinning tool, cutting lightly along the inside
of the skin, Slowly separate the skin from the muscle tissue of the body
or the fillet.
14.2 Filleting the Fish
14.2.1 Using fresh gloves and the specified knife, make a cut behind the entire
length of the gill cover, making sure to cut through the skin, if still
attached, flesh, and as close to the bone as possible. Note: If the fish
samples are small, and it appears difficult to fillet, or if the amount of the
fillet appears to be insufficient for the analysis, consult the Project
Manager prior to filleting. In some cases it may be necessary to
homogenize the whole body.
14.2.2 Make a cut across the base of the tail fin keeping as close to the caudal
fin (tail) as possible. Continue cutting along the underbelly of the fish
moving from the head to the tail.
14.2.3 Go back to the cut made at the beginning at the gill cover and slice down
the entire length of the fish following along the backbone until reaching
the cut previously made across the tail.
14.2.4 Remove the fillet from the fish. Be sure to include the belly flap in each
fillet and do not remove the dark muscle tissue in the vicinity of the
lateral line from the light muscle tissue that makes up the rest of the
muscle tissue mass.
14.2.5 Remove any bones that may be left attached to the fillet. Repeat the fillet
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 11 of 32
-------
steps for the second side of the specimen.
14.2.6 The general procedure recommended for filleting fish is illustrated in
Apendix 1.
14.2.7 Note in the sample processing records in LIMS if the internal organs
were ruptured during freezing or if inadvertent puncture of the internal
organs occurred during the filleting process, rinse the fillet(s) tissue with
Dl water.
14.2.8 Cover the balance with the appropriate clean lining, and weigh the
fillet(s). Record the fillet(s) weight(s) in the processing records.
14.2.9 If the fillet(s) and/or the carcass are to be homogenized immediately,
proceed to Section 14.3. If not, rinse all fish parts with Dl water and store
in the appropriate container; see Section 9.0 for allowable materials.
Note that it may be necessary to chop the fillet(s) or carcass into smaller
pieces, with the appropriately cleaned knife, before storage, and before
homogenization, so the entire sample will fit into the storage container or
the homogenization vessel. If the samples will not be homogenized
immediately, the samples must be placed back into the freezer, until
homogenization.
14.3 Homogenization
14.3.1 Allow the fillet(s), carcass or whole body to partially thaw. Retain all fluids
as part of the sample.
14.3.2 Homogenize whole fish bodies, carcasses, or fish fillets by placing them
into the small or large food processor fitted with the appropriate blades.
The sample may need to be cut into smaller pieces for processing.
Process the sample until it appears to be fully and consistently
homogenous. Continue to grind the sample until there are no chunks
present in the homogenate. The homogenous nature of the sample is
vitally important for reproducible results. Sample should be
homogenized fully and thoroughly.
14.3.3 Individual homogenates may be processed further to prepare composite
homogenates as required by project specifications. Composite
homogenates must be prepared from equal weights of individual
homogenates. All individual weights that make up one composite must
be recorded, if required, or one composite weight may be recorded. If
individual or composite homogenates were frozen prior to
extraction/digestion, these homogenates must be thawed and re-
homogenized by hand mixing prior to being extracted or digested.
14.3.4 Place the individual or composite homogenized samples into the
appropriate glass jars to be frozen pending future extraction/digestion. If
the samples will not be extracted/digested immediately, the samples
must be returned to the freezer until extraction/digestion.
14.3.5 All utensils and equipment must be washed in between samples
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
12 of 32
-------
according to the procedures described previously in Section 12.2.
14.4 Mollusk (Bivalves) Preparation (Mussels. Clams)
14.4.1 Wash all utensils, the cutting board, and surfaces as previously
described in Section 12.2. Obtain samples from freezer.
14.4.2 If required by the project specifications, measure and record the length of
the sample shell. Cover the balance with the proper material as
described in Section 9.0, and weigh and record the sample weight in
LIMS.
14.4.3 Wearing the proper gloves, place the sample on a clean, cutting board.
Samples should be partially thawed. If the sample is frozen, it will be
difficult to break open the shell. If the sample is excessively thawed, the
internal tissue will become soupy and difficult to remove.
14.4.4 If preparing bivalve specimens, use the titanium knife to cut the abductor
muscle by sliding the knife through the crevice where the two shells
meet. Once the abductor muscle is cut the two shell pieces should come
apart easily.
14.4.5 Carefully remove the top shell, and using the Teflon coated spatula,
scoop out the internal tissue that is resting on the mantle.
14.4.6 Cover the balance with the proper material and weigh the amount of
tissue obtained from the sample. Record the weight along with the
information previously recorded on the processing records. The sample
may now be stored pending homogenization in the appropriate jar.
14.4.7 Since the amount of tissue obtained from one bivalve is generally small,
several specimens are frequently combined to make one sample.
Utensils do not need to be rinsed between the individual samples that
comprise one composite, but utensils must always be rinsed in between
each composite sample.
14.4.8 After the tissue has been removed from all of the specimen shells for one
composite or individual sample, place the tissue in the clean small
processor with the titanium blade to be homogenized. Grind the sample
until it appears to be fully and consistently homogenized and there are
no large chunks.
14.4.9 If tissue is being processed for volatile organic carbon (VOC) analysis
the homogenization must be done by hand.
14.4.10 Individual homogenates may be processed further to prepare composite
homogenates as required by project specifications. Composite
homogenates must be prepared from equal weights of individual
homogenates. All individual weights that make up one composite must
be recorded, if required, or one composite weight may be recorded. If
individual or composite homogenates were frozen prior to
extraction/digestion, these homogenates must be thawed and re-
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 13 of 32
-------
homogenized by hand mixing prior to being extracted or digested.
14.4.11 Place the processed samples into the appropriate glass jars to be frozen
for future extraction/digestion, and place back into the freezer.
14.4.12 All utensils and equipment must be washed in between samples
according to the procedures described previously in Section 12.2.
14.5 Crustaceans (Lobsters, Crabs, Shrimp)
14.5.1 Wash all utensils, the cutting board, and surfaces as previously
described in Section 12.2. Obtain samples from the freezer.
14.5.2 If project specifications require gender determination of lobsters, this
must be done prior to dissecting. To determine the gender, hold the
lobster by the thorax, and flip it over to examine the underneath
abdomen, just below the legs and where the abdomen division begins,
there is a first pair of swimmerets. The first pair of swim me rets is what is
used to distinguish the lobster's gender. If the first pair is soft, has small
hairs, and the swimmerets are crossed, it is female. On a male lobster,
the first pair of swimmerets is hard and stiff, and generally do not touch.
14.5.3 If the hepatopancreas of the lobster samples is to be analyzed, the
lobster samples must be received alive. If the samples are frozen prior to
dissection, the hepatopancreas will burst upon thawing making it
impossible to remove. To remove the hepatopancreas, the live lobster
should be placed on a cleaned cutting board. Wearing the proper gloves,
one analyst holds claws out in front of the lobster, while also holding
down the lower abdomen and tail. The second analyst takes a titanium-
coated knife, and places it on the grove in the outer shell, just behind the
head region. Keeping the knife at an angle, the second analyst must
push down and forward, to remove the head. Once the head is removed,
the hepatopancreas can be seen lying just under the carapace and
running the length of the thorax. The hepatopancreas is generally a
greenish-yellow color, but there may be some variation. Using the Teflon
coated spoon, scoop the hepatopancreas out gently trying not to break it
into pieces. Cover the tray of the balance with the proper material, and
weigh and record the weight of the hepatopancreas in the processing
record, and place it into an appropriate sample jar for freezing and future
extraction/digestion.
14.5.4 To remove the edible meat, remove the two claws from the body of the
lobster at the joint. Place a piece of lab mat or paper towel over the claw
and pound with a mallet. Once the shell is crushed, remove the meat,
using the appropriately cleaned tweezers or other tool, making sure to
get all the meat in the joints and arms. Cover the balance tray with the
appropriate material and record the total tissue weight arms. Record this
weight with the previously recorded information from the two claws and
sample processing record.
14.5.5 Remove the abdomen and telson from the rest of the outer shell by
pulling the lobster apart. Using the titanium coated knife, cut through the
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
14 of 32
-------
center underside tissue of the lobster and laterally along the exoskeleton
of the tail. Once the abdomen and tail have been cut open, separate the
shell from the edible meat using cleaned utensils. Any eggs found in the
female lobsters will have to be removed and discarded or sampled
separately. Cover the balance tray with the appropriate material, and
record the weight of the tissue obtained from the abdomen and telson on
the processing record. The sample may now be stored pending
homogenization in the appropriate jar.
14.5.6 If removing tissue from crabs, break off all legs and claws. Squeeze, pull,
or pick all the tissue out of the legs and claws. Pull apart the outer shell.
Scoop out the tissue using a Teflon coated spatula. Cover the balance
tray with the appropriate material, and record the weight of the tissue
obtained from the abdomen and telson on the processing record. The
sample may now be stored pending homogenization in the appropriate
jar.
14.6 Mammals (Mice, Mink, Muskrat, Shrew)
14.6.1 Wash all utensils, the cutting board, and surfaces as previously
described in Section 12.2. Obtain samples from the freezer.
14.6.2 Place the first specimen partially thawed to be processed, and all
equipment needed into the glove box/Bio-hood on a freshly laid out lab
mat (Blue diaper).
14.6.3 Once all materials are in the glove box and set up for use, seal the
transfer box and ensure the motor blower is on. Over tightening of the
outer or inner door knobs is not necessary to achieve a good seal. Place
your hands into the gloves attached to the glove ports and place Latex
gloves over the glove port gloves for use. The outer Latex gloves will
need to be changed in between each sample.
14.6.4 If the gender of the mouse or shrew needs to be determined, turn the
animal over and note the length of the anus and the distance of the anus
from the tail. If the anus is elongated in shape and does not touch the
base of the tail, testicles and a large genital papilla are visible, and there
are no nipples, the animal is male. If the anus is round in shape and
almost touches the base of the tail and/or there are nipples (up to five
sets), the animal is female. If the animal is very small, young or immature
and a gender determination cannot be made, note that the gender is non
determinable. Record the gender observations on the processing
records.
14.7 Organ Dissection/Processing
14.7.1 If the mammal is being dissected for Brain, Liver, Kidney, Heart, Lung, or
Adipose (Fat) tissue, each organ will need to be harvested.
14.7.2 Place the animal on its back with forceps. Pinch the skin at the base of
anus and carefully make an incision at the tail end, and cut just below the
skin along the abdomen and past the chest cavity. Cutting the skin flap
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 15 of 32
-------
at the abdomen cavity carefully separate the adipose tissue from the
muscle tissue. Below it should be a white/yellow material. Take this
material out.
14.7.3 Identify each organ and remove them from the abdomen cavity.
14.7.4 Weigh and record the weight of the mammal organs and place into the
appropriate container.
14.7.5 The rib cage will need to be cut with scissors. Once chest cavity is open,
remove the heart and lungs.
14.7.6 Weigh and record the weight of the mammal organs and place into the
appropriate container.
14.7.7 Since the amount of tissue obtained from one animal may be small,
manually grinding of the organs may need to be done at the time of
extraction.
14.7.8 Place the processed samples into the appropriate glass jars to be frozen
for future extraction/digestion into the freezer.
14.7.9 Before removing any equipment all utensils and equipment must be
washed with Dl water and 10% bleach solution.
14.7.10 All disposable materials must be double bagged for disposal.
14.8 Whole Animal Processing:
14.8.1 If skinning of the mammal is required, carefully make an incision at the
tail end and cut just below the skin along the back, from one hind leg to
the other. Make another cut from one hind leg to one front leg and repeat
the cut on the other side of the animal. Starting from the tail, lift the skin
flap, and carefully separate the skin from the muscle tissue below. Pull
the skin forward from the tail to the head to expose the back tissue of the
animal. Repeat the procedure on the stomach side of the animal. Note: it
may be very difficult to remove the skin from the legs, head, and tail. If
some skin cannot be removed, note this on the processing records.
14.8.2 Weigh and record the weight of the mammal on the processing records.
Depending upon the size of the mammal, it may need to be chopped into
small pieces before being ground. Generally, mice and shrew can be
quartered before homogenization if needed.
14.8.3 Put the whole body or chopped sample into the cup of the grinding unit.
Turn the grinding unit on low speed and gradually increase the speed to
homogenize the sample being careful to minimize any splatter or outside
contamination. Homogenize until a uniform consistency is achieved.
14.8.4 Transfer the homogenized sample from the cup to the pre-labeled
sample jar using the appropriate utensil. Clean the outside of the sample
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
16 of 32
-------
jar with the 10% bleach soaked Kim wipe.
14.8.5 To clean the grinding unit in between samples, remove as much residual
tissue on the blade as possible by operating the unit at low or medium
speed, using Dl water and 10% bleach. Rinse unit with Dl water if
metals are being done and/or hexane or acetone for organics.
14.8.6 Repeat steps 14.9.2 through 14.9.5 until the samples are complete.
14.8.7 Since the amount of tissue obtained from one mouse or shrew may be
small, several specimens may be combined to make one sample, as
required by project specifications. Utensils do not need to be rinsed
between the individual samples that comprise one composite, but
utensils must always be cleaned in between each composite sample.
14.8.8 If several specimens will be composited to make one sample, follow the
applicable Sections of 14.9.2 through 14.9.5, for each of the specimens.
The tissue obtained from each specimen may be weighed and recorded
individually, then totaled for the composite weight. If only one composite
weight is sufficient for the project specifications, weigh the entire
composite and record that weight in LIMS.
14.8.9 Place the processed samples into the appropriate glass jars to be frozen
for future extraction/digestion, placed back into the freezer.
14.8.10 Before removing any equipment all utensils and equipment must be
washed with Dl water and 10% bleach.
14.8.11 All disposable materials must be double bagged for disposal.
14.9 Reptiles and Amphibians (Frogs and Turtles)
14.9.1 Wash all utensils, the cutting board, and surfaces as previously
described in Section 12.2. Obtain samples from the freezer
14.9.2 Wearing the proper gloves, place the turtle sample on the cleaned
cutting board. The turtle should be partially thawed. If the turtle is frozen,
it will be difficult to remove the muscle. If the sample is excessively
thawed, the internal tissue will become soupy and difficult to remove.
14.9.3 Take all project required measurements. The distance between the
anterior and posterior edge of a turtle carapace (top of shell) should be
measured with a ruler and recorded on the processing records. If the
entire mass of the turtle, including the shell, needs to be recorded, cover
the balance with the proper material and weigh and record this weight in
LIMS.
14.9.4 Since the bottom of shell and carapace are extremely dense and difficult
to cut through with normal dissecting tools, the muscle tissue of the turtle
must be removed by cutting the body of the turtle away from the shell.
Insert a knife, made of the proper material, into the skin of the turtle,
close to the shell on the lower half of the body. Slowly, cut along the
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 17 of 32
-------
entire circumference of the shell. Repeat the procedure on the upper half
of the body, on both sides of the shell.
14.9.5 With dissection scissors or a ceramic or titanium paring knife of the
proper material, remove the skin from the hind limbs, tail, and fore limbs
and neck. Remove any visible muscle tissue within the carapace. Most of
this tissue will be found in the upper portion of the carapace around the
pectoral area.
14.9.6 Using the appropriate utensils, remove the muscle tissue from the tail,
neck, hind limbs, and fore limbs, including the feet, leaving bone and
claws behind.
14.9.7 Cover the balance with the proper material and weigh the amount of
tissue of the turtle sample. Record the weight along with the information
previously recorded on the processing records. The sample may now be
stored pending homogenization in the appropriate jar.
14.9.8 If processing frogs, allow the frogs to partially thaw, take the project
specific measurements, and record them in LIMS. The number of frogs
required to make up one sample, and the weight and length of the
individual frogs, must be taken and recorded, if specified. In all cases,
the skin must be removed from the frog prior to processing and chopped
into smaller pieces, due to its thickness. It will then be added to the
processor with the whole body of the frog, or it may be discarded
depending upon the project specifications.
14.9.9 To skin the frog, make an incision, using the proper utensils, and cut into
an area where there is an excess of skin, most likely around the neck.
Slowly, pull the skin off of the frog using dissecting scissors, or a ceramic
or titanium paring knife, as needed. Once skin is removed, chop it up into
tiny pieces using the appropriate knife and set it aside to be processed
with the whole frog body.
14.9.10 Cover the balance with the proper material and weigh the amount of
tissue obtained from the frog samples if the tissue and the whole body
will not be processed. Record the weight along with the information
previously recorded on the processing records. The sample may now be
stored pending homogenization in the appropriate jar.
14.9.11 Since the amount of tissue obtained from one small turtle or frog may be
insignificant, several specimens may be combined to make up one
sample. Utensils do not need to be rinsed between the individual
samples that comprise one composite, but utensils must always be
rinsed in between each composite sample.
14.9.12 If several specimens will be composited to make up one sample, the
tissue obtained from each specimen may be weighed and recorded
individually, then totaled for the composite weight. If only the composite
weight is sufficient for the project specifications, weigh the entire
composite and record that weight in LIMS.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
18 of 32
-------
14.9.13 After the tissue has been removed from all of the specimens,
homogenize the muscle tissue, and skin if required, by placing it into the
small or large food processor fitted with the appropriate blades. The
sample may need to be cut into smaller pieces for processing. Grind the
sample until it appears to be fully and consistently homogenous.
Continue to grind the sample until there are no chunks present in the
homogenate.
14.9.14 Individual homogenates may be processed further to prepare composite
homogenates as required by project specifications. Composite
homogenates must be prepared from equal weights of individual
homogenates. All individual weights that make up one composite must
be recorded, if required, or one composite weight may be recorded. If
individual or composite homogenates were frozen prior to
extraction/digestion, these homogenates must be thawed and re-
homogenized by hand mixing prior to being extracted or digested.
14.9.15 Place the processed samples into the freezer to be frozen for future
extraction/digestion.
14.9.16 All utensils and equipment must be washed in between samples
according to the procedures described previously in Section 12.2.
14.10 Macro Invertebrates (Benthic Worms, Eels, Insects and other Biota)
14.10.1 Wash all utensils, the cutting board, and surfaces as previously
described in Section 12.2. Obtain samples from the freezer.
14.10.2 Cover the balance tray with the appropriate material and record the
weight of the invertebrate sample. Since the weight obtained from one
invertebrate (benthic worm, insect, biota) may be small, several
invertebrates may be combined to make one sample. In many cases,
several invertebrates of the same species and sample location are
delivered to the laboratory in one sample jar. Each specimen from this jar
must be weighed, if requested, and composited to form one
homogenized and unique sample. If only one composite weight is
sufficient for the project specifications, weigh the entire composite and
record that weight. Utensils do not need to be rinsed between the
individual samples or specimens that comprise one composite, but
utensils must always be rinsed between each composite sample.
14.10.3 Invertebrates such as eels must be chopped into smaller pieces before
homogenization. This is generally due to the length of the specimen and
the thickness of the skin.
14.10.4 Place the weighed specimen into the clean small processor with the
titanium blade to be homogenized. Process the sample until it appears to
be fully and consistently homogenized and there are no large chunks.
14.10.5 Individual homogenates may be processed further to prepare composite
homogenates as required by project specifications. Composite
homogenates must be prepared from equal weights of individual
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
19 of 32
-------
homogenates. All individual weights that make up one composite must
be recorded, if required, or one composite weight may be recorded. If
individual or composite homogenates were frozen prior to
extraction/digestion, these homogenates must be thawed and re-
homogenized by hand mixing prior to being extracted or digested.
14.10.6 Place the processed samples into the appropriate glass jars to be frozen
for future extraction/digestion, into the freezer.
14.10.7 All utensils and equipment must be washed in between samples
according to the procedures described previously in Section 12.2.
14.11 Vegetation (Coastal and Wetland Grasses/Plants)
14.11.1 Wash all utensils, the cutting board, and surfaces as previously
described in Section 12.2. Obtain samples from the freezer.
14.11.2 Wearing the appropriate gloves, plants must be rinsed with Dl water to
remove soil, silt, small insects, and other debris. Place the plants in a
stainless steel or plastic strainer, depending on the determinative sample
analysis, and rinse thoroughly with Dl water. If analyzing the sample for
both metals and organic compounds, rinse the plants carefully over a
sink, being sure not to touch the sides of the sink with the plant sample.
14.11.3 Depending on the size and texture of the plants, some may be
homogenized in the small food processor with the titanium blade.
Samples such as long grass will have to be chopped into small pieces
(approximately Vi inch) using titanium or ceramic knives. Leaves can
generally be homogenized in the small food processor without pre-
cutting.
14.11.4 Cover the balance tray with the appropriate material and record the
weight of the plant sample. Since the weight obtained from one plant
may be small, several plants may be combined to make one sample.
Utensils do not need to be rinsed between the individual samples that
comprise one composite, but utensils must always be rinsed in between
each composite sample.
14.11.5 If several plants will be composited to make one sample, the weight of
each specimen may be recorded individually, and then totaled for the
composite weight. If only one composite weight is sufficient for the
project specifications, weigh the entire composite and record that weight
14.11.6 After the plant weight for one composite or individual sample has been
recorded, place the plant(s) in the clean small processor with the titanium
blade to be homogenized, or place them onto the cleaned cutting board
to be chopped. Grind or chop the plants until they appear to be fully
homogenized.
14.11.7 Individual homogenates may be processed further to prepare composite
homogenates as required by project specifications. Composite
in LIMS.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
20 of 32
-------
homogenates must be prepared from equal weights of individual
homogenates. If required, all individual weights that make up one
composite must be recorded, otherwise one weight may be recorded for
the composite. If individual or composite homogenates were frozen prior
to extraction/digestion, these homogenates must be thawed and re-
homogenized by hand mixing prior to being extracted or digested.
14.11.8 Place the homogenized plants back into the freezer to be frozen for
future extraction/digestion.
14.11.9 All utensils and equipment must be washed between samples according
to the procedures described previously in section 12.2.
15.0 CALCULATIONS
15.1 Not Applicable
16.0 METHOD PERFORMANCE
16.1 Not Applicable
17.0 POLLUTION PREVENTION
17.1 Refer to SOP Pace054 and Pace089 for instructions on the disposal of waste
generated during the procedures previously mentioned.
18.0 DATA ASSESSMENT AND ACCEPTANCE CRITERIA FOR QUALITY CONTROL
MEASURES
18.1 Not Applicable
19.0 CORRECTIVE ACTIONS FOR OUT OF CONTROL DATA
19.1 Not Applicable
20.0 CONTINGINCIES FOR HANDLING OUT-OF-CONTROL OR UNACCEPTABLE DATA
20.1 Not Applicable
21.0 WASTE MANAGEMENT
21.1 Refer to SOP Pace054 and Pace089 for instructions on the disposal of waste
generated during the procedures previously mentioned.
22.0 REFERENCES
22.1 NELAP "Quality Systems" Manual, 2005.
22.2 U.S.EPA SW-846 "Test Methods for Evaluating Solid Waste; Volume 1B
Laboratory Manual Physical/Chemical Methods", Office of Solid Waste and
Emergency Response, Third Edition, Final Update III, December 1996.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 21 of 32
-------
22.3 EPA/6OOIR-961027, Guidance for the Preparation of Standard Operating
Procedures (SOPS) for Quality Related Documents, 1996.
22.4 US EPA 823-R-95-007, "Guidance for Assessing Chemical Contaminated Data for
Use in Fish Advisories", Volume 1: Fish Sampling and Analysis 2nd Edition, Office
of Science and Technology, Office of Water, 1995.
22.5 U.S. EPA, 1991 d
23.0 ATTACHMENTS
23.1 Fish Filleting Diagram
23.2 Fish External & Internal Anatomy
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name:
Revision:
Date:
Page:
NE132_07.doc
07
03/29/2011
22 of 32
-------
Fish Filleting Diagram
Scaled Fish
After removing die scales (by
scraping with the edge of a
knife) and rinsing the fish:
i 1 Scaleless Fish
¦
Grasp the skin at the base of the head
(preferably with pliers) and pull toward
the tail.
Note: This step
applies only for
catfish and
other scaleless
species.
Make a shallow cut through the
skin (on either side of the dorsal
fin) from the top of the head to
the base of the tail.
Make a cut behind the entire
length of the gill cover, cutting
through the skin and flesh to the
bone.
Make a shallow cut along the belly
from the base of the pectoral fin to
the tail. A single cut is made from
behind ibe gill cover to the anus
and then a cut is made on both
sides of the anal fin. Do not cut into
the gut cavity as this may
contaminate fillet tissues.
Remove the fillet.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 23 of 32
-------
23.2 FISH EXTERNAL/INTERNAL ANATOMY
EXTERNAL ANATOMY
1. Remove one fish from the storage tank, place in dissecting pan. Make
sure fish is euthanized prior to any dissection.
2. Locate all fins (Figures 1a and 1b):
Paired: pectoral (caudal to head, located ventrolateral^)
pelvic (cranial to anus, located ventrolateral^)
Single: dorsal (caudal to head on dorsal midline)
adipose (caudal to dorsal fin on dorsal midline; salmonids)
Anal: (Caudal to anus on ventral midline)
Figure 1a. Anatomy of a Fish (typical salmonid)
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 24 of 32
-------
Operculum
External
r- , Spines—>
Features K
Rudimentary Ray—>
Isthmus-^1
Abdominal Region
Belly
.V!
-Caudal
Peduncle
Figure 1b. External Anatomy of Striped Bass
3. Find the lateral line located laterally at mid-body running from head to tail.
It arches dorsally over the operculum.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 25 of 32
-------
Body
Depth
Snout
Head
Upper
Jaw
Caudal
Peduncle
Pelvic
Figure 1c. Typical Measurements Locations
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 26 of 32
-------
Common Measurements
Total Length (Maximum Length, mouth shut, tail fin pinched together)
Standard Length (to the fold, where the tail fin begins)
Figure 1d. Typical Measurements of Large Mouth Bass
4. The operculum covers the gills. Lift the opercular flap and identify the
bony gill arches, cartilagenous giii filaments, and primary lamellae
projecting off the gill filaments (Figures 1 f, 1g)
Head
P=Premaxillary bone
M=M axillary bone
D=Oentary
C=Cheek
E=Preopercle
l=lnteropercle
S=Subopercle
0=0percle
B=BranchiostegaJs
U=Upper angle of gill
cleft
A ,yA A/A
Ventricle—>
Rakers —>
Bulbus >
Arteriosus
T=Pharyngeal
Teeth
A=Arch
Esophagus
¦laments
Ventral Aorta
<—Atrium
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 27 of 32
-------
5. Lay the fish on its right side with the head to your left. Open the body cavity
with three cuts (Figure 1h). The first cut should originate just craniad to the anus
and run cranial to a point ventral to the operculum. The second cut originates
from the same point as the first and runs craniad along the dorsum of the body
cavity to a point just dorsal to the operculum. The third cut connects the first two.
All cuts should be made carefully with the blunt tip of the scissors in the body
cavity while applying slight upward pressure to avoid damaging internal organs.
Lift off the body wall.
- FIRST INCISION
SECOND INCISION
THIRD INCISION
Figure 1h. Incisions to Expose Abdominal Cavity
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 28 of 32
-------
INTERNAL ANATOMY
tr dxifW
Identify the gastrointestinal tract (Figures 1 i-1,2). Pass a blunt probe through the
oral cavity, pharynx, esophagus and into the stomach. Many fish species have
pyloric cecae, which are blind sacs projecting from the aborad portion of the
stomach. The stomach empties into the intestine, a long tubular structure
supported by thin membranes called mesenteries. The intestine terminates at
the anus. In fish the intestine is not divided into three distinct regions. The
length and complexity of the intestine is directly proportional to the amount of
plant matter consumed (herbivorous species have longer intestines). Open the
stomach and intestines and note the normal texture and appearance of the lining,
or mucosa. The intestinal mucosa will often exhibit lesions when enteric of
systemic disease is present. The spleen is a small dark red organ attached to
the mesenteries just caudal to the stomach. There may be more than one
spleen. The main auxiliary digestive organs are the liver and pancreas. The liver
is a large, tan, often leaf-shaped organ just caudal to the heart. The liver is a
good site to see many lesions and is also a good site from which to isolate
bacterial and viral pathogens. The location and size of the pancreas varies by
species. The most common location is interspersed within the liver parenchyma.
It may or may not be grossly visible. Cut the intestine near the anus, cut the
esophagus and remove the gastrointestinal tract.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 29 of 32
-------
gas bladder
esophagus stomach intestine
oviduct
urogenital
papilla
anus
brain
pharynx
bladder
7. Locate the gonads, either ovaries or testes. Ovaries will appear as numerous
spherical structures that may comprise up to 70% of body weight. Testes
may comprise up to 12% of body weight. In mature animals they will appear
as a soft white organ suspended from the dorsal body wall. Also, if you don't
see either of these organs, you might be working with an immature specimen.
Note body length and compare to literature on the species/specimen you are
working with.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 30 of 32
-------
The gonads arid kidneys of an Eastern Blue-spotted Flathead. The gonads
(testes) are the large, pale organs and the kidneys are the red tissue either
side of the backbone.
8. Along the dorsum of the body cavity lies the swim bladder. It is a thick-walled
white organ. Occasionally you may see hemorrhages in the swim bladder.
9. The kidneys also lie in the dorsum of the body cavity. The head kidney and
trunk kidney are roughly divided by the swim bladder. In some species (e.g.,
salmonids) the kidneys are almost fused. The kidneys often exhibit lesions, and
the trunk kidney is usually the preferred site for obtaining bacterial and viral
cultures. In most fish we work with in this lab, the head kidney and trunk kidney
appear fused.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 31 of 32
-------
branchial arch
Filament-adductor muscle
Secondary lamellae
Afferent and
efferent arteries
Gill filaments
Capillary bed in gill lamellae
Bony gill support
¦Afferent blood vessel from heart to gill
Efferent blood vessel from gill to body
Bony gill arch
10, The heart lies just caudal to the gills (return to previous figure, Figure 1i).
The heart is enclosed in a thin-walled sac, the pericardium. Open the
pericardium and examine the heart in situ. Blood returns from the body wall
to the sinus venosus, a thin-walled chamber which empties into the atrium.
The sinus venosus might be difficult to identify. The atrium pumps blood to
the ventricle. The atrium lies cranial and dorsal to the ventricle. The ventricle
is the main pump and largest part of the heart. Blood flows from the ventricle
craniad to the bulbus arteriosus. The thick-walled elastic bulbus helps
regulate blood pressure as blood leaves the heart. As the bulbus passes
through the pericardium en route to the gills it becomes the ventral aorta.
PACE ANALYTICAL SERVICES INC.
STANDARD OPERATING PROCEDURE
SOP Name: NE132_07.doc
Revision: 07
Date: 03/29/2011
Page: 32 of 32
------- |