PHASE 1 REPORT - REVIEW COPY
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
EPA WORK ASSIGNMENT NO. 013-2N84
AUGUST 1991
ALTERNATIVE REMEDIAL CONTRACTING STRATEGY (ARCS)
Region II
FOR
HAZARDOUS WASTE REMEDIAL SERVICES
EPA Contract No. 68-S9-2001
VOLUME 1
(BOOK 1 OF 2)
TAMS CONSULTANTS, Inc
and
Gradient Corporation
-------
PHASE 1 REPORT - REVIEW COPY
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
EPA WORK ASSIGNMENT NO. 013-2N84
AUGUST 1991
ALTERNATIVE REMEDIAL CONTRACTING STRATEGY (ARCS)
x
Region II
FOR
HAZARDOUS WASTE REMEDIAL SERVICES
EPA Contract No. 68-S9-2001
VOLUME 1
(BOOK 1 OF 2)
TAMS CONSULTANTS, Inc.
and
Gradient Corporation
-------
TANS CONSULTANTS, INC./GRADIENT CORPORATION
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
ABBREVIATED CONTENTS*
VOLUME 1 (1 OF 2)
EXECUTIVE SUMMARY
INTRODUCTION
PART A:
PART B:
PART C:
REFERENCES
GLOSSARY
LOWER HUDSON CHARACTERIZATION
A.l Physical Site Characteristics
A.2 Sources of PCB Contamination
A.3 Nature and Extent of Contamination
A.4 Review of Lower Hudson PCB Mathematical Model
UPPER HUDSON CHARACTERIZATION
B.l Physical Site Characteristics
B.2 Sources of PCB Contamination
B.3 Nature and Extent of Contamination
B.4 Data Synthesis and Evaluation of Trends
B.5 Sediment Transport Modeling
B.6 Preliminary Human Health Risk Assessment
B.7 Interim Ecological Risk Assessment
PHASE 1 FEASIBILITY STUDY
C.l Phase 1 Objectives
C.2 Remedial Objectives and Response Actions
C.3 Applicable or Relevant and Appropriate Requirements
C.4 Technology and Process Identification
C.5 Innovative Treatment Technologies (USEPA Site Program)
C.6 Initial Screening of Technologies
C.7 Treatability Studies
a:
w
VOLUME 1 (2 OF 2)
TABLES
FIGURES
PLATES
o
o
o
ro
* SEPARATE, DETAILED TABLES OF CONTENTS PRECEDE THE INTRODUCTION, PARTS A,
B AND C, TABLES, FIGURES AND PLATES.
-------
PAGE INTENTIONALLY LEFT BLANK
-------
EXECUTIVE SUMMARY
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
BACKGROUND
For approximately 30 years, two General Electric (GE) facilities, one in Fort Edward and
the other in Hudson Fails, NY, used pofychlorinated biphenyls (PCBs) to make electrical
capacitors. GE discontinued the use of PCBs in 1977, when they ceased to be manufactured
and sold in the United States. From 1957 through 1975, various sources have estimated that
between 209,000 and 1.3 million pounds of PCBs were discharged from these facilities into the
Upper Hudson River. Discharges resulted from washing PCB-containing capacitors and minor
spills.
The PCBs discharged to the river tended to adhere to sediments and subsequently
accumulated with the sediments as they settled in the impounded pool behind the former Fort.
Edward Dam. Because of its deteriorating condition, the dam was removed in 1973. During
subsequent spring floods, PCB-contaminated sediments were scoured and released downstream.
In 1976, the New York State Department of Environmental Conservation (NYSDEC)
issued a ban on fishing in the Upper Hudson River, from Hudson Falls downstream to the
Federal Dam in Troy, because of the potential risk posed by consumption of PCB-contaminated
fish. The ban remains in effect today. A commercial fishing ban on the taking of striped bass
in the Lower Hudson was also imposed by NYSDEC.
E-l
-------
In 1984 the United States Environmental Protection Agency (USEPA) completed a
Feasibility Study that investigated remedial alternatives, including dredging and upland
containment of the contaminated sediments. Later that year USEPA issued a Record of
Decision (ROD) for the Hudson River PCB Superfund Site. The ROD called for: 1) an intern
No Action decision concerning river sediments; 2) in-place capping, containment and monitoring
of remnant deposit (formerly impounded) sediments; and 3) a treatability study to evaluate the
effectiveness of the Waterford Treatment Plant in removing PCBs from Hudson River water.
Since the ROD was signed, the in-place containment remedy for the remnant deposit has
been virtually completed, and the treatability study of domestic water quality from the Waterford
treatment facility concluded that the water supplied meets all current Federal and State
standards.
In December 1989, USEPA announced that the No Action decision for the Hudson river
sediments would be reassessed. This decision was based on several factors.
The Superfund Amendments and Reauthorization Act of1986 (SARA) indicates
a preference for remedies which "'permanently and significantly reduce the
volume, toxicity or mobility of the hazardous substance involved."
USEPA policy calls for a periodic review at least every five years for as long as
hazardous substances, pollutants, or contaminants that may pose a threat to
human health or the environment remain at the site.
Technological advances have been made in processes and techniques for treating
and removing PCB-contaminated sediment.
New York State Department of Environmental Conservation (NYSDEC)
requested a reassessment of the No Action decision.
E-2
-------
The reassessment process consists of a an Interim Characterization and Evaluation which was
previously identified in the Scope of Work as Preliminary Reassessment (Phase 1), Further Site
Characterization and Analysis (Phase 2) and a Feasibility Study (Phase 3).
The Hudson River PCB Superfund site encompasses the Hudson River pom Hudson
Falls to the Battery in New York Harbor, a stretch of nearly 200 river miles. Upper Hudson
refers to that 40-mile stretch of the river upstream of Federal Dam to Fort Edward. Lower
Hudson refers to the portion of the river downstream of Federal Dam to the Battery.
PHASE 1 SCOPE AND OBJECTIVES
The Phase 1 Report is a comprehensive summary and evaluation of all available data
for the site. It is based on a compilation of approximately 30,000 records of data on sediments,
water, fish and aquatic insects, which are now entered into a computerized database. The
purpose of compiling this data is to:
provide as accurate a picture as possible of current levels of PCBs in the river
and changes in these levels since the 1970s;
identify needs for additional data;
allow a preliminary assessment of risks to human health and the environment
posed by the PCBs in the river;
make possible a preliminary assessment of potential remedies and treatment
options for the PCB-contaminated sediments.
E-3
-------
It must be emphasized that the Phase 1 Report presents an interim evaluation only, based
on currently available data. It is not intended to be a definitive characterization of the site or
the risks associated with it nor to suggest any conclusions with respect to what remedies may be
proposed at the end of the reassessment During Phase 2, USEPA will complete
characterization of the site. After completion of Phase 3, the Feasibility Study, USEPA will
determine what remedies, if any, are appropriate.
MEASURING AND REPORTING PCBS IN ENVIRONMENTAL SAMPLES
An assessment of PCB contamination requires an understanding of their chemical
complexity. Polychlorinated biphenyls are not a single chemicaL They are a class of chemicals,
containing from one to ten chlorine atoms per biphenyl molecule, yielding 209 possible
molecular configurations. Laboratory analyses for PCBs are typically reported as Aroclor
mixtures, referring to the manufacturer's trade name for PCB mixtures, with each Aroclor
mixture containing a different amount of chlorine by weight As chlorine content increases, the
PCBs tend to be less soluble, more strongly adsorbed to sediments or bioaccumulated, and less
likely to volatilize into the atmosphere.
Once released into the environment, environmental samples rarely contain PCBs that
reflect the original Aroclor mixture. In addition to physical processes, biological dechlorination
in sediments can alter the chlorine content so that it is different from the Aroclor mixture
originally discharged.
Evaluation of existing PCB data is difficult, because different laboratories have used
various methodologies to measure PCB mixtures. Chemical extraction procedures and analytical
methods can also differ for different media New laboratory methods are now available to
resolve many uncertainties in PCB measurement. Uniform methods of PCB analysis in samples
to be collected in future phases will improve understanding of the specific types of PCBs that
remain in sediments, water, fish and air.
E-4
-------
LOWER HUDSON
The data available regarding PCB contamination in the Lower Hudson are more limited
than data far the Upper Hudson, which has been the subject of intensive monitoring programs
to evaluate PCB contamination.
Physical site characteristics, including basin characteristics, hydrology, water quality and
aquatic resources were reviewed in Phase 1. The review of aquatic resources, relying upon
published studies, demonstrates the presence of a diverse aquatic ecosystem.
The total loading ofPCBs to the Lower Hudson was historically dominated by inputs
from the Upper Hudson, but has also been influenced by other sources of contamination. These
sources include sewage effluent discharges, combined storm/sewer outfalls, storm water outfalls,
i
industrial discharges, atmospheric deposition, landfill leachates and tributaries below Federal
Dam. Contributions of these additional PCB sources to the Lower Hudson, a drainage area
of5,285square miles, are difficult to estimate, because they are poorly identified and quantified.
Various researchers have estimated, however, that these sources currently contribute PCB loads
on the same order of magnitude as the load from the Upper Hudson.
Data are available to document PCBs in sediment cores, surface water and fish in the
Lower Hudson. Sediment core data indicate that maximum PCB deposition in the Lower
Hudson occurred about 1973 and has since declined. A limited number of Lower Hudson
water column measurements from 1978-1981 indicate that PCB levels declined from
approximately 0.17 ug/l (micrograms per liter) in 1978 to approximately 0.07 ug/l in 1981.
NYSDEC has reported PCB measurements for fish from the Lower Hudson, mostly in
the period of1975-1988. The data show that PCB levels in striped bass, the dominant species
monitored, have declined since 1978, with an apparent half-life of approximately five years.
Recent measurements indicate that median PCB levels in striped bass are approximately 5 to
12 mg/kg (milligrams per kilogram) in the Upper Estuary (River Mile 91 to 153) and
-------
approximately 2 to 3 mg/kg in the Lower Estuary (River Mile 12 to 76). Previous investigations
have suggested that PCBs in striped bass of the Lower Hudson are dominated by the highly
chlorinated PCB mixtures. This observation is of significant interest, because sediment data for
the Lower Hudson suggest that there are sources of highly chlorinated PCB mixtures from the
New York City metropolitan area. Further investigations will be needed to assess potential
effects of remedial efforts to reduce PCBs in the Upper Hudson on PCB levels in the Lower
Hudson.
Only a comparative and qualitative human health risk assessment was performed for the
Lower Hudson during this phase. For the Lower Hudson, consumption offish would likely lead
to PCB exposures in human populations that are smaller than or comparable to those in the
Upper Hudson, since PCB concentrations in water and fish from the Lower Hudson are less
than those in the Upper Hudson. The assessment of risks in the Lower Hudson as a result of
PCB loadings those from the Upper Hudson is complicated by the presence of multiple sources
of PCBs within the Lower Hudson.
UPPER HUDSON
An understanding of how PCBs transfer, accumulate and dissipate in sediments, water,
fish and other media is important to the characterization of existing conditions and an
evaluation of the potential benefits of remedial actions in the Upper Hudson. Major items
addressed in Phase 1 are:
potential for redistribution of PCBs in the river sediments;
transfer of PCBs in the sediments to the water column and to fish; and
human health and ecological risks from PCBs in the Hudson River site
E-6
-------
These items will be further addressed in Phase 2 in order to determine appropriate remedial
measures in Phase 3.
Sediment
Data on approximately 2,500 sediment samples are contained in the database, covering
the two principal studies sponsored by NYSDEC in 1977-78 and 1984, as well as data from
other studies by USEPA, GE and the Lamont-Doherty Observatory. The 1978 study covered
approximately 40 river miles, while the 1984 study investigated PCBs in Thompson Island Pool,
a five-mile stretch impounded by the Thompson Island Dam. The Thompson Island Dam is
the first control structure in the river downstream of the GE plants. Several factors, listed below,
hinder both detailed comparisons of results among surveys and analysis ofPCB transfer from
sediments to water and fish.
PCB measurements in sediments exhibit extreme variability over short distances.
The shifting of river sediment deposits confounds comparison of sampling results
at a given river location over time. Too few samples with adequate areal
coverage have been taken to determine trends over time in PCB levels in
sediment deposits.
Because of different laboratory measurement techniques, different methods of
reporting PCB concentrations and lack of sediment and water or sediment and
fish data obtained at the same location, statistical relationships between PCBs in
sediment and PCBs in water or fish have not been developed.
The redistribution of PCBs in sediment is largely determined by sediment scour occurring
during high river flow. The results of flood recurrence computed in this report show that
previously published computations of the magnitude of the 100-year flood or other large floods
E-7
-------
are overestimates. Thus, previous predictions of sediment scouring and PCB loadings will need
to be re-examined based on the new estimates.
Water
The United States Geological Survey (USGS) results of monitoring PCBs in the water
column from 1975-1989 were evaluated in Phase 1. Average total PCB concentrations in water
during the late 1970s ranged from 0.2 to 0.8 ug/l (ppb) and have declined to approximately 0.03
to 0.05 ug/l in the late 1980s.
PCB mass transport in water is analyzed statistically to correct for sampling bias and
account for the numerous data in which PCBs were not detected at the quantification limits set
for the analyses. These findings are summarized below.
Although the greatest number of PCB measurements coincide with high flow
periods, previous estimates have not addressed this bias in the sampling data.
Previous PCB mass transport estimates are, thus, somewhat higher than those
computed in this report.
Mass transport of PCBs from the Upper Hudson at Waterford to the Lower
Hudson has declined pom approximately 3,000 to over 4,000 kg/year in the late
1970s to approximately 150 to 500 kg/yr in recent years.
Since 1983, there has been little, if any, discernible difference in the mass load
carried by the Hudson River pom Rogers Island in Fort Edward to Waterford.
This last observation is significant, because it suggests that in recent years there is little
increase in PCB load in the Hudson River above that already in the river at Rogers Island.
Whether the PCBs in the river at the Rogers Island monitoring station, which is immediately
E-8
-------
downstream from the remnant deposit sediments, are derived from the remnant areas or other
sources remains to be investigated further. PCB results for water samples taken in 1990,
analyses of results for continued USGS water column monitoring and analyses of samples
proposed for Phase 2 of this reassessment are expected to provide more definitive information.
Fish
Just as PCBs in the water column have declined since 1977, detailed analyses of trends
over time indicate a similar decline for PCBs in fish. NYSDEC has reported analyses of PCB
levels in fish sampled in the Upper Hudson since 1975. The dominant species sampled were
largemouth bass, brown bullhead, pumpkinseed and carp. Median PCB levels in fish have
declined from levels ranging from 3 to 143 mg/kg measured in the late 1970s to current levels
i
ranging from 1 to 30 mg/kg. PCB levels in fish have also declined at downstream sample
locations. Statistical analyses presented in this Phase 1 report reveal the following current
conditions and trends.
The current upper-bound, 95 percent confidence limit of the average PCB level
for all fish sampled in the Upper Hudson from 1986 to 1988 is approximately 12
mg/kg.
Lower chlorinated PCBs in fish exhibit a half-life of approximately three to four
years, whereas the higher chlorinated PCBs appear to be declining at a much
slower rate and exhibit half-lives of 7 to 40 years, depending on fish species.
A very strong linear correlation between PCBs in fish tissue and PCBs in the
water colunin is apparent The concentration of PCBs in fish tissue, based on
lipid (fatty) content offish, is on the order of one million times greater than the
PCB concentration in the water column. This ratio is referred to as the
E-9
-------
bioaccumulation factor. Insufficient data are available to relate levels of PCBs
in fish to those in sediment.
Additional fish samples were collected by the NYSDEC in 1990 and the analyses from
these samples will be included in subsequent phases.
Sediment Transport Modeling
Development and calibration of hydraulic and sediment transport models have been
initiated as part of the Phase 1 work. Because PCBs in the river are bound primarily to
sediments, scour of sediments is a crucial mechanism to the movement of PCBs A
mathematical model provides one tool to predict potential scour and redeposition of sediments
containing PCBs. A basic modeling framework has been developed in conjunction with the
analysis of available data in order to determine the type and extent of modeling that may later
be appropriate and feasible.
Preliminary Human Health Risk Assessment
Human exposure to PCBs in the environment or in the workplace generally does not
result in any immediate or acute toxicity, but such exposures are of a public health concern
because of the persistence of PCBs in the environment, their potential to bioaccumulate in
animal and human tissues and their potential for chronic toxicity. Occupational exposures to
relatively high concentrations of PCBs have resulted in effects on liver-function as weU as effects
on the skin. Emerging evidence indicates that PCBs may also be related to other toxic effects
in humans, such as developmental or neurological effects. USEPA has classified PCBs as
probable human carcinogens, based on the induction of cancers in laboratory animals.
A preliminary human health risk assessment for the Upper Hudson was performed during
Phase 1. USEPA considered it important to establish at an early stage the working assumptions
E-10
-------
for the risk assessment in order to allow the public sufficient opportunity to review and comment
upon these assumptions. This preliminary health risk assessment will be modified in the future
as new data become available, or if new information regarding the toxicity ofPCBs is accepted
by USEPA through a scientific review process before completion of the Reassessment
Quantitative health risks associated with PCS exposure and uptake were calculated for
pathways with adequate data. The exposure pathways quantified in the risk assessment are:
consumption offish;
drinking of river water, and
incidental contact with PCBs in water and sediments during recreational activities
associated with the river.
Based on available data, there appear to be unacceptable potential cancer and nan-
cancer risks associated with regular ingestion of fish from the Upper Hudson River. This risk
is based upon the assumption that local residents catch and consume fish from the Upper
Hudson. However, as mentioned above, because of the potential risk posed by PCB
contaminated fish, NYSDEC issued a fishing ban in 1976. The ban remains in effect today.
Recent surveys of angler activity by NYSDEC and fish consumption rates from NYSDEC and
USEPA were utilized, but no specific survey of the population in question has been conducted
to confirm these assumptions.
The risks associated with exposure as a result of drinking river water and recreational
contact were estimated to be within the acceptable range.
E-ll
-------
Other potential exposure pathways, such as inhalation of contaminated air, consumption
of local crops or deary products and ingestion of breast milk by infants, are discussed, but not
quantified. Although some available data indicate low levels ofPCBs in plants and car in the
vicinity of Fort Edward and the Hudson river, these data are insufficient to perform a
quantitative risk assessment Furthermore, the contribution ofPCBs from other sources cannot
be determined.
Interim Ecological Risk Assessment
7Tie interim ecological risk assessment relies upon available data for selected indicator
species in the Upper Hudson River ecosystem. Data are currently insufficient to justify a
quantitative ecological risk assessment. Although thousands of samples have been collected of
sediments, water and fish in the Upper Hudson River over a period of years, there are little data
that relate PCB levels in the River to demonstrated harm to fish or other organisms. Only one
published report was found regarding abnormal cell growth in fish from the contaminated stretch
of the Upper Hudson. Although the researchers conclude that the observed abnormalities may
be attributable to hazardous organics, additional study would be required to establish a causal
relationship and to identify the chemical or chemicals responsible.
The concentration ofPCBs measured in sediment, water, insect larvae and fish, and
estimated in fish-eating birds and mammals, have been compared to published guidelines and
toxicity values for PCBs. Recent PCB levels in water exceed freshwater Ambient Water Quality
Criteria for the protection of aquatic life by two to five-fold. Additional evaluations will be
necessary to assess species health and determine levels of ecological risk.
E-12
-------
FEASIBILITY STUDY
Potential remedial technologies and processes capable of treating river sediments are
identified in this initial step toward a final Feasibility Study. These technologies and processes
are: containment; natural PCB degradation in sediments; removal; disposal; and treatment,
including physical, chemical, thermal and biological treatment techniques. An initial screening
of technologies is presented. No particular technology nor the possibility that no remedy may
be warranted is eliminated from further consideration. Preliminary approaches to remedial
options and recommendations for treatability studies are discussed.
PHASE 2
4
Based on the understanding reached in the Phase 1 process, field sampling and
additional data evaluation are necessary in Phase 2 to provide improved understanding of PCB
levels and transfer mechanisms among sediments, water, air and biota.
In order for the Hudson River PCB Oversight Committee (HROC) and the participants
in the Community Interaction Program (C1P) to provide input into the Phase 2 work, these
groups will be allowed sufficient time to evaluate the Phase 1 Report. Therefore, a full Phase
2 Work Plan will not be developed until after comments are received on the Phase 1 Report.
There are some data, however, that USEPA believes should be collected in Fall 1991, because
the data will be needed to guide subsequent sampling activities and allow the project schedule
to be maintained. These priority sampling activities, which will be conducted this Fall, are
described in the Phase 2A Sampling Plan. Activities included in this plan are geophysical
surveys and water column monitoring in the Upper Hudson, as well as sediment corings in both
the Upper and Lower Hudson.
E-13
-------
When the Phase 2 Work Plan is issued for review, it will include a Phase 2B Sampling
Plan and the description of other activities necessary to complete characterization of the ate, as
well as a summary of the Phase 24 activities. HROC and CIP participants will be allowed to
review and comment on this Plan prior to initiation of Ms second sampling effort
E-14
-------
INTRODUCTION
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
CONTENTS
I. INTRODUCTION
1.1 Purpose of Phase 1 Report
1.2 Purpose of Reassessment RI/FS
1.3 Site History
1.3.1 Prior to 1980
1.3.2 Post 1980
1.4 Guide to Phase 1 Report
1.4.1 Relationship to Phase 1 Work Plan
1.4.2 Organization of Phase 1 Report
PLATES
Plates Located In Volume 1 (2 of 2)
1.2-1 Upper Hudson Site
1.2-2 Lower Hudson Site
1.2-3 Thompson Island Pool and Remnant Deposits
1
-------
THIS PAGE INTENTIONALLY LEFT BLANK
11
-------
I. INTRODUCTION
1.1 Purpose of Phase 1 Report
This Phase 1 report provides an Interim characterization and evaluation of
the Hudson River PCB Superfund site. It summarizes the results of the first
phase in a three-phase Reassessment Remedial Investigation/Feasibility Study
(RI/FS) to reassess the 1984 No Action decision of the United States
Environmental Protection Agency (USEPA) concerning sediments contaminated with
polychlorinated b1phenyls (PCBs) in the Upper Hudson River.
In December 1990, USEPA Issued a Scope of Work for reassessing the No
Action decision for the Hudson River PCB site. The scope of work Identified
three phases:
Phase 1 - Preliminary Reassessment
Phase 2 - Further Site Characterization and Analysis
Phase 3 - Feasibility Study
This report presents the results of Phase 1 only. The Phase 1 report
contains a compendium of background material, discussion of findings where
findings could be made and preliminary assessments of risks.
The material presented here 1s not Intended to characterize the site
definitively nor to draw final conclusions. The principal reason for submitting
this interim report early 1n the reassessment process Is to offer sufficient
background material to the USEPA, other concerned agencies, government officials,
the Hudson River Oversight Committee, the Science and Technical Committee, the
Steering Committee, liaison groups and the public, so that these parties can
reach Informed judgments concerning the technical direction and focus of the
project through the various phases.
1-1
-------
1.2 Purpose of Reassessnent RI/FS
In December 1989, USEPA, Region II announced that It would conduct a
reassessment of Its September 24, 1984 No Action decision concerning the
sediments contaminated with PCBs In the Upper Hudson River.
USEPA decided to reassess the No Action decision, based on the following
events that have occurred since 1984.
t With the Superfund Amendments and Reauthorization Act of 1986 (SARA)
came the Indication that preferred remedies were those which
"permanently and significantly reduce the volume, toxicity or
mobility of the hazardous substance Involved."
USEPA policy 1s to perform periodic review for both pre-and post-
SARA RODs at least every five years for as long as hazardous
substances, pollutants, or contaminants that may pose a threat to
human health or the environment remain at the site.
Technological advances have been made 1n processes and techniques
for treating and removing PCB-contam1nated sediment.
New York State Department of Environmental Conservation (NYSDEC)
requested a reassessment of the No Action decision.
The reassessment Is being performed for the PCBs contained within the
river-bottom sediments of the Upper Hudson between Hudson Falls and Federal Dam
In Troy, New York (see Plate 1.2-1). The Superfund site Itself, however, extends
to the Battery 1n New York Harbor (see Plate 1.2-2).
The reassessment also evaluates the threat of PCBs entering the river from
the remnant area (see Plate 1.2-3) and assesses environmental Impact on the Lower
Hudson. Previously dredged PCB-contam1nated sediments contained 1n upland
disposal facilities will not be addressed In this study.
1-2
-------
1.3 Sitฎ History
1.3.1 Prior to 1980 %
ฆV
PCBs were manufactured by Monsanto Corporation between 1927 and 1977 and
were distributed under the generic nane Aroclor. Two General Electric Company
(GE) capacitor manufacturing plants located 1n Fort Edward and Hudson Falls, New
York began to use PCBs 1n 1946. In-plant sources of PCB discharges have been
characterized as both minor spills and effluent from washing capacitor cans, with
the latter being the major source. Capacitor cans were flood-filled with
dielectric fluid and then washed with detergent and water to remove excess
material. Contaminated wash water was discharged directly to the river. This
practice was discontinued about 1973 (Brown, Jr. et al.t 1984). .
During a 30-year period from 1946 to 1977, PCBs were discharged Into the
Upper Hudson River from the two GE plants. Discharged PCBs adhered to the
sediments 1n the bottom of the river and accumulated 1n areas behind the Fort
Edward Dam. When the dam was removed 1n 1973 because of Its deteriorating
condition, PCB-contam1nated sediments were released downstream, particularly
during large spring floods 1n 1976 and 1983.
PCBs have been associated with a variety of adverse health effects.
Studies performed on rats, mice and monkeys have revealed that various kinds of
toxic effects are associated with PCBs, such as liver damage, reproduction
effects, skin disturbances and cancer.
The first report of PCB contamination In the Hudson River was published in
1970. In 1971, NYSDEC added PCBs to their statewide analyses of pesticide
residues In fish, although no results were released publicly until 1975. After
USEPA Investigations In 1974 of PCB contamination In the Fort Edward area, NYSDEC
Intensified Its PCB sampling program. In 1976, following the 1975-1976 sampling
effort, NYSDEC banned all fishing on the Upper Hudson River, from Albany north
through Fort Edward. The commercial striped bass fishery In the Lower Hudson was
also closed at the same time. The bans are still 1n effect today. In addition,
1-3
-------
the presence of PCBs restricted dredging activities. The New York State
Department of Transportation (NYSDOT) had periodically dredged the river, which
1s prone to sediment buildup, 1n order to maintain a minimum depth to accommodate
river traffic. According to NYSDOT, no channel maintenance dredging has occurr^J
from 1984 to the present (1991).
1.3.2 Post 1980
In the 1980s, site activities diverged 1n two directions. One direction
was pursued by the NYSDEC Project Sponsor Group (PSG) and Included the Hudson
River PCB Reclamation Demonstration Project, which 1n 1989 became the more
comprehensive Project Action Plan. The second direction was pursued by USEPA and
Included the National 011 and Hazardous Substances Pollution Contingency Plan
(NCP) and Comprehensive Environmental Response, Compensation, and Liability Act
(CERCLA) or Superfund process. Only the direction pursued by USEPA under the
NCP/CERCLA process 1s discussed below. The PSG activities are described in
detail in the Community Relations Plan (December 1990.) for this Reassessment.
The 1984 Record of Decision (ROD) stated that a technologically feasible,
cost-effective, remedial response to the river sediments was not available that
would reliably and effectively mitigate and minimize damage to public health,
welfare and the environment. This decision was based on the results of the NUS
Feasibility Study (FS) dated April 1984. At that time 1t was deemed more
appropriate to address the sediments in connection with the Hudson River PCB
Reclamation Demonstration Project being pursued by the PSG. There were several
reasons for this decision: (1) the modeling and sampling data collected at that
time Indicated a decreasing threat to public health and the environment; (2) the
reliability and effectiveness of extant dredging technologies were subject to
considerable uncertainty; (3) the estimated high cost of dredging and disposal
were considered likely to rule out such options, based on fund balancing
considerations, especially given the moderate degree of risk reduction which
might have been achievable. The ROD stated that this decision was to be
reassessed In the future if, during the interim evaluation period, the
reliability and applicability of 1n situ or other treatment methods were to be
1-4
-------
demonstrated or if techniques for dredging of contaminated sediment were further
developed.
The 1984 ROD addressed five remnant deposits (see Plate 1.2-3) behind the
Old Fort Edward Dam and the river sediments. The ROD reflected USEPA's decision
to perform In-place containment, or capping, of the remnants, stabilization of
the associated rlverbanks and revegetatlon of the areas. As stated in the ROD
(USEPA, 1984): "The appropriateness of further remedial action for these sites
will be reexamined, 1f EPA decides at a later date to take additional action with
respect to sediments 1n the river." The construction of the remnant caps Is
essentially complete. No 1n-place containment was required at one of the remnant
deposits (site 1).
The 1984 ROD also included performance of a treatability study to evaluate
the effectiveness of the Town of Waterford's treatment plant 1n removing PCBs
from Hudson River water. The Town of Waterford 1s located 40 miles south of the
remnant deposits and was selected for evaluation, because 1t 1s the northernmost
community downstream that receives Its water supply directly from the Hudson
River. Findings Indicated that PCB levels 1n the water supplied by the Waterford
Water Works did 1n fact meet standards for public water supplies.
Z.4 Guide to Phase 1 Report
1.4.1 Relationship to Phase 1 Work Plan
The Phase 1 Work Plan (January 1991) detailed five main tasks to be
completed during Phase 1.
Task 1 - Site Characterization and Data Synthesis
Task 2 - Evaluation of Fish and Food Chain PCB Bloaccumulation
Task 3 - PCB Transport Model
1-5
-------
Task 4 - Baseline Risk Assessments
Task 5 - Remedial Technology Assessment
Rather than report results of Phase 1 by these Work Plan tasks, the
Information gathered and the findings of this phase are presented here in a
format that 1s more consistent with the CERCLA RI/FS reporting process. This
format will be compatible with the final product of this project, which will be
a Reassessment RI/FS report.
Phase 1 was originally called the Preliminary Reassessment. As a
consequence of discussions at the Hudson River Oversight Committee (HROC) meeting
of April 4, 1991, the term "preliminary reassessment" was considered potentially
misleading and a new designation for Phase 1 was considered appropriate. In order
to stress the fact that additional characterization and evaluation of the site
are required prior to a final characterization and prior to any decision on
reassessment, the title Interim Characterization and Evaluation Report has been
chosen.
1.4.2 Organization of Phase 1 Report
This Phase 1 report 1s Volume 1 of the overall three-phase study. In order
to accommodate the amount of material presented in Phase 1, two books have been
prepared. Book 1 of Volume 1 contains all text, divided as Parts A, B and C.
Book 2 of Volume 1 contains all tables, figures, and plates,
A table of contents Introduces each Part (A, B and C) in Book 1. Each Part
is paginated by section and page number within that section. For example, page
B.6-7 Is page number seven (B.6-Z) within Section B.6 (JLฃ-7). Tables, figures
and plates referenced In the Introduction (I) and Parts A, B and C are designated
so that they can be easily referenced to a report section. For Instance, Table
B.3-9 Indicates that the table was referenced In Section B.3 (B.3-91. The nine
(B.3-2) further Indicates that the table 1s the ninth table within that section.
1-6
-------
Synopses intended to guide the reader In selecting areas of focus or
Interest are provided 1n Book 1. A synopsis of Part A and of Part C is located
at the beginning of each part. Because of the length of Part B, separate
synopses for Sections B.l through B.7 are provided prior to each sectloi^
The interim site characterizations of the Lower Hudson and Upper Hudson are
presented In Book 1, Parts A and B, respectively. Initial emphasis has been
placed on the Upper Hudson, because the basic project premise Is to reassess
options for remediation of the PCB-contaminated sediments of the Upper Hudson.
This approach 1s not intended to diminish the importance of evaluating the
impacts on human health and the environment of the Lower Hudson as a result of
PCB contamination 1n the Upper Hudson. Such Impacts will be pursued sore
emphatically In Phases 2 and 3.
The Interim site characterization for the Lower Hudson (Part A) is based
on a summary of available literature; particular attention is paid to sources of
PCBs into the Lower Hudson and to the river's aquatic ecology. As no
comprehensive data synthesis and evaluation were undertaken at this time, the
reader should not expect new or previously unreported analyses. For those
readers not familiar with the chemical aspects and structure of PCBs, an
explanation 1s provided 1n Section A.2.
The interim site characterization for the Upper Hudson (Part B) 1s based
on comprehensive collection, synthesis and evaluation of available data. The
understanding or Interpretation of the Information and analyses presented for the
Upper Hudson demands varying levels of technical knowledge. To assist the
reader, an overview of the nature of Part B sections 1s provided here.
Sections B.l and B.2, both general In nature, report on the physical site
characteristics and the sources of PCB contamination in the Upper Hudson.
Section B.3 discusses the available data that were collected, reviewed and
synthesized for sediments, water, fish, air, and plants. Parts of this section
utilize statistical techniques, some of which will not be familiar to all
1-7
-------
readers. The conclusion of Section B.3 discusses the adequacy of the data and
presumes some knowledge of analytical chemistry.
Section B.4 analyzes the Interrelationships among data sets and examines
trends In the data over time 1n order to extrapolate trends In PCB levels and
transport Into the future. Although this section relies heavily upon statistical
techniques, readers will gain new Insights Into the data, some of which
contradict and/or extend existing knowledge.
Section B.5, which reports on sediment transport modeling, 1s highly
mathematical and oriented to those readers who will wish to provide comment on
or Input to the modeling effort. Achieving that objective has dictated a
mathematical presentation.
Finally, 1n Part B, Sections B.6 and B.7 provide preliminary human health
and ecological risk assessments, utilizing Information reported 1n the previous
sections. All readers are reminded that the findings presented In these sections
are preliminary since they are based on available Information.
Part C contains Phase 1 of the Feasibility Study for the Upper Hudson.
Presented here 1s a review of regulations, available technologies and processes
to treat PCB sediments, some Innovative technologies under development and the
status of natural dechlorination or blodegradatlon of PCBs In sediments.
Following Parts A, B, and C 1n Book 1 are the 11st of references and a
glossary. Readers are Invited to suggest changes to the list of references and
to recommend additional terms for future expansion of the glossary, as both of
these will form the basis of an expanded reference 11st and glossary In
subsequent documents.
1-8
-------
SYNOPSIS
LOWER HUDSON CHARACTERIZATION
(Sections A.1 through A.4)
Part A provides an interim characterization and evaluation of Lower Hudson River
characteristics pertinent to the Hudson River PCB reassessment. Presented here are physical site
characteristics, sources of PCB contamination, the nature and extent of Lower Hudson PCB
contamination and an overview of a published mathematical model by Thomann et aL (1989)
on PCB dynamics in the Lower Hudson.
The discussion of physical site characteristics (A.1) contains information on basin
characteristics, hydrology, water quality and aquatic resources. The description of basin
characteristics (drainage areas and climate) covers both the Upper and Lower Hudson to
establish a framework for the entire site. The discussions of hydrology and water quality for the
Lower Hudson describe the physical/chemical factors that affect each. The review of aquatic
resources, relying upon published studies, demonstrates a diverse aquatic ecosystem.
There are several sources of PCB contamination (A.2) in the Lower Hudson. PCB
loadings to the Lower Hudson have occurred from the Upper Hudson, but also from sewage
effluent discharges, tributary contributions, combined sewer/storm water and storm water
outfalls, atmospheric deposition, landfill leachate and other sources within the New York City
metropolitan area, all within the Lower Hudson Basin itself. These additional PCB sources are
important to consider, since some have been estimated to contribute PCB inputs of 'similar
magnitude to current loads from the Upper Hudson.
The nature and extent of PCB contamination is analyzed, using available data for
sediments, water and fish (A.3). As demonstrated by dated sediment cores, maximum PCB
deposition in the Lower Hudson occurred around 1973 and has decreased subsequently.
Sediment cores also indicate that sediment influenced by New York City metropolitan area
inputs has recently been accumulating higher PCB levels than those farther upstream. Althoutfi
water column PCB measurements since 1981 are lacking, 1978-81 data show that PCB levels
declined during that period. Studies indicate that PCB concentrations in striped bass have
declined. For migrant/marine fish species and freshwater resident species, data are limited or
dated.
A mathematical model of PCB dynamics in the Lower Hudson (Thomann et aL, 1989)
is examined (A.4). This model considers many aspects of mass transport, geochemistry and
ecology and evaluates the time history of PCB inputs. The model indicates that PCB load to
the Lower Hudson via the Upper Hudson had declined substantially since 1973. Various
assumptions used in the model regarding mass transport estimates, geochemical processes and
ecological parameters are discussed in order to provide perspective on its results.
-------
I
PAGE INTENTIONALLY LEFT BLANK
-------
PART A
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
CONTENTS
Page
A. LOWER HUDSON CHARACTERIZATION A.1-1
A.l Physical Site Characteristics A.1-1
A.1.1 Hudson River Basin Characteristics A.1-1
A. 1.1.1 Drainage Areas A.1-1
A.1.1.2 Climate A.1-2
A.1.2 Hydrology A.1-4
A.1.2.1 Physical Characteristics A.1-4
A.1.2.2 Freshwater Flow and Tributary Inputs A.1-5
A.1.2.3 Circulation A.1-6
A.1.3 Water Quality A.1-7
A.1.3.1 Overview A. 1-7
A.1.3.2 Salinity A.l-8
A.1.3.3 Temperature A.l-8
A.1.3.4 Dissolved Oxygen A.1-9
A.1.3.5 Turbidity and pH A.1-9
A.1.3.6 Municipal Wastewater Discharges A.1-10
A.1.3.7 Phosphates and Nitrates A. 1-10
A.1.3.8 Classification and Use A.1-11
1
-------
PART A
CONTENTS
(continued)
%
Page
A.1.4 Aquatic Resources In the Lower Hudson A.1-12
A.1.4.1 Conceptual Framework A.1-12
A.1.4.2 Physical Constraints A.1-16
A.1.4.3 Trophic Components 1n the Lower Hudson A.1-17
Primary Producers A.1-17
Consumers (Invertebrates and Fish) A.1-20
Synthesis A.1-28
A.2 Sources of PCB Contamination A.2-1
A.2.1 Description of PCBs A.2-1
A.2.2 Lower Hudson PCB Loadings A.2-2
A.2.3 Sewage Effluent Discharges A.2-3
A.2.4 Tributary Contributions A.2-4
A.2.5 Combined Sewer/Storm Water and Storm Water Outfalls A.2-4
A.2.6 Atmospheric Deposition A.2-5
A.2.7 Landfill Leachates A.2-5
A.2.8 Other Sources of PCBs A.2-6
A.3 Nature and Extent of Contamination A.3-1
A.3.1 Sediments A.3-1
A.3.2 Water A.3-5
11
-------
PART A
CONTENTS
(continued)
A.3.3 Fish
P"
A.3-7
A.3.3.1
Overview of Previous Monitoring Programs
A.3-7
A.3.3.2
Striped Bass
A.3-10
A.3.3.3
Other Migrant/Marine Species
A.3-12
A.3.3.4
Resident Freshwater Species
A.3-13
Review of Lower Hudson PCB Mathematical Model
A.4-1
A.4.1 Thomann Model
A.4-1
A.4.1.1
Overview
A.4-1
A. 4.1.2
Mass Transport Estimates
A.4-4
A.4.1.3
Geochemical Processes
A.4-5
A.4.1.4
Ecological Parameters
A.4-6
Uptake of PCBs from Water
"iฉconcentration)
A.4-6
o. c ite of PCBs from Other Sources
(Bioaccumulatlon)
A. 4-7
Elimination of PCBs
A. 4-8
A.4.2 Simulations
Relevant to Upper Hudson Remediation
A.4-8
111
-------
PART A
CONTENTS
(continued)
TABLES
Tables Located In Volume 1 (2 of 2)
%
Mean Annual Flow for Hudson River Tributaries
Public Water Supplies on the Lower Hudson River
Aroclor Mixtures
Summary of Non-Point Source Loads to the Lower Hudson River
Inventory of PCBs In the Sediments of the Lower Hudson River
Comparison of PCB Concentrations in Suspended Matter and Sediment
Core Tops Near River Mile 3
Count of Fish Samples by Year
Hudson River Fish Species and Percent Lipid
Striped Bass, Total PCBs (ppm), Lower Hudson
Striped Bass L1p1d-Adjusted Aroclor Concentrations
Total PCBs (ppm), Various Fish Species, Lower Hudson
FIGURES
Figures Located In Volume 1 (2 of 2)
Annual Precipitation for Albany (1890-1985) and New York City (1826-
1985)
Mean Monthly Flow of the Hudson River at Federal Dam In Water Year
1962
Comparison of Hudson River Upper Basin and Lower Basin Runoff
a: Mean Monthly Flow for Water Year 1986
b: Mean Monthly Flow for Water Year 1984
1 v
-------
PART A
CONTENTS
(continued)
A.1-4 Fresh Water Contributions to the Lower Hudson River
a: Flow Contributions by Tributary
b: Flow Contributions by River Mile
A.2-1 PCB Structure and Group
A.3-1 Total PCB Levels 1n Dated Hudson River Sediment Cores By River Kile
A.3-2 Highly Chlorinated PCB Homologues in Lower Hudson River Sedlnents
A.3-3 Decreasing Sediment PCB Levels 1n Hudson River Sediments Over Tine
A.3-4 L1p1d-Based Aroclor Concentration: Striped Bass, Lower Hudson
A.3-5 Striped Bass (below River Mile 80): Aroclor 1016, Lipid-Based
A.3-6 Total PCBs in Fish, Tappan Zee Bridge: Lipid-Based Values
A.3-7 Total PCBs 1n Fish at Catskl11 (R. M. 114): Lipid-Based Concentra-
tions
PLATES
Plates Located in Volume 1 (2 of 2)
A.1-1 Hudson River Drainage Basin Location Map
A.1-2 Lower Hudson River Surface Water Classifications
v
-------
PART A
CONTENTS
(continued)
PAGE INTENTIONALLY LEFT BLANK
vi
-------
A. LOWER HUDSON CHARACTERIZATION
A.l Physical Site Characteristics
A.1.1 Hudson River Basin Characteristics
A. 1.1.1 Drainage Areas
The Hudson River Basin 1s discussed first to establish a framework for this
discussion of the Lower Hudson (Part A) and subsequent discussion (see Part B)
of the Upper Hudson.
The source of the Hudson River 1s at Lake Tear-of-the-Clouds, a two-acre
pond located on Mount Marcy (Boyle, 1969) 1n the Adirondack Mountains 1n northern
New York State. From the Adirondack headwaters, the Hudson flows 1n a southerly
direction for approximately 315 river miles to the Battery 1n New York City
(River Mile 0) at the southern tip of Manhattan Island. The Hudson River
drainage basin encompasses an area of 13,390 square miles (Plate A.1-1) and has
three distinct parts.
The Upper Hudson River flows from Mount Marcy 1n the Adlrondacks to
the Federal Dam at Green Island, Troy, New York. The drainage area
of this segment 1s approximately 4,640 square miles.
The Mohawk River sub-basin originates 1n the southern Tug Hill
Plateau and flows southeasterly to its confluence with the Hudson
River north of Albany, New York. The drainage area of the Mohawk
sub-basin Is approximately 3,465 square miles.
The Lower Hudson River flows 153.4 miles from the Federal Dam at
Troy to the Battery. The drainage area of the Lower Hudson basin is
approximately 5,285 square miles. This segment of the Hudson is
tidal to the Federal Dam.
A.1-1
-------
A.1.1.2 Climate
Because the Hudson River Basin occupies a substantial portion of New York
State, nearly all of the state's climatic conditions occur within the basin. The
climate of New York State 1s subject to air masses originating from three
principal areas. Masses of cold, dry air frequently arrive from the northern
Interior of the continent. Prevailing winds from the south and southwest
transport warm, humid air, modified by the Gulf of Mexico and adjacent
subtropical waters. The third great air mass flows Inland from the North
Atlantic Ocean and produces cool, cloudy and damp weather. This maritime
Influence, Important to the southernmost portion of the Hudson Basin, Is
secondary to that of the more prevalent air mass flow from the continent.
Nearly all storm and frontal systems moving eastward across the continent
pass through or in close proximity to New York State. Storm systems often move
northward along the Atlantic coast and Influence weather and climate of the Lower
Hudson Basin and Long Island; such systems can also influence weather conditions
1n the more northern portions of the Hudson Valley.
Precipitation 1s variable and is Influenced by topography and proximity to
ocean and lake sources of moisture. Nevertheless, precipitation is quite
uniformly distributed throughout the year within the basin as a whole, with the
least amount generally occurring 1n the winter and the greatest amount occurring
during the warm season. The annual precipitation throughout the Hudson River
Basin varies from about 35 inches in the Albany area to more than 55 Inches in
the higher elevations of the Catsklll and Adirondack Mountains. The Lower Hudson
maximum average annual precipitation 1s approximately 46 inches.
r-
Precipitation records collected at New York City (Central Park) since 1826
and at Albany since 1890 (Figure A.1-1) show that for the period 1826-1985, the
average annual precipitation at New York City was 42.46 Inches and for 1890-1985,
the mean annual precipitation at Albany was 34.24 Inches. There Is wide
variability in annual precipitation for the two areas as shown 1n this figure.
A. 1-2
-------
A five-year moving average for both New York and Albany data, also plotted on the
figure, smooths out the year to year variability. From this plot, the
correlation between trends of high and low precipitation for both New York and
Albany (especially 1n the period 1950-1985) suggest a cyclic, rather than random,
pattern to annual precipitation. y
In a typical year, the highest streamflow throughout the basin occurs 1n
the spring as a result of snowmelt and precipitation. Low streamflow generally
occurs 1n the late summer, when evapo-transplration effects are the greatest.
Floods, however, can occur during any time of the year. They are generally the
result of snowmelt/preclpitatlon during the winter and spring season, hurricanes
during the June to October period, or thunderstorms during the summer.
Floods resulting from snowmelt/preclpitatlon or hurricanes generally affect
larger areas and the larger streams. Thunderstorms with rains of high Intensity
over a small area produce the maximum discharge for a variety of smaller streams.
The Upper Hudson and Mohawk sub-basins contain many lakes and swamps, which
significantly influence flood flow. In contrast to the Upper Hudson, floods in
the Lower Hudson are not significantly affected by storage, but are related to
the slope of the main channel.
Evaporative losses for the Lower Hudson River were calculated by Garvey
(1990) to be about 570 cubic feet per second (cfs) or 16 n'/s during July to 220
cfs (6.3 m3/s) during October. Water loss due to evaporation was most
significant 1n the summer, at about three percent of the mean annual freshwater
flow. In all other seasons, evaporative loss was closer to one percent of the
total flow.
A. 1-3
-------
A.1.2 Hydrology
A.1.2.1 Physical Characteristics
Since the main channel of the Lower Hudson runs fairly straight along a
north/south axis, It permits a rather precise definition of specific river
locations along a north-south axis. Locations 1n the river are usually specified
as river miles (RM). For this study, the Lower Hudson 1s defined as that portion
of the Hudson River from the Federal Dam at Troy (River H1le 153.4) to the .
Battery (River Mile 0) at the southern tip of Manhattan. River miles south of
the Battery are denoted with a negative (-) value.
In contrast to the rather steep gradient of 5.0 meters (m)/m11e north of
Fort Edward and more moderate gradient (1.0 m/m1le) south of Fort Edward to the
Federal Dam at Troy, the Lower Hudson Is considered to be a drowned river valley
with a gradient of only 0.01 m/mlle. With the exception of the Tappan Zee,
Haverstraw and Newburgh Bays, the Lower Hudson has a narrow geometry of less than
0.9 miles In width (Moran and Llmburg, 1986).
The navigational channel Is 32-feet deep from The Battery to Albany and 14
feet deep from Albany to Troy. Although the Lower Hudson has an average depth
of about 27 feet, Stedfast (1980) records a maximum depth of more than 200 feet
1n the vicinity of West Point as the river cuts through the Hudson Highlands.
The total surface area of the Lower Hudson Is about 129 square miles (Hammond,
1975). The total volume of water 1s approximately 0.74 cubic miles (Hammond,
1975) with the greatest volume recorded within the Haverstraw Bay region from
River M11e 25-40 (Texas Instruments, 1977).
The Lower Hudson experiences two tidal cycles dally. The tidal range 1s
about 4.5 feet at the mouth, 2.7 feet at West Point and about 4.7 feet at the
head of tide at Troy. The Increase In tidal range in the upper reaches of the
Lower Hudson results from the constrlctlonal effects of the diminishing
cross-sect1onal area of the river.
A.1-4
-------
A.1.2.2 Freshwater Flow and Tributary Inputs
The mean annual freshwater flow of 19t500 cfs (550 m'/s) 1s fairly large
in comparison to other rivers located 1n the northeastern United States. The
freshwater flow, however, 1s still small 1n comparison to the dally tidal
movements. According to the USGS, the maximum tidal current discharge at
Poughkeepsle has been as high as 240,000 cfs (6900 m'/s), which 1s more than an
order of magnitude greater than the annual mean freshwater discharge past the
Battery.
The runoff pattern for the Lower Hudson drainage basin normally contains
a large seasonal signal 1n spite of the relatively constant precipitation rate
throughout the year (USGS, 1986). This seasonal signal Is a result of melting
winter snow In early spring, producing a major Increase 1n freshwater discharge.
Flow at the Federal 0am can be characterized as having two basic regimes, a low
steady flow of about 5300 to 7100 cfs (150 to 200 m3/s) for nine months of the
year and a large spring surge from March to May resulting from the melting of
winter snow. Figure A.1-2 (from Hammond, 1975) Illustrates flow over the Federal
Dam at Troy for water year 1962. Superimposed on this flow 1s the seasonal
signal from the lower tributaries to the estuary.
The relationship between the flow of a representative Lower Hudson
tributary and the flow at the Federal Dam at Troy Is Illustrated 1n Figure
A.l-3a. For water year 1986, the spring thaw 1n the lower basin, as Indicated
by the Wall kill River flow, 1s evident about one month before the thaw In the
Upper Hudson. Exceptions to this pattern occur, as shown In Figure A.l-3b for
water year 1984. Water year 1984 had high runoff rates occurring several times
throughout the year; the flow rate patterns of the upper and lower basin were
similar because of large storm systems and warmer temperatures, affecting the
upper and lower parts of the drainage basin concurrently.
A.1-5
-------
An Interesting feature of the Lower Hudson River Is the fiow contributions
of tributaries along the river axis (Figure A.l-4a and Table A.1-1). On average,
approximately 50 to 70 percent of the freshwater that enters the Lower Hudson
flows over the Federal Dam at Green Island. As a consequence of the topography
of the Lower Hudson drainage basin, more than 75 percent of all additional
tributary Inputs, other than sewage, occur north of Poughkeepsle at River Nile
75 (Figure A.l-4b).
A.1.2.3 Circulation
The basic features of circulation within the Lower Hudson are well
described as a quasi-two-layer system (Stommel, 1953; Prltchard, 1955, 1969).
Seaward (and southward) flow of the surface layer Is driven by the gradient In
river surface height from the northern end of the Lower Hudson at Troy to The
Narrows, located south of The Battery. This gradient 1s dependent upon the total
freshwater flow. The net movement of the lower layer 1s upstream, northward from
The Narrows, 1n response to the density gradient between the freshwater supply
to the north and seawater, supplied to the Lower Hudson largely through The
Narrows. The net result of these flows superimposed on the tidal surges 1s to
mix the fresh and salt waters, creating a salt distribution Intermediate between
the vertical Isohallnes of a well-mixed system like the Thames of Great Britain
and the nearly flat Isohallnes of a salt wedge estuary like the Mississippi delta
(Deck, 1981). (Isohallne Is defined as a line connecting points of equal
salinity.)
Because of the narrow geometry of the Lower Hudson, the location of the
northern edge of the saline Intrusion (sometimes referred to as the salt front)
becomes a sensitive Indicator of the balance between the freshwater and seawater
flows (Deck, 1981; Prandle, 1981). Typically the saline Intrusion extends to
just above the Hudson Highlands, around River Nile 55. During major spring
runoff events, the salt front can be found below the George Washington Bridge at
River Nile 12. During times of extreme drought, the salt front has been located
as far north as River M11e 75, near Poughkeepsle.
A. 1-6
-------
A.1.3 Water Quality
A.1.3.1 Overview
Water quality parameters, such as dissolved oxygen, nutrients, turbVdlty,
toxic chemicals, heavy metals and pathogens, affect aquatic resources as well as
a variety of recreational activities 1n the Lower Hudson. The maintenance of
acceptable water quality levels 1s Important to the continued viability of the
Lower Hudson River ecosystem.
j
Although overall water quality 1n the Hudson River has Improved In recent
years as a result of the construction of new sewage treatment facilities and the
upgrading of older facilities, there are still many segments of the Lower Hudson
with water quality problems. For example, a variety of heavy metals, Including
copper, lead, mercury, zinc, chromium and cadmium, have been found 1n the water
column of the heavily Industrialized New York metropolitan region (NYSDEC, 1990).
In addition, concentrations of PCBs, cadmium, TCDD (dloxln) and TCDF (furan) 1n
fish and shellfish may exceed levels considered safe for human consumption and
are of concern, because they can be transferred throughout the aquatic food chain
(NYSDEC, 1990).
The following excerpts illustrate the present general condition of the
water quality within the Lower Hudson (NYSDEC, 1990):
"Water quality 1n the River has Improved steadily 1n recent years,
especially 1n the Albany 'pool' due to the completion of high level
secondary treatment plants serving the Albany County and Rensselaer County
Sewer Districts. Surveys and monitoring in this area have shown
relatively good water quality with respect to conventional pollutants."
"Water quality In the mid-Hudson area is best, and It deteriorates In the
last twenty miles above New York Harbor due to the huge population
concentrations and resultant nonpolnt, storm and wastewater discharges on
both the New York and New Jersey sides of the river. A combined sewer
A. 1-7
-------
overflow study done for the City of Yonkers indicates that water quality
standards violations occur 1n the vicinity of Irvlngton. Construction 1s
currently underway to provide treatment of these discharges."
Specific water quality parameters are briefly discussed below. PCBs in the
Lower Hudson are addressed at A.2 and A.3.
A.1.3.2 Salinity
It Is generally recognized that the Lower Hudson River can be described as
having four salinity zones: (1) limnetic or freshwater zone of less than 0.3
parts per thousand (< 0.3 ppt); (2) oligohaline from 0.3 to 5 ppt; (3) mesohallne
from 5 to 18 ppt and (4) polyhaline from 18-30 ppt (Cooper et a/., 1988).
Although the location of each of these zones varies depending on the magnitude
of tidal and freshwater flow, a tidal freshwater region typically occurs above
River Nile 50-55; an oligohaline zone extends from River Mile 25 to 50; and a
mesohallne to polyhaline zone occurs below River Nile 25. Usually the polyhaline
zone 1s limited to the extreme lower reaches of the Lower Hudson In.the vicinity
of Nanhattan. During periods of pronounced drought, however, the polyhaline zone
may extend beyond Nanhattan (R1st1ch et al., 1977).
A.1.3.3 Temperature
Nean monthly temperatures recorded by the USGS at Green Island indicate
that freshwater entering the estuary ranged from a low of 0*C In January to a
high of 29*C in July for the period 1970-1981 (Noran and Llmburg, 1986; Cooper
et a7., 1988). Lower Hudson water temperature patterns are Influenced by
freshwater discharge and ocean waters. For example, In the summer, ocean water
that enters the Lower Hudson 1s considerably cooler than freshwater. This
situation may result 1n temperature differences of as much as 11*C 1n the upper
and lower reaches of the Lower Hudson (Abood et a?., 1976; Garvey, 1990). The
average annual temperature of water in the Lower Hudson 1s 12.3* C (N0AA, 1982).
A. 1-8
-------
A.1.3.4 Dissolved Oxygen
Typical dissolved oxygen levels throughout the Lower Hudson are between 5
and 14 mg/1, depending on spatial and temporal constraints (Cooper et *1., 1988).
Although there 1s considerable variability with season, the highest leveTiy>ccur
In the late winter to early spring (Moran and Mmburg, 1986). Dissolved oxygen
levels are generally undersaturated throughout much of the Lower Hudson during
the sunmer. Supersaturated conditions may, however, occur 1n some shallow bซys
as a consequence of algae blooms (Cooper et *7., 1988). A survey by Garvey
(1990) during the fall of 1985 Indicates that percent dissolved oxygen saturation
values decline from 105 percent at River Mile 153 (supersaturatlon resulting from
photosynthesis) to 71.4 percent at River Mile 6.5. Increased sewage discharges
from the New York City metropolitan region result 1n Increased biological oxygen
demand (BOD) and reduce the dissolved oxygen levels throughout much of the saline
portion of the Lower Hudson compared to the more freshwater regions (Garvey,
1990).
A.1.3.5 Turbidity and pH
The main Influence on turbidity 1n the Hudson River 1s the presence of
s1It/clay particles that are transported either by marine sources or by
terrestrial runoff. Additional silt/clay particles may be resuspended within the
Lower Hudson by erosion or scouring. Turbidity 1n the Lower Hudson Is generally
higher during periods of greatest discharge (Cooper et ซ7., 1988).
Increased turbidity generally decreases light transparency within the water
column, which, 1n turn, limits the extent of the photic (light) zone.
Investigations 1n water bodies such as the Hudson and Delaware River estuaries
have demonstrated that light transparency Is approximately one meter plus or
minus 0.5 meters, depending upon the season (Cantelmo and Wahtola, 1989).
Results of Interpler light transmission studies along the west side of Manhattan
and the Westslde Highway Project Study (NJMSC, 1984; EEA, 1988) indicate that the
photic zone Is generally confined to the upper meter of the water column. These
A.1-9
-------
results are typical for near shore, relatively turbid estuarlne waters. This
situation contributes to a large vertical attenuation of light and Units the
photic zone to the upper 1-1.5 meters.
In the well-buffered Lower Hudson, very little spatial or temporal change
1n pH values occurs. Although historical records reviewed by Cooper et ซ7.
(1988) Indicate that the pH In the Lower Hudson may vary between 6.4 and 8.2,
most measurements were above 7.0 (Moran and Llmburg, 1986). Recent data
collected by Garvey (1990) throughout the entire Lower Hudson also confirms that
the system 1s well-buffered with slight pH variations of 7.6 to 7.8.
A.1.3.6 Municipal Wastewater Discharges
There has been a steady Increase In municipal wastewater discharges since
1952, but dally BOO has declined as a consequence of the construction of new
sewage treatment facilities (HetUng, 1976; Moran and Llmburg, 1986). Recent BOD
loading data are not available for the entire Lower Hudson. HetUng (1976) and
Koran and Llmburg (1986) have calculated 1975 BOD loadings of 55.5 metric
tons/day and 131.5 metric tons/day for River Miles 14-152 and River Miles 0-14,
respectively.
A.1.3.7 Phosphates and Nitrates
It Is well established that sewage 1s the major source of both phosphates
and nitrates to the Lower Hudson (Deck, 1981; Moran and Llmburg, 1986; Cooper et
<7., 1988; Garvey, 1990). Biological uptake of nutrients, such as nitrates and
phosphates by phytoplankton, has been shown to be Insignificant compared to the
total amount of nutrients available within the Lower Hudson (Deck, 1981; Moran
and Llmburg, 1986).
Concentrations of ortho-phosphate above the salt front are generally less
than 95 |ig/l and range between 190-620 |ig/1 below the salt front (Garvey, 1990).
The large Increase 1n ortho-phosphate below the salt front 1s a result of the
A.1-10
-------
large inputs of sewage, which accounts for 73 percent of the total phosphorus
sources to the Lower Hudson (Deck, 1981). Concentrations of phosphates,
generally the highest during low freshwater flow conditions, typically occur
during the late summer (Moran and Llmburg; Cooper et a/., 1986).
The predominant form of nitrogen 1n the Lower Hudson 1s nitrate (Deck,
1981). The only exception occurs below River Nile 18, where ammonia from sewage
1s the dominant form of nitrogen (Deck and Bopp, 1984). Nitrate levels as
nitrogen are typically the highest (560 pg/1) just south of the salt front and
decline 1n the more saline reaches. The largest source of nitrates to the
freshwater reaches of the Lower Hudson enters the system from the Upper Hudson
(Horan and Llmburg, 1986). The nitrates 1n the more saline portion of the Lower
Hudson are governed generally by ammonia additions from sewage and urban runoff
(Deck, 1981).
A.l.3.8 Classification and Use
The Hudson River, like other surface waters in New York State, 1s
classified according to the Intended "best use". The classification scheme for
the Lower Hudson, illustrated in Plate A.1-2, takes into consideration river
flow, water quality, condition of adjacent shorelines, and historic, present and
future uses. Drinking water 1s classified as A; swimming as either B or SB; fish
propagation and fishing as C; and secondary contact recreation (fishing and
boating) as I.
The Hudson River Is used as a source for public water supplies In sections
of the river classified as A. There are nine Lower Hudson facilities that draw
Hudson River water directly for consumption, Including five communities and four
institutions, camps or schools (Table A.1-2). In addition, a water Intake
located at Chelsea (River Mile 66) north of Beacon 1s used by New York City as
an emergency water supply during severe periods of drought (NUS, 1984).
A.1-11
-------
Hudson River water Is also used for Industrial and commercial purposes such
as cooling, manufacturing processes, fire protection and hydroelectric and
thermal power generation. An inventory of such facilities and plants can be
found 1n reports for the Hudson River-Black River Regulating District (Malcolm
Plrnle, 1984) and for the NYSDOT (December 1984).
A.1.4 Aquatic Resources in the Lower Hudson
A.1.4.1 Conceptual Framework
There have been many attempts to conceptualize the structure and function
of stream^ and/or river ecosystems, including those by Hynes (1970), Cummins
(1974), Whltton (1975), Mclntlre and Colby (1978), Cummins and Klug (1979), Moran
and Umburg (1986) and Gladden et al. (1988). Collectively, these studies have
generally emphasized four major categories of organic resources:
Primary Producers phytoplankton, periphyton and macrophytes;
Detritus particulate organic matter and associated microbial biomass;
Dissolved organic matter; and
Consumers microzooplankton, macrozooplankton, benthlc invertebrates and
fish.
f
Primary producers Include the phytoplankton, periphyton and macrophytes.
Phytoplankton are varied microscopic floating plants that have no power of
locomotion and are spatially dispersed by river currents; periphyton are those
algae attached to various substrates, including rocks (ep1l1thic), silt
(epipelic) or other plant species (epiphytic); and macrophytes are macroscopic
forms of aquatic vegetation generally limited to shallow (<10-15 feet) water.
A.1-12
-------
Detritus 1s any non-living particulate organic matter derived fron the
production of living populations of plants and animals. The primary source of
detritus may vary, depending on the nature of the aquatic habitat. For example*
In large bodies of water, most of the detritus 1s derived from aquatic plants;
1n small forest-covered streams, the terrestrial contribution from the
surrounding watershed 1s dominant.
Dissolved organic matter 1s excreted by consumers, released from cells
during feeding, released due to microbial transformations or released by primary
producers. It includes amino acids, glucose, dissolved organic phosphorus and
dissolved organic nitrogen.
Consumers include zooplankton, benthlc invertebrates and fish. Additional-
ly, Invertebrates 1n the Hudson River can be categorized Into major groups,
depending on habitat preference and size. For example, the relatively small
zooplankton, i.e., microzooplankton, are 1n the 50-500 pm size range; the larger
zooplankton, I.e., macrozooplankton, are generally 1n the 500-2,000 |im size
category. The benthos or bottom dwelling Invertebrates may be further subdivided
Into the eplfauna, those organisms primarily Inhabiting the surflclal sediments,
and the Infauna, living predominantly within the sediments. F1sh populations In
the Lower Hudson River can be broadly characterized as freshwater or euryhallne
year-round residents and those that utilize the Hudson during spawning or feeding
migrations.
The nutritional value of the four preceding categories of organic resources
Is determined by a number of factors, Including the C/N (carbon to nitrogen)
ratio, protein content and percentage of refractory (unavailable) nitrogen.
Russell-Hunter (1970) has suggested that the lower limit of nutritional
requirements for organic resource consumers 1s generally 16 percent protein (dry
weight) and a C/N ratio of <17. In addition, Cummins (1979) notes that the C/N
ratio of organic resources varies considerably and some sources with a C/N ratio
<17 may still not be considered "quality" resources. For example, fine detritus
(0.05 - 1.00 mm) may have a significant portion of refractory nitrogen, which is
A.1-13
-------
resistant to microbial transformations and/or direct consumer utilization
(Cummins and Klug 1979). The quality or nutritional value of food 1s especially
Important In temperate 1ot1c freshwater systems, which are more likely to
encounter seasonally high C/N ratios during peak spring detrltal Inputs from the
surrounding watershed.
A number of studies have concluded that various autochthonous, primary
production Inputs (phytoplankton, perlphyton and macrophytes) make Important
contributions to aquatic food webs (see reviews by Wetzel, 1975 and Mann, 1975).
In some relatively fast flowing streams (>2.0 m/sec), perlphyton (algae attached
to various substrates) may be the major contributor to primary production (Mann
1975). As stream velocity decreases, however, phytoplankton and macrophyte
production may dominate autotrophic processes (Taylor, 1971 and Wetzel, 1975).
Inputs provided by organic carbon from the surrounding terrestrial
watershed, Including particulate and dissolved organic matter (allochthonous
Inputs) are also Important food chain constituents for maintenance of trophic
structure (Hynes, 1970). Organic carbon sources within upstream reaches of
various river systems have been shown to make significant contributions to the
lower reaches. For example, the largest source of allochthonous carbon to the
Lower Hudson (66,024 metric tons of carbon/yr) originates from the Upper Hudson
(Gladden et a7., 1988). The following table provides some estimates of the
organic carbon Inputs to the Lower Hudson (Noran and Llmburg, 1986 and Gladden
et al., 1988).
Source
Organic Carbon Inputs
Metric Tons Carbon/yr
Percent
Phytoplankton
Macrophytes
Upper Watershed
Lower Watershed
Sewage
Marine
36,364
5,364
66,024
43,254
57,649
36,898
14.8
2.2
26.8
17.6
23.5
15.0
245,553
100.0
A.1-14
-------
The amount of allochthonous Inputs from upper watershed areas to the lower
reaches of many temperate rivers, Including the Hudson, varies seasonally with
discharge. The greatest discharge 1n the Upper Hudson occurs during the spring
(Cooper et a1.t 1988) and corresponds to the highest seasonal input of
allochthonous carbon to the Lower Hudson (Gladden, per. comm.).
Although the relative Importance of autochthonous versus allochthonous
carbon sources has been debated 1n the scientific literature (Stephens 1967;
Gladden et a/., 1988), this relationship may ultimately depend on the specific
nature of the system being Investigated. For example, F1shฃr and Likens (1973)
showed that 99 percent of the energy Input to a first order stream in New
Hampshire Is from the allochthonous Inputs. In a classic study of energy flow
in Silver Springs, Florida, Odum (1957) found that autochthonous production in
the form of freshwater eel grass was predominant. Gladden et al. (1988) and
Moran and Llmburg (1986) calculated that approximately 65 to 83^ percent of the
total organic carbon In the Lower Hudson came, respectively, from allochthonous
sources 1n the upper watershed and from sewage effluents or from marine sources
1n the lower watershed. Nevertheless, Gladden et al. (1988) speculated that much
of the allochthonous carbon may be refractory and of limited or reduced
nutritional value to the consumers, compared to autochthonous phytoplankton
production.
Studies of small watersheds and larger rivers (Fisher and Likens, 1973;
Fisher, 1977; Sedell et al., 1973; Schaffer 1978) indicate that more organic
carbon may be in a dissolved rather than particulate form. In a review of the
Lower Hudson River ecosystem, Gladden et al. (1988) remark that dissolved forms
of carbon must first be transformed Into microbial biomass before they are
Incorporated by various consumers, Including zooplankton and benthlc Inverte-
brates. Since there Is an additional transformation step, less net energy may
be transferred to the consumers by a dissolved organic carbon route than by a
particulate route.
A.1-15
-------
In most systems Investigated to date, trophic pathways and the efficiency
of organic carbon utilization are not clearly understood, because of the
complexity of Interactions, resource partitioning and changes 1n food preferences
by many organisms during their life cycle. Particulate organic compounds
(allochthonous and autochthonous) grazed by organisms In the water column or
Incorporated Into the sediments may be utilized by zooplankton and benthlc food
resources (Hynes 1970; Gladden et a/., 1988). In many aquatic systems temporal
and spatial dispersal patterns, generation times and population fluctuations of
these lower trophic groups may, thus, exert a pronounced Influence on the
foraging success of juveniles and adult fish populations. As such, the
invertebrates are Important trophic links 1n the food web between organic carbon
sources and fish populations.
A.l.4.2 Physical Constraints
Many studies of lotlc systems have established that a number of physi-
cal/chemical parameters (dissolved oxygen, temperature, nature of the substrate,
etc.) depend primarily on various hydrologlcal features of the river basin.
Energy transfers between trophic levels and blotlc spatial and temporal patterns
are Inextricably linked to a variety of hydrodynamlc factors responsible for
shaping and maintaining the stream channel. Of all the hydrodynamlc factors,
stream current (velocity) Is the most Important physical factor regulating river
biota. Many freshwater river studies have acknowledged that longitudinal
distribution as well as species composition of primary producers and consumers
are greatly Influenced by velocity (for review see Whltton, 1975).
Velocity 1s determined by a number of elements, Including the size, shape
and roughness of the channel, load of suspended sediments and gradient. The
gradient In the Upper Hudson from Fort Edward to the Federal Dam 1s approximately
1 m/m1le (Sanders, 1982), which produces a slow to moderate velocity from 0.3 to
1.3 m/s (Simpson, 1974). Although the gradient 1n the Lower Hudson 1s 0.01
m/mlle (Helslnger and Fledman, 1982), the velocity 1s driven by tidal Influences.
A.1-16
-------
The maximum velocity 1n the Lower Hudson 1s estimated to range from 0.5 to 1.0
m/s (Garvey, per. comm.).
In large measure, stream velocity determines what type of bottom substrate
may be present. For example, erosion of sand and gravel beds generally occurs
at velocities >1.7 m/s, deposition of sand occurs at velocities of 0.3 to 1.2 m/s
and s111 particles normally settle out at velocities <0.3 m/s (Terrell and
Perfettl, 1989). Both the physical nature (average grain size, silt/clay
fraction, range of particle sizes, etc.) and chemical nature (rate of exchange
of compounds and gases across the sediment/water interface, vertical gradients
of Eh, interstitial oxygen, etc.) of the stream bed are major constituents
controlling the distribution of many benthlc organisms.
A.1.4.3 Trophic Components 1n the Lower Hudson
Primary Producers
It 1s well documented that phytoplankton are the dominant primary producers
within the Lower Hudson (Noran and Llmburg 1986; Gladden et a7., 1988). Although
other groups of primary producers, such as the macrophytes, are believed to
represent additional sources 1n estuarine ecosystems (Gladden et al.t 1988),
productivity values are not generally available for the Lower Hudson.
Estimates of phytoplankton gross productivity range between 100-250 g
carbon/nf/year from River Mile 0 to River Mile 76 (Slrois and Fredrick, 1978).
The highest productivity occurred during June and July and decreased dramatically
in August (Slrois and Fredrick, 1978; Gladden et ซ7., 1988), as explained later.
Analysis of spatial trends in the data of Slrois and Fredrick (1978) Indicated
that gross primary production during May through October, 1972 was the highest
within the Tappan Zee and Croton-Haverstraw regions (approximately, River Miles
28-38) and declined farther upriver and farther downriver (Gladden et Ml., 1988).
A.1-17
-------
Gladden et al. (1988) point out that the reasons for the rather large
spatial and temporal variations 1n phytoplankton productivity may be attributed
to the variations in river physical characteristics, flow patterns and light
Intensity. Hudson River phytoplankton are light rather than nutrient limited
(Heffner, 1973; O'Reilly et al.t 1976; Storm and Heffner, 1976). In addition,
phytoplankton may tend to Increase and concentrate in the low-flushing, shallow
regions of the Lower Hudson, such as the vicinity of River Miles 28-38.
Welnsteln (1977) and others (Fredrick et al1976; Storm and Heffner 1976;
McFadden et al.t 1978; Moran and Llmburg, 1986) have indicated that phytoplankton
species are distributed In the Lower Hudson along spatial and temporal gradients.
For example, more marine phytoplankton dominate the lower reaches (River Mile
<25) of the Hudson River, while more brackish water and freshwater species
dominate the middle (River Mile 25-50) to upper reaches (River Mile >50),
respectively. During the wanner sunnier months, phytoplankton species are
dominated by the Chlorophyta (green algae) and Cyanophyta (blue-green algae),
whereas from the late fall through spring, Bad liar iophyceae (diatoms) are often
the most numerous. Typical dominant phytoplankton found throughout the Lower
Hudson (Moran and Llmburg, 1986) are listed below.
Chlorophyta (Green algae)
Chlorophyceae
Pediastrum
Scenedesaus
Anklstrodesaus
Chrysophyta
BadliarIophyceae (01atoms)
Asterlonella
Meloslra
Cyclot el la
Chrysophyceae (golden or yellow-brown algae)
Chrysococcus
Cyanophyta (Blue-green algae)
Anabaena
Anacystls
Pyrrhophyta (Dinoflagellates)
Ceratiun
Porocentrua
A.1-18
-------
In addition to the contribution of phytoplankton to primary production,
perlphyton and aquatic macrophytes may also represent a primary food source for
a variety of consumers. There have been no studies of perlphyton 1n the Lower
Hudson, but a number of studies have described Hudson River macrophytes^
V
The following list of common macrophyte genera In the Lower Hudson has been
compiled from a variety of sources (Klvlat, 1973; McFadden et al., 1978; Noran
and Llmburg, 1986).
Genera
Common Name
Anacharls (Elodea)
Waterweed
Carex
Sedge
Cyperus
Sedge
Eleocharls
Spike Rush
Neteranthera
Water Star-Grass
Lythrum
Purple Loosestrife
Nyrlophyllun
Milfoil
Najas
Nald
Nuphar
Spatterdock
Peltandra
Arrow-arum
Phragmltes
Common Reed
Pontedaria
Pickerelweed
Potanogeton
Pondweed
Saglttaria
Arrowhead
Scripus
Bulrush
Sparganlua
Bur Reed
Spartlna
Cord Grass
Trapa*
Water-chestnut
Typha
Cattail
ValUsneria
Water Celery
Zizanla
Wild Rice
* Introduced species of European or Asian origin.
The salinity regime mediates spatial patterns of macrophytes. For example,
salt-tolerant plants such as Spartlna are common in the lower and middle reaches
of the Hudson River, but are replaced by Peltandra, Nuphar, Typha, ValUsneria,
Nyrlophyllum and Potanogeton 1n the freshwater marshes 1n the upper reaches
(McFadden et <7., 1978; Noran and Llmburg, 1986). At times, some Introduced
A.1-19
-------
species such as water-chestnut and purple loosestrife form extensive monocultures
within the Lower Hudson and displace a variety of more desirable native plants
such as cattails, sedges and bulrushes (Moran and Llmburg, 1986; Stalter, per.,
comm., 1991).
Since productivity estimates for Lower Hudson River macrophytes are not
available, 1t Is not possible to ascertain the exact contribution of macrophytes
to the total organic carbon sources. Moran and Llmburg (1986) estimate that the
7.7 square miles of marshlands 1n the Lower Hudson should produce about 5,364
metric tons of organic carbon annually. Given these estimates, the net
contribution of macrophytes is less than three percent of total organic carbon
sources.
Consumers (Invertebrates and Fish)
Invertebrates
Invertebrate consumers In the Hudson River estuary are a heterogeneous
group of organisms that link estuarlne organic carbon sources and fish
populations. Patterns of Invertebrate temporal and spatial dispersion,
generation times and population fluctuations may exert a pronounced Influence on
foraging success of larvae, juveniles and adult fish populations in the Hudson
River estuary (Gladden et a/., 1988).
Although the two general categories of Invertebrates (zooplankton and
benthos) can be further subdivided depending on habitat preference and size,
I.e., mlcrozooplankton, macrozooplankton, benthic ep1 fauna and benthic infauna,
the following discussion focuses on dominant members of the two general
categories 1n order to develop an overview of how invertebrates are spatially
distributed throughout the Lower Hudson.
A.1-20
-------
There have been numerous studies and reviews (Rlstlch et ซ71977;
Welnsteln, 1977; Gladden et al., 1988; Moran and Llmburg, 1986; ) of Inverte-
brates throughout the Lower Hudson. Mstecf below are some of the dominant taxi
of benthlc (b) and zooplankton (z) Invertebrates found 1n the Lower Hudson.
*
Aquatic Insects
Dlptera [Chlronomids (b)]
Rot1fera
Plolma [Keratella (z), Brachionus (z)]
Mollusks
Gastropoda [Valvata (b/z), Hydrobla (b), Nassirius (b)]
Pelecypoda [Congerla (b), Mya (b), Macoaa (b)jJ
Crustaceans
Amphlpoda
Clrrlpedla
CIadocera
Copepoda
Decapoda
Isopoda
Mysldacea
Annelids
OUgochaeta [Lianodrilus (b), Na1s (b)]
Polychaetes [Scolecolepldes (b), Nereis (b), Boccardia (b)]
Gamarus (b), Leptochelrus (b)]
Balanus (b/z)]
Bosmlna (z), Dlaphanosoaa (z), Holna (z)]
Acartla (z), Euryteaora (z), Teaora (z)j
Crangon (b), Palaeaonetes (b)]
Cyathura (b), Edotea
-------
Unlike the zooplankton populations, benthlc Invertebrates 1n the Hudson
River estuary exhibit much smaller population variations and have longer
generation times than the zooplankton communities (Gladden et al.t 1988).
Benthlc Invertebrates also have a more constant food supply, because most species
consume a variety of food particle sizes. In addition, organic carbon concen-
trations are temporally more stable in sediments (Cantelmo, 1978) than in the
water column (Schaffer, 1978).
Distinct spatial population patterns are, however, observed. The lower
reaches (River Mile <25) support a typical marine assemblage of benthlc
Invertebrates, Including marine ollgochaetes, polychaetes and Crustacea. The
middle reaches (River Nile 25-50) have a mixture of freshwater and marine forms.
The upper reaches (River Mile >50) are dominated typically by freshwater
Insects, snails, ollgochaetes and clams. Greater than 75 percent of the total
benthlc populations 1n many regions may consist of mysids and amphipods. '
Haloph111c Neoaysls aaerlcana usually replaces more freshwater Gamarus species
1n the lower reaches of the Hudson River estuary (Gladden et al., 1988; NYU
Medical Center, 1978; PASNY, 1986).
Simpson et al. (1985) found that the freshwater macrobenthlc fauna of the
main channel of the Hudson River between Glenmont (River Mile 141.1) and New
Hamburg (River Mile 67.4) consisted primarily of ollgochaete worms, midge larvae,
crustaceans, bivalves and gastropods. The most abundant species was the
ollgochaete worm, Llanodrllus hoffmelstert, representing approximately 54-79
percent of the total Individuals collected at all 16 stations sampled. Simpson
et al. (1985) also found that the freshwater macrofauna In the Lower Hudson were
"...more or less typical of a large lowland river" and consisted primarily of
ollgochaetes, midge larvae, crustaceans and bivalves. The most abundant taxa
were the ollgochaetes, which represented approximately 54-79 percent of the total
macrofauna.
A.1-22
-------
Benthlc Invertebrate populations 1n the middle reaches (River Mile 25-50)
of the Lower Hudson River estuary are numerically dominated by ollgochaetes, "
polychaetes, molluscs, and harpactlcold copepods. These taxa account for ซore
than 70 percent of the benthos 1n many regions (Texas Instruments, 1976).
The benthlc invertebrates 1n the lower reaches (< River Nile 25) of the
Hudson are numerically dominated by ollgochaetes and polychaetes. Although EEA
(1988) Identified 80 benthlc taxa during an aquatic study at the Hudson River
Center Site (River Mile 4.0) from April 1986 to March 1988, only two taxa
accounted for 73 percent of the total benthlc macrofauna. These two taxa
(splonld polychaetes and ollgochaete worms) were also found to be the most
dominant organisms 1n other studies. For example, in ecological surveys
conducted at Liberty State Park (River Mile 0), Harborside (River Mile 1.0) and
in the vicinity of the West Side Highway study area (River Miles 1.8, 4.0 and
9.0), the same two taxa were found to dominate the benthlc macrofauna community
(Texas Instruments, 1976; LMS, 1980; Harborslde, 1987; EEA, 1988).
F1sh
Numerous studies and reports have been prepared that document the fisheries
of the Hudson River estuary. Since the historical fish surveys conducted In 1934
and 1936 (Greeley, 1935, 1937) from Albany to the Tappan Zee Bridge, surveys
conducted over the entire length of the estuary have largely concentrated on
concerns over power plant Impacts. McFadden et al. (1978) summarized those
studies conducted from 1936-1975 and Barnthouse et *7. (1988) focused mainly on
surveys conducted from 1976-1980. Beebe and Savldge (1988) updated the 1978
summary and Included data from historical and recent fisheries reports In order
to compile an extensive list of fish fauna within various regions of the Lower
Hudson River. Smith and Lake (1990) have recently documented all known Hudson
River fish within the entire Hudson River basin and Included notes on the
distribution of each species and probable geographic origins.
A.1-23
-------
Major recent reviews by Moran and Llmburg (1986) and Gladden et <7. (1988)
Included fisheries components as part of their evaluation of the Lower Hudson
River ecosystem. The following description of the overall fisheries of the Lower
Hudson primarily utilizes Information synthesized by Horan and Llmburg (1986),
Gladden et *7. (1988) and Beebe and Savldge (1988).
The Lower Hudson River supports some 140 species of resident and migrant
fish (Beebe and Savldge, 1988). Some 66 native, freshwater residents (Gilbert,
1980) and a variety of Introduced freshwater species occur 1n the freshwater
tidal areas with less than three parts per thousand salinity (< 0.3 ppt). The
common species are listed below.
B1ueglll
Brown bullhead
Common carp
Eastern silvery minnow
Emerald shiner
Golden shiner
Goldfish
Largemouth bass
Pumpklnseed
Redbreast sunflsh
Spottail shiner
Tesselated darter
White catfish
White sucker
Yellow perch
Those resident species consistently observed within ollgohaline regions
(0.3-5 ppt), mesohallne regions (5-18 ppt) or, at times, throughout the entire
Lower Hudson Include:
Atlantic sllverslde
Foursplne stickleback
Hogchoker
Longhom sculpln
Mumraichog
Shortnose sturgeon
Tidewater sllverslde
White perch
Many species living throughout the Lower Hudson are considered euryhallne
and are adapted to the wide variations In salinity described at A.1.3.2.
A. 1-24
-------
The resident species occur mainly 1n the shore and the channel bottom zones
and typically exploit the more stable benthlc habitat (Gladden et a1., 1988). 'Hie
adult freshwater residents, such as common carp, golden shiner, yellow perch or
pumpklnseed, and euryha11ne/mar1ne residents, such as mummlchog, hogchoker and
longhorn sculpln, are broadly characterized as fish that feed at the lower end
of the trophic level, I.e. omnivorous. Regions of peak abundance in the Hudson
estuary shift from year to year (Gladden et a1.t 1988).
The Lower Hudson supports abundant euryhallne and marine populations as
suggested by Beebe and Savldge (1988):
"...extensive marine and brackish-water fish faunas found 1n the Hudson
are a result of the mid-Atlantic location of the estuary and Its proximity
to the Gulf Stream. Hany marine species ride the Gulf Stream Into coastal
nurseries, Including the Hudson river estuary. In addition, several
tropical or pelagic species stray Into the estuary during the summer
(Smith, 1985), and deep-water species occasionally enter the estuary
during winter. Individuals of these species have been collected
frequently 1n the lower, more saline, portion of the estuary
Euryhallne species may be found throughout the estuary, often In large
numbers 1n their preferred habitats.
Approximately eight fish species utilize the Lower Hudson River estuary as
a migratory pathway for spawning activities (dladromous species), Including:
Alewlfe
American eel
American shad
Atlantic sturgeon
Atlantic tomcod
Blueback herring
Rainbow smelt
Striped bass
Of these eight adult dladromous species, seven complete a part of their
life cycle 1n the ocean and spawn 1n freshwater (anadromous). The American eel,
the only catadromous species 1n the Lower Hudson, spawns 1n the Sargasso Sea;
juveniles (elvers) enter the Lower Hudson and move upstream Into the ollgohaline
to freshwater reaches (Boyce Thompson Institute, 1977).
A.1-25
-------
Some Migratory fish species that are not dladromous nay periodically enter
the Lower Hudson during seasonal feeding cycles In order to consume zooplankton
or juvenile fish. Others are considered more permanent marine species. The
seasonal and permanent marine species (Smith, 1990) Include:
Atlantic menhaden Fourbeard rockllng
Bay anchovy Northern pipefish
Blueflsh Weakfish
Gladden et al. (1988) found that most of the adult migrant species can be
found in all habitat types but only for short periods of time. The anadromous
species as well as many seasonal marine species are primarily carnivorous and
feed on zooplankton and other fish species. The timing of the runs of various
migrants, although flow and temperature dependent, Is closely linked with the
abundance of food for the newly hatched larval forms and juveniles of the species
(Cushlng, 1975).
The numbers of species and their respective abundance vary seasonally. The
numbers tend to Increase 1n the spring through mid-summer when they reach a
maximum. Abundance then decreases through the fall and reaches a minimum 1n the
winter. Gladden et al. (1988) and others (Barnthouse et al., 1988) have examined
various regions of the Hudson River estuary and attempted to correlate food
preferences and depth (shore versus channel) 1n order to explain spatial
distribution. Trophic, spatial and temporal partitioning by resident and
migrant fish populations may reduce competition (Gladden et al., 1988) and
Increases the likelihood of maintaining a diverse fisheries resource in the Lower
fi
Hudson.
The following 11st of common and scientific names of the dominant or
abundant species of migratory and resident species has been compiled from Moran
and ^Imburg (1986); Gladden et al. (1988); and Beebe and Savldge (1988).
t
A.1-26
-------
Alewlfe
American eel
American shad
Atlantic menhaden
Atlantic sllverslde
Atlantic sturgeon
Atlantic tomcod
Banded kill 1 fish
Bay anchovy
Black crapple
Blueback herring
Blueflsh
Blueglll
Brown bullhead
Common carp
Eastern silvery minnow
Emerald shiner
Fourbeard rockllng
Fourspine stickleback
Golden shiner
Goldfish
Hogchoker
Largemouth bass
Longhorn sculpln
Nummlchog
Northern pipefish
Pumpkinseed
Rainbow smelt
Redbreast sunfish
Shortnose sturgeon
Spottall shiner
Striped bass
Summer flounder
Tesselated darter
Tidewater sllverslde
Weakfish
White catfish
White perch
White sucker
Winter flounder
Yellow perch
Alosa pseudoharengus
Angullla rostrata
Alosa sapldlsslaa
Brevoorlta tyrannus
Menidia aenldla
Aclpenser oxyrhynchus
Mlcrogadus tomcod
Fundulus dlaphanus
Anchova aHchUll
Poaoxls nlgroaaculatus
Alosa aestivalis
Pomatoaus saltatrlx
Lepoais aacrochlrus c
Ictalurus nebulosus
Cyprinus carpio
Hybognathus reglus
Notropls cornutus
Enchelyopus clabrlus
Apeltes quadracus
Noteaigonus crysoleucas
Carassius auratus
Trlnectes aaculatus
Mlcropterus salaoides
Myoxocephalus octodeceaspfnosus
Fundulus heteroclltus
Syngnathus fuscus
Lepomis gibbosus
Osmerus aordax
Lepoais auritus
Aclpenser brevirostrua
Notropls hudsonius
Mo rone saxatllls
Parallchthys dentatus
Etheostoaa ol lasted 1
Menidia peninsulae
Cynoscion regal Is
Ictalurus catus
Mo rone aaerfcana
Catostoaus coaaersonl
Pseudopleuronectes aaerlcanus
Perca flavescens
A.1-27
-------
Of the 140 species of fish In the Hudson (Beebe and Savldge, 1988), the
striped bass has received a disproportionate amount of attention, because of Its
commercial and recreational value. In addition, most power plant Impact studies
have focused largely on striped bass populations and have seldom Included an
ecosystem approach to the Lower Hudson. Although the striped bass Is an
Important species, the Lower Hudson 1s more than just a single species system.
It has been shown to contain one of the most diverse fisheries found throughout
Atlantic coastal systems (Beebe and Savldge, 1988).
Synthesis
Differences In the life history strategies and characteristics of
zooplankton and benthlc Invertebrates may exert a major Influence on the fish
populations. The exploitation of Invertebrates by Hudson River fish populations
Is well documented and leads to a pattern of trophic partitioning among the '
dominant fish species. For example, Gladden et al. (1988) has shown that
migratory adult fish species predominantly feed on pelagic zooplankton and other
fish, whereas, the resident adult fish species predominantly exploit the more
stable benthlc Invertebrate populations. Since the shallow shore and shoal zone
habitats generally support the most extensive resident fish populations 1n the
Hudson River (Beebe and Savldge, 1988), the benthos 1n these regions Is closely
linked to resident fish populations with low risk/persistence strategies
(Gladden, et a/., 1988).
In addition, Thurow (1974) has shown that the year class strength for some
demersal fish may be linked to the availability of food and the foraging success
of larval and juvenile fish. It has been proposed that the Increased food
demands by larval fish must coincide with Increased populations of Invertebrates
to Insure a strong year class (Cushlng, 1975). Thus, the variability 1n fish
numerical abundance and spawning success may be, 1n part, explained In terms of
spatial and temporal variability of the Invertebrate trophic base. Invertebrate
consumer populations, 1n turn, are linked to the availability of energy supplied
by other organic resources, Including the primary producers, detritus and
dissolved organic matter.
A. 1-28
-------
A.2 Sources of PCB Contamination
A.2.1 Description of PCBs
Polychlormated b1phenyls (PCBs) are a class of chlorinated, aromatle
hydrocarbons. Each PCB consists of two connected rings of six carbons each (a
biphenyl) to which one or more chlorine atoms are attached at any of 10 available
sites (Figure A.2-1). Positions on the blphenyl molecule not filled w1tt
chlorine atoms have hydrogen atoms 1n their place. Mono- through deca-
chlorlnated biphenyls make up a homologous series In which each successive
homologue group contains compounds with one more chlorine atom than the preceding
group. Within each homologue group, the PCBs containing the same number ol
chlorine atoms but differing 1n their structural arrangement on the phenyl rings,
are referred to as Isomers. For example, 2-chlorob1 phenyl and 4-chlorob1 phenyl,
both monochloroblphenyls, are Isomers of each other, each containing one chlorine
atom per blphenyl molecule, yet differing In the position of chlorine substitu-
tion. Figure A.2-1 also lists the ten different PCB homologue groups and show!
the number of possible Isomers within each group. The Isomers in all of the
homologue groups are generlcally referred to as congeners. There are a total ol
209 theoretically possible PCB congeners.
PCBs are produced commercially by the chlorlnatlon of blphenyl using ferric
chloride or iodine as a catalyst. This produces a mixture of congeners, whlct
1s usually distilled to produce somewhat simpler mixtures with des1re<
properties. The distilled mixtures have been marketed under various trade name:
Including Aroclor (Monsanto, US), Clophen (Bayer, West Germany), Kanechloi
(Kanegafuchi, Japan), Phenochlor (Caffaro, Italy), Pyralene (Prodelec, France]
and Sovol (U.S.S.R.). In the US, each Aroclor product marketed was given a foui
digit numerical nomenclature; the first two digits (12) Indicate the number o1
carbon atoms per molecule and the last two digits represent the approximate
percent chlorine by weight 1n the mixture. (Aroclor 1242, for example, Is a US-
produced polychlorinated blphenyl mixture, containing approximately 42 percenl
chlorine by weight.) An exception to the nomenclature Is Aroclor 1016. Table
A.2-1 lists the weight percentages of different congener groups In various
A.2-1
-------
Aroclor mixtures.
In the environment, PCBs have a high affinity for the organic carbon
fraction of soil, sediments and suspended matter and a high tendency to
accumulate 1n biota. PCBs (especially the lesser chlorinated congeners) also
have a moderate tendency to volatilize from water Into the atmosphere. PCB
mixtures also have characteristics Imparted by their Individual PCB congeners,
each possessing specific, unique physical and chemical properties and each
behaving somewhat differently 1n the environment. The multiple congener
composition of Aroclors 1s an Important consideration 1n Interpreting environmen-
tal data.
A.2.2 Lower Hudson PCB Loadings
Total loading of PCBs to the Lower Hudson has been historically dominated
by inputs from the Upper Hudson. The TAMS/Grad1ent analysis of PCB loading from
1977 to 1989 (see B.4) Indicates that approximately 15,000 kg of PCBs were
released to the Lower Hudson over the Federal Dam at Troy during this period.
This amount represents only a portion of the total historic loading of 178,000
kg (Thomann et a/., 1989) from the Upper Hudson to the Lower Hudson (see A.4).
Based on the sediment measurements obtained by Bopp (1979), Bopp et al. (1983)
and Bopp and Simpson (1989), the PCB loading from the Upper Hudson has been
continuous from the early 1950s at levels comparable to and often greater than
those measured over the period 1977 to 1989. As confirmed by sediment core PCB
measurements, estimated maximum PCB loading occurred around 1973, the year the
Fort Edward Dam was removed. PCB loads for that year were estimated to be on the
order of 5,000 kg/yr or five times the annual loading 1n 1980.
A closer review of the loadings for the period 1977 to 1989 suggest a
diminution of PCB releases. For example, 410 kg of PCBs have been estimated to
be released over the Federal Dam at Troy 1n 1986 compared to 210 kg 1n 1989 (see
B.4). Current estimates Indicate that a substantial portion of the approximately
85,000 kg of PCBs estimated to remain 1n the sediments of the Lower Hudson can
be attributed to the Upper Hudson loadings (Bopp and Simpson, 1989).
A.2-2
-------
Although historically dominated by the Upper Hudson PC8 loadings, total
loading of PCBs to the Lower Hudson has also been a function of other sources.
These sources have been Identified and estimated by various Investigators over
the last 15 years and Include: sewage effluent discharges, tributary Inputs below
the Federal Dam, combined sewer/storm water outfalls, storn water outtalls,
atmospheric deposition and landfill leachates. These additional PCB sources are
Important to consider, since some may contribute Inputs of similar magnitude to
those sources currently originating uprlver (Bopp and Simpson, 1984).
%
Precise data describing the exact quantity of PCB loadings from each of the
preceding additional sources are generally not available. Most parameters, such
as flow rate and PCB concentration, must be estimated 1n order to predict the
total PCB loadings. The lack of data 1s, In part, a result of the dispersed
nature of such sources. Table A.2-2 summarizes the range of estimates for these
loads, based upon currently available Information.
A.2.3 Sewage Effluent Discharges
Extensive data has been collected on sewage flow to the river by a number
of agencies and at least two sets of measurements exist on PCB levels 1n New York
City (NYC) metropolitan area effluent (MacLeod et al.t 1981 and Hydroqual, 1991
for NYCDEP).
The MacLeod data set consists of five separate measurements of flow-
weighted composite samples of sewage effluent from three major treatment plants
and several untreated outfalls. The measured total PCB levels averaged 0.3 and
0.97 i>g/l for treated and raw sewage, respectively. Data collected by NYCDEP In
1989 from fourteen sewage treatment plants showed all non-detect levels for PCBs
1n both sewage Influent and effluent. This analysis, on an Aroclor basis, had a
detection limit of 0.33 |tg/l. Based on the Hydroqual (1991) estimate of NYC
metropolitan area sewage flow for 1989 of 1,750 mgd (76.6 m'/s) and the MacLeod
et ป1. (1981) estimates, the 1989 PCB loading would be 4.6 lb/day (2.1 kg/day).
In addition, Thomann et a7. (1989) estimate a somewhat lower sewage effluent
concentration than either Hydroqual (1991) or MacLeod et a7. (1981) and report
A.2-3
-------
a PCB loading for 1980 conditions of 3 lb/day (1.4 kg/day). All of the above
estimates would collectively yield a range of PCB loadings of approximately 3
lb/day (1.4 kg/day) to 4.6 lb/day (2.1 kg/day) for sewage effluent discharges.
A.2.4 Tributary Contributions
Estimates of PCB loadings from tributaries to the Lower Hudson can all be
characterized as poor. Although flow and suspended matter measurements exist for
most major tributaries, there are essentially no measurements of PCB concentra-
tions 1n the tributary flow. Tributary PCB loadings to the Lower Hudson were
estimated by Mueller et al. (1982) and Thomann et <7. (1989), based on literature
data and USGS flow and suspended matter measurements. PCB loadings for the Lower
Hudson 1n 1980 estimated by Thomann et al. (1989) were 2.3 lb/day (1 kg/day),
using a mean tributary PCB concentration of 0.05 pg/1. Based on sediment data
collected for the Passaic, Raritan, Hackensack, Elizabeth and Rahway Rivers,
Mueller et al. (1982) estimated that tributary concentrations were an order of
magnitude lower. These estimates would collectively yield a range of PCB
loadings of approximately 0.2 lb/day (1 kg/day) to 2.3 lb/day (1 kg/day) for the
Lower Hudson tributaries.
A.2.5 Combined Sewer/Storm Hater and Storm Hater Outfalls
Combined sewer-storm water drainage systems 1n the NYC metropolitan area
have long been a source of pollutants to the Lower Hudson. Overflow occurs after
rainfall events and results 1n the release of diluted, untreated sewage directly
to the river. In addition, effluent from storm water collection systems,
draining residential and Industrial areas, also reaches the Lower Hudson
untreated. Estimates of flow via these pathways are based on modeling efforts
with relatively little field data. Mueller et al. (1982) and Thomann et al.
(1989) estimate respectively that storm water runoff and combined sewer outfalls
contribute about 2 lb/day (1 kg/day) to 3 lb/day (1.4 kg/day).
A.2-4
-------
A.2.6 Atmospheric Deposition
Atmospheric loading to the Lower Hudson can result from direct deposition
of particle-bound PCBs, precipitation and gas exchange with the overlying air.
Because the Lower Hudson has a substantial concentration of PCBs In thi water
column relative to the overlying air, gas exchange with the atmosphere results
1n a net loss of PCBs from the river (Bopp, 1983). Particle deposition and
precipitation both can result 1n net PCB transfer to the river 1n spite of the
gas exchange flux. These processes are considered here to the extent that they
act on the river surface Itself. PCB loads resulting from uarticle deposition
and precipitation 1n the remainder of the Lower Hudson basin are Included In the
estimates of tributary Input and storm water runoff.
Like most of the other sources of PCBs, there are few useful measurements
of PCB deposition or precipitation levels. Mueller et al. (1982) estimate tjte
total Input to the Lower Hudson to be 0.2 to 2 lb/day (0.09 to 0.9 kg/day), based
on a study of atmospheric PCB loadings In the Great Lakes by Galloway et al.
(1980). Since the Mueller et a7. (1982) analysis covers portions of the New York
Bight and Long Island Sound as well as the Lower Hudson, those estimates should
probably be reduced by a factor of two to four 1n order to adjust for surface
area differences. Thomann et a7. (1989) also estimated the atmospheric loading
to the Lower Hudson and arrived at a figure of 0.5 lb/day (0.2 kg/day), based on
a review of Mueller et <7. (1982) and other data 1n the literature.
A.2.7 Landfill Leachites
Estimates of leachate loadings of PCBs to the Lower Hudson River from
bordering landfills, examined by Mueller et <7. (1982), are based on a minimal
number of measurements and on a simple model of leachate transport. A total of
5,650 acres (23 km*) of landfill area were Included In the estimate of landfill
leachate generation. A range of 0 to 1.3 lb/day (0 to 0.57 kg/day) was estimated
as the PCB load from these landfills to the Lower Hudson, New York Bight and Long
A.2-5
-------
Island Sound (Mueller et a1., 1982). Although the load to the Lower Hudson
Itself was not developed, It 1s expected that 1t would represent about half of
the total landfill leachate load.
A.2.8 Other Sources of PCBs
There are five facilities with SPDES permits that may provide additional
sources of PCBs, to the Lower Hudson (NYSDEC, March 7, 1991 list of facilities
with SPDES permits). Four of these (Carlyle Plermont Corporation, IBM East
Flshklll Facility, Norllte Corporation and Columbia Corporation) are currently
permitted to discharge PCBs within the Lower Hudson River Basin, but not directly
to the Lower Hudson River. The fifth permitted facility, Metro-North Commuter
Railroad North Harmon Shops 1n Westchester County, discharges PCBs directly into
the Lower Hudson River. Because estimates of flow are not available, the PCB
loading to the Lower Hudson cannot be ascertained. However, according to the '
SPDES permit, the allowable dally average PCB concentration 1s 1.0 ppb (mg/1)
with a dally maximum of 2.0 ppb.
There may be additional Incidental releases of PCBs to the Lower Hudson as
a result of accidental spills and illegal dumping activities. The extent and
total PCB loading of these releases to the Lower Hudson River remain unknown.
A.2-6
-------
A.3 Nature and Extent of Contamination
A. 3.1 Sediments
Intensive studies of PCB geochemistry and transport In the Lower Hudson by
Investigators from Lamont-Doherty Geological Observatory began In the mid-1970s
and have continued to the present day. As a part of those studies, measurements
were made of PCB levels 1n sediments, riverine and estuarlne suspended matter and
dissolved constituents. In addition, accurate radionuclide measurement techniques
were developed, which permitted the accurate dating of core samples taken from
recent (post-1950) sediments.
The dating technique uses natural and anthropogenic radionuclides and It
1s possible to determine the year of deposition of a given layer of sediment
within the sediment core (Olsen, 1979; Bopp, 1979; Bopp et ah, 1982). Since
sediments deposited on the river bottom are derived from the sediments carried
by the river, sediments deposited within a given area reflect the nature of the
river sediment that has traveled past that area. By collecting a core of
undisturbed river sediments and sectioning 1t into layers, which approximate an
annual load of sediment accumulation, 1t is possible to analyze the annual load
of sediments carried by the river and their associated contaminants.
On the basis of dated * res, Investigators were able to examine the annual
variations in sediment PC. evels and derive water-related PCB transports at
locations throughout the Lower Hudson. Figure A.3-1 plots the year of deposition
and total PCBs of dated cores at several Hudson River locations. River Hile
188.5 represents a location 1n the Thompson Island Pool 1n the Upper Hudson.
River M11e 143.4 represents the Lower Hudson sediments 1n the Albany area. River
H1les 88.6 and 91.8 are located near Kingston, New York. River Nile 53.8 Is
located 1n a cove near Cornwall-on-Hudson, New York. River Miles -1.65 and -1.7
are in upper New York Bay. River Miles 53.8, -1.65 and -1.7 are ซil located at
or below the salt front and are subject to NYC metropolitan area Impacts.
Collectively, these plots illustrate the temporal trends of sediment PCB levels
A.3-1
-------
from the mid 1950s to the late 1980s (Bopp, 1979; Bopp et al., 1982; Bopp et al
1984 and Bopp and Simpson, 1989).
As Indicated In Figure A.3-1, there 1s a gradual Increase 1n the sediment
PCB concentration from about 1954 to about 1970. This loading 1s largely
attributed to General Electric PCB releases that escaped capture behind the dams
of the Upper Hudson. In 1973, the removal of the Fort Edward Dam followed by
major river flood events resulted In a substantial Increase 1n the PCB loading
to the Lower Hudson. As Indicated by all the sediment cores, maximum PCB
deposition occurred throughout the Lower Hudson around 1973. The decrease 1n
sediment PCB levels from 1977 to the most recent measurement 1s attributed to the
reworking, resuspenslon and gradual dispersion by the river of the sediments
released after the Fort Edward Dam was removed. In addition, the discontinued
use of PCBs by GE since 1977 may have also contributed to decreased PCB sediment
levels.
The Lamont-Doherty studies also demonstrated that the sediments record the
characteristics of the PCBs being transported 1n the Lower Hudson (Bopp et a/.,
1982; Bopp et ซ7., 1984; Bopp and Simpson, 1989). The analysis of the cores
showed that the nature of the PCBs stored 1n the sediments was not constant
throughout the river and that an additional source of highly chlorinated PCBs
must be located 1n the saline region of the Lower Hudson. Bopp and Simpson
(1989) conclude that 1986 loadings of PCBs In the NYC metropolitan area sewage
are comparable to the 1986 PCB loading from the Upper Hudson.
Figure A.3-2 shows the variation 1n some of the PCB homologues in two core
locations, one above the salt front (River Miles 88.6 and 91.8) and one 1n upper
New York Bay (River Miles -1.65 and -1.7). The cores taken at River Miles 88.6
and 91.8 are located well above the salt front and are free of any Influence from
NYC metropolitan area PCB loadings. The maximum values for the highly
chlorinated homologues In these cores are attributed to the use of more highly
chlorinated Aroclor mixtures 1n the early years of production at the General
Electric facility. If the General Electric Inputs were the only significant PCB
releases to the Hudson River, then the PCB homologue variations with time should
A.3-2
-------
be similar throughout the entire Hudson. When the sediments at River Miles -1.65
and -1.7 are examined, however, 1t 1s clear that the homologue variations wltli
time are quite different from those at River Miles 88.6 and 91.8. The downriver
cores show maximum values roughly 10 to 15 years later than those collected above
the salt front. In addition, the absolute concentrations of these homologues are
higher down river. Based on the preceding data, Bopp and Simpson (1989) conclude
that an additional source or sources of highly chlorinated PCBs must be located
1n the lower portion of the Lower Hudson.
Figure A.3-3, an expanded view of the cores at River Miles 88.6 and -1.65,
offers additional supporting evidence for the Importance of the NYC metropolitan
area as a source of PCBs (Bopp and Simpson, 1989). The results for the sediments
at River Mile 88.6 show an exponential decrease in the sediment PCB concentration
from 1973 to about 1986. The curve appears to be asymptotic to zero with the PCB
concentration of annually deposited sediments decreasing by a factor of two every
3.5 years. This finding suggests that the annual loading of PCBs to this part
of the river 1s decreasing at the same rate. The general decrease 1n sediment
PCB concentrations with time 1s consistent with the decrease 1n PCB concentra-
tions recorded at the US6S Upper Hudson monitoring stations.
The results for the sediment core at River M11e -1.65 represent sediments
accumulating in upper New York Bay, where the influence of both uprlver and NYC
metropolitan area Inputs should be seen. As seen 1n Figure A.3-3, the PCB trend
with time appears to have the same exponential decay rate as the uprlver core,
but Is asymptotic to 0.5 mg/kg Instead of zero. As of 1986, It appears that
sediments Influenced by the NYC metropolitan area Inputs were accumulating with
higher PCB levels than those found further upstream beyond the Influence of the
metropolitan region. Based on the absolute concentrations In the sediments at
these two coring locations, Bopp and Simpson (1989) also concluded that the NYC
metropolitan area related inputs in 1986 were of similar magnitude to those
originating uprlver. NYCDEP (1987) records of PCB levels Indicate that Lower
Hudson River stations from the New York-Bronx County Line to the Narrows had an
average concentration of 0.488 mg/kg from 1983 to 1987, which 1s comparable to
the Bopp and Simpson (1989) 0.5 mg/kg asymptote.
A.3-3
-------
The PCB sediment studies of Lamont-Doherty also Indicate the lack of
blodegradatlon In the sediments of the Lower Hudson. Based on core dating, 30-
year old sediment layers show little variation In the patterns of congeners
constituting the PCB levels. Thus, sediments still retain patterns that can be
readily described by the standard Aroclor mixtures and there 1s no apparent shift
toward the less chlorinated PCBs (Bopp et <7., 1982). In addition, Bopp et <7.
(1982) observed Increased ratios of some of the more highly chlorinated PCB
peaks, a finding that may Indicate the presence of Aroclor 1254.
On the basis of their sediment analysis, Bopp (1979) and Bopp and Simpson
(1989) estimated the' Inventory of PCBs 1n the sediments of the Lower Hudson,
Including the dredge spoils excavated 1n recent years. These estimates are based
on the combination of dated cores collected throughout the Lower Hudson and the
matching PCB analysis. The Inventory was estimated by determining the sediment
accumulation rates based on the radionuclide records and by measuring the PCB
concentrations 1n those sediments. These calculations yield a PCB burden of
187,000 lb (85,000 kg) In the sediments of the Lower Hudson and an additional
82,000 lb (37,000 kg) 1n the dredge spoils removed to the New York Bight. Table
A.3-1 summarizes the PCB Inventory estimates on an area-specific basis. The
authors estimate the uncertainty of these estimates to be about a factor of two.
Several studies of the Lower Hudson also Included suspended matter PCB
measurements .Unlike the sediment studies which provide time-Integrated samples,
the water column samples are snapshots of Instantaneous riverine conditions.
Nonetheless, they provide a useful confirmation of the sediment records as well
as a means to relate the sediment records to the water column Inventory.
Although the majority of samples collected for suspended matter PCB analysis were
taken for the Uppe^ Hudson (Bopp et a7., 1985), some results were obtained around
River M11e 3 (Bopp and Simpson, 1984). The results are presented In Table A.3-2
and Indicate that water column suspended matter correlates extremely well with
the sediment core tops.
A. 3-4
-------
Lamont-Doherty Investigators also addressed the partitioning of PCBs
between suspended matter and water. The partitioning can be described by a
constant (K.) which relates the suspended matter PCB concentration to the
dissolved PCB concentration. A K. value greater than 1,000 generally Indicates
that the compound 1n question tends to strongly associate with the st&^ended
matter. The first study consisted of In situ measurements. The second study
involved a laboratory simulation, determining K. under a broad range of
conditions. On the basis of 1n situ measurements of dissolved and suspended
matter PCB concentrations (Bopp, 1979) and laboratory simulations (Warren et ซ7.,
1987), the K. for PCBs was shown to vary with several factors. In general, K,
Increased with Increasing PCB chlorine content, decreasing dissolved organic
carbon concentration and decreasing temperature. Based on these findings, Bopp
(1979) and Bopp et <7. (1985) conclude that total water column PCB concentration
varies directly with suspended matter levels 1n the water column. The range of
measured K/s from Bopp (1979) was 1.5 x 10* to 4 x 10s for dlchloroblphenyls to
hexachloroblphenyls, respectively. Warren et ซ7. (1987) found a similar relative
range for these PCBs, although the absolute values shifted, depending on
conditions.
A.3.2 Water
The USGS monitored total (dissolved and suspended matter) PCBs In the Lower
Hudson water column from 1978-81, but they have discontinued their efforts since
that period. Five locations on the Lower Hudson were monitored, with an average
of five samples taken yearly at each location (Schroeder and Barnes, 1983a). The
mean concentrations reported by Schroeder and Barnes are given below.
Water column measurements for the Lower Hudson show consistently lower PCB
concentrations than the Waterford (Upper Hudson) values. In particular, the
measurements at Castleton, the nearest downstream station to Waterford, can be
taken to show the effects of dilution by the Mohawk River, which enters the
Hudson below Waterford, and the Increased watershed drainage area. Like the
Upper Hudson, the PCB levels 1n the Lower Hudson water column showed a declining
concentration trend 1n time over the monitoring period.
A.3-5
-------
Mean PCB Concentrations From Five Lower Hudson Sites
Compared to Upper Hudson Samples at Waterford*
Location
1978
Mean
1979
PCB Concentration
(M/J)
1980
1981
Waterford
0.45
0.36
0.25
0.14
Castleton
0.23
0.20
0.11
0.08
Catskill
0.17
0.17
0.16
0.05
Staatsburg
0.14
0.16
0.19
0.07
Clinton Point
0.15
0.12
0.17
0.08
Highland Falls
0.17
0.07
'From Schroeder and Barnes (1983a) Table 6; Lower Hudson averages based on
approximately 5 samples per year.
Schroeder and Barnes (1983a) caution that limited significance should be
attached to the mean.concentration values summarized above. These mean values
are based on very few data points (typically no more than 5) and recent (1981)
values are very close to analytical detection limits, such that they are subject
to large uncertainty. Current levels of PCBs in the Lower Hudson water column
are uncertain as is the relative contribution of uprlver versus lower river PCB
Inputs.
Although few data exist on total water column PCB concentrations 1n the
Lower Hudson after 1981, it 1s possible to make a rough estimation based on core
top PCB levels as determined by Bopp and Simpson (1989). For the purposes of
these calculations, 1t was assumed that the water column suspended matter PCB
levels were equal to the PCB levels measured in the core tops on a mass basis,
i.e., suspended matter PCB concentration 1n mg/kg of suspended matter equals
sediment PCB concentration In mg/kg of sediment. Sediment cores collected at
River Miles 88.6 and -1.7 recorded sediment PCB levels of 0.8 and 1.5 mg/kg,
respectively. In addition, Bopp and Simpson (1984) obtained measurements of PCBs
in suspended matter River Mile 3 of 1.7 to 1.9 mg/kg 1n 1984.
A.3-6
-------
Estimates were made of total water column PCB levels using the preceding
PCB levels 1n sediment cores and suspended matter. The water column estimates
assumed a partition coefficient of 10* (Bopp et al.t 1985) and suspended matter
of 6 to 40 pg/l from 1984-1986 (Garvey, 1990). For 1984 the water column levels
at River Nile 3 were estimated at 0.02 to 0.08 pg/l; at River Nile -1.7 they were
estimated at 0.02 to 0.06 ug/1. In 1986, the water column levels at River Nile
88.6 were estimated at 0.01 to 0.04 |ig/l. These values are generally consistent
with the decreasing trend seen 1n the water column levels 1n earlier (1978-1981)
USGS Survey data (Schroeder and Barnes, 1983). These estimates are subject to
some degree of uncertainty, since the partition coefficient utilized for the more
saline locations (River Niles 3 and -1.7) was originally based on data from
freshwater systems. To date, no measurements of PCB partition coefficients are
available for saline waters.
A.3.3 Fish
A.3.3.1 Overview of Previous Monitoring Programs
The Lower Hudson River historically supported a valuable commercial fishery
of striped bass and other species. Although the Lower Hudson is considered to
contain one of the most diverse fisheries found throughout the Atlantic Coastal
systems (Beebe and Savldge, 1988), the striped bass fishery has been closed since
1976 because of elevated PCB levels. Since the early 1970s, NYSDEC has
extensively monitored PCBs 1n fish 1n the Lower Hudson. Most of the NYSDEC
studies of PCBs placed a major emphasis on the striped bass fishery as a
consequence of the well documented commercial and recreational value of this
resource.
The first report of PCB contamination In striped bass In the Lower Hudson
was published in 1970 (Boyle, 1970). In 1971 NYSDEC added PCBs to their
statewide analyses for pesticide residues In fish, including the analysis of the
1970 samples. F1sh data collected and analyzed for PCBs in the 1970-74 period
are summarized in Spagnoli and Skinner (1977).
A.3-7
-------
Following the 1974 USEPA investigation of PCB contamination In the Fort
Edward area (Nadeau and Davis, 1974), NYSDEC modified and intensified Its PCB
sampling program 1n Hudson River fish. Results of a total of 440 Hudson River
fish samples analyzed in 1975-6 have been provided by NYSDEC and were included
1n the TAMS/Gradlent computerized database.
Sampling and analytical procedures for the 1975-1976 collections are
documented In Spagnoll and Skinner (1977). The 1975-76 fish collections were
made by regional NYSDEC F1sh and Wildlife personnel who were Instructed on
specific species and sizes of fish desired, location of stations and timetables
for collection. Target species for the Lower Hudson Included smallmouth bass
(Hlcropterus dolonlenl), largemouth bass (Mlcropterus salaoldes), brown bullhead
(Ictalurus nebulosus), goldfish (Carasslus auratus), white sucker (Catastomus
coamersonl), striped bas$ (Morone saxatills) and various other Lower Hudson River
species. Other species were occasionally obtained as available. Attempts were
made to sample small, medium (minimum legal) and large representatives of each
species.
Analyses conducted by several different state laboratories were reported
against standards for Aroclor 1242 or 1016 and Aroclor 1254. Aroclor 1221 was not
analyzed. The nominal detection limit of the method was 0.01 ppm, although some
of the labs reported results only as low as 0.1 ppm. .
Following the 1975-1976 sampling effort, the commercial striped bass
fishery 1n the Hudson was closed 1n February 1976 as a consequence of elevated
PCB levels. As a condition of the 1975 Settlement Agreement signed by GE and
NYSDEC, GE provided funds for continued PCB monitoring 1n fish collected annually
at various locations below the discharge. A fish sampling program, redesigned
in part to address the need for reassessment of closure of the commercial
fishery, was implemented in Spring 1977.
In 1980, NYSDEC made a commitment to commercial fishermen to review the
annual PCB data prior to mid-November of each year. This date was established
to provide an adequate amount of time to prepare any necessary changes in the
A.3-8
-------
regulations regarding the closure of the commercial fishery and to provide
commercial fishermen with the opportunity to purchase new nets In the event that
the fishery were reopened (Horn and Sloan, 1984). This mandate has resulted In
periodic reports on the status of the striped bass fishery (Horn and Sloan* 1984;
Sloan and Horn, 1986; Sloan et *1., 1986; Sloan et <7., 1987; Sloan tV #7.,
1988).
Methodology for the sampling effort since 1977 1s described in Armstrong
and Sloan (1981), Sloan and Horn (1986) and Sloan et <7. (1W8). Collection of
resident species was as close as possible to the same time each year in order to
minimize the effects of seasonal variation on PCB testing. Striped bass were
typically collected during the spring migration (April, May and sometimes as late
as June). Distinct summer and fall collections were made 1n several years'to
Investigate seasonal variation. Because striped bass were caught during spring
migration, the location at which they were caught probably bears little or no
relationship to the PCBs In sediment and water at that location.
Aroclor concentrations were determined by comparison to standards for
Aroclor 1221, Aroclor 1016 and Aroclor 1254 obtained from the Monsanto
Corporation. The method utilized was not able to distinguish between Aroclors
1242 and 1016. The Aroclor detection limit for each Aroclor tested was 0.1 ppป
wet weight through 1986; the detection limit for Aroclor 1221, which was never
a large component of the total PCBs detected 1n striped bass, was changed to 0.05
ppm in 1987. For each sample, the percent lipid was determined as the percent
by weight of tissue soluble 1n petroleum ether.
Since 1978, striped bass have been the species most commonly monitored,
although a number of other species have been sampled as well. Table A.3-3
provides a summary of the major fish species sampled by year and the type of fish
tissue analyzed as Included 1n the NYSDEC database. Table A.3-4 presents the
overall average lipid content of fish samples from the Lower Hudson. The most
recent fish samples for which PCB results are available were collected In 1988.
Fish samples collected and prepared for analyses In 1990 have been shipped
recently to the analytical laboratory for PCB analyses (R. Sloan, pers. comm.).
A.3-9
-------
A.3.3.2 Striped Bass
Time series trends of total PCBs, Aroclor 1016, and Aroclor 1254 on a ppm
wet weight basis 1n the spring-collected striped bass from the Lower Hudson show
a large decline from 1978 to 1979. Using all striped bass data below River Nile
80, no significant change (at the 95 percent confidence level) 1n arithmetic mean
PCB levels was found for the period 1979 to 1987. A statistically significant
decline was observed, however, 1n the geometric mean PCB concentrations (Sloan
et ah, 1988). Geometric means since 1983 for striped bass showed a consistent
decrease In both total PCBs and Aroclor 1254, but levels of Aroclor 1016 were
found to remain relatively constant between 1983 and 1987. Sloan et a/, conclude
that "a general, significant decline of total PCB occurred from 1978 through
1981. In 1982 and 1983, concentration Increased. Since 1983, levels have shown
a slow but apparently steady reduction with the 1987 values ... similar to those
observed 1n 1981." Lower PCB levels were also noted 1n samples which were taken
from locations lower In the river.
Table A.3-5 summarizes data for striped bass from the Lower Hudson,
obtained from the Lower Estuary (River Nile 12 to 76) and the Upper Estuary
(River Nile 91 to 153). (The 1977 data for one sample only are omitted from this
analysis.) The results for River Nile 27 (the Tappan Zee Bridge) and 153
(Federal Oam) are each shown separately 1n this table. The data for both
locations are also Included 1n the overall lower and upper estuary summaries In
the same table. In the years 1986-1988, the median PCB concentrations 1n striped
bass from the lower estuary have ranged from 2.4 - 3.0 ppm, whereas the median
PCB levels 1n striped bass from the upper estuary have ranged from 4.8 - 11.2 ppm
during this period.
Because of the tendency of PCBs to concentrate 1n lipids, some studies have
analyzed data that have been based on lipid content of the samples. In the first
four years of sampling (Sloan and Armstrong, 1980), data were generally found to
be much more consistent when based on lipid content. Other studies have not been
based on lipid content. Sloan et a7. (1988) state that "expressing PCB data on
a lipid basis (I.e. tig PCB/g-Hpid) has not been useful in the past to explain
A.3-10
-------
variability in PCB concentrations in migrant species." Therefore, in their study
of striped bass, results were not based on lipid content.
In the present analysis, PCB data 1n fish have been evaluated on a lipid
basis to provide consistency with results from the Upper Hudson. Time trends in
I1p1d-based means for Aroclor 1016, Aroclor 1254 and their ratio for striped bass
are shown 1n Figure A.3-4. Since 1978, the arithmetic means Illustrate a drop
1n concentration, although the change 1s not as dramatic as the Sloan et ซ7.
(1988) geometric mean plots. The data summarized In Spagnoll and Skinner (1977)
report average Aroclor 1254 concentrations (only) 1n strlpSjd bass in the whole
Lower Hudson as approximately 15 ppm wet weight (or approximately 200 pg/g-Hpid,
assuming an average 6.8 percent lipid 1n striped bass).
While the averages show only slight declines In recent years, exponential
decay functions can be fitted to the plots. Aroclor 1016 levels for striped bass
below River Nile 80, excepting 1975-6 data, fit an exponential decay very
closely, with an apparent half-Hfe of 2.6 years (Figure A.3-5, 95 percent
confidence bounds on regression shown by dotted lines). Aroclor 1254 levels do
not provide quite as smooth a fit, but show an apparent half-Hfe of 7.1 years.
The apparent half-life for total PCBs (sum of Aroclor 1016, 1221 and 1254) is 5.3
years.
A comparison of 11p1d-based Aroclor 1016 and Aroclor 1254 levels 1n all
Lower Hudson striped bass samples Is shown in Table A.3-6. Although the
concentration of both Aroclor mixtures has decreased dramatically since
1977/1978, the lower chlorinated congeners quantltated as Aroclor 1016 have
decreased 1n concentration faster than the higher chlorinated congeners
quantltated as Aroclor 1254, leading to an Increase 1n the ratio of Aroclor 1254
to Aroclor 1016. Sloan et <7. (1988) conclude that Aroclor 1016 levels In
striped bass have declined to a level where "the 'Aroclor 1254' component 1s the
determinant for the fate of PCB 1n Hudson River striped bass." This observation
1s significant, because sewage Inputs of PCBs to the estuary from the NYC
metropolitan area are thought to be characterized by a larger percentage of
higher chlorinated congeners than the Input to the estuary from the Upper Hudson.
A.3-11
-------
A.3.3.3 Other Migrant/Harlne Species
Other n1grant/marine species monitored 1n significant numbers, but not
sampled since 1985/1986, Include American eel (Angull7a rostrata), American shad
(Alosa sapidisslma), Atlantic tomcod (Nlcrogadus tomcod), alewlfe (Alosa
pseudoharengus), and blueback herring (Alosa aestivalis).
In general, a correlation Is expected between lipid content of fish and the
concentration of PCBs, because of the lipophilic nature of the compound. An
exception to this rule 1s the American shad where a significant correlation
between total lipid and total PCBs was found 1n only one of 20 sample sets.
There may be a lack of correlation for shad, because shad are transient 1n the
estuary or do not feed there. For other migrant species Including alewlfe,
blueback herring and rainbow smelt, PCBs appear to accumulate at a rate related
to body size, I.e. surface area to volume ratio (Sloan and Armstrong, 1980). PCB 1
concentrations 1n marine species such as Atlantic tomcod, Immature blueflsh,
Atlantic sturgeon and American eel are reported as showing significant
correlations with 11p1d content, but not with length.
For the years from 1978 through 1981, a significant decrease 1n PCB
concentrations 1n fish was observed for all species, but "most of the decline 1n
PCB concentrations of migrant/marine species has been primarily due to the
reduction of Aroclor 1016" (Sloan and Armstrong, 1980). The average percent
decline In Aroclor 1016 calculated over those years was 42 percent, compared to
five percent Aroclor 1254.
A relatively long time series of observations for a few migrant/marine
species 1n the estuary are available at the Tappan Zee Bridge (River Mile 27).
Trends 1n I1p1d-based PCB concentrations at this location are shown In Figure
A.3-6 for striped bass and American shad, with 95 percent confidence limits on
the arithmetic means. This data set from the lower estuary does not show the
sharp drop off 1n PCB concentrations from 1978 to 1980 typical of the complete
Lower Hudson data set. The shad show substantially lower bloaccumulatlon than
the striped bass, reflecting their short residence 1n the estuary.
A.3-12
-------
A.3.3.4 Resident Freshwater Species
A significant population of resident fish species, present 1n the
freshwater portions of the river, form a valuable resource. Significant numbers
(1 n space and time) of PCB analyses have been undertaken for the freshwater
species largemouth bass (Mcropterus salaoides), pumpklnseed (Lepoaus gibbosus),
redbreasted sunflsh (Lepoaus aurltus), white perch (Horone anerl carta), and yellow
perch (Perca flavescens). The greatest number of samples 1n the upper estuary
are available for largemouth bass and pumpklnseed. For these fish, no clear
trend for PCB levels 1s apparent since 1980. Median PCB levels In largeaouth
bass at River Miles 112-114 go from a high of 31.2 ppm In 1978, to a low of 0.9
ppm 1n 1980 and rise again 1n 1986 to 10.2 ppm (Table A.3-7).
L1p1d-based trends In total PCB concentrations 1n fish at Catsklll (River
Miles 112-114) are shown 1n Figure A.3-7. In contrast to the time series at
Tappan Zee, the drop 1n PCB burdens from 1978 to 1980 1s very clear In the
largemouth bass population. Since 1980, total PCB concentrations 1n largemouth
bass at Catsklll have remained approximately steady. Althouoh there are less
data for yellow perch (the next most frequently sampled fish at this location),
a similar pattern exists.
A large number of fish have been sampled in the upper estuary from Catsklll
ver Miles 112-114) and just below the Federal 0am at Troy (River M11e 153).
Although technically 1n the Lower Hudson, fish species at River Mile 153 are very
similar to those found just upstream. Furthermore, the P loading here 1s
likely to come from upstream only and can be estimated from lo... measurements at
Waterford. Therefore, the fish sampling just below the Federal 0am Is aost
conveniently reviewed together with samples from the Upper Hudson and 1s
discussed 1n conjunction with NYSDEC fish monitoring for the Upper Hudson River
presented In Part B.
A.3-13
-------
PA6E INTENTIONALLY LEFT BLANK
f
A.3-14
-------
A.4 Review of Lower Hudson PCB Mathematical Model
A variety of models developed within the past thirteen years have
collectively attempted to link PCBs 1n the Upper Hudson to those found In the
Lower Hudson. There have been efforts to develop specific transport models of
PCBs (LMS, 1978, 1979; Aplcella and Zlmmle, 1978), models of ecosystem fate of
PCBs (Hydrosclence, 1979) and physio-chemical and striped bass food web models
of PCB homologues (Thomann et a7., 1989).
Such a complex undertaking 1s not without difficulties and some models 1n
the past have been criticized for not using "state-of-the-art" approaches (USEPA,
1981). Llmburg (1986) has summarized and reviewed the successes and failures of
both the LMS (1978; 1979) and Hydrosclence (1979) models. For Phase 1, a
preliminary understanding of some aspects and assumptions of the Thomann et ซ7.
(1989) model 1s appropriate 1n order to understand the Intricate linkages of
various physical, chemical and ecological parameters within the Lower Hudson.
A.4.1 Thomann Model
A.4.1.1 Overview
Thomann et al. (1989) considered many aspects of PCB transport, geochemis-
try and blogeochemlstry In order to construct a fairly complete description of
the PCB budget for the Lower Hudson. The model was designed to consider PCB
dynamic patterns over long time and large spatial scales. The model Includes
water column transport, sediment Interactions, degradation, dredging, gas
exchange, biological Interactions and tidal dispersion. In addition, the model
considers Individual PCB homologues and their potential environmental fate.
The model evaluates the time history of PCB Inputs from 1946 to 1987 and
calculates annual PCB budgets for the Lower Hudson. The work Integrates and
summarizes the effects of major PCB sources and sinks over time and provides
Insight concerning recent loads to the Lower Hudson. The mass of PCB Input to
the system on a source by source basis was calculated and the mass of PCBs lost
A.4-1
-------
from the system via gas exchange, burial 1n the sediments, transport to the shelf
waters and degradation was estimated. Thus, the model provides a means of
understanding the current conditions 1n the river, how they developed, and how
they may change 1n the future.
On the basis of the model, several conclusions. were drawn on the
environmental fate of PCBs 1n the Lower Hudson, New York Bight and Long Island
Sound. The study concluded that the maximum PCB load to the Lower Hudson via the
flow at the Federal Dam was 150 lb/day (68 kg/day), corresponding to the removal
of the Fort Edward Dam In 1973. Since that time, the load decreased exponential-
ly, to an estimate of 3 lb/day (1.4 kg/day) 1n 1987. The load of PCBs from the
Upper to the Lower Hudson was dominated by the d1- and trlchloroblphenyls (40
percent) and by the tetrachloroblphenyls (40 percent). The remainder consisted
of heavier PCB homologues.
The PCB loadings to the Lower Hudson from sewage discharges, atmospheric
deposition and runoff also reached a maximum 1n the early 1970s at 30 lb/day (14
kg/day) and declined steadily since that time. However, the decline of these
sources has not been as rapid as that of the uprlver source. These additional
sources dominate the PCB Input to the entire system and represent about 46
percent of the total PCB loading to the Lower Hudson 1n 1980.
A mass' balance for the Lower Hudson and estimates of the relative
contribution of various sources and sinks were presented. A total of 595,000 lb
(270,000 kg) were estimated to have been discharged into the Lower Hudson River
through 1987. Of this total, 66 percent was estimated to have entered via the
flow at the Federal Dam, 20 percent via tributaries and urban runoff, 12 percent
from municipal sewage discharges and 2 percent from atmospheric deposition.
PCB losses from the Lower Hudson River through 1987 were estimated as
follows: 66 percent was lost via gas exchange to the atmosphere, 19 percent was
lost via transport to Long Island Sound and the New York Bight, 9 percent was
lostyla dredging of sediments, while 6 percent remained stored 1n the sediments
of the Lower Hudson. In addition, the loss terms were homologue-dependent, with
A.4-2
-------
gas exchange being more Important for the lower chlorinated homologues and
transport to the surrounding waters playing a more significant role for the
highly chlorinated homologues.
The storage of PCBs 1n the Lower Hudson sediments did not always rq^esent
a net loss of PCB from the water column. The study concluded that in the 1970s
the sediments acted as a net sink of PCBs. By 1987, however, the sediments were
releasing PCBs to the water column with a resulting loss of about 10 percent of
the maximum sediment PCB Inventory. Although the sediments remain a net sink,
their Importance to the PCB mass balance as a sink term will continue to decrease
with time.
As part of the Investigation, a striped bass food web model of PCB
homologues was constructed. In order to assemble the model, the authors' analyzed
the existing striped bass, PCB data base, formulated equations based on uptake
and accumulation components and specified food web Interactions and physiological
parameters. Finally, they compared (calibrated) model output of total PCBs and
PCB homologues to observed levels In certain age classes of striped bass and
white perch collected 1n selected regions of the Lower Hudson.
The striped bass food web model encompassed a variety of trophic groups
Including the phytoplankton, zooplankton (e.g. Gamarus), "small fish" and white
perch. The age-dependent (17 age classes) striped bass model was driven by
outputs from the geochemlcal transport model and was "...successfully calibrated
to white perch and striped bass 1n the mid and Lower Hudson using data from
1978-1987" (Thomann et ซ7., 1989).
Some of the major results of the striped bass food web model Indicate that
peak concentrations of PCBs (45 |tg/g, wet weight) In striped bass occurred 1n the
mid 1970s and have declined at an exponential rate of 0.057/yr since 1980-1982
(approximate half-life of 12 years). In addition, PCB concentrations In striped
bass are linear to the phytoplankton uptake of PCBs from the water column
(commonly termed bloconcentratlon). Furthermore, greater than 90 percent of the
PCBs In striped bass are due to food web bloaccumulatlon and only a minor
A.4-3
-------
percentage (approximately 10 percent) 1s accounted for by an uptake of PCBs from
the water column.
The effort by Thomann et <7. (1989) represents a vast undertaking, because
so many parameters and fluxes were directly linked 1n order to produce an
Improved understanding of PCB dynamics 1n the Lower Hudson. Specific results of
the modeling with respect to PCB mass transport estimates, geochemlcal processes
and ecological parameters are presented below.
A.4.1.2 Mass Transport Estimates
PCB mass loads were estimated from PCB and river discharge measurements
taken at the US6S monitoring station In Uaterford. For the period 1977-1983,
Thomann et a7. (1989) estimated that approximately 19,000 kg (2,700 kg/yr on
average) of PCBs were transported past Uaterford from the Upper to Lower Hudson.
The reported estimate 1s approximately 35 percent higher than the quantity
computed by TAMS/Grad1ent, which Is presented at B.4.
Mass loads of PCBs from the Upper to Lower Hudson prior to 1976 cannot be
estimated from water column PCB concentrations, because none existed. (USGS
monitoring began 1n 1976.) Prior to 1976, the model based mass load calculations
on an approximate dating method applied to the PCBs measured 1n NYSDEC sediment
core samples as reported by Hetllng et al. (1978). Apparently, a constant
sedimentation rate was assumed, which may be unrealistic 1n light of the
historical channel destablUzatlon and sediment scour following the removal of
the Fort Edward Dam. Thomann et al. (1989) estimate that approximately 1,000
kg/yr of PCBs were transported from the Upper Hudson In the early to mid 1960s.
A maximum PCB discharge of 24,600 kg/yr was estimated for 1973 (Thomann et <7.,
1989), which is approximately five-fold higher than that suggested by radiologi-
cal ly dated cores from the Lower Hudson (Bopp and Simpson, 1989).
A. 4-4
-------
A.4.1.3 Geochemlcal Processes
In the construction of a PCB budget for the Lower Hudson River, the Thomann
model was required to simulate several Important geochemlcal processes. In
particular, the model needed to simulate PCB partitioning 1n the water column
between dissolved and suspended matter forms. It also simulated the gas exchange
of PCBs between river water and the overlying ambient air. Although there 1s a
need to simulate the above processes, estimates of parameters used in simulations
may be subject to some uncertainty and, 1n some cases, may have been underesti-
mated.
Of particular Importance are the suspended matter (or sediment) to water
partition coefficients (Ko's) and the gas exchange coefficients. Thomann used
a model of water-sediment interactions for the Ko values, which consistently
predicts K,, values below those measured by the Lamont-Doherty Geological
Observatory (Bopp et al., 1985; Warren et ซ7., 1987). This consistent bias in
the Ko will result 1n the overestimate of the water-borne transport from the
Lower Hudson, since estimates of the total water column concentration are
directly dependent on the partition coefficient. It will also result in the
overestlmatlon of any flux, such as gas exchange, which Is dependent upon the
dissolved concentration of PCBs.
The model assumed a constant value for the gas exchange coefficient,
derived from oxygen exchange across all congeners found in the river. This
assumption Ignores several important factors related to the large chemical
differences among individual PCBs and between PCBs and oxygen and may overesti-
mate the actual gas exchange rate by a factor of two or more. Depending upon the
type of model used to simulate gas exchange, Ignoring the differences among
congeners can result 1n errors of 20 to 40 percent In the gas exchange
coefficient. Furthermore, the lack of Incorporation of additional gas exchange
resistance related to the very low Henry's Law constant values for PCBs (about
10'2 to 10ฐ, unltless) may underestimate the gas exchange coefficient (Bopp,
1983).
A.4-5
-------
The gas exchange rate utilized as an Input to the model Is probably high,
given literature estimates for PCBs (Bopp, 1979). In addition, the gas exchange
problem Is compounded by the underestimation of K as previously discussed.
Since model calculations show gas exchange loss to represent the most important
single flux of PCBs from the Lower Hudson, the potential error 1n all of the mass
balance calculations Is presumably high.
A.4.1.4 Ecological Parameters
The core of the striped bass food web model of PCB homologues links the
change in concentration of PCBs 1n striped bass to three components: (1) uptake
of PCBs from the water; (2) uptake of PCBs from other sources, mainly predatlon;
and (3) elimination (depuration) of PCBs. Food chain uptake parameters which
Include bloconcentration factors (BCF) and bloaccumulatlon factors (BAF), are
discussed below and defined 1n the Glossary.
Uptake of PCBs from Water (Bloconcentration)
Thomann et al. assigned the BCF for all levels above the phytoplankton as
equivalent to the log octanol/water partition coefficient (K.J. Since the
primary route for accumulation from the water Is by exchange across lipoprotein
cell membranes, it appears valid to assign BCFs based on K.., although regression
relationships between K^, and BCF values would be preferred. It should be noted,
however, that the efficiency of uptake Increases with increased K., until a
plateau is reached and then declines at log Rvalues greater than 6, which is
in the range of the pentachloroblphenyl and hexachloroblphenyl PCB homologues.
The phytoplankton BCF was assigned a constant value of 30 7/g(w) [liters
per gram (wet weight)], based largely on the results of Oliver and N11m1 (1988).
The constant value assigned 1s for PCB homologues dichlorobiphenyl through
hexachloroblphenyl, which actually ranged from 14 to 61 7/g(w). Other PCB
homologs such as octachlorobiphenyl had a BCF of 100 7/g(w), according to the
data presented by Oliver and N11m1 (1988) for Lake Ontario plankton.
A. 4-6
-------
Thomann et i7. use Oliver and NHml's (1988) plankton data as Indicative
of phytoplankton BCF when, 1n fact, their plankton samples contain a Mixture of
phytoplankton and zooplankton.. In addition, studies on the uptake kinetics of
PCB by phytoplankton suggest a great deal of species-spedf 1c variation (ฃosper,
1991 per. comm.) Thus, a reassessment of the phytoplankton BCF 1s probably
warranted.
Uptake of PCBs from other Sources (Bloaccunulatlon)
The main avenue of PCB accumulation 1n fish Is via consumption of food
containing PCBs. As an animal consumes food, Its growth will be less than that
projected by the content of the food. One gram of food does hot produce one gram
of growth, since there are associated respiratory losses, which Thomann et <7.
(1989) correctly recognize. An important variable here 1s the assimilation rate,
which diminishes with log above 6, but remains greater than zero at all times
(Thomann et a/., 1989). Thus, the assimilation efficiencies used for the various
homologues seem reasonable.
As Thomann points out, most studies Indicate that striped bass shift from
a primarily Invertebrate diet to a primarily piscivorous mode of feeding at
approximately two years of age or older. Depending on salinity, either Gunarus
or Neonysls makes up the largest percentage of the diet for the young striped
bass (age classes 0 to 2). Contrary to the Thomann assumption, however, white
perch have not been shown to be "an important food source" of Hudson River
striped bass populations and It 1s not appropriate to consider white perch as the
"middle trophic level" for striped bass populations. Rather, 1t Is documented
that Hudson River striped bass (age classes > 2 years) are opportunistic
plsclvores and consume a variety of middle level trophic prey such as blueback
herring, Atlantic tomcod, bay anchovy and mummlchogs (Klauda and Setzler-
Hamllton, per. comm.). The relative percent contribution of the various species
to total striped bass dietary consumption 1s affected by seasonal changes and
constraints In prey populations.
A.4-7
-------
Another potential problem 1n the food web Interactions presented by Thomann
concerns the link between the phytoplankton and zooplankton compartments.
Thomann et *7. (1989) state that the phytopl ankton are preyed upon by the
zooplankton, which 1s represented by Ganaarus. However, It 1s well established
that Gammarld amphlpods, Including the genus Gvmarus, are essentially bottom
dwelling eplbenthlc organisms and commonly Inhabit the surflclal sediments.
Gammarlds are classified as detritus feeders or scavengers and do not utilize
phytoplankton as a primary dietary source. In fact, Bek (1972) has shown that
dally consumption of detritus by juvenile and adult gammarlds may range from
60-100 percent of total body weight. These food links and consumption rates are
Important considerations In bloaccumulatlon assessments, since sedimentary
concentrations of PCB are often orders of magnitude greater than those In the
water column.
Elimination of PCBs
The term excretion 1s somewhat misleading, since what should be understood
by the term are all those factors that result In a reduction of body concentra-
tion of PCBs. Such a change could come about from actual excretion, metabolism
or dilution as the animal grows. In view of the extremely low water solubility
of PCBs, the excretory route Is probably low, as Indicated by Thomann. On the
other hand, animals do grow, leading to a possible dilution of the Ingested PCBs.
A.4.2 Simulations Relevant to Upper Hudson Remediation
As part of their Investigation, Thomann et ซ7. (1989) developed a fairly
Intricate model of the food web, which was coupled to a geochemical transport
model. Their effort Is one of the first attempts to model the breadth and
complexity of PCBs 1n the Lower Hudson River, New York Bight and Long Island
Sound. Undoubtedly, the most Important use of the model 1s for estimating future
PCB levels 1n striped bass under various remedial scenarios.
A. 4-8
-------
The model was used to simulate the time 1t would take for 50 percent and
95 percent of the striped bass to be below the FDA threshold of 2 |ig/g (wet
weight) of PCBs under two scenarios: (1) complete elimination of Upper Hudson PCB
sources and (2) status quo or a no action alternative. Conclusions based on
these scenarios are a direct result of their estimates on the relative Importance
of the uprlver source and how other Inputs to the Lower Hudson will vai7 with
tine.
Under the no action scenario, the authors conclude that by 1992, the median
concentration 1n three to six-year old striped bass wou'ij be below the FDA
threshold of 2 //g/g (wet weight). If the uprlver source were completely removed
as of 1987, they conclude that this median concentration would be achieved a few
years earlier (1990). When conditions are extrapolated to achieve a goal of 95
percent of striped bass below the 2 yg/g threshold level, both scenarios seem to
yield the same result by the year 2004. Because Thomann et ซ7. estimate that the
Upper Hudson load contributes only 10 percent to PCB levels in striped bass, the
Lower Hudson PCB Inputs and sediment releases may be the most critical factor
influencing the time required for PCBs in the striped bass population to drop
below the 2 iปg/g FDA threshold.
The assignment of various geochemical and physiological parameters to the
model should be periodically reinvestigated in light of extant data. This
preliminary assessment of the Thomann et <7. (1989) model agrees with some of the
final comments that "... simulation results should therefore be viewed as
Indicators of overall trends only." In any model, there 1s often a great deal
of skepticism concerning the various simulations. As acknowledged by Thomann ซt
ซ7. (1989), spatial and temporal variability in PCB water concentrations, loading
estimates, and variable striped bass migration patterns may contribute to overall
model uncertainty.
A. 4-9
-------
THIS PAGE INTENTIONALLY LEFT BLANK
A.4-10
-------
PART B
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
CONTENTS
Page
B. UPPER HUDSON CHARACTERIZATION
SYNOPSIS (Section B.l)
B.l Physical Site Characteristics B.11
B.l.l Hydrology B.11
B.l.2 Water Quality and Use B.l-2
B.l.2.1 Water Quality B.l-2
B.l.2.2 Use B.l-5
B.l.3 Population and Land Use B.l-7
B.l.4 Fisheries B.l-8
SYNOPSIS (Section B.2)
B.2 Sources of PCB Contamination B.2-1
B.2.1 GE Discharges (To 1977) B.2-1
B.2.2 Current Permitted Discharges B.2-2
B.2.3 Other Sources B.2-2
1
-------
PART B
CONTENTS
(continued)
ฃaafi\
SYNOPSIS (Section B.3)
B.3 Nature and Extent of Contamination B.3-1
B.3.1 Overview of Sources and Database B.3-1
B.3.2 Sediment B.3-5
B.3.2.1 1976-1978 NYSDEC Sampling B.3-5
B.3.2.2 1984 NYSDEC Sampling B.3-7
Methods and Procedures B.3-7
PCB Results B.3-8
Comparison of 1976-78 and 1984 Studies B.3-10
B.3.2.3 Lamont-Doherty Geological Observatory B.3-11
Investigations
B.3.2.4 Other Studies B.3-12
1983 USEPA Study B.3-12
GE 1989 Baseline Studies B.3-13
(Remnant Remediation Project)
GE 1990 Sediment Sampling B.3-13
(B1oremed1at1on Investigations)
B.3.2.5 Other Chemicals in Sediments B.3-14
B.3.2.6 Discussion B.3-15
11
-------
PART B
CONTENTS
(continued)
Paoe
B.3.3 Surface Water Monitoring B.3-17
B.3.3.1 USGS Flow Records B.3-17
B.3.3.2 Suspended Sediments Monitoring B.3-19
B.3.3.3 USGS PCB Monitoring B.3-20
Methods and Procedures B.3-20
PCB Results B.3-21
Current Full-Year Average PCB Concentrations B.3-23
in Water Column
Summer Average PCB Concentrations B.3-24
B.3.3.4 Other Sources of Water Column Data B.3-25
Waterford Treatment Plant Data B.3-25
Lamont-Doherty Study of 1983 B.3-25
NYSDOH Water Column PCBs B.3-27
Remnant Deposit Containment Monitoring " B.3-27
Program
B.3.4 Fish and Other Aquatic Biota
B.3.4.1 Fish Sampling
Samples Prior to 1975
Samples 1975-1976
Samples 1977-1988
B.3-28
B.3-29
B.3-30
B.3-31
B.3-31
Hi
-------
PART B
CONTENTS
(continued)
Paoe
PCB Levels 1n F1sh B.3-32
L1p1d-Based PCB Concentrations B.3-34
B.3.4.2 Other Chemicals In Fish B.3-35
B.3.4.3 NYSDOH MacroInvertebrate Studies B.3-36
Long-Tern B1omon1tor1ng Study B.3-36
Short-Term Blomonltoring Study B.3-38
B.3.5 PCB Concentrations 1n Air and Plants B.3-40
B.3.5.1 Air B.3-40
Monitoring Near Fort Edward B.3-40
Lamont-Doherty Investigations B.3-42
B.3.5.2 PCB Uptake By Plants B.3-43
Early Studies B.3-43
More Recent Studies B.3-45
B.3.6 Other Media B.3-47
B.3.7 Adequacy of PCB and Aroclor Measurement B.3-48
B.3.7.1 Overview B.3-48
PCB - Aroclors, Congeners, and B.3-48
Analysis Methods
Interpreting Reported Aroclor Results B.3-51
1 v
-------
PART B
CONTENTS
(continued)
B.3.7.2 Discussion of Data Quality Assurance B.3-52
1976-1978 Sediment Survey B.3-52
1984 Sediment Survey B.3-55
Other Sediment Data B.3-57
Hudson River Fish Samples B.3-58
USGS Water Column Data B.3-60
B.3.7.3 Summary B.3-61
SYNOPSIS (Section B.4)
B.4 Data Synthesis and Evaluation of Trends B.4-1
B.4.1 Phase 1 Objectives ฃ.4-1
B.4.2 Flood Flow and Sediment Transport B.4-1
B.4.2.1 Flood Frequency Analysis B.4-1
B.4.2.2 Suspended Sediment Discharge B.4-7
Empirical Trend Analysis B.4-10
B.4.3 PCBs in the Water Column and Mass Discharge B.4-11
B.4.3.1 PCB-D1scharge Relationships B.4-12
Regression Analyses B.4-12
Non-parametric Tests of Trend for B.4-16
Water Column PCBs
v
-------
PART B
CONTENTS
(continued)
PaaeV
B.4.3.2 Mass Transport Estimates B.4-17
't
Regression Approach B.4-18
Direct Estimation Approach B.4-20
B.4.3.3 Discussion of Mass Transport from B.4-26
Upper to Lower River
B.4.4 Analysis of PCBs in Fish B.4-30
B.4.4.1 Evaluation of Time Trends B.4-31
Non-Parametric Trend Test B.4-31
Apparent Aroclor Half-Lives In the F1sh B.4-32
Population
B.4.4.2 Projected PCB Concentrations in Fish B.4-35
B.4.4.3 Relation Between PCB Concentrations in B.4-37
Fish and Water
B.4.5 Summary B.4-40
SYNOPSIS (Section B.5)
B.5 Sediment Transport Modeling B.5-1
B.5.1 Overview B.5-1
B.5.2 Previous Modeling Studies B.5-3
vi
-------
PART B
CONTENTS
(continued)
Page
B.5.3 Hydrodynamic Model Description B.5-7
B.5.3.1 Use of WASP4 Family of Models B.5-7
B.5.3.2 Governing Equations B.5-7
B.5.3.3 Model Implementation B.5-11
B.5.3.4 Model Setup for Thompson Island Pool B.5-14
B.5.3.5 Model Calibration B.5-16
B.5.4 Sediment Transport Model B.5-18
B.5.4.1 Streambed Erosion and Deposition B.5-18
Sediment Transport Capacity B.5-19
Active Bed Layer B.5-20
Bed Erosion B.5-21
Sediment Deposition B.5-23
Bed Elevation Change B.5-24
B.5.4.2 Streambank Erosion B.5-24
Sediment Routing B.5-27
B.5.4.3 Initial Calibration Efforts B.5-30
B.5.5 Summary B.5-30
vi 1
-------
PART B
CONTENTS
(continued)
Paoe
SYNOPSIS (Section B.6)
Preliminary Human Health Risk Assessment B.6-1
B.6.1 Phase 1 Objectives B.6-1
B.6.2 Exposure Assessment B.6-2
B.6.2.1 Introduction B.6-2
B.6.2.2 Dietary Intake B.6-4
Fish Consumption B.6-4
Snapping Turtles B.6-10
Ingestion of Agricultural or Home Garden Crops B.6-11
Ingestion of Beef or Dairy Products B.6-12
Breast M1lk B.6-12
Drinking Water B.6-14
B.6.2.3 Inhalation Exposures B.6-15
Exposure From Air B.6-15
B.6.2.4 Recreational Exposures B.6-16
Dermal Absorption From Contact With Sediments B.6-16
Incidental Ingestion of River Sediments B.6-19
Dermal Absorption From Water B.6-20
vlii
-------
PART B
CONTENTS
(continued)
Page
B.6.3 Toxicity Assessment B.6-23
B.6.3.1 Introduction B.6-23
B.6.3.2 Noncarc1nogen1c Effects B.6-24
B.6.3.3 Carcinogenic Effects B.6-26
Definition B.6-26
USEPA Cancer Slope Factor B.6-27
Other Cancer Slope Factor Studies B.6-28
B.6.3.4 Toxicity of Specific PCB Congeners B.6-30
B.6.3.5 Epidemiological Studies B.6-31
Cancer Effects B.6-32
Non-Cancer Effects B.6-32
B.6.3.6 Other Health-Based Regulatory Limits B.6-33
or Guidelines
FDA: Tolerance for PCBs in F1sh B.6-33
USEPA: Drinking Water B.6-34
USEPA: Ambient Water B.6-34
New York State: Ambient Water B.6-35
USEPA Advisory Levels for PCB Superfund B.6-35
Clean-up
National Academy of Sciences: Suggested B.6-36
No-Adverse-Response Level
1x
-------
PART B
CONTENTS
(continued)
%
Page V
Standards for Occupational Exposures B.6-36
)
New York State: Ambient Air Guidelines B.6-37
B.6.4 Risk Characterization B.6-37
B.6.4.1 Definition B.6-37
B.6.4.2 Dietary Intake B.6-39
Fish B.6-39
Drinking Water B.6-39
Other B.6-40
B.6.4.3 Inhalation Exposures B.6-40
B.6.4.4 Recreational Exposures B.6-40
Dermal Exposure to River Sediment B.6-40
Incidental Ingestion of River Sediment B.6-41
Contact With River Water B.6-41
B.6.4.5 Risk Characterization Compared B.6-41
to Human Studies
B.6.4.6 Analysis of Uncertainties B.6-42
B.6.5 Lower Hudson Discussion B.6-45
x
-------
PART B
CONTENTS
(continued)
^pas.
SYNOPSIS (Section B.7)
B.7 Interim Ecological Risk Assessment B.7-1
B.7.1 Phase 1 Objectives B.7-1
B.7.2 Ecosystem Description B.7-2
B.7.2.1 Terrestrial Habitats B.7-2
Habitats B.7-2
Vegetation B.7-3
Birds B.7-3
Mammals B.7-4
Amphibians and Reptiles B.7-4
Threatened and Endangered Species B.7-5
B.7.2.2 Aquatic Ecosystem B.7-6
Conceptual Ecosystem Framework B.7^7
Phytoplankton B.7-8
Perlphyton B.7-8
Macrophytes B.7-9
Invertebrate Community B.7-9
Fish B.7-14
Summary of Aquatic Ecosystem B.7-16
x1
-------
PART B
CONTENTS
(continued)
Page
B.7.3 PCB Exposure Assessment B.7-19
B.7.3.1 Exposure Pathways B.7-19
B.7.3.2 Identification of Indicator Species B.7-21
B.7.3.3 Exposure Quantification B.7-22
Ambient Water and Sediment Exposures B.7-23
Ch1rononm1d Larvae B.7-24 <
F1sh B.7-25
Estimated Fish Dietary Intake B.7-27
Herring Gull B.7-28
Mink B.7-30
B.7.4 Toxicity Assessment B.7-30
B.7.4.1 Types of Toxicity B.7-30
B.7.4.2 Toxicity Literature Review B.7-33
Plants B.7-33
Planktonlc Species B.7-34
Aquatic Macroinvertebrates and Insects B.7-34
Fish B.7-35
Birds B.7-36
Mammals B.7-37
xi 1
-------
PART B
CONTENTS
(continued)
Ease
B.7.4.3 Proposed Criteria and Guidelines B.7-38
Ambient Water Quality Criteria B.7-38
Sediment Quality Guidelines B.7-39
Guidelines for PCBs in Fish B.7-40
Guidelines for PCBs 1n Birds B.7-41
Guidelines for PCBs in Mammals B.7-42
B.7.5 Risk Characterization B.7-42
B.7.5.1 Ambient Water B.7-42
B.7.5.2 Sediment B.7-43
B.7.5.3 Fish B.7-43
B.7.5.4 Fish-Eating Birds B.7-45
B.7.5.5 Mammals B.7-46
B.7.5.6 Summary B.7-46
xi 11
-------
PART B
CONTENTS
(continued)
V
TABLES
Tables Located 1n Volume 1 (2 of 2)
B.11 Water Quality Rating Criteria
B.l-2 Public Water Supplies on the Upper Hudson River
B.l-3 Fish Species Occurrence Summary Between Fort Edward and the Federal
Dam
B.2-1 Current Permitted PCB Discharges, Upper Hudson River Drainage Basin
B.2-2 Inactive Disposal Sites Located Near Upper Hudson River
B.3-1 Studies of PCB Contamination In the Hudson
B.3-2 Hudson River Sediment Database Summary
B.3-3 Comparison of Sediment Samples By River Mile
B.3-4 PCB Levels 1n 1976-1978 Sediment Samples
B.3-5 1984 Thompson Island Pool Sediment Summary
B.3-6 Texture Classifications From 1984 Sediment Study
B.3-7 GE Baseline Remnant Remediation Sediment Monitoring
B.3-8 Total PCBs In Sediments - GE's 1990 Study and Comparison to Earlier
Studies
B.3-9 Dally Flows For Upper Hudson USGS Gauging Stations
B.3-10 Suspended Sediment Monitoring Summary
B.3-11 Total PCBs 1n the Water Column - USGS Stations
B.3-12 Current (1986-89) Average Water Column PCB Concentrations - USGS
Stations
xlv
-------
PART B
CONTENTS
(continued)
B.3-13 Summer Average Water Column PCB Concentrations (pg/1) - USGS
Monitoring Stations
B.3-14 Upper Hudson Yearly F1sh Count
B.3-15 Average Aroclor Levels 1n Upper Hudson F1sh
B.3-16 Total PCBs (ppm) 1n Largemouth Bass: Upper Hudson - NYSDEC
Monitoring
B.3-17 Total PCBs (ppm) 1n Pumpklnseed: Upper Hudson - NYSDEC Monitoring
B.3-18 Total PCBs (ppm) 1n Brown Bullhead: Upper Hudson - NYSDEC Monitoring
B.3-19 Lipid-Based Total PCBs for All Fish Species: NYSDEC Database
B.3-20 Other. Chemicals In F1sh Samples
B.3-21 PCBs In A1r
B.3-22 PCBs In Plants
B.4-1 Flood Recurrence Intervals at Fort Edward
B.4-2 Regression Analysis: PCBs 1n Water Column
B.4-3 Published PCB Mass Loading Past Waterford (kg/yr)
B.4-4 Estimated TAMS/Gradient Yearly Average PCB Loads (kg/yr)
B.4-5 Trends 1n Aroclor Concentration at River Mile 175 (|ig/l)
B.6-1 Exposure Assumptions: F1sh Ingestion
B.6-2 Exposure Assumptions: Dermal Contact with Sediments
B.6-3 Exposure Assumptions: Sediment Ingestion
B.6-4 Exposure Assumptions: Dermal Contact with River Water
B.6-5 Cancer Risk Estimates
xv
-------
PART B
CONTENTS
(continued)
B.6-6 Hazard Quotient Estimates
B.6-7 Epidemiological Studies: PCB Carcinogenicity 1n Humans
B.6-8 Epidemiological Studies: Non-Cancer PCB Effects in Humans
B.7-1 Estimated Ecological PCB Exposure Levels for Indicator Species
B.7-2 Summary of Observed PCB Effects in Biota
B.7-3 Summary of Proposed Ecological Guidelines for PCBs
FIGURES
Figures Located 1n Volume 1 (2 of 2)
B.1-1 Mean Monthly Flow in the Upper Hudson River, Water Year 1986
B.12 Mean Monthly Flow 1n the Upper Hudson River, Water Year 1984
B.2-1 Reported General Electric PCB Usage
B.3-1 Total PCBs in Surface Sediments - 1976-78
B.3-2 PCB Concentration vs. Texture Relationship - Gravel
B.3-3 PCB Concentration vs. Texture Relationship - Fine Sand
B.3-4 PCB Concentration vs. Texture Relationship - Find Sand/Wood Chips
B.3-5 PCB Concentration Frequency Comparison
B.3-6 Correlation of Sediment Aroclor 1242 Levels in Upper and Lower
Hudson Sediment Cores
B.3-7a Upper Hudson Dally Average Flows, 1973-1981
B.3-7b Upper Hudson Daily Average Flows, 1982-1990
xvi
-------
PART B
CONTENTS
(continued)
B.3-8 Total PCBs in Water Column: Fort Edward
B.3-9 Total PCBs 1n Water Column: Schuylervllle
B.3-10 Total PCBs in Water Column: Stillwater
B.3-11 Total PCBs 1n Water Column: Waterford
B.3-12 Summer (June - September) Average PCB Concentrations in Water
B.3-13 Mean Total PCBs in Brown Bullhead
B.3-14 Mean Aroclor Trends 1n Fish (River Mile 175)
B.3-15 Trends in Mean Lipid-Based Aroclor Levels in Fish (River Mile 175)
B.3-16 Lipid-Based Aroclor Trends: Largemouth Bass, River Mile 175
B.3-17 Lipid-Based Aroclor Trends: Brown Bullhead, River Mile 153
B.3-18 Total PCBs in Multiplate/Caddisfly - Fort Miller
B.3-19 Total PCBs in Multiplate/Caddisfly Data, PCB-7, Stillwater
B.3-20 PCBs in Multiplate/Caddisfly - All Stations
B.3-21 PCB Trends in Multiplate/Caddisfly Data
B.3-22 Gas Chromatogram Peaks for Three Aroclor Standards
B.4-1 Conceptual Reassessment Framework
B.4-2 Upper Hudson Flow Duration Curve
B.4-3a Comparison of Estimated and Measured Flows at Hadley
B.4-3b Annual Maximum Daily Flows Below Sacandaga River
B.4-4 Suspended Sediment Rating Curve: Fort Edward at Rogers Island,
1975-1989
xvii
-------
PART B
CONTENTS
(continued)
B.4-5 Suspended Sediment Rating Curve: Hudson River at Schuylerv1llซ
B.4-6 Suspended Sediment Rating Curve: Hudson River at Stillwater
B.4-7 Suspended Sediment Rating Curve: Hudson River at W. terford
B.4-8 Sediment Load, Hudson River at Fort Edward
B.4-9 Sediment Load, Hudson River at Schuylervllle
B.4-10 Flows at Fort Edward and PCBs at Fort Edward
B.4-11 Flows at Fort Edward and PCBs at Schuylervllle
B.4-12 Total PCBs 1n Water vs. Flow: Fort Edward
B.4-13 Total PCBs 1n Water vs. Flow: Schuylervllle
B.4-14 Total PCBs 1n Water vs. Flow: Stillwater
B.4-15 Total PCBs 1n Water vs. Flow: Waterford
B.4-16 Suspended Sol Ids vs. Total PCBs: Stillwater
B.4-17 PCB Load at Non-Scour1ng Flows, Stillwater, 1983
B.4-18 Flow-PCB Observation Pairs: Stillwater
B.4-19 PCB Mass Transport Corrected Hean Method Estimates
B.4-20 PCB Mass Transport Past Waterford: Corrected Mean Estimates
B.4-21 PCB Mass Transport, Fort Edward and Stillwater: Corrected Mean
Estimates
B.4-22 Estimated PCB Load Past Waterford
B.4-23 Aroclor 1016 1n Largemouth Bass (Lipid): River Mile 175
B.4-24 Simulated Average Total PCBs 1n Fish: Upper Hudson River, 1991-2020
xviil
-------
PART B
CONTENTS
(continued)
B.4-25 Total PCBs in Yearling Pumpkinseed vs Summer PCB Concentrations In
the Water Column at Stillwater
B.4-26 Aroclor Levels in Yearling Pumpkinseed vs. Sumner Water-Column Total
PCBs
B.4-27 Total PCBs 1n Yearling Pumpkinseed vs. Summer PCB Concentrations in
the Water Column at Schuylerville
B.4-28 Total PCBs in Largemouth Bass vs. Summer PCB Concentrations in Water
Column at Stillwater
B.4-29 Total PCBs 1n Brown Bullhead vs. Sumner PCB Concentrations in Water
Column at Stillwater
B.5-1 Model Nodes and Links
B.5-2 Nodal Areas
B.5-3 Preliminary Hydraulic Calibration, 1-D Model, Thompson Island Pool
B.6-1 Potential Exposure Pathways
PLATES
Plates Located in Volume 1 (2 of 2)
B. 11
Upper
Hudson River USGS Monitoring Stations
B.l-2
Upper
Hudson River Water Surface Profile
B.l-3
Upper
Hudson River Surface Water Classifications
B.l-4
Upper
Hudson River Land Use
B.3-1
. Upper
Hudson River Sediment Core Locations
xix
-------
PART B
CONTENTS
(continued)
PAGE INTENTIONALLY LEFT BLANK
xx
-------
SYNOPSIS
PHYSICAL SITE CHARACTERISTICS
(Section B.1)
The hydrology of the Upper Hudson River is described in detail (B.1.1), building upon
the basic description oif Hudson River hydrology given in Part A. Flow characteristics of the
Upper Hudson generally show a strong seasonal dependence, with maximum flows during the
annual spring thaw. Flows are partially regulated by wetlands as well as the Sacandaga
Reservoir. Four major tributaries to the Upper Hudson below Fort Edward combine with flaw
upstream to produce an average annual flow of 7100 cfs at Waterford, above its confluence with
the Mohawk. This basic flow regime governs the transport ofPCBs in the Upper Hudson.
Water quality (B.1.2) is described according to New York State water quality
classifications assigned on the basis of "best usages" and the results of water quality sampling
data from a 1987-1988 survey. The water quality of the Upper Hudson at Fort Edward and
Schuylerville was rated as poor, partly because of the fishing ban, as a result of historic PCB
discharges.
Only the town of Waterford draws its drinking water from the Upper Hudson below Fort
Edward. More commonly, the river water is used for industrial and commercial purposes, such
as power-generation, and for domestic and agricultural use, such as watering lawns, gardens or
crops. The Upper Hudson River is a navigational waterway; from Waterford to Fort Edward,
it is co-incident with the Champ lain Canal
Population in the four counties bordering the Upper Hudson between Albany and Glens
Falls is over half a million and land use is predominantly agricultural (B.1.3). Dairy farming
is the principal form of agriculture. The region is also host to a number of industries, generally
located in the vicinity of the population centers.
The Upper Hudson River represents a diverse fisheries resource (B.1.4); six fish surveys
from 1933 to 1985 are reviewed. In general, these surveys show that the majority offish species
historically found in the Upper Hudson continue to reside there. The data also indicate a
qualitative improvement in the fisheries resource.
-------
PAGE INTENTIONALLY LEFT BLANK
-------
B. UPPER HUDSON CHARACTERIZATION
B.l Physical Site Characteristics
B.l.l Hydrology ^
The Upper Hudson River flows southerly from Its source at Lake Tear-of-the-
Clouds near Mt. Marcy in the Adlrondacks to Its confluence with the Mohawk River
near the Federal Dam at Green Island, Troy, NY. The drainage area of this
segment, shown 1n Plate A.1-1, 1s 4,630 square miles (Wagner, 1982). (Sone
discussion of overall Hudson River basin hydrology, presented 1n Part A, Is not
repeated here.)
The Upper Hudson River drains a major portion of the southern and central
Adlrondacks. Along Its course, the main channel Is Intersected by several
tributary branches, the most significant of which are the Sacandaga River, the
Batten Kill, the Fish Creek and the Hooslc River (see Plate B.l-1). Water flow
In this segment 1s regulated by several dams on the Hudson Itself (see Plate B.l-
2), as well as on tributary branches. Flow 1s further controlled by abundant
wetlands located throughout the basin, which act as a buffer for high and low
flow conditions.
The total mean annual fresh water flow from the Upper Hudson at Its
confluence with the Hohawk near Waterford Is about 7,100 cfs. This flow
represents more than a two-fold increase 1n flow from that at Hadley, NY. Before
the river reaches the Bakers Falls - Fort Edward area, It Is joined by the
Sacandaga River, the largest single tributary In this area. The mean annual flow
at Fort Edward is roughly 3,800 cfs, about 54 percent of the total flow at
Waterford. Downstream of this location, the remaining tributaries are fairly
evenly spaced, at roughly 10 to 15 miles between tributary junctions. The
combined total of these tributaries doubles the flow of the Upper Hudson by the
time It reaches Waterford. Of particular Importance 1s the Hooslc River, which
represents about 15 percent of the total drainage area south of Hadley.
B.l-1
-------
Flow 1n the Hudson Basin is seasonally dependent, with flow patterns
similar to those seen throughout the basin. The typical regime Is one of fairly
steady flow throughout nine months of the year. During the spring, flows 1n the
Upper Hudson Increase substantially 1n response to the melting of winter snow.
The maximum flow at Waterford occurs one month before the maximum at Fort Edward;
melting of winter snows in the southern portion of the Upper Hudson basin tends
to occur earlier than 1n the portion of the Upper Hudson basin above Fort Edward.
Figure B.l-1 shows mean monthly flows for water year 1986 at Fort Edward and
Waterford, respectively. This seasonal pattern Is duplicated 1n the flows at the
Federal Dam at Green Island for the same year.
Flows for an atypical water year (1984) are shown in Figure B.l-2 for Fort
Edward and Waterford, respectively. That water year was characterized as having
an unusually warm winter with many major storm events throughout the year. The
flow patterns at Fort Edward and Waterford 1n that year were also present at the
Federal Dam and on the Wallkill River, Implying that these unusual conditions
were felt throughout the entire Hudson basin.
B.1.2 Water Quality and Use
B.1.2.1 Water Quality
New York State has classified its surface waters according to "best usages"
and has established numerical water quality criteria (standards) to which those
waters should conform. Waters that conform to the numerical criteria are
considered suitable for their Intended best use. Water quality classifications
applicable to the Upper Hudson are listed In New York State's environmental
regulations (6NYCRR700, et seq.) and are Illustrated on Plate B.l-3. In summary,
the Upper Hudson has assigned to Its different reaches the following classes and
uses:
Class A - drinking water (water supply)
Class B - primary contact recreation (swimming), fishing, fish propagation
Class C - fishing, fish propagation, swimming
Class D - fishing, fish passage, swimming
B.l-2
-------
Numerical quality standards for each of the above classifications are found
in the State's rules at 6NYCRR700-705. A recent NYSDEC Technical and
Operational Guidance Series (TOGS) memo (September 25, 1990) has augmented the
rules, particularly with regard to toxic constituents. The standards encompass
conventional pollutant parameters, such as conform levels, dissolved oxygen,
turbidity and pH, as well as toxic constituents, such as heavy metals and
organlcs. The toxic substance standards have been derived from health and
environmental risk assessments performed either at the state level or derived
from those performed by USEPA.
Historically, the state has used Information from diverse sources to
ascertain the condition of Its surface waters and to evaluate their attainment
of the designated best uses. Currently, a program of rotating Intensive basin
studies (RIBS) exists whereby NYSDEC monitors all surface waters on a six year
cycle and uses data from other programs to provide continuity of Information when
RIBS sampling 1s not occurring. An Intensive basin survey was conducted within
the Upper Hudson during 1987 and 1988 and Initial reports from that effort have
now become available (NYSDEC, 1990).
As part of the Upper Hudson RIBS effort, NYSDEC evaluated water col win
conditions, toxics 1n bottom sediments and contaminant uptake by macroln-
vertebrates and fish. Table B.1-1 provides the six main parameters/media used
1n the RIBS program to rate water quality. Samples were collected at five
locations along the river's main stem at North Creek, Corinth, Fort Edward,
Schuylervllle and Waterford. Although NYSDEC did not report direct conclusions
concerning attainment of water quality standards 1n Its RIBS document, a
qualitative evaluation of overall conditions within particular river reaches 1s
provided. Conclusions pertinent to those reaches Incorporating the Fort Edward
and Schuylervllle sampling locations are summarized here.
Both the Fort Edward and Schuylervllle water column samples exhibited
elevated copper levels at a sufficient frequency to warrant considering copper
a parameter of concern. Similarly, Iron was found to be a parameter of concern
at Schuylervllle. No other trace constituents were detected at elevated
B.l-3
-------
concentrations with sufficient frequency to be considered parameters of concern.
Copper, a pervasive constituent of Upper Hudson River water, 1s also found In
samples from the relatively pristine North Creek and Corinth river reaches. While
sediment samples were obtained and analyzed from the Schuylervllle location, su<^i
samples were not obtained at Fort Edward. According to RIBS, Schuylervllle
sediments were at the upper limit of background for cadmium and were slightly
above background for lead and mercury. NYSDEC did not Identify the background
levels upon which their conclusions were based.
The 1987-1988 RIBS program found no water column toxicity to Cer1odaphn1*
at either Fort Edward or Schuylervllle. Tissue from caddlsfHes collected at the
two sites did not exhibit elevated levels of either heavy metals or of PCBs, 1n
contrast to previous studies. Apparently macrolnverterbrate tissue collected In
1987 at Hudson Falls (location not specified) did exhibit elevated lead and
manganese levels. At both Fort Edward and Schuylervllle, NYSOEC assessed overall
water quality as poor, based on the RIBS results and the fishing ban, a result
of historic PCB discharges to this river reach. At one point the RIBS document
concludes that water quality 1s rated as poor primarily due to the fishing ban
(NYSDEC,1990,p.77).
As a result of several amendments to the Clean Water Act, states are
required to report 1n specific terms conditions of their surface waters. Clean
Water Act Section 305(b) mandates that states submit water quality condition
reports to USEPA every two years. The 305(b) report evaluates surface waters 1n
relationship to their ability to sustain primary contact recreation and fish
propagation uses. In addition, Clean Water Act Section 304(1) requires that
states generate lists of surface waters that fall to meet water quality standards
because of toxic pollutants, 1n general, and toxics from point sources, in
particular.
New York's most recent 305(b) report was published 1n April 1990. That
document provides, 1n part, a summary of waters wherein contamination in fish
exceeds either FDA levels or other guidelines. For the Upper Hudson reach from
Hudson Falls to Federal Dam, the only contaminant Identified as exceeding either
B.l-4
-------
the FDA levels or other guidelines 1s PCB (NYSDEC, 1990, Table 20). The 305(b)
report also Identifies PCB 1n sediment as the sole toxic that Is responsible for
use impairment 1n the Upper Hudson (NYSDEC, 1990, Table 17). Similarly, the
state's 304(1) lists identify priority organ1cs (PCBs) as being the toxic
responsible for Upper Hudson use Impairment.
B.1.2.2 Use
The Hudson River 1s used as a source for public water supplies (municipal
and Institutional drinking water) 1n sections of the river classified as Class
AA or A. Along the Upper Hudson, three communities draw directly Hudson River
water. Of these, Queensbury and Waterford have current average uses of more than
1 mgd (NYSDOH, 1991), as shown 1n Table B.1-2. The Waterford Intake Is located
at the base of the Upper Hudson Basin near Lock 1. The Queensbury Intake is
located near Sherman Island Dam 1n Warren County. The Wlnebrook Hills Water
District, the third Upper Hudson water supply drawing from Hudson, 1s located at
the headwaters of the Hudson 1n Newcomb, Essex County.
A more common use of Hudson River water 1s for Industrial and commercial
purposes such as cooling, manufacturing processes and fire protection. Hudson
River water 1s also extensively used for hydroelectric and thermal power
generation. An inventory of facilities and plants that utilize Hudson River
water can be found 1n reports for the Hudson River-Black River Regulating
District (Malcolm Pirnie, 1984a) and for the NYSD0T (1984).
Hudson River water is also used for domestic (watering lawns and gardens)
and agricultural purposes (irrigating crops). There are currently no records of
water withdrawal for agricultural uses. Unlike the other Intakes, permits are not
needed to withdraw water from the Hudson for irrigation purposes (pers. conn.,
NYSDEC and NYSDOH, 1991).
The NYSDEC Division of Water, Source Surveillance Section provided a
listing (March 7, 1991) of all significant active facilities With SPDES permits
In the Upper Hudson River Basin. This search revealed that 27 facilities
B. 15
-------
discharge into the Upper Hudson Basin, with 15 discharging directly Into the
Hudson River. Five of the 15 facilities are municipal wastewater treatment
plants, Including Corinth Sewage Treatment Plant, Glens Falls Wastewater
Treatment Plant, Saratoga County Sewer District #1 Wastewater Treatment Plant at
Mechanlcvllle, Stillwater Sewage Treatment Plant and Washington County Sewer
District #2 Water Pollution Control Plant at Fort Edward.
The Champlaln Canal Is coincident with portions of the Hudson River; It
extends from Waterford, New York on the Hudson River to Whitehall at the southern
end of Lake Champlaln. The Champlaln Canal 1s part of the New York State Barge
Canal System, also comprised of the Erie Canal, Oswego Canal and Cayuga-Seneca
Canals. This network of waterways connects the Atlantic Ocean with the Great
Lakes and the Saint Lawrence Seaway. The Champlaln Canal 1s 60 miles in length,
Including 37 miles of canalized Hudson River from Waterford to Fort Edward and
23 miles of land-cut sections. The canal diverges from the river at Fort Edward
just below Lock 7 and proceeds 1n a northeasterly direction to Lake Champlaln.
Additional land-cut areas exist at Stillwater, Northumberland and Fort Miller.
Natural flows provide a considerable portion of the water supply needs for
the Hudson River portion of the Champlaln Canal. The Hudson River at Fort Edward
provides an average discharge of 5,244 cfs (USGS average for 1979-1990 water
years) to the canal at the confluence below Lock 7. This flow Is significantly
Influenced by regulation of flows from Great Sacandaga Lake. Below this point,
the river 1s canalized to provide natural flows for the canal. Water Is also
supplied to the Champlaln Canal via the Glens Falls Feeder Canal. The Feeder
'Canal diverts approximately 100 cfs of water from the Hudson River upstream of
the Feeder Dam west of Glens Falls to the Champlaln Canal Summit Level between
Locks 8 and 9 on the canal near Smith's Basin. The Summit Level 1s the point of
highest elevation on the canal and allows for the gravity flow of water In both
directions as lockages are made (south to the Hudson River, north to Lake
Champlaln). Approximately 25 percent of the diverted water returns to the Hudson
River (Halcolm P1m1e, 1984).
B. 1-6
-------
Commercial traffic has declined on the Champlaln Canal and other canals In
the Barge Canal system as a result of "the unreliability of the syste* for
waterway transport" and competition from other modes of transportation for bulk
products (Malcolm Pirnle, 1984b, US Ars\y Corps of Engineers, 1977). Unlike the
other three canals In the system, the Champlaln Canal shows a steady decline 1n
recreational use along the entire stretch of the canal on both the canalized
Hudson River and the land-cut section north of Fort Edward (Malcolm Pirnle,
1984).
*
B.1.3 Population and Land Use
Four counties (Albany, Washington, Rensselaer and Saratoga) 11e adjacent
to the Upper Hudson between Albany and Glens Falls. All counties experienced
growth between 1980 and 1990 with Saratoga having the greatest increase over the
period and Rensselaer the lowest. Total population of these counties In 1990 was
over 500,000.
Land use within a zone adjacent to the Upper Hudson River, depicted on
Plate B.1-4, is mostly agricultural. Portions lie within New York State
Agricultural Districts and include parcels considered to be prime farmland.
Dairy farming 1s the major agricultural Industry. The majority of the crops
grown, such as corn and hay, are used for forage; small quantities of cash crops,
such as oats and wheat, are produced. Industrial use 1s typically located near
population centers. Major non-agricultural Industries within the study area
Include: an Industrial demolition company; several paper mills; hydroelectric
plants; a grocery warehouse; and manufacturers of garden equipment, brake
linings, brushes, paints, wallpaper, paper products, gun barrels, silicone
products, abrasives, brass fittings and clothing. Forested and recreational land
uses are scattered.
Existing recreational uses Include Schaghtlcoke Canal Park at Lock 4 of the
Champlaln Canal and two town parks, which lie along the river In Fort Edward.
Proposed for the Fort Edward area are a marina, to be located on the south end
of Rogers Island, and a marina, trails and picnic areas to be located one mile
B. 17
-------
south of Fort Edward on the former Champlaln Canal. Saratoga National Historic
Park lies on the western bank of the River 1n the Town of Stillwater. Moreau
State Park 1s located south of Glens Falls. At the confluence of the Mohawk and
Hudson Rivers are Peebles Island State Park and the Van Schalk Island Counter
Club. Several parks and/or country clubs also front the River.
B.1.4 Fisheries
Fishery resources within the Hudson River from Federal Dair to Fort Edward
are Influenced by different physical features, as well as man-made structures
such as locks, dams, guard gates, bridges, spillways and submerged power lines
and cables. This array of physical features produces the following variety of
different fish habitats:
Outlets of streams and rivers;
Shallow water areas (wetland and non-wetland);
Designated ship channels of the canal (canalized river);
Steep embankment areas with relatively swift current;
Landcut (artificial) portions of the canal;
Wet dumping grounds or spoil areas; and
Various alternate channels separated from the main channel
by an Island.
This wide variation In habitats expands spatial heterogeneity and results In a
complex fishery resource.
The New York State Conservation Department (Greeley, J.R. and Bishop, 1933)
conducted an early and comprehensive fish Inventory of the portion of the Hudson
River between Hudson Falls and the mouth of the Hooslc River. Forty-one species
of fish (Table B.l-3) were recorded. (The American shad, one might note, was
listed as extinct.) The historical data developed as part of this 1933 fish
Inventory documented that there was an Imbalance between the juvenile and adult
B.l-8
-------
game fish. There appeared to be abundant juvenile game species from a few miles
below Fort Edward to the mouth of the Hooslc River, but adult populations of game
fish were uncommon.
An anonymous report, prepared by the Conservation Department 1n June 1960
(see Table B.1-3) at the request of the Stillwater Rod and Gun Club, are the only
fish data that exist from 1933 to 1960. This 1960 report contains the
observation that game fish resources declined between 1949 to 1959. Prior to
1949, angling success was supposedly satisfactory for "...bass, walleyes,
northerns and pickerel." Several theories for this decline were advanced, such
as Industrial pollution from the Glens Falls and Fort Edward areas, effluent
process changes at a Glens Falls paper and pulp mill, expanded boating activity
and over-explo1tat1on.
The next major fish Inventory was conducted by Lane (1970) who found 13
species of fish. Lane stated the somewhat remarkable conclusion that: "The
collection data Indicate the absence of any significant fishery between Lock No.
1 and Fort Edward." With the exception of goldfish, few larger fish were found.
Game fish collected 1n the river channel were represented largely by juvenile
populations. Similar to 1933 observations, Lane noted a diminished amount of
aquatic vegetation 1n the river channel. Although It 1s generally recognized
that an adequate supply of aquatic vegetation 1s Important for adult fish
population maintenance and productivity, few historical and/or current data are
available. One of the most unexpected results of Lane's survey was the absence
of common carp. Although Lane concluded that "...conditions within the study
area are not suitable for carp," later studies (Shupp, 1975; Makarewicz, 1983)
do document the presence of the common carp within this section of the Upper
Hudson.
Subsequently, NYSDEC began collecting fish for PCB analysis. Shupp (1975)
collected fish samples 1n the 40-m1le stretch of the Upper Hudson from Lock No.
1 to Hudson Falls and found an Improvement, which he attributed to upgrading of
treatment facilities and tougher regulations concerning Industrial discharges.
Although Shupp reported approximately 24 species of fish compared to the 13
B. 1-9
-------
species listed by Lane (1970), the fishing from Lock No. 1 to Hudson Falls was
still considered poor, because of the overall low standing crop of fish and low
numbers of adult fish compared to juveniles (Shupp, 1975). The reported
preponderance of juvenile fish was similar to data from the 1933 and 1970
surveys. Sheppard (1976) Indicated that "...some unknown factor Is causing the
exodus or demise of the mature segment of certain fish populations Including the
rock bass, pumpklnseed, yellow perch, walleye and chain pickerel." NYSDEC (R.
Sloan, per. comm., 1991) has recently observed a greatly diminished number of
both pumpklnseed and yellow perch populations during routine PCB assessments of
resident fish 1n the Upper Hudson (Fort Edward to Federal Dam).
Since 1975, NYSDEC has continued to collect fish between Federal Dam and
Fort Edward as part of their ongoing assessment and monitoring of PCB levels 1n
fish flesh. The principal species collected and analyzed within this reach have
been the brown bullhead, goldfish, largemouth bass, pumpklnseed and yellow perch.
Because of the demise of the yellow perch and goldfish, current collection
efforts have focused on the brown bullhead, common carp and largemouth bass (R.
Sloan, per. comm.).
One of the most extensive fishery surveys since the 1933 survey was
conducted approximately eight years ago by Makarewlcz (1983). He surveyed 85
stations along the entire length of the Hudson between Federal Dam and Whitehall
as part of the New York State Barge Canal Maintenance Dredging Program 1985-1995
for NYSDOT (Malcolm Plrnle, 1984b). The sampling stations Included nine sampling
reaches from Federal Dam to Fort Edward. A total of 46 species, Including four
migratory species (American eel, blueback herring, sea lamprey and striped bass),
were found. Of the 42 resident freshwater species, the panflsh ere the most
prevalent (40 percent); demersal fish were second In abundance (22 percent);
forage fish were the third most abundant group (14 percent); and game fish had
the lowest relative abundance (9 percent). Dominant panflsh members were
blueglll, pumpklnseed, rock bass and yellow perch; demersal dominants were black
bullhead and brown bullhead, common carp and white sucker; forage dominants were
golden shiner, spotfln shiner and spottall shiner; and game fish dominants were
B.1-10
-------
the largemouth bass and smallmouth bass. Collectively, these 13 species
accounted for 85 percent of all resident freshwater species collected.
The most recent fish survey data available are from a study conducted by
Green (1985) between Stillwater (Lock 4) and Schuylervllle (Lock 5), covering
approximately 13 river miles. Fewer overall species were taken (20) compared to
the more Intensive biological surveys 1n 1933 (41 species) and 1983 (46 species),
both of which covered more of the river (Fort Edward to Federal Dam) than that
covered 1n 1985 (Stillwater to Schuylervllle).
Length at age comparisons for both smallmouth and largemouth bass Indicate
that growth rates were comparable to the New York Bass Study's average to fast
growth rates (Green, 1985), an Indication that some of the historical observa-
tions regarding the preponderance of juvenile fish and the paucity of adult game
fish may no longer be valid for bass populations.
As shown 1n Table B.1-3, the list of species of fish 1n the 1983 study
agrees quite well with the 11st for 1933. Although comparative percent
contributions of the dominant game fish, panflsh, forage fish and demersal fish
with those species recorded 1n 1933 1s not possible, because quantitative
Information are lacking 1n the historical study, all the above dominant species
(with the exception of the spotfln shiner and black bullhead, see note 1, Table
B.1-3) were also recorded In 1933. In addition, 31 species were similarly
reported 1n both studies. An analysis of these two studies (Greeley and Bishop,
1933 and Makarewlcz, 1983), which have spanned nearly 50 years, reveals
considerable qualitative similarity of the fishery within the reach from Federal
Dam to Fort Edward.
The construction of Federal Dam and various locks as part of the Champlaln
Canal Section of the New York State Barge Canal System blocked major upstream
spawning migrations for a number of anadromous species, Including the American
shad, alewlfe, blueback herring, sturgeons and striped bass. Some migrants, as
documented by Smith and Lake (1990), may be found periodically upstream of the
B.1-11
-------
Federal Dam and Lock. Population pulses nay enter and leave the lower region of
the Upper Hudson through the Interconnecting system of locks.
The majority of fish species listed in Table 6.1-3 are freshwater resident
of the Hudson River. Some migratory species, such as the striped bass, blueback
herring, sea lamprey and American eel, still attempt to utilize sections of the
Upper Hudson as migratory routes. With the exception of black bullhead, johnny
darter, pearl dace and northern redbelly dace, all the fish species listed have
also been found 1n various regions of the Lower Hudson River (Beeb'e and Savldge,
1988; Smith and Lake, 1990). Whereas many of the fish species are year-round
freshwater residents, they are not unique to the Federal Dam/Fort Edward section
of the Upper Hudson.
Some additional fish species, which were not found during the reported fish
surveys summarized 1n Table B.l-3, have been reported 1n various sections of the
entire Upper Hudson Region (Smith and Lake, 1990). These include:
American shad (Alosa sapldissiaa)
Stonecat (Noturus flavus)
Longnose sucker (Catostoaus catostomus)
Lake chub (Coueslus plumbeus)
Brassy minnow (Hybognathus hankinsonl)
Blacknose shiner (Notropis heterolepis)
Finescale dace (Phoxlnus neogaeus)
Lake herring (Coregonus artedi)
Lake whiteflsh (Coregonus cTupeaforals)
Round whiteflsh (Prosoplun cyllndraceum)
Lake trout (Salvellnus nanaycush)
Rainbow smelt (Osaerus mordax)
Tessellated darter (Etheostoaa olmstedl)
Although not found during the major fish surveys conducted within the Federal
Dam/Fort Edward Region, the American shad has been known to occur within this
particular region (per. comm., R. Sloan). Others such as the lake chub, lake
herring, lake whiteflsh and lake trout are not commonly found 1n freshwater
riverine systems and are not expected to occur to any great extent within this
section of the Upper Hudson.
B. 1-12
-------
All studies reviewed to date Indicate that the majority of species
historically present 1n the lower section of the Upper Hudson continue to reside
1n this particular reach of the Hudson River. According to Information submitted
by Shupp (1987), the section of the Upper Hudson River between the Federal Dam
and Fort Edward can support a diverse and high quality fishery resource. Vshupp
also cited evidence gathered from some NYSOEC studies between Mechanlcvllle and
Schuylervllle, which suggested a "vast Improvement" 1n small mouth and largemouth
bass stocks and other fish species from the early 1960s to the late 1980s. Shupp
has stated, "Since 1984, the greatly Improved warm water fish community 1n the
Fort Edward to Troy (Upper Hudson) reach has stimulated Interest 1n reopening the
fishery."
Angler reports to The Warrensburg F1sh Management Unit Indicate a somewhat
Improved fishery for bass, yellow perch, black crapple and brown bullhead from
1969 to 1975 (Shupp, 1975). Analysis of the data presented by Lane (1970),
Makarewlcz (1983) and Green (1985) and prepared testimony statements by Shupp
(1987) also suggest a qualitative Improvement within the past twenty years.
B. 113
-------
i
PAGE INTENTIONALLY LIFT BLANK
B.1-14
-------
SYNOPSIS
SOURCES OF PCB CONTAMINATION
(Section B2)
General Electric discharged PCBs from plants at Fort Edward and Hudson Falls between
1946 and 1977 (B.1.1). The total amount of PCBs released during 1957 to 1975, a period for
which estimates can be made using historic data, ranges from 209,000 to 1,330,000 pounds.
<
Currently, six New York State facilities are permitted to discharge PCB-cantaminated
waste water to the basin of the Upper Hudson River (B.Z2). A facility in western Massachusetts
discharges to a Hudson River tributary, the Hoosic River.
Other potential sources of PCBs to the Upper Hudson (B.3.3) are discussed
-------
PAGE INTENTIONALLY LEFT BLANK
-------
B.2 Sources of PCB Contamination
B.2.1 6E Discharges (To 1977)
*
The two 6E plants at Fort Edward and Hudson Falls, New York began to use
PCBs 1n 1946 and discontinued their utilization In 1977. In-plant sources of PCB
discharges have been characterized as both minor spills and effluent from washing
capacitor cans, with the latter being the major source. Capacitor cans were
flood filled with dielectric fluid and then washed with detergent and water to
remove excess material. Contaminated wash water was finally discharged,
untreated, to the Hudson River (Brown, Jr. et ซ/., 1984).
Estimates of PCB releases at the two capacitor plants have been made on the
basis of GE's overall usage of the chemical and considering discharges allowed
pursuant to USEPA's discharge permit for the facilities. Figure B.2-1
Illustrates the company's PCB usage, by Aroclor type, for the period 1946 to
1977. That figure shows the trend 1n Aroclor usage to be from relatively highly
chlorinated forms 1n the mid-1950s (Aroclor 1254) to less chlorinated homologues
in the 1960s (Aroclor 1242) and the 1970s (Aroclor 1016).
By using actual GE purchase records for the years 1966 thru 1975 and
approximating GE purchases for the prior period on the basis of Monsanto's
production records, GE's total PCB consumption for the period 1957 to 1975 has
been estimated at 133,000,000 pounds (Umburg, 1985). Plant discharges during
the 1960s have been approximated at 5 metric tons per year (Sofaer, 1976), a rate
which is roughly compatible with that allowed by GE's 1975 discharge permit (30
pounds per day or about 11,000 pounds per year). Sanders (1989) provides
anecdotal evidence of plant releases being less than one percent of plant
consumption or less than 1,330,000 pounds from 1957 to 1975. Thus, one estimate
of the range of releases to the river would be 209,000 to 1,330,000 pounds over
the period 1957 to 1975, where the lower quantity is based on a continuous
discharge of 30 pounds per day (or 5 metric tons/year) for a nineteen-year
period.
B.2-1
-------
B.2.2 Current Permitted Discharges
Six facilities 1n New York State, Including GE, are permitted to discharge
PCBs 1n the Upper Hudson Basin. NYSDEC (March 19, 1991) provided SPPDES permits
and discharge monitoring reports (PCBs only) for each of these facilities. Table
B.2-1 identifies these facilities, receiving waters and relevant Information on
PCB limits and measurements. Two facilities (GE, Fort Edward and James River
Corporation, South Glens Falls Mill) are permitted to discharge PCBs directly
Into the Upper Hudson River, while one (GE, Old Fort Edward Site Remediation
Project) discharges Into the Old Champlaln Canal 1n the vicinity of Fort Edward.
In most cases, the concentration of PCBs In the final effluent 1s limited to the
minimum reliable detection 11m1t based on USEPA Method 608.
According to available Commonwealth of Massachusetts SPDES records, Sprague
Electric Company 1s permitted to discharge PCBs (0.01 mg/7) directly Into the
Hooslc River, which flows Into the Upper Hudson.
B.2.3 Other Sources
Table B.2-2 Identifies Inactive hazardous waste disposal sites located near
the Upper Hudson River (above Federal Dam at Troy) In which PCBs have been
dumped. This tabulation was obtained from NYSDEC, Division of Hazardous Waste
Remediation, utilizing their annual inventories of disposal sites 1n New York
State (April 1990). The NYSDEC priority classification codes stated In the table
are: Code 2 - Significant threat to the public health or environment and action
1s required; Code 2a - Temporary classification assigned to sites with
Inadequate or Insufficient data for Inclusion 1n any of the other classes; Code
3 - Does not present a significant threat to the public health or environment,
and action may be deferred; Code 4 -Site 1s properly closed and requires
continued management.
The release of some contaminants from these Inactive sites adds to the
total PCB loadings 1n the Upper Hudson. In many Instances, flow of surface water
and groundwater from the sites are towards the Hudson River. "Unknown" material
B.2-2
-------
was listed as being disposed of at many sites and 1s generally not Included 1n
the table. None of the sites identified by NYSDEC are classified as an Imminent
danger to the public or environment (Code 1). Many sites are classified as Code
2, suggesting that these sites may be potential sources of pollutlom to the
Hudson River.
B.2-3
-------
PAGE INTENTIONALLY LEFT BLANK
B.2-4
-------
SYNOPSIS
NATURE AND EXTENT OF CONTAMINATION
(Section B3)
Available environmental data on the distribution ofPCBs in the sediments, water, fish,
air and plants of the Upper Hudson River as well as supporting data on flow and sediment
transport are summarized and evaluated. As a foundation for continued analyses, available
data have been compiled into a computerized, relational database management system (B.3.1).
Data on PCB concentrations in river-bottom sediments (B.3.2) are drawn primarily from
the 1976-1978 NYSDEC sampling efforts and the 1984 Thompson Island Pool investigation,
along with several other sources. Sediments are the major environmental repository for PCBs
in the Upper Hudson, but there is a high degree of spatial variability in PCB concentrations.
The 1984 study covered only the Thompson Island Pool and relatively little data have been
collected since. It is difficult to determine the current mass and distribution of PCBs in
sediments without further investigation.
The discussion of surface water monitoring (B.3.3) concentrates on data collected by the
USCS. Transport of PCBs is affected by hydrologic processes, particularly flood events. A
discussion of flow monitoring is followed by presentation of time series data, to the extent
available, for suspended sediment and PCBs in the water column. Current full-year and
summer average PCB concentrations are calculated, taking into account the problem of
numerous measurements below analytical detection limits.
NYSDEC has monitored Upper Hudson fish on a regular basis since 1975; data are
presently available for PCBs in fish through 1988. The extensive data collected in this program
(nearly 3,000 Upper Hudson samples) are discussed (B.3.4). Total PCB burdens in fish declined
sharply from 1978-1981. Levels of the higher chlorinated congeners in fish appear to have
remained relatively constant since 1982. Results ofNYSDOH macroinvertebrate monitoring are
also described.
PCB monitoring data for air and plants near the Upper Hudson (B.3.5) are generally
insufficient to assess the impact from PCBs in the river. Isolating the contribution of the river
from other possible PCB sources is a particularly difficult problem.
For various other media there is a notable lack of monitoring data (B.3.6). Only limited
groundwater sampling has been performed and surface soils near the river have not been
monitored.
Data quality and analysis methods for the various monitoring programs are evaluated.
PCBs have many different variations in chemical structure and differing physical properties.
Uncertainties surrounding PCB measurement, particularly the specific variations in PCBs, results
in considerable difficulties in interpreting the results. Furthermore, differing PCB measurement
methods used for water, sediments or fish confound direct comparisons among them.
-------
PAGE INTENTIONALLY LEFT BLANK
-------
B.3 Nature and Extent of Contamination
B.3.1 Overview of Sources and Database
Site data defining the current understanding of the nature and extent of
PCB contamination, based on previous studies, are summarized in this section.
Data synthesis efforts have focused on:
obtaining the most complete, and current, data sources available;
compiling these data Into a computerized database;
evaluating the PCB data for the media sampled;
Identifying current trends and relationships; and
determining the adequacy of the existing data.
During the early 1970s, NYSDEC and several other agencies began the first
comprehensive monitoring studies for PCBs in the Upper Hudson. Fish, which were
some of the earliest environmental samples analyzed, showed high concentrations
of PCBs. These early Investigations began what is now over two decades of
studies on PCBs in water, sediments, fish and other media affected by PCB
discharges to the Upper Hudson. Table B.3-1 summarizes the major investigations.
Past USEPA documents, Including the 1984 Feasibility Study and the EIS,
have been reviewed for this work. Emphasis, however, was given to reviewing
additional, more recent data and evaluating that along with the long-tera
monitoring record contained in the previous studies.
Previous Investigations at the Hudson River PCB site have examined the
nature and extent of contamination in several media, Including fish, sedinents,
river water and, to a lesser extent, air. Each medium Is discussed below.
Additional studies, currently being performed by GE at the remnant deposit sites,
are not available for this report. Before discussing each medium, a brief
overview of the TAMS/Gradient database 1s provided below.
B.3-1
-------
As the foundation for these Phase 1 data evaluation efforts as well as
continued analyses during subsequent phases of the project and possibly future
projects, the data gathered during Phase 1 have been compiled into a relational
database management system using PC-based Paradox software. This databas^
currently contains approximately 30,000 records of information, primarily on
sediments, water, fish and some other biota obtained from numerous sources (see
Table B.3-1). Data input, verification and database management have been
conducted by the TAMS/Gradient project personnel.
In the current sediment database, there are nearly 2,500 samples for the
period 1976 through 1990. The fish database contains approximately 8,000 samples
for the period 1973 through 1988. Additional data for other aquatic biota
(macroinvertebrates and multiplate data) account for several hundred.additional
samples. Water column data, including daily and peak flow, suspended sediment
and total PCBs, comprise the bulk of the remaining data in the database. The
database contains a small number of samples summarizing PCB data for air and crop
plants.
Data for separate media are linked primarily through sample date and
location information. For each medium, data are organized such that a unique
sample identification number links information among tables. Individual data
tables are grouped by medium to contain similar kinds of information. As an
illustration of the database format developed specifically for this Reassessment
and its contents, excerpts from the sediment database tables are summarized below
1n four tables: Sample Information, Core Section, Chemical and Non-Chemical.
Each of these tables Is linked by a unique sample Identification number, e.g.,
sample ID numbers 30000, 30016 and 30032 shown here, such that sample IDs in each
table correspond to data about a single sample.
The Sample table contains Information about sample date, location (River
Mile distance from bank and northing and easting coordinates, where available),
sample type (grab versus core), the agency or investigator responsible for the
analytical method(s), the reference report/location of the original data and
other Information as available.
B.3-2
-------
Database Table Example: Sample Information Table
Sampt*
ID
Typ*
M/D/YR
Nv*r
Mil*
Fwtfr.
<|f, at
win
Bwik
Northing
(ft)
Easting
(ปป)
Sampler
Water
Depth
(ft)
B*v.
W
R*f
Ag*ncy
30000
Grab
5/21/77
188.8
330.0
1071785
100
2
O'Brien
30016
Cor*
3/18/77
188.4
100.0
1183740
868870
100
M
118.6
1
OWm
tQm
Cor*
3/18/77
183.4
eo.o
1140410
668040
100
Z2
102.4
1
O'Brien
AGere
Core samples in the Sample table are linked with the Core Section table,
which Identifies the length of each core sample section and the depth beneath the
river bottom, i.e., the depth of sample penetration for the top and bottoa of
each section.
Sediment Database Example: Cora Sactlon Table
Sampi* 10
Cor* S*otion No.
Bottom of Section (in.)
Top of Section On.)
30016
1
1
0
30016
2
2
1
30016
12
12
11
30032
1
1
0
30032
8
8
8
Selecting a sample ID from the Core Sample and Section tables and locating
the same ID 1n the Chemical data table shows either the Aroclor results for an
entire grab sample or section by section results for core samples. Additional
Information describing analytical measurement methods, I.e., extraction aethod,
are contained 1n the database as available. The Chemical data table also
contains non-PCB chemical data, such as metals analyses (not shown here), where
available.
B.3-3
-------
Sediment Databaae Example: Chemical Date Table
Sample 10
Parameter
Cor* Section No.
Extraction Method
Concentration (ppm)
30000
Arodor 1016
ahaka
1.0
90000
Arodor 1221
ahaka
1.0
3000Q
Arodor 1254
ahaka
1.0
30016
Arodor 1016
4
aoxhlet
6.0
30016
Arodor 1254
12
aoxhlat
0.1
30032
Arodor 1016
5
aoxhlet
234.0
30032
Arodor 1254
5
aoxhlat
163.0
Finally, non-chemical data, such as sediment texture class, percent
volatile versus total solids, are contained in the Non-Chemical table.
Sediment Database Example: Non-Chemical Data Table
Sample ID
Core Section No.
Parameter
Value
90000
* total aoOda
78.83
3Q000
% volatile solid*
0.85
30016
1
texture
GRAVEL
30016
4
% total sotkJ*
85.97
30016
4
% volatile aolkfa
Z2
ซ
30016
12
% total aoUda
80.23
1
texture
CL-WC
90032
5
* volatile aoUda
25.39
B.3-4
-------
B.3.2 Sediment
Two primary sources of Information provide the largest amount of data on
sediment contamination 1n the Upper Hudson: (1) printed results (>1,000 samples)
of the 1976-1978 NYSDEC sampling survey and (2) computer files (>2,000 samples)
of the 1984 Thompson Island Pool survey. These data were entered into a
computerized database (referenced here as the TAMS/Gradlent database). Tables
6.3-2 and B.3-3 provide a summary of these sediment samples 1n the TAMS/Grad1ent
database. In addition to these two major NYSDEC sediment surveys, USEPA, the
Lamont-Doherty Geological Observatory and GE have sampled sediments in the Upper
Hudson.
B.3.2.1 1976-1978 NYSDEC Sampling
As reported by Tofflemire and Qulnn (1979), NYSDEC conducted several
sediment sampling surveys in the Hudson River between 1976 and 1978. Details
about the sampling and analysis procedures for these studies are summarized in
NYSDEC Technical Report No. 56 (Tofflemire and Qulnn, 1979).
The data provided by NYSDEC contained a total of 1,167 sediment samples
(396 cores and 771 grabs) taken during 1976, 1977 and 1978; 1,770 PCB analyses
were reported for the 1,167 samples. The overwhelming majority of samples (1,091
of the 1,167 samples) from the 1976-78 data set were collected 1n the Upper
Hudson River. Only five samples in this data set were obtained in the Lower
Hudson River and all of these five were from River Hile 153, just south of the
Troy Lock. Another sample 1n this set was Identified as from the Lower Hudson,
but other descriptions placed it 1n the Upper Hudson, while 70 samples had no
Information regarding location.
Aroclors 1016, 1221 and 1254 were identified as the PCB mixtures detected
1n the 1976-1978 sediment sampling effort. Total PCB concentrations were
reported as the sum of these three aroclors. Analytical detection limits were
not reported in this data nor was any Indication given about a sample's
detectable or non-detectable concentrations of PCBs. Because several concentra-
B.3-5
-------
tions (1 ppm, 5 ppm, 10 ppm) occur an inordinate number of times in this data
set, these concentrations are the probable detection limits for these samples.
Table B.3-4 summarizes the Aroclor concentrations for grab and coi^
samples, respectively, within each of the nine river reaches sampled. In
summarizing the mean and median PCB values for the core samples, each core was
counted as a single sample, /.e., the statistics are calculated over the entire
core length. The reported N is the count of core sections, not of samples. One
of the most striking aspects of the 1976-1978 results is the enormous variation
of PCB concentration over very short distances. Samples taken only a few feet
apart may have PCB levels varying by orders of magnitude. This extreme
variability is highlighted in Figure B.3-1, which plots total PCB concentration
in surface sediments (grab samples and the top section of all cor&s) by river
mile. Some trend with river miles is shown by the median. The median PCB
concentration by reach is highest 1n the Thompson Island Pool (River Mile - 188 -
194), decreasing for several miles downstream, then showing an increase from
River Mile - 175 to River Mile - 160. This pattern points to a relative scarcity
of depositlonal areas between Stillwater and Northumberland.
Based on the results of the 1976-1978 survey, NYSDEC identified 40 hot-
spots, areas containing more than 50 ppm total PCBs. The 1984 Thompson Island
Pool survey re-evaluated the hot-spot locations and revised the designation to
a series of over 100 polygons containing PCBs greater than 50 ppm. (The "hot-
spot" term 1s used here only as a frame of reference to earlier studies and
sediment areas previously defined as containing high levels of PCBs.) Subsequent
to the 1976-1978 study, questions about the adequacy of the data were raised
(NUS, 1984), as noted below:
PCB concentrations exhibited wide spatial variability; samples along
single transects not uncommonly ranged from non-detectable PCB
concentrations to greater than 1,000 ppm.
t The sampling density was so low that, given the extreme spatial
variability of the PCB concentrations, the accuracy of the hot-spot
delineation is questionable.
B.3-6
-------
Changes 1n sediment deposits caused by the dynamics of the river
greatly complicates comparisons between PCB concentrations In
similar locations in different years.
A major flood event that occurred 1n 1979 redistributed sediments
significantly, again calling into the question the usefulness of the
1976-1978 sediment data, other than general purposes.
The analytical techniques used to quantify PCBs have improved since
this data was collected.
B.3.2.2 1984 NYSDEC Sampling
Methods and Procedures
In 1984, NYSDEC again undertook an extensive sediment sampling program.
This effort focused on the Thompson Island Pool (M. P. Brown et a/., 1988b). The
objective of this study was to Identify areas of contaminated sediments that
would be removed during the Hudson River PCB Reclamation Demonstration Project.
Primarily these areas were the 20 hot-spots previously Identified in the Thompson
Island Pool and other areas with known or suspected high PCB concentrations.
The investigators identified 1,260 sampling locations in the approxinately
five-mile reach of the river. Many of these locations were determined by
imposing a 125-foot triangular grid on previously defined hot-spots and areas
that had PCB concentrations in excess of 50 ppm during the 1983 USEPA survey
(NUS, 1984). In addition, sample locations were selected based on known or
suspected sediment deposltlonal areas, as indicated by location 1n the river and
bathymetry measurements. Sample locations 1n the field were deterained
electronically by using a microwave locating system and generally agreed with
predetermined locations.
Samples were collected by Normandeau Associates, Inc. between August 24,
1984 and November 30, 1984. In addition, 21 cores were collected during February
1-4, 1985. These later samples were taken through the ice on the river at
locations that had been inaccessible by boat.
B.3-7
-------
Whenever possible, the Investigators collected core samples. At those
locations where Insufficient sediment was available for core samples, grab
samples were attempted. At 80 locations, bedrock or coarse material precluded
either sample method. In all, 1,016 locations throughout the Thompson Island
Pool were sampled and yielded 674 grab and 408 core samples.
The depth of penetration for the core samples was fairly uniform, with an
average depth of approximately 31 Inches. The investigators divided core samples
into sections based on a desire to define contaminated layers without introducing
dilution from adjacent less contaminated layers. Sections were also chosen based
on potential dredging considerations and a need to limit the number of chemical
analyses.
NYSDOH and Versar, Inc. measured physical and chemical parameters of the
sediments collected 1n this study. NYSDOH determined lengths of cores and
sections, percent dry solids, dry specific weight (density) and textures, which
were determined visually. Versar measured percent volatile solids and performed
the gas chromatograph analyses for PCBs.
PCB Results
Versar reported PCBs as Aroclors 1242, 1254 and 1260 using the method of
Webb and McCall (1973). Although the data contained a total PCB quantification,
no mention 1s made 1n M. P. Brown et al. (1988b) of the method used to quantify
or calculate this total. Examination of the data received indicates that the
total was not simply the sum of the three Aroclor mixtures quantitated.
Wide variations 1n PCB concentrations in sediments were observed 1n the
1984 NYSDEC study throughout the Thompson Island Pool, even though sampling
concentrated on areas of known contamination. The discrepancies between means
and medians for both grabs and cores indicate that PCB concentrations have a
highly skewed distribution over the area of the Thompson Island Pool.
(Environmental monitoring results frequently exhibit this skewed pattern and a
log-normal distribution is often a good approximation of the data.) Both grab
B.3-8
-------
and core samples had significantly higher concentrations in the least chlorinated
fraction that was quantified (Aroclor 1242) than in the more chlorinated
fractions. Table B.3-5 provides Aroclor and total PCB summary statistics for
both grab and core samples. Aroclor 1242 was the predominant Aroclor reported
for these samples, with lower levels of Aroclors 1254 and 1260 also identified.
On average, total PCBs in the samples were approximately 55 ppm, with maximum
levels of >1,000 ppm detected in several samples.
Table B.3-6 presents the results of texture classifications determined by
NYSDOH. Considering grab and core samples together, Thompson Island Pool
sediments were classified most often as either gravel or fine sand, with a
significant fraction of fine sand/wood chips and clay samples, particularly for
core samples.
Because of their high adsorption (partition) coefficients, PCBs are
generally expected to associate with the organic carbon fraction of the
sediments. Although no measurements of organic carbon content were made as part
of this study, organic carbon and organic matter content, which were measured
frequently in the study as percent volatile solids, can be correlated. Thus,
comparing PCB levels with organic matter content (volatile solids) provides a
surrogate for comparing PCB concentration to sediment organic carbon content.
Figures B.3-2 through B.3-4 show the relationships between PCB concentra-
tions and percent volatile solids within a texture classification for the three
most commonly occurring textures (gravel, fine sand, fine sand/wood chips). Very
little relationship appears to exist between total PCBs and volatile solids
measured in the gravel texture class (Figure B.3-2). The fine sand (Figure B.3-
3) and fine sand/wood chip (Figure B.3-4) categories exhibit a better correlation
between PCB concentration and percent volatile solids, although percent volatile
solids would still make a poor predictive measure of PCB concentration.
B.3-9
-------
Comparison of 1976-78 and 1984 Studies
As mentioned previously, the 1984 study focused on the 20 hot-spot areas
in the Thompson Island Pool as defined by areas exceeding 50 ppm PCBs in tft|
1976-1978 survey. N. P. Brown et al. (1988b) describe in detail the differences
found between the two surveys. The PCB concentrations in sediments of the
Thompson Island Pool exhibited lower concentrations in the 1984 survey than in
the earlier study, as shown by the somewhat higher frequency of PCBs detected at
<25 ppm (Figure B.3-5). A direct comparison of the relative frequency plot for
the two studies 1s hindered by the fact that the 1984 survey specifically
targeted potentially contaminated areas and areas with fine-grained sediments,
which were thought to contain more PCBs. Thus, 1984 samples were potentially
more heavily biased in a statistical sense to those areas of PCB contamination.
If samples had been taken randomly from the Thompson Island Pool, the results
would likely have yielded an even larger relative number of samples with lower
concentrations.
Because of the different scope and sampling density of the two surveys and
their different sampling and analytical methods, M. P. Brown et a7. (1988b)
Indicate that direct quantitative comparisons between samples collected in
similar areas are problematic. M. P. Brown et al. found that areas of high PCB
concentration determined in the 1976-1978 survey appeared to be generally
confirmed by the 1984 survey. Based on area-weighted average PCB concentrations
1n 138 polygon areas, these Investigators calculated a total PCB mass of 23,200
kg (51,040 lb) in the top 1.5 m (59 inches) of Thompson Island Pool sediments
(M.P. Brown et a7., 1988b). This mass estimate compares to Malcolm Pirnie's
estimate in 1978 of approximately 61,000 kg (133,670 lb) for all of the Pool and
approximately 48,000 kg (104,870 lb) for the 20 hot spots (Malcolm Pirnie, 1978).
M. P. Brown et al. (1988b) offer as possible explanation for the varied estimates
the differences in analytical PCB methods, depth Integration/area averaging and
sediment densities used.
B.3-10
-------
The importance of random variability and sampling density for either the
1976-1978 or 1984 studies also complicates comparisons between the two surveys
and affects mass calculation comparisons. For example, GE Indicates that they
have taken 30 samples from polygon 5 of the Thompson Island Pool (General
Electric, John Claussen letter to USEPA, March 29, 1991). This polygon was
estimated by M. P. Brown et aJ. (1988b) to have contained approximately seven
percent of the PCB mass in the Thompson Island Pool based on two samples
containing 39.7 and 6,587.8 ppm PCBs, yielding an average concentration of 2,437
ppm. GE has Indicated that the average, based on their 30 samples in this
polygon, is less than 20 ppm. (Results are not yet available to the project
study team.) Although PCB mass differences in this one small area of the
Thompson Island Pool do not necessarily mean that the overall conclusions of the
1984 survey are Incorrect, they do suggest that:
wide variations in PCB concentrations occur over relatively short
distances;
direct quantitative comparisons of PCB levels in samples frm
different years are problematic; and
the mass and distribution of PCBs in the Upper Hudson are difficult
to quantify.
Lamont-Doherty Geological Observatory Investigations
The Lamont-Doherty Geological Observatory, under contract to the NYSDEC,
conducted a survey of PCB levels in the sediments, suspended matter and water
column of the Upper Hudson River during 1983 and 1984 (Bopp et a/., 1985). The
survey, a coring effort, collected 16 cores, covering the Upper Hudson froa above
Hudson Falls to the Albany area (Plate B.3-1), and analyzed them for radio-
nuclides. Procedures were the same as those described 1n Section A.3 for the
Lower Hudson. Many sections of these cores were analyzed for PCB levels, with
an emphasis on homologue and congener-specific information. The investigation
also Involved PCB analyses of surface water samples (see Section B.3.3).
B.3.2.3
B.3-11
-------
On the basis of these data, Bopp et al. (1985) were able to draw a number
of Important conclusions concerning the fate of PCBs in the Upper Hudson. In
cores showing interpretable radionuclide chronologies, the occurrence of a
maximum PCB concentration in sediments deposited circa 1973 (the removal of the
Fort Edward Dam) could be seen in all areas of the Hudson, including the Thompson
Island Pool (Figure B.3-6). This data demonstrated that the sediments of the
Upper Hudson could be used to determine PCB transport history.
Analysis of Upper Hudson sediments revealed that very recent sediment
deposits (1980-1983) contained PCB congeners 1n ratios very similar to Aroclor
1242 and 1016, whereas older sediments, typically under anaerobic conditions,
showed substantially different congener ratios. In general, these older
sediments contained higher levels of mono through tetrachlorobiphenyls and lower
levels of the higher chlorinated congeners, relative to a standard Aroclor 1242
mixture. Bopp et a7. (1985) concluded on the basis of these results, that under
anaerobic conditions, biologically driven dechlorination must be occurring.
However, little or no dechlorination was occurring under the aerobic conditions
of the uppermost sediment layers.
B.3.2.4 Other Studies
1983 USEPA Study
In August 1983, USEPA conducted a limited study to collect samples from
locations that had been sampled 1n 1976-1978 (NUS, 1984). Sixty-six samples, of
which 54 were core and 12 were grab samples, were collected within a nine-mile
stretch of the river south of Rogers Island, including the Thompson Island Pool.
Forty-two samples were collected from within or on the border of previously
determined hot-spots. The results of this study tended to show that areas of
high PCB concentrations in 1976-1978 exhibited high concentrations In 1983 as
well. Nevertheless, direct comparisons between samples taken within 50 - 100
feet of each other during the two surveys Indicated that variations by two orders
of magnitude were not uncommon (NUS, 1984). In general, the concentrations
observed 1n 1976-1978 were greater than those seen in 1983 at corresponding locations.
B. 3-12
-------
6E 1989 Baseline Studies (Remnant Remediation Project)
As part of the Remnant Remediation Project, General Electric conducted
baseline pre-remediat1on sediment monitoring. Sediment samples were collected
at five locations in the vicinity of the remnants: one location near Rogers
Island; one location far upstream; one location between the remnants and Bakers
Falls; and two downstream locations near Lock 6 and Waterford. Median PCB
concentrations in river sediments 1n the remnant areas ranged from 0.47 ppm to
34 ppm, with a maximum of 99 ppm (Table B.3-7). Downstream samples contained
median PCB concentrations of 0.64 to 2.1 ppm, whereas the control location had
median PCB levels (seven samples) of 0.11 ppm. The sample location between
Bakers Falls and the remnants had a median PCB concentration of 1.4 ppm. With
the exception of the two downstream locations, PCBs were detected 1n all samples.
The chromatograms were compared against Aroclor mixtures 1221, 1232, 1016, 1242,
1248, 1254 and 1260; Aroclor mixtures in the samples were reported to be Aroclors
1242 and/or 1254.
GE 1990 Sediment Sampling (B1oremed1at1on Investigations)
General Electric has been conducting extensive research on biological
dechlorination and/or degradation processes occurring within the river, which may
have altered the composition of the PCB Aroclor patterns within the sediments.
In conjunction with these studies, GE has recently collected samples from
selected areas of the Upper Hudson for more detailed evaluation. General
Electric provided preliminary results of their sediment sampling activities
(Claussen, 1991b).
Harza Engineering, GE's contractor, collected 103 core samples from 12 hot-
spots during 1990 and reported 275 PCB analyses. From three to eight cores were
collected at most locations, with the exception of GE's H-7 location where 62
cores on a 12 x 12 foot grid were sampled. Samples were analyzed for PCB
homologue groups and five Aroclors (1221, 1242, 1254, 1260 and 1268). The
results of this sampling are summarized 1n Table B.3-8. With the exception of
H-7 and Location 4, the median PCB concentration at the four other locations
B.3-13
-------
within the Thompson Island Pool exceeded 100 ppm. Median PCB concentrations for
the aforementioned two locations and the locations downstream of the Thompson
Island Pool were less than 100 ppm.
\
Coincident samples and PCB measurements from the same hot-spots in the
NYSDEC 1976-78, USEPA 1983, and GE 1990 samples are available for only six
locations. The average PCB concentrations for each of these three surveys are
summarized in the lower portion of Table B.3-8. In four out of $ix locations,
the GE samples indicate average PCB levels above both the 1976-1978 and 1983
values; the remaining two locations show 1990 PCB levels lower than the 1976-1978
and/or 1983 results. Because of the very small sample sizes, few coincident
locations and difficulty 1n determining whether these samples represent similar
sediment zones over time, these results are Inadequate to suggest a clear trend.
Qualitatively, the results document the continued presence of PCB* in areas
originally defined in 1976-1978 to be contaminated.
B.3.2.5 Other Chemicals in Sediments
In addition to PCBs In river sediments, other chemicals, particularly heavy
metals, have been measured during 1976-1978 (Tofflemire and Quinn, 1979), 1984
Thompson Island Pool study (M. P. Brown et al., 1988b) and by other investiga-
tors. Lead, cadmium, zinc, chromium, mercury and other metals have been
measured. M.P. Brown et al. (1988b) indicate that anthropogenic sources,
Including a pigment manufacturer 1n Glens Falls, elevate lead, cadmium and
chromium In river sediments above their naturally occurring levels. Based on
their 1984 study, M.P. Brown et al. (1988b) reported mean metals concentrations
1n sediments for lead (217 jig/g), cadmium (21.6 pg/g), chromium (475 jig/g) and
mercury (1.96 |ig/g) along with several other chemicals. M. P. Brown et al. found
that although lead and cadmium are the two metals most frequently found in
sediments, standard leaching tests, e.g., EP Toxicity, suggest they are not
readily leachable.
B.314
-------
Relatively few sediment samples have been tested for other organic priority
pollutants. Four sections of two cores collected 1n 1983 by Dr. Richard Bopp
from River Miles 188.5 and 191.1 (Thompson Island Pool) were submitted to NYSDOH
and analyzed for dloxln and dlbenzofurans. Six sediment samples collected in
1987 from three hot-spots were analyzed for dioxins, dlbenzofurans, volatile and
semi-volatile organlcs and pesticides (M. P. Brown et a/., 1988). With the
exception of dlbenzofurans, none of the other organic parameters were detected
1n the 1987 samples.
* >.<
As reported by N. P. Brown et <7. (1988b) tetrachlorod1benzo-p-diox1n
(TCDD) and tetrachlorodlbenzofuran (TCDF) as well as their 2,3,7,8- Isomers were
detected at less than part per billion levels In the 1983 samples. Total TCDD
1n these 1983 samples ranged from non-detected to 0.135 ppb; total TCDFs ranged
from non-detected to 0.731 ppb. In two of the six 1987 samples, total TCDFs were
detected at <0.2 ppb levels; TCDDs were not detected In the 1987 samples;
detection limits ranged from 0.012 - 0.058 ppb. M. P. Brown et a7. Indicate that
possible sources of the TCDFs 1n sediment include residual fall-out from coal and
wood combustion, discharge from wood processing plants (by-product of chlorophe-
nol pyrolysis) and discharge of chemical-wastes containing TCDDs and TCDFs as
trace contaminants. Industrial PCB mixtures are known to contain trace levels
of TCDFs.
B.3.2.6 Discussion
The study team encountered some difficulty 1n matching the contents of the
data entries 1n the TAMS/Grad1ent database, especially the 1976-1978 sediment
data, with data summaries provided 1n previous reports (Tables B.3-2 and B.3-3).
The 1976-1978 raw data 1n printed form did not contain Identification designa-
tions cross-referenced to date and sample location and no report containing such
a cross-referenced summary was found. Sample Identification numbers are shown
on a marked-up copy of the 1977 Normandeau Associates, Inc. map 1n the NYSDEC
offices, but do not cross-reference dates of the samples or laboratory
Identification numbers. The team could assess the completeness of the data set
only by comparing sample dates and river location associated with the samples
B.3-15
-------
with summaries in published reports. As shown in Table B.3-2, these comparisons
Identify some inconsistencies between the data in the TAMS/Gradient database and
the number of samples reported by Tofflemire and Quinn (1979) in NYSDEC Technical
Paper No. 56. For example, the TAMS/Gradient database contains 254 grab samples
and 21 core samples reportedly collected in 1976, whereas Tofflemire and Quinn
report 24 cores and 80 unspecified sediment samples for 1976. For 1977, the
TAMS/Gradient database contains 446 grab and 246 core samples compared to 692
grab and 208 core samples reported by Tofflemire and Quinn (1979). For all
samples collected between 1976 and 1978, the database contains 1,092 samples
compared to 1,404 indicated by Tofflemire and Quinn.
Differences in the overall number of samples may be accounted for
approximately by noting that:
approximately 202 of the 672 summer 1977 grab samples taken by ,
Normandeau Associates, Inc. (NAI) were not analyzed for PCBs and
were not provided in the printed data summaries supplied for the
TAMS/Gradient database;
200 spring 1978 remnant samples from Malcolm Pirnie, Inc. (MPI) were
not provided for use in the TAMS/Gradient database.
If these 402 samples are subtracted from the Tofflemire and Quinn total of
1,404, there would be approximately 1,000 samples, more closely approximating the
1,092 samples in the TAMS/Gradient database. A comparison of the number of
samples by river mile (Table B.3-3) also indicates differences in total numbers
of samples reported by Tofflemire and Quinn and the data in the database.
Because 15 years have passed since the 1976-1978 samples, their use for
identifying precise locations and concentrations of contaminated sediments is
limited.
The data summaries contained inconsistencies within samples and between
samples. For example, the sums of Individual core lengths did not always match
the total core length; for approximately 3 percent of the samples, the data
fields were Incorrect; approximately 10 percent of the samples had no information
B.3-16
I
-------
about northing-easting coordinates. The team could not evaluate independently
the reason for the discrepancies nor the accuracy of the original data archival
process/data summaries. These discrepancies 1n the sediment database are not
considered to be significant at this time and an effort to resolve them will
continue during the course of the project.
Differences between the TAMS/Gradlent database and the results reported by
N. P. Brown et a/. (1988b) for the 1984 Thompson Island Pool survey appear to be
slight. The database contains 1,141 samples, whereas M. P. Brown et <7. report
1,205 for this survey, a difference of -5 percent in total samples.
B.3.3 Surface Hater Monitoring
Numerous surface water monitoring stations along the Upper Hudson are
maintained by the USGS. These stations have monitored flow, suspended sediment,
PCBs and other water quality parameters. The USGS data, obtained from WATSTORE
and the Albany USGS office, provide the longest and most comprehensive record of
surface water data for the Upper Hudson.
B.3.3.1 USGS Flow Records
The USGS has collected river discharge (flow) and water quality data at
various points along the Upper Hudson River (Plate B.l-1). The USGS records of
the monitoring stations located on the Hudson between Hadley, well above Fort
Edward, and Green Island, below the confluence with the Mohawk River at Troy,
were obtained for use 1n this investigation.
The majority of the USGS flow monitoring stations on the Upper Hudson have
periods of record beginning in the 1970s, although continuous monitoring 1s
available at Hadley since 1921. A lack of widespread flow measurements for
earlier periods presents difficulties In analyzing the longer-term flow regime
and flood probabilities. In particular, the flow record at Fort Edward, at the
upper end of the Thompson Island Pool, commences only in 1976. No USGS
monitoring 1s available at the Thompson Island Dam. Barge Canal stage data are
B.3-17
-------
available at the guard gate at Crockers Reef (Gauge #118) in the Hudson
River/Champlaln Canal approximately parallel to the Thompson Island Oam, at the
lower end of the pool. At the northern end of the Thompson Island Pool, Barge
Canal stage data are available below Lock 7 (Gauge #119). These gauges report
water elevations in reference to the Barge Canal datum. They have not been
calibrated to river discharges and provide only qualitative data regarding flood
discharges.
In order to extend the record of flow data for the Upper Hudson, It is
necessary to move upstream to the confluence of the Hudson and Saeandaga Rivers,
near Hadley (see Plate B.l-1). A monitoring station has been maintained on the
Hudson at Hadley since July 1921. The Saeandaga River, a major tributary
entering just below Hadley, has been monitored since 1907 at Stewarts Bridge near
Its confluence. By adding these two stations, USGS provides an estimate of the
flow in the Hudson below the confluence with the Saeandaga. Between this point
and Fort Edward there are several dams, but there are few additional tributaries.
The drainage area above Fort Edward is 2,817 square miles, while that of the
Hudson River at Hadley plus Saeandaga River at Stewarts Bridge is 2,719 square
miles, representing only a 3.6 percent increase in contributing area. Estimates
of flow in the Hudson below the Saeandaga, thus, provide an accurate estimate of
the magnitude of flow at Fort Edward.
The dally average flow value records for USGS Upper Hudson stations are
summarized 1n Table B.3-9, from upstream to downstream. USGS flow monitoring 1s
available since 1976 only at the upstream end of the Thompson Island Pool (Fort
Edward at Rogers Island). The closest functioning USGS monitoring station
downstream 1s at Stillwater, which 1s 26 miles and three dams south. Thus, Fort
Edward monitoring 1s most Informative of hydraulic conditions in the Thompson
Island Pool. Mean dally flow at Fort Edward 1s 5,244 cfs; dally flows range from
652 cfs to 34,100 cfs for the period of record. Additional inputs from
tributaries and runoff Increase the average dally flow to 7,933 cfs at Waterford
and 13,642 cfs at Green Island, below the confluence of the Hohawk River.
B.3-18
-------
Figures B.3-7a and B.3-7b display the dally average flows for 1973-1990.
As Fort Edward monitoring commenced In water year 1977, the 1973-1976 flows are
estimated from the calculated flows below Hadley, representing 97 percent of thie
contributing watershed area at Rogers Island. This record reveals the presence
of several major flood events, which were associated with mass erosion of the
remnant deposits. These events occurred on: April 2, 1976, when a dally average
flow of 39,340 cfs was reported below Sacandaga River; April 29, 1979, when a
flow of 31,700 cfs was reported at Rogers Island; and Hay 2, 1983, when 32,600
cfs was reported at Fort Edward. During these flood events flows were even
higher for shorter time periods (peak flows). The maximum peak flows monitored
at Fort Edward since December 1976 are 34,000 cfs on April 29, 1979 and 35,200
cfs on May 3, 1983.
The dally flows show evidence of a strong weekly periodicity. The seven-
day cycle Is the result of regulation of the Sacandaga Reservoir to supply power
plants during the week, while maintaining the weekend recreational pool in
Sacandaga Lake.
B.3.3.2 Suspended Sediments Monitoring
Information on time trends 1n suspended sediment data as well as the
relationship between sediment and discharge 1s provided by USGS monitoring
stations (Plate B.l-1). Several water quality stations were established on the
Upper Hudson 1n 1969, but measurements of suspended sediment did not commence
until 1975. Monitoring 1s not continuous or on a set schedule and there has been
a tendency to focus on spring flood periods, with little data available for the
winter months. Lack of a more extensive database and of regular time series
creates difficulties 1n analyzing sediment data as well as other water quality
parameters.
Summary statistics on the USGS suspended sediment monitoring for stations
between Fort Edward and Waterford are given 1n Table B.3-10. The median
suspended sediment concentration 1n the Upper Hudson above the confluence of the
Mohawk ranges from 4 to 12 mg/7 and Increases downstream. Relationships between
B.3-19
-------
suspended sediment levels and river flow are discussed In 8.4.
B.3.3.3 USGS PCB Monitoring
Methods and Procedures
Regular monitoring of PCBs 1n the water column In the Upper Hudson was
Instituted by the USGS In late 1975 at Waterford and expanded to other upstream
stations In 1977. Most other sampling programs, discussed later 1n this section,
have been of short term duration. A recent search of the STORET database reveals
that limited water-column PCB measurements are also available for some of the
tributary rivers to the Upper Hudson. These data have not been reviewed. The
USGS data are, thus, the primary source of time series information indicative of
trends 1n water-column PCB concentrations.
USGS observations of PCB concentrations 1n the water column have been made
at most of the same water quality stations as for sediment data (Plate B.l-1).
Data sets of significant size are available at Fort Edward (River Mile 194.5),
Schuylervllle (River Mile 181), Stillwater (River Mile 168) and Waterford (River
Mile 156.5), with a limited record at Fort Miller (River Mile 187). In addition
background samples are taken upstream at Glens Falls (River Mile 200).
The original purpose of the USGS monitoring was to gather several years of
data on PCB concentrations prior to removal of contaminated sediments (Schroeder
and Barnes, 1983). Although this dredging plan has been delayed, monitoring has
continued. Data are now available on UATSTORE, a USGS-maintained computerized
database, through the end of water year 1989 (September 1989). Data for water
year 1990 were collected by the USGS and have been turned over to NYSDEC for
analysis, but the results are not yet available.
Methods of data collection and analysis are summarized in Turk and Troutman
(1981) and Schroeder and Barnes (1983). According to the latter source, samples
from the Upper Hudson were collected from bridges using depth-Integrating
samplers. The sampler held a wide-mouth glass bottle, which was lowered and
B.3-20
-------
raised through the water column to obtain a depth-integrated sample (about 1
liter) for PCB analysis.
The USGS National Water Quality Laboratory in Doraville, GA performed the
PCB analyses. Comparison was made to standard Aroclor mixtures. ft&ults,
however, were reported as total PCBs. Schroeder and Barnes (1983) reported that
PCBs in the Hudson were "almost always 1n the composition range from [Aroclor]
1232 to 1248," but recognized that natural processes had likely altered the
Congener composition of the original Aroclor mixture.
'j
Although the USGS laboratory reports a theoretical detection limit of 0.01
|ig/7 through water year 1983, the practical quantitation limit was considered to
be 0.1 yg/7, because of the small size of the water sample (Bopp et al., 1985).
Data for this period recorded on UATSTORE contain both values entered as 0 and
values coded as <0.10 jig/7. Apparently these are both intended to represent non-
detects at the 0.1 detection level and the inconsistency is unintentional
(Rogers, pers. comm., 1991). With water year 1984, the practical detection limit
was lowered to 0.01 pg/7. Nevertheless, the 1984 and 1985 data are reported on
WATST0RE as if they adhere to the previous detection limit of 0.1 jig/7. In 1986,
the detection limit began to be reported as 0.01 pg/7 in WATST0RE.
PCB Results
The USGS monitoring station at Glens Falls provided upstream background
levels of PCBs in the Hudson for 1977-1983. Of 45 observations for total PCBs,
only two had detectable levels of PCBs. These observations occurred on December
5, 1978 and September 28, 1980 and were both reported as 0.1 |ig/7.
Summary statistics for the USGS monitoring of total PCBs in the water
column between Fort Edward and Waterford are given in Table B.3-11. A remarkable
fact evident from this table is that there is rather little variation in measured
PCB concentration by river mile between Schuylervllle and Waterford. At all
stations, PCBs were detected in more than 60 percent of the water samples, with
detection frequencies ranging from 63 percent to 89 percent. Average PCB
B.3-21
-------
concentrations below Fort Miller range from a high of 0.29 jig/7 at Stillwater to
0.23 pg/7 at Waterford. Averages at Fort Miller are not directly comparable to
other stations because of the short period of record. The average at Fort Edward,
Inflated by one very high measurement (77 pg/7), was calculated 1n Table
by omitting this outlier. This procedure yields a long-term average approximate-
ly half of that observed downstream. Concentrations observed at Schuylervllle,
Stillwater and Waterford are approximately constant, although the average at
Vaterford is lower than the up-river values, primarily because of dilution from
the Hooslc River. Indeed, the average concentration at Stillwater (0.29 pg/7)
1s slightly higher than that at the Schuylervllle station upstream (0.26 pg/7).
This finding suggests that there may be relatively little loss of water-column
PCBs during transit In the Upper Hudson.
At the Green Island station, downstream of the confluence of the Mohawk
River, the contributing watershed area 1s nearly double that at Waterford and PCB .
concentrations are correspondingly diluted. Twelve samples were analyzed for
total PCBs between 1978 and 1985 and all were non-detects at the 0.1 pg/7 level.
No PCB measurements have been reported at Green Island since March 1985.'
Figures B.3-8 through B.3-11 show the time series of PCB observations at
Fort Edward, Schuylervllle, Stillwater and Waterford, respectively. Observations
reported as non-detect (or zero) are plotted at the detection limit (0.1 pg/7
through September 1986 and 0.01 |tg/7 thereafter). While much of the variability
observed near the detection limit may represent analytical noise, there Is a
clear similarity apparent between the PCB time-series plots at Schuylervllle,
Stillwater and Waterford (B.3-9 through B.3-11), particularly 1n the marked
response to the 1979 spring flood. Despite the fact that samples were taken at
somewhat erratic Intervals, field personnel seem to have frequently visited each
Upper Hudson station In succession, so that many samples for the whole reach,
while not contemporaneous, are close together In time. All three stations are
below the Thompson Island Pool. Although the response to floods appears somewhat
different at Fort Edward, located below the remnant deposits but above the so-
called hot-spots 1n the Thompson Island Pool, the 1979 event stands out.
B.3-22
-------
Also notable 1n these time series is a general decline in water-column PCB
levels from about 1979-1986. The question of whether there has been any genuine
trend in PCB loading to water over time, or whether the apparent year to year
trends are actually due to variability 1n the hydrologic regime, is discussed in
Section B.4.
Current Full-Year Average PCB Concentrations 1n Water Column
Estimates of current average concentration of total PCBs in the water
column are needed for the assessment of potential baseline health risks and other
environmental impacts. As noted previously, seasonal variability and lack of
continuous sampling confound the estimation of this average.
It 1s common practice when working with data 1n time series containing non-
detects to develop an average based on treating the non-detects as if they were
equal to one half of the detection limit. A more sophisticated way to approach
this question is to use the Adjusted Log Normal Maximum Likelihood method of Cohn
(1988), which overcomes the bias in the sample collection and variance in the
data. Another method for including non-detects 1n the estimate of the mean,
which however does not address the problem of sampling bias, is to use a log-
probit analysis (Helsel and Cohn, 1988). Under the assumption that the logs of
the data are normally distributed, they will fall on a straight line when plotted
on a probit (probability) scale. The log-probit analysis essentially uses
regression to extend this line past the detection limit to predict the values of
the non-detect samples predicted by the observable part of the distribution.
In order to examine the sensitivity of the mean to the non-detects, the
above three methods of calculation were used here: (1) simple mean with non-
detects at 1/2 the detection limit; (2) the Adjusted Maximum Likelihood method;
and (3) log-probit analysis. Results of all three methods and the upper 95
percent confidence limit on the estimates of the means, which 1s used 1n the risk
analysis, are shown 1n Table B.3-12.
B.3-23
-------
Little difference 1n estimates of the mean and the 95 percent upper
confidence limit on the mean is produced by any of the three methods. Mean PCB
levels in recent 1986-89 water samples are on the order of 0.05 pg/7 at Fort
Edward and drop to approximately 0.03 |ig/7 at Waterford. Upper 95 percent
confidence limits on these means are not much higher, approximately 0.075 pg/7
at Fort Edward and 0.035 |ig/7 at Waterford.
Summer Average PCB Concentrations
Average water-column PCB concentrations can be calculated from monitoring
data based on either whole-year monitoring, including flood periods, or based on
low-flow or seasonal monitoring only. Assessing average concentrations for the
summer period, after the spring floods, 1s likely to be of greatest interest for
assessing biological impact. This period has maximum biological production and
1s also the season during which most of the fish samples have been collected.
There is also evidence to suggest that spring flood PCBs are largely sorbed on
sediment particles, whereas concentrations associated with low flows are
primarily dissolved or sorbed to very fine particles, i.e., pass a 0.45-pm
filter. Thus, they are more readily available to enter the food chain (Bopp et
al1985).
Average PCB concentrations during the summer (June-September) have been
calculated. In doing so, the presence of many non-detects among the samples was
addressed. The robust log-prob1t analysis method (Helsel and Cohn, 1988) was
used to estimate averages in the presence of non-detects. No systematic bias
toward higher concentration events is expected to apply to the summer observa-
tions, although this may be the case for spring observations. In 1986 there were
no summer observations at Stillwater or Waterford, while summer observations at
Schuylervllle 1n this year as well as at Waterford 1n 1985 were all less than or
equal to the detection limit of 0.1 |tg/7. In the latter case the log-probit
method cannot be used and the average has been arbitrarily set to one-half the
detection limit. The calculated summer average PCB concentrations are shown in
Figure B.3-12 and summarized in Table B.3-13. Recent 1988-1989 summer average
PCB concentrations are fairly uniform for all locations and are on the order of
B.3-24
-------
0.03 - 0.04 ug/1. Summer average concentrations at Fort Edward tend to be less
than or equal to those downstream, whereas for the full year concentrations,
including spring runoff, the average concentration at Fort Edward is higher than
that found downstream.
B.3.3.4 Other Sources of Water Column Data
Waterford Treatment Plant Data
The City of Waterford operates a water works serving a population of
approximately 12,000 persons in the Towns of Waterford and Halfmoon and the
Village of Waterford. This 1s the first water treatment facility downstream of
Fort Edward drawing water from the Hudson. In 1975, when the USGS began
collecting PCB data 1n the river at Waterford, they also began collecting raw
water input and finished water output data at the Waterford treatment plant, in
cooperation with the Board of Water Commissioners of the Town of Waterford and
the NYSDEC (Schroeder and Barnes, 1983b). The water for the treatment plant is
drawn from a location 0.5 km upstream of the US Highway 4 bridge, where Hudson
River water samples are also taken.
Data collected in cooperation with the USGS run through the end of water
year 1983. In addition, approximately bimonthly data for November 1983 -
February 1985 and March 1987 - October 1989 were available from the Waterford
Water Works (Metcalf & Eddy, 1990). The data through 1983 were often collected
concurrently with samples from the Hudson River at Waterford and can be used to
provide a check on that data. Since September 1982, no PCBs have been reported
above the detection limit (generally 0.1 yg/7), 1n either raw intake water or
treated water.
Lamont-Doherty Study of 1983
A detailed study of PCB transport in the Upper Hudson, conducted by
researchers of the Lamont-Doherty Geological Observatory in 1983 (Bopp et <7.,
1985), involved an investigation of spring/summer 1983 PCB transport in the Upper
B.3-25
-------
Hudson, which was a period of relatively high flows. The Lamont-Doherty study
Included the collection of data not available from USGS sampling. In addition
to sediment cores, this study included 20 large-volume filter samples of
suspended matter and fifteen 9-20 liter unfiltered samples containing water aad
suspended matter, collected from Troy to Glens Falls. Unlike USGS monitoring^
detailed component analysis was undertaken for these samples. The data collected
were used to form an empirical transport model, which used as input USGS measures
of suspended sediment and flow, to predict total PCB load.
PCBs 1n samples of suspended matter were found to match standard Aroclor
mixtures, e.g., Aroclor 1242, reasonably well. Samples taken during a high-flow
period were found to have significantly higher PCB levels (6.33 ppm) than those
taken during a low-flow period (0.69 ppm). Some of the drop 1n PCB levels at
locations downstream of Fort Edward was attributed to dilution from tributaries
joining the Hudson.
Water samples were filtered and both filtrate and particles were tested for
PCBs. Comparisons of water versus suspended matter PCB concentrations were made
to derive a distribution coefficient. An inconsistency was found in that the
concentration of PCBs on suspended matter filtered from the 9-20 liter samples
was generally two to four times higher than the levels from the large-volume
filter samples collected at the same locations on the river. This discrepancy
could possibly be accounted for by an equilibration between water and suspended
matter during storage of water samples prior to filtration and testing. Another
possible explanation was the difference in filter size used to collect the
particulate sample (1.2 microns) and to separate the water samples (0.7 micron).
Seasonal variations In concentrations of PCBs in water were found. Pre-
spring runoff showed the lowest PCB levels. Summer samples showed higher PCB
concentrations, accounted for by boat traffic and increased use of locks, which
would resuspend sediments.
B.3-26
-------
Bopp et al. (1985) were able to derive in situ partition coefficients for
PCBs on a quasi-homologue basis, based on packed column analysis. These analyses
also Indicated the possibility that water-column PCB distributions may not
reflect equilibrium conditions and that dissolved phase PCB concentrations say
be higher than those predicted by equilibrium. The homologue distribution of
PCBs on suspended matter was readily Interpreted as Aroclor 1242-11ke, similar
to PCBs in the aerobic sediment layers. The dissolved phase homologue
distributions were not as readily explained, although preferential partitioning
of the lower chlorinated homologues to the dissolved phase appeared to be a
'i
likely possibility, i.e., lower chlorinated homologues have lower partition
coefficients and, therefore, higher levels in the dissolved phase. The greater
volatility of the lower chlorinated PCBs and their production In the anaerobic
sediment layers could possibly confound this Interpretation. The general
homologue pattern agreement between water column and surface sediments led Bopp
et a7. (1985) to conclude that little or no release of PCBs from the anaerobic
sediments was occurring on a substantive basis in comparison to the mixing and
resuspension of the surflcial sediments.
Bopp et al. concluded that PCB transport 1s tied to suspended Batter
transport, with the majority of PCB transport to the Lower Hudson occurring
during the 10 to 20 days per year of major sediment transport.
NYS00H Hater Column PCBs
As part of their macroinvertebrate sampling program, NYSDOH also collected
water samples and analyzed them for PCBs. This and other recent data received
from Dr. Bush at NYSDOH, require additional evaluation. The macroinvertebrate
studies are discussed 1n a later section.
Remnant Deposit Containment Monitoring Program
As part of remedial activities at the PCB remnant deposits at Fort Edward,
General Electric is conducting a baseline environmental monitoring program, which
will continue through and follow the in-place containment of Remnant Deposits 2,
B.3-27
-------
3, 4 and 5 In accordance with the Administrative Order on Consent II CERCLA-
90224. The baseline monitoring includes sampling of many environmental matrices,
including water column PCB concentrations. Results of the first phase of this
effort cover August through December 1989 (Harza Engineering, 1990).
During the baseline monitoring, samples were taken weekly or biweekly at
ten water quality stations. Two of these stations are upstream of Fort Edward,
one in the Sherman Island Pool and one just below Bakers Falls, while five are
1n the area of the remnant deposits at Fort Edward above Rogers Island. Station
E-5 was located downstream of the Route 197 bridge on Rogers Island and is just
below the US6S monitoring station. The remaining two stations were at Channel
Marker 175, below Fort Miller Dam and Lock 6 (E-6), and at Channel Marker 13, two
miles north of a NYSDEC boat ramp at the Erie Canal and Hudson River confluence,
near Uaterford (E-7). Raw water samples were analyzed for PCBs with an
approximate analytical detection limit of 0.1 |ig/7. In addition, dialysis bags, ,
filled with 4 ml of hexane to concentrate PCBs, were suspended 1n the water
column and analyzed biweekly. As the concentration factor for the dialysis bags
is unknown, they can be used to indicate qualitatively the presence of PCBs, but
not ambient concentrations.
The 1989 monitoring program unfortunately missed the spring runoff period.
All raw water samples were reported to be below the detection limit of 0.1 iปg/7.
This result 1s consistent with USGS monitoring data for water year 1989 at Fort
Edward, which had a lower detection limit and showed detectable concentrations
1n the 0.01-0.1 pg/7 range. During the same period the dialysis bag concentra-
tors occasionally detected PCBs at all stations except the uppermost (Sherman
Island). All the detects were Identified as Aroclor 1242.
B.3.4 F1sh and Other Aquatic Biota
Substantial declines In average PCB burden in Upper Hudson fish were
observed in the years after 1978 (Sloan et a J., 1983, 1984; M. P. Brown et a7.,
1985). Analysis of the data reveals that these declines 1n concentration have
proceeded at a slower rate in more recent years. It Is unclear to what degree
B.3-28
-------
the abnormally low spring floods of the 1980s have affected PCB levels and may
be responsible for the observed declines. It does appear that total PCB con-
centrations in fish on a lipid (fat)-basls can be closely predicted from summer
average water column PCB concentrations.
Because PCBs are typically stored (bioaccumulate) in fatty (lipid) tissues,
it 1s sometimes useful to normalize the PCB levels In fish and express them on
a lipid basis, i.e., PCB content 1n fish expressed as ug-PCB/g-flsh lipid.
Whether or not normalized to lipid content, levels of the higher chlorinated
congeners 1n largemouth bass were approximately stable from 1981-1988. Reported
Aroclor 1016 levels, representing less chlorinated congeners, appear to have
continued to decline for all species during this time period. Given the slow
rate of reduction of Aroclor 1254, it is unclear when or 1f natural processes
will reduce the PCB burden 1n fish to acceptable levels. Furthermore, potential
changes 1n PCB levels in the water column and sediment caused by possible scour
and resuspenslon of sediments would likely cause at least temporary Increases 1n
PCB levels 1n fish and aquatic biota.
In addition to fish data, some monitoring of invertebrates is also
available for the Hudson. From 1976-1985 multiplate samples and chlronomid
larvae have been monitored by NYSDOH. These data are discussed at the end of
this section, following the discussion of the fish monitoring program.
B.3.4.1 Fish Sampling
Data on concentrations of PCBs in Upper Hudson River fish collected by
NYSDEC between 1975 and 1988 were used In this study. While over 30 species of
fish are represented In the data, the majority (75 percent) of the samples are
from half a dozen species Including striped bass, largemouth bass, brown
bullhead, pumpkinseed, American shad, and American eel. Approximately two-thirds
of the samples tested were standard fillet samples.
B.3-29
-------
Samples Prior to 1975
While PCBs are known to have been discharged Into the Upper Hudson River
since the 1940s, no testing for PCBs 1n fish Is known to have been undertak^i
before 1970. In that year, a nationwide survey of chemical pollutants 1n game
fish conducted by a popular magazine Included a sample of spawning striped bass
from the Hudson estuary In which 4.5 to 5 ppm PCBs In flesh and 11 to 12 ppm PCBs
In eggs were reported (Boyle, 1970). NYSDEC had been analyzing fish for DOT and
other pesticides statewide since the early 1960*s. In 1971, NYSDEC added PCBs
to their analyses, although no results were released publicly until 1975
(Sanders, 1989).
F1sh data collected and analyzed for PCBs 1n the 1970-74 ;per1od are
summarized by Spagnoll and Skinner (1977). These 1970-74 Hudson River samples
Include one smallmouth bass collected at Warrensburg (Upper Hudson) and 146 fish
from 11 species collected below the Federal 0am (Lower Hudson). The highest
observed concentration from below the Federal Dam was 1n a largemouth bass,
reported as containing 53.81 ppm wet weight total PCBs 1n the 1970-72 period.
This sample was taken prior to the removal of the Fort Edward Dam.
In August 1974, a USEPA team obtained water, sediment and fish samples from
upstream and downstream of the GE discharge at Fort Edward. A sample of 42
shiner minnows from below the GE discharge showed an average concentration of 78
M9/9 (ppm) PCB as Aroclor 1242, while one rock bass was reported with 342 pg/g
(Nadeau and Davis, 1974). It should be noted that samples collected at control
Station 0 above Bakers Falls were not non-detects. PCB levels at Station O were
reported as 7.0 pg/gm in shiner minnows and as 17.0 pg/gm 1n yellow perch. The
latter level Is higher than the average for all NYSDEC yellow perch samples in
the Upper Hudson. Few other samples have been reported from Bakers Falls.
B.3-30
-------
Samples 1975-1976
Following the USEPA Investigation, NYSDEC undertook more detailed
monitoring of PCBs in fish from both the Upper and Lower Hudson. A total of 440
Hudson River (Upper and Lower) fish samples were analyzed in 1975-1976, the
results of which NYSDEC -provided for Incorporation into the IMS/Gradient
database.
The 1975-1976 fish collections were made by regional USFUS personnel who
were instructed on specific species and sizes of fish desired, location of
stations and time tables for collection. Target species for the Hudson included
smallmouth bass {Nlcropterus dolomienl), largemouth bass (Nlcropterus salmoldes),
brown bullhead (Ictalurus nebulosus), goldfish (Carassius auratus), white sucker
(Catastomus comersoni), striped bass (Horone saxatilis) and various other
estuarlne species. Other species were occasionally obtained as available.
Attempts were made to sample small, medium (minimum legal) and large representa-
tives of each species.
Analyses were conducted by several different state laboratories, apparently
using the methodology of Bush and Lo (1973) and reported against standards for
Aroclor 1242 or 1016 and Aroclor 1254. Aroclor 1221 was not analyzed. The
nominal detection limit of the method was 0.01 ppm, although some of the labs
reported results only as low as 0.1 ppm.
Samples 1977-1988
In 1977-1979 NYSDEC monitoring methods were refined and standardized.
Collection has continued to the present, although no results after 1988 are
available. As of 1988, NYSDEC changed the frequency of sampling from yearly to
every other year. Samples from 1990 were collected, but PC8 analyses have not
been completed. NYSDEC provided data covering the period of 1977 through 1988,
which contains 7,373 Upper and Lower Hudson fish or fish composite analyses.
These samples have been collected on a regular basis, with the intent of sampling
given species at predetermined locations within two weeks of a specified date in
B.3-31
-------
order to minimize potential seasonal effect. Records are very limited for 1981
and 1987. Table 6.3-14 provides a summary of the total number of fish sampled
for the Upper Hudson from 1975-1988.
Information in the fish database, in addition to chemical analyses, usually
includes species, method of preparation, weight, length, percent lipid and, if
determined, sex and age. In addition to a descriptive sampling location, river
mile numbers and geographic designations are associated with each sample.
Sample collection, preparation and analytical methods are described in
Armstrong and Sloan (1981). The desired sample size for each species collected
was 30 Individuals, although the availability'of fish did limit sample size in
some cases. A single fish sample generally consisted of a composite of one to
three fish. For fish longer than 150 mm, standard fillets of whole sides of
scaled fish were prepared for analysis. For brown bullhead samples, the skin was
removed from the fillet. Samples of fish shorter than 150 mm were analyzed whole
with head and viscera removed.
Aroclor concentrations 1n fish were determined by comparison to commercial
Aroclor standards. This method was not able to distinguish between Aroclors 1242
and 1016. The detection limit for each Aroclor tested was 0.1 ppm. For each
sample the percent lipid was determined as the percent by weight of tissue
soluble 1n petroleum ether.
PCB Levels 1n Fish
This section summarizes the Upper Hudson PCB monitoring results for fish.
Additional statistical analyses are presented In Section B.4. Samples collected
at River Nile 153, just below the Federal Dam, are included with those of the
Upper Hudson, since they represent a resident, freshwater* rather than estuarine,
population, which is exposed to PCBs transported over the dam.
B.3-32
-------
Although over 30 fish species have been sampled 1n the Hudson River, a few
of these species account for the majority of the samples collected. These
species are pumpkinseed, largemouth bass, brown bullhead, goldfish (carp), white
perch and yellow perch. Overall average Aroclor levels for these species In
River Miles 153 to 195 of the Upper Hudson for 1975-1988 are provided tt^Table
B.3-15. Aroclors 1016 and 1254 are the dominant PCB mixtures reported; Aroclor
1221 represents the smallest PCB fraction. Overall, the highest average PCB
concentrations have been found in goldfish (carp) with 1975-1988 average Aroclor
levels of 32.0 ppm (Aroclor 1254) and 91.6 ppm (Aroclor 1016). That goldfish
have the highest average PCB levels Is not altogether surprising as they also
have the highest lipid content (9.3 percent). The average 1975-1988 Aroclor
levels 1n largemouth bass, perch and bullhead are <30 ppm, although total PCB
levels 1n Individual largemouth bass have been reported as high as 370 ppm.
The average of total PCB levels in all fish for recent years (1986-1988)
also shown in Table B.3-15, is 10.9 ppm. The upper 95 percent confidence found
on this mean, used in the preliminary human health risk assessment (B.6), is 12.0
ppm.
Tables B.3-16 through B.3-18 provide summary statistics by river aile
sampled for largemouth bass, pumpkinseed, and brown bullhead, respectively.
Recent data for 1986-88 show that median PCB levels in fish range from 1 to 30
ppm depending on species and river mile. Additionally, the variability of PCBs
1n fish from a given location and year, as measured by the standard error of the
estimate of the mean, has dropped as the PCB levels have declined.
Figure B.3-13, which plots mean, total PCB levels 1n brown bullhead
(skinned standard fillets) as a function of year and river mile, provides an
overall perspective of the general decline of PCB levels over time and by river
mile. PCB levels in the vicinity of River Mile 190 are the highest, but samples
from only a few years are available here.
B.3-33
-------
Surprisingly little fish data are available from the Thompson Island Pool
area (River Mile 188-94). Regular monitoring there began In the latter half of
the 1980s. The most complete data set available for the Upper Hudson above the
Federal Dam is that at River Nile 175. Data here begin 1n 1975 and continue,
with some Interruptions, through 1988. Figure B.3-14 summarizes the trends
mean Aroclor levels for four fish species at this location as ppm wet weight
Aroclor 1254 and Aroclor 1016. Aroclor 1016 levels are shown on a log scale, as
they were extremely high for certain species 1n the 1977-78 time period. Aroclor
1221 was quantltated also, but generally at very low levels compt-ftd to the other
two Aroclors. A notable trend 1s that during the 1977-78 perlQd the ratio of
Aroclor 1016/1254 levels 1n fish are elevated, /.e., the lower chlorinated
congeners (1016) dominated the higher chlorinated congeners (1254) during the
late 1970s. This elevation may be attributable to scour of dechlorlnated
sediments from depth. Another possible explanation for the shift from Aroclor
1016 dominance to Aroclor 1254 1n recent years 1s that the lower chlorinated
congeners were more rapidly released and dissipated 1n the late 1970s and early
1980s, whereas the higher chlorinated congeners have tended to bloaccumulate more
and are less rapidly released by fish.
L1p1d-Based PCB Concentrations
Differences between PCB levels 1n different fish species are probably due
to both differential feeding patterns and 11p1d content. F1sh lipid (fat)
typically accumulates PCBs more than other, less fatty tissues, because PCBs are
highly lipophilic compounds, I.e., PCBs are more soluble 1n fat than water.
Normalizing PCB levels on a fish lipid weight basis has sometimes proved useful
1n comparing measurements among species. The average percent lipid of goldfish
was 9.2, pumpklnseed was 3.3, brown bullhead was 2.9 and largemouth bass was 1.4,
with skewed distributions. There appears to be little correlation between weight
and PCB body burdens in most species. Age of the fish sampled was not often
determined.
B.3-34
-------
Besides accounting for species differences, accounting for lipid content
1s Important because fish lipid content appears to have changed from samples
caught from one year to the next. Dividing the PCB concentration (wet weight)
by the measured 11 p1d content (g-lipid/g-fish) of the sample, one obtains the PCB
concentration per gram (mass) of fish-Hpid (ug-PCB/g-Hp1d). Looking at all
samples for all species at all Upper Hudson sampling locations on a lipid basis,
the median PCB levels have declined from 1829 ug-PCB/g-Hp1d In 1977 to 271 ug-
PCB/g-1ip1d 1n 1988, with a 95 percent upper confidence bound estimate on the
mean of 484 (Table B.3-19).
L1p1d-based means of Aroclor 1254 and 1016 concentrations by year In the
predominant fish species sampled at River M11e 175 are shown 1n Figure B.3-15.
On a lipid basis, largemouth bass have usually shown the highest Aroclor 1254
levels. This may reflect their position as top carnivores In the aquatic food
chain or their low fat content. Further, the Upld-based Aroclor 1254 levels 1n
most species appear to have been relatively constant since 1982. Error bars are
omitted, from the multiple species plots for legibility. Trends 1n largemouth
bass, with error bars, are shown 1n Figure B.3-16. (An error bar shown as a
vertical line at each year representing the 95 percent upper and lower confidence
bounds on the mean.)
In addition to the monitoring at River Mile 175, there 1s a good continuous
record of sampling of brown bullhead at River Mile 153 just below the Federal
Dam, for the period 1977-1988. PCB levels here were on average much lower than
those observed at River Mile 175, presumably due to dilution of PCB concentra-
tions by the flow of the Mohawk River. Average lipid-based PCB concentrations
in brown bullhead show a regular exponential decline for Aroclor 1016 components
and a less dramatic decline for Aroclor 1254 (Figure B.3-17).
B.3.4.2 Other Chemicals 1n Fish
The NYSDEC database contained analyses for chemicals 1n fish other than
PCBs. Summary statistics for these other chemicals are shown In Table B.3-20.
B.3-35
-------
Netals such as mercury, cadmium, chromium, copper and zinc were frequently
detected. Median mercury and zinc concentrations were on the order of 0.4 ppm
and 10 ppm, respectively. Organic compounds frequently detected Include
pesticides such as DDT, heptachlor, and dleldrln. Hexachlorobenzene and
hexachlorocyclohexane were also detected frequently.
B.3.4.3 NYSDOH MacroInvertebrate Studies
As part of the Hudson River PC8 Reclamation Demonstration Project, the New
York State Department of Health (NYSDOH) conducted b1omon1tor1ng studies from
1976 to 1985 using caddlsfly larvae, multlplate samples and chlronomld larvae
(Simpson et a?., 1986). These studies Included long-term blomonltorlng efforts
from 1976 to 198S as well as two short-term biological uptake studies 1n July and
September of 1985.
Long-Term Blomonltorlng Study
From 1976 through 1985, artificial substrate samplers.(multlplates) were
placed at 17 sites along the Hudson River from Hudson Falls to Nyack, New York
(Novak et a1.t 1988). These samplers were collected each year after a period of
five weeks during the months of July, August and September. PCBs in the samples
were reported as Aroclors 1016 and 1254. The resulting PCB concentrations 1n
the multlplate samples represented a composite of sediment, algae, plankton and
various macrolnvertebrates. Invertebrates collected 1n the multlplate samplers
Included the following taxonomlc groups: Chlronooridae, OUgochataeta,
Trlchoptera, Ephemeroptera, Amphlpoda and Ellmldae. Chlronomld larvae and pupae
were the most abundant Invertebrate component from Fort Edward to Saugertles,
comprising up to 86 percent of the total macroInvertebrate population at Fort
Miller and Waterford.
From 1978 to 1985, caddlsfly larvae were collected by hand-picking
Individuals from rocks at five designated sites: Hudson Falls, Fort Edward, Fort
Miller, Stillwater and Waterford. Caddlsfly collections were made 1n June, July,
August and September of each year.
B.3-36
-------
Measured PCB levels 1n the 1985 multiplate samples for September ranged
from 0.25 ppm at Hudson Falls (the control site) to over 6 ppm at Fort Edward.
Multiplate monitoring from Fort Miller to Waterford resulted 1n PCB levels of 4
to 5 ppm. The multiplate samples at any one site appeared to show a consistent
decline 1n PCB concentrations from early summer to later summer 1n any particular
year. Larger scale trends or relationships In either time or with sample
location are difficult to detect, because of the extremely wide variation In the
sample results. Average PCB concentrations 1n multiplate samples generally
showed a decline from 1976 - 1980. Nevertheless, average PCB concentrations
Increased in 1981 and remained high through 1985. Multiplate samples from the
Thompson Island Pool and downstream showed significantly higher PCB concentra-
tions than samples taken upstream of Fort Edward. Yet, no significant trends are.
apparent when comparing Fort Edward results with those at Waterford.
The results of the caddlsfly blomonltorlng efforts show a decline 1n total
PCB concentrations from 1978 to 1980. As 1s true of the multiplate data, spatial
trends are not readily apparent from the caddlsfly samples. Measured PCB levels
1n macroinvertebrate tissues generally ranged from 20 to 60 ppm (dry weight) 1n
1979 and from 20 to approximately 40 ppm (dry weight) In 1985. During the same
collection periods, macroinvertebrates collected at Hudson Falls, the control
site, exhibited PCB tissue residues less than 10 ppm.
The data from this study exhibit a great deal of random variability. The
bulk of the data are from multiplates, which collect sediment as well as living
matter. In theory, 1t should be possible to distinguish sediment versus
biologically-based PCBs by adjusting the observations to a lipid basis; lipid
content of samples was reported. An additional factor contributing to the
variability is that 1n almost every year a downward trend by month was observed
at most stations, based generally on three samples. The cause of this phenomenon
is not known, but the limited number of observations in this yearly cycle m*y
have obscured the Influence of other factors.
B.3-37
-------
Examination of the complete run of samples at the Fort Miller multlplate
and caddisfly station (PCB-5) shows what appears to be a largely random pattern
of total PCBs on a dry weight basis (Figure B.3-18). The larger mean PCB
concentration for the period 1981-82 1s largely attributed to the Increased
variability of the data for this period. Adjusting to a lipid basis actually
increases the total variability represented by outlying data points. Only a
slight trend 1s suggested 1n this figure, where PCB levels may have declined from
1976 to 1980, then remained approximately stable. At Stillwater, the Upld-based
values appear to be almost entirely random (Figure B.3-19). i*
A comparison between the confidence Intervals on the overall means for all
sample years by sample locations (Figure B.3-20) shows differences between the
low values observed above Fort Edward and values at downstream stations. Llpld-
based means at all stations by year (Figure B.3-21) appear to show a decline from
1976 to 1979, then relatively constant levels except for a jump upward In 1982.
Again, this apparent trend 1s difficult to confirm as a consequence of the
limited number of samples and possible differences 1n sampling or analytical
protocols throughout the duration of the monitoring.
Short-Term Blomonltorlng Study
Short-term blomonltorlng Investigations using the chlronomld larvae,
Chironomus tentans, were also performed'by the NYSDOH during July and September
1985 (Novak et a/., 1990). The monitoring method consisted of placing 25
1 aboratory-reared chlronomld larvae 1n nylon mesh envelopes or packets that were
exposed to the water column. Envelopes were placed, 1n groups of ten, 1n steel
mesh baskets at the primary collection site and monitored at 0, 1, 2, 4, 8, 12,
24, 48, 72 and 96 hours. Chlronomlds were placed at four sites, Including two
at Thompson Island Pool, one at Bakers Falls and one at F1sh Creek, and monitored
at 96 hours. Packets of chlronomlds exposed to the sediment at a collection site
located on the eastern shore of the Thompson Island Pool were also collected at
96 hours. Water column samples were obtained during the same collection
intervals for each site.
B.3-38
-------
Results of this Investigation Indicated that chlronomlds accumulated PCBs
ranging from 0.1 to 7 ppm after 1 to 96 hours of continuous exposure, whereas
larvae exposed to sediments near the Thompson Island Pool for 96 hours contained
over 100 ppm. Water column PCB levels were in the range of 0.03 - C^.l pg/f
during the experiment. The ratio of the PCB levels 1n the chlronomlds (in ppb)
to the ambient PCB concentrations 1n the water column (defined as the bioaccue-
ulation factor or BAF) were on the order of 10* to 105.
A significant conclusion from this study was that the PCB congener patten
found in tissues of chironomid larvae differed substantially from the congener
pattern observed in water. Using capillary column gas chromatography, the
investigators were able to isolate PCB congeners in both the water column and
chlronomlds. The most abundant congeners in chironomid tissues were 2,4,2',5*-
tetrachlorobiphenyl and 2,3,6,4'-tetrachlorobiphenyl. In contrast, the
predominant congeners in water were 2,2'-d1chlorobiphenyl, 2,6-d1chlorobiphenyl,
and 2,6,2'-tr1chlorob1phenyl. These findings suggest a number of possible
explanations. One explanation could be that chironomid larvae may selectively
b1oconcentrate the more highly chlorinated congeners which are present at
relatively low concentrations in water. Another factor which could explain the
observed results is that the lower chlorinated congeners were present In the
water but below detection limits.
Although this study presents some very interesting congener-specific
results, they are too limited in scope to provide clear indications of either
congener-specific biฉaccumulation or congener-specific comparisons of PCBs 1b
sediments. (No congener-specific sediment data were obtained.)
B.3-39
-------
B.3.5 PCB Concentrations in Air and Plants
B.3.5.1 A1r
Monitoring Near Fort Edward
A1r monitoring efforts for PCBs and other air toxics have been conducted
1n the Upper Hudson River study area from late 1976 to 1982 by NYSDEC/NYSDOH and
various researchers and as recently as 1989/90 by contractors (Harza/Yates-
Auberle) for General Electric.
From January through August of 1977, NYSDOH collected air samples at five
locations 1n the Upper Hudson Valley to determine ambient PCB concentrations.
While the Glens Falls and Warrensburg samples showed no detections above the
0.020 |ig/m3 detection limit, results from the Hudson Falls and the Fort Edward
i
samples demonstrated high levels of total PCBs, ranging from 0.060 |ig/m3 to 3.26
|ig/m3 (Malcolm-Plrnie, 1978). Atmospheric levels of PCBs in the Fort Edward area
were reported to decrease from 1 iig/ra3 down to 0.3 |&g/m3 after the cessation of
PCB use by General Electric in their Hudson Falls and Fort Edward capacitor
plants 1n 1977 (Llmburg, 1984).
In 1979, NYSDEC conducted an air monitoring survey for PCBs around various
dumps and landfills (Caputo and Fort Miller dumps, Remnant Area, Moreau and Site
3a and Buoy 212) In the Hudson Falls/Fort Edward area bordering the Upper Hudson
River (see Table B.3-21). Values ranged from 5 to 15 |ig/m3 total PCBs at the
Moreau and remnant areas and 24 to 300 yg/m3 total PCBs at the Fort Miller and
Caputo dumps, respectively. At the Caputo dump, where the soil was reported to
contain 5,000 mg/kg PCBs, air monitoring for PCBs before and after capping of the
site showed that average ambient PCB concentrations decreased from 118 |ig/m3
before capping to 0.26 |ig/m\ once the site was capped (Shen, 1982).
Two air samples taken over Lock 6 1n the summer of 1980 yielded Aroclor
1242 concentrations ranging from 0.11 i>g/m3 to 0.52 |ig/m3 (Table B.3-21).
Aroclors 1221 and 1254 were not detected (NYSDEC, 1981). Ambient PCB levels
B.3-40
-------
monitored at farm fields in 1981 near the Hudson River, however, showed low PCB
concentrations of approximately 0.005 iig/m1 (NUS, 1984).
In the early 1980s, NYSDEC and the Boyce Thompson Institute for Plant
Research of Cornell University conducted a joint air/plant monitoring effort near
the tallwater of Lock 6 to determine 1f volatilization of PCBs from the water
column was occurring (Buckley and Tofflemlre, 1983 and 1984). Between August
1981 and September 1981, seven air samples were taken. Aroclor 1242 was detected
in all seven samples, ranging from 0.031 to 0.06 |ig/m'. Aroclor 1254 was
detected in three of the samples at levels up to 0.0013 jig/m\ During this
study, a vertical PCB gradient was also noticed, when airborne PCBs were measured
simultaneously at heights of 1 and 4.5 meters above the water.
In August 1986, NYSDEC collected three sets of ambient grab samples in
duplicate at the proposed containment site (Site G), the Fort Edward Landfill,
the Bourgoyne Avenue School and Lock 7 of the Champlaln Canal (USEPA/NYSDEC,
1987). The highest ambient PCB concentration measured was 0.083 |ปg/m3 at Lock
7. Site G and the Fort Edward landfill samples contained PCB levels below the
detection limit of 0.007 jปg/m\ The Bourgoyne Avenue School sample Information
was not available at the time of the report.
Again in 1987, NYSDEC conducted air monitoring from April 2, 1987 to July
16, 1987 at the Kingsbury Landfill, located north of Fort Edward. In 76 of 105
samples taken over April and May 1987, Aroclor 1016/1242 was detected with a
maximum concentration of 0.49 |ig/m3. Aroclor 1248 was detected in 5 of 105
samples with a maximum concentration of 0.52 iig/m3. Both the Kingsbury and Fort
Edward municipal landfills were the burial site of several thousand tons of PCBs.
Neither these landfills, nor others in the area, are part of the current
investigation of the Hudson River Superfund Site. The results are presented here
as evidence of PCBs having been monitored 1n the air within the vicinity of the
site.
B.341
-------
Most recently, 1n connection with Its Remnant Remediation Project, GE
conducted baseline pre-remediation air monitoring from August to November 1989
(Harza, 1990). Fixed air monitoring stations were planned at five locations
along the river: two residential areas, one upwind location, one downwi^
location and a farming receptor location two miles south of the remnant area.
Because of site access problems, only three of the five sites were monitored.
Once every three days, two 1-liter/mlnute, 24-hour air samples were taken using
a two-channel (two separate samples taken simultaneously) air monitoring station
three feet above the ground. The samples were analyzed by NIOS/H Method 5503
(desorption of fluorisll tubes with solvent followed by GC-ECD analysis) with a
detection limit of 0.05 |ig/sample. In total, 84 samples were collected. Seven
samples showed levels of PCBs above the detection limit (0.05 |ig/samp1e), with
a maximum value of 0.23 yg/m\ Of these seven detects, three were from a
residential area (location A2), two from a downwind receptor (location A4) and
one from the farm area (location A5). Although PCBs were detected 1n this
investigation, the two sample channels often gave inconsistent results; one
channel contained PCBs while the other did not. This occurrence may have been
due to sampling or analytical problems or both.
Maximum ambient background PCB concentrations 1n air measured by New York
State during a statewide monitoring effort, listed at the bottom of Table B.3-21,
provide a perspective on PCB levels 1n air in the vicinity of the site. Maximum
ambient air PCB concentrations measured in this effort ranged from 0.002 yg/m3
1n Syracuse and Rensselaer (urban areas) to 0.007 |ig/m3 1n Staten Island (NYSDEC,
1982-4). These maxima are one to two orders of magnitude lower than the maxima
detected during GE's baseline monitoring study and values measured by NYSDEC near
Lock 7 in August 1986.
Lamont-Ooherty Investigations
Three studies of PCB volatilization and properties governing volatilization
were conducted at the Lamont-Doherty Geological Observatory. The first two of
these studies (Bopp, 1979 and Bopp, 1983) dealt with the estimation of PCB
properties that govern volatilization and the application of these properties to
B.3-42
-------
/etals such as mercury, cadmium, chromium, copper and zinc were frequently
.ed. Median mercury and zinc concentrations were on the order of 0.4 ppm
10 ppm, respectively. Organic compounds frequently detected include
clcides such as DOT, heptachlor, and dieldrin. Hexachlorobenzene and
xachlorocyclohexane were also detected frequently.
6.3.4.3 NYSDOH HacroInvertebrate Studies
As part of the Hudson River ?cr.. Reclamation Demonstration Project, the New
York State Department of Health (NYSDOH) conducted biomonitoring studies from
1976 to 1985 using caddisfly larvae, nปU1plate samples and chironomid larvae
(Simpson et a7., 1986). These studies included long-term biomonitoring efforts
from 1976 to 1985 as well as two short-term biological uptrke studies in July and
September of 1985.
Long-Tern Biomonitoring Study
From 1976 through 1985, artificial substrate samplers.(mult1plates) were
*d at 17 sites along the Hudson River from Hudson Falls to Nyack, New York
et <7., 1988). These samplers were collected each year after a period of
ks during the months of July, August and September. PCBs 1n the samples
V s Aroclors 1016 and 1254. The resulting PCB concentrations In
samples represented a composite of sediment, algae, plankton and
nvertebrates. Invertebrates collected In the multlplate samplers
Allowing taxonomlc groups: Chironomidae, Ollgochataeta,
eroptera, Amphlpoda and EHmidae. Chironomid larvae and pupae
vปt Invertebrate component from Fort Edward to Saugerties,
*rcent of the total macrolnvertebrate population at Fort
caddisfly larvae were collected by hand-picking
designated sites: Hudson Falls, Fort Edward, Fort
Caddisfly collections were made in June, July,
**.3-36
-------
Besides accounting for species differences, accounting for lipid content
1s important because fish lipid content appears to have changed from samples
caught from one year to the next. Dividing the PCB concentration (wet weight)
by the measured lipid content (g-11pid/g-f1sh) of the sample, one obtains t^e PCB
concentration per gram (mass) of f1sh-Hp1d (ug-PCB/g-lipid). Looking at all
samples for all species at all Upper Hudson sampling locations on a lipid basis,
the median PCB levels have declined from 1829 ug-PCB/g-lip1d 1n 1977 to 271 ug-
PCB/g-1ipid in 1988, with a 95 percent upper confidence bound estimate on the
mean of 484 (Table B.3-19).
L1pid-based means of Aroclor 1254 and 1016 concentrations by year 1n the
predominant fish species sampled at River Mile 175 are shown in Figure B.3-15.
On a lipid basis, largemouth bass have usually shown the highest Aroclor 1254
levels. This may reflect their position as top carnivores In the aquatic food
chain or their low fat content. Further, the Hpid-based Aroclor 1254 levels in
most species appear to have been relatively constant since 1982. Error bars are
omitted, from the multiple species plots for legibility. Trends in largemouth
bass, with error bars, are shown 1n Figure B.3-16. (An error bar shown as a
vertical line at each year representing the 95 percent upper and lower confidence
bounds on the mean.)
In addition to the monitoring at River M11e 175, there is a good continuous
record of sampling of brown bullhead at River Mile 153 just below the Federal
Dam, for the period 1977-1988. PCB levels here were on average much lower than
those observed at River Mile 175, presumably due to dilution of PCB concentra-
tions by the flow of the Mohawk River. Average lipid-based PCB concentrations
1n brown bullhead show a regular exponential decline for Aroclor 1016 components
and a less dramatic decline for Aroclor 1254 (Figure B.3-17).
B.3.4.2 Other Chemicals In F1sh
The NYSDEC database contained analyses for chemicals in fish other than
PCBs. Summary statistics for these other chemicals are shown in Table B.3-20.
B.3-35
-------
Examination of the complete run of samples at the Fort Miller multiplate
and caddisfly station (PCB-5) shows what appears to be a largely random pattern
of total PCBs on a dry weight basis (Figure B.3-18). The larger mean PCB
concentration for the period 1981-82 is largely attributed to the increased
variability of the data for this period. Adjusting to a lipid basis actually
increases the total variability represented by outlying data points. Only a
slight trend is suggested 1n this figure, where PCB levels may have declined from
1976 to 1980, then remained approximately stable. At Stillwater, the lipid-based
values appear to be almost entirely random (Figure B.3-19).
A comparison between the confidence intervals on the overall means for all
sample years by sample locations (Figure B.3-20) shows differences between the
low values observed above Fort Edward and values at downstream stations. Lipid-
based means at all stations by year (Figure B.3-21) appear to show a decline from
1976 to 1979, then relatively constant levels except for a jump upward in 1982.
Again, this apparent trend 1s difficult to confirm as a consequence of the
limited number of samples and possible differences in sampling or analytical
protocols throughout the duration of the monitoring.
Short-Term Biomonitorlng Study
Short-term biomonitorlng investigations using the chironomld larvae,
Chlronoaus tentans, were also performed'by the NYSDOH during July and September
1985 (Novak et a/., 1990). The monitoring method consisted of placing 25
laboratory-reared chironomld larvae 1n nylon mesh envelopes or packets that were
exposed to the water column. Envelopes were placed, in groups of ten, in steel
mesh baskets at the primary collection site and monitored at 0, 1, 2, 4, 8, 12,
24, 48, 72 and 96 hours. Chlronomids were placed at four sites, including two
at Thompson Island Pool, one at Bakers Falls and one at Fish Creek, and monitored
at 96 hours. Packets of chlronomids exposed to the sediment at a collection site
located on the eastern shore of the Thompson Island Pool were also collected at
96 hours. Water column samples were obtained during the same collection
intervals for each site.
B.3-38
-------
Measured PCB levels 1n the 1985 multlplate samples for September ranged
from 0.25 ppm at Hudson Falls (the control site) to over 6 ppm at Fort Edward.
Multiplate monitoring from Fort Miller to Waterford resulted in PCB levels of 4
to 5 ppm. The multlplate samples at any one site appeared to show a consistent
decline 1n PCB concentrations from early summer to later summer In any particular
year. Larger scale trends or relationships 1n either time or with sample
location are difficult to detect, because of the extremely wide variation 1n the
sample results. Average PCB concentrations in multlplate samples generally
showed a decline from 1976 - 1980. Nevertheless, average PCB concentrations
Increased 1n 1981 and remained high through 1985. Multlplate samples from the
Thompson Island Pool and downstream showed significantly higher PCB concentra-
tions than samples taken upstream of Fort Edward. Yet, no significant trends are
apparent when comparing Fort Edward results with those at Waterford.
The results of the caddisfly biomonitoring efforts show a decline in total
PCB concentrations from 1978 to 1980. As Is true of the multlplate data, spatial
trends are not readily apparent from the caddisfly samples. Measured PCB levels
in macroinvertebrate tissues generally ranged from 20 to 60 ppm (dry weight) In
1979 and from 20 to approximately 40 ppm (dry weight) in 1985. During the same
collection periods, macroinvertebrates collected at Hudson Falls, the control
site, exhibited PCB tissue residues less than 10 ppm.
The data from this study exhibit a great deal of random variability. The
bulk of the data are from multlplates, which collect sediment as well as living
matter. In theory, it should be possible to distinguish sediment versus
biologically-based PCBs by adjusting the observations to a lipid basis; lipid
content of samples was reported. An additional factor contributing to the
variability is that in almost every year a downward trend by month was observed
at most stations, based generally on three samples. The cause of this phenomenon
is not known, but the limited number of observations In this yearly cycle may
have obscured the influence of other factors.
B.3-37
-------
B.3.5 PCB Concentrations In Air and Plants
B.3.5.1 Air
Monitoring Near Fort Edward
A1r monitoring efforts for PCBs and other air toxics have been conducted
in the Upper Hudson River study area from late 1976 to 1982 by NYSDEC/NYSDOH and
various researchers and as recently as 1989/90 by contractors (Harza/Yates-
Auberle) for General Electric.
From January through August of 1977, NYS00H collected air samples at five
locations in the Upper Hudson Valley to determine ambient PCB concentrations.
While the Glens Falls and Uarrensburg samples showed no detections above the
0.020 jig/m3 detection limit, results from the Hudson Falls and the Fort Edward
samples demonstrated high levels of total PCBs, ranging from 0.060 ng/m3 to 3.26
pg/m3 (Malcolm-Pirnie, 1978). Atmospheric levels of PCBs in the Fort Edward area
were reported to decrease from 1 iปg/m3 down to 0.3 |ig/m3 after the cessation of
PCB use by General Electric in their Hudson Falls and Fort Edward capacitor
plants 1n 1977 (Limburg, 1984).
In 1979, NYSDEC conducted an air monitoring survey for PCBs around various
dumps and landfills (Caputo and Fort Miller dumps, Remnant Area, Moreau and Site
3a and Buoy 212) In the Hudson Falls/Fort Edward area bordering the Upper Hudson
River (see Table B.3-21). Values ranged from 5 to 15 |xg/m3 total PCBs at the
Moreau and remnant areas and 24 to 300 yg/m3 total PCBs at the Fort Miller and
Caputo dumps, respectively. At the Caputo dump, where the soil was reported to
contain 5,000 mg/kg PCBs, air monitoring for PCBs before and after capping of the
site showed that average ambient PCB concentrations decreased from 118 iปg/ป3
before capping to 0.26 |ig/m3, once the site was capped (Shen, 1982).
Two air samples taken over Lock 6 in the summer of 1980 yielded Aroclor
1242 concentrations ranging from 0.11 |tg/m* to 0.52 |ig/m3 (Table B.3-21).
Aroclors 1221 and 1254 were not detected (NYSDEC, 1981). Ambient PCB levels
B.3-40
-------
Results of this Investigation Indicated that chlronomlds accumulated PCBs
ranging from 0.1 to 7 ppm after 1 to 96 hours of continuous exposure, whereas
larvae exposed to sediments near the Thompson Island Pool for 96 hours contained
over 100 ppm. Water column PCB levels were in the range of 0.03 - Ovl iป9/I
during the experiment. The ratio of the PCB levels in the chlronomlds (& ppb)
to the ambient PCB concentrations in the water column (defined as the bloaccuB-
ulatlon factor or BAF) were on the order of 10* to 10s.
A significant conclusion from this study was that the congener pattern
found 1n tissues of chironomid larvae differed substantially from the congener
pattern observed in water. Using capillary column gas chromatography, the
investigators were able to isolate PCB congeners in both the water column and
chironomids. The most abundant congeners in chironomid tissues were 2,4,2',5*-
tetrachlorobiphenyl and 2,3,6,4'-tetrachlorobiphenyl. In contrast, the
predominant congeners 1n water were 2,2*-dichlorobiphenyl, 2,6-dichlorobiphenyl,
and 2,6,2'-trichlorobiphenyl. These findings suggest a number of possible
explanations. One explanation could be that chironomid larvae may selectively
bioconcentrate the more highly chlorinated congeners which are present at
relatively low concentrations in water. Another factor which could explain the
observed results is that the lower chlorinated congeners were present 1n the
water but below detection limits.
Although this study presents some very Interesting congener-specific
results, they are too limited 1n scope to provide clear Indications of either
congener-specific bioaccumulatlon or congener-specific comparisons of PCBs in
sediments. (No congener-specific sediment data were obtained.)
B.3-39
-------
Most recently, 1n connection with Its Remnant Remediation Project, GE
conducted baseline pre-reraediation air monitoring from August to November 1989
(Harza, 1990). Fixed air monitoring stations were planned at five locations
along the river: two residential areas, one upwind location, one downwih^
location and a farming receptor location two miles south of the remnant area.
Because of site access problems, only three of the five sites were monitored.
Once every three days, two l-l1ter/m1nute, 24-hour air samples were taken using
a two-channel (two separate samples taken simultaneously) air monitoring station
three feet above the ground. The samples were analyzed by NIClH Hethod 5503
(desorptlon of fluorlsll tubes with solvent followed by GC-ECD analysis) with a
detection limit of 0.05 pg/sample. In total, 84 samples were collected. Seven
samples showed levels of PCBs above the detection limit (0.05 |ig/sample), with
a maximum value of 0.23 pg/m3. Of these seven detects, three were from a
residential area (location A2), two from a downwind receptor (location A4) and
one from the farm area (location A5). Although PCBs were detected 1n this
Investigation, the two sample channels often gave Inconsistent results; one
channel contained PCBs while the other did not. This occurrence may have been
due to sampling or analytical problems or both.
Maximum ambient background PCB concentrations 1n air measured by New York
State during a statewide monitoring effort, listed at the bottom of Table B.3-21,
provide a perspective on PCB levels 1n air 1n the vicinity of the site. Maximum
ambient air PCB concentrations measured 1n this effort ranged from 0.002 yg/m3
in Syracuse and Rensselaer (urban areas) to 0.007 |ig/m3 in Staten Island (NYSDEC,
1982-4). These maxima are one to two orders of magnitude lower than the maxima
detected during GE's baseline monitoring study and values measured by NYSDEC near
Lock 7 in August 1986.
Lamont-Doherty Investigations
Three studies of PCB volatilization and properties governing volatilization
were conducted at the Lamont-Doherty Geological Observatory. The first two of
these studies (Bopp, 1979 and Bopp, 1983) dealt with the estimation of PCB
properties that govern volatilization and the application of these properties to
B.3-42
-------
monitored at farm fields in 1981 near the Hudson River, however, showed low PCB ~
concentrations of approximately.0.005 iปg/m3 (NUS, 1984).
In the early 1980s, NYSOEC and the Boyce Thompson Institute for Plant
Research of Cornell University conducted a joint air/plant monitoring effort near
the tailwater of Lock 6 to determine if volatilization of PCBs from the water
column was occurring (Buckley and Tofflemlre, 1983 and 1984). Between August
1981 and September 1981, seven air samples were taken. Aroclor 1242 was detected
in all seven samples, ranging from 0.031 to 0.06 iig/m1. Aroclor 1254 was
detected in three of the samples at levels up to 0.0013 jปg/m\ During this
study, a vertical PCB gradient was also noticed, when airborne PCBs were measured
simultaneously at heights of 1 and 4.5 meters above the water.
In August 1986, NYSDEC collected three sets of ambient grab samples in
duplicate at the proposed containment site (Site G), the Fort Edward Landfill,
the Bourgoyne Avenue School and Lock 7 of the Champlain Canal (USEPA/NYSDEC,
1987). The highest ambient PCB concentration measured was 0.083 |ig/m3 at Lock
7. Site G and the Fort Edward landfill samples contained PCB levels below the
detection limit of 0.007 ng/m3. The Bourgoyne Avenue School sample information
was not available at the time of the report.
Again in 1987, NYSDEC conducted air monitoring from April 2, 1987 to July
16, 1987 at the Kingsbury Landfill, located north of Fort Edward. In 76 of 105
samples taken over April and Nay 1987, Aroclor 1016/1242 was detected with a
maximum concentration of 0.49 |ig/m3. Aroclor 1248 was detected in 5 of 105
samples with a maximum concentration of 0.52 pg/m3. Both the Kingsbury and Fort
Edward municipal landfills were the burial site of several thousand tons of PCBs.
Neither these landfills, nor others in the area, are part of the current
investigation of the Hudson River Superfund Site. The results are presented here
as evidence of PCBs having been monitored 1n the air within the vicinity of the
site.
B.3-41
-------
PHASE 1 REPORT - REVIEW COPY
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
EPA WORK ASSIGNMENT NO. 013-2N84
AUGUST 1991
ALTERNATIVE REMEDIAL CONTRACTING STRATEGY (ARCS)
Region II
FOR
HAZARDOUS WASTE REMEDIAL SERVICES
EPA Contract No. 66-S9-2001
VOLUME 1
(BOOK 1 OF 2)
TAMS CONSULTANTS, Inc.
and
Gradient Corporation
-------
PART B
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
CONTENTS
Paoe
B. UPPER HUDSON CHARACTERIZATION
SYNOPSIS (Section B.l)
B.l Physical Site Characteristics B.1-1
B.l.l Hydrology B.l-1
B.l.2 Water Quality and Use B.l-2
B.l.2.1 Water Quality B.l-2
B.l.2.2 Use B.l-5
B.l.3 Population and Land Use B.l-7
B.l.4 Fisheries B.l-8
SYNOPSIS (Section B.2)
B.2 Sources of PCB Contamination B.2-1
B.2.1 GE Discharges (To 1977) B.2-1
B.2.2 Current Permitted Discharges B.2-2
B.2.3 Other Sources B.2-2
1
-------
PART B
CONTENTS
(continued)
SYNOPSIS (Section B.3)
B.3 Nature and Extent of Contamination B.3-1
B.3.1 Overview of Sources and Database B.3-1
B.3.2 Sediment B.3-5
B.3.2.1 1976-1978 NYSDEC Sampling B.3-5
B.3.2.2 1984 NYSDEC Sampling B.3-7
Methods and Procedures B.3-7
PCB Results B.3-8
Comparison of 1976-78 and 1984 Studies B.3-10
B.3.2.3 Lamont-Doherty Geological Observatory B.3-11
Investigations
B.3.2.4 Other Studies B.3-12
1983 USEPA Study B.3-12
GE 1989 Baseline Studies B.3-13
(Remnant Remediation Project)
GE 1990 Sediment Sampling B.3-13
(B1oremed1at1on Investigations)
B.3.2.5 Other Chemicals 1n Sediments B.3-14
B.3.2.6 Discussion B.3-15
11
-------
PART B
CONTENTS
(continued)
Page
B.3.3 Surface Water Monitoring B.3-17
B.3.3.1 USGS Flow Records - B.3-17
B.3.3.2 Suspended Sediments Monitoring B.3-19
B.3.3.3 USGS PCB Monitoring B.3-20
Methods and Procedures B.3-20
PCB Results B.3-2I
Current Full-Year Average PCB Concentrations B.3-23
1n Water Column
Summer Average PCB Concentrations B.3-24
B.3.3.4 Other Sources of Water Column Data B.3-25
Waterford Treatment Plant Data B.3-25
Lamont-Doherty Study of 1983 B.3-25
NYSDOH Water Column PCBs B.3-27
Remnant Deposit Containment Monitoring " B.3-27
Program
B.3.4 F1sh and Other Aquatic Biota B.3-28
B.3.4.1 Fish Sampling B.3-29
Samples Prior to 1975 B.3-30
Samples 1975-1976 B.3-31
Samples 1977-1988 B.3-31
11 i
-------
PART B
CONTENTS
(continued)
*
PaoeV
PCB Levels 1n Fish B.3-32
Lipid-Based PCB Concentrations B.3-34
B.3.4.2 Other Chemicals In F1sh B.3-35
B.3.4.3 NYSDOH MacroInvertebrate Studies B.3-36
Long-Term Blomonitorlng Study B.3-36
Short-Term Blomonitorlng Study B.3-38
B.3.5 PCB Concentrations 1n Air and Plants B.3-40
B.3.5.1 Air B.3-40
Monitoring Near Fort Edward B.3-40
Lamont-Doherty Investigations B.3-42
B.3.5.2 PCB Uptake By Plants B.3-43
Early Studies B.3-43
More Recent Studies B.3-45
B.3.6 Other Media B.3-47
B.3.7 Adequacy of PCB and Aroclor Measurement B.3-48
B.3.7.1 Overview B.3-48
PCB - Aroclors, Congeners, and B.3-48
Analysis Methods
Interpreting Reported Aroclor Results B.3-51
1 v
-------
PART B
CONTENTS
(continued)
Page
B.3.7.2 Discussion of Data Quality Assurance B.3-52
1976-1978 Sediment Survey B.3-52
1984 Sediment Survey B.3-55
Other Sediment Data B.3-57
Hudson River Fish Samples B.3-58
USGS Water Column Data B.3-60
B.3.7.3 Summary B.3-61
SYNOPSIS (Section B.4)
B.4 Data Synthesis and Evaluation of Trends B.4-1
B.4.1 Phase 1 Objectives B.4-1
B.4.2 Flood Flow and Sediment Transport B.4-1
B.4.2.1 Flood Frequency Analysis B.4-1
B.4.2.2 Suspended Sediment Discharge B.4-7
Empirical Trend Analysis B.4-10
B.4.3 PCBs in the Water Column and Mass Discharge B.4-11
B.4.3.1 PCB-D1scharge Relationships B.4-12
Regression Analyses B.4-12
Non-parametric Tests of Trend for B.4-16
Water Column PCBs
v
-------
PART B
CONTENTS
(continued)
*
Page V
B.4.3.2 Mass Transport Estimates B.4-17
Regression Approach B.4-18
Direct Estimation Approach B.4-20
B.4.3.3 Discussion of Mass Transport from B.4-26
Upper to Lower River
B.4.4 Analysis of PCBs 1n F1sh B.4-30
B.4.4.1 Evaluation of Time Trends B.4-31
Non-Parametric Trend Test B.4-31
Apparent Aroclor Half-Lives 1n the F1sh B.4-32
Population
B.4.4.2 Projected PCB Concentrations 1n F1sh B.4-35
B.4.4.3 Relation Between PCB Concentrations In B.4-37
Fish and Water
B.4.5 Summary B.4-40
SYNOPSIS (Section B.5)
B.5 Sediment Transport Modeling B.5-1
B.5.1 Overview B.5-1
B.5.2 Previous Modeling Studies B.5-3
v1
-------
PART B
CONTENTS
(continued)
Page
B.5.3 Hydrodynamlc Model Description 6.5-7
B.5.3.1 Use of WASP4 Family of Models B.5-7
B.5.3.2 Governing Equations B.5-7
B.5.3.3 Model Implementation B.5-11
B.5.3.4 Model Setup for Thompson Island Pool B.5-14
B.5.3.5 Model Calibration B.5-16
B.5.4 Sediment Transport Model B.5-18
B.5.4.1 Streambed Erosion and Deposition B.5-18
Sediment Transport Capacity B.5-19
Active Bed Layer B.5-20
Bed Erosion B.5-21
Sediment Deposition B.5-23
Bed Elevation Change B.5-24
B.5.4.2 Streambank Erosion B.5-24
Sediment Routing B.5-27
B.5.4.3 Initial Calibration Efforts B.5-30
B.5.5 Summary B.5-30
vii
-------
PART B
CONTENTS
(continued)
Emป\
SYNOPSIS (Section B.6)
Preliminary Human Health Risk Assessment B.6-1
B.6.1 Phase 1 Objectives B.6-1
B.6.2 Exposure Assessment B.6-2
B.6.2.1 Introduction B.6-2
B.6.2.2 Dietary Intake B.6-4
Fish Consumption B.6-4
Snapping Turtles B.6-10
Ingestion of Agricultural or Home Garden Crops B.6-11
Ingestion of Beef or Dairy Products B.6-12
Breast Milk B.6-12
Drinking Water B.6-14
B.6.2.3 Inhalation Exposures B.6-15
Exposure From A1r B.6-15
B.6.2.4 Recreational Exposures B.6-16
Dermal Absorption From Contact With Sediments B.6-16
Incidental Ingestion of River Sediments B.6-19
Dermal Absorption From Vater B.6-20
viii
-------
PART B
CONTENTS
(continued)
ฃasฃ
8.6.3 Toxicity Assessment B.6-23
B.6.3.1 Introduction B.6-23
B.6.3.2 Noncarcinogenic Effects B.6-24
B.6.3.3 Carcinogenic Effects B.6-26
Definition B.6-26
USEPA Cancer Slope Factor B.6-27
Other Cancer Slope Factor Studies B.6-28
B.6.3.4 Toxicity of Specific PCB Congeners B.6-30
B.6.3.5 Epidemiological Studies B.6-31
Cancer Effects B.6-32
Non-Cancer Effects B.6~32
B.6.3.6 Other Health-Based Regulatory Limits B.6-33
or Guidelines
FDA: Tolerance for PCBs in Fish B.6-33
USEPA: Drinking Water B.6-34
USEPA: Ambient Water B.6-34
New York State: Ambient Water B.6-35
USEPA Advisory Levels for PCB Superfund B.6-35
Clean-up
National Academy of Sciences: Suggested B.6-36
No-Adverse-Response Level
1x
-------
PART B
CONTENTS
(continued)
Eaas V
Standards for Occupational Exposures 6.6-36
>
New York State: Ambient Air Guidelines. 6.6-37
6.6.4 Risk Characterization 6.6-37
6.6.4.1 Definition 8.6-37
6.6.4.2 Dietary Intake 8.6-39
Fish 8.6-39
Drinking Water B.6-39
Other 8.6-40
B.6.4.3 Inhalation Exposures B.6-40
6.6.4.4 Recreational Exposures 8.6-40
Dermal Exposure to River Sediment B.6-40
Incidental Ingestion of River Sediment B.6-41
Contact With River Water 6.6-41
8.6.4.5 Risk Characterization Compared 8.6-41
to Human Studies
B.6.4.6 Analysis of Uncertainties B.6-42
8.6.5 Lower Hudson Discussion B.6-45
x
-------
PART B
CONTENTS
(continued)
Paoe
SYNOPSIS (Section B.7)
B.7 Interim Ecological Risk Assessment B.7-1
B.7.1 Phase 1 Objectives B.7-1
B.7.2 Ecosystem Description B.7-2
B.7.2.1 Terrestrial Habitats B.7-2
Habitats B.7-2
Vegetation B.7-3
Birds B.7-3
Mammals B.7-4
Amphibians and Reptiles B.7-4
Threatened and Endangered Species B.7-5
B.7.2.2 Aquatic Ecosystem B.7-6
Conceptual Ecosystem Framework B.7=-7
Phytoplankton B.7-8
Periphyton B.7-8
Nacrophytes B.7-9
Invertebrate Community B.7-9
Fish B.7-14
Summary of Aquatic Ecosystem B.7-16
xi
-------
PART B
CONTENTS
(continued)
%
PaoeV
6.7.3 PCB Exposure Assessment B.7-19
B.7.3.1 Exposure Pathways B.7-19
B.7.3.2 Identification of Indicator Species B.7-21
B.7.3.3 Exposure Quantification B.7-22
Ambient Water and Sediment Exposures B.7-23
Ch1rononm1d Larvae B.7-24
F1sh B.7-25
Estimated F1sh Dietary Intake B.7-27
Herring Gull B.7-28
Mink B.7-30
B.7.4 Toxicity Assessment B.7-30
B.7.4.1 Types of Toxicity B.7-30
B.7.4.2 Toxicity Literature Review B.7-33
Plants B.7-33
Planktonlc Species B.7-34
Aquatic Hacrolnvertebrates and Insects B.7-34
F1sh B.7-35
Birds B.7-36
Mammals B.7-37
x1i
-------
PART B
CONTENTS
(continued)
Page
B.7.4.3 Proposed Criteria and Guidelines B.7-38
Ambient Water Quality Criteria B.7-38
Sediment Quality Guidelines B.7-39
Guidelines for PCBs in Fish B.7-40
Guidelines for PCBs in Birds B.7-41
Guidelines for PCBs in Mammals B.7-42
B.7.5 Risk Characterization B.7-42
B.7.5.1 Ambient Water B.7-42
B.7.5.2 Sediment B.7-43
B.7.5.3 Fish B.7-43
B.7.5.4 Fish-Eating Birds B.7-45
B.7.5.5 Mammals B.7-46
B.7.5.6 Summary B.7-46
xiii
-------
PART B
CONTENTS
(continued)
TABLES
Tables Located 1n Volume 1 (2 of 2)
B.1-1 Hater Quality Rating Criteria
B.l-2 Public Water Supplies on the Upper Hudson River
B.l-3 Fish Species Occurrence Summary Between Fort Edward and the Federal
Dam
B.2-1 Current Permitted PCB Discharges, Upper Hudson River Drainage Basin
B.2-2 Inactive Disposal Sites Located Near Upper Hudson River
B.3-1 Studies of PCB Contamination 1n the Hudson
B.3-2 Hudson River Sediment Database Summary
B.3-3 Comparison of Sediment Samples By River Mile
B.3-4 PCB Levels 1n 1976-1978 Sediment Samples
B.3-5 1984 Thompson Island Pool Sediment Summary
B.3-6 Texture Classifications From 1984 Sediment Study
B.3-7 GE Baseline Remnant Remediation Sediment Monitoring
B.3-8 Total PCBs 1n Sediments - GE's 1990 Study and Comparison to Earlier
Studies
B.3-9 Dally Flows For Upper Hudson USGS Gauging Stations
B.3-10 Suspended Sediment Monitoring Summary
B.3-11 Total PCBs In the Water Column - USGS Stations
B.3-12 Current (1986-89) Average Water Column PCB Concentrations - USGS
Stations
xlv
-------
PART B
CONTENTS
(continued)
6.3-13 Summer Average Water Column PCB Concentrations (|ig/1) - USGS
Monitoring Stations
B.3-14 Upper Hudson Yearly Fish Count
B.3-15 Average Aroclor Levels 1n Upper Hudson F1sh
B.3-16 Total PCBs (ppm) 1n Largemouth Bass: Upper Hudson - NYSDEC
Monitoring
B.3-17 Total PCBs (ppm) 1n Pumpkinseed: Upper Hudson - NYSDEC Monitoring
B.3-18 Total PCBs (ppm) 1n Brown Bullhead: Upper Hudson - NYSDEC Monitoring
B.3-19 Lipid-Based Total PCBs for All F1sh Species: NYSDEC Database
B.3-20 Other Chemicals 1n Fish Samples
B.3-21 PCBs 1n A1r
B.3-22 PCBs 1n Plants
B.4-1 Flood Recurrence Intervals at Fort Edward
B.4-2 Regression Analysis: PCBs In Water Column
B.4-3 Published PCB Mass Loading Past Waterford (kg/yr)
B.4-4 Estimated TAMS/Grad1ent Yearly Average PCB Loads (kg/yr)
B.4-5 Trends 1n Aroclor Concentration at River Mile 175 (|ig/1)
B.6-1 Exposure Assumptions: Fish Ingestion
B.6-2 Exposure Assumptions: Dermal Contact with Sediments
B.6-3 Exposure Assumptions: Sediment Ingestion
B.6-4 Exposure Assumptions: Dermal Contact with River Water
B.6-5 Cancer Risk Estimates
xv
-------
PART B
CONTENTS
(continued)
B.6-6 Hazard Quotient Estimates
B.6-7 Epidemiological Studies: PCB Carcinogenicity 1n Humans
B.6-8 Epidemiological Studies: Non-Cancer PCB Effects 1r Humans
B.7-1 Estimated Ecological PCB Exposure Levels for Indicator Species
B.7-2 Summary of Observed PCB Effects 1n Biota
B.7-3 Summary of Proposed Ecological Guidelines for PCBs
FIGURES
Figures Located 1n Volume 1 (2 of 2)
B.1-1 Mean Monthly Flow 1n the Upper Hudson River, Water Year 1986
B.1-2 Mean Monthly Flow in the Upper Hudson River, Water Year 1984
B.2-1 Reported General Electric PCB Usage
B.3-1 Total PCBs 1n Surface Sediments - 1976-78
B.3-2 PCB Concentration vs. Texture Relationship - Gravel
B.3-3 PCB Concentration vs. Texture Relationship - Fine Sand
B.3-4 PCB Concentration vs. Texture Relationship - Find Sand/Wood Chips
B.3-5 PCB Concentration Frequency Comparison
B.3-6 Correlation of Sediment Aroclor 1242 Levels in Upper and Lower
Hudson Sediment Cores
B.3-7a Upper Hudson Dally Average Flows, 1973-1981
B.3-7b Upper Hudson Dally Average Flows, 1982-1990
xv 1
-------
PART B
CONTENTS
(continued)
B.3-8 Total PCBs In Water Column: Fort Edward
B.3-9 Total PCBs 1n Water Column: Schuylervllle
B.3-10 Total PCBs 1n Water Column: Stillwater
B.3-11 Total PCBs 1n Water Column: Waterford
B.3-12 Summer (June - September) Average PCB Concentrations 1n Water
B.3-13 Mean Total PCBs 1n Brown Bullhead
B.3-14 Mean Aroclor Trends 1n Fish (River Mile 175)
B.3-15 Trends In Mean Upid-Based Aroclor Levels 1n Fish (River Mile 17S)
B.3-16 L1p1d-Based Aroclor Trends: Largemouth Bass, River Mile 175
B.3-17 Upid-Based Aroclor Trends: Brown Bullhead, River Mile 153
B.3-18 Total PCBs in Multiplate/Caddisfly - Fort Miller
B.3-19 Total PCBs in Multiplate/Caddisfly Data, PCB-7, Stillwater
B.3-20 PCBs in Multiplate/Caddisfly - All Stations
B.3-21 PCB Trends 1n Multiplate/Caddisfly Data
B.3-22 Gas Chromatogram Peaks for Three Aroclor Standards
B.4-1 Conceptual Reassessment Framework
B.4-2 Upper Hudson Flow Duration Curve
B.4-3a Comparison of Estimated and Measured Flows at Hadley
B.4-3b Annual Maximum Daily Flows Below Sacandaga River
B.4-4 Suspended Sediment Rating Curve: Fort Edward at Rogers Island,
1975-1989
xvii
-------
PART B
CONTENTS
(continued)
B.4-5 Suspended Sediment Rating Curve: Hudson River at Schuylervllle
B.4-6 Suspended Sediment Rating Curve: Hudson River at Stillwater
B.4-7 Suspended Sediment Rating Curve: Hudson River at Wrterford
B.4-8 Sediment Load, Hudson River at Fort Edward
B.4-9 Sediment Load, Hudson River at Schuylervllle
B.4-10 Flows at Fort Edward and PCBs at Fort Edward
B.4-11 Flows at Fort Edward and PCBs at Schuylervllle
B.4-12 Total PCBs 1n Water vs. Flow: Fort Edward
B.4-13 Total PCBs in Water vs. Flow: Schuylerville
B.4-14 Total PCBs in Water vs. Flow: Stillwater
B.4-15 Total PCBs 1n Water vs. Flow: Waterford
B.4-16 Suspended Solids vs. Total PCBs: Stillwater
B.4-17 PCB Load at Non-Scour1ng Flows, Stillwater, 1983
B.4-18 Flow-PCB Observation Pairs: Stillwater
B.4-19 PCB Hass Transport Corrected Mean Method Estimates
B.4-20 PCB Mass Transport Past Waterford: Corrected Mean Estimates
B.4-21 PCB Mass Transport, Fort Edward and Stillwater: Corrected Mean
Estimates
B.4-22 Estimated PCB Load Past Waterford
B.4-23 Aroclor 1016 1n Largemouth Bass (Lipid): River Mile 175
B.4-24 Simulated Average Total PCBs in Fish: Upper Hudson River, 1991-2020
xviii
-------
B.4-25
B.4-26
B.4-27
B.4-23
B.4-29
B.5-1
B.5-2
B.5-3
B.6-1
B. 11
B. 12
B. 1-3
B.l-4
B.3-1
PART B
CONTENTS
(continued)
Total PCBs in Yearling Pumpkinseed vs Sumner PCB Concentrations in
the Water Column at Stillwater
Aroclor Levels in Yearling Pumpkinseed vs. Summer Water-Column Total
PCBs
Total PCBs In Yearling Pumpkinseed vs. Summer PCB Concentrations in
the Water Column at Schuylerville
Total PCBs In Largemouth Bass vs. Summer PCB Concentrations 1n Water
Column at Stillwater
Total PCBs in Brown Bullhead vs. Summer PCB Concentrations 1n Water
Column at Stillwater
Model Nodes and Links
Nodal Areas
Preliminary Hydraulic Calibration, 1-D Model, Thompson Island Pool
Potential Exposure Pathways
PLATES
Plates Located in Volume 1 (2 of 2)
Upper Hudson River USGS Monitoring Stations
Upper Hudson River Water Surface Profile
Upper Hudson River Surface Water Classifications
Upper Hudson River Land Use
Upper Hudson River Sediment Core Locations
xix
-------
PART B
CONTENTS
(continued)
PAGE INTENTIONALLY LEFT BLANK
xx
-------
SYNOPSIS
PHYSICAL SITE CHARACTERISTICS
(Section B.1)
The hydrology of the Upper Hudson River is described in detail (B.l.l), building upon
the basic description of Hudson River hydrology given in Part A. Flow characteristics of the
Upper Hudson generally show a strong seasonal dependence, with maximum flows during the
annual spring thaw. Flows are partially regulated by wetlands as well as the Sacandaga
Reservoir. Four major tributaries to the Upper Hudson below Fort Edward combine with flow
upstream to produce an average annual flow of 7100 cfs at Waterford, above its confluence with
the Mohawk. This basic flow regime governs the transport ofPCBs in the Upper Hudson.
Water quality (B.1.2) is described according to New York State water quality
classifications assigned on the basis of "best usages" and the results of water quality sampling
data from a 1987-1988 survey. The water quality of the Upper Hudson at Fort Edward and
Schuylerville was rated as poor, partly because of the fishing ban, as a result of historic PCB
discharges.
Only the town of Waterford draws its drinking water from the Upper Hudson below Fort
Edward. More commonly, the river water is used for industrial and commercial purposes, such
as power-generation, and for domestic and agricultural use, such as watering lawns, gardens or
crops. The Upper Hudson River is a navigational waterway; from Waterford to Fort Edward,
it is co-incident with the Cham plain Canal
Population in the four counties bordering the Upper Hudson between Albany and Glens
Falls is over half a million and land use is predominantly agricultural (B.1.3). Dairy farming
is the principal form of agriculture. The region is also host to a number of industries, generally
located in the vicinity of the population centers.
The Upper Hudson River represents a diverse fisheries resource (B.1.4); six fish surveys
from 1933 to 1985 are reviewed. In general, these surveys show that the majority offish species
historically found in the Upper Hudson continue to reside there. The data also indicate a
qualitative improvement in the fisheries resource.
-------
PAGE INTENTIONALLY LEFT BLANK
-------
B. UPPER HUDSON CHARACTERIZATION
B.l Physical Site Characteristics
B.l.l Hydrology
The Upper Hudson River flows southerly from its source at Lake Tear-of-the-
Clouds near Mt. Marcy 1n the Adlrondacks to Its confluence with the Mohawk River
near the Federal Dam at Green Island, Troy, NY. The drainage area of this
segment, shown In Plate A.1-1, 1s 4,630 square miles (Wagner, 1982). (Sooe
discussion of overall Hudson River basin hydrology, presented 1n Part A, Is not
repeated here.)
The Upper Hudson River drains a major portion of the southern and central
Adlrondacks. Along Its course, the main channel 1s Intersected by several
tributary branches, the most significant of which are the Sacandaga River, the
Batten Kill, the Fish Creek and the Hooslc River (see Plate B.1-1). Water flow
In this segment Is regulated by several dams on the Hudson itself (see Plate B.l-
2), as well as on tributary branches. Flow Is further controlled by abundant
wetlands located throughout the basin, which act as a buffer for high and low
flow conditions.
The total mean annual fresh water flow from the Upper Hudson at its
confluence with the Hohawk near Waterford Is about 7,100 cfs. This flow
represents more than a two-fold Increase 1n flow from that at Hadley, NY. Before
the river reaches the Bakers Falls - Fort Edward area, 1t 1s joined by the
Sacandaga River, the largest single tributary 1n this area. The mean annual flow
at Fort Edward Is roughly 3,800 cfs, about 54 percent of the total flow at
Waterford. Downstream of this location, the remaining tributaries are fairly
evenly spaced, at roughly 10 to 15 miles between tributary junctions. The
combined total of these tributaries doubles the flow of the Upper Hudson by the
time 1t reaches Waterford. Of particular Importance 1s the Hooslc River, which
represents about 15 percent of the total drainage area south of Hadley.
B. 11
-------
Flow in the Hudson Basin is seasonally dependent, with flow patterns
similar to those seen throughout the basin. The typical regime is one of fairly
steady flow throughout nine months of the year. During the spring, flows 1n the
Upper Hudson increase substantially in response to the melting of winter sntM.
The maximum flow at Waterford occurs one month before the maximum at Fort Edward;
melting of winter snows in the southern portion of the Upper Hudson basin tends
to occur earlier than in the portion of the Upper Hudson basin above Fort Edward.
Figure B.1-1 shows mean monthly flows for water year 1986 at Fort Edward and
Waterford, respectively. This seasonal pattern is duplicated in i-ne flows at the
Federal Dam at Green Island for the same year.
Flows for an atypical water year (1984) are shown in Figure B.1-2 for Fort
Edward and Waterford, respectively. That water year was characterized as having
an unusually warm winter with many major storm events throughout the year. The
flow patterns at Fort Edward and Waterford in that year were also present at the
Federal Dam and on the Wallkill River, implying that these unusual conditions
were felt throughout the entire Hudson basin.
B.1.2 Water Quality and Use
B.1.2.1 Water Quality
New York State has classified Its surface waters according to "best usages"
and has established numerical water quality criteria (standards) to which those
waters should conform. Waters that conform to the numerical criteria are
considered suitable for their Intended best use. Water quality classifications
applicable to the Upper Hudson are listed 1n New York State's environmental
regulations (6NYCRR700, et seq.) and are Illustrated on Plate B.l-3. In summary,
the Upper Hudson has assigned to Its different reaches the following classes and
uses:
Class A - drinking water (water supply)
Class B - primary contact recreation (swimming), fishing, fish propagation
Class C - fishing, fish propagation, swimming
Class D - fishing, fish passage, swimming
B.l-2
-------
Numerical quality standards for each of the above classifications are fotmd
In the State's rules at 6NYCRR700-705. A recent NYSDEC Technical and
Operational Guidance Series (TOGS) memo (September 25, 1990) has augmented the
rules, particularly with regard to toxic constituents. The standards encompass
conventional pollutant parameters, such as collform levels, dissolved oxygen,
turbidity and pH, as well as toxic constituents, such as heavy metals and
organlcs. The toxic substance standards have been derived from health and
environmental risk assessments performed either at the state level or derived
from those performed by USEPA.
Historically, the state has used Information from diverse sources to
ascertain the condition of Its surface waters and to evaluate their attainment
of the designated best uses. Currently, a program of rotating intensive basin
studies (RIBS) exists whereby NYSDEC monitors all surface waters on a six year
cycle and uses data from other programs to provide continuity of Information when
RIBS sampling Is not occurring. An Intensive basin survey was conducted within
the Upper Hudson during 1987 and 1988 and Initial reports from that effort have
now become available (NYSDEC, 1990).
As part of the Upper Hudson RIBS effort, NYSDEC evaluated water column
conditions, toxics 1n bottom sediments and contaminant uptake by macroln-
vertebrates and fish. Table B.1-1 provides the six main parameters/media used
1n the RIBS program to rate water quality. Samples were collected at five
locations along the river's main stem at North Creek, Corinth, Fort Edward,
Schuylervllle and Waterford. Although NYSDEC did not report direct conclusions
concerning attainment of water quality standards 1n Its RIBS document, a
qualitative evaluation of overall conditions within particular river reaches Is
provided. Conclusions pertinent to those reaches Incorporating the Fort Edward
and Schuylervllle sampling locations are summarized here.
Both the Fort Edward and Schuylervllle water column samples exhibited
elevated copper levels at a sufficient frequency to warrant considering copper
a parameter of concern. Similarly, Iron was found to be a parameter of concern
at Schuylervllle. No other trace constituents were detected at elevated
B. 13
-------
concentrations with sufficient frequency to be considered parameters of concern.
Copper, a pervasive constituent of Upper Hudson River water, is also found 1n
samples from the relatively pristine North Creek and Corinth river reaches. While
sediment samples were obtained and analyzed from the Schuylervllle location, suc^i
samples were not obtained at Fort Edward. According to RIBS, Schuylervllle
sediments were at the upper limit of background for cadmium and were slightly
above background for lead and mercury. NYSDEC did not identify the background
levels upon which their conclusions were based.
The 1987-1988 RIBS program found no water column toxicity to CerlodtphnU
at either Fort Edward or Schuylervllle. Tissue from caddlsflles collected at the
two sites did not exhibit elevated levels of either heavy metals or of PCBs, in
contrast to previous studies. Apparently macroInverterbrate tissue collected In
1987 at Hudson Falls (location not specified) did exhibit elevated lead and
manganese levels. At both Fort Edward and Schuylervllle, NYSDEC assessed overall
water quality as poor, based on the RIBS results and the fishing ban, a result
of historic PCB discharges to this river reach. At one point the RIBS document
concludes that water quality 1s rated as poor primarily due to the fishing ban
(NYSDEC,1990,p.77).
As a result of several amendments to the Clean Water Act, states are
required to report In specific terms conditions of their surface waters. Clean
Water Act Section 305(b) mandates that states submit water quality condition
reports to USEPA every two years. The 305(b) report evaluates surface waters 1n
relationship to their ability to sustain primary contact recreation and fish
propagation uses. In addition, Clean Water Act Section 304(1) requires that
states generate lists of surface waters that fall to meet water quality standards
because of toxic pollutants, 1n general, and toxics from point sources, 1n
particular.
New York's most recent 305(b) report was published 1n April 1990. That
document provides, In part, a summary of waters wherein contamination 1n fish
exceeds either FDA levels or other guidelines. For the Upper Hudson reach from
Hudson Falls to Federal Dam, the only contaminant Identified as exceeding either
B. 1-4
-------
the FDA levels or other guidelines 1s PCB (NYSDEC, 1990, Table 20). The 305(b)
report also Identifies PCB In sediment as the sole toxic that 1s responsible for
use Impairment 1n the Upper Hudson (NYSDEC, 1990, Table 17). Similarly, the
state's 304(1) lists Identify priority organlcs (PCBs) as being the toxic
responsible for Upper Hudson use Impairment.
B.1.2.2 Use
The Hudson River 1s used as a source for public water supplies (municipal
and Institutional drinking water) 1n sections of the river classified as Class
AA or A. Along the Upper Hudson, three communities draw directly Hudson River
water. Of these, Queensbury and Uaterford have current average uses of more than
1 mgd (NYSDOH, 1991), as shown 1n Table B.l-2. The Waterford Intake Is located
at the base of the Upper Hudson Basin near Lock 1. The Queensbury Intake Is
located near Sherman Island Dam 1n Warren County. The Wlnebrook Hills Hater
District, the third Upper Hudson water supply drawing from Hudson, 1s located at
the headwaters of the Hudson In Newcomb, Essex County.
A more common use of Hudson River water 1s for Industrial and commercial
purposes such as cooling, manufacturing processes and fire protection. Hudson
River water 1s also extensively used for hydroelectric and thermal power
generation. An Inventory of facilities and plants that utilize Hudson River
water can be found 1n reports for the Hudson River-Black River Regulating
District (Malcolm P1rn1e, 1984a) and for the NYSDOT (1984).
Hudson River water 1s also used for domestic (watering lawns and gardens)
and agricultural purposes (Irrigating crops). There are currently no records of
water withdrawal for agricultural uses. Unlike the other Intakes, permits are not
needed to withdraw water from the Hudson for Irrigation purposes (pers. comm.,
NYSDEC and NYSDOH, 1991).
The NYSDEC Division of Water, Source Surveillance Section provided a
listing (March 7, 1991) of all significant active facilities With SPDES permits
1n the Upper Hudson River Basin. This search revealed that 27 facilities
B. 15
-------
discharge into the Upper Hudson Basin, with 15 discharging directly Into the
Hudson River. Five of the 15 facilities are municipal wastewater treatment
plants, Including Corinth Sewage Treatment Plant, Glens Falls Wastewater
Treatment Plant, Saratoga County Sewer District #1 Wastewater Treatment Plant ^
Mechanlcvllle, Stillwater Sewage Treatment Plant and Washington County Sewer
District #2 Water Pollution Control Plant at Fort Edward.
The Champlaln Canal is coincident with portions of the Hudson River; 1t
extends from Waterford, New York on the Hudson River to Whitehall at the southern
end of Lake Champlaln. The Champlaln Canal 1s part of the New York State Barge
Canal System, also comprised of the Erie Canal, Oswego Canal and Cayuga-Seneca
Canals. This network of waterways connects the Atlantic Ocean with the Great
Lakes and the Saint Lawrence Seaway. The Champlaln Canal 1s 60 miles in length,
Including 37 miles of canalized Hudson River from Waterford to Fort -Edward and
23 miles of land-cut sections. The canal diverges from the river at Fort Edward
just below Lock 7 and proceeds 1n a northeasterly direction to Lake Champlaln.
Additional land-cut areas exist at Stillwater, Northumberland and Fort Miller.
Natural flows provide a considerable portion of the water supply needs for
the Hudson River portion of the Champlaln Canal. The Hudson River at Fort Edward
provides an average discharge of 5,244 cfs (USGS average for 1979-1990 water
years) to the canal at the confluence below Lock 7. This flow 1s significantly
influenced by regulation of flows from Great Sacandaga Lake. Below this point,
the river is canalized to provide natural flows for the canal. Water is also
supplied to the Champlaln Canal via the Glens Falls Feeder Canal. The Feeder
Canal diverts approximately 100 cfs of water from the Hudson River upstream of
the Feeder Dam west of Glens Falls to the Champlaln Canal Summit Level between
Locks 8 and 9 on the canal near Smith's Basin. The Summit Level 1s the point of
highest elevation on the canal and allows for the gravity flow of water in both
directions as lockages are made (south to the Hudson River, north to Lake
Champlaln). Approximately 25 percent of the diverted water returns to the Hudson
River (Malcolm Plmie, 1984).
B.l-6
-------
Commercial traffic has declined on the Champlaln Canal and other canals 1i>
the Barge Canal system as a result of "the unreliability of the system for
waterway transport" and competition from other modes of transportation for bulk
products (Malcolm P1rn1e, 1984b, US Army Corps of Engineers, 1977). Unlike the
other three canals 1n the system, the Champlaln Canal shows a steady decline in
recreational use along the entire stretch of the canal on both the canalized
Hudson River and the land-cut section north of Fort Edward (Malcolm Plrnte,
1984).
B.1.3 Population and Land Use
Four counties (Albany, Washington, Rensselaer and Saratoga) He adjacent
to the Upper Hudson between Albany and Glens Falls. All counties experienced
growth between 1980 and 1990 with Saratoga having the greatest Increase over the
period and Rensselaer the lowest. Total population of these counties 1n 1990 was
over 500,000.
Land use within a zone adjacent to the Upper Hudson River, depicted on
Plate B. 1-4, 1s mostly agricultural. Portions He within New York State
Agricultural Districts and include parcels considered to be prime farmland.
Dairy farming 1s the major agricultural Industry. The majority of the crops
grown, such as corn and hay, are used for forage; small quantities of cash crops,
such as oats and wheat, are produced. Industrial use 1s typically located near
population centers. Major non-agricultural Industries within the study area
Include: an Industrial demolition company; several paper mills; hydroelectric
plants; a grocery warehouse; and manufacturers of garden equipment, brake
linings, brushes, paints, wallpaper, paper products, gun barrels, silicone
products, abrasives, brass fittings and clothing. Forested and recreational land
uses are scattered.
Existing recreational uses Include Schaghticoke Canal Park at Lock 4 of the
Champlaln Canal and two town parks, which He along the river 1n Fort Edward.
Proposed for the Fort Edward area are a marina, to be located on the south end
of Rogers Island, and a marina, trails and picnic areas to be located one mile
B.l-7
-------
south of Fort Edward on the former Champlaln Canal. Saratoga National Historic
Park lies on the western bank of the River In the Town of Stillwater. Moreau
State Park 1s located south of 61 ens Falls. At the confluence of the Mohawk and
Hudson Rivers are Peebles Island State Park and the Van Schalk Island Country
Club. Several parks and/or country clubs also front the River.
B.1.4 Fisheries
Fishery resources within the Hudson River from Federal Dair to Fort Edward
are Influenced by different physical features, as well as man-Bade structures
such as locks, dams, guard gates, bridges, spillways and submerged power lines
and cables. This array of physical features produces the following variety of
different fish habitats:
Outlets of streams and rivers;
Shallow water areas (wetland and non-wetland);
Designated ship channels of the canal (canalized river);
Steep embankment areas with relatively swift current;
Landcut (artificial) portions of the canal;
Wet dumping grounds or spoil areas; and
Various alternate channels separated from the main channel
by an Island.
This wide variation 1n habitats expands spatial heterogeneity and results In a
complex fishery resource.
The New York State Conservation Department (Greeley, J.R. and Bishop, 1933)
conducted an early and comprehensive fish Inventory of the portion of the Hudson
River between Hudson Falls and the mouth of the Hooslc River. Forty-one species
of fish (Table B.I-3) were recorded. (The American shad, one might note, was
listed as extinct.) The historical data developed as part of this 1933 fish
Inventory documented that there was an Imbalance between the juvenile and adult
B.l-8
-------
game fish. There appeared to be abundant juvenile game species from a few Biles
below Fort Edward to the mouth of the Hooslc River, but adult populations of game
fish were uncommon.
An anonymous report, prepared by the Conservation Department 1n June 1960
(see Table B.1-3) at the request of the Stillwater Rod and Gun Club, are the only
fish data that exist from 1933 to 1960. This 1960 report contains the
observation that game fish resources declined between 1949 to 1959. Prior to
1949, angling success was supposedly satisfactory for "...bass, walleyes,
northerns and pickerel." Several theories for this decline were advanced, such
as Industrial pollution from the Glens Falls and Fort Edward areas, effluent
process changes at a Glens Falls paper and pulp mill, expanded boating activity
and over-exp1o1tat1on.
The next major fish Inventory was conducted by Lane (1970) who found 13
species of fish. Lane stated the somewhat remarkable conclusion that: "The
collection data Indicate the absence of any significant fishery between Lock No.
1 and Fort Edward." With the exception of goldfish, few larger fish were found.
Game fish collected 1n the river channel were represented largely by juvenile
populations. Similar to 1933 observations, Lane noted a diminished amount of
aquatic vegetation In the river channel. Although 1t 1s generally recognized
that an adequate supply of aquatic vegetation Is Important for adult fish
population maintenance and productivity, few historical and/or current data are
available. One of the most unexpected results of Lane's survey was the absence
of common carp. Although Lane concluded that "...conditions within the study
area are not suitable for carp," later studies (Shupp, 1975; Makarewicz, 1983)
do document the presence of the common carp within this section of the Upper
Hudson.
Subsequently, NYSDEC began collecting fish for PCB analysis. Shupp (1975)
collected fish samples 1n the 40-m1le stretch of the Upper Hudson from Lock No.
1 to Hudson Falls and found an Improvement, which he attributed to upgrading of
treatment facilities and tougher regulations concerning Industrial discharges.
Although Shupp reported approximately 24 species of fish compared to the 13
B. 1-9
-------
species listed by Lane (1970), the fishing from Lock No. 1 to Hudson Falls was
still considered poor, because of the overall low standing crop of fish and low
numbers of adult fish compared to juveniles (Shupp, 1975). The reported
preponderance of juvenile fish was similar to data from the 1933 and 1^0
surveys. Sheppard (1976) Indicated that "...some unknown factor Is causing the
exodus or demise of the mature segment of certain fish populations Including the
rock bass, pumpklnseed, yellow perch, walleye and chain pickerel." NYSDEC (R.
Sloan, per. comm., 1991) has recently observed a greatly diminished number of
both pumpklnseed and yellow perch populations during routine PCS'assessments of
resident fish 1n the Upper Hudson (Fort Edward to Federal Dam);
Since 1975, NYSDEC has continued to collect fish between Federal Dam and
Fort Edward as part of their ongoing assessment and monitoring of PCB levels In
fish flesh. The principal species collected and analyzed within this reach have
been the brown bullhead, goldfish, largemouth bass, pumpklnseed and yellow perch.
Because of the demise of the yellow perch and goldfish, current collection
efforts have focused on the brown bullhead, common carp and largemouth bass (R.
Sloan, per. comm.).
One of the most extensive fishery surveys since the 1933 survey was
conducted approximately eight years ago by Makarewlcz (1983). He surveyed 85
stations along the entire length of the Hudson between Federal Dam and Whitehall
as part of the New York State Barge Canal Maintenance Dredging Program 1985-1995
for NYSDOT (Malcolm Pirnle, 1984b). The sampling stations included nine sampling
reaches from Federal Dam to Fort Edward. A total of 46 species, Including four
migratory species (American eel, blueback herring, sea lamprey and striped bass),
*
were found. Of the 42 resident freshwater species, the panflsh ere the most
prevalent (40 percent); demersal fish were second In abundance (22 percent);
forage fish were the third most abundant group (14 percent); and game fish had
the lowest relative abundance (9 percent). Dominant panflsh members were
blueglll, pumpklnseed, rock bass and yellow perch; demersal dominants were black
bullhead and brown bullhead, common carp and white sucker; forage dominants were
golden shiner, spotfin shiner and spottail shiner; and game fish dominants were
B.1-10
-------
the largemouth bass and sraallmouth bass. Collectively, these 13 species
accounted for 85 percent of all resident freshwater species collected.
The most recent fish survey data available are from a study conducted by
Green (1985) between Stillwater (Lock 4) and Schuylervllle (Lock 5), covering
approximately 13 river miles. Fewer overall species were taken (20) compared to
the more Intensive biological surveys 1n 1933 (41 species) and 1983 (46 species),
both of which covered more of the river (Fort Edward to Federal Dam) than that
covered 1n 1985 (Stillwater to Schuylervllle).
Length at age comparisons for both smallmouth and largemouth bass Indicate
that growth rates were comparable to the New York Bass Study's average to fast
growth rates (Green, 1985), an Indication that some of the historical observa-
tions regarding the preponderance of juvenile fish and the paucity of adult game
fish may no longer be valid for bass populations.
As shown 1n Table B.l-3, the list of species of fish 1n the 1983 studSy
agrees quite well with the 11st for 1933. Although comparative percent
contributions of the dominant game fish, panfish, forage fish and demersal fish
with those species recorded In 1933 1s not possible, because quantitative
Information are lacking 1n the historical study, all the above dominant species
(with the exception of the spotfln shiner and black bullhead, see note 1, Table
B.l-3) were also recorded 1n 1933. In addition, 31 species were similarly
reported 1n both studies. An analysis of these two studies (Greeley and Bishop,
1933 and Makarewlcz, 1983), which have spanned nearly 50 years, reveals
considerable qualitative similarity of the fishery within the reach from Federal
Dam to Fort Edward.
The construction of Federal Dam and various locks as part of the Champlain
Canal Section of the New York State Barge Canal System blocked major upstreaa
spawning migrations for a number of anadromous species, Including the American
shad, alewlfe, blueback herring, sturgeons and striped bass. Some migrants, as
documented by Smith and Lake (1990), may be found periodically upstream of the
B.111
-------
Federal Dam and Lock. Population pulses nay enter and leave the lower region of
the Upper Hudson through the Interconnecting system of locks.
The majority of fish species listed 1n Table B.l-3 are freshwater resident
of the Hudson River. Some migratory species, such as the striped bass, blueback
herring, sea lamprey and American eel, still attempt to utilize sections of the
Upper Hudson as migratory routes. With the exception of black bullhead, johnny
darter, pearl dace and northern redbelly dace, all the fish species listed have
also been found In various regions of the Lower Hudson River (Beeoe and Savldge,
1988; Smith and Lake, 1990). Whereas many of the fish species are year-round
freshwater residents, they are not unique to the Federal Dam/Fort Edward section
of the Upper Hudson.
Some additional fish species, which were not found during the reported fish
surveys summarized 1n Table B.l-3, have been reported 1n various sections of the
entire Upper Hudson Region (Smith and Lake, 1990). These Include:
American shad (Alosa sapldlsslaa)
Stonecat (Noturus flavus)
Longnose sucker (Catostonus catostonus)
Lake chub (Coueslus plunbeus)
Brassy minnow (Hybognathus hankinsonl)
Blacknose shiner (Notropis heterolepis)
Finescale dace (Phoxlnus neogaeus)
Lake herring (Coregonus artedi)
Lake whlteflsh (Coregonus clupeaforals)
Round whlteflsh (Prosopiua cyHndraceun)
Lake trout (SalveTinus naaaycush)
Rainbow smelt (Osaerus atordax)
Tessellated darter (Etheostoaa olastedl)
Although not found during the major fish surveys conducted within the Federal
Dam/Fort Edward Region, the American shad has been known to occur within this
particular region (per. comm., R. Sloan). Others such as the lake chub, lake
herring, lake whlteflsh and lake trout are not commonly found In freshwater
riverine systems and are not expected to occur to any great extent within this
section of the Upper Hudson.
B.l-12
-------
All studies reviewed to date Indicate that the majority of species
historically present 1n the lower section of the Upper Hudson continue to reside
1n this particular reach of the Hudson River. According to information submitted
by Shupp (1987), the section of the Upper Hudson River between the Federal Dam
and Fort Edward can support a diverse and high quality fishery resource. Shupp
also cited evidence gathered from some NYSOEC studies between Mechanlcvllle and
Schuylervllle, which suggested a "vast Improvement" In smallmouth and largemouth
bass stocks and other fish species from the early 1960s to the late 1980s. Shupp
has stated, "Since 1984, the greatly Improved warm water fish community in the
Fort Edward to Troy (Upper Hudson) reach has stimulated Interest In reopening the
fishery."
Angler reports to The Uarrensburg F1sh Management Unit Indicate a somewhat
Improved fishery for -bass, yellow perch, black crapple and brown bullhead from
1969 to 1975 (Shupp, 1975). Analysis of the data presented by Lane (1970),
Makarewicz (1983) and Green (1985) and prepared testimony statements by Shupp
(1987) also suggest a qualitative Improvement within the past twenty years.
B. 1-13
-------
PAGE INTENTIONALLY LEFT BLANK
B.1-14
-------
SYNOPSIS
SOURCES OF PCB CONTAMINATION
(Section BJ)
General Electric discharged PCBs pom plants at Fort Edward and Hudson Falls between
1946 and 1977 (B.1.1). The total amount ofPCBs released during 1957 to 1975, a period for
which estimates can be made using historic data, ranges from 209,000 to 1,330,000 pounds.
<
Currently, six New York State facilities are permitted to discharge PCB-contaminated
waste water to the basin of the Upper Hudson River (B.Z2). A facility in western Massachusetts
discharges to a Hudson River tributary, the Hoosic River.
Other potential sources ofPCBs to the Upper Hudson (B.3.3) are discussed.
-------
PAGE INTENTIONALLY LEFT BLANK
-------
B.2 Sources of PCB Contamination
B.2.1 6E Discharges (To 1977)
The two GE plants at Fort Edward and Hudson Falls, New York began to use
PCBs 1n 1946 and discontinued their utilization In 1977. In-plant sources of PCB
discharges have been characterized as both minor spills and effluent from washing
capacitor cans, with the latter being the major source. Capacitor cans were
flood filled with dielectric fluid and then washed with detergent and water to
remove excess material. Contaminated wash water was finally discharged,
untreated, to the Hudson River (Brown, Jr. et ซ7., 1984).
Estimates of PCB releases at the two capacitor plants have been made on the
basis of GE's overall usage of the chemical and considering discharges allowed
pursuant to USEPA's discharge permit for the facilities. Figure B.2-1
Illustrates the company's PCB usage, by Aroclor type, for the period 1946 to
1977. That figure shows the trend In Aroclor usage to be from relatively highly
chlorinated forms 1n the mid-1950s (Aroclor 1254) to less chlorinated homologues
In the 1960s (Aroclor 1242) and the 1970s (Aroclor 1016).
By using actual GE purchase records for the years 1966 thru 1975 and
approximating GE purchases for the prior period on the basis of Monsanto's
production records, GE's total PCB consumption for the period 1957 to 1975 has
been estimated at 133,000,000 pounds (Umburg, 1985). Plant discharges during
the 1960s have been approximated at 5 metric tons per year (Sofaer, 1976), a rate
which Is roughly compatible with that allowed by GE's 1975 discharge permit (30
pounds per day or about 11,000 pounds per year). Sanders (1989) provides
anecdotal evidence of plant releases being less than one percent of plant
consumption or less than 1,330,000 pounds from 1957 to 1975. Thus, one estimate
of the range of releases to the river would be 209,000 to 1,330,000 pounds over
the period 1957 to 1975, where the lower quantity 1s based on a continuous
discharge of 30 pounds per day (or 5 metric tons/year) for a nineteen-year
period.
B.2-1
-------
B.2.2 Current Remitted Discharges
Six facilities 1n New York State, Including GE, are permitted to discharge
PCBs in the Upper Hudson Basin. NYSDEC (March 19, 1991) provided SPPDES permits
and discharge monitoring reports (PCBs only) for each of these facilities. Table
B.2-1 Identifies these facilities, receiving waters and relevant Information on
PCB limits and measurements. Two facilities (GE, Fort Edward and James River
Corporation, South Glens Falls H111) are permitted to discharge PCBs directly
into the Upper Hudson River, while one (GE, Old Fort Edward Site Remediation
Project) discharges Into the Old Champlaln Canal 1n the vicinity of Fort Edward.
In most cases, the concentration of PCBs 1n the final effluent Is limited to the
minimum reliable detection limit based on USEPA Method 608.
According to available Commonwealth of Massachusetts SPDES records, Sprague
Electric Company 1s permitted to discharge PCBs (0.01 mg/7) directly Into the
Hooslc River, which flows Into the Upper Hudson.
B.2.3 Other Sources
Table B.2-2 Identifies Inactive hazardous waste disposal sites located near
the Upper Hudson River (above Federal Dam at Troy) 1n which PCBs have been
dumped. This tabulation was obtained from NYSDEC, Division of Hazardous Waste
Remediation, utilizing their annual Inventories of disposal sites 1n New York
State (April 1990). The NYSDEC priority classification codes stated In the table
are: Code 2 - Significant threat to the public health or environment and action
Is required; Code 2a - Temporary classification assigned to sites with
Inadequate or Insufficient data for Inclusion 1n any of the other classes; Code
3 - Does not present a significant threat to the public health or environment,
and action may be deferred; Code 4 -Site 1s properly closed and requires
continued management.
The release of some contaminants from these inactive sites adds to the
total PCB loadings 1n the Upper Hudson. In many Instances, flow of surface water
and groundwater from the sites are towards the Hudson River. "Unknown" material
B.2-2
-------
was listed as being disposed of at many sites and 1s generally not Included In
the table. None of the sites Identified by NYSDEC are classified as an Imminent
danger to the public or environment (Code 1). Many sites are classified as Code
2, suggesting that these sites may be potential sources of pollution to the
Hudson River.
B.2-3
-------
PAGE INTENTIONALLY LEFT BLANK
B.2-4
-------
SYNOPSIS
NATURE AND EXTENT OF CONTAMINATION
(Section B.3)
Available environmental data on the distribution ofPCBs in the sediments, water, fish,
air and plants of the Upper Hudson River as well as supporting data on flow and sediment
transport are summarized and evaluated. As a foundation for continued analyses, available
data have been compiled into a computerized, relational database management system (B.3.1).
Data on PCB concentrations in river-bottom sediments (B.3.2) are drawn primarily from
the 1976-1978 NYSDEC sampling efforts and the 1984 Thompson Island Pool investigation,
along with several other sources. Sediments are the major environmental repository for PCBs
in the Upper Hudson, but there is a high degree of spatial variability in PCB concentrations.
The 1984 study covered only the Thompson Island Pool and relatively little data have been
collected since. It is difficult to determine the current mass and distribution of PCBs in
sediments without further investigation.
The discussion of surface water monitoring (B.3.3) concentrates on data collected by the
USGS. Transport of PCBs is affected by hydrologic processes, particularly flood events. A
discussion of flow monitoring is followed by presentation of time series data, to the extent
available, for suspended sediment and PCBs in the water column. Current full-year and
summer average PCB concentrations are calculated, taking into account the problem of
numerous measurements below analytical detection limits.
NYSDEC has monitored Upper Hudson fish on a regular basis since 1975; data an
presently available for PCBs in fish through 1988. The extensive data collected in this program
(nearly 3,000 Upper Hudson samples) are discussed (B.3.4). Total PCB burdens in fish declined
sharply from 1978-1981. Levels of the higher chlorinated congeners in fish appear to have
remained relatively constant since 1982. Results of NYSDOH macroinvertebrate monitoring an
also described.
PCB monitoring data for air and plants near the Upper Hudson (B.3.5) are generally
insufficient to assess the impact from PCBs in the river. Isolating the contribution of the river
from other possible PCB sources is a particularly difficult problem.
For various other media there is a notable lack of monitoring data (B.3.6). Only limited
groundwater sampling has been performed and surface soils near the river have not been
monitored.
Data quality and analysis methods for the various monitoring programs are evaluated.
PCBs have many different variations in chemical structure and differing physical properties.
Uncertainties surrounding PCB measurement, particularly the specific variations in PCBs, results
in considerable difficulties in interpreting the results. Furthermore, differing PCB measurement
methods used for water, sediments or fish confound direct comparisons among them.
-------
PAGE INTENTIONALLY LEFT BLANK
-------
B.3 Nature and Extent of Contamination
B.3.1 Overview of Sources and Database
Site data defining the current understanding of the nature and extent of
PCB contamination, based on previous studies, are summarized 1n this section.
Data synthesis efforts have focused on:
obtaining the most complete, and current, data sources available;
compiling these data Into a computerized database;
evaluating the PCB data for the media sampled;
Identifying current trends and relationships; and
determining the adequacy of the existing data.
During the early 1970s, NYSDEC and several other agencies began the first
comprehensive monitoring studies for PCBs 1n the Upper Hudson. F1sh, which were
some of the earliest environmental samples analyzed, showed high concentrations
of PCBs. These early Investigations began what 1s now over two decades of
studies on PCBs 1n water, sediments, fish and other media affected by PCB
discharges to the Upper Hudson. Table B.3-1 summarizes the major Investigations.
Past USEPA documents, Including the 1984 Feasibility Study and the EIS,
have been reviewed for this work. Emphasis, however, was given to reviewing
additional, more recent data and evaluating that along with the long-tera
monitoring record contained 1n the previous studies.
Previous Investigations at the Hudson River PCB site have examined the
nature and extent of contamination 1n several media, Including fish, sediments,
river water and, to a lesser extent, air. Each medium 1s discussed below.
Additional studies, currently being performed by GE at the remnant deposit sites,
are not available for this report. Before discussing each medium, a brief
overview of the TAMS/Grad1ent database 1s provided below.
B.3-1
-------
As the foundation for these Phase 1 data evaluation efforts as well as
continued analyses during subsequent phases of the project and possibly future
projects, the data gathered during Phase 1 have been compiled into a relational
database management system using PC-based Paradox m software. This databas^
currently contains approximately 30,000 records of information, primarily on
sediments, water, fish and some other biota obtained from numerous sources (see
Table B.3-1). Data input, verification and database management have been
conducted by the TAMS/Gradient project personnel.
In the current sediment database, there are nearly 2,500 samples for the
period 1976 through 1990. The fish database contains approximately 8,000 samples
for the period 1973 through 1988. Additional data for other aquatic biota
(macroinvertebrates and multiplate data) account for several hundred/additional
samples. Water column data, including daily and peak flow, suspended sediment
and total PCBs, comprise the bulk of the remaining data in the database. The
database contains a small number of samples summarizing PCB data for air and crop
plants.
Data for separate media are linked primarily through sample date and
location information. For each medium, data are organized such that a unique
sample identification number links information among tables. Individual data
tables are grouped by medium to contain similar kinds of information. As an
illustration of the database format developed specifically for this Reassessment
and its contents, excerpts from the sediment database tables are summarized below
in four tables: Sample Information, Core Section, Chemical and Non-Chemical.
Each of these tables is linked by a unique sample identification number, e.g.,
sample ID numbers 30000, 30016 and 30032 shown here, such that sample IDs in each
table correspond to data about a single sample.
The Sample table contains information about sample date, location (River
Mile distance from bank and northing and easting coordinates, where available),
sample type (grab versus core), the agency or investigator responsible for the
analytical method(s), the reference report/location of the original data and
other information as available.
B.3-2
-------
Database Table Example: Sample Information Table
Sampta
O
Typ*
M/D/YR
rซv*r
MUl
FMtfr.
WMt
Bank
Northing
w
Easting
fft)
Samptor
Wkter
Depth
fft)
Bev.
(ft)
Raf
Aeanoy
30000
Grab
5/21/77
iesa
330.0
1071755
685695
100
2
O'Brien
&Gara
30016
Cora
3/18/77
188.4
100.0
1163740
698870
100
54
119.6
1
O'Brien
aQara
30032
Cora
3/18/77
183.4
60.0
1140410
669040
100
Z2
102.4
1
O'Brien
& Oara
Core samples In the Sample table are linked with the Core Section table,
which Identifies the length of each core sample section and the depth beneath the
river bottom, i.e., the depth of sample penetration for the top and bottom of
each section.
Sediment Database Examplt: Cora Section Table
Sampla ID
Cora Section No.
Bottom of Section (in.)
Top of Section (in.)
30016
1
1
0
30016
2
2
1
e
.
30016
12
12
11
30032
1
1
0
30032
9
9
8
Selecting a sample ID from the Core Sample and Section tables and locating
the same 10 1n the Chemical data table shows either the Aroclor results for an
entire grab sample or section by section results for core samples. Additional
Information describing analytical measurement methods, i.e., extraction method,
are contained 1n the database as available. The Chemical data table also
contains non-PCB chemical data, such as metals analyses (not shown here), where
available.
B.3-3
-------
Sediment Database Example Chemical Data Table
Sampte 10 .
Parameter
Cora Saction No.
Extraction Mathod
Concentration (ppm)
30000
Arodor 1016
ahaka
1j0
30Q(X)
Arodor 1221
shako
1^
90000
Aroetor 1254
ahaka
1.0 *
30016
Arodor 1016
4
aoxhlat
6.0
ป
30016
Arodor 1254
12
aoxhlat
> 0.1
30032
Arodor 1016
5
aoxhlat
234.0
30032
Arodor 1254
5
aoxhlat
163.0
Finally, non-chemical data, such as sediment texture class, percent
volatile versus total solids, are contained in the Non-Chemical table.
Sediment Database Example: Non-Chemical Data Table
Sample ID
Cora Saction No.
Parameter
Value
30000
% total toilets
78.83
30000
% volatil* aolida
0.85
30016
1
taxtura
GRAVEL
30016
4
% total aolida
86.97
30016
4
% volatile aolida
Z2
30016
12
% total aolida
88.23
30032
1
I
3
CL-wc
30032
5
* volatile aolida
2S.30
B.3-4
-------
B.3.2 Sediment
Two primary sources of Information provide the largest amount of data on
sediment contamination 1n the Upper Hudson: (1) printed results (>1,000 samples)
of the 1976-1978 NYSDEC sampling survey and (2) computer files (>2,000 samples)
of the 1984 Thompson Island Pool survey. These data were entered Into a
computerized database (referenced here as the TAMS/Grad1ent database). Tables
B.3-2 and B.3-3 provide a summary of these sediment samples 1n the TAMS/Gradlent
database. In addition to these two major NYSDEC sediment surveys, USEPA, the
Lamont-Doherty Geological Observatory and GE have sampled sediments 1n the Upper
Hudson.
B.3.2.1 . 1976-1978 NYS0EC Sampling
As reported by Tofflemlre and Qulnn (1979), NYSDEC conducted several
sediment sampling surveys 1n the Hudson River between 1976 and 1978. Details
about the sampling and analysis procedures for these studies are summarized In
NYSDEC Technical Report No. 56 (Tofflemlre and Qulnn, 1979).
The data provided by NYSDEC contained a total of 1,167 sediment samples
(396 cores and 771 grabs) taken during 1976, 1977 and 1978; 1,770 PCB analyses
were reported for the 1,167 samples. The overwhelming majority of samples (1,091
of the 1,167 samples) from the 1976-78 data set were collected In the Upper
Hudson River. Only five samples In this data set were obtained 1n the Lower
Hudson River and all of these five were from River M11e 153, just south of the
Troy Lock. Another sample In this set was Identified as from the Lower Hudson,
but other descriptions placed 1t 1n the Upper Hudson, while 70 samples had no
Information regarding location.
Aroclors 1016, 1221 and 1254 were Identified as the PCB mixtures detected
1n the 1976-1978 sediment sampling effort. Total PCB concentrations were
reported as the sum of these three aroclors. Analytical detection limits were
not reported 1n this data nor was any Indication given about a sample's
detectable or non-detectable concentrations of PCBs. Because several concentra-
B.3-5
-------
tions (1 ppm, 5 ppm, 10 ppm) occur an Inordinate
set, these concentrations are the probable detecti
Table B.3-4 summarizes the Aroclor concen
samples, respectively, within each of the nine
summarizing the mean and median PCB values for the
counted as a single sample, I.e., the statistics a
core length. The reported N 1s the count of core s
of the most striking aspects of the 1976-1978 resu
of PCB concentration over very short distances. !
apart may have PCB levels varying by orders o
variability is highlighted 1n Figure B.3-1, which
in surface sediments (grab samples and the top se
mile. Some trend with river miles is shown by
concentration by reach is highest in the Thompson I
194), decreasing for several miles downstream, t
River Mile - 175 to River Mile 160. This pattern
of depositlonal areas between Stillwater and Nortl
Based on the results of the 1976-1978 surve
spots, areas containing more than 50 ppm total PCI
Pool survey re-evaluated the hot-spot locations ai
a series of over 100 polygons containing PCBs gre;
spot" term 1s used here only as a frame of refere.
sediment areas previously defined as containing high '
to the 1976-1978 study, questions about the adequa<
(NUS, 1984), as noted below:
PCB concentrations exhibited wide spat
single transects not uncommonly ran
concentrations to greater than 1,000 ;
The sampling density was so low tha'
variability of the PCB concentrations,
delineation 1s questionable.
B.3-6
-------
Changes 1n sediment deposits caused by the dynamics of the river
greatly complicates comparisons between PCB concentrations In
similar locations in different years.
t A major flood event that occurred In 1979 redistributed sediments
significantly, again calling into the question the usefulness of the
1976-1978 sediment data, other than general purposes.
The analytical techniques used to quantify PCBs have Improved since
this data was collected.
B.3.2.2 1984 NYSDEC Sampling
Methods and Procedures
In 1984, NYSDEC again undertook an extensive sediment sampling program.
This effort focused on the Thompson Island Pool (M. P. Brown et a7., 1988b). The
objective of this study was to identify areas of contaminated sediments that
would be removed during the Hudson River PCB Reclamation Demonstration Project.
Primarily these areas were the 20 hot-spots previously Identified in the Thompson
Island Pool and other areas with known or suspected high PCB concentrations.
The investigators identified 1,260 sampling locations 1n the approximately
five-mile reach of the river. Many of these locations were determined by
Imposing a 125-foot triangular grid on previously defined hot-spots and areas
that had PCB concentrations in excess of 50 ppm during the 1983 USEPA survey
(NUS, 1984). In addition, sample locations were selected based on known or
suspected sediment depositlonal areas, as Indicated by location 1n the river and
bathymetry measurements. Sample locations in the field were determined
electronically by using a microwave locating system and generally agreed with
predetermined locations.
Samples were collected by Normandeau Associates, Inc. between August 24f
1984 and November 30, 1984. In addition, 21 cores were collected during February
1-4, 1985. These later samples were taken through the ice on the river at
locations that had been Inaccessible by boat.
B.3-7
-------
Whenever possible, the investigators collected core samples. At those
locations where insufficient sediment was available for core samples, grab
samples were attempted. At 80 locations, bedrock or coarse material precluded
either sample method. In all, 1,016 locations throughout the Thompson Island
Pool were sampled and yielded 674 grab and 408 core samples.
The depth of penetration for the core samples was fairly uniform, with an
average depth of approximately 31 Inches. The investigators divided core samples
into sections based on a desire to define contaminated layers without introducing
dilution from adjacent less contaminated layers. Sections were also chosen based
on potential dredging considerations and a need to limit the number of chemical
analyses.
NYSDOH and Versar, Inc. measured physical and chemical parameters of the
sediments collected In this study. NYSDOH determined lengths of cores and
sections, percent dry solids, dry specific weight (density) and textures, which
were determined visually. Versar measured percent volatile solids and performed
the gas chromatograph analyses for PCBs.
PCB Results
Versar reported PCBs as Aroclors 1242, 1254 and 1260 using the method of
Webb and HcCall (1973). Although the data contained a total PCB quantification,
no mention 1s made 1n M. P. Brown et al. (1988b) of the method used to quantify
or calculate this total. Examination of the data received indicates that the
total was not simply the sum of the three Aroclor mixtures quantitated.
Wide variations 1n PCB concentrations in sediments were observed in the
1984 NYSDEC study throughout the Thompson Island Pool, even though sampling
concentrated on areas of known contamination. The discrepancies between means
and medians for both grabs and cores indicate that PCB concentrations have a
highly skewed distribution over the area of the Thompson Island Pool.
(Environmental monitoring results frequently exhibit this skewed pattern and a
log-normal distribution is often a good approximation of the data.) Both grab
B.3-8
-------
and core samples had significantly higher concentrations in the least chlorinated
fraction that was quantified (Aroclor 1242) than in the more chlorinated
fractions. Table B.3-5 provides Aroclor and total PCB summary statistics for
both grab and core samples. Aroclor 1242 was the predominant Aroclor reported
for these samples, with lower levels of Aroclors 1254 and 1260 also identified.
On average, total PCBs in the samples were approximately 55 ppm, with naximun
levels of >1,000 ppm detected in several samples.
Table B.3-6 presents the results of texture classifications determined by
NYSDOH. Considering grab and core samples together, Thompson Island Pool
sediments were classified most often as either gravel or fine sand, with a
significant fraction of fine sand/wood chips and clay samples, particularly for
core samples.
Because of their high adsorption (partition) coefficients, PCBs are
generally expected to associate with the organic carbon fraction of the
sediments. Although no measurements of organic carbon content were made as part
of this study, organic carbon and organic matter content, which were measured
frequently 1n the study as percent volatile solids, can be correlated. Thus,
comparing PCB levels with organic matter content (volatile solids) provides a
surrogate for comparing PCB concentration to sediment organic carbon content.
Figures B.3-2 through B.3-4 show the relationships between PCB concentra-
tions and percent volatile solids within a texture classification for the three
most commonly occurring textures (gravel, fine sand, fine sand/wood chips). Very
little relationship appears to exist between total PCBs and volatile solids
measured 1n the gravel texture class (Figure B.3-2). The fine sand (Figure B.3-
3) and fine sand/wood chip (Figure B.3-4) categories exhibit a better correlation
between PCB concentration and percent volatile sol Ids, although percent volatile
solids would still make a poor predictive measure of PCB concentration.
B.3-9
-------
Comparison of 1976-78 and 1984 Studies
As mentioned previously, the 1984 study focused on the 20 hot-spot areas
1n the Thompson Island Pool as defined by areas exceeding 50 ppm PCBs 1n th|
1976-1978 survey. M. P. Brown et a7. (1988b) describe in detail the differences
found between the two surveys. The PCB concentrations in sediments of the
Thompson Island Pool exhibited lower concentrations 1n the 1984 survey than In
the earlier study, as shown by the somewhat higher frequency of PCBs detected at
<25 ppm (Figure B.3-5). A direct comparison of the relative frecwency plot for
the two studies 1s hindered by the fact that the 1984 survey specifically
targeted potentially contaminated areas and areas with fine-grained sediments,
which were thought to contain more PCBs. Thus, 1984 samples were potentially
more heavily biased in a statistical sense to those areas of PCB contamination.
If samples had been taken randomly from the Thompson Island Pool, the results
would likely have yielded an even larger relative number of samples with lower
concentrations.
Because of the different scope and sampling density of the two surveys and
their different sampling and analytical methods, M. P. Brown et al. (1988b)
Indicate that direct quantitative comparisons between samples collected in
similar areas are problematic. M. P. Brown et al. found that areas of high PCB
concentration determined in the 1976-1978 survey appeared to be generally
confirmed by the 1984 survey. Based on area-weighted average PCB concentrations
1n 138 polygon areas, these investigators calculated a total PCB mass of 23,200
kg (51,040 lb) 1n the top 1.5 m (59 inches) of Thompson Island Pool sediments
(M.P. Brown et *7., 1988b). This mass estimate compares to Malcolm Pirnie's
estimate in 1978 of approximately 61,000 kg (133,670 lb) for all of the Pool and
approximately 48,000 kg (104,870 lb) for the 20 hot spots (Malcolm Pirnie, 1978).
M. P. Brown et al. (1988b) offer as possible explanation for the varied estimates
the differences 1n analytical PCB methods, depth Integration/area averaging and
sediment densities used.
B.3-10
-------
The importance of random variability and sampling density for either the
1976-1978 or 1984 studies also complicates comparisons between the two surveys
and affects mass calculation comparisons. For example, GE Indicates that they
have taken 30 samples from polygon 5 of the Thompson Island Pool (General
Electric, John Claussen letter to USEPA, March 29, 1991). This polygon was
estimated by M. P. Brown et al. (1988b) to have contained approximately seven
percent of the PCB mass 1n the Thompson Island Pool based on two samples
containing 39.7 and 6,587.8 ppm PCBs, yielding an average concentration of 2,437
ppm. GE has indicated that the average, based on their 30 samples in this
polygon, 1s less than 20 ppm. (Results are not yet available to the project
study team.) Although PCB mass differences 1n this one small area of the
Thompson Island Pool do not necessarily mean that the overall conclusions of the
1984 survey are incorrect, they do suggest that:
wide variations in PCB concentrations occur over relatively short
distances;
direct quantitative comparisons of PCB levels 1n samples froป
different years are problematic; and
the mass and distribution of PCBs in the Upper Hudson are difficult
to quantify.
Lamont-Doherty Geological Observatory Investigations
The Lamont-Doherty Geological Observatory, under contract to the NYSDEC,
conducted a survey of PCB levels 1n the sediments, suspended matter and water
column of the Upper Hudson River during 1983 and 1984 (Bopp et al., 1985). The
survey, a coring effort, collected 16 cores, covering the Upper Hudson from above
Hudson Falls to the Albany area (Plate B.3-1), and analyzed them for radio-
nuclides. Procedures were the same as those described in Section A.3 for the
Lower Hudson. Many sections of these cores were analyzed for PCB levels, with
an emphasis on homologue and congener-specific Information. The Investigation
also Involved PCB analyses of surface water samples (see Section B.3.3).
B.3.2.3
B.3-11
-------
On the basis of these data, Bopp et al. (1985) were able to draw a number
of important conclusions concerning the fate of PCBs 1n the Upper Hudson. In
cores showing Interpretable radionuclide chronologies, the occurrence of a
maximum PCB concentration in sediments deposited circa 1973 (the removal of t^e
Fort Edward Dam) could be seen in all areas of the Hudson, including the Thompson
Island Pool (Figure B.3-6). This data demonstrated that the sediments of the
Upper Hudson could be used to determine PCB transport history.
Analysis of Upper Hudson sediments revealed that very recent sediment
deposits (1980-1983) contained PCB congeners 1n ratios very similar to Aroclor
1242 and 1016, whereas older sediments, typically under anaerobic conditions,
showed substantially different congener ratios. In general, these older
sediments contained higher levels of mono through tetrachloroblphenyls and lower
levels of the higher chlorinated congeners, relative to a standard Aroclor 1242
mixture. Bopp et al. (1985) concluded on the basis of these results, that under
anaerobic conditions, biologically driven dechlorination must be occurring.
However, little or no dechlorination was occurring under the aerobic conditions
of the uppermost sediment layers.
B.3.2.4 Other Studies
1983 USEPA Study
In August 1983, USEPA conducted a limited study to collect samples from
locations that had been sampled 1n 1976-1978 (NUS, 1984). Sixty-six samples, of
which 54 were core and 12 were grab samples, were collected within a nine-mile
stretch of the river south of Rogers Island, including the Thompson Island Pool.
Forty-two samples were collected from within or on the border of previously
determined hot-spots. The results of this study tended to show that areas of
high PCB concentrations in 1976-1978 exhibited high concentrations in 1983 as
well. Nevertheless, direct comparisons between samples taken within 50 - 100
feet of each other during the two surveys indicated that variations by two orders
of magnitude were not uncommon (NUS, 1984). In general, the concentrations
observed 1n 1976-1978 were greater than those seen in 1983 at corresponding locations.
B.3-12
-------
6E 1989 Baseline Studies (Remnant Remediation Project)
As part of the Remnant Remediation Project, General Electric conducted
baseline pre-remedlation sediment monitoring. Sediment samples were collected
at five locations in the vicinity of the remnants: one location near Rogers
Island; one location far upstream; one location between the remnants and Bakers
Falls; and two downstream locations near Lock 6 and Uaterford. Median PCB
concentrations in river sediments 1n the remnant areas ranged from 0.47 ppm to
34 ppm, with a maximum of 99 ppm (Table B.3-7). Downstream samples contained
median PCB concentrations of 0.64 to 2.1 ppm, whereas the control location had
median PCB levels (seven samples) of 0.11 ppm. The sample location between
Bakers Falls and the remnants had a median PCB concentration of 1.4 ppm. With
the exception of the two downstream locations, PCBs were detected 1n all samples.
The chromatograms were compared against Aroclor mixtures 1221, 1232, 1016, 1242,
1248, 1254 and 1260; Aroclor mixtures in the samples were reported to be Aroclors
1242 and/or 1254.
6E 1990 Sediment Sampling (B1oremed1at1on Investigations)
General Electric has been conducting extensive research on biological
dechlorination and/or degradation processes occurring within the river, which may
have altered the composition of the PCB Aroclor patterns within the sediments.
In conjunction with these studies, GE has recently collected samples from
selected areas of the Upper Hudson for more detailed evaluation. General
Electric provided preliminary results of their sediment sampling activities
(Claussen, 1991b).
Harza Engineering, GE's contractor, collected 103 core samples from 12 hot-
spots during 1990 and reported 275 PCB analyses. From three to eight cores were
collected at most locations, with the exception of GE's H-7 location where 62
cores on a 12 x 12 foot grid were sampled. Samples were analyzed for PCB
homologue groups and five Aroclors (1221, 1242, 1254, 1260 and 1268). The
results of this sampling are summarized in Table B.3-8. With the exception of
H-7 and Location 4, the median PCB concentration at the four other locations
B.3-13
-------
within the Thompson Island Pool exceeded 100 ppm. Median PCB concentrations for
the aforementioned two locations and the locations downstream of the Thompson
Island Pool were less than 100 ppm.
\
Coincident samples and PCB measurements from the same hot-spots in the
NYSDEC 1976-78, USEPA 1983, and GE 1990 samples are available for only six
locations. The average PCB concentrations for each of these three surveys are
summarized in the lower portion of Table B.3-8. In four out of ,six locations,
the GE samples Indicate average PCB levels above both the 1970 1978 and 1983
values; the remaining two locations show 1990 PCB levels lower than the 1976-1978
and/or 1983 results. Because of the very small sample sizes, few coincident
locations and difficulty 1n determining whether these samples represent similar
sediment zones over time, these results are inadequate to suggest a clear trend.
Qualitatively, the results document the continued presence of PCBs in areas
originally defined in 1976-1978 to be contaminated.
B.3.2.5 Other Chemicals in Sediments
In addition to PCBs in river sediments, other chemicals, particularly heavy
metals, have been measured during 1976-1978 (Tofflemire and Quinn, 1979), 1984
Thompson Island Pool study (H. P. Brown et a?., 1988b) and by other investiga-
tors. Lead, cadmium, zinc, chromium, mercury and other metals have been
measured. M.P. Brown et al. (1988b) indicate that anthropogenic sources,
Including a pigment manufacturer 1n Glens Falls, elevate lead, cadmium and
chromium in river sediments above their naturally occurring levels. Based on
their 1984 study, H.P. Brown et al. (1988b) reported mean metals concentrations
in sediments for lead (217 |ig/g), cadmium (21.6 ปปg/g), chromium (475 jig/g) and
mercury (1.96 |ig/g) along with several other chemicals. M. P. Brown et al. found
that although lead and cadmium are the two metals most frequently found in
sediments, standard leaching tests, e.g., EP Toxicity, suggest they are not
readily leachable.
B.3-14
-------
Relatively few sediment samples have been tested for other organic priority
pollutants. Four sections of two cores collected 1n 1983 by Dr. Richard Bopp
from River Miles 188.5 and 191.1 (Thompson Island Pool) were submitted to NYSDOH
and analyzed for dloxln and dibenzofurans. Six sediment samples collected 1n
1987 from three hot-spots were analyzed for dloxlns, dlbenzofurans, volatile and
semi-volatile organlcs and pesticides (M. P. Brown et al., 1988). With the
exception of dlbenzofurans, none of the other organic parameters were detected
In the 1987 samples.
As reported by M. P. Brown et al. (1988b) tetrachlorod1benzo-p-diox1n
(TCDD) and tetrachlorodlbenzofuran (TCDF) as well as their 2,3,7,8- isomers were
detected at less than part per billion levels in the 1983 samples. Total TCDD
in these 1983 samples ranged from non-detected to 0.135 ppb; total TCDFs ranged
from non-detected to 0.731 ppb. In two of the six 1987 samples, total TCDFs were
detected at <0.2 ppb levels; TCDDs were not detected in the 1987 samples;
detection limits ranged from 0.012 - 0.058 ppb. M. P. Brown et al. indicate that
possible sources of the TCDFs in sediment include residual fall-out from coal and
wood combustion, discharge from wood processing plants (by-product of chlorophe-
nol pyrolysis) and discharge of chemical-wastes containing TCDDs and TCDFs as
trace contaminants. Industrial PCB mixtures are known to contain trace levels
of TCDFs.
B.3.2.6 Discussion
The study team encountered some difficulty 1n matching the contents of the
data entries in the TAMS/Gradlent database, especially the 1976-1978 sediment
data, with data summaries provided 1n previous reports (Tables B.3-2 and B.3-3).
The 1976-1978 raw data in printed form did not contain identification designa-
tions cross-referenced to date and sample location and no report containing such
a cross-referenced summary was found. Sample identification numbers are shown
on a marked-up copy of the 1977 Normandeau Associates, Inc. map in the NYSDEC
offices, but do not cross-reference dates of the samples or laboratory
Identification numbers. The team could assess the completeness of the data set
only by comparing sample dates and river location associated with the samples
B.3-15
-------
with summaries In published reports. As shown 1n Table B.3-2, these comparisons
Identify some Inconsistencies between the data 1n the TAMS/Grad1ent database and
the number of samples reported by Tofflemlre and Qulnn (1979) 1n NYSDEC Technical
Paper No. 56. For example, the TAMS/Grad1ent database contains 254 grab samples
and 21 core samples reportedly collected 1n 1976, whereas Tofflemlre and Qulnn
report 24 cores and 80 unspecified sediment samples for 1976. For 1977, the
TAMS/Gradient database contains 446 grab and 246 core samples compared to 692
grab and 208 core samples reported by Tofflemlre and Qulnn (1,979). For all
samples collected between 1976 and 1978, the database contain^ 1,092 samples
compared to 1,404 Indicated by Tofflemlre and Qulnn.
Differences In the overall number of samples may be accounted for
approximately by noting that:
approximately 202 of the 672 summer 1977 grab samples taken by
Normandeau Associates, Inc. (NAI) were not analyzed for PCBs and
were not provided 1n the printed data summaries supplied for the
TAMS/Grad1ent database;
200 spring 1978 remnant samples from Malcolm P1rn1e, Inc. (MPI) were
not provided for use 1n the TAMS/Grad1ent database.
If these 402 samples are subtracted from the Tofflemlre and Qulnn total of
1,404, there would be approximately 1,000 samples, more closely approximating the
1,092 samples 1n the TAMS/Gradient database. A comparison of the number of
samples by river mile (Table B.3-3) also Indicates differences In total numbers
of samples reported by Tofflemlre and Qulnn and the data In the database.
Because 15 years have passed since the 1976-1978 samples, their use for
identifying precise locations and concentrations of contaminated sediments 1s
limited.
ป
The data summaries contained inconsistencies within samples and between
samples. For example, the sums of Individual core lengths did not always match
the total core length; for approximately 3 percent of the samples, the data
fields were Incorrect; approximately 10 percent of the samples had no Information
B.3-16
-------
about northing-easting coordinates. The team could not evaluate independently
the reason for the discrepancies nor the accuracy of the original data archival
process/data summaries. These discrepancies in the sediment database are not
considered to be significant at this time and an effort to resolve them will
continue during the course of the project.
Differences between the TAMS/Gradient database and the results reported by
M. P. Brown et a/. (1988b) for the 1984 Thompson Island Pool survey appear to be
slight. The database contains 1,141 samples, whereas M. P. Brown et al. report
1,205 for this survey, a difference of *>5 percent 1n total samples.
B.3.3 Surface Nater Monitoring
Numerous surface water monitoring stations along the Upper Hudson are
maintained by the US6S. These stations have monitored flow, suspended sediment,
PCBs and other water quality parameters. The USGS data, obtained from WATSTORE
and the Albany USGS office, provide the longest and most comprehensive record of
surface water data for the Upper Hudson.
B.3.3.1 USGS Flow Records
The USGS has collected river discharge (flow) and water quality data at
various points along the Upper Hudson River (Plate B.l-1). The USGS records of
the monitoring stations located on the Hudson between Hadley, well above Fort
Edward, and Green Island, below the confluence with the Mohawk River at Troy,
were obtained for use 1n this investigation.
The majority of the USGS flow monitoring stations on the Upper Hudson have
periods of record beginning in the 1970s, although continuous monitoring is
available at Hadley since 1921. A lack of widespread flow measurements for
earlier periods presents difficulties 1n analyzing the longer-term flow regime
and flood probabilities. In particular, the flow record at Fort Edward, at the
upper end of the Thompson Island Pool, commences only in 1976. No USGS
monitoring 1s available at the Thompson Island Dam. Barge Canal stage data are
6.3-17
-------
available at the guard gate at Crockers Reef (Gauge #118) in the Hudson
River/Champlaln Canal approximately parallel to the Thompson Island Oam, at the
lower end of the pool. At the northern end of the Thompson Island Pool, Barge
Canal stage data are available below Lock 7 (Gauge #119). These gauges report
water elevations 1n reference to the Barge Canal datum. They have not been
calibrated to river discharges and provide only qualitative data regarding flood
discharges.
In order to extend the record of flow data for the Upper Hudson, it is
necessary to move upstream to the confluence of the Hudson and Sacandaga Rivers,
near Hadley (see Plate B.l-1). A monitoring station has been maintained on the
Hudson at Hadley since July 1921. The Sacandaga River, a major tributary
entering just below Hadley, has been monitored since 1907 at Stewarts Bridge near
its confluence. By adding these two stations, USGS provides an estimate of the
flow in the Hudson below the confluence with the Sacandaga. Between this point
and Fort Edward there are several dams, but there are few additional tributaries.
The drainage area above Fort Edward is 2,817 square miles, while that of the
Hudson River at Hadley plus Sacandaga River at Stewarts Bridge 1s 2,719 square
miles, representing only a 3.6 percent increase in contributing area. Estimates
of flow 1n the Hudson below the Sacandaga, thus, provide an accurate estimate of
the magnitude of flow at Fort Edward.
The daily average flow value records for USGS Upper Hudson stations are
summarized 1n Table B.3-9, from upstream to downstream. USGS flow monitoring is
available since 1976 only at the upstream end of the Thompson Island Pool (Fort
Edward at Rogers Island). The closest functioning USGS monitoring station
downstream 1s at Stillwater, which is 26 miles and three dams south. Thus, Fort
Edward monitoring is most informative of hydraulic conditions in the Thompson
Island Pool. Mean dally flow at Fort Edward is 5,244 cfs; daily flows range from
652 cfs to 34,100 cfs for the period of record. Additional inputs from
tributaries and runoff increase the average dally flow to 7,933 cfs at Waterford
and 13,642 cfs at Green Island, below the confluence of the Mohawk River.
B.3-18
-------
Figures B.3-7a and B.3-7b display the daily average flows for 1973-1990.
As Fort Edward monitoring commenced in water year 1977, the 1973-1976 flows are
estimated from the calculated flows below Hadley, representing 97 percent of the
contributing watershed area at Rogers Island. This record reveals the presence
of several major flood events, which were associated with mass erosion of the
remnant deposits. These events occurred on: April 2, 1976, when a daily average
flow of 39,340 cfs was reported below Sacandaga River; April 29, 1979, when a
flow of 31,700 cfs was reported at Rogers Island; and May 2, 1983, when 32,600
cfs was reported at Fort Edward. During these flood events flows were even
higher for shorter time periods (peak flows). The maximum peak flows monitored
at Fort Edward since December 1976 are 34,000 cfs on April 29, 1979 and 35ป200
cfs on Nay 3, 1983.
The daily flows show evidence of a strong weekly periodicity. The seven-
day cycle is the result of regulation of the Sacandaga Reservoir to supply power
plants during the week, while maintaining the weekend recreational pool in
Sacandaga Lake.
B.3.3.2 Suspended Sediments Monitoring
Information on time trends in suspended sediment data as well as the
relationship between sediment and discharge is provided by USGS monitoring
stations (Plate B.l-1). Several water quality stations were established on the
Upper Hudson in 1969, but measurements of suspended sediment did not commence
until 1975. Monitoring is not continuous or on a set schedule and there has been
a tendency to focus on spring flood periods, with little data available for the
winter months. Lack of a more extensive database and of regular time series
creates difficulties 1n analyzing sediment data as well as other water quality
parameters.
Summary statistics on the USGS suspended sediment monitoring for stations
between Fort Edward and Waterford are given in Table B.3-10. The median
suspended sediment concentration 1n the Upper Hudson above the confluence of the
Mohawk ranges from 4 to 12 mg/7 and Increases downstream. Relationships between
B.3-19
-------
suspended sediment levels and river flow are discussed in B.4.
B.3.3.3 USGS PCB Monitoring
V
Methods and Procedures
Regular monitoring of PCBs in the water column in the Upper Hudson was
instituted by the USGS in late 1975 at Uaterford and expanded to other upstream
stations 1n 1977. Most other sampling programs, discussed later iiii this section,
have been of short term duration. A recent search of the STORET database reveals
that limited water-column PCB measurements are also available for some of the
tributary rivers to the Upper Hudson. These data have not been reviewed. The
USGS data are, thus, the primary source of time series information indicative of
trends 1n water-column PCB concentrations.
USGS observations of PCB concentrations in the water column have been made
at most of the same water quality stations as for sediment data (Plate B.l-1).
Data sets of significant size are available at Fort Edward (River Mile 194.5),
Schuylerville (River Mile 181), Stillwater (River Mile 168) and Waterford (River
Mile 156.5), with a limited record at Fort Miller (River Mile 187). In addition
background samples are taken upstream at Glens Falls (River Mile 200).
The original purpose of the USGS monitoring was to gather several years of
data on PCB concentrations prior to removal of contaminated sediments (Schroeder
and Barnes, 1983). Although this dredging plan has been delayed, monitoring has
continued. Data are now available on UATSTORE, a USGS-maintained computerized
database, through the end of water year 1989 (September 1989). Data for water
year 1990 were collected by the USGS and have been turned over to NYSDEC for
analysis, but the results are not yet available.
Methods of data collection and analysis are summarized in Turk and Troutman
(1981) and Schroeder and Barnes (1983). According to the latter source, samples
from the Upper Hudson were collected from bridges using depth-integrating
samplers. The sampler held a wide-mouth glass bottle, which was lowered and
B.3-20
-------
raised through the water column to obtain a depth-integrated sample (about 1
liter) for PCB analysis.
The USGS National Water Quality Laboratory in Doraville, GA performed the
PCB analyses. Comparison was made to standard Aroclor mixtures. Results,
however, were reported as total PCBs. Schroeder and Barnes (1983) reported that
PCBs in the Hudson were "almost always 1n the composition range from [Aroclor]
1232 to 1248," but recognized that natural processes had likely altered the
congener composition of the original Aroclor mixture.
Although the USGS laboratory reports a theoretical detection limit of 0.01
pg/7 through water year 1983, the practical quantitation limit was considered to
be 0.1 pg/7, because of the small size of the water sample (Bopp et a7., 1985).
Data for this period recorded on UATSTORE contain both values entered as 0 and
values coded as <0.10 |ig/7. Apparently these are both intended to represent non-
detects at the 0.1 detection level and the inconsistency is unintentional
(Rogers, pers. comm., 1991). With water year 1984, the practical detection limit
was lowered to 0.01 pg/7. Nevertheless, the 1984 and 1985 data are reported on
WATSTORE as if they adhere to the previous detection limit of 0.1 |ig/7. In 1986,
the detection limit began to be reported as 0.01 yg/7 in WATSTORE.
PCB Results
The USGS monitoring station at Glens Falls provided upstream background
levels of PCBs in the Hudson for 1977-1983. Of 45 observations for total PCBs,
only two had detectable levels of PCBs. These observations occurred on Deceaber
5, 1978 and September 28, 1980 and were both reported as 0.1 |ig/7.
Summary statistics for the USGS monitoring of total PCBs in the water
column between Fort Edward and Waterford are given in Table B.3-11. A remarkable
fact evident from this table Is that there 1s rather little variation 1n measured
PCB concentration by river mile between Schuylerville and Waterford. At all
stations, PCBs were detected in more than 60 percent of the water samples, with
detection frequencies ranging from 63 percent to 89 percent. Average PCB
B.3-21
-------
concentrations below Fort Miller range from a high of 0.29 jig/7 at Stillwater to
0.23 itg/7 at Waterford. Averages at Fort Miller are not directly comparable to
other stations because of the short period of record. The average at Fort Edward,
Inflated by one very high measurement (77 ng/7), was calculated in Table B.3-^
by omitting this outlier. This procedure yields a long-term average approximate-
ly half of that observed downstream. Concentrations observed at Schuylerville,
Stillwater and Waterford are approximately constant, although the average at
Waterford is lower than the up-river values, primarily because of dilution from
the Hoosic River. Indeed, the average concentration at Stillwa-f^r (0.29 }ig/I)
is slightly higher than that at the Schuylerville station upstream (0.26 |ig/7).
This finding suggests that there may be relatively little loss of water-column
PCBs during transit in the Upper Hudson.
At the Green Island station, downstream of the confluence of the Mohawk
River, the contributing watershed area is nearly double that at Waterford and PCB
concentrations are correspondingly diluted. Twelve samples were analyzed for
total PCBs between 1978 and 1985 and all were non-detects at the 0.1 yg/7 level.
No PCB measurements have been reported at Green Island since March 1985.'
Figures B.3-8 through B.3-11 show the time series of PCB observations at
Fort Edward, Schuylerville, Stillwater and Waterford, respectively. Observations
reported a& non-detect (or zero) are plotted at the detection limit (0.1 |ig/7
through September 1986 and 0.01 yg/7 thereafter). While much of the variability
observed near the detection limit may represent analytical noise, there is a
clear similarity apparent between the PCB time-series plots at Schuylerville,
Stillwater and Waterford (B.3-9 through B.3-11), particularly in the marked
response to the 1979 spring flood. Despite the fact that samples were taken at
somewhat erratic Intervals, field personnel seem to have frequently visited each
Upper Hudson station in succession, so that many samples for the whole reach,
while not contemporaneous, are close together in time. All three stations are
below the Thompson Island Pool. Although the response to floods appears somewhat
different at Fort Edward, located below the remnant deposits but above the so-
called hot-spots in the Thompson Island Pool, the 1979 event stands out.
B.3-22
-------
Also notable in these time series 1s a general decline In water-column PCB
levels from about 1979-1986. The question of whether there has been any genuine
trend 1n PCB loading to water over time, or whether the apparent year to year
trends are actually due to variability 1n the hydrologic regime, 1s discussed 1n
Section B.4.
Current Full-Year Average PCB Concentrations 1n Water Column
Estimates of current average concentration of total PCBs 1n the water
column are needed for the assessment of potential baseline health risks and other
environmental impacts. As noted previously, seasonal variability and lack of
continuous sampling confound the estimation of this average.
It 1s common practice when working with data 1n time series containing non-
detects to develop an average based on treating the non-detects as 1f they were
equal to one half of the detection limit. A more sophisticated way to approach
this question 1s to use the Adjusted Log Normal Maximum Likelihood method of Cohn
(1988), which overcomes the bias in the sample collection and variance in the
data. Another method for including non-detects in the estimate of the mean,
which however does not address the problem of sampling bias, is to use a log-
probit analysis (Helfeel and Cohn, 1988). Under the assumption that the logs of
the data are normally distributed, they will fall on a straight line when plotted
on a probit (probability) scale. The log-prob1t analysis essentially uses
regression to extend this line past the detection limit to predict the values of
the non-detect samples predicted by the observable part of the distribution.
In order to examine the sensitivity of the mean to the non-detects, the
above three methods of calculation were used here: (1) simple mean with non-
detects at 1/2 the detection limit; (2) the Adjusted Maximum Likelihood method;
and (3) log-probit analysis. Results of all three methods and the upper 9S
percent confidence limit on the estimates of the means, which 1s used in the risk
analysis, are shown 1n Table B.3-12.
B.3-23
-------
Little difference in estimates of the mean and the 95 percent upper
confidence limit on the mean 1s produced by any of the three methods. Mean PCB
levels 1n recent 1986-89 water samples are on the order of 0.05 p.g/7 at Fort
Edward and drop to approximately 0.03 jปg/7 at Waterford. Upper 95 perc^t
confidence limits on these means are not much higher, approximately 0.075 pg/7
at Fort Edward and 0.035 jig/7 at Waterford.
Summer Average PCB Concentrations
Average water-column PCB concentrations can be calculated-from monitoring
data based on either whole-year monitoring, including flood periods, or based on
low-flow or seasonal monitoring only. Assessing average concentrations for the
summer period, after the spring floods, 1s likely to be of greatest Interest for
assessing biological Impact. This period has maximum biological production and
1s also the season during which most of the fish samples have been collected.
There 1s also evidence to suggest that spring flood PCBs are largely sorbed on
sediment particles, whereas concentrations associated with low flows are
primarily dissolved or sorbed to very fine particles, i.e., pass a 0.45-jxm
filter. Thus, they are more readily available to enter the food chain (Bopp et
a7., 1985).
Average PCB concentrations during the summer (June-September) have been
calculated. In doing so, the presence of many non-detects among the samples was
addressed. The robust log-probit analysis method (Helsel and Cohn, 1988) was
used to estimate averages in the presence of non-detects. No systematic bias
toward higher concentration events is expected to apply to the summer observa-
tions, although this may be the case for spring observations. In 1986 there were
no summer observations at Stillwater or Waterford, while summer observations at
Schuylervllle 1n this year as well as at Waterford in 1985 were all less than or
equal to the detection limit of 0.1 tig/7. In the latter case the log-probit
method cannot be used and the average has been arbitrarily set to one-half the
detection limit. The calculated summer average PCB concentrations are shown in
Figure B.3-12 and summarized 1n Table B.3-13. Recent 1988-1989 summer average
PCB concentrations are fairly uniform for all locations and are on the order of
B.3-24
-------
0.03 - 0.04 ug/1. Summer average concentrations at Fort Edward tend to be less
than or equal to those downstream, whereas for the full year concentrations.
Including spring runoff, the average concentration at Fort Edward 1s higher than
that found downstream.
B.3.3.4 Other Sources of Nater Coluim Data
Uaterford Treatment Plant Data
The City of Uaterford operates a water works serving a population of
approximately 12,000 persons 1n the Towns of Uaterford and Halfmoon and the
Village of Waterford. This is the first water treatment facility downstream of
Fort Edward drawing water from the Hudson. In 1975, when the USGS began
collecting PCB data in the river at Waterford, they also began collecting raw
water input and finished water output data at the Uaterford treatment plant, in
cooperation with the Board of Uater Commissioners of the Town of Uaterford and
the NYSDEC (Schroeder and Barnes, 1983b). The water for the treatment plant is
drawn from a location 0.5 km upstream of the US Highway 4 bridge, where Hudson
River water samples are also taken.
Data collected 1n cooperation with the USGS run through the end of water
year 1983. In addition, approximately bimonthly data for November 1983 -
February 1985 and March 1987 - October 1989 were available from the Uaterford
Uater Uorks (Metcalf & Eddy, 1990). The data through 1983 were often collected
concurrently with samples from the Hudson River at Uaterford and can be used to
provide a check on that data. Since September 1982, no PCBs have been reported
above the detection limit (generally 0.1 |ig/7), in either raw Intake water or
treated water.
Lamont-Doherty Study of 1983
A detailed study of PCB transport 1n the Upper Hudson, conducted by
researchers of the Lamont-Doherty Geological Observatory 1n 1983 (Bopp et ?.,
1985), involved an investigation of spring/summer 1983 PCB transport in the Upper
B.3-25
-------
Hudson, which was a period of relatively high flows. The Laraont-Doherty study
included the collection of data not available from USGS sampling. In addition
to sediment cores, this study included 20 large-volume filter samples of
suspended matter and fifteen 9-20 liter unfiltered samples containing water aad
suspended matter, collected from Troy to Glens Falls. Unlike USGS monitoring^
detailed component analysis was undertaken for these samples. The data collected
were used to form an empirical transport model, which used as input USGS measures
of suspended sediment and flow, to predict total PCB load.
PCBs 1n samples of suspended matter were found to match standard Aroclor
mixtures, e.g., Aroclor 1242, reasonably well. Samples taken during a high-flow
period were found to have significantly higher PCB levels (6.33 ppm) than those
taken during a low-flow period (0.69 ppm). Some of the drop in PCB levels at
locations downstream of Fort Edward was attributed to dilution from tributaries
joining the Hudson.
Water samples were filtered and both filtrate and particles were tested for
PCBs. Comparisons of water versus suspended matter PCB concentrations were made
to derive a distribution coefficient. An inconsistency was found in that the
concentration of PCBs on suspended matter filtered from the 9-20 liter samples
was generally two to four times higher than the levels from the large-volume
filter samples collected at the same locations on the river. This discrepancy
could possibly be accounted for by an equilibration between water and suspended
matter during storage of water samples prior to filtration and testing. Another
possible explanation was the difference in filter size used to collect the
particulate sample (1.2 microns) and to separate the water samples (0.7 micron).
Seasonal variations in concentrations of PCBs in water were found. Pre-
spring runoff showed the lowest PCB levels. Summer samples showed higher PCB
concentrations, accounted for by boat traffic and increased use of locks, which
would resuspend sediments.
B.3-26
-------
Bopp et al. (1985) were able to derive in situ partition coefficients for
PCBs on a quasi-homologue basis, based on packed column analysis. These analyses
also indicated the possibility that water-column PCB distributions may not
reflect equilibrium conditions and that dissolved phase PCB concentrations may
be higher than those predicted by equilibrium. The homologue distribution of
PCBs on suspended matter was readily interpreted as Aroclor 1242-1 ike, similar
to PCBs in the aerobic sediment layers. The dissolved phase homologue
distributions were not as readily explained, although preferential partitioning
of the lower chlorinated homologues to the dissolved phase appeared to be a
likely possibility, I.e., lower chlorinated homologues have lower partition
coefficients and, therefore, higher levels in the dissolved phase. The greater
volatility of the lower chlorinated PCBs and their production in the anaerobic
sediment layers could possibly confound this interpretation. The general
homologue pattern agreement between water column and surface sediments led Bopp
et al. (1985) to conclude that little or no release of PCBs from the anaerobic
sediments was occurring on a substantive basis 1n comparison to the mixing and
resuspension of the surficial sediments.
Bopp et a7. concluded that PCB transport 1s tied to suspended matter
transport, with the majority of PCB transport to the Lower Hudson occurring
during the 10 to 20 days per year of major sediment transport.
NYSDOH Hater Column PCBs
As part of their macro in vertebrate sampling program, NYSDOH also collected
water samples and analyzed them for PCBs. This and other recent data received
from Dr. Bush at NYSDOH, require additional evaluation. The macroInvertebrate
studies are discussed in a later section.
Remnant Deposit Containment Honitoring Program
As part of remedial activities at the PCB remnant deposits at Fort Edward,
General Electric 1s conducting a baseline environmental monitoring program, which
will continue through and follow the 1n-place containment of Remnant Deposits 2,
B.3-27
-------
3, 4 and 5 in accordance with the Administrative Order on Consent II CERCLA-
90224. The baseline monitoring includes sampling of many environmental matrices,
Including water column PCB concentrations. Results of the first phase of this
effort cover August through December 1989 (Harza Engineering, 1990).
During the baseline monitoring, samples were taken weekly or biweekly at
ten water quality stations. Two of these stations are upstream of Fort Edward,
one 1n the Sherman' Island Pool and one just below Bakers Falls, while five are
1n the area of the remnant deposits at Fort Edward above Rogers Inland. Station
E-5 was located downstream of the Route 197 bridge on Rogers Island and 1s just
below the USGS monitoring station. The remaining two stations were at Channel
Marker 175, below Fort Miller Dam and Lock 6 (E-6), and at Channel Marker 13, two
miles north of a NYSDEC boat ramp at the Erie Canal and Hudson River confluence,
near Uaterford (E-7). Raw water samples were analyzed for PCBs with an
approximate analytical detection limit of 0.1 yg/7. In addition, dialysis bags,
filled with 4 ml of hexane to concentrate PCBs, were suspended in the water
column and analyzed biweekly. As the concentration factor for the dialysis bags
is unknown, they can be used to Indicate qualitatively the presence of PCBs, but
not ambient concentrations.
The 1989 monitoring program unfortunately missed the spring runoff period.
All raw water samples were reported to be below the detection limit of 0.1 yg/7.
This result 1s consistent with USGS monitoring data for water year 1989 at Fort
Edward, which had a lower detection limit and showed detectable concentrations
1n the 0.01-0.1 yg/7 range. During the same period the dialysis bag concentra-
tors occasionally detected PCBs at all stations except the uppermost (Sherman
Island). All the detects were Identified as Aroclor 1242.
B.3.4 Fish and Other Aquatic Biota
Substantial declines in average PCB burden In Upper Hudson fish were
observed 1n the years after 1978 (Sloan et al., 1983, 1984; M. P. Brown et al.,
1985). Analysis of the data reveals that these declines 1n concentration have
proceeded at a slower rate 1n more recent years. It 1s unclear to what degree
B.3-28
-------
the abnormally low spring floods of the 1980s have affected PCB levels and may
be responsible for the observed declines. It does appear that total PCB con-
centrations 1n fish on a lipid (fat)-basls can be closely predicted from summer
average water column PCB concentrations.
Because PCBs are typically stored (bioaccumulate) in fatty (lipid) tissues,
1t 1s sometimes useful to normalize the PCB levels 1n fish and express them on
a lipid basis, i.e., PCB content in fish expressed as ug-PCB/g-flsh lipid.
Whether or not normalized to lipid content, levels of the higher chlorinated
congeners 1n largemouth bass were approximately stable from 1981-1988. Reported
Aroclor 1016 levels, representing less chlorinated congeners, appear to have
continued to decline for all species during this time period. Given the slow
rate of reduction of Aroclor 1254, 1t is unclear when or if natural processes
will reduce the PCB burden in fish to acceptable levels. Furthermore, potential
changes in PCB levels in the water column and sediment caused by possible scour
and resuspension of sediments would likely cause at least temporary Increases In
PCB levels 1n fish and aquatic biota.
In addition to fish data, some monitoring of invertebrates is also
available for the Hudson. From 1976-1985 multlplate samples and chironomld
larvae have been monitored by NYSDOH. These data are discussed at the end of
this section, following the discussion of the fish monitoring program.
B.3.4.1 Fish Sampling
Data on concentrations of PCBs 1n Upper Hudson River fish collected by
NYSDEC between 1975 and 1988 were used in this study. While over 30 species of
fish are represented 1n the data, the majority (75 percent) of the samples are
from half a dozen species Including striped bass, largemouth bass, brown
bullhead, pumpktnseed, American shad, and American eel. Approximately two-thirds
of the samples tested were standard fillet samples.
B.3-29
-------
Samples Prior to 1975
While PCBs are known to have been discharged into the Upper Hudson River
since the 1940s, no testing for PCBs in fish 1s known to have been undertak^
before 1970. In that year, a nationwide survey of chemical pollutants In game
fish conducted by a popular magazine Included a sample of spawning striped bass
from the Hudson estuary 1n which 4.5 to 5 ppm PCBs 1n flesh and 11 to 12 ppm PCBs
1n eggs were reported (Boyle, 1970). NYSDEC had been analyzing fish for ODT and
other pesticides statewide since the early 1960's. In 1971, NYSDEC added PCBs
to their analyses, although no results were released publicly until 1975
(Sanders, 1989).
Fish data collected and analyzed for PCBs 1n the 1970-74 period are
summarized by Spagnoli and Skinner (1977). These 1970-74 Hudson River samples
include one smallmouth bass collected at Warrensburg (Upper Hudson) and 146 fish
from 11 species collected below the Federal 0am (Lower Hudson). The highest
observed concentration from below the Federal Dam was in a largemouth bass,
reported as containing 53.81 ppm wet weight total PCBs 1n the 1970-72 period.
This sample was taken prior to the removal of the Fort Edward Dam.
In August 1974, a USEPA team obtained water, sediment and fish samples from
upstream and downstream of the GE discharge at Fort Edward. A sample of 42
shiner minnows from below the GE discharge showed an average concentration of 78
1*9/9 (ppm) PCB as Aroclor 1242, while one rock bass was reported with 342 pg/g
(Nadeau and Davis, 1974). It should be noted that samples collected at control
Station 0 above Bakers Falls were not non-detects. PCB levels at Station O were
reported as 7.0 |ig/gm in shiner minnows and as 17.0 |tg/gm in yellow perch. The
latter level 1s higher than the average for all NYSDEC yellow perch samples in
the Upper Hudson. Few other samples have been reported from Bakers Falls.
B.3-30
-------
Samples 1975-1976
Following the USEPA investigation, NYSDEC undertook more detailed
monitoring of PCBs in fish from both the Upper and Lower Hudson. A total of 440
Hudson River (Upper and Lower) fish samples were analyzed 1n 1975-1976, the
results of which NYSDEC provided for incorporation Into the TAMS/Gradient
database.
The 1975-1976 fish collections were made by regional USFUS personnel who
were instructed on specific species and sizes of fish desired, location of
stations and time tables for collection. Target species for the Hudson included
smallmouth bass (Hicropterus dolonieni), largemouth bass (Nicropterus salmoldes),
brown bullhead (Ictalurus nebulosus), goldfish (Carasslus auratus), white sucker
(Catastonus comersoni), striped bass (Ho rone saxatilis) and various other
estuarine species. Other species were occasionally obtained as available.
Attempts were made to sample small, medium (minimum legal) and large representa-
tives of each species.
Analyses were conducted by several different state laboratories, apparently
using the methodology of Bush and Lo (1973) and reported against standards for
Aroclor 1242 or 1016 and Aroclor 1254. Aroclor 1221 was not analyzed. The
nominal detection limit of the method was 0.01 ppm, although some of the labs
reported results only as low as 0.1 ppm.
Samples 1977-1988
In 1977-1979 NYSDEC monitoring methods were refined and standardized.
Collection has continued to the present, although no results after 1988 are
available. As of 1988, NYSDEC changed the frequency of sampling from yearly to
every other year. Samples from 1990 were collected, but PCB analyses have not
been completed. NYSDEC provided data covering the period of 1977 through 1988,
which contains 7,373 Upper and Lower Hudson fish or fish composite analyses.
These samples have been collected on a regular basis, with the intent of sampling
given species at predetermined locations within two weeks of a specified date in
B.3-31
-------
order to minimize potential seasonal effect. Records are very limited for 1981
and 1987. Table B.3-14 provides a summary of the total number of fish sampled
for the Upper Hudson from 1975-1988.
*
.V
Information in the fish database, in addition to chemical analyses, usually
includes species, method of preparation, weight, length, percent lipid and, if
determined, sex and age. In addition to a descriptive sampling location, river
mile numbers and geographic designations are associated with each sample.
s>
Sample collection, preparation and analytical methods are described in
Armstrong and Sloan (1981). The desired sample size for each species collected
was 30 individuals, although the availability of fish did limit sample size in
some cases. A single fish sample generally consisted of a composite of one to
three fish. For fish longer than 150 mm, standard fillets of whole sides of
scaled fish were prepared for analysis. For brown bullhead samples, the skin was
removed from the fillet. Samples of fish shorter than 150 mm were analyzed whole
with head and viscera removed.
Aroclor concentrations in fish were determined by comparison to commercial
Aroclor standards. This method was not able to distinguish between Aroclors 1242
and 1016. The detection limit for each Aroclor tested was 0.1 ppm. For each
sample the percent lipid was determined as the percent by weight of tissue
soluble in petroleum ether.
PCB Levels 1n Fish
This section summarizes the Upper Hudson PCB monitoring results for fish.
Additional statistical analyses are presented in Section B.4. Samples collected
at River Mile 153, just below the Federal Dam, are Included with those of the
Upper Hudson, since they represent a resident, freshwater, rather than estuarlne,
population, which 1s exposed to PCBs transported over the dam.
B.3-32
-------
Although over 30 fish species have been sampled 1n the Hudson River, a few
of these species account for the majority of the samples collected. These
species are pumpklnseed, largemouth bass, brown bullhead, goldfish (carp), white
perch and yellow perch. Overall average Aroclor levels for these species In
River Miles 153 to 195 of the Upper Hudson for 1975-1988 are provided in Table
B.3-15. Aroclors 1016 and 1254 are the dominant PCB mixtures reported; Aroclor
1221 represents the smallest PCB fraction. Overall, the highest average PCB
concentrations have been found 1n goldfish (carp) with 1975-1988 average Aroclor
levels of 32.0 ppm (Aroclor 1254) and 91.6 ppm (Aroclor 1016). That goldfish
have the highest average PCB levels is not altogether surprising as they also
have the highest lipid content (9.3 percent). The average 1975-1988 Aroclor
levels 1n largemouth bass, perch and bullhead are <30 ppm, although total PCB
levels 1n Individual largemouth bass have been reported as high as 370 ppm.
The average of total PCB levels 1n all fish for recent years (1986-1988)
also shown 1n Table B.3-15, 1s 10.9 ppm. The upper 95 percent confidence found
on this mean, used 1n the preliminary human health risk assessment (B.6), is 12.0
ppm.
Tables B.3-16 through B.3-18 provide summary statistics by river aile
sampled for largemouth bass, pumpklnseed, and brown bullhead, respectively.
Recent data for 1986-88 show that median PCB levels In fish range from 1 to 30
ppm depending on species and river mile. Additionally, the variability of PCBs
in fish from a given location and year, as measured by the standard error of the
estimate of the mean, has dropped as the PCB levels have declined.
Figure B.3-13, which plots mean, total PCB levels In brown bullhead
(skinned standard fillets) as a function of year and river mile, provides an
overall perspective of the general decline of PCB levels over time and by river
mile. PCB levels in the vicinity of River Mile 190 are the highest, but samples
from only a few years are available here.
B.3-33
-------
Surprisingly little fish data are available from the Thompson Island Pool
area (River H1le 188-94). Regular monitoring there began 1n the latter half of
the 1980s. The most complete data set available for the Upper Hudson above the
Federal Dam 1s that at River Mile 175. Data here begin 1n 1975 and continue,
with some Interruptions, through 1988. Figure B.3-14 summarizes the trends to
mean Aroclor levels for four fish species at this location as ppm wet weight
Aroclor 1254 and Aroclor 1016. Aroclor 1016 levels are shown on a log scale, as
they were extremely high for certain species In the 1977-78 time period. Aroclor
1221 was quantltated also, but generally at very low levels compc-ijd to the other
two Aroclors. A notable trend 1s that during the 1977-78 period the ratio of
Aroclor 1016/1254 levels 1n fish are elevated, I.e., the lower chlorinated
congeners (1016) dominated the higher chlorinated congeners (1254) during the
late 1970s. This elevation may be attributable to scour of dechlorinated
sediments from depth. Another possible explanation for the shift from Aroclor
1016 dominance to Aroclor 1254 In recent years 1s that the lower chlorinated
congeners were more rapidly released and dissipated 1n the late 1970s and early
1980s, whereas the higher chlorinated congeners have tended to bioaccumulate more
and are less rapidly released by fish.
L1p1d-Based PCB Concentrations
Differences between PCB levels In different fish species are probably due
to both differential feeding patterns and lipid content. Fish lipid (fat)
typically accumulates PCBs more than other, less fatty tissues, because PCBs are
highly lipophilic compounds, i.e., PCBs are more soluble in fat than water.
Normalizing PCB levels on a fish lipid weight basis has sometimes proved useful
in comparing measurements among species. The average percent lipid of goldfish
was 9.2, pumpklnseed was 3.3, brown bullhead was 2.9 and largemouth bass was 1.4,
with skewed distributions. There appears to be little correlation between weight
and PCB body burdens 1n most species. Age of the fish sampled was not often
determined.
B.3-34
-------
Besides accounting for species differences, accounting for 11p1d content
1s Important because fish lipid content appears to have changed from samples
caught from one year to the next. Dividing the PCB concentration (wet weight)
by the measured lipid content (g-lipid/g-fish) of the sample, one obtains the PCB
concentration per gram (mass) of f1sh-11p1d (ug-PCB/g-11p1d). Looking at all
samples for all species at all Upper Hudson sampling locations on a lipid basis,
the median PCB levels have declined from 1829 ug-PCB/g-11p1d 1n 1977 to 271 ug-
PCB/g-1 1p1d In 1988, with a 95 percent upper confidence bound estimate on the
mean of 484 (Table B.3-19).
L1p1d-based means of Aroclor 1254 and 1016 concentrations by year 1n the
predominant fish species sampled at River Mile 175 are shown 1n Figure B.3-15.
On a lipid basis, largemouth bass have usually shown the highest Aroclor 1254
levels. This may reflect their position as top carnivores in the aquatic food
chain or their low fat content. Further, the I1p1d-based Aroclor 1254 levels In
most species appear to have been relatively constant since 1982. Error bars are
omitted, from the multiple species plots for legibility. Trends 1n largemouth
bass, with error bars, are shown 1n Figure B.3-16. (An error bar shown as a
vertical line at each year representing the 95 percent upper and lower confidence
bounds on the mean.)
In addition to the monitoring at River Mile 175, there Is a good continuous
record of sampling of brown bullhead at River Mile 153 just below the Federal
Dam, for the period 1977-1988. PCB levels here were on average much lower than
those observed at River Mile 175, presumably due to dilution of PCB concentra-
tions by the flow of the Mohawk River. Average lipid-based PCB concentrations
1n brown bullhead show a regular exponential decline for Aroclor 1016 components
and a less dramatic decline for Aroclor 1254 (Figure B.3-17).
B.3.4.2 Other Chemicals In Fish
The NYSDEC database contained analyses for chemicals in fish other than
PCBs. Summary statistics for these other chemicals are shown 1n Table B.3-20.
B.3-35
-------
Metals such as mercury, cadmium, chromium, copper and zinc were frequently
detected. Median mercury and zinc concentrations were on the order of 0.4 ppm
and 10 ppm, respectively. Organic compounds frequently detected include
pesticides such as DDT, heptachlor, and dleldrln. Hexachlorobenzene a^d
hexachlorocyclohexane were also detected frequently.
B.3.4.3 NYSDOH Macroinvertebrate Studies
As part of the Hudson River PCB Reclamation Demonstration Project, the New
York State Department of Health (NYSDOH) conducted b1omon1tor1ng studies from
1976 to 1985 using caddlsfly larvae, multlplate samples and chlronomld larvae
(Simpson et a7., 1986). These studies Included long-term b1omon1tor1ng efforts
from 1976 to 1985 as well as two short-term biological uptake studies 1n July and
September of 1985.
Long-Term B1omon1tor1ng Study
From 1976 through 1985, artificial substrate samplers.(multlplates) were
placed at 17 sites along the Hudson River from Hudson Falls to Nyack, New York
(Novak et aJ.t 1988). These samplers were collected each year after a period of
five weeks during the months of July, August and September. PCBs in the samples
were reported as Aroclors 1016 and 1254. The resulting PCB concentrations 1n
the multlplate samples represented a composite of sediment, algae, plankton and
various macrolnvertebrates. Invertebrates collected 1n the multlplate samplers
Included the following taxonomlc groups: Chlronomldae, OUgochataeta,
Trlchoptera, Ephemeroptera, Amphlpoda and EHmldae. Chlronomld larvae and pupae
were the most abundant Invertebrate component from Fort Edward to Saugertles,
comprising up to 86 percent of the total macroinvertebrate population at Fort
Miller and Waterford.
From 1978 to 1985, caddlsfly larvae were collected by hand-picking
Individuals from rocks at five designated sites: Hudson Falls, Fort Edward, Fort
Miller, Stillwater and Waterford. Caddlsfly collections were made 1n June, July,
August and September of each year.
B.3-36
-------
Measured PCB levels 1n the 1985 multiplate samples for September ranged
from 0.25 ppm at Hudson Falls (the control site) to over 6 ppm at Fort Edward.
Multiplate monitoring from Fort Miller to Waterford resulted in PCB levels of 4
to 5 ppm. The multiplate samples at any one site appeared to show a consistent
decline In PCB concentrations from early summer to later summer In any particular
year. Larger scale trends or relationships in either time or with sample
location are difficult to detect, because of the extremely wide variation In the
sample results. Average PCB concentrations 1n multiplate samples generally
showed a decline from 1976 - 1980. Nevertheless, average PCB concentrations
increased in 1981 and remained high through 1985. Multiplate samples from the
Thompson Island Pool and downstream showed significantly higher PCB concentra-
tions than samples taken upstream of Fort Edward. Yet, no significant trends are
apparent when comparing Fort Edward results with those at Waterford.
The results of the caddisfly biomonitorlng efforts show a decline in total
PCB concentrations from 1978 to 1980. As Is true of the multiplate data, spatial
trends are not readily apparent from the caddisfly samples. Measured PCB levels
1n macroinvertebrate tissues generally ranged from 20 to 60 ppm (dry weight) in
1979 and from 20 to approximately 40 ppm (dry weight) in 1985. During the sane
collection periods, macroinvertebrates collected at Hudson Falls, the control
site, exhibited PCB tissue residues less than 10 ppm.
The data from this study exhibit a great deal of random variability. The
bulk of the data are from multlplates, which collect sediment as well as living
matter. In theory, It should be possible to distinguish sediment versus
biologically-based PCBs by adjusting the observations to a lipid basis; lipid
content of samples was reported. An additional factor contributing to the
variability 1s that 1n almost every year a downward trend by month was observed
at most stations, based generally on three samples. The cause of this phenomenon
is not known, but the limited number of observations 1n this yearly cycle may
have obscured the Influence of other factors.
B.3-37
-------
Examination of the complete run of samples at the Fort Miller multiplate
and caddisfly station (PCB-5) shows what appears to be a largely random pattern
of total PCBs on a dry weight basis (Figure B.3-18). The larger mean PCB
concentration for the period 1981-82 is largely attributed to the increased
variability of the data for this period. Adjusting to a lipid basis actually
increases the total variability represented by outlying data points. Only a
slight trend is suggested in this figure, where PCB levels may have declined from
1976 to 1980, then remained approximately stable. At Stillwater, the Hpid-based
values appear to be almost entirely random (Figure B.3-19).
A comparison between the confidence intervals on the overall means for all
sample years by sample locations (Figure B.3-20) shows differences between the
low values observed above Fort Edward and values at downstream stations. Lipid-
based means at all stations by year (Figure B.3-21) appear to show a decline from
1976 to 1979, then relatively constant levels except for a jump upward in 1982.
Again, this apparent trend Is difficult to confirm as a consequence of the
limited number of samples and possible differences 1n sampling or analytical
protocols throughout the duration of the monitoring.
Short-Tern Biomonltorlng Study
Short-term biomonltoring investigations using the chlronomid larvae,
Chlronoaus tentans, were also performed'by the NYSDOH during July and September
1985 (Novak et a/., 1990). The monitoring method consisted of placing 25
laboratory-reared chlronomid larvae in nylon mesh envelopes or packets that were
exposed to the water column. Envelopes were placed, In groups of ten, in steel
mesh baskets at the primary collection site and monitored at 0, 1, 2, 4, 8, 12,
24, 48, 72 and 96 hours. Chlronomids were placed at four sites, including two
at Thompson Island Pool, one at Bakers Falls and one at F1sh Creek, and monitored
at 96 hours. Packets of chlronomids exposed to the sediment at a collection site
located on the eastern shore of the Thompson Island Pool were also collected at
96 hours. Water column samples were obtained during the same collection
Intervals for each site.
B.3-38
-------
Results of this Investigation indicated that chironomids accumulated PCBs
ranging from 0.1 to 7 ppm after 1 to 96 hours of continuous exposure, whereas
larvae exposed to sediments near the Thompson Island Pool for 96 hours contained
over 100 ppm. Water column PCB levels were 1n the range of 0.03 - 0.1 pg/l
during the experiment. The ratio of the PCB levels 1n the chironomids (in ppb)
to the ambient PCB concentrations 1n the water column (defined as the bloaccu*-
ulation factor or BAF) were on the order of 10* to 10*.
A significant conclusion from this study was that the PCB congener pattern
found 1n tissues of chironomid larvae differed substantially from the congener
pattern observed in water. Using capillary column gas chromatography, the
Investigators were able to isolate PCB congeners in both the water column and
chironomids. The most abundant congeners in chironomid tissues were 2,4,2',5'-
tetrachlorobiphenyl and 2,3,6,4'-tetrachlorobiphenyl. In contrast, the
predominant congeners 1n water were 2,2'-d1chlorobiphenyl, 2,6-dichloroblphenyl,
and 2,6,2'-trichlorobiphenyl. These findings suggest a number of possible
explanations. One explanation could be that chironomid larvae may selectively
bioconcentrate the more highly chlorinated congeners which are present at
relatively low concentrations in water. Another factor which could explain the
observed results is that the lower chlorinated congeners were present in the
water but below detection limits.
Although this study presents some very Interesting congener-specific
results, they are too limited 1n scope to provide clear Indications of either
congener-speclf1c biฉaccumulation or congener-specific comparisons of PCBs in
sediments. (No congener-specific sediment data were obtained.)
B.3-39
-------
B.3.S PCB Concentrations in Air and Plants
B.3.5.1 Air
Monitoring Near Fort Edward
A1r monitoring efforts for PCBs and other air toxics have been conducted
1n the Upper Hudson River study area from late 1976 to 1982 by NYSDEC/NYSDOH and
various researchers and as recently as 1989/90 by contracto "u (Harza/Yates-
Auberle) for General Electric.
From January through August of 1977, NYSDOH collected air samples at five
locations 1n the Upper Hudson Valley to determine ambient PCB concentrations.
While the Glens Falls and Warrensburg samples showed no detections above the
0.020 yg/m3 detection 11m1t, results from the Hudson Falls and the Fort Edward
samples demonstrated high levels of total PCBs, ranging from 0.060 yg/m3 to 3.26
yg/m3 (Malcolm-Pirnie, 1978). Atmospheric levels of PCBs 1n the Fort Edward area
were reported to decrease from 1 yg/m3 down to 0.3 yg/m3 after the cessation of
PCB use by General Electric In their Hudson Falls and Fort Edward capacitor
plants 1n 1977 (Llmburg, 1984).
In 1979, NYSOEC conducted an air monitoring survey for PCBs around various
dumps and landfills (Caputo and Fort Miller dumps, Remnant Area, Moreau and Site
3a and Buoy 212) 1n the Hudson Falls/Fort Edward area bordering the Upper Hudson
River (see Table B.3-21). Values ranged from 5 to 15 yg/m3 total PCBs at the
Moreau and remnant areas and 24 to 300 yg/m3 total PCBs at the Fort Miller and
Caputo dumps, respectively. At the Caputo dump, where the soil was reported to
contain 5,000 mg/kg PCBs, air monitoring for PCBs before and after capping of the
site showed that average ambient PCB concentrations decreased from 118 yg/m3
before capping to 0.26 yg/m3, once the site was capped (Shen, 1982).
Two air samples taken over Lock 6 In the summer of 1980 yielded Aroclor
1242 concentrations ranging from 0.11 yg/m3 to 0.52 yg/m3 (Table B.3-21).
Aroclors 1221 and 1254 were not detected (NYSDEC, 1981). Ambient PCB levels
B.3-40
-------
monitored at farm fields in 1981 near the Hudson River, however, showed low PCB
concentrations of approximately 0.005 |ig/m3 (NUS, 1984).
In the early 1980s, NYSOEC and the Boyce Thompson Institute for Plant
Research of Cornell University conducted a joint air/plant monitoring effort near
the tailwater of Lock 6 to determine if volatilization of PCBs from the water
column was occurring (Buckley and Tofflemlre, 1983 and 1984). Between August
1981 and September 1981, seven air samples were taken. Aroclor 1242 was detected
in all seven samples, ranging from 0.031 to 0.06 pg/m3. Aroclor 1254 was
detected in three of the samples at levels up to 0.0013 yg/m\ During this
study, a vertical PCB gradient was also noticed, when airborne PCBs were measured
simultaneously at heights of 1 and 4.5 meters above the water.
In August 1986, NYSDEC collected three sets of ambient grab samples in
duplicate at the proposed containment site (Site G), the Fort Edward Landfill,
the Bourgoyne Avenue School and Lock 7 of the Champlaln Canal (USEPA/NYSDEC,
1987). The highest ambient PCB concentration measured was 0.083 yg/m' at Lock
7. Site G and the Fort Edward landfill samples contained PCB levels below the
detection limit of 0.007 jig/ra3. The Bourgoyne Avenue School sample information
was not available at the time of the report.
Again In 1987, NYSDEC conducted air monitoring from April 2, 1987 to July
16, 1987 at the Kingsbury Landfill, located north of Fort Edward. In 76 of 105
samples taken over April and May 1987, Aroclor 1016/1242 was detected with a
maximum concentration of 0.49 }ig/m3. Aroclor. 1248 was detected in 5 of 105
samples with a maximum concentration of 0.52 iปg/m3. Both the Kingsbury and Fort
Edward municipal landfills were the burial site of several thousand tons of PCBs.
Neither these landfills, nor others In the area, are part of the current
Investigation of the Hudson River Superfund Site. The results are presented here
as evidence of PCBs having been monitored in the air within the vicinity of the
site.
B.3-41
-------
Most recently, 1n connection with Its Remnant Remediation Project, GE
conducted baseline pre-remediatlon air monitoring from August to November 1989
(Harza, 1990). Fixed air monitoring stations were planned at five locations
along the river: two residential areas, one upwind location, one downwlh^l
location and a farming receptor location two miles south of the remnant area.
Because of site access problems, only three of the five sites were monitored.
Once every three days, two l-Hter/m1nute, 24-hour air samples were taken using
a two-channel (two separate samples taken simultaneously) air monitoring station
three feet above the ground. The samples were analyzed by NIGlH Method 5503
(desorptlon of fluorlsll tubes with solvent followed by GC-ECD analysis) with a
detection limit of 0.05 |ig/sample. In total, 84 samples were collected. Seven
samples showed levels of PCBs above the detection 11m1t (0.05 iig/sample), with
a maximum value of 0.23 pg/m1. Of these seven detects, three were from a
residential area (location A2), two from a downwind receptor (location A4) and
one from the farm area (location A5). Although PCBs were detected 1n this
Investigation, the two sample channels often gave Inconsistent results; one
channel contained PCBs while the other did not. This occurrence may have been
due to sampling or analytical problems or both.
Maximum ambient background PCB concentrations In air measured by New York
State during a statewide monitoring effort, listed at the bottom of Table B.3-21,
provide a perspective on PCB levels 1n air 1n the vicinity of the site. Maximum
ambient air PCB concentrations measured in this effort ranged from 0.002 pg/m3
1n Syracuse and Rensselaer (urban areas) to 0.007 jig/m3 in Staten Island (NYSOEC,
1982-4). These maxima are one to two orders of magnitude lower than the maxima
detected during GE's baseline monitoring study and values measured by NYSDEC near
Lock 7 in August 1986.
Lamont-Doherty Investigations
Three studies of PCB volatilization and properties governing volatilization
were conducted at the Lamont-Doherty Geological Observatory. The first two of
these studies (Bopp, 1979 and Bopp, 1983) dealt with the estimation of PCB
properties that govern volatilization and the application of these properties to
B.3-42
-------
actual conditions 1n the river. In particular, Bopp developed estimates for the
molecular dlffusivlty, Henry's Law constant and gas exchange rate of PCB
homologues at Troy and Poughkeepsle. Assuming a 20-day water transit time
between the two points, a simple model of gas exchange based on the derived
parameters worked quite well. The data and model both showed gas exchange rates
for PCBs 1n the Hudson that varied Inversely with the degree of chlorlnation.
For low chlorinated homologues (d1- and trlchloroblphenyls), Bopp (1983)
estimates that about 40 percent 1s lost, because of gas exchange, as a given
parcel of water travels from Troy to Poughkeepsle. For the tetra- and
pentachloroblphenyls, 10 to 20 percent is lost as a result of gas exchange. For
the higher chlorinated homologues, Bopp (1983) calculated even lower loss rates
and, in fact, was unable to measure any water column loss attributable to gas
exchange. This finding has Important ramifications for the fate of the nore
highly chlorinated homologues in the Lower Hudson. Since gas exchange and
blodegradation did not appear to remove these PCBs, they would remain In the
sediments of the Lower Hudson or be transported out to the New York Bight.
The remaining study from the Lamont Doherty Geological Observatory was
conducted by Warren et a?. (1985) and consisted of a series of laboratory studies
to refine earlier estimates of Henry's Law constants for PCBs on a quasi-congener
basis. Warren et a7. (1985) determined Henry's Law constants for individual PCB
congeners or pairs of similar congeners under a range of temperature conditions,
which are directly applicable to conditions formed 1n the Hudson.
B.3.5.2 PCB Uptake by Plants
Early Studies
In 1977, PCBs were found in vegetation growing on dump sites in the Fort
Edward area in a monitoring study performed by Weston Environmental Consultants
(Weston, 1977). Following the discovery of high PCB concentrations in vegetation
at the Fort Edward dump, NYSDEC and the Boyce Thompson Institute for Plant
Research at Cornell University conducted a number of studies from 1978 to 1981
to measure airborne-PCB uptake in several vegetative species 1n the Fort
B.3-43
-------
Edward/Hudson Falls vicinity. Their research evaluated PCB levels in trembling
aspen along easterly transects from the Fort Edward dump, the Buoy 212 dredge
spoil site, and a riffle area near Lock 6 (Tofflemire et al., 1981). PCB
measurements 1n aspen leaves ranged from 180 mg/kg at the dump, decreasing to
0.15 mg/kg at a distance of 820 meters from the dump. A similar declining trend
was reported for the Buoy 212 dredge spoil site and the riffle area.
From 1978-1980, total background PCB concentrations were measured in
goldenrod and trembling aspen within Washington and Saratoga Counties and were
found to be decreasing with time (Buckley, 1983). Average PCB concentrations 1n
goldenrod decreased from 0.32 mg/kg (ppm dry weight) 1n 1978 to 0.18 mg/kg in
1980, whereas average PCB levels in trembling aspen decreased from 0.12 mg/kg in
1978 to 0.07 mg/kg In 1980. Also 1n Washington and Saratoga counties, background
levels of total PCBs were measured 1n crops such as hay, corn, timothy grass,
perennial rye, brome grass, and orchard grass. Average total PCB background
concentrations ranged from 0.02 mg/kg (corn/silage) to 0.12 mg/kg (brome
grass/hay), as shown 1n Table B.3-22.
In September 1979, total PCB concentrations were measured in aspen, sumac
and goldenrod at five sites located at various distances (<1,200 m) and
directions from the Patterson Road PCB dump in Fort Miller, New York (Buckley,
1982). PCB levels 1n aspen ranged from 0.1 mg/kg to 58.2 mg/kg. PCB levels In
sumac suggested similar PCB uptake, with concentrations ranging from 0.11 mg/kg
to 68.6 mg/kg. Measurements for goldenrod ((0.26 mg/kg to 182 mg/kg) showed
approximately twice the rate of PCB uptake at the same sites as the aspen and
sumac measurements. This result suggests that PCB uptake by vegetation may be
species-dependent.
During September 1980, vapor-phase PCB accumulation in vegetation was
measured 1n the leaves of two varieties of sumac near an abandoned PCB dump 1n
the Fort Edward/Hudson Falls area (Buckley and Tofflemire, 1983). The data,
shown 1n Table B.3-22, demonstrate PCB concentrations higher than background
levels with PCB concentrations ranging from 0.97 mg/kg to 5.2 mg/kg. The lowest
PCB concentrations were found in samples furthest (230m) from the source, whereas
B.3-44
-------
the highest PCB levels were in samples nearest (60m) the source.
In 1978 and 1981, air and vegetation measurements for total PCBs were made
near the tailwater of Lock 6. This study was conducted to investigate whether
PCB concentrations in air and vegetation were elevated above background
concentrations measured in Saratoga and Washington counties, because of PCB
volatilization from turbulent sediment resuspenslon at the tailwater of the dam.
The data indicate a general decrease in foliar PCB concentrations from 1978 to
1981, although the small sample size precludes a conclusive result. The PCB
concentrations 1n vegetation near the tailwater of the Lock 6 dam, which ranged
from 0.23 to 1.07 mg/kg total PCBs, are considerably higher than background
levels in the same species measured in 1979.
More Recent Studies
In July and August 1984, researchers from NYSDOH performed translocation
and transplantation studies on purple loosestrife at two locations to determine
PCB uptake from soil and air (Bush et al., 1986). The study site was located
alongside the Upper Hudson River, near Albany. The control site was two miles
from the river. Their results indicated that the main route of PCB uptake in
purple loosestrife is via the root system. Plants transplanted from the control
site to the Hudson River site showed an Increase 1n PCB levels from 0.010 mg/kg
to 0.210 mg/kg total PCBs. Additional evidence that soil, as opposed to air,
served as the major pathway of PCB uptake was provided by congener analyses. The
congeners in the plants that showed an appreciable increase over time were those
present in the soil samples, but not in air samples.
Although the primary uptake of PCBs In purple loosestrife appeared to be
systemic through the roots, PCB uptake at the air-leaf interface also occurred.
This pathway is suggested by the Increased PCB concentrations measured in purple
loosestrife plants that were translocated 1n plastic bags from the control site
with control site soil to the experimental site at the river. Further field and
laboratory Investigations revealed that 1n areas with high PCB levels 1n air
(0.140 |ig/m3), plants scavenged PCB congeners from the air. At lower ambient PCB
B.3-45
-------
levels (0.008 |ig/m3), monochloro- and dichlorobiphenyl were emitted from the
plant. This finding suggests that at least two processes are possible; PCB
accumulation from the air into the lipophilic, waxy sections of the leaf and
volatilization of the less chlorinated PCBs from the leaf to the air. ^
A similar research project Involving PCB uptake in crops and their fruits
was undertaken during July 1985 by the NYSDOH (Shane and Bush, 1989). The study
site was Patroon Island, situated 1n the Upper Hudson River, near Albany. Soil
analysis for PCBs at the study site revealed an average PCB concer' ration in soil
of 0.145 mg/kg. The control site was located in a rural area near Guiderland,
with an average total PCB soil concentration of 0.009 mg/kg.
Corn and soybeans were planted at both sites, but corn did not take root
at the study site. Strlngbeans and pinto beans, already growing at.the study
site, also were analyzed during the study. Corn plants were both transplanted
and translocated from the control site to the experimental site. Composite leaf
samples were taken from the end of July through mid - September for PCB analysis.
At the end of the experiment, plant fruits, e.g. beans and corn ears, were also
collected.
The average PCB concentrations In the 17 sets of samples taken from the
study site ranged from 0.001-0.050 mg/kg total PCBs, but were generally less than
0.025 mg/kg. PCB levels in fruits were 0.00013 mg/kg for corn kernels and
0.00055 mg/kg for soybeans, approximately 100 times lower than the leaf
concentrations. The results of the study Indicated that all four plant species
preferentially accumulated eight mono-, di-, tri- and tetrachlorobiphenyl
congeners, but not 1n the same order. Mono- and dichlorobiphenyls (2-mono-,
2,4'-di- and 2,2'-dichlorobiphenyl) were present as the highest percentages in
all species; the remaining five congener percentages varied by species.
None of the bean species yielded a significant increase in accumulation
over time. In addition, for both the translocated and transplanted corn plants,
a significant decrease over time in the content of the major congeners (mono- and
dichlorobiphenyl) was recorded. Shane and Bush (1989) note that although leaves
B.3-46
-------
of corn contain lipophilic waxy cutin, which would be likely to accumulate PCBs,
a greater elimination route for the less chlorinated* more water-soluble,
congeners exists by way of transpiration through the stomates of the corn leaf.
Thus, Shane and Bush concluded that although PCB accumulates in corn seedlings,
emissions or metabolism reduce the PCB load to the same levels as uncontaninated
plants 1n approximately three weeks of growth after emergence.
Very low PCB concentrations 1n vegetation were found In the studies by
Shane and Bush (1986) and Bush et aJ. (1989). Their results and the data froa
the earlier studies by NYSDEC and Boyce Thompson Institute demonstrate that crop
plants potentially accumulate PCBs from either soil, air or both. The older
studies show more dramatic uptake in response to the relatively high historical
PCB levels in air near concentrated dump sources.
B.3.6 Other Hedla
Few monitoring data for PCBs are available for other media of possible
concern. Limited groundwater sampling has been performed in the late 1970s, but
no more recent data were located. Surface soils near the river and adjacent
croplands have not been monitored. Virtually no data exists for upland biota.
As discussed in the 1984 FS (NUS, 1984), Weston Environmental Consultants
(1978) collected groundwater samples in 1977 from several dredge spoil and PCB
dump areas. Groundwater samples from the spoil and dump areas reportedly
contained PCBs at levels ranging from 16.7 jxg/J (ppb) near Site 212 to as high
as 693 |tg/7 in groundwater at the Old Fort Edward landfill. Weston calculated
PCB migration potential in groundwater from dredge spoil sites to be on the order
of 10'4 to 10'1 lb/year. These estimates were several orders of magnitude lower
than potential surface runoff/erosion losses of PCBs from dredge spoil areas,
which were estimated to be as much as 24.2 lb/yr (Weston Environmental
Consultants, 1978, as reported 1n NUS, 1984).
B.3-47
-------
B.3.7 Adequacy of PCB and Aroclor Measurement
B.3.7.1 Overview
In general, the sediment, fish and water column PCB monitoring data sets
reviewed here are acceptable for assessing trends 1n PCB concentrations.
Nevertheless, it 1s difficult to draw comparisons between data sets, because
different studies selected different approaches for quantltatlng (measuring) PCBs
and reporting them as various Aroclor mixtures. Most of the reported Aroclor
concentrations are based on the concentrations of only a portion of the congeners
present 1n the sample mixture. Furthermore, the actual distribution of congeners
1n an environmental sample, collected years or decades after the original
release, bear little resemblance to a commercial Aroclor mixture. For most, If
not all, the reported Aroclor results, e.g. Aroclor 1016, 1242, 1254, etc., the
data should be interpreted as representative of lower versus higher chlorine
containing PCB congeners, rather than as Indicative of a true commercial Aroclor-
like mixture.
PCBs - Aroclors, Congeners, and Analysis Methods
PCBs are a class of chemicals theoretically consisting of 209 different
congeners, of which approximately half are found 1n various Aroclor mixtures.
A single Aroclor mixture may consist of dozens of congeners. The number of
different congeners reportedly found 1n an Aroclor mixture may vary considerably,
depending upon the type of analysis performed and the quantity analyzed (A1 ford-
Stevens, 1986). For example, In a review article by A1ford-Stevens, the number
of congeners reportedly found In Aroclor 1242 ranged from 27 to 74. In a more
recent study, Schulz et a7. (1989) report a total of 77 different congeners
detected 1n Aroclor 1242 along with weight percent Information.
The weight percent of each congener or group of congeners reported In the
literature for a given Aroclor mixture has also varied. The weight percent
Information for homologue groups found 1n different Aroclors cited by USEPA
(1983) differs from the weight percent information reported by Webb and McCall
B.3-48
-------
(1973). The extent to which these differences can be attributed to analytical
technique or methodology and batch to batch variation in Aroclor mixtures is
unknown. O'Brien and Gere (1977) report that the results of an inter-laboratory
analysis of reference samples suggest that "ostensibly pure Aroclor mixtures from
different sources contained different percentages of PCB homologs." The Aroclor
standards were obtained commercially and directly from Monsanto.
Early determinations of Aroclor mixture compositions involved separation
of the mixture on a packed column gas chromatograph (GC) followed by quantitation
using an electron capture detector (ECD). A single Aroclor mixture can be
resolved on a packed column GC into two to eighteen separate peaks depending upon
the Aroclor mixture analyzed. Due to the poor resolution of PCBs on a packed
column, each packed column GC peak may represent as many as eight or more
individual congeners. The majority of the PCB analyses conducted for the Hudson
River Investigations have been packed column gas chromatograph (GC) methods.
Only recently have more precise, capillary column GC techniques capable of
resolving Individual congeners been employed. These techniques still generally
use the ECD for quantitation.
The sensitivity of the ECD to an individual congener varies directly with
the degree of chlorlnatlon. In fact, the ECD response may vary by as much as a
factor of 100 over the entire range of PCB congeners. Thus, the peak area
generated by a mono- or dichlorobyphenyl may be 100 times less than that
generated by a nona- or decachlorobiphyenyl, if all are present at the same
concentration in the sample. This finding has important ramifications when
analysis considers total peak area and not individual peaks to quantify PCB
levels or Aroclor mixtures. Minor changes 1n the highly chlorinated PCB
concentrations will be more readily reflected In the total peak area than nore
substantial changes in the lower chlorinated PCB concentrations.
In 1973, Webb and McCall formalized an approach for quantifying Aroclor
mixture concentrations using GC-ECD and the weight percent of PCBs represented
by each packed column GC peak. Using elemental analysis, GC-ECD and GC-ซass
spectrometry
-------
of PCBs contained 1n commercial solutions of Aroclors 1221 through 1260. The
weight percent of packed column peaks corresponding to Aroclors 1242, 1254 and
1260 are shown 1n Figure B.3-22. Individual peaks are Identified by their
retention time relative to the Internal standard pp'-DDE which 1s assigned^
value of 100. With this information, individual response factors can be
calculated for each packed column GC peak in a standard Aroclor mixture. The
total amount of PCBs present in any sample can then be calculated from the sum
of the amounts of all the individual peaks.
ฆj
Webb and HcCall peaks corresponding to Aroclor mixtures quantitated for
Hudson River samples are summarized in the tabulation that follows.
Webb and McCall Peaks Used to Quantitate Aroclor Mixtures
Source
Aroclor
1221
Aroclor
1016
Aroclor
1242
Aroclor
1254
ArocVor
1260
1977
Sediment
Survey
11 or
21
28, 47, 58
8 peaks
ID unknown
1984
Sediment
Survey
28, 47, 58
See text
See text
Fish
Monitoring
11
37, 40
125, 146,
174
1983
USEPA (in
NUS FS,
1984)
See text
See text
See text
1985
Bopp et
ซ7.
Analyzed for Total PCBs as
Sum of Peaks
28 - 174
3985
Bopp et
al.
47
B.3-50
-------
Interpreting Reported Aroclor Results
The Webb and McCall approach works well for estimating concentrations of
total PCBs in environmental samples Including weathered samples where gas
chromatograms bear little resemblance to standard Aroclor mixtures. A problem
arises, however, when total PCBs are calculated in a weathered sample using the
Webb and HcCall approach, but then reported as the concentration of a single
Aroclor mixture or the combination of two or more Aroclor mixtures.
Weathered samples may be defined as samples containing PCB mixtures no
longer resembling pure Aroclor standards. Weathering of Aroclors occurs
naturally 1n the environment, since each PCB congener possesses different
physical and chemical properties. For example, Henry's Law constants nay vary
by a factor of ten (Brunner, 1990). Association constants may vary by over two
orders of magnitude (Lara, 1989). Solubility may vary over several orders of
magnitude (Dickhut, 1986) depending upon the PCB congener. Overall, the lesser
chlorinated PCBs are generally more soluble and subject to losses through
volatilization and photolysis. The more chlorinated PCBs are generally less
soluble and more likely to accumulate 1n sediments and biota. Even though
natural transformation processes may alter PCB compositions so they no longer
resemble commercial Aroclor mixtures, PCB concentrations in environmental samples
continue to be reported as Aroclor mixtures. For weathered samples, Aroclor
designations are no longer descriptive of the congeners present and their
relative amounts.
The Webb and McCall approach for analyzing environmental samples containing
more than one Aroclor fosters the tendency to report PCB concentrations in all
samples, even weathered samples, as Aroclor concentrations. Peaks with relative
retention times 11 to 70 are compared to peaks in Aroclor 1242; peaks with
relative retention times 84 to 174 are associated with Aroclor 1254; and peaks
with relative retention times longer than 174 are associated with Aroclor 1260.
Some exceptions to this division of peak assignments are presented by Webb and
McCall along with a flow chart for assigning Aroclor standards to be used for
quantitation. The scheme presented for assigning Aroclor standards to quantltate
B.3-51
-------
peaks found 1n specified regions of the chromatogram may lead to an incorrect
assumption of congeners present, if the concentrations are reported as Aroclor
mixture concentrations or 1f GC operating conditions are altered, such as using
a different stationary phase in the column. Webb and McCall used a SE-30
stationary phase and noted that DC-200, OV-17, and OV-101 phases are also
appropriate. QF-1 and OV-225 stationary phases are not appropriate for this
technique, since the PCBs will elute 1n a different order and the peak
assignments made by Webb and McCall will no longer be valid.
B.3.7.2 Discussion of Data Quality Assurance
1976-1978 Sediment Survey
O'Brien and Gere conducted the PCB analyses for the sediment samples from
the 1976-1978 survey. The O'Brien and Gere (1977) report contains a summary of
the quality control (QC) data produced during analyses of sediment samples
collected during the 1977 sediment survey. Method blanks (sodium sulfate and
sand) were incorporated with every batch of samples analyzed. Two different
extraction procedures were employed in the study soxhlet extraction and a two
hour shaker method. Extraction efficiencies were evaluated by performing a
second extraction on the same sample and assuming that all recoverable PCBs were
extracted after two extractions. This approach does not account for PCBs which
may be irreversibly bound to the matrix. Duplicate analyses where samples are
split into two portions and analyzed as two samples in the same batch were
employed to evaluate the reproducibility of results. In addition, an inter-
laboratory study of reference samples was conducted between five different
laboratories to assess the variability between laboratories.
PCB contamination ( > 1 ppm) was observed 1n 5 out of 37 soxhlet extraction
blanks and 0 out of 120 shaker extraction blanks. Sediment samples analyzed in
the same batch as unacceptable blank samples were re-analyzed. Extraction
efficiencies averaged over 80 percent for both methods. Sediment analyses were
not corrected for extraction efficiency, since extraction efficiencies varied as
much as duplicate analyses on the same sample.
B.3-52
-------
The relative percent difference between duplicate analyses of sedlnent
samples extracted using the soxhlet extraction technique averaged 125 percent for
Aroclor 1016, 160 percent for Aroclor 1254 and 174 percent for Aroclor 1221.
These differences are quite large compared to the relative percent differences
observed for the shaker method ~ 105 percent for Aroclor 1016, 85 percent for
Aroclor 1221 and 71 percent for Aroclor 1254. The target, relative percent
difference for the study was 100 percent. After rejecting outliers, the relative
percent difference between duplicates Improved considerably so that only Aroclor
1221 measured with.the soxhlet technique remained outside the target level with
a value of 111 percent (Tofflemire et ah, 1979). The large relative percent
differences averaged over all sediment samples may have been due to the extreae
heterogeneities of the samples. Bopp et a7. (1985) noted that one sediment
sample collected from the Upper Hudson River contained centimeter-sized pieces
of paper containing about 6,000 ppm PCBs.
The overall relative standard deviation (RSD) of PCB concentrations
measured in the inter-laboratory study was 49 percent. During the inter-
laboratory study, it was determined that pure Aroclor mixtures from different
sources contained different amounts of PCB congeners. Since these pure Aroclor
mixtures were used as standards for quantitation, some of the variation between
laboratories may be attributed to the use of Aroclors from different sources.
Unfortunately, there 1s no indication of the magnitude of this effect.
In the 1977 sediment survey, the O'Brien and Gere laboratory reported PCBs
as Aroclor concentrations. Aroclors 1221, 1016 and 1254 were used as standards.
GC peaks were identified by their actual retention times and not retention tines
relative to pp'-DDE. As a result, peak Identities according to the Webb and
HcCall scheme are not certain. O'Brien and Gere note that Aroclor chromatograa
patterns may be altered drastically in environmental samples, making the
Identification of specific Aroclors difficult, If not impossible. Nevertheless,
the laboratory reported PCB levels as Aroclor concentrations.
B.3-53
-------
Aroclor 1221 was not always identified In the samples analyzed by O'Brien
and Gere. When Aroclor 1221 was identified, quantitation was based on a single
peak (most likely Webb and McCall peak 11 and possibly peak 21). Peak. 11
accounts for 32 percent of the PCBs found in Aroclor 1221 and only one percqฃ
of PCBs found in Aroclor 1242. As Aroclor 1221 was rarely identified and peak
11 is difficult to quantitate, little importance should be attached to Aroclor
1221 levels reported in the 1977 sediment survey.
Aroclor 1016 levels were quantitated based on three diffe nt peaks (Webb
and McCall peaks 28, 47, and 58). Aroclor 1254 levels were quantitated based on
eight peaks (possibly Webb and NcCall peaks 70, 84, 98, 104, 125, 146, 178, and
203). For each peak, a separate estimate was derived of the respective Aroclor
mixture and the average for the Aroclor was reported for the sample. Because
some of the peaks used to quantitate Aroclor 1016 and 1254 occur in-other PCB
mixtures, e.g., peaks 47 and 58 occur in both Aroclor 1016 and 1254, an
overestimation error results from the use of these peaks to quantitate any given
Aroclor mixture. If equal amounts of Aroclors 1016 and 1254 were contained in
a sediment sample, then the Aroclor 1016 concentration would be overestimated by
almost 30 percent. The higher the relative amount of Aroclor 1254 in the sample,
the greater the overestimation.
The same type of situation exists for quantitation of Aroclor 1254. If
equal amounts of Aroclor 1242 and 1254 were contained in a sample, estimated
Aroclor 1254 levels would be higher by almost 15 percent and total PCB levels
(Aroclor 1242 plus Aroclor 1254) would be overestimated by approximately 22
percent. Once again, the degree of overestimation would depend upon the relative
levels of Aroclors 1242 and 1254 in the sample.
The degree of overestimation would also likely depend upon the extent of
weathering in the sample. Aroclor levels and total PCB levels, based on the sum
of the individual Aroclors in lightly weathered samples, are most likely to be
overestimated. The weathering process is not well enough known to predict the
probable direction of error in estimating PCBs in heavily weathered samples.
B.3-54
-------
In spite of the large errors Introduced by the above approach, the error
In estimating Aroclor and total PCB levels using the above approach Is small
compared to the large relative percent difference measured between duplicate
analyses obtained by O'Brien and Gere. The large variability 1n duplicate
measurements may be due to the heterogeneity of the sediments.
1984 Sediment Survey
Sediment samples collected 1n the 1984 sediment survey were analyzed by
Versar, Inc. Versar's internal quality assurance program included evaluation of
duplicate analyses and Aroclor 1242 spike recoveries. In addition, blind samples
were submitted to Versar 1n duplicate but 1n different batches. M. P. Brown et
a/. (1988b) provide a summary of Versar duplicate and blind duplicate analyses
as well as Aroclor 1242 spike recoveries.
The average relative percent difference between Versar-submitted duplicate
analyses was 27 percent. This value Is much lower than the values reported fay
O'Brien and Gere for the 1977 sediment survey. The average relative percent
difference between blind duplicate samples was almost twice as large at 52
percent, but was still lower than the percent difference values reported by
O'Brien and Gere.
The mean percent recovery of Aroclor 1242 was 86 percent. M. P. Brown et
a7. suggest that the reason for this low recovery might be the method used for
quantifying Aroclor 1242. Versar originally used the method of Webb and McCall
(1973) to quantify Aroclor 1242 levels, but only used peaks with relative
retention times of 21 to 84. The omission of peaks with relative retention tiaes
of 11, 16, 98, 104, 125, and 146 would result in an underestimation of total PCB
levels by approximately 11 percent for a true Aroclor 1242 mixture. An actual
recovery of 100 percent would result 1n a calculated recovery of only 89 percent.
This value of 89 percent 1s very close to the mean percent recovery of 85
percent. Thus, a recovery reported to be approximately 100 percent would
actually overpredict PCB concentrations of the sample by approximately 11 percent
to 14 percent. The data were not corrected for the percent recovery, so it is
B.3-55
-------
anticipated that the reported levels are fairly close to the true levels.
In the 1984 sediment survey, Versar estimated PCB concentrations as
Aroclors 1254 and 1260 using the method of Webb and McCall. It should be noted
that there was not a strict correspondence between relative retention times
reported by Versar and those of Webb and McCall, because of differences 1n the
Instrument packed columns and GC operating conditions.
Aroclor 1254 quantitation is uncertain for these samples because peaks 98
and 104 were not accurately quantltated (detection frequency of 2.6 and 2.9
percent, because of Interference from the internal standard pp'-DDE which, by
definition, elutes at a relative retention of 100. Peaks 98 and 104 contain
approximately 21 percent of the PCBs found 1n Aroclor 1254. Thus, Aroclor 1254
levels are likely underestimated.
No Aroclor 1221 levels were reported in the 1984 sediment survey and,
indeed, Aroclor 1221 is difficult to quantltate using 6C-EC0. This situation
results from the relatively poor sensitivity of the method to mono- and
dichloroblphenyls, the major constituents of Aroclor 1221. Early eluting peaks
(Webb and McCall peaks 11 and 16) were identified in only two to three percent
of the samples analyzed. Early eluting peaks were also rarely Identified In the
1977 sediment survey.
Versar also reported total PCB levels for the 1984 sediment survey. As the
methods used to calculate total PCB levels were not described, it is impossible
to evaluate the data quality. Based on the reported Aroclor and total PCB
levels, it was clear that total PCB levels were not simply the sum of the
quantified Aroclor mixtures.
Packed column GC peak area data for the 1984 sediment survey are contained
in the TAMS/Grad1ent database. (Access to the NYSDEC data was provided by GE.)
This peak area data allows for an evaluation of the PCB quantitation procedure
with respect to weathered samples. The average normalized peak areas for peaks
47 and 58 in sediment extracts are approximately 35 percent greater than the
B.3-56
-------
average normalized peak areas for the same peaks 1n an Aroclor 1242 standard.
The average normalized peak areas for peak 28 are approximately equal In extracts
and standard. This finding suggests that using these peaks will tend to
overestimate Aroclor 1242 levels to a greater extent 1n a weathered sample
compared to a standard Aroclor 1242 mixture. This factor may account for the 40
percent higher Aroclor 1242 estimates obtained by H. P. Brown et al. (1988b)
using a revised Webb and McCall procedure compared to the Initial estimates
obtained by Versar using the standard Webb and McCall procedure. It 1s not known
to what extent this overestlmatlon may offset the underestimation In Aroclor 1254
levels due to Interference from the Internal standard.
Other Sediment Data
In addition to the two major sediment sampling events just discussed, a
number of other sediment surveys have been conducted. In 1983, the USEPA
collected 66 sediment samples from the Upper Hudson River. These data were
summarized In Volume 1 of the Feasibility Study for the Hudson River PCBs Site
published by NUS (1984). Two different methodologies were employed to quantltate
Aroclor concentrations. If the sample appeared to be characterized by a single
Aroclor, then the sum of the areas of all PC6 peaks was used to calculate the
Aroclor concentration. If more than one Aroclor appeared to be present In the
sample, then the Webb and McCall method was used.
Using the sum of the areas of all PCB peaks to quantltate a single Aroclor
mixture will likely overestimate total PCBs present, 1f the sample 1s enriched
1n more highly chlorinated PCBs. This procedure will underestimate total PCBs
present, if the sample 1s enriched 1n lesser chlorinated PCBs, because of the
relative ECO response factors for lower and higher chlorinated PCBs. As
mentioned earlier, the Webb and McCall procedure provides the best estimate of
total PCBs In weathered samples.
In another study, sixty-five core sections were collected from the Upper
Hudson River, extracted and analyzed for PCBs by Bopp et a7. (1985), at the
Lamont Doherty Geological Observatory. Total PCBs were quantltated using the
B.3-57
-------
method of Webb and McCall (1973). Individual peak concentrations were calculated
for peaks with relative retention times of 28 to 174. Webb and McCall-analyzed
Aroclor 1242 and 1254 standards were used in the analyses to eliminate the
uncertainty in using weight percent data supplied by Webb and McCall with Aroclflr
standards from other sources. In addition, Bopp et al. estimated original
Aroclor 1242 concentrations 1n some samples by comparing a single peak (peak 47)
1n the sample to the same peak found 1n the standard, since Peak 47 was observed
to be relatively stable in all core section extracts. v
Total PCB estimates using the Webb and McCall procedure should be
reasonably accurate. However, only peaks 28 to 174 were quantltated. Assuming
that Aroclors 1242 and 1254 are found in the sediments, 15 percent of Aroclor
1242 may be omitted by neglecting peaks 11 to 21 and 3 percent of Aroclor 1254
may be omitted by neglecting to quantitate peaks 203 and 232. Observation of a
major shift to higher relative amounts of lesser chlorinated PCBs in subsurface
samples may increase the error In estimating total PCBs, since lower chlorinated
PCBs eluting in peaks 11, 16 and 21 were not quantltated.
Hudson River F1sh Samples
PCB concentrations have been measured 1n numerous species of fish collected
throughout the Hudson River. Hazleton Laboratories America Inc. performed all
the fish analyses since 1977 in an attempt to maintain some consistency in the
analyses. PCB sample results of the project have been periodically published in
numerous reports (Sloan et <7., 1983, 1984, 1986; Armstrong and Sloan, 1980a,
1980b) and are also in the TAMS/Gradient database.
Quality control data generated as part of the Hudson River Fish Monitoring
Project consisted of: 1) analysis of periodic performance samples to assess
accuracy; 2) analysis of one blank sample for every 20 fish samples; 3) analysis
of one duplicate sample for every set of 20 fish samples; and 4) evaluation of
the spike recovery of three Aroclors (1221, 1016, and 1254) In one sample for
every 20 fish samples reported.
B.3-58
-------
Summary quality control data generated between the years 1978 - 1981 of the
Hudson River Fish Monitoring Project are presented 1n an appendix to a report
prepared by Sloan and Armstrong (1980). All blank analyses were found to contain
no PCBs. The average relative percent difference In Aroclor concentrations
reported for duplicate analyses was small and varied from approximately 10 to 20
percent. The average recovery of Aroclors 1221, 1016 and 1254 spiked in fish
flesh varied from 65 to 98 percent. In general, spike recoveries were highest
for Aroclor 1254 and lowest for Aroclor 1221.
Aroclor mixtures 1n fish samples have been reported as concentrations of
Aroclors 1221, 1016 and 1254. Quantitation was done by comparing the areas of
one or several peaks to those produced by the respective Aroclor. Aroclor 1221
levels were calculated based on a single peak (Webb and McCall peak 11). Aroclor
1016 levels were quantitated based on two peaks (Webb and McCall peaks 37 and
40). Interestingly, Aroclor 1242 standards were used to quantitate Aroclor
"1016" levels. Early in the project, it was determined that Aroclor 1242
patterns obtained from the fish extract were indistinguishable from that produced
by Aroclor 1016 (Armstrong and Sloan, 1980). As a result, all Aroclor 1016/1242
levels have been reported as Aroclor 1016. Aroclor 1254 levels were quantitated
based on three peaks (Webb and McCall peaks 125, 146 and 174).
Estimation errors of PCBs in the fish monitoring program can be evaluated
in a fashion analogous to the evaluation of errors in the 1977 sediment survey.
Peaks 37 and 40 were used to quantify Aroclor 1242 and are not found in Aroclor
1254. Peaks 125, 146 and 174, used to quantify Aroclor 1254, contain approxi-
mately 34 percent of the PCBs contained in Aroclor 1254 and only 2.6 percent of
the PCBs found 1n Aroclor 1242. Peak 11 was used to quantify Aroclor 1221 in the
fish samples. It represents 32 percent of the PCBs in Aroclor 1221 and only one
percent of those found 1n Aroclor 1242. This selection of peaks for Aroclor
quantitation appears to be a much better selection of peaks than those used 1n
the 1977 and 1984 sediment surveys, because there is little overlap in Aroclors
1221, 1242 and 1254. Thus, on this basis, the reported levels 1n fish are
anticipated to be fairly accurate to the extent that the chromatographic elutlon
profile found in the fish actually resembled the Aroclor mixture quantified. If,
B.3-59
-------
however, the chromatographic elution profiles resemble a heavily altered sample,
there 1s little means of assessing the adequacy of the data.
US6S Hater Column Data
The US6S began regular monitoring of PCB concentrations water samples
collected from the Hudson River at Uaterford In 1975. In 1977, water samples
collected at Rogers Island near Fort Edward, Schuylervllle and Stillwater were
also Included In the monitoring program. All samples were shipped chilled on 1ce
to the USGS National Water Quality Laboratory 1n Doravllle, 6a., where all PCB
analyses were done.
Analysis for total recoverable PCB concentration was performed on
unflltered samples. Results, therefore, Include the dissolved as well as the
suspended fraction. Dissolved PCB concentrations were determined on samples
filtered through a 0.45 pm silver oxide filter. Dissolved concentrations were
determined on only a small percentage (<5 percent) of the samples.
Elapsed time from sample collection to laboratory analysis was reported as
two to six weeks 1n 1984. Analysis of split samples retained 1n Albany for up
to seven months Indicated that storage time had no discernible effect on
concentration. It 1s further contended that allowance of several days' contact
time between solvent and water directly 1n the sample bottle 1s desirable because
of PCB's affinity for particulates (Schroeder and Barnes, 1983).
Total concentrations were calculated by dividing the area of a sample's
Identified PCB peaks by the area of all peaks for an Aroclor standard, then
multiplying this ratio by the concentration of the Aroclor standard. Require-
ments for a calculation were that at least 60 percent of the peaks 1n a standard
be present 1n a sample and both relative peak ratios and column detention time
must match. In a few samples, chromatographic peaks resembled a mixture of two
Aroclors. In such cases, calculations were based on a standard containing two
Aroclors. The data are generally reported as concentration of total PCBs,
supplemented by notes as to presence or absence of specific Aroclors, but with
B.3-60
-------
quantitation to specific Aroclors not given.
The uncertainty introduced by the use of the entire sample peak area to
calculate the total PCB concentration in the USGS samples will vary to the degree
that the elutlon profile does not match the standard profiles. It is not clear
how great an error a 60 percent match will Introduce, but 1t is clear that the
error will Increase the more the sample pattern 1s shifted to the lower
chlorinated congeners peaks relative to the standard.
Concentrations were reported uncorrected for Incomplete extraction.
Schroeder and Barnes, however, contend that extraction efficiency 1s high (>80
percent) for Hudson River water, because the river 1s relatively low in suspended
sediment and dissolved organic carbon concentrations. Extraction efficiency ซay,
however, be an issue for periods of high suspended sediment.
B.3.7.3 Summary
PCB concentrations in Hudson River sediment, water and fish samples have
been reported as Aroclors, despite the fact that heavily weathered samples may
bear little resemblance to original Aroclor mixtures. A number of methods have
been devised to quantltate PCBs as Aroclors and considerable variation exists
between the methods. Most methods tend to overpredict total PCBs present In the
sample.
Sediment Survey (1977-1978)
PCB extraction efficiencies were on the order of 80 percent which
leads to an underpredlctlon of total PCBs.
Aroclor 1242 levels were likely overestimated by basing quantitation
on peaks containing congeners also found In Aroclor 1254; Aroclor
1254 levels were likely overestimated by basing quantitation on
peaks containing congeners also found in Aroclor 1242.
Lower chlorinated PCBs (Aroclor 1221) were rarely detected.
Higher chlorinated PCBs (Aroclor 1260) were not quantltated.
Summing aroclors may have led to an overpredict ion of total PCB con-
B-3-61
-------
centrations.
Sediment Survey(1984)
Yป
It is difficult to predict whether Aroclor concentrations presented
in the 1984 sediment survey overpredicted or underpredicted total
PCB levels.
Mean Aroclor 1242 extraction efficiency was 86 percent, possibly
leading to an underprediction of Aroclor 1242.
Lower chlorinated PCBs (Aroclor 1221) were not quar'stated.
Aroclor 1254 levels were likely underestimated, because of Interfer-
ence from the internal standard pp'-DDE.
Total PCB levels were quantified, but the method of quantification
could not be evaluated.
Fish Survey (1977-1988)
The Aroclor measurements and quantitation were performed by one
laboratory, giving what should be a consistent set of results.
Aroclor results appear to be reliable.
USGS Water Column Data (1975-1989)
Total PCBs were reported (not Aroclor mixtures), thus avoiding the
Inherent ambiguities and uncertainties of modelling an environmen-
tally sampled PCB with an Aroclor mixture. (USGS reported that most
samples were in the Aroclor 1232 to 1248 composition range.)
Few (<5 percent) of the water samples were analyzed for dissolved
PCBs.
Data may be subject to uncertainties via quantification based on
total peak area.
Other Data Sources
Other data sources that provide information- on PCB levels 1n
sediment at various points in time exist.
These records do not constitute a continuous monitoring program, but
can provide a cross-reference on other data.
B.3-62
-------
The data examined here are adequate for assessing gross trends. In most
cases direct comparisons of Aroclors reported for different media are problemat-
ic, because of the different procedures used to measure Aroclors in each nediim.
Aroclor concentrations reported for the same medium are generally adequate for
assessing temporal or spatial trends in the data, but different quantitation
methods used by different laboratories complicate direct comparisons. Congener-
specific PCB analyses would alleviate problems in estimating total PCBs and allow
comparisons of data among groups.
B.3-63
-------
PAGE INTENTIONALLY LEFT BLANK
*
B.3-64
-------
SYNOPSIS
DATA SYNTHESIS AND EVALUATION OF TRENDS
(Section B.4)
Detailed interpretation and analyses of the Upper Hudson monitoring data have focused
on the potential for migration and redeposition of PCBs in sediments and evaluation of
statistical relationships among PCB concentrations in sediments, water and fish.
Three questions posed as the objectives of these analyses concern the cycling of PCBs in
sediments and water and their impact on the fish population (B.4.1).
Flood flow and sediment transport (B.4.2) are addressed first. USGS data are used to
analyze flood recurrence intervals and the relationships between flow and sediment load. The
flood frequency analysis suggests that other investigations may have overestimated the magnitude
of the 100-year flood in the Thompson Island Pool by 25 percent This finding implies that the
potential for scour of contaminated sediments may be less than previously estimated Analyses
also suggest that a decline in suspended sediment load has occurred over time
An investigation of the relationship between PCB concentrations and flow and estimation
of mass loading from the Upper to Lower Hudson is presented. PCBs in the water column and
mass discharge (B.4.3) are difficult to evaluate, because relatively few samples are taken at a
station in a typical year, whereas PCB concentrations may change rapidly with changes in river
flow. PCB concentrations and trends must, therefore, be inferred from an incomplete time series
of measurements. PCB concentrations in the Upper Hudson have shown a bimodal relationship
to flow, increasing at both high and low river flows. Separate multiple regression models, fit for
high and low flow regimes at each station, do not yield great predictive strength. A negative
correlation between PCB concentration and year is found at all stations, indicating a gradual
decline of PCB concentrations in the water column over time Estimates of PCB mass loading
from the Upper to the Lower Hudson are evaluated PCB measurements are biased toward high
flow events. Mass load, not measured directly, must be estimated statistically. To correct the
sampling bias, this analysis adopted a new method of estimating the load. In the period of
1984-1989, little increase in total load between Fort Edward and Waterford appears to have
occurred This finding led to the unexpected conclusion that much of the load in recent years
appears to come from locations upstream of the Thompson Island Pool Use of the new
analysis also indicates that of PCB loading from the Upper to the Lower Hudson may have
been overestimated previously.
PCB levels in fish have declined over the last ten years, exhibiting approximate
exponential decline patterns, with a leveling out or stabilization in recent years (B.4.4). For the
11-year sampling record at River Mile 175, levels of a less chlorinated mixture of PCBs (Aroclor
1016) in fish exhibit an apparent half-life of 3 to 4 years. The rate of decline of the higher
chlorinated congeners (Aroclor 1254) appears to be much slower, with an apparent half-life of
7 to 40 years. Time-trend regression equations are used to obtain an approximate estimate of
-------
total PCB levels in fish over the next 30 years, assuming the current declining trend continues.
PCB mass transport and PCB levels in fish in the Upper Hudson both exhibit generally
declining trends over time. Despite the large number of sediment samples that have titen
analyzed, shifting sediments, widely disparate sampling densities and uncertainties in PCB
measurement methods all confound the interpretation of the sediment sample results. Available
data are insufficient to relate PCB concentrations in fish to PCBs in sediments. In order to
understand better the exchange of PCBs between sediment, water and fish, detailed PCB
analyses related to specific forms of PCBs (congeners) will be necessary. Among the questions
still to be answered are whether the PCB levels will continue their observed. decline and what
specific conditions would alter their decline.
-------
B.4 Data Synthesis and Evaluation of Trends
B.4.1 Phase 1 Objectives
In general terms, a benefit/risk analysis of existing conditions and
possible remedial actions In the Upper Hudson depends on an understanding of how
PCBs transfer, accumulate and dissipate 1n sediments, water, fish and other
media. Perhaps the most visible net effect of the exchange of PCBs in this
system has been PCB uptake 1n fish. Use of the PCB levels 1n fish as a barometer
of the overall status of PCBs 1n the Upper Hudson 1s, therefore, considered
appropriate and may ultimately provide a means to evaluate benefits and risks of
remedial measures versus current trends. If the fish data are to be utilized as
a future barometer to evaluate the effects of remedial actions, then the
following questions need to be answered:
4
(1) What is the potential for migration and redepositlon of PCBs in
sediments?
(2) How are PCBs in sediments transferred to the water-column?
(3) What is the effect of (1) and (2) on biฉaccumulation of PCBs by
fish?
These three areas of investigation, shown schematically in Figure B.4-1,
provide an initial framework within which to reassess the PCBs in the Hudson
River and focus attention on those questions most important to the reassessment.
To meet this end, initial results of scour and flow relationships and correla-
tions of PCBs in water and fish are examined.
B.4.2 Flood Flow and Sediment Transport
B.4.2.1 Flood Frequency Analysis
This section evaluates USGS flow data to derive flood recurrence
probabilities for the Fort Edward and the Thompson Island Pool areas. This
B.4-1
-------
evaluation suggests that earlier investigations of flood flows significantly
overestimated the magnitude of the 100-year flood for the upper river.
Analysis of both average flow and flood recurrence intervals is important
for assessing the probability of future high flow events, which could scour PCB-
contaminated river sediments. Figure B.4-2 shows the flow-duration curves
generated from 13 years of daily average flow observations at Fort Edward. The
curves show the exceedance probability for any specified flow. For instance, at
Fort Edward, only 1 percent of observed daily average flows exceeded 20,800 cfs,
while 50 percent of the flows exceeded 4,060 cfs. The period of record at Fort
Edward is relatively short (13 years), which could lead to inaccurate estimates
for extreme events. A 69-year record upstream at the Hudson River below the
confluence with the Sacandaga River, reported by USGS from flows at Hadley plus
flows ai Stewarts Bridge in the Sacandaga, is also shown in Figure B.4.2. This
record provides better estimates for extreme events.
To estimate a flood recurrence interval it is necessary to form a partial
duration series of the flood events. Series were formed for both daily average
and peak value flows. The partial duration series of interest here is the annual
maxima series (largest flood event in each year of record). This evaluation
provides an estimate of the annual flood recurrence interval or return period.
Using observed annual peak flows, this recurrence interval can be calculated as
(N+l)/M, where M is the rank order of the observed flood (e.g., for the largest
flood M-l) and N is the total number of years of observation. The time series
available at Fort Edward is not sufficiently long to provide a reliable estimate
of the extremes of the distribution. The occurrence of floods can be estimated
from the longer record (the upstream station of the Hudson River at Hadley to
which discharges from the Sacandaga River have been added) and the floods can be
translated to the Thompson Island Pool and other areas downstream (Figure B.4-
3a).
B.4-2
-------
Before proceeding with the flood frequency analysis for the Thompson Island
Pool, two caveats must be stated:
(1) The longer (69-year) monitoring period for Hadley is inadequate to
estimate extreme floods with recurrence intervals exceeding the
length of the record {e.g., >70 years). A Log Pearson Type III
extreme value distribution provides a means to estimate floods of
longer recurrence intervals (USGS, 1982). Figure B.4-3a plots the
empirical flood recurrence values for the 69-year record, together
with estimates of larger (100-year) floods at Hadley.
(2) A plot of the annual maxima by year (Figure B.4-3b) reveals that the
three highest flows on record all occurred in the 1920s. This
finding suggests that the construction of the Sacandaga dam in 1930
significantly altered the flow regime and reduced the magnitude of
floods.
Flood frequencies calculated using a Log Pearson Type III distribution
(USGS, 1982) on the post-1930 data only yield estimates of daily floods as shown
below for the Hudson River below Sacandaga River.
Estimated (Log Pearson Type III) Daily Flood Events,
Hudson below Sacandaga (Post 1930)
Recurrence Period
T (yrs)
1931-1989 Flood Flow
Qt (cfs)
5
27,964
10
32,027
25
36,876
50
40,314
100
43,621
During the 14 years following removal of the Fort Edward Dam, the maxinum
daily flood observed in the Upper Hudson occurred in April 1976, with a daily
average flow of 39,340 cfs reported below Sacandaga. The above analysis implies
that this flow is less than a 50-year flood. Frequency analysis with a log
normal distribution (Bras, 1990) yields a similar result, with the 50-year flood
estimated at 41,144 cfs for a daily average flow.
B.4-3
-------
Peak flows (as opposed to maximum dally flows) are of most Interest for the
assessment of flood damage and the determination of maximum erosional shear
stresses leading to sediment scour. Peak flows were not regularly monitored at
Fort Edward until 1976. USGS does not measure peak flows for the Hudson
immediately below the confluence with Sacandaga River (see Plate B.l-1). Because
the Sacandaga is controlled, peaks on the two rivers are unlikely to coincide.
Therefore, a method had to be developed to estimate peak flow at Fort Edward from
peak flow data for the Hudson River at Hadley for the period prior to 1976. The
100-year peak flow flood at Fort Edward has been estimated at 41, ^-3 cfs (Malcolm
Pimie, 1975). This estimate was obtained using a direct translation of the 100-
year flood at Hadley downstream, without proration for drainage area and further
assuming that the flow 1n the Sacandaga would be zero during flooding 1n the
Hudson. Examination of flow records 1n the Sacandaga shows that the zero flow
assumption is not always valid. The method of the Federal Emergency!Management
Administration (FEMA, 1982 and 1984) for estimating flood recurrence at Fort
Edward based on peaks at Hadley was to assume that the Sacandaga reservoir would
contribute 8,000 cfs of flow.to the Hudson River during extreme flood events.
This discharge results from the opening of one control valve, a procedure often
followed during major storms to prevent topping of the dam. With this
assumption, FEMA used a Log Pearson III distribution, fit to the period of record
for the Hudson at Hadley, and accounted for the Increased drainage area between
Hadley and Fort Edward to predict flood flows at Fort Edward. Assuming a
constant 8,000 cfs contribution from the Sacandaga is likely to overestimate the
magnitude of floods, 1s a conservative assumption for a flood Insurance study.
Examination of the record shows that dally average flow In the Sacandaga
at Stewarts Bridge was often near zero on the days of peak flow at Hadley; It
exceeded 8,000 cfs only twice between 1930 and 1976. For the present study, a
modified approach was undertaken. Peak flows at Fort Edward prior to the period
of record were estimated from peak flows at Hadley plus measured daily flows in
the Sacandaga at Stewarts Bridge, prorated downstream according to the following
equation:
B.4-4
-------
0ซr - 10(H) + 0MiS)l *
where
-------
Estimation of the magnitude and recurrence interval of future expected
floods is of significance in assessing the future potential erodibility of the
remnant and hot spot deposits. Examining the flood recurrence interval estimates
in Table B.4-1 suggests that the magnitude of expected floods may have been
overestimated by past researchers. For instance, this TAMS/Gradient estimate of
the 100-year recurrence flood expected at Fort Edward is approximately 8,000 cfs
less than that estimated by FEMA (1984).
The largest peak flow experienced at Fort Edward since the removal of the
dam was that of April 1-2, 1976. Although a gauge was not in operation, it was
estimated by FEMA (1984) that the peak flows in 1976 were greater than 50,000 cfs
and approximated a 100-year flood. This analysis may not be accurate.
Translating the combined flows in the Hudson River at Hadley and in the Sacandaga
River at Stewarts Bridge downstream suggests that the flood peak at Fort Edward
in 1976 should have been in the neighborhood of 45,000 cfs. This would be
approximately a 100-year flood by TAMS/Gradient estimates, but less than a 50-
yซar flood by FEMA estimates.
There is considerable doubt concerning the estimated peak flow magnitude
of floods prior to the institution of actual gauging. The estimated daily
average flow at Fort Edward during the flood event of April 2, 1976 (about 40,750
cfs) was also less than a 100-year daily flood. The next downstream station
operating at this time was at Green Island (below the Mohawk River), where an
annual peak flow of 106,000 was reported on April 2, 1976. On this date, flow
in the Mohawk at Cohoes was 52,100 cfs, leaving only 53,900 cfs attributable to
the upstream Hudson. This flow would correspond to a flow at Fort Edward in the
neighborhood of 39,000 cfs. The data suggest that the magnitude of the 1976 peak
and daily flood events may have been overestimated. Possibly, an extreme high
stage observed at Fort Edward in 1976 did not correspond to the estimated flow,
which would be derived from the stage-discharge curve developed in 1977 by the
USGS.
B.4-6
-------
The magnitude of major flood events has considerable significance for the
analysis of erodibility of contaminated sediments 1n the Thompson Island Pool.
An attempt was made to assess erodibility in this area in 1985 (Zimmie, 1985),
using the HEC-6 computer model. In Zimmie's study, the 10-year and 100-year peak
flood events for the Thompson Island Pool were assumed to be 46,600 and 63,700
cfs, respectively, citing the Fort Edward flood insurance studies (FEMA, 1980,
1984). Actually, these figures are the peak discharge estimates proposed by FEMA
(1980, 1982) for the downstream corporate limits of the Town of Fort Edward;
these limits are located below Fort Miller, two dams downstream of the Thompson
Island Pool. FEMA (1982) estimates for the 10 and 100-year peak discharges
within the Thompson Island Pool, upstream of the confluence with Moses Kill, are
41,900 and 56,800 cfs, about 10 percent less in magnitude. The FEMA calculations
were based on estimates of 10 and 100-year peak discharges at the Village of Fort
Edward of 38,800 and 52,400 cfs. The current analysis, with the benefit of a
longer record, suggests that these values are overestimates.
Using the TAMS/Grad1ent peak flood calculations at Fort Edward, peak
discharges for the 10 and 100-year events In the Thompson Island Pool, upstream
of the confluence with Moses Kill, would be about 37,000 and 49,000 cfs
respectively. Discharges modeled by Zimmie may have overestimated the 100-year
peak discharge in the Thompson Island Pool by about 14,000 cfs and, indeed, were
in excess of the expected 500-year peak discharge, using values computed for this
study and presented in Table B.4-1.
B.4.2.2 Suspended Sediment Discharge
Understanding the movement of sediments in the Upper Hudson Is essential
to evaluating the movement of PCBs adsorbed to the sediments. The natural rate
of sediment transport 1n the Upper Hudson River 1s relatively low compared to
many other eastern North American rivers of similar size, because of comparative-
ly low rates of eroslonal sediment Input, combined with extensive sediment
trapping behind the many dams on the river. Nevertheless, a period of extensive,
PCB-contaminated sediment scour and transport occurred after removal of the Fort
Edward Dam in 1973.
B.4-7
-------
Important early data on sediment transport 1n the Upper Hudson was
documented in the engineering studies investigating conditions following the
removal of the Fort Edward Dam (Malcolm Pirnie, 1975, 1976). This dam, completed
in September 1822, was reconstructed in 1898 as a rock-filled timber cr^,
approximately 586 feet long, with an average height of approximately 19 feet.
The dam created a pool about two and one-half miles long and about 400-800 feet
wide. Extensive sediment deposits, largely derived from lumber industry debris,
accumulated in the pool behind the dam.
Because of severe deterioration of the structure, the Federal Power
Commission granted permission to remove the dam and it was removed in July-
October 1973. With removal of the dam, the impounded pool disappeared and the
river eroded a channel into the entrapped sediments, leaving large sediment
deposits or remnants exposed along the river banks. Now called remnant deposits,
these are shown in Plate 1.2-3. Spring floods of 1974 mobilized large volumes
of sediment and debris from the former dam pool; by the beginning of the
navigation season in 1974, the channel was blocked by debris to about one quarter
mile below Lock 7. It is estimated that 850,000 cubic yards of debris were
scoured from the former dam pool between July 1973 and July 1974, but that
775,000 cubic yards of this total were deposited between the dam site and Lock
7, from which they were dredged in 1974-75 (Malcolm Pirnie, 1976). It was later
determined that these sediments contained large amounts of PCBs.
The sediments behind the former Fort Edward Dam were not typical natural
sediments, but contained a significant proportion of sawdust particles and other
wood debris, derived from historical Adirondack timber operations. This organic
material was characterized by a low specific gravity and a low settling rate,
which had impact on subsequent transport. Borings behind the dam in 1970 (Dames
and Moore, quoted in Malcolm Pirnie, 1975) indicated that the bottom of the pool
was covered by a weak, compressible organic silt and paper waste from 2 to 14
feet deep. Another 1970 study, designed to assess metal contamination behind the
dam, concluded that the sediments behind the dam consisted of a "transient
deposit" of brown fibrous sludge, covering a black sandy silt and stated: "No
B.4-8
-------
appreciable amount of organic silt was encountered" (Clarkeson & Clough, quoted
in Malcolm Pirnie, 1975).
Major spring floods (peak of approximately 100-year recurrence interval)
occurred between March 31 and April 8, 1976. Malcolm Pirnie estimated that an
additional 36,000 tons of sediment were scoured from the remnant deposits and
transported into the Thompson Island Pool by this flood, while about 45,000 tons
of sediment were transported out of the lower end of the Thompson Island Pool
(Malcolm Pirnie, 1976). The net increase in sediment load leaving the pool
likely included substantial scour of sediments deposited in 1974.
The 10-year recurrence daily average flood at Fort Edward is estimated to
be about 31,200 cfs (see Table B.4-1). Daily floods in excess of this magnitude
have been experienced only twice since 1976, on April 29, 1979 (31,700 cfs) and
on May 2, 1983 (32,600 cfs).
Information on time trends in suspended sediment data, as well as
discharge, are provided by USGS monitoring stations (see B.3). Suspended
sediment measurements did not commence until 1975 nor is monitoring continuous
or on a set schedule. Measurements have focused on spring flood periods, with
little data available for the winter months. Lack of a more extensive data base
and of a regular time series creates difficulties in analyzing sediment data and
other water quality parameters.
Use of the monitoring records to establish relationships between suspended
sediment concentration and discharge assumes that conditions have not changed
over time. Figures B.4-4 through B.4-7 show plots of suspended sediaent
concentration (mg/1) versus discharge. The rating curves show a steady increase
in sediment concentration moving downstream from Fort Edward to Waterford. The
relationship appears to exhibit breakpoints or sills, previously noted by Ziarfe
(1985), which occur when suspended sediment concentrations remain at a rather
steady, low level until a critical (threshold), flow velocity is reached.
Thereafter, concentrations increase as a function of discharge. Such behavior
is thought to represent an approximate critical shear stress for sediments in the
B.4-9
-------
river. At Fort Edward, the sill appears to extend to about 10,000-12,000 cfs
(283-340 m3/sec). At Schuylerville, the relationship is not as clear, perhaps,
because the station is just below the confluence with Batten Kill. Breakpoints
in the sediment response also appear further downstream at Stillwater and
Waterford, with an apparent increase to a range around 19,000 cfs by Uaterford.
The destabilization of the channel following the removal of the Fort Edward
Dam in 1973 suggests that a decline in average suspended sediment levels was to
have been expected as the river gradually recovered to a more equilibrium
condition and the remnant remediation was completed. Time trends of observed
sediment load at Fort Edward and Schuylerville, shown in Figures B.4-8 and B.4-9,
suggest a decline over time, particularly at Schuylerville. Although high
sediment loads typically occur during spring flood periods, greater sediment load
is shown for the moderate floods of 1981-1982 than for the major floods of 1979
and 1983. It could be that a limited sampling schedule missed the sediment
transport peak during the major flood years. Between 1984 and 1989, daily
average flows greater than 28,000 cfs occurred only in spring 1987 and a clear
sediment load peak is evident in response to this event.
Empirical Trend Analysis
Time trends in suspended sediment concentration, corrected for discharge,
can be examined through multiple regression relating total suspended sediment
concentration (TSS) to discharge and year to better understand sediment transport
relationships. A log transformation of concentration is necessary to stabilize
the residual variance. Hodels can then be fit in the following form:
LN {TSS *1) - a + (PixQ) ซ (P2 * Yz)
where TSS is the sediment concentration (mg/7), Q is the measured or estimated
Instantaneous discharge (cfs) and Yr is the years since 1900. (Because a log
transformation is used for TSS, a "1" is added to TSS, to handle zeros in the
B.4-10
-------
record.) Very similar models can be fit for Fort Edward, Schuylerville and
Stillwater.
Fort Edward
LN I TSS ~ 1) - 2.87 ~ 9.5XlO'a X 0 - 0.0223 * Yz
R2 - 70 percent
Schuylerville
LN (TSS ~ 1) - 4.61 ~ 8.6X10*5 x 0 - 0.0386 * Yz
R2 - 63 percent
Stillwater
LN (TSS + 1) - 3.80 + 9.9x20"s x Q - 0.0291 x Yz
R2 - 68 percent
For a constant discharge, the relationships imply that sediment concentra-
tions show an exponential decline against year. All the regression coefficients
In these equations are significant at the 95 percent level. The average slope
against year is -0.03, which indicates sediment load corrected for variability
in flow has declined since 1977 with an apparent half-life of 23 years. Such a
decline may to some degree represent a depletion of readily erodible sediments
from behind the former Fort Edward Dam. The decline may also be a consequence
of somewhat lower than average flows in the mid 1980s or a combination of these
and other factors.
B.4.3 PCBs in the Hater Column and Hass Discharge
This section presents extended statistical analyses of water column
monitoring data, with particular emphasis on exploring relationships to aid in
understanding and predicting the concentrations and mass transport of PCBs In the
water column.
B.4-11
-------
B.4.3.1 PCB-D1scharge Relationships
Regression Analyses
&
As was the case with sediment data, PCB measurements at USGS monitoring
stations are gathered at Irregular intervals 1n time, so that standard time
series statistical tests cannot be applied.
The time series of PCB concentrations In the water col1 -t show a high
degree of similarity among Schuylervllle, Stillwater and Waterford. The
coincidence of peaks among these records would naturally seem to be controlled
by hydrologlc events, particularly flood/scour episodes. This relationship is
shown in Figures B.4-10 and B.4-11, which plot daily average flows at.Fort Edward
together with the total PCB measurements at Fort Edward and at Schuylej*ville, the
nearest long-established station downstream of the Thompson Island Pool. The
time series show a clear response in PCB levels to the major flood event of 1979
as well as to the smaller flood events in 1981. The extreme flood event of 1983
seems, however, to have produced only a modest response at Schuylervllle'in PCB
concentrations; subsequent floods also seem to have produced less of a response
than 1s found in the 1979-1980 period. One explanation is that the amount of
PCB-contaminated sediment available for scour and suspension into the water
column has been steadily reduced. In particular, the 1979 flood may have removed
much of the easily erodlble, contaminated sediment in the remnant deposits and
Thompson Island Pool, while burying other contaminated sediments.
There are some other functional differences between the periods before and
after 1979. For instance, between 1972 and 1979 there was channel maintenance
dredging yearly in the reaches between Schuylerville and Fort Edward, whereas no
dredging has occurred there since 1980. Dredging around Lock 7 and Fort Edward
occurred 1n 1974, 1976, 1977, 1978 and 1979 (Ryan, 1991). Maximum volumes were
removed from this area 1n 1974 (351,000 cubic yards) and 1979 (66,930 cubic
yards), following the floods that redistributed the remnant deposits.
Destab111zation of the channel margins following dredging may have excaberated
the erosion of contaminated sediments. Furthermore, 1974-75 dredge spoils placed
B.4-12
-------
on Rogers Island at Fort Edward and may have provided a source of contaminated
washoff into the river.
Variations in PC6 concentrations in the water column are related to
variations in river discharge, but the relationship is complex. Various authors
(Turk and Troutman, 1981; Schroeder and Barnes, 1983; NUS, 1984; Bopp et ซ/.,
1985) have proposed that total PCB concentrations in the water column may have
a blmodal relationship to flow, with measured concentrations increasing at both
high and low flow extremes. The high flow peak has been attributed to Increased
scour of contaminated sediments, whereas the low flow peak has been attributed
to decreased dilution. Increasing PCB concentrations at low flows presume a
relatively constant rate of loading of PCBs, either by desorption from the
sediments or in base flow, accompanied by little or no erosional input, yielding
concentrations Inversely related to flow.
Measurements of both dissolved and adsorbed PCBs in the water column,
involving filtration of raw river water samples and subsequent analysis of the
two fractions (Bopp, 1979); Tofflemire, 1980; Turk and Troutman, 1981; Bopp et
a?., 1985) indicate that PCBs are often detectable in the water column in both
adsorbed and dissolved form. The dissolved component actually refers to
concentrations that will pass a fine filter and, thus, includes both truly
dissolved constituents and PCBs sorbed to tiny organic particles (Bopp et al.
1985). The dissolved component is thought to predominate at low flows,
exhibiting "surprisingly high" concentrations (NUS, 1984) above what would be
predicted for solution equilibrium and up to 0.5 ppb (Tofflemire, 1980).
At high river flows, the PCBs in the water column are more likely to be
predominantly sorbed to sediment. This occurrence is theorized to represent the
scour of contaminated sediments during high flow events, resulting in transient
mass fluxes of PCBs 1n the water column. At such times the dissolved and
adsorbed PCB fractions may not be in equilibrium with the sediments and water.
6.413
-------
Previous analysis of the PCB concentration-discharge relationship has been
based primarily on observations prior to 1980. Figures B-4-12 through B.4-15
plot PCB concentrations versus daily average flows at Fort Edward, Schuylerville,
Stillwater and Waterford for the full period of record (1976-1989). A visual
inspection of these plots reveals surprisingly little obvious relationship
between flow and concentration, with non-detects occurring at almost any flow
level. This occurrence is particularly true at Fort Edward. At the stations
below Thompson Island, the bimodal relationship of concentration to flow is more
evident, with peaks found in both the low flow and high flow parts of the graph.
While high concentrations are not found at middle level flows, low concentrations
continue to be found at all flow levels, indicating that flow alone is not a
particularly good predictor of PCB concentration.
As shown on these plots, the apparent bimodal relationship to flow is most
apparent in the pre-1985 observations. The more recent PCB concentration data
seem to bear little clear relationship (direct or inverse) to flow alone, without
correcting for other influences. Similarly, the relationship between levels of
suspended solids levels and PCB concentrations in the water is surprisingly weak
(see Figure B.4-16 for the relationship at Stillwater). Nevertheless, it does
appear that the highest PCB concentrations in the water column generally occur
at high flows and accompany high sediment loads.
The relationship between PCB concentrations and flow may be obscured by
joint correlation with other related variables. This possibility was explored
through the use of stepwise multiple regression, in which flow, the inverse of
flow, sediment concentration, year and month in year were considered as possible
regression variables. A transformation to the log of PCB concentration was used
to stabilize the residual variance for high flows. The bimodal hypothesis
suggests that PCB concentrations should begin to rise with flow when the shear
stress at the sediment water Interface reaches a certain critical value,
resulting in the resuspenslon of PCB-contam1nated sediments. This hypothesis
implies that separate models should be fit for observations above and below some
critical flow value, distinguishing scouring versus non-scouring flows. From the
sediment rating curves, the non-scour/scour breakpoints appear to be at about
B.4-14
-------
11,000 cfs at Fort Edward, 12,000 cfs at Schuylerville, 16,000 cfs at Stillwater
and 19,000 cfs at Uaterford. The relative values correspond to increasing
drainage area moving downstream and are in agreement with the conclusion of NUS
(1984) that "the transition from one form of PCB (transport) to the other varies
at flows ranging from 10,000 cfs to 20,000 cfs."
Regression models were tested for both high and low flow regimes. Non-
detected PCB values were set at one half the detection limit for the regressions.
The F-test statistic at a 95 percent significance level was set as the criterion
to retain a variable as "significant" in these regressions. Table B.4-2
summarizes the variables that are significant in these regression equations. As
shown by the relatively low correlation (R2) coefficients in this table (ranging
from 0.25 to 0.70), these regression equations do not have great predictive
strength. The regression coefficients on the variables, however, are significant
at a 95 percent confidence level and, thus, there are strong correlations between
these variables and total PCB concentrations. For all four stations, high-flow
PCB concentrations are positively correlated with either discharge or sediment
concentration and these two variables are strongly correlated with one another.
A negative correlation between PCB concentration and year is also noted. This
relationship may result, in part, from the decrease in the detection limit froa
0.1 to 0.01 |ig/7 at the start of water year 1987. While the peak PCB concentra-
tions 1n the water column of 1979-80 have not been observed since, the models
still show a significant response of concentrations to high flow/erosional
events. For low flows at all stations, there 1s also a decline with time as well
as an inverse relationship to discharge (the dilution effect). The decline with
time may represent a gradual depletion of the more readily soluble congeners in
the upper layers of sediment.
The regression coefficients with time (year) for the Fort Edward station
are somewhat lower than the respective coefficients at the other downstream
stations. This outcome suggests a somewhat less significant decline of PCB
concentrations in water over time at Fort Edward.
B. 415
-------
Non-parametric Tests of Trend for Water Column PCBs
As noted above, the negative correlation detected with time in the
regression equations may be partly accounted for by the change in detect^
limit. Nevertheless, there has been a statistically significant downward trend
in concentrations during the period of monitoring. For instance, if all values
less than 0.1, detect or non-detect, are set to 0.1, the regression still shows
a significant negative correlation between concentration an? year at the
downstream stations. The interpretation of "significant" depenc on assumptions
regarding the distribution of the parameter, i.e., normal, Isgnormal, etc.
Further, the extreme variability of concentrations during flood events may
obscure the trend. For this reason, NUS (1984) advocated assessing trends with
time on low flow concentrations only. They indicated that low flow concentra-
tions at all stations had decreased with time, with the decrease being
statistically significant between 1979 and 1980.
A non-parametric test, which makes no distributional assumptions and is not
overly sensitive to the variability of high concentration events, was undertaken
in order to confirm the trend. The non-parametric test used here is that
advocated for water quality data by Lettenmaier et a7. (1991), which was applied
to the time series of PCB observations at Schuylerville. The presence of
multiple detection limits is treated by: 1) setting all data flagged as non-
detect to the highest detection limit; and 2) setting all measured concentrations
below the highest detection limit to the highest detection limit. The non-
parametric Spearman Rank Correlation test was then applied to examine the
correlation between PCB concentration and discharge. As this test did not show
significance at the five percent level, flow-adjusted concentrations were not
needed for the test.
The time series of yearly averages was then examined using the Mann-Kendall
test (Gilbert, 1987), a non-parametric test of trend in which missing values are
allowed to occur. For the 1977-89 series of annual mean PCB concentrations in
the water column at Schuylerville, the Mann-Kendall test results indicate a
statistically significant decline (>95 percent confidence) in PCB concentration
B.4-16
-------
with time. Similar results hold for the other stations downstream of the
Thompson Island Pool. The only exception to this result is mean annual PCB
concentrations at Rogers Island (1978-1989) for which the trend is not
significant with time at the 95 percent confidence level. In other words, the
robust, non-parametric test confirms that the trend at Fort Edward is, indeed,
toward lower mean levels, but the evidence is not sufficiently strong to conclude
that this is not due to random variation.
B.4.3.2 Mass Transport Estimates
Estimates of PCB mass transport rates are critical to assessing the Impact
of the Upper Hudson contamination on the PCB problem in the estuary and to
evaluating remedial actions in the Upper Hudson. Where such rates can be
determined, they reveal the magnitude of the problem and the relative contribu-
tion of various source areas.
Accurate estimation of PCB mass transport is a difficult problem in light
of the available data. While river flow has been monitored on a continuous
basis, PCB measurements in the water column consist of 10-50 small-volume samples
per year, taken at irregular time intervals. Major portions of the yearly load
may be transported during a few brief flood events. The problem lies in filling
the gaps in the PCB monitoring record in order to estimate the integrated total
mass over an entire year. For this reason, estimates of annual PCB load have
varied widely (summarized in NUS, 1984). Two approaches to estimating annual PCB
load were considered here, each with particular problems. One approach is to
develop a regression analysis on the available data, relating observed PCB
concentrations to flow, thus creating a continuous PCB record. A second approach
is to develop a direct estimate of the yearly load. As will be seen below, the
first approach is inherently biased. A bias 1s also present 1n the monitoring
data, which would tend to an overestlmation of load. Robust methods of
estimating the average can correct for this problem and were developed for this
study. The results are expected to be unbiased, but, for the periods of high
transport, they have wide uncertainty ranges, /.e., large error bars.
B.4-17
-------
Regression Approach
In the first approach, regression equations (previously discussed) would
relate PCB concentration to time, flow and sediment concentration. Sediment
load, which 1s not monitored continuously, cannot be used 1n the regression to
be useful for continuous prediction. Furthermore, PCB load is not' measured
directly; Instead, it must be estimated from the PCB concentration multiplied by
the discharge (Q). Thus, it 1s not appropriate to develop a regression equation
relating PCB load (L) to discharge, because discharge would appear on both sides
of such an equation. Because PCB concentration must also be predicted from flow,
the residuals from such an equation will not meet Independent normality
assumptions, but will be dependent in magnitude on discharge. A substantial
proportion of the yearly total mass transport 1s dependent on a few high flow-
high PCB concentration events, which will fall Into the statistical noise of the
regression/extrapolation procedure. An estimate of yearly load based on the
regression of concentration on flow would tend to underestimate the total load
contributed by high flow events and, thus, underestimate the yearly totals.
Indeed, use of the regression equations derived for concentration to predict
loads from flows results in estimates of total load that are consistently smaller
in magnitude than estimates obtained through a robust estimate of the mean.
Schroeder and Barnes (1983a), building on earlier work of Turk and Troutman
(1981), estimated low flow PCB loads using regression analysis. They divided the
flow regime at each monitoring station into scouring and non-scouring segments,
defined by a 600 m3/sec flow (21,186 cfs) at Uaterford. The high flow (scouring
regime) PCB load was identified as highly variable and dependent on the specific
source areas contributing flows to the event. Therefore, no model was proposed
for this portion of the load, even though 1t may be the dominant contributor to
total mass, and total mass calculations were not made. Schroeder and Barnes did
propose yearly (water year) linear regression models for non-scouring-flow loads,
for which they reported statistically significant relationships of PCB transport
rates to inverse of discharge and relatively high R2 values (50-98 percent).
They adopted a no-intercept model, 1n which PCB concentration (C) during non-
scouring flows is modeled as inversely related to discharge (Q):
B.4-18
-------
where 0 is the regression coefficient. Imposition of a no-intercept model
essentially assumes that concentration declines to zero at infinite flow, via
dilution. This assumption 1s dubious, as the model is explicitly restricted to
no-scour situations and the maximum non-scouring flow does not exhibit infinite
dilution. By fitting the following non-zero Intercept model:
C * a + -fi-
0
the intercept (ซ) 1s found to be more significant than the slope parameter (0*),
and the R2 value for the regression declines to very low levels (typically <10
percent). The implication is that the significance of the relationship is
focused in the constant intercept value. Indeed, what Schroeder and Barnes have
actually done 1s to fit a constant load model to the non-scouring regime,
wherein the load (L) 1s not dependent upon discharge at all:
This model implies that under no-scour conditions a constant load of PCBs
enters the river, perhaps through baseflow and constant desorption/diffusion from
the sediments, so that concentration varies only by dilution. This 1s a simple
and appealing paradigm, but the approach reduces the regression problem to one
equivalent to determining a mean mass loading rate solely as a function of mean
concentration, although treating only a restricted flow range. This method is
also likely to result 1n biased estimates 1f used to calculate cumulative
loading, because the regression approach makes the assumption that concentrations
are normally distributed. Loads, however, will not be normally distributed, but
will be a function of Q, which will be approximately log-normally distributed.
For instance, the distribution of loads under non-scouring flows observed at
Stillwater 1n 1983 is shown in Figure B.4-17. The arithmetic mean of these
observations 1s 1.74 kg/day, whereas the regression approach yields a 0 value of
B.4-19
-------
1.0 kg/day. In essence, j) is estimated as the central point or median of the
distribution, but this estimate is not equivalent to the mean in a skewed
distribution. Therefore, use of the Schroeder and Barnes model to estimate non-
scouring loads will result in an underestimation of total loading over the cour^
of time.
Direct Estimation Approach
The second approach to estimating PCB loading over time is o use the USGS
measurements of PCB concentration (C,) and instantaneous discharge (QJ directly.
Their product yields a series of estimates of load rate, i.e., L4 - C, x Q<.
Because PCB levels were not measured continuously, an estimate of yearly load
provided by summing the Instantaneous loads does not yield an unbiased load
estimate. PCB measurements have not been spread randomly throughout the year,
but Instead are focused in the months of April and August, with the apparent
intent of providing good coverage of the high and low flow regimes. For
instance, at Stillwater, there are 85 PCB observations in April, but only 3 in
January (see Figure B.4-18). Measurements at other stations follow a similar
pattern. The problem, as partially recognized by NUS (1984), is that the
available monitoring data provide a biased estimate of the load and the bias is
toward high flow events. At Schuylerville, the PCB observations are associated
with discharges having a mean of 11,136 cfs and a median of 6,422 cfs, whereas
the expected mean discharge, prorated from the Fort Edward gauge, is about 4,900
cfs. Because the PCB concentration 1s highly variable, the mean is likely to be
influenced by outlying extreme high values, so that it is unlikely that an
accurate estimate can be obtained from a small number of samples. The uneven
distribution of measurements during the year will tend to overestimate the true
average load rate.
Additional problems arise because the PCB concentration samples are
(practically) Instantaneous rather than continuous. Thus, the observed
concentration response to a flood event 1s likely to be dependent on whether the
observation caught the rising or falling Umb of a flood wave. Determination of
the mean 1s also Influenced by the presence of non-detects in the concentration
B.4-20
-------
measurements. Various methods can be used to address estimation 1n the presence
of non-detects in a sample, e.g. Helsel and Cohn (1988). No straightforward
method is available to address the problem of estimating the effect of such non-
detects on the estimation of load, which is the product of the weakly correlated
variables, concentration and discharge.
Most authors who have attempted to estimate the total yearly PCB load
(e.g., Tofflemire, 1980; Brown and Werner, 1985; Barnes, 1987) seem to have used
similar methods, based on the observation that PCB concentrations exhibit a
bimodal distribution, against discharge and begin to increase, on average, beyond
a certain critical flow value. The bias inherent in over-sampling high flow
episodes can then be corrected by taking yearly averages for both scouring and
non-scouring flow regimes, then weighting these by the actual rate of occurrence
of daily average flows above and below the assumed scour limit during the course
of the water year or calendar year.
USGS estimates of PCB mass transport are summarized in Barnes (1987), who
analyzed data through water year 1983. As with earlier authors, the transport
regime was divided into scouring and non-scouring flows. For non-scouring flows,
annual loads were calculated by direct averaging of instantaneous loads
calculated from PCB concentration and water discharge data. Barnes found that
PCB transport rates at Schuylerville, Stillwater and Waterford through 1983, were
approximately equal to one another and greater than those at Fort Edward (Rogers
Island). No confidence limits were attached to the estimates nor was It stated
how non-detects were treated. For PCB transport during high flow, Barnes notes
that PCB yield at Waterford depends on the percent of discharge originating froa
the sub-basin above Fort Edward. He then presents a graph of annual transport
by both scouring and non-scouring flows across Federal Dam. The method of
calculating the total annual scouring loads is not documented nor are the exact
figures given.
Other, earlier efforts to calculate total PCB loading to the estuary are
summarized 1n NUS (1984), including calendar year estimates by Brown and Werner
(1983). None of these estimates appears to include confidence limits or any
B.4-21
-------
detailed statistical analysis of the computation of the scouring flow loads.
Results of Tofflemire (1980), Barnes (1987) and Brown and Werner (1983) are
summarized in Table B.4-3.
The foregoing methods of PCB load calculation remain highly sensitive to
mis-estimation of high flow loads, as a few such events may contribute a major
part to the total calculation. One way around this problem is to adopt a robust
estimation method, based on the fact that the observed loads, for both scouring
and non-scouring regimes, appear to follow lognormal distributions.
The current analysis uses a corrected mean method, which is derived from
the median PCB concentration. The median is much more resistant to bias
introduced by outliers in a distribution than is the mean. In addition, the data
appear to be well described by lognormal distributions. The arithmetic mean for
a lognormal distribution is related to the median through:
- f ฐ2lax
X ซ mK exp I^
where % is the mean, m is the median of the untransformed data and o21nx is the
variance of the natural logs of the data.
Calculation by this formula yields an alternative estimate of the
arithmetic mean, which is relatively insensitive to outliers in the data. The
arithmetic variance can also be estimated from the log-space parameters as:
a2x ซ 3? (exp[oy - l).
To implement this method, the data for each year are first divided into
scouring and non-scouring flow regimes. These regimes are defined by flow values
determined for each monitoring station, rather than by simply determining high
flow days by flows at Waterford. The same critical flow values as were selected
B.4-22
-------
for the regression equations earlier in this section are used, except in the case
of Waterford where the figure proposed by Schroeder and Barnes (1983) is used,
I.e., 21,000 cfs at Waterford. Log-space parameters for each flow regime were
calculated and used to determine adjusted arithmetic parameters by the equations
given above. These were then weighted by the actual number of scouring and non-
scouring flow days observed in a given year to obtain the corrected nean. The
variances were also weighted and pooled. Non-detects among the concentration
observations are included in the load calculations as one-half the concentration
detection limit times the flow.
Annual PCB load estimates using the corrected and uncorrected mean methods
are summarized in Table B.4-4. In general, the corrected mean method yields
lower estimates of PCB load than use of average annual PCB concentration
multiplied by average annual flow, particularly in the earlier years. For
instance, in 1983 the uncorrected mean estimates of loads past Fort Edward and
Waterford are 4200 and 3900 kg, respectively, whereas the corrected mean
estimates are 1700 and 980 kg.
Total mass of PCBs transported per year at all monitoring stations, except
for the short run at Fort Miller, is plotted in Figure B.4-19. Figures B.4-20
and B.4-21 provide the 95 percent confidence intervals on the load calculations
for Waterford and Fort Edward, respectively. In general, the error bounds are
quite large for years in which there were a significant number of scouring flows,
due to the high variability of PCB loads in these flows.
A number of interesting inferences can be drawn from the plots of annual
PCB loads. For the early years, through 1979, it is clear that there was a
substantial gain in PCB load over the length of the Thompson Island Pool,
reflected in the differences between loads at Fort Edward and downstream
stations. This regime seems to have been altered by the significant flood of
April 1979 (34,000 cfs at Rogers Island), which may have removed much of the
readily erodible PCB-contaminated sediment in the Thompson Island Pool. For
1980 through 1982 the load gain from Fort Edward downstream is less dramatic,
with annual loads at Stillwater (see Figure B.4-21) about twice those at Fort
B.4-23
-------
Edward (an increase in the range of 300 to 800 kg/yr). The spring flood in 1983
(35,200 cfs) was even greater than that, of 1979 and PCB loads increased sharply
during this year. Loads since 1983 have continued a downward trend, with only
a moderate increase shown for the high flow in 1987. ^
After 1983, there appears to have been little or no gain in annual PCB load
between Fort Edward and downstream stations. This finding suggests that, at
least for the flows experienced in this period, the Thompson Island Pool has not
contributed any significant increase to the PCB load above t ' load already
present upstream at Rogers Island, presumably because most of the easily erodible
contaminated sediments were removed by earlier floods. That period, however, was
one of lower than average spring floods. The only significant spring flood event
from 1984-1989 was that of 1987, which did produce an apparent gain from Fort
Edward to Schuylerville and further downstream. Regardless of whether sediment
scour has been less during this period, it appears that a significant PCB load
is in the river upstream of the hot spot areas (see Figure B.4-19).
For the observations at Waterford, average PCB concentrations are lower,
due to dilution, and flows higher than those at stations upstream. Nevertheless,
the annual PCB load estimated at Waterford very closely tracks that estimated for
Stillwater and Schuylerville (Figure B.4-19). Evidently there is no significant
difference in load between Schuylerville (the first station with a long-term
record downstream of Thompson Island Pool) and Waterford (the last station before
Federal Dam), implying that most PCBs mobilized in the Upper Hudson are
transported through to the Lower Hudson. These observations fit with the
relative annual water-column PCB concentrations (see B.3). The contributing
watershed area at Waterford (4,611 square miles) Is 1.6 times that at Fort Edward
(2,817 square miles) and 1.3 times that at Schuylerville (3,440 square miles).
If the load 1s simply throughput from Fort Edward past Waterford, then concentra-
tions should decline, by dilution, as the Inverse ratio of the contributing area.
That 1s, concentrations at Waterford should be 77 percent of those at
Schuylerville and 63 percent of those at Fort Edward. The estimated current
(1986-1989) annual (full year) average concentrations (Table B.3-12) imply that
concentrations at Waterford are 81 percent of those at Schuylerville and 62
B.4-24
-------
percent of those at Fort Edward, fitting the prediction very closely. It should,
however, be noted that the same relationship is not apparent for summer low flows
(Table B.3-13).
The calculations of total PCB mass passing Uaterford presented here are
somewhat higher than those of either Brown and Werner (1983) or Barnes (1987).
Their estimates are still well within the confidence limits of the present
estimates. It is also possible that the earlier estimates Included sane
consistent underestimation of total mass.
An independent analysis of mass transport rates 1s provided by the detailed
study of PCB transport in the Upper Hudson for Spring-Summer 1983 conducted by
Bopp et a7. (1985), supplemented by collection of data not available from USGS
sampling. Bopp et a7. developed equations to estimate PCB transport, utilizing
USGS monitoring of mean dally discharge and mean suspended matter concentrations,
together with experimentally determined PCB dissolved/adsorbed distribution
coefficients and a measure of PCB component concentration on suspended utter
that was determined from ten large-volume water samples, which varied signifi-
cantly from season to season. The PCB loads estimated are, thus, deterained
Independently of the USGS monitoring of total PCB concentrations. Using this
method, Bopp et al. estimated the total PCB mass transport past Stillwater for
March 1 to September 30, 1983 to be 940 kg, of which 840 kg was transported in
spring runoff, between April 11 and June 10. The TAMS/Gradient mean estimate for
the March to September time period is 830 kg total PCBs transported past the
Stillwater station on a total of 21 scouring flow days (>18,000 cfs). This
result is in close agreement with the figure given by Bopp. The TAMS/Gradient
mean estimate of total transport during this period, 1s 1,400 kg. The
TAMS/Gradient estimate of non-scouring transport is, thus, much higher cotpared
to that of Bopp and yields 570 kg rather than 100 kg. However, the 95 percent
confidence limits on the estimate for this period are also very large (tl.QOO
kg). For the same time period, the mean estimated transport at Schuylenrille Mas
only about 800 kg.
B.4-25
-------
One interpretation of PCB mass load estimates presented here is that the
Thompson Island Pool has not been a major source of PCB mass in recent years (see
Figure B.4-29). The bulk of the PCB mass presently transported in the river is
already present in the water column at Rogers Island (above the Thompson Island
Pool) and is simply throughput downstream without significant loss. This finding
is reflected in the fact that recent PCB concentrations observed in the water
column at Fort Edward are higher than in any station downstream of the Thompson
Island Pool. Presumably the source of the load passing Rogers Island has been
the area beginning adjacent to the Hudson Falls Plant and continuing to Rogers
Island, unless there is some as yet undetected contaminated sediment source in
the Hudson Falls area. This hypothesis should be investigated through continued
monitoring, subsequent to the remediation of the remnant deposits, of PCB
concentrations at Fort Edward, Fort Miller and Schuylerville. Additionally, the
probability of a future significant flood event remobilizing large quantities of
buried contaminated sediments must be evaluated.
B.4.3.3 Discussion of Mass Transport from Upper to Lower River
PCBs In the Upper Hudson enter the Lower Hudson system by passing over the
Federal Dam at Troy. The rate of mass transport over the dam is thus a factor
in the impact of PCBs in the estuarine system. There are also other sources of
PCB contamination in the estuary (see Part A). Assessing the relative importance
of these various sources for current PCB contamination In the estuary presents
a difficult problem.
It has been estimated that 6E capacitor plants discharged between 209,000
and 1.33 million pounds of PCBs to the Hudson River between 1957 and 1975 and an
unknown quantity prior to 1957. Much of that total was sorbed to sediments 1n
the pool behind the former Fort Edward Dam and redistributed following the
removal of that dam in 1973. Large portions of the contaminated sediment were
removed by dredging in 1973-1974, but M. P. Brown et al. (1988b) estimated that
approximately 23,000 kilograms (51,000 pounds) of PCBs remain in the Thompson
Island Pool sediments, whereas the Draft Supplement to the Final EIS (USEPA,
1986) estimated that PCB mass 1n the lower reaches of the Upper Hudson were on
B.4-26
-------
the order of 50,000 kg (110,000 pounds) and that the remnant deposits contained
on the order of 21,000 kg (46,000 pounds).
The PCBs presently stored 1n the sediments of the Upper Hudson, dredge
spoil and containment sites constitute only a fraction of the total releases.
Unknown portions of the total mass have been lost by volatilization to the
atmosphere. Tofflemire and Quinn (1979) estimated volatilization losses of about
600 kg/yr total PCBs from the Upper Hudson. Most of the balance of the mass
released Into, the Upper Hudson has been transported across the Federal Dam and
Into the Lower Hudson estuary and, ultimately, the Atlantic Ocean.
Based on the PCB loads past Waterford, as developed previously, one can
estimate that approximately 15,000 kilograms of PCBs were transported from the
Upper to the Lower Hudson between 1977 and 1989. This estimate is obtained by
summing the annual load obtained by the corrected mean method; it is somewhat
higher than would be calculated using the annual loads determined by Barnes
(1987). The 95 percent confidence limits on the sum are approximately ฑ3,000 kg
(i.e, 12,000 to 18,000 kg total). Of this total, about 13,000 kg are estimated
to have been transported between calendar years 1977 and 1983. There has been
a steady decline 1n the rate of loading, reflecting both the decline 1n water
column PCB concentrations and the generally lower flows of the mid 1980s. Figure
B.4-22 shows the estimated annual loads past Waterford (calendar year basis).
Although there is year to year variability, the loads appear to have declined
exponentially, with a half-life of about three years.
For the period before 1977 there 1s no water column monitoring of PCBs
available. A record of the sediment transport history is, however, preserved in
undisturbed sediment cores. Extensive work on the sediment record of contamina-
tion in the Hudson, particularly the Lower Hudson (see Part A), have been carried
out by the Lamont Doherty Geological Observatory (summarized 1n Bopp and Simpson,
1989).
B.4-27
-------
Study of dated sediment cores from the Lower Hudson above the salt wedge
shows a maximum total PCB concentration In suspended sediment 1n the early to mid
1970s. This occurrence reflects the large volume of contaminated sediment
transported downstream following the removal of the Fort Edward Dam in 1973
several cores the peak concentration appears to be at or very near 1973. A core
collected at River Mile 88.6 (Bopp and Simpson, 1989) exhibits an exponential
decrease in sediment concentration by year since 1973 with a half-life of about
3.5 years, which is very similar to the three-year half-life decline in mass
loads past Waterford. The core data do not directly yield histo^ > water column
FCB concentrations, but do provide information on relative concentration from
year to year. The cores suggest that peak concentrations circa 1973 may have
been about five times those experienced around 1980, i.e. on the order of 5,000
kg/yr. Prior to the peak, the cores show a gradual, consistent increase in PCB
deposition rates from about 1950 on. Core concentrations dated around 1960
appear to be similar to those found circa 1980, suggesting that in this period
the mass flux of total PCBs from the Upper Hudson to the estuary was on the order
of 500 to 1,000 kg/year.
Another analysis of the yearly loading past Waterford is contained in a
modeling study of PCBs In the Hudson River estuary (Thomann et al., 1989). This
study used sediment data to estimate loads through 1975 and USGS monitoring at
Waterford to estimate loads for 1976-87. The results they obtained from analysis
of the USGS data differ from those presented in the previous section and show
significantly higher transport rates than estimated here for the period before
1984. For the year 1980, Thomann et al. estimate average daily loading across
the Federal Dam, based on Waterford monitoring, as 3.8 kg/day, which is 90
percent higher than this report's estimate of 2.0 kg/day. Thomann et al.
estimate about 20.3 kg/day PCB transport 1n 1977, while this report's estimate
is 11.2 kg/day. The primary reason for the difference is that Thomann et al. did
not attempt to correct for the fact that the USGS monitoring 1s intentionally
biased toward higher flow (and possibly higher PCB concentration) events, while
the winter months are under-represented. Thomann et al. (1989) do not state how
non-detects in the data were handled.
B.4-28
-------
The loads estimated by Thomann et a7. are very close to the uncorrected
mean loads presented for Waterford, e.g., 4.1 kg/day for 1980, 21.7 kg/day for
1977. By neglecting the bias 1n the data, Thomann et a7. have likely overesti-
mated the PCB loading from the Upper to the Lower Hudson. Compared to the
TAMS/Gradient estimate that -13,000 kg of PCBs were transported past Waterford
from 1977-83, Thomann et a7. (1989) estimated total loading across the Federal
Dam for the same period to be -19,000 kg, or about 35 percent more.
Thomann et a/. (1989) also used sediment data to attempt to estimate PCB
mass transport prior to 1976. Their estimates are on the order of 1,000 kg/year
around 1960-1965, but then indicate a very sharp increase to a peak of 24,600
kg/yr in 1973. This peak 1s nearly five times that suggested by the Lower Hudson
cores. The calculation was based on analysis of several Upper Hudson sediment
cores reported by Hetling et a7. (1978). Approximate, dated annual PCB
concentrations were established 1n these cores, which were taken to represent the
PCB concentrations in suspended sediment during the same year. The dating seas
to have been based on the assumption of a constant sedimentation rate, which is
probably unrealistic in terms of the historic channel destabilization and
sediment mobilization following the removal of the Fort Edward Dam.
For the analysis of Thomann et al.t annual sediment PCB concentrations from
the cores in river reaches 1-5 (see Plate B.1-2) were prorated to the expected
concentrations at the exit of Reach 1. The dally PCB load was then calculated
as:
WT ซ 5.391 x 10*6 ra? i"1
ฃpi
in which WT is the total PCB discharge (lb/day), Q 1s the annual average flow
over the Federal Dam (cfs), m, 1s the estimated solids concentration in the water
column at Waterford, r2 is the sediment PCB concentration from the core data, and
f 1s the fraction of total PCB associated with the solids (Thomann et *?.,
1989). Estimation via this equation 1s obviously highly sensitive to the value
of f.
B.4-29
-------
Thomann et al. use a value of fplซ0.637, based on a partition coefficient
of 20,500 7/kg. The dissolved portion of the mass was therefore taken to
represent a constant, relatively large fraction of 36 percent. During some
periods of high flood scour, however, the bulk of PCB mass transported is sorbed
to particulate matter, without being in equilibrium with the water column (Bopp
et al.f 1985). In addition, the major portion of the total PCB mass transported
in a given year may be accounted for by scouring transport on a few high flow
days. Thus, use of the equilibrium partition coefficient means that the
dissolved portion and, thereby, the total mass of PCBs transported may be
substantially overestimated. This overestimate would result particularly for the
period Immediately after 1973, when large volumes of sediment from the former
Fort Edward Dam pool were moved by spring floods. The calculations of Thomann
et al. further assume a single suspended sediment concentration value of 85.6
mg/7 (undocumented), whereas USGS observations for 1975-1989 show a median
sediment concentration of 12 mg/1 and an average of 63.3 mg/1 at Waterford.
Taken together these two assumptions could result in overestimation of PCB load
by a factor of around two. Finally, the adjustment factor normalizing Hetling's
(1978) data to the exit of Reach 1 is also highly uncertain.
A more reliable estimate of pre-1976 loading over the Federal Dam Is
provided by examining the radionucllde-dated core samples from below the dam and
comparing these to later monitoring. 4
B.4.4 Analysis of PCBs 1n Fish
Plots of concentration versus time for fish in the Upper Hudson (see B.3)
indicate that PCB levels in all fish species appear to have declined in recent
years. This trend was discussed 1n Sloan et al. (1983), Sloan et al. (1984) and
M. P. Brown et al. (1985), who noted that the levels of all Aroclors In fish In
the Upper Hudson have tended to decline since about 1980. Data from recent years
suggest that this trend alone, one of apparent exponential decay, will not be
sufficient to reduce PCB burdens in many fish species to acceptable levels in a
reasonable period of time.
B.4-30
-------
There have not been any major flood erosion events since 1976; the effect
of such an event In the future on fish PCB levels remains to be assessed.
In monitoring since 1975, peak PCB levels in most species of fish in the
Upper Hudson seem to have occurred in 1977-78. The plots of Aroclor concentra-
tions versus time show that the ratio of Aroclor 1016 to Aroclor 1254 in fish was
also elevated in this period, typically to around five versus an earlier ratio
near one. Presumably this result represents a significant input from burled,
dechlorlnated sediments to the water column during this period or the extensive
use of Aroclor 1016 by General Electric from 1972 to 1976. Since 1980, however,
1016/1254 ratios have been around one and slowly declining in most fish species.
This result appears to represent a gradual depletion of lower congeners from the
aquatic system, resulting from the absence of significant new input from the deep
sediments and/or the mixing of older and newer sediments, resulting in the return
to the dominant, pre-1972 PCB mixture in the riverine system. No similar pattern
was found for Aroclor 1221 relative to 1016 or 1254; levels of this Aroclor
(representing mono and dichloro congeners) were consistently low.
B.4.4.1 Evaluation of Time Trends
Non-Parametr1c Trend Test
Many tests used to examine the significance of trends exhibited 1n tine-
series data, such as the fish data, rely on assumptions regarding the underlying
probability distribution (e.g., normal, lognormal, etc.) of the data. Thus, the
results of such tests depend on the validity of the basic assumption that the
data adequately follow the probability distribution selected. Non-parametric
tests offer an alternate means of examining the significance of time-series
trends without requiring that the data follow any particular probability
distribution.
B.4-31
-------
To test whether the apparent decline in Aroclor concentrations 1n fish is
significant, the non-parametric Mann-Kendall test was used on the medians of the
yearly data (Gilbert, 1987). This is the same test used previously in evaluating
the time-series trends for water-column PCB levels. ^
The entire time series of Aroclor 1254 and Aroclor 1016 measurements at
River Mile 175 for largemouth bass, brown bullhead, pumpkinseed and goldfish
(Table B.4-5) was examined. Although the negative values obtained from the Mann
Kendall statistics confirm the visual impression of a general dowJ ward trend for
all concentrations, the trends are not significant at the 95 percent confidence
level for either Aroclor 1254 or 1016 In brown bullhead, nor for Aroclor 1254 1n
largemouth bass nor Aroclor 1016 in pumpkinseed. This result does not mean that
real declines are not occurring in these cases, but Indicates that the trend
cannot be distinguished at the 95 percent confidence level from random
variability 1n the data without making further assumptions regarding the
probability distribution of the data.
Apparent Aroclor Half-Lives in the Fish Population
i
The fact that the Aroclor 1016/1254 ratio has declined fairly continuously
since about 1980 1n most species suggests the absence of any major resuspension
of buried, lightly chlorinated or dechlorinated sediments since that time. The
1980-1988 data are, thus, appropriate for estimating the rate of removal of the
steady state, Aroclor 1254-11 ke PCB components from the system. These components
are taken to represent the more persistent fraction of the PCBs. Estimates of
removal rate or half-life depend on multiple factors, many or most of which may
be unknown or unquantified. This empirical determination is limited to the range
of conditions found 1n the river over the course of the monitoring. In
particular, there is a strong connection between average summer PCB concentra-
tions in water and concentrations in fish. PCB concentration in water, in turn,
may depend on discharge and the 1980s were generally notable for low flows.
Thus, the rate of decay of Aroclor concentrations observed In recent fish data
may, 1n part, reflect only the temporary variations in climate experienced during
this period.
B.4-32
-------
Oust as an exponential decline and half-life were estimated for PCB mass
transport 1n the previous section, an exponential decay pattern tends to hold for
Aroclor concentrations 1n Hudson River fish. In chemical data characterized by
exponential decay patterns, a plot of the log of Aroclor concentration versus
time Is linear:
LN(Aroclor) - < - (P t)
where c 1s a regression constant, f) Is the linear rate of change 1n the LN of
Aroclor concentration relative to t (1/tlme) and t 1s time (years). The half-
life (t1/2) of the compound can be calculated as LN(2)/p. For 1980-1988 fish
data, the Aroclor 1254 decay rate factors and half-Hfe values-at River Mile 175
are shown below.
Aroclor 1254 Half-Lives In Upper Hudson Fish (River H11e 175)
Species
Time Constant, 0
(1/years)
Half-Life, t1/2
(years)
Largemouth Bass
0.046
15.1
Pumpklnseed
0.095
7.3
Brown Bullhead
0.017
40.7
When PCB levels are based on lipid content, the half-Hves for Aroclor 1254
do not change much for the 1980-88 data, yielding 11.7 years for largemouth bass,
9.4 years for pumpklnseed and 37.3 years for brown bullhead. The slope of the
regression for largemouth bass, however, 1s entirely the result of the relatively
high 1980 measurements. For 1981-1988 there has actually been a slightly
increasing trend 1n llpid-based Aroclor 1254 levels in largemouth bass.
The half-Hfe values calculated above can, in theory, be used to estimate
the length of time it will take to reduce levels of the more res1stent higher
chlorinated congeners 1n fish populations, assuming the continuation of recent
hydrologlcal patterns and lack of additional major Input from deep sediments.
For example, the most recent (1988) observed Aroclor concentrations in fish (wet
weight, not lipld-based) are shown below.
B.4-33
-------
Mean Aroclor 1254 Concentrations (ppro), 1988
Species
RM 190
(fort Edward)
RM 175
(Stillwater)
RM 153
(Federal Dam)
Brown Bullhead
6.9
[6.3]
4.1
[3.8]
1.2
[1.2]
Goldfish
33.1
[30.9]
Pumpkinseed
4.2
[3,9]
2.2
[2.1]
Largemouth Bass
1.3
f3.31
1.9
2.8
mi
ILM
Vatuw in 11 ift gtonwtric ww vปtyซ.
Based on the half-life estimates, it is expected to take 70 years to reduce
the Aroclor 1254 component in brown bullhead to below 2 ppm at River Mile 190.
On the other hand, largemouth bass might reach such levels of Aroclor 1254 in
about 25 years.
Aroclor 1016 levels in fish have shown a steady decline since 1980. Figure
B.4-23 shows the exponential regression line for Aroclor 1016 decay in largemouth
bass (lipid-based). Calculated Aroclor 1016 half-lives at River Mile 175 were
rather similar for different species, yielding 3.5 years for largemouth bass, 4.5
years for pumpklnseed and 3.8 years for brown bullhead.
This discussion of half-life calculations is for River Mile 175 only and
depends on whether or not lipid-based values are used. A fuller utilization of
the data, but still not considering variations 1n hydrology, may be accomplished
by considering the following multiple regression relating measured Aroclor
content to these variables: year of measurement (Y), river mile (M), percent
lipid (LP) and sample weight in grams (W):
LN(Aroclor) - ซ + Y + pt M + p, (LP) + (W).
Fish sample data from 1980-1988 taken at Troy and Fort Edward for
largemouth bass, pumpklnseed and brown bullhead were used in these multiple
regressions. As a weighted average, the apparent half-life for Aroclor 1254
B.4-34
-------
components in all fish is 9.3 years, while the apparent half-life for Aroclor
1016 components is 4.1 years. The R2 values of these regressions are fairly low,
ranging from 28 percent to 60 percent. Better performing models (R2-80 to 84
percent) can be constructed by considering total PCBs, rather than individual
Aroclors.
B.4.4.2 Projected PCB Concentrations in F1sh
A 30-year projected average PCB concentration in fish is necessary for the
evaluation of baseline health risks. This average is calculated through a
regression approach similar to that just described above. There are extensive
data sets of PCB concentrations in largemouth bass, brown bullhead and
pumpkinseed in the Upper Hudson. The pumpkinseed data are not very useful for
assessing average concentrations in terms of potential human consumption, since
an effort was made to gather in September primarily yearling pumpkinseed, which
are smaller than commonly edible size. PCB concentrations in largemouth bass and
the bullhead data were, therefore, averaged to obtain a representative value for
human exposure.
The multiple regression here is based on the natural log of the sua of
Aroclors 1016, 1221 and 1254 taken to represent total PCBs. The log of percent
lipid, rather than percent lipid, provided a better fit in these regressions.
Regression equations used Aroclor measurements for brown bullhead and largemouth
bass sampled for River Miles 153 to 190. The best fit equations are:
LN{PCBl^) ซ 12.45 + 0.031 M - 0.188 Y + 0.997 P * 0.0003 W
R2 - 86 percent
SE - 0.553
LW(PCBB>JUJJA#ad) ซ 6.4 + 0.069 M - 0.200 Y + 0.894 P + 0.00045 W
R2 ป 81 percent
B.4-35
-------
where
H
Y
P
U
SE
SE - 0.520
River Nile;
Years since arbitrary datum of 1900 (Year - 1900); ^
Natural log of percent lipid content of sample;
Weight of fish (grams); and
Standard error of regression estimate.
Using these equations, the 30-year projected median PCB eventrations In
these species at River Mile 175 are 0.19 ppm for largemouth bass and 0.16 ppm for
brown bullhead.1 The average PCB concentration over the entire 30-year period
will be greater. Because the regressions use lognormal variables, a straightfor-
ward expression for this average cannot be developed. Thus, a slmpl&Honte Carlo
sampling procedure was employed to provide an estimate of this 30-year average.
Steps 1n the Monte Carlo procedure are outlined below.
Incrementally select years (e.g. 30 years) ranging from 1991 to
2020.
Uniformly sample fish between Federal 0am and Fort Edward.
Select a random sample for percent 11pid (P) and weight (W) assuming
they are lognormally distributed, I.e. LN(P) and LN(J/) are normally
distributed with the following parameters:
LN(/>; LN(V)
mean s, mean s
-0.2925 1.1597 6.2666 0.6186 (bass)
0.7918 0.9575 5.8370 0.4269 (bullhead).
Calculate the annual average PCB levels 1n the fish using the
samples of P and V 1n the above regression equations, which also
Involves adding a random error term, which 1s modeled using the
standard error (SE) of the regression equations.
'Values 1n the equations are P0.29, W-608.867 (bass); P-0.681, W-373.754
(bullhead); years since 1900 Y-105 and Mป175.
B.4-36
-------
Repeat the calculation a large number of times (500 here) and
calculate the overall 30-year mean PCB estimate for brown bullhead
and largemouth bass.
The frequency distribution of the annual average PCB concentration after
500 simulations (300,000 fish realizations) is shown in Figure B.4-24. Note that
over a 30-year time horizon the distribution remains skewed and 1s best described
as lognormal. The simulated average of largemouth bass and brown bullhead has
a grand mean (mean of the means) of 1.23 ppm and a standard deviation of 0,148.
The upper 95 percent confidence limit on the grand mean 1s 1.55 ppm. Statistics
on the 30-year average (7) for Individual species follow:
Largemouth Bass: * - 0.99; s, - 0.18; X.K (upper 95 percent confidence limit) -
1.38 ppm.
Brown Bullhead: 1.47; s, - 0.24; X.,, (upper 95 percent confidence limit)
2.00 ppm.
B.4.4.3 Relation Between PCB Concentrations 1n Fish and Nater
An Important determinant of PCB concentrations 1n fish, although not
necessarily the only one, Is likely to be ambient water concentrations. In
particular, ambient PCB concentrations 1n water during the summer low flow season
(June-September), which 1s also the period of maximum biological production, are
likely to provide a good Indicator.
M. P. Brown et al. (1985) provided what appeared to be a nearly exactly
linear relationship between 1977-1983 mean summer water column PCB concentration
and median I1p1d-based total PCB concentrations 1n yearling pumpklnseed collected
annually In September at River Nile 175. Total PCBs were calculated by adding
concentrations calculated for Individual Aroclor mixtures and the median was
employed as a robust Indicator of central tendency. (Mean and median are very
close to one another 1n this data set.) Only yearling pumpklnseed were Included,
which necessitated the exclusion of a few older Individuals collected In 1981-
1983.
B.4-37
-------
Figure B.4-25 provides a plot similar to that of M. P. Brown et al. (1985)
extended to Include the 1984-1988 data. Data for fish in 1987 and data for
summer PCB concentrations 1n water for 1986 could not be Included as they do n^l
exist. L1p1d-based total PCB concentrations 1n yearling pumpklnseed at River
Nile 175 are plotted against summer PCB concentrations 1n water at River Mile
168. The general linearity of the relationship 1s maintained, although the 1984
and 1985 data points do not 11e directly on the earlier line.; The apparent
strength of the relationship 1s largely due to the 1979 and 1980 observations,
when summer PCB concentrations in the water column were high. Water column PCB
concentrations have been much lower 1n subsequent years and It 1s not necessarily
clear that the relationship In the lower part of the scale Is linear.
Consideration of Individual Aroclor measurements 1n pumpklnseed (Figure
B.4-26) suggests that both Aroclor 1016 and Aroclor 1254 1n the fish are linearly
related to total PCB concentrations measured in the summer water column at River
Mile 168, although with differing slopes. Measurements of Aroclor 1254 1n
particular look as though there 1s an asymptotic leveling off 1n the lower part
of the water concentration range. This observation may indicate a non-linear
b1oaccumu1at1on effect. As water column concentrations of PCBs become very low,
the body burden of PCBs 1n fish may begin to decline at a slower rate, perhaps,
as a consequence of some direct food pathway from sediment to fish.
The plots 1n Figures B.4-25 and B.4-26 have compared the PCB concentrations
1n pumpklnseed against concentrations in the water column at River Mile 168,
downstream from River Mile 175 where the pumpklnseed were collected. The data
can also be compared to PCB concentrations 1n the water column at Schuylervllle,
upstream from River Mile 175, with rather similar results (Figure B.4-27). Use
of the Schuylervllle data allows a point for 1986 to be plotted.
Yearling pumpklnseed may be expected to provide the best relationship among
the fish monitored to ambient PCB concentrations 1n the water column as they are
sedentary, relatively low on the trophic chain and all the same year class.
Total 11p1d-based PCBs In all largemouth bass sampled at River Mile 175 bear a
B.4-38
-------
rather similar relationship to water column concentrations as shown in Figure
B.4-28. A single sample from 1979 is omitted here as possibly anomalous. There
appears to be even more Indication of an asymptotic leveling off of body burden
in the lower part of the water column concentration range. ^
Total lipld-based PCBs in brown bullhead collected at River Mile 175 also
exhibit a strong linear relationship to summer water concentration as shown in
Figure B.4-29, except for two points that lie outside the 95 percent confidence
limits.
>
Given the apparent linear relationship, simple regressions can be developed
to indicate dependence of fish PCB burden on summer water concentration, in the
form:
PCB ซ a + P S
where PCB is the median Hpid-based total PCB concentration (ppm) at River Mile
175 for a given species in a given year, and S 1s the average Summer water
concentration (itg/7) at River Mile 168 for the same year. The results for
pumpkinseed, largemouth bass and brown bullhead are shown below.
Relationship of L1
pid-Based PCBs in Fish and Summer PCB Concentrations in Vater
Species
ซ
u
R2 (X)
SE
p value
Pumpkinseed
131.0
(P-.004)
1602.8
(P-.00001)
97.1
53.1
.00001
Largemouth Bass
-76.50
(P-.722)
6804.3
(P-.00002)
96.0
381.2
.00002
Brown Bullhead
134.45
(P-.431)
2407.6
(P-.0015)
83.6
296.6
.0015
The i coefficients (a constant) for the largemouth bass and brown bullhead
are not statistically significant at the 95 percent confidence interval, which
would be consistent with fish PCB concentrations going to zero as summer water
concentrations go to zero. The |1 coefficients, all highly significant, differ
from one another by up to more than a factor of 3. These coefficients represent
the species I1p1d-based bioaccumulation factor (BAF) for total PCBs relative to
B.4-39
-------
water column PCB concentration. (Note that BAF ซ P x 1,000, since PCBs in fish
are 1n ppm and water concentrations are in ppb.)
An interesting question is whether these BAFs, calculated at River Nile
175, also apply at other locations. If PCB concentrations in fish respond both
to PCB concentrations in water and sediment, perhaps via a benthic food chain
pathway, the answer might well be yes. The data are not sufficient to resolve
this question. Despite recent data for several species of fish 1n the Thompson
Island Pool (River Mile 190), there 1s no water column PCB monitoring at the same
location. The fish data from River Mile 190 are all from 1984 or later, by which
time PCB concentrations in fish, in general, had declined. If largemouth bass
data at River Mile 190 (1984, 1985, 1986 and 1988) are plotted against downstream
PCB concentrations in the water column at Stillwater, all the data points lie
slightly above the regression line. Similar results are found for limited
samples of other species. As noted earlier, water column PCB concentrations
increase upstream. Thus, it is not clear 1f the relationship at River Mile 190
1s any different from that observed at River Mile 175, particularly as no
observations are available from the high PCB concentration periods before 1980.
Except for the data at River Mile 175, there are no other significantly
long runs of observation (>5 years) on a single species at a single location in
the Upper Hudson. There is a long record (1977-1988) for samples of brown
bullhead at River Mile 153 just below the Federal Dam. Here the limited PCB
samples from the water column were all non-detects at the 0.1 pg/1 level, so the
relationship between fish PCB burden and water concentration can not be directly
determined.
B.4.S Summary
Summarized here are the questions posed at the beginning of Section B.4 and
the Information currently developed to help answer them.
(1) What 1s the potential for migration and redeposition of PCBs 1n
sediments?
B.4-40
-------
Flood frequency analysis results have yielded improved
estimates of the high flow recurrence intervals, which may
govern sediment scour and mass redistribution of PCBs in
sediments.
Sediment scour appears to be influenced by minimum scouring
flows necessary to suspend appreciable sediment loads.
Minimum scouring flows are approximately >10,000 - 20,000 cfs.
Few large spring floods have occurred in the 1980s, but the
high flows in spring 1983 and somewhat lower high flow in
spring 1987 both resulted in somewhat increased PCB loads.
The potential for resuspenslon of significant amounts of PCBs
during a future major flood remains an important unanswered
question.
(2) What is the relationship between PCBs in sediments and PCBs in the
water column?
Multiple regression analyses suggest PCBs are directly related
to flow at high flows and inversely related to flow at low
flows, although the relationships do not have very great
predictive strength.
It is unclear from available data under what flow conditions
PCBs in the water column are predominantly dissolved or
adsorbed to suspended sediment. It appears, however, that the
major portion of the annual PCB transport occurs during the
highest flows, a finding that suggests that PCBs attached to
sediments are a major component of the PCB load.
The TAMS/Gradient estimate of PCB mass transport from the
Upper to Lower Hudson from 1977-1989 is approximately 15,000
kg. Estimates prepared by previous investigators of PCB
stored in the Upper Hudson sediments total approximately
94,000 kg: 23,000 kg in the Thompson Island Pool; 50,000 kg in
the other reaches of the Upper Hudson; and 21,000 kg in the
remnant areas.
Empirical time trends suggest that the PCB loads, transported
in the water column, have diminished since the late 1970s,
with a half-life of approximately three years.
(3) What is the effect of (1) and (2) on the levels of PCBs in fish?
B.4-41
-------
PCB levels 1n fish have declined from the high levels 1n the
late 1970s; however, the rate of decline has been very low 1n
recent years.
A strong linear relationship between PCBs in the water colunfg
and PCBs 1n fish holds for the range of PCB concentrations
observed in both water and fish. Field BAF values relating
lipid-based PCB levels 1n fish to levels 1n water range from
1.6 million to 6.8 million.
Assuming current trends, the projected 30-year average (1991 -
2020) PCB concentrations 1n fish is on the orfer of 1.5 ppm.
This projected average 1s a "best-case" estimate in that 1t
assumes no major resuspenslon of PCBs 1n sediments.
Lack of sufficient paired fish and sediment samples (1n time
and space) precludes an analysis of the correlation between
PCBs in sediments and PCBs in fish.
The long record of PCB measurements in fish and surface water provide an
extensive and reliable database, with measurements as recent as 1988 and 1989,
respectively. Although sediments have been sampled by GE as recently as 1990,
the monitoring record for sediments lacks the extended coverage 1n both time and
space of the water and fish samples. Thus, the sediment data provide a somewhat
less reliable database from which to extrapolate trends and relationships with
PCBs 1n water and fish. Other media, including other aquatic biota, air and
upland plants, have not been studied as extensively as fish, surface water and
sediments.
B.4-42
-------
SYNOPSIS
SEDIMENT TRANSPORT MODELING
(Section B.5)
Development and calibration of hydraulic and sediment transport models have been
initiated. Because PCBs in the river are bound primarily to sediments, scour of sediments is a
crucial mechanism to the movement of PCBs. A mathematical model provides one tool to
predict potential scour and redeposition of sediments containing PCBs. A basic modeling
framework has been developed contemporaneously with the data analysis in order to determine
what type of modeling may later be appropriate and feasible. The current modeling effort is
limited to implementation and calibration of a hydraulic model of the Thompson Island Pod
and the preliminary development of a sediment transport model for this reach of the Upper
Hudson.
The WASP4 family of models (B.5.1) is selected for use, because of its flexible format
and data handling capabilities. Although it can accept sediment transport information, it does
not have a sediment scour or transport routine. Thus, one objective is to develop sediment
transport routines that can work in conjunction with WASP4.
Previous attempts at sediment transport modeling in the Thompson Island Pool
undertaken with the HEC-6 model (B.5.2) are reviewed. There are difficulties, such as the
presence of cohesive organic sediment, that make modeling sediment transport in the Upper
Hudson difficult. This difficulty and problems in calibrating hydrodynamic models have resulted
in a low degree of success for previous efforts by other researchers.
DYNHYD5, the hydrodynamic module ofWASP4, is described (B.5.3). A detailed two-
dimensional link-node mathematical representation of the Thompson Island Pool is developed
using a Geographic Information System. The model is calibrated to river stage data recorded
in the barge canal at both the upper and lower ends of Thompson Island Pool Comparison
of results predicted by the model with measured stage observations yielded a fit to within
approximately one-tenth of a meter for several flood episodes.
A sediment transport model, STREAM, is described (B.5.4). This model is used to
simulate stream bed and streambank erosion, sediment deposition and resuspension and
sediment transport. The potential application of this model to the Upper Hudson is still in
development. Although preliminary results indicate an improvement in prediction over previous
HEC-6 modeling, calibration of this model has not been completed.
Continued modeling efforts will depend on obtaining additional data to define current
sediment bed deposits and grain-size distribution and to measure suspended sediment at both
upstream and downstream sections of the Thompson Island Pool
-------
PAGE INTENTIONALLY LEFT BLANK
-------
B.5 Sediment Transport Modeling
B.S.I Overview
PCB loads in surface water are highest during high river flows, although
the magnitude of response appears to have diminished in recent years. It is
currently uncertain 1f this apparent declining response of PCBs in the water
column is a result of the generally low flows of the 1980s and/or if higher than
normal and flood flows could cause significant erosion, downstream transport, and
increased PCB concentrations 1n the water column and fish. Potential sediment
scour, resuspension, and redeposition, in response to high flows or floods must
be evaluated in order to extrapolate from current conditions (e.g., PCB levels
in sediment, water fish) and assess likely future trends. Furthermore, PCB
dissolution and dispersion as a consequence of possible maintenance or remedial
dredging actions must be evaluated. Both of these evaluations are facilitated
by hydraulic and sediment transport models, which can be operated to test the
response of sediments to flood flows and evaluate changes in dissolved and
adsorbed PCB concentrations in the water column.
In contrast to mathematical transport models, evaluation of historical
monitoring data provides information on the Interrelationships of PCBs 1n
sediment, water and biota (fish) based on the observed data. Thus, the
relationship between flow and sediment discharge suggests a possible threshold
scouring flow phenomena. Other examples of empirical relationships supported by
an evaluation of the data are: regression and time-trend relationships among
flow, suspended sediment concentration and PCBs; or correlations between PCB
levels in fish and those 1n the water column. The data record may be too short
or data from all media insufficient, however, to prove accurate as a predictive
indication of future trends for conditions other than those covered by the
historical record. In Phase 1, available flow, suspended sediment and water
column PCB data have been evaluated to assess completeness and analyze
statistical trends. In addition, some initial hydraulic and sediment transport
model testing have been performed to assess whether detailed modeling is feasible
and appropriate.
B. 5-1
-------
The ability of transport models to provide useful results depends on a
number of factors. Sufficient data must be available to calibrate the models and
determine whether the models adequately mimic observed flows, sediment transport,
and PCB concentrations In the water column. The model must be sufficient^
detailed to capture the Important physical processes Influencing flow and
sedlment/PCB transport. The models should be tested and supported, yet flexible
enough to allow modifications for site-specific conditions.
The WASP4 family of models was chosen for Upper Hudson hydraulic and PCB
transport models. This modeling package is widely used and supported by the
USEPA Center for Exposure Assessment Modeling (CEAM) in Athens, GA. It provides
a flexible framework for incorporating modified submodel components as necessary.
As currently distributed by CEAM, the WASP4 package consists of Independent, but
fully compatible components DYNHYD and WASP. DYNHYD Is a 11nk-node hydrodynamic
model that provides Input to WASP. WASP is the water quality module and can be
compiled by substitution of submodels as EUTRO, a eutrophicatlon and conventional
pollutant model, or TOXI, a toxic pollutant model. In addition, the package also
contains a Beta test version of FCHAIN, a food chain model.
The WASP4 package 1s based on a hydrodynamic model of sufficient complexity
to provide useful sediment transport parameters, yet is straightforward enough
to be Implemented and calibrated with available data. Use of DYNHYD to model two
dimensional hydrodynamic flows provides the capability of emulating hydraulics
of at least trapezoidal (non-rectangular) reaches.
Although WASP4 1s designed to accept sediment transport Information, the
CEAM package notably lacks any sophisticated sediment scour/transport routines.
Thus, an important facet of modeling efforts 1s to develop sediment transport
routines, appropriate to conditions in the Upper Hudson, which can take the
hydrodynamic model DYNHYD as Input and, in turn, provide input to the water
quality model WASP. Summarized here are hydraulic and sediment transport model
development and calibration results, as developed in Phase 1.
B.5-2
-------
B.5.2 Previous Modeling Studies
Attempts to model sediment and PCB transport from the Thompson Island Pool
using the HEC-6 computer model have been made by Lawler, Matusky & Skeller (LHS)
in 1978 and 1979 and Zimmle in 1985. LMS (1979) modeled sediment transport from
Fort Edward into the Lower Hudson and modeled reaches from Lock 7, just below
Rogers Island, to the Federal Dam. Zimmle (1985) applied the same model to the
Thompson Island Pool area, following the extensive resurvey of sediments there
in 1984.
Both efforts were based on the HEC-6 model (US Army Corps of Engineers,
1977) and both are subject to limitations.
1. The HEC-6 model is designed to address transport of cohesionless
sediments, and cannot explicitly model cohesive organic sediments.
Such sediments may play an Important role 1n Hudson River PCB
transport.
2. HEC-6 is a one-dimensional sediment model and, thus, cannot
replicate the lateral variability 1n sediment composition In the
Thompson Island Pool. This Imposes limitations on the ability to
model flood effects on specific hot spots.
3. It is unclear whether the calibration data are sufficient to
accurately fit a model of sediment transport.
4. As discussed at B.4, Ziramie (1985) appears to have overestimated the
magnitude of flood flows and resulting probability of sediment
scour.
The purpose of the LMS study was to predict PCB transport downstream under
no action and mitigated (dredged) conditions. The focus was on expected impacts
of average rather than extreme conditions. Thus, hydraulic simulation was
undertaken over observed flows of the 20-year period 1958-1977, rather than
generating specific flood recurrence events. The conclusion of LMS (1979) was
that the hot spot dredging project would reduce the time needed to deplete the
input of significant amounts of PCBs to the estuary from 69 to 44 years.
B.5-3
-------
A detailed critique of the LMS application of the HEC-6 model 1s provided
1n the Feasibility Study (NUS, 1984). The hydraulic submodel (HEC-2) was
considered adequate for the application; similarly, the one dimensional nature
of the model would not have Introduced significant difficulties, given the
relative length of reaches and heights of dams between Fort Edward and Troy.
Significant problems were, however, noted in the calibration of the hydraulic
submodel. For the Thompson Island Pool, the only calibration data used were
apparently a limited set of 1976-77 observations at Lock 7, which 1s the upstream
end of the reach. Because the grade over the reach 1s quite low, elevations at
Lock 7 should be closely controlled by the downstream boundary at the Thompson
Island Dam, although calibration data at both the upstream and downstream ends
of the reach would of course be preferable. The NUS study notes that the
predicted model water surface elevations at Lock 7 consistently exceeded
observations and, indeed, "the surface elevation within the drawdown curve at the
dam already exceeds the observed elevation at the upstream end of the reach."
An error 1n specification of the rating curve at the dam was thought to be the
cause. The result 1s that flow rates corresponding to a given water surface
elevation are overestimated by about 30 percent, which would imply that the
average velocities driving the sediment model are overestimated.
More severe problems appear to apply to the sediment transport portion of
the LMS modeling effort. Upstream sediment loading was simply calculated by a
regression relation at Glens Falls. The first sediment data available for
calibration were at Lock 4, Stillwater. Hodel predictions here consistently
underestimated measured sediment concentrations by about a factor of two for high
flows and up to an order of magnitude for low flows. Further, the HEC-6
application predicted sediment deposition throughout the reaches between Lock 7
and Lock 4 at all flows below the one percent exceedance value. NUS (1984)
demonstrated that significantly better performance was achieved by simply taking
the Glens Falls sediment loads and routing It through to Lock 4 with no scour or
deposition. Thus, the sediment transport predictions of the model appear to be
unreliable and, 1n turn, cast doubt on PCB transport predictions.
B.5-4
-------
Among the reasons for poor performance of the sediment transport model may
be a lack of adequate calibration data. The problem may also have been
compounded by assigning the entire silt component to the coarsest model category
(0.032 to 0.062 mm). More generally, the Inability of the model to reproduce
observed sediment behavior may be a result of neglecting the role of cohesive
organic sediments. In a recent, detailed study of a river in Wisconsin, which
possesses extensive PCB contamination together with Input from pulp and paper
Industries, Ga1lan1 et a?. (1991) demonstrate that both scour and deposition of
PCB-contaminated organic sediments must be modeled as time-dependent processes.
The rate of settling is dependent on the state of flocculatlon of the cohesive
sediments, a time-dependent process, while resuspenslon 1s controlled by the
state of compaction, which also Increases over time. The critical shear stress
for resuspenslon of these sediments was found to vary by an order of magnitude
with time, since deposition. Together these factors create a situation in which
sediments, once suspended, have a tendency to keep on moving, with rates of
deposition less than would be predicted from a model that does not treat cohesive
sediments.
Zlmmle (1985) also used the HEC-6 model to analyze the Thompson Island
Pool. Using approximately 50,000 depth soundings of the Thompson Island Pool
obtained by Raytheon 1n 1982 and a computer program to convert these soundings
to average cross-sectional profiles, Zlmmie defined 31 reaches, from 634 to 1,267
feet in length, between Lock 7 and the Thompson Island Dam. The cross-sectional
average geometry of the Thompson Island Pool was thus very fine, although still
one-dimens1onal. In addition, Zimmie used detailed sediment size data for each
model segment. Z1mm1e's results indicated that water discharges of less than
32,000 cfs would not cause bed changes of more than 0.14 feet 1n any modeled
segment. A 63,700 cfs discharge event (taken to represent a 100-year flood) also
predicted only small amounts of local bed changes, with a maximum scour of 1.5
feet In one reach.
Zimmle's application was subject to the same theoretical limitations of the
HEC-6 model as was the LMS (1979) study. Most notably, the transport of cohesive
organic sediments was not addressed. A number of practical difficulties In
B.5-5
-------
implementation, similar to those of IMS, may be of even more significance. The
first concern 1s the adequacy of calibration. Zimraie's HEC-6 calibration was to
observed stage data at the staff gauge at Lock 7, at the mouth of the Champl$1n
Canal, just above the uppermost reach of the model. Observed errors 1n HEC-ซ
predictions of stage at this gauge are ฃ3 percent, which 1s a substantial
improvement over the LMS (1978) calibration. Nevertheless, it 1s unclear why
Zimmie did not employ similar barge canal stage data from the Crockers Reef Guard
Gate, at the downstream end of the pool. Use of such data would have enabled
further refinement of the calibration, I.e., a better fit could have been
obtained by using variable, Manning's roughness coefficients for model segments.
Calibration of the sediment model was problematic; 1977 and 1984 data were
available for grain size distribution of bottom sediments, but no suspended
sediment monitoring was available at the downstream end of the model. Thus,
predictions of sediment output from the Thompson Island Pool cannot be considered
to be calibrated.
Other serious problems apply to predictions of scour expected to occur
under given recurrence Interval floods. Zimmie assumed that the 10-year
recurrence peak discharge event was 46,600 cfs and the 100-year recurrence event
was 63,700 cfs. These flows were applied at the upstream end of the Thompson
Island Pool and routed through the system with no additional inflow. As noted
at B.4, these values are actually those proposed by FEMA for the downstream
corporate limits of the Town of Fort Edward, located below Fort Miller. The FEMA
estimates for 10 and 100-year discharges at the Village of Fort Edward (above
Lock 7) are 38,800 and 52,400 cfs. Additional analysis in Section B.4 suggests
that even these values are too high and that the 10 and 100-year discharges at
Lock 7 are actually on the order of 33,300 and 44,300 cfs. This finding implies
that the 10-year recurrence event modeled by Zimmie was actually greater thfan the
100-year recurrence event. The likelihood of significant scour in the Thompson
Island Pool may, thereby, be significantly overestimated.
B.5-6
-------
B.5.3 Hydrodynamic Model Description
B.5.3.1 Use of HASP4 Family of Models
*
TAMS/Gradient modeling efforts are founded on the HASP4 family of models.
The current hydrodynamic model in WASP4 Is DYNHYD, Version 5.02 (0YNHYD5). This
is an updated version of DYNHYD4, documented 1n Ambrose et *1. (1988). Important
modifications for Version 5 include the ability to accommodate (mildly)
trapezoidal channels and evaporation and precipitation (Wool;, 1990). The model
is an enhancement of the Potomac Estuary hydrodynamic model, DYNHYD2 (Roesch et
a/., 1979).
DYNHYD5 1s constructed as a link-node model, which 1s essentially a finite
volume approach, allowing a two-dimensional problem to be reduced-to a series of
one-dimensional transport problems. Essentially, the nodes may be conceived of
as junctions which store water, while the links are idealized channels which
convey water between a pair of junctions. Taken together, the junctions account
for all the water volume in the river, while the channels account for all the
water movement in the river. This approach yields a computationally efficient
system; at each time step, the equation of motion 1s solved on all the links,
giving flow velocities, while the equation of continuity Is solved at the nodes,
giving water elevations (hydraulic head).
B.5.3.2 Governing Equations
DYNHYD5 1s based on the equations of continuity and momentum, which
describe the propagation of a long wave through a shallow water system. The full
equations of continuity and momentum for flow 1n an open channel are usually
referred to as the St. Venant equations, which can be derived on a unit control
volume (e.g., Bras, 1990). The momentum, or motion, equation is:
B.5-7
-------
where
V - local velocity of flow In direction x;
A - channel cross-sectional area;
g - acceleration of gravity;
y ป depth to the centrold of the flow cross section;
q - spatially averaged lateral Inflow rate;
af - frlctlonal acceleration;
af - gravitational acceleration; and
aM - wind acceleration.
The terms on the left hand side of the equation represent, respectively,
the local acceleration In velocity; the Bernoulli acceleration, or rate of
momentum change by mass transfer; the momentum change Induced by pressure
differentials related to large water-surface changes; and the momentum change
caused by the Incoming mass of lateral Inflow.
The second equation, the continuity or mass balance equation, 1s given by:
dQ dA ~ /9\
Si * Tt "ฆ m
For the DYNHY05 solution of the momentum equation It 1s first assumed that
the flood wave length 1s significantly greater than the depth and, therefore, the
momentum change Induced by pressure differentials 1n (1) can be Ignored. For
this reason, the model 1s not appropriate to dam break situations. It 1s also
assumed that lateral Inflow occurs only at nodes and not at the links, so that
the fourth term on the left side of (1) Is also zero.
The right hand side of (1) contains terms for the gravitational and
frlctlonal acceleration. In DYNHYD5 a separate term 1s developed for wind
acceleration. Although It 1s an Important factor for flow 1n estuaries or large
B.5-8
-------
lakes, it is Insignificant in the Upper Hudson. The wind acceleration tern will,
therefore, be ignored 1n this application.
Gravitational acceleration a, is driven by the slope of the water surface
K-
ar ป -gr sin (hL). (3)
When hL 1s small, s1n(hL) may be replaced by hL, which can then be written
1n terms of the rate of change 1n head, H, yielding
ป <4>
Frictional acceleration af is equal 1n magnitude and opposite in sign to
the gravitational acceleration aa during steady flow. During unsteady flow, 'it
must also balance other changes 1n momentum and, thus, can be written in a form
similar to a9, but 1n terms of the energy slope, S, as:
a/-?-งf- (5) -
The assumption 1s then made that the flow is nearly steady, so that the
energy slope can be determined 1n the same manner as 1n steady, uniform flow.
In particular, the empirical Manning equation for steady, uniform flow states,
for V 1n units of meters/second, is:
<>
1n which n 1s Manning's roughness coefficient and R is the hydraulic radius.
Solving this equation for the energy gradient and substituting Into (5)
yields the following expression for the friction acceleration:
B.5-9
-------
in which V2 has been replaced with V* |V| to Insure that the friction acceleration
will always oppose the direction of flow.
The complete momentum equation can then be rewritten, neglecting the wind
acceleration, as:
'j
Sv _ 3v aBH _ _ n* ,,,
Si V-& 9~te V,V" <8)
The equation of continuity (2) 1s likewise applied to channels by assuming
that lateral Inflow occurs only at nodes. For flow In a rectangular channel of
constant width B, (2) may be rewritten as:
dH _ 1 dQ
Ht B dx
(9)
in which .the term on the left is the rate of water surface elevation change with
respect to time (m/sec) and the terms on the right represent the rate of water
volume change with respect to distance per unit width (m/sec).
For mildly trapezoidal channels, such as that of the Upper Hudson, 1n which
the width B 1s much greater than the depth, the depth can be taken as nearly
equal to the hydraulic radius R and the solution for the trapezoidal channel can
then be undertaken on an approximately equivalent rectangular channel. For BปR
and moderate side slopes, this method Introduces negligible error into the
solution. v
In addition to the equations of momentum and continuity, solution requires
the Imposition of boundary conditions. For a river these typically Include an
upstream Inflow condition and a downstream boundary or outflow condition.
Variable Inflows can be specified as plecewlse linear functions, which are
Interpolated 1n time.
B.5-10
-------
0YNHYD5 1s set up to specify downstream boundary conditions as: 1) an
outflow specification; 2) a fixed head boundary; or 3) a tidal boundary. None
of these conditions 1s appropriate for study of the Thompson Island Pool, where
the downstream boundary consists of a broad-crested dam. Flow over this dam 1s
not constant nor fixed head, but depends on the elevation behind the dam.
Therefore, the model was modified to accept a dynamic discharge condition. That
1s, outflow 1s specified as head-dependent and calculated via the Hulslng
equation for flow over dams used as weirs:
0 = mbH3'2 (10)
1n which Q 1s outflow (cfs); b 1s the effective length of the dam crest (760
feet); and H Is the head on the weir (feet), defined as the height of the level
liquid surface above the crest. In this equation m 1s an experimental factor and
taken as a calibration parameter. The theoretical value for a trapezoidal weir,
with 4:1 end slopes, 1s ซ-3.367.
B.5.3.3 Model Implementation
The link-node computational network Is solved by writing the governing
equations 1n a finite difference form, yielding one equation for each channel and
one equation for each junction. A solution 1s then obtained explicitly via a
modified Runge-Kutta method.
The momentum equation (8) can be written 1n finite difference form as:
v^-v' g Vi J v/S (11,
AC 1 Axt s4*, (ii/)"5
In which the subscript 1 refers to channel; the superscript t refers to the
present time step of length At (sec); and Ax, Is the length of channel 1
(meters).
B.5-11
-------
The water surface gradient AH^Ax, is computed from the junction heads at
either end of the channel. However, the velocity gradient cannot be computed
directly from upstream and downstream channel velocities, because of the possible
branching 1n the network. Therefore, an expression for the velocity gradient
with the channel must be derived from the continuity equation (2). On
substituting V*A-Q, the velocity gradient 1s:
dv _ 1 dA V dA (12)
dx Ad t A dx'
Writing BปR-A and B*AH-&A, the velocity gradient can be expressed in finite
difference form as:
AVj
Ax,
A Hj
At
A Ht
Ax,
(13)
wherein AH,/At can be computed as the average water surface elevation change
between the junctions at each end of channel 1 during a given time step.
Substituting (13) Into (11) yields the explicit solution for the equation of
motion 1n each channel for each time step, again neglecting terms for wind
acceleration:
Vi*1 . y/ + At
V/ AH/
At
(Vj)2
Ri
-9\
A H'
Ax,
- 9
n<
(J?/)4/3
Vi
t I TT T |
I Vj I
The equation of continuity (9) can be written in finite difference form as:
A 0/
rปe*l u c
Hi ~Hi
At
Bi Ax,
(15)
B.5-12
-------
in which j is the junction number. AQ_, 1s the total sum of all 1-1 to n flows
entering and leaving junction j, while B/AXj 1s equal to the surface area of the
junction ASj.
This yields an explicit solution for the head at each junction at each time
step:
ฃ Qป tl6\
hj*x = af - At
Asf
The solution implementation uses a modified Runge-Kutta procedure. To
proceed from time t to time t + At, the following steps are taken.
1. For time t + At/2, predict the mean velocity in each channel using
the channel velocities, cross sectional areas and junction heads
from time t.
2. Predict the flow in each channel for time t + At/2, using the
velocity obtained in (1) and the cross-sectional area from time t.
3. Compute the head at each junction at time t + At/2, using the flows
derived 1n (2).
4. Compute the cross-sectional area of each channel at time t + At/2,
using the heads computed in (3).
5. Predict the mean velocity for each channel at time t + At, using the
velocities, cross-sectional areas and junction heads computed for
time t + At/2.
6. Compute the flow in each channel at time t + At, using the velocity
for time t + At computed 1n (5) and the cross sectional area
computed for t + At/2 1n (4).
7. Compute the head at each junction at time t + At, using the flows
computed in 6).
8. Compute the cross-sectional area of each channel at time t + At
computed in (7).
B.5-13
-------
B.5.3.4 Modal Setup for Thompson Island Pool
Initial modeling efforts have focused on the Thompson Island Pool. ^If
results from these efforts prove promising, they will be expanded to cover mor^
of the river during Phase 2. In order to Implement the hydrodynamlc flow model,
the geometry of this reach of the river was defined or dlscretlzed and the flow
model was calibrated to observed USGS flow records.
4
The links in the DYNHY05 model will later form the model segments for
Implementation of the transport model (WASP4). During Phase 1, both one-
dimensional and two-dimensional versions of the DYNHYD5 were Implemented.
Although a two-dimensional flow and sediment model will likely be required to
investigate the potential erodlblHty of specific areas, the sediment model was
developed and calibrated 1n one dimension, which required a one-dimensional
hydraulic model. The link-node discretization was defined in such a way that the
one-d1mens1onal Implementation 1s a direct approximation of the two-dimensional
model. This approximation 1s accomplished by collapsing lateral nodes -in the
two-dimensional discretization into single nodes for the one-dimensional model.
The two-dimensional 1 ink-node network is, therefore, described first.
The river cross-sections, bathymetry and initial geometric discretization
of the Thompson Island Pool were based on the work 1n this reach by Zimmie
(1985). Although some specific limitations of that effort, notably the
limitations of Its calibration were addressed earlier, the basic geometry
employed by Z1mm1e Was adopted for the DYNHYD5 model. Adoption of this geometry
was necessary to avoid the extensive effort of recalculating pool bathymetry from
raw data. The cross-sections developed by Zimmie were used 1n recognition of the
fact that current river bathymetry may differ from that determined 1n 1982 and
that additional data may be needed during Phase 2.
In Zlmmle's effort, the Thompson Island Pool was described by 32 averaged
cross sections, from River H1le 188.5 (Thompson Island Oam) to River Nile 193.64
(just below the mouth of the Champlaln Canal near Lock 7 and the tip of Rogers
Island). The geometry of the river bed was defined by approximately 50,000 depth
B.5-14
-------
soundings taken 1n late spring of 1982 by Raytheon for NYSDEC. (Data are
apparently now available only 1n a hardcopy printout.) According to Zimmle
(1985), these data were processed by computer to generate cross sections every
115 to 35 feet along the river axis. The cross sections were then grouped Into
sub-reaches of like physical dimensions and hydraulic characteristics, then
interpolated by hand, with the Intention of preserving major bedforms, to
represent sections varying from 634 to 1267 feet 1n length. (Sections were drawn
to Barge Canal datum, which 1s 1.177 feet higher than the National Geodetic
Vertical Datum.) Overbank elevations were then approximated from FEMA studies.
The upstream end of each cross section was assigned a river mile station ID;
thus, the true position of each averaged cross section Is between the transect
lines shown by Z1mm1e. Host of the river mile transect lines appear to
correspond to sediment sampling transects shown by Gahagan and Bryant (1982).
The Z1imn1e cross sections do not represent exact, discrete cross-sections,
but Instead represent average conditions throughout a short reach. As such, they
should be appropriate for hydrodynamlc modeling. Because they are average cross-
section descriptions, when they are transferred onto an actual map of the river,
It 1s often necessary to do minor shrinking or stretching to fit the river width
at the modeled cross-section.
Adopting Zimmie's average cross-section description represents a
compromise. On the one hand, these cross-sections were based on detailed
bathymetry of the Thompson Island Pool. On the other hand, the original sounding
data are not readily available and the accuracy of the Interpolations cannot
readily be checked. It is possible that re-discretization of the reach will be
needed for future work.
To implement a two-dimensional 11nk-node network, the cross-sections were
divided Into up to three subsections to reflect bedforms. For example, a given
reach might contain a relatively deep central channel and shallower, near-shore
areas. Some cross-sections did not indicate any clear change 1n bedform across
the width (e.g., a channel of uniform depth, without shallow near-shore areas),
in which case only a single horizontal node was used. A* model node was
B.5-15
-------
In order to calibrate the model 1t 1s desirable to have at least two
reference points. These are provided by the staff gauges maintained during the
navigation season on the Hudson R1ver/Champ1a1n Canal by NYSDOT. Gauge #119 Is
located at the mouth of the Champlaln Canal below Lock 7c and, thus, Is
approximately coincident with the uppermost node of the link-node model. Gauge
#118, Crockers Reef, 1s located at the guard gate 1n the land cut bypassing the
Thompson Island Dam and 1s approximately parallel with the eastern section of the
dam. The canal entrance 1s, however, approximately 0.5 miles upstream of the
dam. Given the low energy gradients normally experienced 1n the Thompson Island
Pool, levels at this gauge are expected to be approximately equal to those at
Node 13 of the two-dimensional model (Node 5 of the one-d1mens1ona1 model). Both
gauges are normally read twice a day from m1d-Apr1l to mid-December and the two
values averaged. (Observations are sometimes omitted during floods.)
Initial calibration, using both the two-dimensional and one dimensional
networks, was undertaken on flows of Spring 1978. As this year was a relatively
low flow year, the model was relatively Insensitive to specification of
calibration parameters. In other words, moderate flows pass through the pool
with little dynamic character; specification of typical Manning roughness
coefficient values of 0.027 for the uppermost, dredged channel reaches and 0.030
for the remaining reaches suffices to give an excellent fit to the staff gauge
data. Additional calibration of the one dimensional model was performed on the
major spring flood of April-Nay 1983. In this year, a dally average flow of
32,600 cfs was reported at Fort Edward on Hay 2, 1983 and a peak flow of 35,200
cfs was reported on May 3, 1983. These flows are on the order of 20-year
recurrence Interval events, but are the largest flood events on record since
monitoring began at Fort Edward 1n December 1976. Inputs to the model were taken
as the dally average flows recorded at Fort Edward, because the full record of
hourly peaks was not readily available. The hydrodynamlc model was run on a six
second time step, with the Inflow rates Interpolated between dally midpoints.
At these high flow rates, model performance 1s more sensitive to the
calibration parameters. Reasonable calibration was obtained by setting the
Manning roughness coefficient to 0.026 1n the uppermost two nodes, 0.027 In the
B.5-17
-------
next three lower nodes, 0.028 in the node below that, and 0.029 in the remaining
nodes down to the Thompson Island Dam. In addition, the weir discharge parameter
was set to 3.36 to fine tune stage elevations 1n the lower end of the pool.
Comparison of predicted and measured stage observations for the flood
period and for the subsequent declining limb of the flood is shown in Figure B.5-
3. A reasonable fit (usually within a tenth of a meter) Is provided, considering
that: 1) only daily average flow and not full hydrographs were used as Input; and
2) there may be noticeable differences in stage between the main channel and
barge canal gauges during extreme floods.
Additional calibration of the hydrodynamic model 1s expected 1n Phase 2,
depending on the data needs of the sediment model.
B.5.4 Sediment Transport Hodel
The sediment transport model STREAM (Borah et al.t 1982a; Borah and
Bordoloi, 1989a, 1989b, 1991) is used here to simulate bed and bank scour,
sediment deposition and resuspension, and sediment transport for the Thompson
Island Pool. Spatial and temporal variations of the flow conditions and
hydraulic parameters are obtained from the output of DYNHY05, described 1n the
previous section, and provide the hydraulic parameters, e.g., flow, velocity,
etc., needed for the sediment model.
B.5.4.1 Streambed Erosion and Deposition
The amount of sediment transported in, deposited in or eroded from an
alluvial stream bed is the result of Imbalances between sediment transport
capacity of the flow and the Incoming sediment. Such an imbalance 1s determined
by considering local conservation of mass. During an erosion condition, particle
entrapment occurs, if the particles on the bed surface are transportable with
the existing flow conditions. Otherwise the particles remain on the bed surface
as part of an armor layer. These processes are simulated using the algorithms
discussed in the subsections below. The sediment 1s divided Into small size
B.5-18
-------
groups, based on the particle size distributions with each group represented by
an average diameter. Each sediment group 1s considered Individually and the
total response 1s determined by adding responses of all the groups.
Sediment Transport Capacity y
The sediment transport capacity of a flow may be expressed as a function
of the flow parameters, such as depth and velocity and the particle size. There
are many sediment transport formulas available today; a review of these is given
by Vanoni (1975) and evaluations of some are made by Alonso ci a7. (1981). Such
formulas are directly applicable when modeling the transports uniform sediment;
they are not directly applicable for simulating transport of nonuniform sediment.
A given flow has a characteristic capacity for transporting different
sediment size groups; sediment transport capacity is calculated-separately for
each particle size group. Thus, as the transport capacity for each size group
1s calculated, transport capacity for the remaining size groups Is reduced. A
0
variable called the residual transport capacity accounts for this Incremental
transport capacity calculation:
T.i - T,a (17)
(W)
the residual transport capacity for size group 1;
the sediment transport capacity for group 1;
the volumetric concentration of sediment group j; and
the total number of sediment size groups considered 1n the
simulation.
where
Tr, -
T, -
C, -
N
B.5-19
HRP 001 0800
-------
The term 0 1n Equation (17) Is used as an Indicator for entra1nment/eros1on
when Q>0, deposition when Q<0 or equilibrium when Q-0.
Active Bed Layer
The materials available for entrapment are essentially those exposed at
the bed surface. The model assumes entrapment of particles starting from the
smallest size group exposed at the surface. Entrapment then continues with
larger sizes, deeper Into the bed until either Q-0 or the bed Is covered with,
non-transportable materials forming an armor layer. This process 1s assumed to
be taking place from an upper bed layer called the active layer. The thickness,
porosity and particle size distribution of this layer can vary throughout the
simulation, but the layer 1s assumed homogeneous at all times. For bed materials
having enough non-transportable sizes to form an armor layer, the thickness of
the active layer Is expressed as:
"
-------
until the flow achieves equilibrium. Under such conditions, the following
expression is used to compute the active layer thickness:
ฆ TOT (20)
in which - the thickness of active layer under non-armoring conditions; and
du - the sediment size under which 85 percent particles are finer.
In general, Equation (19) 1s used only if dL
-------
Ftj ซ 0, for j>i (23)
where
1 - the matrix row representing a sediment size group exposed
the bed surface;
j ป the matrix column representing a sediment group hidden under
group 1; and
Fu - the element of the entrapment frequency matrix representing
probable frequency of entrapment of group j 'after group 1
(1>j) 1s entrained.
Volumes of potential entrapment from different size groups, based on the
entrapment frequency matrix, are computed and arranged 1n the corresponding
elements of another matrix called the volume entrapment matrix. Erosion or
entrapment volumes for different size groups during a time Interval are computed
from this matrix and are expressed as:
Ei * ฃ y" (24)
,._Sh52l
" t ** ซ
Jc-J
VJ " JC viJ " (26>
where
Ej - the erosion volume of sediment group j per unit le""*h during
a time interval;
B.5-22
-------
e ซ sediment erodibility parameter, Osesl, ranging from 0 for non-
erodible material to 1 for detached and easily erodible -
sediments;
v4J - the element of the volume entrainment matrix representing
volume of size group j in the active layer per unit length
which 1s exposed at the surface due to the removal of volume
v of group i;
N - the row number in the volume entrainment matrix v4J for 0-0;
Vj - the volume of group j present in the active layer per unit
length; and
W - the active bed width; and t ป the active layer thickness 1n
Equation (19) or (20).
The computation of entrainment volume 1n Equation (24) 1s accomplished
through a search procedure starting from the first row of the volume entrainment
matrix. Every time entrainment from an element 1s computed, the term Q 1s
updated using Equation (18) after increasing Cj by the eroded volume.
Entrainment computation continues towards higher rows until 0-0 or the bed 1s
armored (T< - 0).
The sediment erodibility parameter (e) represents the resistance to erosion
due to cohesion or other bonding properties (Borah et al., 1982a). While testing
STREAM with Little and Mayer's (1972) experimental data, Borah and Bordoloi
(1989) found this parameter to be 1.0 for noncohesive bed materials. For Hudson
River sediments, it will serve as an Initial calibration parameter for modeling
the effects of cohesive sediments. More detailed cohesive models may be
necessary after further model testing.
Sediment Deposition
Whenever Q<0, deposition 1s assumed and particles beginning with the
largest size group are dropped out. The simulation continues until the flow is
no longer overloaded (Q-0) or all the material from different size groups present
1n the flow are settled. Thus, the deposition volume during a time interval 1s
computed as:
B.5-23
-------
Dj ป ACj, if *Txj\ 2 Cj or Tj ซ 0
(27)
D, - A \ Tzj\,
if \TZ}\ < Cj
(28)
1n which Dj - the deposition volume of size group j per unit length during a time
Interval. The amount of sediment reaching the bed depends on the particle
settling velocity. Therefore, Dj Is adjusted by a correction factor of 2WjAt*/h
(<1) to account for slow deposition of particles with low settling velocities,
where - the average fall velocity of size j; At* - the time Interval computed
in the characteristic solution of sediment continuity equation (discussed later);
and h ฆ the average flow depth.
Bed Elevation Change
The change in bed elevation during a computational time Interval 1s
computed as:
1n which AZ - the bed elevation change and At the computational time Interval
selected for the simulation.
B.5.4.2 Streambank Erosion
The lateral erosion of a cohesive rlverbank 1s given by Arulanandan et al.
(1980) and is expressed as (Osman and Thorne, 1988):
(29)
B.5-24
-------
AW. At
(30)
yxe
R ซ 0.0022 xc e'ฐ'l3Xc (31)
where
AW - the bank erosion distance (1n m) 1n one of the banks;
tc ซ the critical shear stress (1n dynes/cm2) for cohesive soils
with given sodium adsorption ratio, pore fluid salt concentra-
tion and dielectric dispersion;
Y - the soil unit weight (In KN/m1);
t - the average shear stress (1n dynes/cm2); and
At - the computational time Interval (In minutes).
The rate at which the eroded bank material 1s added to the sediment load
1s computed by introducing a new linear relation.
The bank height (H) Is updated by subtracting AZ from Its Initial value.
The bank height above the zone of lateral erosion (H') Is computed by subtracting
AW tan0 from the Initial bank height, where 0 Is the bank angle. Using the ratio
of these heights (H/H'), the critical height ratio for unstable bank 1s computed
as (Osman and Thorne, 1988):
(32)
Xx ซ (1 -K2) (sinp cos0 - cos2P tan0) (33)
B.5-25
-------
X2 = 2(1 -K)
2 1H'
(34)
%
sinfl cosfl tan8 - sin2P
tan8
(35)
f>
(1-K2) tan0
* <|>
(36)
where
H
the
bank height;
H'
the
bank height above zone of lateral erosion;
(H/H')e -
the
critical height ratio for an unstable bank;
8
the
Initial bank angle;
ฆ
the
failure plane angle;
~
the
angle of Internal friction;
o ฆ
the
cohesion;
K
the
ratio of crack depth to bank height (y/H); and
y
the
depth of tension cracking.
If the bank height ratio (H/H'). Is approximately equal to the critical
bank height ratio (H/H')e, bank failure would occur. The failure block width and
volume are computed as (Osman and Thorne, 1988):
BW =
- H ~
tan
y _ H'
p tan8
(37)
B.5-26
-------
(38)
in which BW - the width of failed bank on one side; and VB - the volume of failed
block per unit length on one side.
After the bank has initially failed, the new bank slope becomes 0 and
remains the same for parallel bank retreat. The subsequent bank failure width
and volume are computed using Equations (32) through (38) after substituting 0
ฆ K - 0, y - 0 and X3 - 1,.
The volumes of materials generated from lateral bank erosion (AW)ztan0 and
bank failure VB are not readily available to the entire cross sectional flow.
Therefore, these materials are stored at the bank toes and the volume of lateral
sediment contribution (Equation 44) during a computational time Interval is
computed by multiplying this storage by the fraction 2AW/(W+2AW). This procedure
1s a modification of the original procedure of Osman and Thome (1988), where the
total sediment transport capacity was used to transport these materials.
Sediment Routing
The routing algorithm of nonuniform sediment is primarily based on the
conservation of mass for the sediment load and materials on the bed per size
group. The total sediment load (concentration or discharge) 1s computed by
adding the corresponding values from all the size groups. The continuity
equation for a size group may be written as:
(39)
where
A
the flow area
c
the volumetric concentration of the sediment load
B.5-27
-------
X
ft
t
Q,
q,
the volumetric sediment discharge;
the volumetric rate of lateral sediment Inflow per unit
length;
the volumetric rate of sediment deposition per unit length;
the volumetric rate of sediment entrapment per unit length;
time; and
the longitudinal distance.
Assuming constant flow conditions and constant q ft and q within small
intervals of time and space and assuming sediment moving with the same velocity
as water, Equation (39) is expressed 1n a quasi-linear hyperbolic form, governing
the propagation of sediment load waves, and solved by the method of characteris-
tics which yields (Borah et *7., 1982a):
(40)
(41)
Ej - tjj Af
(42)
Dj ซ dj At*
(43)
2 AW [ (AW) 2tan6 + 2 VB] Pj
(W +2 AVI) At
(44)
B.5-28
-------
In which m - the subscript representing space Increment node; n - the superscript
representing time Interval node, Ax - the space Increment (x^, - xj; At* - the
tine Increment (t - t") given by Equation 45; Q ซ the water discharge flow rate;
and V - the average water velocity.
As Indicated earlier, Equation (44) Is a new addition to the original
procedure. After the first bank failure, the angle 6 1s replaced by 0 (Equation
He) and kept the same afterwards.
In the above solution, Q, V and A are constant within each space element
(Ax) and time Increment (At*). The concentration (Equation 40) occurs at
x^, and t"+At*. Concentration c"*^,.., at x^, and t"*1 - t"+At 1s computed by
Interpolating or extrapolating c"Jt~i (initial value) and cJtIK, between t" and
t"+At*. This concentration is multiplied by Q to compute sediment discharge for
the size group J. Adding sediment discharges for all the groups, the total
sediment discharge 1s computed. From these values, size distribution of the
sediment load 1s computed.
The bed material volume Vj per unit length of size group j in the active
layer 1s updated by adding Dj and subtracting Ej. From all the updated bed
material volumes, the new size distribution of active layer is computed based on
which new active layer thickness 1s computed. During bed erosion, the new
thickness may be deeper than the remaining thickness from the previous time
interval. In such cases, additional parent bed materials from the excess
thickness are added to the current active layer materials and size distribution
of the active layer 1s revised.
The bed elevation within Ax 1s updated by adding AZ in Equation (29) to the
initial value. The bed porosity 1s computed in the model using the following
empirical relation given by Komura and Simmons (1967):
B.5-29
-------
(46)V
1n which dM - the median particle size In cn.
*
B.5.4.3 Initial Calibration Efforts
The Initial calibration efforts for the STREAM sediment transport model are
ongoing. These Initial efforts for the Thompson Island Pool section of the Upper
Hudson are being applied to a one-dimensional segmentation of the river reach,
containing 31 segments 1n the f1ve-m1le river reach. Two scenarios-are being
Investigated: 1) a moderate flow event which occurred 1n mid May 1983; and 2)
a flood event In late April 1983.
In order to complete the sediment model calibration, additional field data
defining suspended sediment loads over the Thompson Island Dam and further bed
sediment data are needed. Continuing calibration efforts and on-going modeling
may be performed 1n subsequent phases, depending on the need.
B.5.5 Summary
During Phase 1, transport modeling efforts have focused on proposing and
selecting an appropriate transport model and defining preliminary hydraulic and
sediment transport scenarios upon which to base the calibration of the models.
At present, the hydraulic model has been calibrated with results showing an
Improvement over previous HEC-6 model efforts by others. The sediment transport
model has been Implemented, but final calibration requires additional data.
0.245 ~ ฐ-0864
<*o>
0.21
B.5-30
-------
SYNOPSIS
PRELIMINARY HUMAN HEALTH RISK ASSESSMENT
(Section B.6)
*
V
This section sets forth the objectives, method and results of a preliminary baseline human
health risk assessment The main objective of the preliminary assessment (B.6.1) is to examine
the quality of the available site data for risk assessment purposes and identify where additional
data are needed to perform a more complete assessment of potential health risks. For the
purposes of this assessment, only risks associated with exposure to PCBs are evaluated. Future
land uses in the area are assumed to be similar to current land uses. ^
There are several potential pathways by which people might be exposed to PCBs
originating from the Hudson River (B.6.2). Potential exposure to PCBs originating from the
Hudson River via dietary intake includes exposure from ingestion of fish, home-garden crops,
beef or dairy products, human breast milk and drinking water. Of these dietary intake pathways,
only potential exposures from ingestion offish and drinking water are quantified Remaining
dietary intake pathways could not be quantified, because insufficient data exist to determine
whether or from what source PCB exposure may occur. While potential exposure as a
consequence of inhalation of PCBs in ambient car is discussed, exposures occurring via this
pathway could not be evaluated quantitatively. Sampling data are too sparse and/or inadequate
to determine: 1) representative PCB concentrations in ambient air; or 2) the contribution of
volatilization of PCBs from the Hudson River as opposed to contributions from other sources.
Recreational exposures include dermal contact with sediments and river water as well as
incidental sediment ingestion during recreational activities, all of which were quantified. The
analysis has revealed that estimated PCB intake through consumption of fish from the Hudson
River appears to be the most significant, potential pathway of human exposures to PCBs from
the site.
A discussion of the current understanding of potential carcinogenic and non-carcinogenic
toxic effects associated with exposure to PCBs (B.6.3) summarizes methods used by USEPA to
derive toxicity values and estimate potential health risks associated with exposures to PCBs.
Quantitative exposure estimates are evaluated in conjunction with the toxicity information
in order to predict the potential for human health effects (B.6.4) associated with exposure to
Hudson River PCBs. Two types of health risk evaluations are presented: non-carcinogenic
health effects and carcinogenic risks. The potential health risks associated with all quantified
pathways other than the ingestion offish are estimated to be within the acceptable range.
Based on available data, there appear to be unacceptable potential cancer and non-
cancer risks associated with regular ingestion offish from the Upper Hudson River. Assuming
consumption of PCB-contaminated fish for 30 years, cancer risks are estimated to be as high
as 2 in an exposal population of 100. With respect to non-cancer risks, the average daily
exposure to PCBs resulting from consumption offish from the Upper river may be as high as
51 times the reference dose. This evaluation shows that the population that regularly consumes
-------
fish from the Upper Hudson River is at risk from PCB exposure. The working assumption is
that people still consume fish, despite the fishing ban. This assumption may require further
quantification, because, as the operative risk, it will be useful to ascertain the effectiveness of
the fishing ban.
-------
B.6 Preliminary Hunan Health Risk Assessment
B.6.1 Phase 1 Objectives
USEPA has prepared this preliminary, baseline human health risk assessment
for several reasons. First, the 1984 ROD did not Include any quantitative
Information concerning the risks to human health from PCBs in the Hudson River
and USEPA would like to Initiate early public review of this complex topic.
Second, USEPA needed an evaluation of those pathways, such as airborne PCBs or
ingestion of food crops or livestock, where sufficient empirical data are not
available, so as to determine an appropriate data acquisition program. Third,
USEP' needed to know whether the historical assumption that fish were the
significant pathway of potential human exposure was empirically verifiable in
order to consider whether institutional control, i.e., a fishing ban, might be
required, regardless of other concurrent remedial alternatives, or might be u^ed
because 1t Is most protective. Further, USEPA wanted to disclose all current
working assumptions used in this risk assessment, because various reviews are
underway that may require subsequent modifications to this baseline assessment.
By identifying these areas of uncertainty now, the final result will be more
easily understood, regardless of whether any adjustments are necessary or not.
This preliminary evaluation is based on appropriate conservative exposure
assumptions (as defined by USE guidance), the most recent toxic potencies
estimated by appropriate USEPA offices and the most current regulatory criteria
and standards of NYSDEC and USEPA. The potential baseline risks addressed are
the current public health risks, which might result from taking no remedial
action, /.e., the no action alternative for the river. Risks associated with any
proposed remedial action for the river or natural biodegradatlon of contaminants
will be assessed in Phase 3 - Feasibility Study.
As PCBs have been Identified as the major sources of health risks
associated with exposure to site related media, this preliminary assessment 1s
limited to evaluating potential health risks associated with PCBs.
B.6-1
-------
Unlike many NPL sites where current use differs significantly from
plausible use 1n the future, 1t 1s assumed here that future use would be the same
as present use along the river and would not alter the potential for exposure^of
area residents. Therefore, no alternate analysis for a hypothetical fututa
exposure scenario 1s provided. Also, this baseline assessment does not consider
the potential occurrence of a major scour event.
The scope of the preliminary evaluation 1s limited to addressing those
risks stemming from PCB contamination that currently exists in the Hudson River
or from PCB contamination that can be directly attributed to the river. For
example, potential risks due to exposures to PCB-contarainated fish, sediments and
surface water are evaluated 1n this assessment, since there is little question
that this contamination can be quantitatively and directly attributed to riverine
levels. Potential risks stemming from exposures to PCB-contam1nate~d soils or
products produced on those soils can not be evaluated, since sufficient data are
not available, the source of the PCBs in soils is unclear and, 1f present, such
PCBs may not be directly attributed to the river.
In. Interpreting the findings of this report, it 1s important to note that
new data regarding PCB concentrations, e.g., in fish or other media, in the river
system will be utilized as they become available. In addition, ongoing studies
regarding the toxicology of PCBs and human risks associated with PCB exposures
will be Incorporated, 1f accepted by USEPA through a scientific review process
before the completion of the Reassessment RI/FS. Subsequent phases of the Hudson
River reassessment process will incorporate such new information and analyses.
B.6.2 Exposure Assessment
r.
B.6.2.1 Introduction
Following Identification of PCBs as the chemical of concern, an examination
of potential exposure routes, potentially exposed populations and exposure
pathways was performed. To the extent data were available, the magnitude of
B.6-2
-------
exposures via each pathway was estimated in a two step process, considering both
contaminant concentration and human exposures.
Contaminant concentrations 1n each of the environmental media of concern
(e.g., water, sediments, air, etc.) are determined at relevant receptor points.
Determination of media of concern is based on analyses of mechanisms of
contaminant release from the site and environmental fate and transport as well
as consideration of locations and mechanisms of human contact with site
contaminants.
As suggested 1n USEPA's recent Risk Assessment Guidance for Superfund
(1989b), "reasonable maximum" individual exposure concentrations are calculated
to the extent appropriate. Geographic variations in environmental concentrations
are considered in determining appropriate exposure point concentrations.
Duration of exposure and the likelihood of exposure pathways occurring are also
evaluated.
Human exposures to PCBs are quantified using the environmental
concentrations together with estimates of media intake. These scenarios, under
which exposures are evaluated, include assumptions regarding physiological
parameters, such as body weights, media Intake rates, such as soil Ingestion
rates, and activity patterns, such as frequency of contact at the site. In soae
instances, standard exposure assumptions are included in the assessment. For
example, throughout this assessment, a 30-year duration of residence in the
Hudson River area, and a 70-year lifespan are Incorporated Into exposure
calculations, based on recent USEPA guidance (USEPA, 1989a). Similarly, a
lifetime average body weight of 70 kg Is assumed. In other cases, assumptions
are tailored to s1te-spec1f1c conditions as appropriate.
The population of concern 1n the evaluation of the Upper Hudson River
consists of the inhabitants of the towns, cities, and rural areas surrounding the
River. Exposure by these populations to PCBs present along the river and to PCBs
that have migrated from the River could occur via a number of potential pathways,
as Illustrated in Figure B.6-1.
B.6-3
-------
Several gaps 1n the available Information emerged during the process of
quantifying an exposure dose and precluded a thorough, quantitative exposure
assessment for some exposure pathways. Concentration data for PCBs 1n some of
the media of concern were either non-existent, out of date, or of questionable
applicability. In addition to data limitations, several pathways cannot be
quantitatively assessed, because they are not considered complete at this time.
Rather than calculating exposures (and associated risks) from data of
questionable relevance, 1t 1s considered more appropriate to point out the
limitations and suggest possible means of acquiring better, more relevant data.
The potential exposure pathways considered In this analysis and the type of
evaluation performed for each pathway are summarized in the tabulation on the
following page.
B.6.2.2 Dietary Intake
F1sh Consumption
Because fish effectively bioaccumulate PCBs, fish provide a pathway,
frequently the predominant pathway, for human exposures to PCBs. Studies
conducted on Michigan residents established that those who regularly ate Lake
Michigan fish had serum PCB levels up to 30 times greater than those who did not
eat these fish (Humphrey, 1987). Data on PCBs in Hudson River fish, discussed
in Section B.3, clearly indicate that fish consumption can result 1n human
exposures to PCBs.
Recent studies (NYSDEC, 1990) Indicate that the Hudson River continues to
draw a significant number of anglers. Estimates are that 26,870 (ฑ3,440)
Individuals fish the Hudson River for a total of 232,110 (ฑ51,310) angler days.
Over 38 percent of these Individuals claim to fish the Upper Hudson (section
north of Federal Dam at Troy) for an estimated 87,060 (ฑ22,090) angler days along
B.6-4
-------
Summary of Potential Upper Hudson Exposure Pathways
&poซn tahway/Medtam
Potentialy fitpoaed PopuUon
Type of Evaluation
Rth Inoestion
Adult looal fishing population -
Upper Hudson from Albany to
Port Edward
Reasonable maximum exaqsyre - Two
Soenarios:
(1) Reoent fish data 1969-1988
(2) 30-year projections based on current trends
60" percentile fish ingestion rats for local
anglers ("50 half-pound servings per year).
ฆftfriMnaYitoir
Adult residents using the Hver
as a potential regular source of
drinking water
Screening tvoe assessment: Possible future drinkina
water use near Fort Edward with no PC8 removal in
water treatment
Dermal PCB Absonrtion from
Sediments
Adults and children during
summertime recreational
activities on the River -
Thompson Island Pool
Wsdlno: Hands and feet expoeed.
Dermal PCB Absonrtion from
Water
Adults and children during
summertime recreational
activities on the Hver-
Thompson Island Pool
Swimmino: FuH bodv exoosed.
trwestion of PCBs from
Sediments
Adults and children during
summertime recreational
activities on the River -
Thompson Wand Pool
Incidental Uptake: Hand to mouth activity
Pathways Not Evaluated Due to Inadequate Data
Inhalation of PCBa In Air
Limited recent air measurements are available. It is difficult to determined whether PCBe
in air are from the river, dump sites, or other (unknown) sources.
hwestion of PCB* In Local
SOttgiOfl Tufttff
A few snapping turtles in the Upper Hudson have been found to contain PCS*. It is
uncertain whether the small sample is representative and further unknown whether turtles
are caught and oonsumed by local residents.
Inoestion of Garden or
Aartcultural Croos
Historical studies have shown some PC8 contamination In plants (e.g., frees and pasture)
near the river, reoent data are lacking. It Is difficult to determine whether PCBe from the
river are the major or only source of possible crop contamination.
Inoestion of Milk or Meat
Npfttf areavfiltble.
Inoestion of Breast Milk
No data are available.
B.6-5
-------
this portion of the river. (The source of this Information does not specifically
indicate the proportion of days 1n the Fort Edward to Federal Dam reach.) From
the NYSOEC study, 1t appears that local populations do most of the fishing along
the river and the average distance traveled per fishing trip 1s approximately i*
miles.
NYSDOH has Issued a Health Advisory that recommends against consumption of
any fish taken from the Upper Hudson River (Hudson Falls to Federal Oam) and
limited Intake of many species from the Lower Hudson. It 1s, however, USEPA
policy not to assume that fishing bans or other similar types of Institutional
controls have any significant long-term effectiveness In reducing the Intake of
contaminated fish. Therefore, this baseline risk assessment evaluates the
possible risks that would be associated with consumption of fish and the
potential risk that would arise from consumption of fish at a rate that might be
expected to occur 1n the absence of or despite the Health Advisory. This
approach is prudent 1n light of information from the NYSDEC angler survey
indicating that Individuals may be unaware that they are consuming potentially
contaminated fish, even though they are aware of fishing advisories (NYSDEC,
1990). Or individuals may feel that their fish are safe to eat, even though the
water from which the fish come are known to be polluted (Belton et al., 1986).
Specific examples of disregard for fishing advisories have been addressed
in a study conducted by the New Jersey Department of Environmental Protection
(NJDEP) and Rutgers University (Belton, 1986). In this survey of anglers along
the Lower Hudson River, New York Bay, and Newark Bay, observations by NJDEP
personnel Indicated that fishing advisories were being ignored. While half of
the anglers surveyed claimed knowledge of warnings, two-thirds of those who
admitted to eating their catch considered them to be totally safe to eat, while
others considered them to be polluted but not harmful. The fishermen were found
to hold several beliefs that made them resistant to advisories. For example,
some stated that fish were able to avoid pollution through a self-cleansing
mechanism; other fishermen felt that they could determine whether a fish was safe
to eat based on the physical appearance, smell, taste, or behavior of the fish.
B.6-6
-------
Almost half of the respondents Indicated that the fish must be safe to eat, "if
you ate them and had no reaction within a day or two."
Given the presumption that people do fish 1n the Upper Hudson below Hudson
Falls, a key Ingredient 1n estimating exposure via this route 1s the mount of
fish caught and eaten by local anglers. USEPA (1989a) suggests use of the Intake
rate of 30 g/day or 140 g/day for the 50th and 90th percentile values,
respectively, of annual average dally Intake. These numbers represent values
from studies by fuffer (1981) and Pierce (1981), which Investigated the
consumption of fish by Individuals who were known to fish regularly from piers
1n southern California and Puget Sound. While these study locations are quite
distinct from the Upper Hudson River, the values generated 1n these studies are
considered by USEPA to be more representative of actual annual consumption rates
for recreational anglers (USEPA, 1989a) than other values available 1n the
literature. Other values are usually based on the amount of fish that enter
-------
of the days spent fishing on the Hudson River (Upper and Lower) are spent fishing
for bass. Another 6.5 percent of days 1s spent fishing for brown trout. Over
22 percent of the days are spent in pursuit of "no specific type" of fish, and
33 percent of days are spent for "other" species. No information specific to the
Upper Hudson 1s available. According to data regarding species preferences on
a statewide basis, bass are the fish most frequently sought after, followed
sequentially by brown trout, rainbow/steel head trout, yellow perch, and walleye
(yellow pike). The statewide values appear to echo the data specific to the
Hudson.
The only available PCB data for the species preferred by fishermen are for
PCBs in bass. Since a large percentage of the Individuals polled in the NYSDEC
survey Indicated that they fished for other species or no specific type of
species, all species sampled were used to provide the best characterization of
the exposure concentration. Additionally, 1n evaluating the data it became clear
that any differences in concentrations among species were not of sufficient
magnitude to necessitate treating each species separately.
PCB concentrations in fish fillet or eviscerated fish were used according
to available data. The fillet represents the part of the fish most commonly
consumed. Although concentrations of some organochlorlne compounds in fish may
decrease while cooking (Stachiw, 1988), the data for PCBs are not consistent.
One study reported a small decrease in PCB concentrations with cooking (Zabik,
1982), while another study reported a wide range of PCB decrease after cooking
(Cordle, 1982). Because, no specific value can be derived based on the available
data and no information on PCB concentrations after cooking the species of
concern from the Hudson are available, no adjustments were made.
Because of the uncertainty in predicting future trends from current levels
of PCBs in fish, potential exposures from PCBs in fish are evaluated for two
scenarios: 1) using the most recent concentration data, /.e., 1986-1988, to
represent appropriate exposure concentrations, and 2) using a 30-year average
concentration, estimated by assuming that the observed exponential rate of
decline in tissue concentrations of PCBs continues from 1991 to 2020.
B.6-8
-------
For the first scenario, the exposure concentration for PCBs in fish is the
95th percent confidence limit of the arithmetic average PCB concentration in fish
sampled from 1986 to 1988, across all species sampled from River Miles 153 to
195. This 95th percent upper confidence limit of 12.0 ppm (12.0 mg/|cg) was
determined to represent a reasonable upper bound estimate of exposure
concentration for several reasons. First, guidelines for assessing baseline
human health risks for Superfund sites state that for "any estimate of exposure
concentrations, the upper confidence limit (/.e., the 95 percent upper confidence
limit) on the arithmetic average will be used for this variable" (USEPA, 1989b).
Second, fish tend to migrate at least over a limited distance and thus, exposure
concentrations would be Integrated over distance. Third, anglers fishing the
Hudson River over the 30-year exposure duration would typically fish more than
one location over time. Taken together, the second and third points indicate
that It would noi be reasonable to assume that either the most contaminated fish,
nor fish from the most contaminated location are consumed continuously over a
long period of time.
For the second scenario, the exposure concentration Is based, on future
average concentrations of PCBs in fish over the period of the next 30 years, as
calculated from the available data on largemouth bass and brown bullhead. A
value of 1.5 ppm 1s considered predictive of the 95th percent confidence interval
of the 30-year mean, barring any unforeseen redistribution of PCBs within the
rlvfer. The full derivation of this value 1s described at B.4.4. The validity
of the 30-year concentration value should be assessed as new Information on
potential resuspension and redistribution of PCBs in sediments 1s developed.
In summary, exposure to PCBs from ingestion of contaminated fish is
calculated assuming an annual average dally Intake rate of fish of 30 g/d and
exposure concentrations of PCBs 1n fish of either 12.0 ppm or 1.5 ppm, considered
to reasonably approximate the range of concentrations. Fish consumption 1s
assumed to occur each year, over a 30-year residence time 1n the vicinity of the
river. Also assumed are a lifetime average bodyweight of 70 kg and 100 percent
gastrointestinal absorption of PCBs from the fish.
B.6-9
-------
These exposure assumptions are summarized 1n Table B.6-1. The calculation
used to estimate exposure to PCBs from Ingestion of fish Is:
Human Intake
of PCBs from - FIR x FC x GI x EF x ED x CF
fish Ingestion BW x AT
(mg/kg-day)
where
FIR - F1sh ingestion rate, 30 g/day (USEPA, 1989a)
FC - F1sh concentration of PCBs (1.5 mg/kg and 12.0 mg/kg)
GI - Gastrointestinal absorption, 100%
EF - Exposure frequency, 365 days/year (Ingestion rate 1s annual
dally average)
ED - Exposure duration (30 years; national upper-bound time at one
residence; USEPA, 1989b)
CF - Conversion factor (0.001 kg flsh/g fish)
BU - Body weight (70 kg)
AT - Averaging t1me(365 days/year x 70 years for carcinogenic
effects; 365 days/year x 30 years for non-cancer effects)
Under these assumptions, annual average dally exposure over the assumed 30-
year exposure duration 1s calculated to be 5.1 x 10"J mg/kg-d, assuming current
average concentrations of PCBs In fish of 12.0 ppm, and 6.4 x 10'4 mg/kg-d, using
the estimated future average concentration of 1.5 ppm. Averaged over a 70-year
lifetime, the chronic dally Intake (COI) corresponding to the two exposure
concentrations 1s 2.2 x 10'3 mg/kg-d or 2.8 x 10'4 mg/kg-d, respectively. The
average dally exposures and chronic dally Intakes for this pathway are listed 1n
Tables B.6-5 and B.6-6.
Snapping Turtles
Available literature Indicates that snapping turtles exhibit an exceptional
ability to bloaccumulate PCBs from their environment (Stone et a/., 1980). Since
these animals may be consumed by some human populations, they can constitute a
possible exposure pathway.
B.6-10
-------
Neither recent data showing specific concentrations of PCBs in Hudson River
turtles nor intake rates, as in turtle soup, are available. Therefore, no
exposure calculation is made.
Ingestion of Agricultural or Home Garden Crops
Information available in the literature indicates that plants may become
contaminated with PCBs. It has been established that some plants may become
contaminated with PCBs by root uptake from contaminated soil or water or by
uptake of volatilized PCBs from air (see B.3). It is possible that vegetables
from vegetable gardens or farms 1n the vicinity of the Hudson River may contain
PCBs and pose a route for human exposure. There are, however, no data on soil
concentrations of PCBs in this study area for this computation. Water used for
crops or backyard gardens 1s not believed to be a significant concern, given the
levels of PCBs in the water. Thus, the data are Insufficient to predict tyie
amount of plant uptake of PCBs from soil or water. Other data Indicate that air
In the Hudson River area may contain PCBs and that uptake of PCBs from air could
also occur, but a Hudson River source of the air-borne PCBs cannot be Isolated
from other sources.
There are many uncertainties in the estimation of human health risks
related to the Ingestion of PCB-contam1nated vegetables. Uncertainties
associated specifically with this pathway of Ingestion Include, first, that the
background data from the area on the concentration of PCBs in crops, collected
in the 1970s and early 1980s, are not necessarily representative of current
conditions. Some reports indicate that the fruits of plants (corn and beans)
contain lower concentrations of PCBs than the leaves by an order of 100-fold
(Shane and Bush, 1989). In addition, different species of plants may be more
efficient 1n eliminating PCBs congeners (Shane and Bush, 1989). PCB
concentrations 1n plants are declining over time (Buckley, 1983). For example,
the levels of PCBs measured in aspen decreased about seven-fold over the period
of 1978 to 1981, while sumac PCB concentrations declined about two-fold (Buckley
and Tofflemlre, 1983).
B.6-11
-------
In the absence of data demonstrating that the Hudson River site contributes
to the concentration of PCBs in local crop species, any quantitative estimate of
exposures from this pathway would be speculative and, therefore, have not been
made.
Ingestion of Beef or Dairy Products
Should locally raised livestock consume PCB-contaminated feedstock or
water, they would tend to accumulate the PCBs in their adipose tissue and could
present a pathway for human exposure, 1f the meat or milk from these animals were
consumed. In the Hudson River area, water fed to animals is, for the most part,
from streams or wells on individual properties, which are not known to be
contaminated with PCBs. In addition, no data regarding the concentration of PCBs
in soils in the vicinity are available.
Many uncertainties exist 1n the analysis of risks related to these exposure
pathways. As described above, PCB concentrations in plants are apparently
declining over time; thus, the use of available data (concentrations measured in
1979) 1s likely to overestimate potential exposures. Also lacking are data on
the proportionate consumption of supplemental feed versus feed grown locally,
especially in the case of dairy cows, where the amount of supplemental feed is
probably significant (Fries, 1982).
Breast H1lk
Because human breast milk is high in lipid content, lipophilic PCBs may
also be present 1n breast milk and constitute a source of exposure for a nursing
Infant. Some researchers (Wolff, 1983; Yakushijl, 1978; Brilliant, 1978) have
suggested that the concentrations of organochlorine residues (such as PCBs) in
breast milk are close to those theoretically predicted by diffusion directly from
adipose tissue. That 1s, the level in milk fat 1s practically the same as that
in adipose tissue, with a ratio of adipose concentration to milk fat
concentration of approximately 1.1:1. This transfer of contaminants from adipose
tissue to milk fat, which occurs indirectly via the blood, 1s very efficient
B.6-12
-------
because blood flow to both mammary tissue and adipose tissue greatly exceeds silk
production. Therefore, adipose and milk fat attain equilibrium concentrations
of dissolved organochlorine residues (Wolff, 1983).
Studies conducted on fish-consuming populations have specifically
correlated fish consumption with PCB levels 1n human breast milk (Schwartz, 1983)
or in human adipose tissue (Rogan, 1986), and indicate that breast milk from
women who regularly consume fish from contaminated waters may contain PCBs at an
average concentration as high as 1.8 tig PCB/g milk fat (54 ng/g whole milk)
" (Rogan, 1986). There is limited evidence that the specific PCB congeners that
bloaccumulate in fish may be detected at elevated levels in the breast milk of
women who eat those fish (Bush, 1991).
One study, conducted on breast milk from residents of two communities of
Upstate New York, found that there was no difference 1n the congener profile of
breast milk collected from a population of Oswego county (near the Great Lakes
region) than in breast milk collected from Albany (along the Hudson River), a
finding that makes determination of the source of exposure difficult to
distinguish. This study also Indicates that the mean concentration of PCBs in
breast milk among women resident 1n Upstate New York is in the range of 26.5 ng/g
(whole milk), a concentration that could result in a young child approaching an
unsafe level of Intake (Cordle, 1982).
No studies have been conducted specifically to assess the PCB
concentrations present in breast milk among the residents along the Hudson River.
Similarly, there are no known human biomonltorlng data for other tissue types for
this population that might allow extrapolation to or modeling of breast nilk
concentrations. Consequently, It is Impossible to provide an adequate evaluation
of the potential risks that might be presented to nursing Infants via this route
of exposure.
A NYSDOH study Is characterizing the PCB level 1n breast milk and other
human tissues among residents of the St. Regis Reservation in Massena, NY (Bush,
1991), because of concerns relating to exposure from the St. Lawrence River.
B.6-13
-------
Human biological samples are being collected from this population as well as
populations along the Great Lakes and control populations from Warren, Schoharie,
Rensselaer, and Oswego counties. Information from this study could provide
useful Information regarding the background concentrations of PCBs 1n human
tissues as well as PCB concentrations In other populations of concern, such as
residents of the Indian reservation and residents of northern New York state.
Results of this NYSDOH study are expected to be released 1n the summer or fall
of 1991.
Drinking Hater
Data regarding the concentrations of PCBs In water from the Upper Hudson
River are available for five locations: Fort Edward, Fort Miller, Schuylervllle,
Stillwater, and Waterford. The concentrations of PCBs 1n these water samples
have been consistently low. Specifically, based on data collected from 1986 to
1989, the adjusted mean concentrations from these different locations does not
exceed 0.05 ppb. The highest concentrations were detected 1n Fort Edward, where
the upper 95th percent confidence Interval on the adjusted mean 1s 0.06 ppb;
almost one full order of magnitude below the current Maximum Contaminant Level
(MCL) of 0.5 ppb (see B.6.3.5). Consistently, the lowest measured levels were
found at the Waterford sampling station, where the upper 95th percent confidence
interval on the adjusted mean 1s 0.034 ppb.
The only populations between Fort Edward and Poughkeepsle known to use the
Hudson River as a source of municipal water are the populations of Waterford and
Halfmoon, which are served by the Waterford Water Works (Berger, personal
communication). Recently, an extensive study of the quality of the water at this
Intake found that the PCB concentration 1n Hudson River water and treated water
has been below the 0.1 ppb detection limit since September 1983 (Metcalf and
Eddy, 1990). Thus, these measurements are below the federal MCL of 0.5 ppb and
the NYSDOH action level of 0.1 ppb.
B.6-14
-------
The USGS monitoring station at Waterford has detected PCBs above their 0.0]
ppb detection limit. In recent years, the concentration of PCBs measured by USGJ
at this station has averaged on the order of 0.03 ppb. Despite the consistently
low levels of PCBs in water samples collected in the recent past and the facl
that PCB concentrations fall below regulatory limits for PCBs in water, ar
initial screening level exposure and risk estimate was calculated for drinkinc
water. In this screening step, the exposure scenario utilized the upper 9!
percent confidence limit on the mean concentration from recent (1986-1989) river
water samples at Rogr Island as the exposure concentration (0.06 mg/1). The
exposure evaluation assumes a dally intake of 2 liters of water for 30 years.
Exposure by a 70 kg adult was assumed. The exposed population was assumed to
drink untreated water from the Fort Edward area, where the highest concentrations
of PCBs were consistently detected. PCBs taken 1n via this pathway are assumed
to be 100 percent absorbed. All the exposure assumptions employed in evaluating
exposure through drinking water ingestion are consistent with EPA guidance'or
evaluating exposures using reasonable maximum assumptions.
The annual average dally exposure to PCBs from this exposure scenario 1s
estimated to be 1.7 x 10"' (mg/kg-day) for adults. The chronic dally Intake over
a 70-year lifetime 1s calculated to be 7.3 x 10*7 (mg/kg-day). Any drinking
water exposures that may occur are likely to be lower than the screening level
exposures presented here due to water treatment.
B.6.2.3 Inhalation Exposures
Exposure from Air
A1r monitoring (see B.3) for PCBs has been conducted 1n the Upper Hudsor
River study area from the late 1970s to as recently as 1989. None of the
sampling efforts to date, however, has adequately characterized the contributor
of the Hudson River to PCBs 1n ambient air of resident populations. The currenl
database raises concerns about the possibility of inhalation exposure to PCBs,
but the inadequacy of the data proscribes any sound quantitative evaluation o1
inhalation exposure.
B.6-15
-------
B.6.2.4 Recreational Exposures
Direct contact with contaminated sediments 1s a possible route of
contaminant exposure, particularly where such sediments occur close to
residential and recreational areas. Routes of exposure during a visit to a
contaminated site Include absorption following dermal contact with sediments, the
Incidental Ingestion of sediments during subsequent hand to mouth contact and
denial contact with water. Under a recreational use scenario, considered the
predominant use along the Hudson, such exposures are likely to occur occasionally
during the course of a resident's life.
Deraal Absorption From Contact With Sediments
Individuals wading along the Hudson River may be exposed to PCBs by dermal
contact with contaminated sediments. The dose of PCBs absorbed through the skin
depends on many factors, including the skin surface area exposed to sediments,
adherence of sediments to the skin, the frequency and duration of exposure, the
concentration of PCBs In the sediments, and the amount of PCBs transferred from
the sediment through the skin. The derivation of the values used for each of
these exposure factors is provided below and summarized in Table B.6-2.
Contact with river sediments is most likely to occur during recreational
activities. 'Because activity patterns change with the age of the population,
exposure by young children (ages 1-6), older children and teenagers (ages 7-18),
and adults (age 18 and above) were considered separately. It is assumed that
infants under one year of age would not come into direct contact with river
sediments. For children aged 1-6 and for adults, it is assumed that use of the
river occurs approximately seven days per year (USEPA, 1988). For older children
and teenagers, use of the river 1s estimated to occur 24 days per year, based on
the assumption of use twice per week for the three summer months. Age-spec1f1c
body weights are also used. The skin surface area available for contact with
sediments depends on clothing. For children between the ages of 1-18, 1t Is
assumed that legs, feet, arms, and hands are available for sediment contact.
These areas correspond to roughly 55 percent of the total body surface area for
B.6-16
-------
children. For adult exposures, 1t Is assumed that lower legs, feet, forearms,
and hands are available for sediment contact. Following USEPA guidance (USEPA,
1969b), 50th percentile body surface areas were used to estimate dermal contact
rates. The specific surface area values, found 1n the USEPA's Exposure Factors
Handbook (USEPA, 1989a), are presented 1n Table B.6-2.
The rate of adherence of sediments to skin and the absorption of PCBs froa
the sediment across the skin are based on a review of the literature. The values
of 1 mg/cm2 and 3 percent are used for these parameters, respectively.
Data regarding so1l-to-sk1n adherence factors (AF) are limited. USEPA
(1989b) currently recommends default values of 1.45 mg/cm* (potting soil) and
2.77 mg/cm2 (kaolin clay) based on experimental values for soil-related dust
adherence reported by the Toxic Substance Control Commission of the State of
Michigan (Harger, 1979). However, data presented by other Investigators suggest
that the average amount of soil adhering to human skin may be as much as 10-fold
lower. AF values ranging between 0.2 and 0.7 mg/cm2 have been reported by Lepov
et <7. (1975), Roels et <7. (1980), Que Hee et ซ7. (1985), and USEPA (1989c),
wherein USEPA calculated an upper-bound (95th percentile) so11-to-sk1n adherence
factor of 1.0 mg/cm2. Based on data from these four studies, 1.0 mg/aR2 was
chosen to be the most appropriate AF for Hudson River sediments. It was selected
because 1t 1s an upper-bound adherence value derived from studies of diverse soil
types, a condition likely to be representative of areas surrounding the Hudson
River, and Is based on the most recent USEPA evaluation of this Issue (USEPA,
1989c).
Data regarding absorption (AB) of PCBs through the skin are limited. PCB
absorption (bioavailability) can be estimated by comparison to absorption values
for 2,3,7,8-tetrachlorodlbenzo-p-dloxln (TCDD), since both chemicals have similar
chemical properties (Oak Ridge National Laboratory, 1989). Based on a study of
dermal absorption of TC00 by Polger and Schlatter (1980) 1n rats, USEPA (Schatm,
1984) recommended using three percent as an upper-bound AB for PCBs 1n humans.
This value 1s consistent with recent guidance from USEPA Indicating derail
absorption rates of between one and ten percent and Is supported by data
B.6-17
-------
presented 1n other Investigations (Shu, 1988; USEPA 1989c) together with the
slallarlty of the chemical properties of TCDD and PCBs.
Because of the spatially and temporally disjointed nature of the available
data on sediment concentrations of PCBs 1n the Hudson River, choosing a best
exposure concentration for use In the risk assessment 1s difficult. The most
recent data for surface sediments would appear to be reasonable, but few recent
data are available. GE conducted a limited sampling effort 1n 1989 where three
to eight cores from previously Identified hotspots were obtained (see B.3).
Their results Indicate depth-averaged PCB levels of <1 ppm to 918 ppm. A review
of the 1984 Thompson Island Pool survey, the most recent large-scale monitoring
effort, reveals that approximately 80-85 percent of the samples 1n the Thompson
Island Pool had PCB levels less than 100 ppm. The highest concentrations
measured exceed 1,000 mg/kg in the Thompson Island Pool. Taking all surface
(Including all grab and top core section samples - 569 samples) sediment samples
from the 1984 Thompson Island Pool survey, the 95 percent upper confidence bound
on the mean Is 66.2 mg/kg. This estimate provides a reasonable maximum exposure
level for sediment-bound PCBs. Subsequent analyses 1n Phase 2 and 3 may focus
attention on evaluating sediment exposures as a function of river reach and areas
of particularly high PCB concentrations.
The dose of PCBs absorbed through the skin was calculated as:
Absorbed Dose (mg/kg/day) - CS x CF x SA x AF x AB x EF x ED
BU x AT
where
CS - PCB Concentration In Sediments (mg/kg)
CF - Conversion Factor (10** kg/mg)
SA ฆ Skin Surface Area Available for Contact (3,931, 7,420 and 5,170 cm2
for ages 1-6, 7-18 and adults, respectively)
AF - Sed1ment-to-Sk1n Adherence Factor (1.0 mg/cm2)
AB - Absorption Factor (0.03)
EF - Exposure Frequency (7 days/year for ages 1-6 and adults; 24 days/yr
ages 7-18)
ED - Exposure Duration (30 years)
BU - Body Weight (15, 42 and 70 kg for ages 1-6, 7-18 and adults,
respectively)
AT - Averaging Time (70 years x 365 days/year).
B.6-18
-------
The annual average dally exposure to PCBs from this pathway Is estimated
to be 1.0 x 10*', 2.3 x 10"*, and 2.8 x 10** mg/kg-d for young children, older
children, and adults, respectively. The chronic dally Intake over a 70-year
lifetime 1s calculated to be 5.3 x 10"* mg/kg-d. These values are listed 1n
Tables B.6-5 and B.6-6.
Incidental Ingestion of River Sedlnents
Contaminated sediments can adhere to the skin of Individuals swimming or
wading 1n the Hudson River. Subsequent hand-to-mouth activities, particularly
prevalent 1n young children, will lead to the Ingestion of adhered sediments and
contaminating PCBs. The extent of human exposure that will result from sediment
Ingestion depends on the concentration of PCBs 1n sediment, the bioavailability
of PCBs from the sediment, sediment Ingestion rates, exposure frequency and
i
duration, and body weight. The specific values used to calculate exposures via
Ingestion of contaminated sediment are presented In Table B.6-3.
All of the values used 1n calculating sediment Ingestion exposures are
discussed above for dermal exposure to sediment, with the exception of the
sediment Intake rate. Based on guidance from USEPA (Porter, 1989), dally
sediment Ingestion rates are 200 mg/day for children 1 to 6 years of age and 100
mg/day for Individuals 7 or more years of age. In both cases, 100 percent of the
PCBs In the sediments are assumed to be absorbed. Although the sedlaent
Ingestion rates are expressed as a dally rate, most soil and sediment Ingestion
1s expected to occur during outdoor play. This exposure assessment assises that
the entire dally sediment intake occurs at the Hudson River.
The Intake of PCBs due to the Incidental Ingestion of sediments was
calculated as follows (USEPA, 1989b):
Intake (mg/kg/day) - CS x CF X IR X EF X ED
BW x AT
B.6-19
-------
where
CS - PCB Concentration 1n Sediments (mg/kg)
CF - Conversion Factor (10** kg/mg)
IR - Sediment Ingestion Rate (200 and 100 mg/day for ages 1-6 and >age 6,
respectively)
EF ป Exposure Frequency (7 days/yr for ages 1-6 and adults; 24 days/yr
for ages 7-18)
ED - Exposure Duration (30 years)
BW - Body Weight (15, 42 and 70 kg for ages 1-6, 7-18 and adults,
respectively)
AT - Averaging Time (70 years x 365 days/year).
Annual average dally exposure to PCBs resulting from the Ingestion of
sediments Is calculated to be 1.7 x 10'* mg/kg-d, 1.0 x 10'* mg/kg-d, and 1.8 x
10"* mg/kg-d for young children, older children, and adults, respectively. Over
the assumed 30-year duration of residence near the Hudson River and a 70-year
lifetime, the chronic dally Intake (CDI) of PCBs from sediments 1s calculated to
be 3.5 x 10"* mg/kg-d. These values are listed 1n Tables B.6-5 and B.6-6.
Dermal Absorption from Water
Another possible exposure pathway 1s dermal absorption of PCBs directly
from water while swimming or wading In the Hudson River. The amount of exposure
(dose) from swimming activities 1s dependent on the concentration of PCBs 1n the
water, the frequency and duration of human contact with the water, the skin
surface area available for contact, the amount of PCBs that cross the skin
(permeability) and the body weight of the exposed Individual. The values used
for these exposure conditions are listed In Table B.6-4 and addressed below.
Exposures from swimming are assumed to occur during recreational use of the
river as described for exposure from dermal contact and Ingestion of sediment,
except that the entire surface of the body 1s assumed to contact the water. The
values used for bodywelght, exposure frequency and exposure duration are age-
speclflc and are the same as those described above. The skin permeability
constant of 3.2 x 10'* cm/hr 1s derived from the literature. The duration of a
swimming event 1s assumed to be 2.6 hours per day (USEPA, 1989a). The dermal
B.6-20
-------
permeability constant for PCBs Is unknown, but 1s estimated to be 3.2 x 10**
an/hr based on the considerations described below.
Dermal permeability has been found to be well described by Flck's
first law, which states that the steady state flux across the skin
Is proportional to the concentration gradient, the permeaBlllty
coefficient, and the reciprocal of skin thickness (Scheuplein,
1977). The concentration of permeating chemicals within the body is
conservatively assumed to be zero so that the full external
concentration of the chemical contaminant appears as the
concentration gradient. Flck's second law predicts that a lag time
ranging from minutes to hours occurs before th:? flux through the
skin builds up to Its steady state value. To be health protective
In calculations, It 1s assumed that the full-steady state flux
exists Instantly at the Initiation of expsoure.
The permeability coefficient 1s a crucial factor controlling uptake
of chemicals v-ia dermal exposure. For pure liquids of a given
family (e.g., the alcohols, methanol, ethanol,' propanol, and
decanol), the permeability coefficient goes down tilth Increasing
molecular weight. For chemicals presented In solution, the
characteristics of solute, solvent and skin are all Important in
determining the magnitude of the permeability coefficient. The
stratum corneum Is lipophilic (1.e.t hydrophobic), so that, for
chemicals presented 1n aqueous solution, 1t 1s experimentally found
that over a range of octanol-water partition coefficients (K^'s),
the chemical-specific dermal permeability coefficient (K,), 1s
directly related to that chemical's K^, (Flynn, 1990).
Hawker and Connell (1988) have reported that PCBs have log values
ranging between 4.5 (for monochloroblphenyls; MW-189) and 8.2 (for
decachlorobiphenyl; MW - 499). Flynn (1990), who has tabulated and
modeled the permeability coefficients of a wide variety of
substances, predicts a value of K, 1.6 x 10"* cm/hr for all PCBs.
Flynn also suggests that the fitted curve be given a bias moving it
upward 1n permeability so that 85 percent of all the values fall
below 1t 1n order to derive an upper bound estimate of dermal
permeability. This process assigns a value of K, ซ 3.2 x 10** to the
skin permeability coefficient for PCBs. This value 1s unlikely to
underestimate the health risks posed by PCBs in water and, hence, K,
- 3.2 x 10'* cm/hr 1s used as the PCB-spec1f1c dermal permeability
constant for the calculations below. (This value 1s within the
range of those values presented in USEPA's recently released Interim
Guidance for Dermal Exposure Assessment.)
B.6-21
-------
Concentrations of PCBs 1n Hudson River water have been evaluated and
discussed 1n Section B.3. For reasons described In that section and to be
consistent with current USEPA guidance on determination of exposure concentration
(USEPA 1989b), this exposure assessment uses the 95th percent confidence 11m1t
value of the adjusted log normal maximum likelihood estimate of the mean value.
Since the concentration of PCBs 1n water at Fort Edward Is consistently higher
1 than for other sampled locations, data from that location were selected for use
1n the exposure assessment. Specifically, the exposure concentration of 0.06
pg/1 Is Incorporated into the exposure calculations.
The dose of PCBs absorbed through the skin from direct contact with Hudson
River water Is calculated as follows (USEPA, 1989b):
CW PCB Concentration In Water (0.06 ug/1)
CF Conversion Factor (10"* 1/cnr)
SA - Skin Surface Area Available for Contact (100%, or 6,880, 13,100 and
18,150 cm2 for ages 1-6, 7-18 and adults, respectively)
K- Chemical-Specif1c Dermal Permeability Constant (3.2 x 10 cm/hr)
DE ซ Duration of Event (2.6 hr/day)
EF ซ Exposure Frequency (7 days/year for ages 1-6 and adults; 24
days/year for ages 7-18)
ED - Exposure Duration (30 years)
BW ฆ Body Weight (15, 42, and 70 kg for ages l-ป6, 7-18 and adults,
respectively)
AT - Averaging Time (70 years x 365 days/year).
Under these assumptions, annual average dally exposure to PCBs resulting
from dermal absorption is calculated to be 4.4 x 10** mg/kg-d, 1.0 x 10'7 mg/kg-d,
and 2.5 x 10** mg/kg-d for young children, older children, and adults,
respectively. Over the assumed 30-year duration of residence near the Hudson
River and a 70 year lifetime, the chronic dally Intake (CDI) of PCBs from
sediments 1s calculated to be 2.6 x 10** mg/kg-d. These values are listed In
Tables B.6-5 and B.6-6.
Absorbed Dose (mg/kg/day) -
where
B.6-22
-------
B.6.3 Toxicity Assessment
B.6.3.1 Introduction
PCBs generally have low acute toxicity but are of public health concern due
to their persistence In the environment, the potential to bloaccumulate In animal
and human tissues, and their potential for chronic or delayed toxicity. The
major target organs of PCB toxicity are the liver and the skin. Occupational
exposures to relatively high concentrations of PCBs have resulted In changes In
serum levels of liver enzymes and skin effects such as chloracne (ATSDR, 1987).
In Individuals who accidentally consumed PCBs In contaminated rice oil In Japan
(Yusho patients), routine liver function tests were abnormal. PCBs have also
been shown to cause some developmental effects and neurological effects 1n Yusho
patients, occupationally exposed Individuals, and 1n Individuals exposed via the
consumption of contaminated fish.
USEPA has developed several sets of toxicity values to provide quantitative
estimates of the potency of chemicals and resultant toxic effects. The reference
dose (RfD) and the cancer slope factor (CSF) are the toxlcologlcal values of
relevance for this assessment. The RfD and the CSF are fundamentally different
In their assumptions of the relationship between dose and response. For
carcinogenic effects, It 1s assumed that there Is no threshold below which no
effect will occur. Some risk, however small, Is associated with every level of
exposure. In contrast, RFDs for non-carcinogenic effects assume that there 1s
a threshold dose below which there will be no deleterious effect.
Verified RfDs and CSFs are available on USEPA's Integrated Risk Information
System (IRIS). These toxicity values and other health risk assessment
Information are Included In IRIS after a comprehensive review of chronic toxicity
data by work groups of USEPA scientists. Verified RfDs and CSFs are considered
to be the most reliable basis for estimating noncarclnogenlc and carcinogenic
risks resulting from chronic chemical exposures. If toxicity values are not
available for the chemicals of concern in IRIS, EPA's secondary source known as
the Health Effects Assessment Summary Tables (HEAST) may be reviewed. This 1s
B.6-23
-------
a quarterly review of toxicity values for carcinogenic or noncarclnogenic
effects. If no toxicity values are available 1n HEAST, guidance documents
published by USEPA's Environmental Criteria and Assessment Office (ECAO) may
provide useful toxicity Information.
Because PCBs comprise a class of chemicals, but are largely considered to
be one compound as far as regulatory approaches to their evaluation, additional
discussion regarding the characteristics of the diverse class of compounds, and
the possible Impact on regulation 1s also provided.
B.6.3.2 Noncarclnogenic Effects
Reference Doses (RfDs) provide a benchmark for the dally dose (with a
confidence level of an order of magnitude) to which humans, Including sensitive
populations such as children or pregnant women, may be subjected without an
appreciable risk of deleterious effects during a lifetime of exposures. The same
unit system commonly used for dose (mg chemical/kg body weight - day) is also
used for RfDs.
The basis of an RfD calculation is usually the highest dose level that did
not cause observable adverse effects (/.e., the No Observed Adverse Effect Level,
or NOAEL) after chronic exposure 1n animal experiments. The NOAEL 1s then
divided by uncertainty (safety) factors and, occasionally, an additional
modifying factor to obtain the RfD. In general, the uncertainty factor 1s
determined by multiplying by a factor of 10 to account for interspecies variation
and a factor of 10 to account for sensitive human populations. Additional
factors of 10 are Included in the uncertainty factor, when the RfD 1s based on
the Lowest Observed Adverse Effect level (LOAEL) Instead of the NOAEL, if an
experiment was conducted over less than a lifetime of exposure, or if there are
inadequacies 1n the database.
For PCBs, an oral RfD of 1 x 10*4 mg/kg/day for Aroclor 1016 was developed
in 1987. Other RfDs for other Aroclor mixtures are not available, and thus 1t
is assumed that the same RfD applies to all PCB mixtures. The basis of the RfD
B.6-24
-------
was a study In rhesus monkeys where offspring of a group exposed to 1 ppm PCBs
in food were significantly smaller than the controls. The NOAEL In this study
was 0.25 ppm 1n the diet, equivalent to a dose of 0.0105 mg/kg-d. An uncertainty
factor of 100 was applied to this NOAEL to generate an RfD of 1 x 10"4 mg/kg-d.
The most recent consultation of the IRIS database, in which PCBs Here
updated as of January 1990, revealed that no oral RfD for PCBs currently exists.
USEPA Is in the process of reviewing the RfD for PCBs. Recent conversations with
and publications by USEPA staff (Dourson, personal communication; Clark, personal
communication; Dourson and Clark, 1990) Indicate that this RfD of lxIO"4 mg/kg-d
is being considered for use by some USEPA regional offices and 1s under
consideration for adoption by the Environmental Criteria Assessment Office of
USEPA.
Other, more restrictive, RfDs for PCBs are currently being proposed by
state agencies. For example, based on research conducted by Fein (1984), which
correlates consumption of PCB-contaminated fish with lower birth weight, smaller
head circumference, shorter gestational age, and poorer neuromuscular maturity
in infants, the Minnesota Department of Health has proposed an RfD for human
reproductive effects of 5 x 10'* mg/kg-d (Shubat, 1991). Researchers 1n human
development have suggested a threshold for the developmental effects of PCBs in
the range of 1 to 3.4 ppm PCBs 1n breast milk (fat basis). Application of
pharmacokinetics modeling to these threshold values indicates a threshold dose
for a 60 kg woman on the order of 3 x 10*4 to 1 x 10ฐ mg/kg-d. Incorporating a
10-fold uncertainty factor for 1ntra-species variability suggests an RfD based
on human studies in the range of 3 x 10'* to 1 x 10"4 (Shubat, 1991).
There 1s some concern that an RfD based on studies of Aroclor 1016 in
monkeys may underestimate the risks associated with human exposure to PCBs. Not
only do human studies indicate a greater possible sensitivity In humans than
animals, but Aroclor 1016 is also one of the less highly chlorinated PCB mixtures
generated commercially. Until the RfDs based on human studies can undergo
verification by USEPA, 1t Is anticipated that the RfD of 1 x 10"4 would provide
an adequate indication of potential non-cancer toxicity, at least within an order
B.6-25
-------
of magnitude range (Clark, personal communication). Consequently, this RfD 1s
used here to evaluate non-carc1nogen1c risks from PCB exposure.
B.6.3.3 Carcinogenic Effects
Definition
PCBs have been classified by USEPA as B2 carcinogens, or probable human
carcinogens. The carcinogen classification system 1s based on the strength of
the evidence that a compound Is a human carcinogen. Compounds for which there
1s sufficient human evidence of carcinogenicity are 1n Class A. Chemicals for
which there 1s limited human evidence and sufficient evidence of carcinogenicity
1n animals are In Class Bl; those for which there 1s Inadequate evidence 1n
humans, but sufficient evidence In animals are 1n Class B2. Chemicals 1n classes
Bl and B2 are considered to be probable human carcinogens.
USEPA's Carcinogen Assessment Group (CAG), which reviews human, animal, and
1n vitro data on suspected chemical carcinogens, has calculated cancer slope
factors (CSFs) for those determined to be carcinogenic. CSFs are used to
estimate the excess cancer risk due to continuous exposure to a chemical
throughout the course of a 70-year lifetime. CSFs are based on data from
lifetime animal bloassays, although human data are used when available. Since
the evidence of carcinogenicity from human studies 1s Inconclusive for PCBs, the
CSF 1s based on studies conducted In animals. For animal data, USEPA uses a
mathematical, linearized multistage model to extrapolate from the high doses used
1n the bloassay to the low doses expected to result from environmental exposure.
The CSF represents the upper 95 percent confidence 11m1t of the slope of the
linear portion of the dose-response curve. Excess carcinogenic risk for the
experimental animal 1s extrapolated from this slope to the excess carcinogenic
risk expected for humans.
B.6-26
-------
USEPA Cancer Slope Factor
USEPA has determined that the cancer slope factor (CSF) for all PCB
mixtures 1n humans Is 7.7 (mg/kg-day)"1. This CSF Is based on a study by Norback
and Weltman (1985) 1n which rats (70 nale and 70 female) were fed a diet
containing Aroclor 1260 mixed in corn oil for two years (100 ppm for 16 months
and 50 ppm for 8 months) followed by a standard diet alone (0 ppm) for another
five months. Controls were fed a diet containing com oil for 18 months followed
by a standard diet alone for another five months. Among animals surviving more
than 18 months, the Incidence of hepatocellular carcinomas 1n females and males
was 91 percent (43/47) and 4 percent (2/46) respectively; the Incidence In
corresponding controls was 0 percent. Among these survivors, an additional 9
percent of the females (4/47) and 11 percent of the males (5/46) had neoplastic
nodules; the Incidence of total neoplasms 1n the corresponding controls was 2
percent (1/49) and 0 percent (0/32) for females and males, respectively. ,
During the course of this study, concurrent liver morphology studies were
performed on tissue samples obtained by partial hepatectomles of exposed rats (3
male and 3 female) at eight time points. These studies demonstrated the
sequential progression of liver lesions to hepatocellular carcinomas. This
progression from nodules to benign tumors to carcinomas has been used by USEPA
to justify Its calculation of CSFs on the basis of both malignant tumors plus
benign tumors and nodules. Formerly, CSFs were determined on the basis of
malignant tumors alone and thus were lower.
USEPA views the Norback and Weltman study as a positive study since: (a)
1t used an adequate number of animals (* 50/group); (b) 1t spanned the natural
life span of. the rat; and (c) Sprague-Dawley rats have a low Incidence of
spontaneous hepatocellular neoplasms. Confidence In this study Is apparently
enhanced by the fact that the current CSF for Aroclor 1260 1s similar to the old
value of 3.9 (mg/kg-day)'1, based on data from an earlier study (Klmbrough,
1975).
8.6-27
-------
USEPA views the CSF for Aroclor 1260 as representative of all other PCB
mixtures. First, there Is no Information about which Aroclor congeners are
carcinogenic and, second, the Norback Weltman study using Aroclor 1260 Is
superior to all other studies of PCB mixtures (USEPA, 1988). Although this may
be a conservative assumption (USEPA, 1990), USEPA believes that given the data
available, the public health 1s best protected by assuming that Aroclor 1260 1s
representative of all other mixtures (Mukerjee, personal communication).
In this assessment, USEPA protocol 1s used and the evaluation of
carcinogenic risk uses the CSF of 7.7 (mg/kg-d)'1.
Other Cancer Slope Factor Studies
USEPA headquarters has calculated a CSF for Aroclor 1254 of 2.6 (mg/kg-
day)'1 (USEPA, 1988) 1n contrast to the value of 7.7 (mg/kg-d)'1 calculated for
Aroclor 1260. This value was based on a 1978 National Cancer Institute (NCI)
study 1n which statistically significant, dose-related Increases In hepatic
neoplastic nodules and carcinomas were seen 1n female (Fischer 344) rats fed a
diet containing Aroclor 1254. The separate CSF for Aroclor 1254 has not been
promulgated by headquarters due to uncertainties 1n the underlying study and so
1s not used 1n the current assessment.
In establishing the CSF for Aroclor 1260, USEPA (1988a) chose to exclude
a rat study by Schaffer et aJ. (1984). That study evaluated the carcinogenicity
of Clophens 60 and 30, commercial PCB mixtures with a congener profiles similar
to Aroclors 1260 and 1016 respectively (Schaeffer, 1988; ATSDR, 1987).
USEPA excluded this study because there was a discrepancy 1n the published
tables, which prevented a determination of the number of animals at risk, and the
study used male (HIstar) rats. Norback and Weltman showed that female (Sprague-
Dawley) rats are more sensitive than males. Although these objections have
merit, the study by Schaeffer, In many ways, resembles that of Norback and
Helton. Both used an adequate number of animals (2 50/group), were of similar
duration (lifespan of the rat), used rats with a low Incidence of spontaneous
B.6-28
-------
hepatocellular neoplasms (less than 10 percent), and used doses which did not
Increase mortality, as shown below.
NORBACK & WELTMAN
SCHAEFFER FT AL
Strain of male rat
Sprague-Dawley
Wstar
Exposure route (vehicle)
Oral (diet)
Oral (diet)
Tumor site (type)
Liver (carcinoma and
neoplastic nodules)
Liver (carcinoma and
neoplastic nodules)
Aroclor/Clophen
none 1260
none 30 60
Nominal dose 1n feed (ppm)
TWA-D, (mg/kg/day)
0 100
0 3.45
0 100 100
0 5.0 5.0
D* (ซg/kg/day)
0 0.59
0 0.85 0.85
Tumor Incidence
0/32 7/46
6/122 42/130 122/125
Tumor Incidence ( percent)
0 15
5 32 98
Exposure duration
720 days (1/2 dose
days 480 to 720)
832 days
Observation period
days 540 to 870
days 501 to 832
Observation duration
330 days
331 days
Study duration
870 days
832 days
If females are Indeed the most sensitive sex, the data of Schaeffer et al.
(1984) on male rats could not be used to determine CSFs for PCBs. However, a
comparison of tumor Incidence between rats exposed to Clophens 30 and 60 does
Indicate that less chlorinated Aroclors may also be less carcinogenic. The CSF *
for Aroclor 1016 could be estimated given the relative change 1n CSFs between
Clophens 60 and 30 as well as the CSF for Aroclor 1260.
B.6-29
-------
B.6.3.4 Toxicity of Specific PCB Congeners
Few congener-specific analysis of PCBs 1n environmental samples from the
Hudson River have been performed. Because of the different chemical character-
istics of various PCB congeners, different congeners will partition differently
throughout environmental media. Therefore, the congeners to which humans may be
exposed may differ from the congeners present 1n the original source of
commercial PCB formulations and may also differ from one exposure route to
another.
In December 1990, USEPA hosted a workshop to evaluate the possibility of
shifting the toxicity assessment of PCBs from an Aroc1or-spec1f1c approach to a
congener-specific approach, which would permit a more specific toxicity
assessment for Aroclor mixtures as well as for environmental mixtures of PCBs.
The consensus was that Information 1s currently Insufficient to develop a Toxic
Equivalency Factor (TEF) approach for PCBs similar to that used for polychlor-
Inated dlbenzo-dloxlns and -furans. Some of the Issues related to. toxicity are
listed below.
PCBs do not Induce one set of toxic endpolnts via the same mechanism
of action. To the contrary, 1t appears that various structural
classes of PCBs may exert a number of different toxic effects via
different mechanisms. As a result, a TEF scheme would have to be
developed for each mechanism which produces an effect of concern.
Some PCBs (substituted 1n both the para and one meta position of
both rings) produce effects similar to dloxln and act via the same
mechanism. A proposed TEF scheme for these dloxln-llke effects
ranks the toxicity of these PCBs relative to dloxln, based on the
ability of each congener to Induce aryl hydrocarbon hydroxylase
(AHH) activity or exert other d1ox1n-Hke effects.
Coplanar PCBs (meta and para substituted but not ortho substituted)
are the most similar to dloxln and are given the highest TEF values
for dloxln-llke effects.
Increasing ortho-substitution of these PCBs decreases their ability
to Induce the AHH enzyme system and decreases their similarity to
dloxln. Such PCBs are given lower TEF values.
B.6-30
-------
Some PCBs seem to promote cancer via a mechanism different than the
d1ox1n-Hke AHH enzyme Induction. The current understanding Is that
PCBs substituted 1n both the para and at least one meta position on
each ring and having some degree of ortho substitution are potent
carcinogen promoters by Inhibiting Intercellular communication or
stimulating cell proliferation. The carcinogenic potential of most
commercial PCB mixtures 1s probably attributed to this mechanism.
No TEF approach has yet been proposed for these effects.
PCB metabolites, such as arene oxides, may exert still different
toxic effects via different mechanisms.
Recent studies have shown that lightly chlorinated ortho-substltuted
PCB congeners may be associated with developmental neurotoxicity,
e.g. statistically significant decrease 1n scores on psychomotor
development tests associated with exposure of human Infants to PCBs
/n uteroor via breast milk. Lightly chlorinated congeners were
found to accumulate 1n the central nervous system of monkeys, and to
reduce cellular dopamine concentrations. For humans, this neurolog-
ical effect endpolnt may be more sensitive than the cancer endpolnt
for PCBs, which do not seem to be complete carcinogens.
i
Recent findings Indicate that there may be reproductive effects,
e.g., sperm motility or estrogenic effects, to which humans may be
very sensitive.
B.6.3.S Epidemiological Studies
Epidemiological studies of accidental or occupational exposures to PCBs are
few and for the most part Inconclusive, because of small study size or low
Incidence rates for the endpolnts being evaluated. Several epidemiological
studies do point to an association between exposure to PCBs and some forms of
cancer and other adverse effects. Most of the studies are Inconclusive, either
because they lack statistical power or because exposures are often not to PCBs
alone. Nevertheless, findings are usually consistent with those from animal
research.
Some of these studies contain enough Information to estimate exposure or
dose. In subsequent phases of the study process, further examination of the
epidemiological studies may provide Insights Into the human dose-response
relationships for PCBs.
B.6-31
-------
Cancer Effects
Table B.6-7 outlines the findings of some epidemiological studies that have
examined cancer mortality 1n PCB-exposed populations. Malignant melanoma, brain
cancer, pancreatic cancer, rectal cancer, liver cancer, gall bladder and biliary
tract cancer, hematologic neoplasms, gastrointestinal tract cancer, and stomach
cancer have each been associated with PCB exposure 1n at least one study.
Reports of liver cancer are most common, most statistically significant, and most
consistent with the results of animal studies. As these studies vary 1n quality,
a more Intensive critique of their study designs 1s necessary before considering
the significance of the results.
Non-Cancer Effects
Table B.6-8 summarizes the results of some recent studies that have
examined the non-cancer effects of PCB exposures 1n humans. It Is well
documented from studies of workplace or environmental exposures that high levels
of PCB exposure causes chloracne and dermal effects. Workers exposed to high
levels of PCBs have also complained of lassitude, loss of appetite, loss of
libido, burning of eyes, nose and throat, nausea, and dizziness. Associations
have been made between PCB exposure and Increased blood pressure, liver Injury
and abnormal liver function, and reduced gestational periods. People exposed to
PCBs, along with chlorinated furans and quaterphenyls 1n the Yusho and Yucheng
poisonings reported dermal effects, Itching, swelling, eye discharge, jaundice,
weakness, numbness, fever, and hyperemlc conjunctivae. In many of the patients,
respiratory symptoms persisted and many have chronically Infected airways.
Delayed or Impaired Immune response has been found 1n some of the poisoning
victims (Klmbrough, 1987).
Even for studies that demonstrate a statistically significant association
between PCB exposure and a toxic endpolnt, the results must be Interpreted with
caution since the studies are generally unable to control for exposures to other
chemicals or to the dloxlns and furans which are common contaminants of
commercial PCBs.
B.6-32
-------
B.6.3.6 Other Health-Based Regulatory Limits or Suldelines
Reviewed below are several of the established regulations and guidelines
for PCBs In various media, which were developed with considerations of the^health
effects of PCBs. V
FDA: Tolerance for PCBs In Fish
The Food and Drug Administration (FDA) promulgated a regulation lowering
'j
the tolerance level for PCBs 1n fish destined for Interstate commerce from 5 ppm
to 2 ppm (44 FR 38330, 1979). This regulation, proposed In 1977 and promulgated
In 1979, was stayed 1n the courts and did not become effective until 1984. The
tolerance level was based on weighing the results of a risk assessment against
the magnitude of potential food loss resulting from a lowered tolerance level.
It Is Important to point out that the methodology of the FDA risk assessment
precludes application of Its results to the Hudson River risk assessment for fish
Ingestion. The FDA tolerance was not developed using methodology consistent with
current USEPA guidance for risk assessment. Additionally, the FDA specifically
states that this tolerance 1s Intended to apply to fish entering Interstate
commerce and that this level may not be protective for locally caught fish from
contaminated areas.
ป
To arrive at a tolerance of 2 ppm, the FDA considered national per capita
fish consumption, looking at the general distribution of PCB levels 1n fish for
sale across the U.S. The FDA risk assessment was performed by assuming that with
a tolerance level of 2 ppm this level would be the maximum concentration In fish
encountered by the heavy fish consumer and that PCB concentrations in fish
consumed would be distributed below 2 ppm 1n a manner reflecting a mix of fish
from diverse sources (Cordle, 1982). The tolerance is not based on the
assumption that all fish consumed contains 2 ppm PCBs. Because the distribution
of PCB concentrations in fish caught in the Hudson River by local anglers is
likely to be different from the distribution of PCB concentrations In fish for
sale across the United States, the risk associated with regularly eating Hudson
B.6-33
-------
River fish will differ from the risks associated with the FDA assessment for a
2 ppm tolerance, even If Hudson River'fish do not exceed 2 ppm PCBs.
The FDA has derived a cancer slope factor of 0.34 (mg/kg-day)"1 from the
NCI bloassay using Aroclor 1254. A USEPA derivation of a cancer slope factor
based on the NCI bloassay calculated a slope factor of 2.6 (mg/kg-day)"1 (USEPA,
1988). The reasons for the difference cannot be evaluated without additional
details on the FDA methodology.
USEPA: Drinking Water
USEPA used a cancer slope factor of 7.7 (mg/kg-day)"1 to estimate that
Ingestion of drinking water containing 0.5, 0.05, and 0.005 |ig/7 PCBs corresponds
to increased cancer risks of 10'4, 10'*, and 10"', respectively. The CSF used 1n
this determination 1s consistent with the most recent value recommended by USEPA
and 1s based on the study by Norback and Weltman.
The Maximum Contaminant Level (HCL) promulgated by USEPA for drinking water
1s 0.5 |tg/7 PCBs, which corresponds to a lifetime risk of 10'4 assuming lifetime
Ingestion of 2 liters of water per day, and a CSF of 7.7 (mg/kg-day)"1. The HCL
Is set equal to the practical quantitation level for PCBs, which USEPA has
determined reflects the level that can be measured by good laboratories under
normal operating conditions within specified limits of precision and accuracy (54
FR 22062, 1989).
USEPA: Ambient Hater
USEPA has Issued ambient water quality criteria for PCBs of 7.9 x JO'*
jig/7, 7.9 x 10"* yg/7, and 7.9 x 10'7 pg/7 corresponding to lifetime risks of 10'*,
10"*, and 10'7, based on the ingestion of fish and shellfish and Ingestion of
drinking water. The risks are primarily attributable to Ingestion of fish and
remain constant whether Ingestion of drinking water Is considered or not. These
values were derived using a previous CSF of 4.34 (mg/kg-day)"1. Ambient water
quality criteria for consumption of fish and shellfish assumes a fish consumption
B.6-34
-------
rate of 6.5 g/day, a 3 percent average lipid content for the edible portion of
fish and shellfish consumed 1n the United States, and a biฉconcentration factor
for fish of 10,385 per percent lipid In the fish.
New York State: Ambient Hater
New York State has Issued ambient PCB water criteria for surface waters.
These criteria are 0.001 pg/7 for waters where ingestion of fish and shellfish
may occur and 0.01 pg/7 for waters where Ingestion of drinking water 1s the only
source of human exposure. The NYS 0.001 pg/7 ambient water quality criteria is
slightly higher than the USEPA-derlved ambient water quality criteria (0.00079
lig/7 as described above), which corresponds to a lifetime cancer risk of 10'*
based on consumption of fish.
USEPA Advisory Levels for PCB Superfund Cleanup < ,
In 1990, USEPA's Office of Solid Waste and Emergency Response issued a
guidance document containing preliminary cleanup goals, or action levels, for
various media at Superfund sites with PCB contamination. These concentrations
represent "the level above which unrestricted exposure may result in risks
exceeding protective levels." The guidance establishes the MCL (0.5 ug/L) as an
action level for groundwater and an action level of 19 ug PCBs per gram of
organic carbon for freshwater sediments (USEPA, 1990). It should be noted that
action levels are not cleanup levels.
Guidelines for PCB cleanup in sediments are currently being reviewed by
USEPA. A proposed action level for freshwater sediments (19 ug PCBs per gram
organic carbon) is based on a USEPA ambient water quality criteria of 0.014 ug/L
for protection of aquatic life (0SWER, 1990). This 1s derived through the use
of an equilibrium partitioning approach which relates the PCB concentration In
sediment to that in the Interstitial pore water of the sediment. The 1990 OSWER
guidance states that where sediment values exceed the listed action levels, a
monitoring program of indigenous biota should be instituted. This guidance does
-------
not derive an action level for PCBs In freshwater sediments based on human
consumption of contaminated fish.
National Academy of Sciences: Suggested No-Adverse-Response Level
The Safe Drinking Water Committee calculated Suggested No-Adverse-Response
Levels (SNARLs) for 24-hour and 7-day exposures to PCBs In drinking water In 1977
and updated them 1n 1980. The SNARLs were derived by looking at Induction of
mixed-function oxidase enzymes In the liver of mammals as shown 1n studies by
Grant (1974) and Bruckner (1977) 1n which rats were fed Aroclor 1254. Based on
these studies, a no-adverse-response level was determined to be 1 mg/kg In
animals; an uncertainty factor of 100 was applied. The resulting SNARL for PCBs
1s 0.35 mg/1 water (0.01 mg/kg-d) for 24-hour exposures, and 0.05 mg/1 for 7-day
exposures (0.0014 mg/kg-d). The'NAS determined that a reliable chronic SNARL
could not be calculated for PCBs, because they are suspected carcinogens and
represent a complex mixture of Isomers and Impurities having various biological
activities and environmental fates.
Standards for Occupational Exposures
Both the Occupational Safety and Health Administration (0SHA) and the
American Council of Government and Industrial Hyg1en1sts (ACGIH) have recommended
8-hour time-weighted average exposure limits for PCBs 1n air of 1 mg/m3 for
Aroclor 1242 and 0.5 mg/m3 for Aroclor 1254 (ACGIH, 1986; CFR, 1990). These
limits appear to be based on preventing skin Irritation and chloracne, although
the ACGIH documentation paper notes that the 1 mg/m3 standard for Aroclor 1242
"will offer reasonably good protection against systemic Intoxication, but may not
guarantee complete freedom from chloracne." The 0SHA levels are legally
enforceable 1n the workplace. These exposure limits are not based on protecting
workers from possible carcinogenicity. Assuming an inhalation rate of 10 m3 per
8-hour work day and exposure to PCBs at the enforceable limit (1 mg/m3) for 40
hours per week over 30 years, the resulting estimate of cancer risk Is 3.4 x 10*1
from PCB exposure, based on USEPA's CSF of 7.7 (mg/kg-day)*1.
B.6-36
-------
The National Institute for Occupational Safety and Health (NIOSH) has
recommended a 10-hour time-weighted average occupational exposure 11m1t of 0.001
mg/m' or less (NIOSH, 1977). The NIOSH recommended exposure levels are
guidelines and are not legally enforceable 1n the workplace.
New York State: Ambient Air Guidelines
The State of New York has Issued guidelines for both short-term and long-
term ambient air concentrations for PCBs. The short-term guideline concentration
for ambient air is 0.5 mg/m' PCBs, which is based on and equivalent to the ACGIH
Time Weighted Average-Threshold Limit Value (TWA-TLV) for exposure to Aroclor
1254. The annual ambient air guideline concentration for long-term exposures to
PCBs 1n air is 1.19 ug/m* (NYSDEC, 1989). The TLV does not consider carcinogenic
effects. The preamble to the TLV specifically states that these values are not
to be applied to the general population and several Investigators (Calabrese,
1988; (JSEPA Inhalation RfD documentation; Jarabek, 1990) have specifically
cautioned against the use of the TLV for the general population.
B.6.4 Risk Characterization
B.6.4.1 Definition
The risk characterization step of the risk assessment process defines the
potential threats to human health posed by the contaminants In the Hudson River.
In characterizing and presenting such risks, carcinogenic and noncarcinogenic
health risks are described for a reasonable maximum exposure scenario where
possible.
In contrast to the quantitative assessment of carcinogenic health risk,
assessment for non-cancer endpolnts Involves comparison of average daily exposure
levels with established references doses (RfDs) to determine whether estimated
exposures exceed recommended limits. Typically this comparison Is expressed as
a Hazard Quotient (HQ), which is the ratio of the estimated exposure to the RfD:
B.6-37
-------
HQ - Daily Exposure
RfD
When HQ exceeds unity, unacceptable exposures nay be occurring. Both
exposure levels and RfDs are typically expressed In units of mg of PCB Intake per
kg of body weight per day (mg/kg-d). Actual dally exposures over the assumed
exposure period of 30 years, rather than the llfetlne (70 year) average
exposures, are of concern when evaluating noncarclnogenlc effects. Thus, unlike
evaluation of carcinogenic effects, exposures of less than lifetime duration are
not averaged over an entire lifetime.
Quantitative assessment' of carcinogenic risks Involves evaluation of
lifetime average dally exposure levels and applications of toxicity factors which
reflect carcinogenic potency. Excess lifetime cancer risk 1s calculated as:
Risk - CSF x CDI
where CSF 1s the carcinogenic slope factor and CDI 1s the chronic dally Intake
(averaged over a lifetime) of PCBs from the river. Exposure levels are expressed
1n units of mg of PCB Intake per kg of human body weight per day (mg/kg-d). When
exposure occurs only over some portion of a lifetime, the lifetime average
exposure, which 1s believed to reflect the contribution of the exposure to
lifetime risk, Is approximated by dividing the total PCB Intake over the period
of exposure by the total lifetime (assumed to be 70 years). Since carcinogenic
slope factors (CSFs) are expressed 1n units that are the reciprocal of those for
exposure (I.e., the CSF is expressed as (mg/kg-d)'1, multiplication of the
exposure level by the CSF yields a unit!ess estimate of cancer risk.
The risk estimate for cancer reflects the Incremental Increase In the
probability of getting cancer following site-specific exposure compared to the
background probability or the probability associated without exposure to site
contaminants. For example, a risk of 2.0 x 10'* means that an additional 2
people In an exposed population of 100,000 people who are actually receiving the
B.6-38
-------
dose from which the risk was calculated are estimated to manifest cancer during
a lifetime of exposure.
Table B.6-5 summarizes the carcinogenic risks from exposure to PCBs from
the Hudson River, for all exposure pathways. Non-carc1nogen1c risks, in the foro
of the Hazard Quotient, are summarized 1n Table B.6-6. These risks are discussed
below, by exposure pathway.
B.6.4.2 Dietary Intake
F1sh
The risk calculations for recreational anglers assume that a person
consumes contaminated fish from the Hudson River for 30 years. The assumptions
presented 1n Table B.6-1 are the basis of the risk calculations. The risks are
calculated using the 95th percent confidence limits of mean PCB concentrations
In fish for an estimated 30-year projection and for current conditions (1986-1988
data). For the projected 30-year mean concentration, the excess cancer risk 1s
estimated to be 2 x 10'1 and the HQ 1s 6. The estimated excess cancer risk for
exposure under current conditions 1s 2 x 10"2 and the Hazard Quotient 1s 51.
Drinking Water
Recent monitoring data on Hudson River water Indicate that surface water
PCB concentrations are below regulatory guidelines for drinking water. Thus,
only a screening level risk calculation for the drinking water pathway was
performed. This scenario used the 95 percent upper confidence limit of the mean
PCB concentration (0.06 yg/1) based on samples taken at Rogers Island from 1986 -
1989. Using this exposure concentration and assuming direct consumption of water
without treatment yields an upper bound risk estimate of 6 x 10**. The non-
carcinogenic hazard quotient for this scenario 1s less than unity (HQ<1).
B.6-39
-------
Other
Risks that nay be associated with consumption of vegetables, meat, dairy
nllk or breast milk were not calculated, because Insufficient data are availably
(see B.6.2).
B.6.4.3 Inhalation Exposures
Air concentration data for PCBs are Inadequate to permit a quantitative
assessment of human exposure and risk from the Hudson River site. To do so,
would require monitoring concentrations of contaminants In areas along the Hudson
River and at distances from 1t and relating measured concentrations to the Hudson
River as the source.
B.6.4.4 Recreational Exposures
Dermal Exposure to River Sediment
The risk calculations assume that an Individual 1s dermally exposed to
river sediments over a 30-year period. The magnitude of this exposure changes
over time because of different behavioral and physical characteristics at
different ages. The risks are calculated assuming that the PCB level In sediment
1s 66.2 ppm.
The estimated lifetime excess cancer risk from exposure Is 4 x 10'*. The
Hazard Quotient for dermal exposures never exceeds unity for any exposed age
group. There are many uncertainties regarding the current PCB levels In
sediments. Because most recent sediment sampling efforts have focused on
previously defined hot spots, the 66.2 ppm RME value Is probably biased toward
sediments with relatively higher PCB concentrations as neither a random sampling
program nor sampling of beach sediments was employed. Thus, this excess cancer
risk value must be considered only a rough estimate.
B.6-40
-------
Incidental Ingestion of River Sediment
As with the evaluation of dermal exposures, risk calculations for
Incidental ingestion of river sediment assume that an Individual 1s exposed to
river sediments over a 30-year period. The magnitude of this exposure changes
over time due to different behavioral and physical characteristics at different
ages.
The estimated lifetime excess cancer risk for Ingestion exposure 1s 2x10**.
The Hazard Quotient for Ingestion exposures never exceeds unity for any exposed
age group. As noted above, the limitations 1n the current database for PCB
levels 1n sediment result 1n a cancer risk that should be considered only a rough
estimate.
Contact with River Water
The risk calculations assume that an Individual 1s dermally exposed to
river water during recreational outings over a 30-year period. The magnitude of
this exposure changes over time due to different behavioral and physical
characteristics at different ages. The risks are calculated using the 95tfa
percent confidence limit of the mean water concentration from Fort Edward.
The estimated lifetime excess cancer risk from water exposure 1s 2.0x10"*. %
The Hazard Quotient for exposures to water never exceeds unity for any exposed
age group. Since the exposure concentration Incorporated Into the risk
calculations Is from the sampling location that demonstrates the highest level
of PCBs of any sampling location, 1t can be anticipated that risks associated
with exposure to water from other locations would be lower.
B.6.4.5 Risk Characterization Compared to Human Studies
The risk estimates are based on studies of the effects of PCBs on various
animal species given very high doses 1n controlled laboratory tests. Specifical-
ly, the evaluation of cancer risk 1s based on studies performed on rats; the
B.6-41
-------
evaluation of the risk of non-cancer endpolnts 1s based on studies 1n monkeys.
There 1s considerable uncertainty 1n extrapolating from animals with high doses
to humans with low doses. Ideally, an assessment of risks to humans from low-
level environmental exposures would be based on toxlcologlcal studies from human
populations exposed at low doses. Although low-exposure human studies are
virtually non-existent, epidemiological studies conducted on humans exposed 1n
occupational settings can provide Information useful to evaluating risks of low
level exposure.
Due to the lack of adequate exposure data and the Inability to control
adequately for other chemical exposures 1n epidemiological studies, It 1s
Impossible to establish whether the quantitative estimate of risk that 1s derived
from animal studies Is consistent with findings from exposed human populations.
A review of available studies does, however, demonstrate that PCBs have been
associated with adverse health effects among human populations and that the
endpolnts of concern from animal studies are consistent with observed human
toxicities (see B.6.3.5 and Tables B.6-7 and B.6-8).
B.6.4.6 Analysis of Uncertainties
The process of estimating human health risks posed by PCBs 1n the Hudson
River contains multiple steps. Inherent in each step are uncertainties that
ultimately affect the final risk estimate. Generally, these uncertainties belong
to one of the major categories of risk assessment: hazard Identification,
exposure assessment, and toxicity assessment.
In this assessment, no formal hazard Identification process was undertaken,
since previous Investigations of the site had Identified PCBs as the predominant
contaminant. The absence of sufficient other chemical data precludes an
evaluation of the interaction among PCBs and other chemicals. These Interactions
nay ameliorate or exacerbate the toxicity of PCBs. An Interaction 1n the latter
category would be the co-presence of PCBs and a primary carcinogen. Unlike
primary carcinogens, evidence 1s accumulating that PCBs do not initiate
neoplastic growths; rather they appear to promote the development of latent
B.6-42
-------
neoplasms. Thus, 1f PCBs and a primary carcinogen were both present In the
Hudson River, an evaluation of PCBs alone would underestimate the cancer risks
associated with river contact.
The Intake quantity of PCBs has uncertainties associated with the variables
used to make the intake calculation. Of these variables, the concentration of
PCBs 1n individual media and the rate at which human populations contact these
media are major sources of uncertainty. For example, contact rates are based on
standardized national statistics that may not apply to the local population.
Because PCBs are a unique class of chemicals, yet are regulated as one
compound, significant uncertainties also exist 1n the toxlcologlcal values used
In this analysis. Cancer and non-cancer toxicity values for PCBs have been
derived from studies using mixtures of PCBs. These values are based on anljnal
studies where doses administered are high relative to typical environmental
exposures experienced by humans. In the extrapolation for high laboratory doses
to low environmental doses, uncertainties arise. In addition, these toxicity
values are assumed to be representative of the toxicity values of all mixtures
of PCBs and of Individual PCB congeners, even though the toxic potency of
different PCB congeners can vary significantly. In the future, specific cancer
slope factors may be assigned to each Aroclor mixture and possibly to environmen-
tal mixtures or specific congeners.
Recent changes 1n the liver tumor classification system may also Influence
conclusions regarding carcinogenicity, both qualitative and quantitative, of
PCBs. In 1986 the National Toxicology Program developed a new classification
system for proliferative liver lesions 1n rats. This change occurred after the
pathological criteria used to classify some of the lesions had evolved based upon
new knowledge. For example, the older liver tumor classification systems
Included the classification of "neoplastic nodules." A more recent understanding
of the pathogenesis of cancer has resulted 1n a division of this single category
Into two categories (Maronpot, 1986); 1) foci of cellular alteration, which are
non-neoplastlc changes in liver cells, and 2) hepatocellular adenomas, a benign
form of liver tumor which may progress towards cancer.
B.6-43
-------
A revaluation of the carcinogenicity of dloxlns (TCDD) was recently
completed using the new classification (Halne Department of Human Services,
1990). In this reread of pathology slides, the total number of liver tumors
declined from 33 to 18 In the high dose group. In addition, 1t was observed tha^
carcinogenicity appeared to be associated with liver toxicity. Thus, cancer
slope factor estimates, which are based 1n part upon neoplastic nodules of the
liver, are likely to change with review of the slides. General Electric is
currently 1n the process of funding a similar revaluation of the pathology
slides from PCB cancer bloassays (Neal, personal communication}:' Because the
cancer slope factor for PCBs 1s also based on evidence of liver cancer, 1t is
possible that Information from this revaluation could support a downward
adjustment of the CSF for PCBs.
Using the example of the revaluation of the carcinogenicity of TCDD
discussed above, the reduction 1n the upper 95 percent confidence limit of the
linearized multistage (LMS) model, as applied by the Halne Bureau of Health,
showed less than an order of magnitude decrease 1n the cancer slope factor of
TCDD. In fact, the upper 95 percent confidence 11m1t declined only by a factor
of about four. In contrast, the maximum likelihood estimate showed a much
greater decline with the reread of the slides - well over an order of magnitude
demonstrating the insensltivlty of the upper 95 percent confidence limit to
changes 1n the dose-rsponse curve.
Such a finding reflects the degree to which cancer slope factors, based on
the upper 95 percent confidence limit of the LMS, are strongly Influenced by the
dose levels (typically based on the Maximum Tolerated Dose or MTD) of the
particular experiment. Since selection of the MTD (the basis for dose selection
in many chronic bloassays) reflects, in part, acutely toxic effects, 1t is not
surprising that a correlation exists between the dose levels of a study and the
carcinogenic slope factor estimate.
Two additional Issues are Important to recognize. First, based on the
currntly available data, the primary risks estimated to exist rsult from
consumption of contaminated fish by anglers who either disregard the fishing
B.6-44
-------
advisory Issued by the NYSDOH or are unaware of 1t. Second, the risks associated
with some potential pathways of exposure were not quantified 1n this assessment
due to Inadequate data. Preliminary analysis Indicates that risks might be
associated with these other pathways and that further data are needed to derive
a total risk across all exposure pathways. Table B.6-1 summarizes potential
exposure pathways, Including those for which quantitative risk assessments were
not performed.
B.6.5 Lower Hudson Discussion
The extent of area encompassed by the Lower Hudson River and estuary and
the volume of data that would be required to conduct a quantitative exposure and
risk assessment for this area combine to make such an effort difficult to
accomplish with the available data. Issues such as potential PCB exposure
differences 1n the fresh versus salt water portions of the Lower Hudson,
Including assessing the Impacts of metropolitan and Industrial sources of PCBs,
will have an Important bearing on a risk assessment for the Lower Hudson. At
this time, data for the Lower Hudson are sufficient to provide only some general
comparisons with the preliminary risk assessment for the Upper Hudson.
t Fish have been sampled extensively at various locations In the fresh
and salt water portions of the Lower Hudson. PCBs transported from
the Upper to Lower Hudson contribute to the PCBs found In the
freshwater fish population of the Lower Hudson. A comparison of the
concentrations of PCBs 1n the fish from this freshwater portion of
the Lower Hudson Indicate that overall concentrations 1n the Lower
Hudson are slightly below those 1n fish from the Upper Hudson, but
they are on the same order of magnitude. Therefore, risks associat-
ed with human consumption of these fish assuming exposure
patterns are similar for the Upper and Lower Hudson would be
similar to the risks associated with consumption of fish from the
Upper Hudson.
The available data on PCBs 1n river water from the Lower Hudson are
both fragmentary and out of date. The data available are limited to
a few samples collected between 1978 and 1981 (Schroeder and Barnes,
1983). These data are not adequate to characterize possible
exposures that might occur via river contact.
B.6-45
-------
In the Lower Hudson, samples were taken from river bottom sediments,
which 1n many cases were 1n areas unlikely to receive human
exposure. This situation makes 1t difficult to apply the data to
human exposure scenarios. In order to provide an adequate charac-
terization of the human health risks associated with exposure to
river sediments In the Lower Hudson, sampling of sediments from
areas of human contact, such as river banks, would be required.
Data on PCB levels 1n air of the Lower Hudson are even more United
than for the Upper Hudson. Furthermore, It 1s likely that there are
sources of PCBs In the airshed of the Lower Hudson that are not
adequately characterized and It may be difficult to correlate
airborne PCBs with a specific source.
A substantial effort would be required to evaluate how exposure
pathways along the Lower Hudson differ from those of the Upper
Hudson. This effort 1s particularly difficult for the Lower Hudson,
because exposures may be unique not only 1n comparison to the Upper
Hudson, but may change within different parts of the Lower Hudson,
because of the differing Influences of salt water and freshwater.
The preliminary assessment of the human health risks associated with the
Upper Hudson indicates that the risks posed by exposure to PCBs from consumption
of fish outweigh risks associated with other routes of exposure by several orders
of magnitude. It Is likely that exposures from consumption of fish taken from
the Lower Hudson will similarly present the greatest risk, because extensive
bioaccumulatlon of PCBs occurs 1n fish. Since exposures via fish consumption
overshadow other exposures to such a great extent, the total risks from PCBs 1n
the Lower Hudson would not be expected to change significantly even with
collection of adequate data for additional exposure pathways. If, however, a
comparison or evaluation of the possible risks from other uses of the river
(recreational contact, drinking water, etc.) are desired, additional data must
be collected.
B.6-46
-------
SYNOPSIS
INTERIM ECOLOGICAL RISK ASSESSMENT
(Section B.7)
The objective of this interim ecological risk assessment (B.7.1) is to present available
information considered pertinent to a subsequent more definitive assessment of risks to non-
human populations exposed to PCBs derived from the Upper Hudson River. This assessment
is considered interim, because sufficient data are not available to provide a comprehensive and
fully quantitative risk assessment. The ecosystem is defined as the river (Fort Edward to Troy)
and adjacent uplands.
As background, a description of the ecosystem is given (B.7.2), drawing from previous
environmental investigations and other published information. The PCB exposure assessment
(B.7.3) presents a conceptual food chain of PCB exposure, uptake and transfer pathways. A
simplified aquatic food chain (benthic insect, benthic fish, carnivorous fish, bird and mammal)
is developed and five indicator species (chironomid larvae, brown bullhead, largemouth bass,
herring gull and mink) are selected, based upon data available. PCB exposure concentrations
and possible PCB uptake and transfer via sediment, water and food intake are discussed. The
analysis utilizes measurements of PCBs in sediments, water, fish and some aquatic insects.
The toxicity assessment (B.7.4) relies upon a review of the literature on PCB toxicity in
aquatic species, fish-eating birds and mammals as well as proposed criteria and guidelines. The
risk characterization (B.7.5) compares monitored and estimated PCB exposure levels of selected
indicator species to published toxicity information and proposed PCB guidelines.
Based upon the limited available data, it is premature to conclude whether ecological
risks specifically attributable to PCB contamination from the Upper Hudson River exist In spite
of the many years of PCB monitoring data for water and fish and sporadic monitoring of
sediments, only one known study of potential adverse ecological effects of PCBs in the Upper
Hudson has been published, but could not pinpoint PCBs as the causal agent
-------
PAGE INTENTIONALLY LEFT BLANK
-------
B.7 Interim Ecological Risk Assessment
B.7.1 Phase 1 Objectives
The objective of this Interim Ecological Risk Assessment 1s to present
Information pertinent to a subsequent and more definitive ecological risk
assessment of PCBs In the Upper Hudson River. This preliminary discussion
adheres to current USEPA Superfund guidance for ecological risk assessments
(USEPA, 1989d) and 1s the first step 1n the ecological risk assessment procedure.
Depending on subsequent analyses, Interpretations and conclusions will be
presented 1n following phases. In the ensuing discussion, the term ecosystem
refers to the river and adjacent upland areas from the vicinity of Hudson
Falls/Fort Edward to the Federal Dam, but does not refer to the entire Upper
Hudson drainage basin.
Although considerable data have been gathered on PCBs In the Upper Hudson
River, the monitoring programs have been conducted by a variety of agencies and
researchers. Therefore, the data gathered do not follow a singular, orchestrated
plan to assess ecological risk per se. Although 1t may be tempting to assine
that the many years of monitoring should provide definitive conclusions on the
ecological health of the Upper Hudson site, the fact remains that very little
data have been gathered to relate the measured PCB concentrations 1n sediment,
water and biota to observed ecological effects. Thus, this preliminary
assessment relies on published Information concerning PCB toxicity 1n relation
to measured PCB concentrations In Upper Hudson River sediments, water and biota.
A comprehensive ecological risk assessment, including population, community and
ecosystem Interactions 1n response to PCB exposure, Is not possible with the
available monitoring data. Continued efforts 1n subsequent phases of the
reassessment will evaluate the need for additional ecological monitoring.
PCB contamination 1n the Upper Hudson River ecosystem Is most evident in
the river sediments, water column and fish. Although fish have been the most
extensively monitored component, some PCB data for aquatic Invertebrates exist
and a small amount of data on PCBs 1n crop/plant species have been collected (see
B.71
-------
discussion at B.3). Because very limited data exist concerning PCB levels In
piscivorous (f1sh-eat1ng) birds, mammals, and other wildlife, a semi-quantitative
evaluation of possible PCB exposure for these groups Is presented. Indicator
species have been chosen on the basis of species occurrence, PCB sensitivity
and/or availability of toxicity and PCB monitoring data. This approach provides
an Initial evaluation of potential ecological risks for selected species.
B.7.2 Ecosystem Description
B.7.2.1 Terrestrial Habitats
Habitats
The terrestrial ecosystem Is characterized by mixed deciduous, and
coniferous forests, croplands, pasture, marshes, and seasonally or semi-
permanently flooded evergreen and deciduous forests (see Plate B.l-4). The
forested area 1s representative of northeastern transition woodlands (Andrle and
Carroll, 1988). Forested areas are particularly apparent along the eastern
shoreline and nearby uplands. Forests west of the river are sparser and
discontinuous, but are apparent along much of the shoreline and surrounding area
north of the Saratoga National Historical Park (north of Stillwater). Woodland
species common to these areas Include the white-tailed deer, raccoon, opossum,
various squirrels and other small mammals.
Agricultural land use along much of the Upper Hudson 1s dominated by
pasture and croplands that support the local dairy Industry. Deciduous forests
as well as meadows and grasslands containing a variety of herbaceous shrubs,
grasses and wildflowers often border the farm fields. Such diversity in cover
types provides an abundance of food and suitable habitat for white-tailed deer,
foxes, birds and small mammals. Several species of raptors, Including the red-
tailed hawk and the great homed owl, frequent these open areas.
B.7-2
-------
The New York State Department of Environmental Conservation has defined and
designated "significant habitat," considered to be areas of special concern for
aquatic and terrestrial species. Within the Upper Hudson River area, deer
wintering areas are listed as significant habitats (Malcolm Plrnie, 1984b).
Other, possibly sensitive or Important habitats are fish spawning areas,
waterfowl wintering areas, pine barrens, and bog-wetlands. Wetlands surrounding
the Upper Hudson are largely riverine as classified by Cowardln et a1. (1979).
Vegetation
The Upper Hudson River lies 1n the Hudson Lowlands. Vegetative classes of
the pockets of wetlands along the river Include: emergent, moss-lichen, floating
leaved bed and submergent bed plant varieties. At higher elevations, there is
a transition to a northern hardwood forest with species such as sugar maple,
beech, yellow birch, hemlock and white pine (Malcolm Pirnie, 1984b).
Birds
The Hudson River Valley Is part of the Atlantic flyway migratory route and
Champlaln-Hudson Valley crossing (Schlerbaum et <7., 1959). Rivers and wetlands
along this route provide ample bird feeding and breeding habitat for such species
as black duck, mallard, wood duck, both blue and green winged teal and Canada
goose (Malcolm Pirnie, 1984b).
Resident or breeding species include numerous songbirds, wintering
waterfowl and raptors. Based on Andrle and Carroll's (1988) atlas of all
breeding birds within New York State, the following 1s a representative list of
suspected or confirmed breeding species within the Upper Hudson River ecosystem.
Mallard Red-winged Blackbird
American Black Duck European Starling
Northern Pintail Northern Flicker
Blue-winged Teal Eastern Meadowlark
Belted Kingfisher Common Nighthawk
American Bittern Wild Turkey
Spotted Sandpiper Red-necked Pheasant
B.7-3
-------
Herring Gull Ruffed Grouse
American Crow Red-tailed Hawk
Mourning Dove American Kestrel
Northern Cardinal Great Horned Owl
Maunals
Mammals are especially sensitive to human encroachment as range size
decreases and competition for food Increases. Nevertheless, habitat, such as
deciduous and evergreen forests bordered by openlands and wetlands, has
sufficient diversity to support a moderate mammalian population. Game species,
such as cottontail rabbit and white tall deer, abundant muskrat populations and
moderate populations of mink, beaver and weasels are found (Malcolm P1rn1e,
1984b). Wildlife field guides for the Upper Hudson provide the following 11st
of other species likely to Inhabit the vicinity.
Red and Gray Fox Hoary Bat
Raccoon Striped Skunk
Red and Gray Squirrel Uoodchuck
Eastern Chipmunk Meadow Vole
Southern Bog Lemming River Otter
White-footed Mouse Muskrat
Amphibians and Reptiles
Several species of amphibians (frogs, toads and salamanders) Inhabit the
river's shores, shallow tributaries and areas of dense forest litter.
Salamanders common to the area are the spotted, dusky and two-Hned salamanders.
The American toad as well as bullfrogs, green frogs and wood frogs are also
common throughout the surrounding wetlands.
Many reptilian species are found along the river and adjacent uplands.
Characteristic aquatic species Include the snapping turtle, painted turtle, brown
snake and the northern water snake. Terrestrial or upland species Include the
B.7-4
-------
wood turtle, smooth green snake, eastern ribbon snake and the rare tlrter
rattlesnake.
Threatened and Endangered Species
A number of species found 1n the Upper Hudson River Valley are listed by
New York State as endangered, threatened or species of special concern
(Bufflngton, 1991). These protected species are named below by category.
Endangered
Bald Eagle*
Peregrine Falcon*
Shortnose Sturgeon*
Bog Turtle
Threatened
Mud Sunflsh
Osprey
Timber Rattlesnake
Red-shouldered Hawk
Northern Harrier
Special Concern
Least Bittern
Cooper's Hawk
Black Rail
Upland Sandpiper
Common Barn Owl
Common Nlghthawk
Henslow's Sparrow
Grasshopper Sparrow
Vesper Sparrow
New England Cottontail
Small-footed Bat
Southern Leopard Frog
Spotted Salamander
Banded Sunflsh
Blackchln Shiner
*A1so considered endangered under federal regulations.
B.7-5
-------
In addition to being listed as endangered by New York State, the bald
eagle, peregrine falcon, and shortnose sturgeon are currently listed as
endangered by the federal government (USFWS, 1991). No existing field data were
found to document the presence of either the bald eagle or peregrine falcon
within the site area. Although the shortnose sturgeon's range 1s generally
United to the Lower Hudson south of Troy, 1t may access the upper reaches of the
river through the series of locks.
B.7.2.2 Aquatic Ecosystem
Studies conducted 1n the Upper Hudson by the New York State Conservation
Department (now NYSDEC) as early as 1933 document a number of major pollution
problems and associated blotlc perturbations (Farrell, 1933). Pulp and sewage
contaminants were often cited as major factors contributing to a shift In the
benthlc (bottom) community structure from pollution Intolerant taxa, e.g., '
mayflies and caddlsfHes, to pollution tolerant taxa, e.g., midges and
tubificids. This structural shift was also reflected 1n the fish populations
with many more tolerant species, e.g., bullheads, and suckers, caught 1n
extremely polluted areas. Although Farrell (1933) gave a very general account
of pollution sources, 1t was one of the earliest attempts to document the biotlc
Impacts of pollution 1n the Upper Hudson River.
Boyle (1969), 1n his book The Hudson River, described the general
characteristics of the Upper Hudson between Fort Edward and Troy in a rather
blunt narrative fashion:
The character and appearance of the Hudson change entirely when the river
leaves the foothills of the Adlrondacks at Fort Edward, the head of
navigation. Locks and dams now choke the flow and turn the river Into a
forty-mile chain of sluggish lakes. From Fort Edward down to Troy, the
river serves mainly as a highway for pleasure boats and self-propelled
barges...
Ecologically, the canalized Hudson Is sort of an entity unto itself, a
world apart from the rushing Adirondack parent and sealed off by locks and
dams from the estuary to the south. Before the construction of the Troy
Dam 1n 1826, fishes from the Atlantic used to make their way upstream as
far north as Glen Falls, a journey of 209 miles. Shad were found in the
B.7-6
-------
tributaries of the Upper Hudson River such as Battenklll...In colonial
times, sturgeon were abundant 1n this stretch of the Hudson. ...Striped
bass, too, came 1n numbers...Nowadays, a stray striper or shad may work
its way through the lock system, but they are markedly rare above Troy.
The most conspicuous fishes are strangersblack bass... and carp...
In the canalized Hudson, especially where the river slows, the aquatic
Insects differ from those in the rushing river of the Adlrondacks. There
are, 1n certain clean coves and backwaters, a profusion of dragonflles and
damsel flies...There presence indicates, by rough rule of thumb, whether or
not the water Is badly polluted. Alas, the canalized Hudson probably does
not have as many dragonflles as 1t did in times of the past. This stretch
of the river has been greatly despoiled and disfigured by pollution, much
of which Is from pulp and paper mills...thick, gray mats of pulp
wastes...drift downstream, where they sink and pile up against dams.
...Instead of dragonflles and fishes...one may find..."Index organ-
isms"... sludge worms,...leeches,..and rattall maggots.
From the time Boyle's book was published 1n 1969 to the present, positive
changes 1n the Upper Hudson have taken place. A number of sewage treatment
facilities have been upgraded and Industrial discharges are more stringently
regulated (Shupp, 1975).
Earlier in this document (see A.l), the conceptual framework of an aquatic
food chain for the Hudson was presented. Because PCBs bioaccumulate through the
food chain, that approach 1s also used here to provide a foundation for
evaluating the ecological exposure and risks posed by PCBs in the Upper Hudson
ecosystem. Where necessary, information from the previous discussion is
summarized briefly.
Conceptual Ecosystem Framework
Recent evaluations of the Hudson River ecosystem by Limburg et <7. (1986)
and Gladden et a7. (1988) discuss the four major categories of organic resources
in the aquatic ecosystem:
primary producers phytoplankton, periphyton and macrophytes;
detritus particulate organic matter and associated microbial
biomass;
B.7-7
-------
dissolved organic matter; and
consumers microzooplankton, macrozooplankton, benthlc Inverte-
brates and fish.
Phytoplankton
There are very limited data on resident phytoplankton populations 1n the
Upper Hudson River. Data collected by NYSDEC (data provided by R. Bode) Indicate
that species of green algae, blue-green algae, diatoms and dlnoflagellates were
found 1n the Upper Hudson from 1974-1978. Not surprisingly, the dominant genera
recorded in the Upper Hudson were also found throughout the freshwater portion
of the Lower Hudson River (Frederick et a?., 1976; Storm and Heffner, 1976;
Weinsteln, 1977).
During a seasonal survey of phytoplankton throughout the Hudson River,
Storm and Heffner (1976) found that representatives of the diatom genus
Cyclotella and unidentified green nannoplankton were the most common taxa. Of
the 16 river stations surveyed by Storm and Heffner (1976), only one was located
1n the Upper Hudson, approximately two miles above the Federal Dam 1n the
vicinity of Uaterford. Although an overall species 11st by station was not
Included, Storm and Heffner (1976) found that during the warmer months the
phytoplankton communities In the vicinity of Waterford shifted from diatoms to
those dominated by green and blue-green algae. Temporal patterns of diatom
dominance from late fall through spring and green/blue-green algae dominance
during the late summer months are typical for most temperate aquatic systems,
Including the Lower Hudson (McFadden et al.t 1978).
Perlphyton
The growth of perlphyton Is linked to the velocity of the river. Some
species need much slower current than others to maintain adequate population
density (Whltton, 1975). For example, many of the benthlc diatoms, e.g.,
Navlcula, are attached to silt particles and may form large mats, while others,
B.7-8
-------
such as Cladophora, usually occur 1n faster flowing water and form long resistant
filaments. No perlphyton data are known to be available within the Fort Edward
to Federal Dam river reaches.
V
Hacrophytes
There are very limited data on resident macrophyte (macroscopic forms of
aquatic vegetation) populations. Muenscher (1933) surveyed the "abundant" and
"common" aquatic vegetation from the mouth of Snook Kill to 4echan1cville. This
study found plckerelweed (Pontedarla), arrowhead (Sag1ttar1a), and dense beds of
floating heart (Nyaphoides) 1n relatively shallow, slow-flowing coves and in
areas behind islands adjacent to the various landcut portions of the Champlain
Canal. Since this survey, there have been no additional inventories of aquatic
macrophytes within this part of the Upper Hudson. Thus, the present composition
and extent of macrophytes cannot be determined. In a recent assessment (Feldmafi,
1991, SUNY Binghampton, pers. comm.) of macrolnvertebrate taxa associated with
aquatic macrophyte species, Trapa (Water Chestnut) and Va1l1sner1a (Water Celery)
were collected In the vicinity of the Thompson Island Pool. Given the paucity
of data, the contribution of macrophytes to primary production in the Upper
Hudson is not discernible.
Invertebrate Community
Invertebrates can be categorized 1n three major groups, based on habitat
preference and size. These are: microzooplankton, macrozooplankton and benthlc
invertebrates.
The zooplankton components of most freshwater systems are dominated by
rotifers, pelagic cladocerans and copepods. Although considered an Important
food resource for a variety of plankton-feeding fish, Including emerald shiners,
spottall shiners and tesselated darters (Gladden et a7., 1988), no surveys have
been conducted on resident zooplankton populations within the Upper Hudson.
B.7-9
-------
The benthic invertebrates comprise a large and heterogeneous assemblage of
organisms. Some of the major groups of benthic Invertebrates found 1n many
freshwater systems, likely to occur In the Upper Hudson, are listed below.
Aquatic Insects
DragonfUes (Odonata)
Damsel flies (Odonata)
Mayf 1 les (Ephemeroptera)
Stoneflles (Plecoptera)
Caddisflies (Trlchoptera)
Hidges (Diptera)
Mollusks
Bivalve clams and mussels (Pelecypoda)
Univalve snails (Gastropoda)
Annelids
Aquatic earthworms (01igochaetes)
Roundworms (Nematodes)
Crustaceans
Scuds (Amphipods)
Sowbugs (Isopods)
Seed shrimp (Ostracods)
Surveys of benthic invertebrates during 1972 (Simpson et *7., 1974), 1977
(Bode, 1979) and 1987-1988 (RIBS Sampling, 1987-1988), conducted within selected
parts of the Fort Edward to Federal Dam section of the Upper Hudson, indicate
that the region 1s dominated by midge larvae (Chlronomidae). All the above-cited
studies utilized multlplate samplers (Hester and Dendy, 1962) suspended 1n the
water column. These samplers consisted of tempered hardboard plates and spacers,
which were held together with an aluminum turnbuckle (Simpson et a7., 1974).
Artificial multiplate samplers are biased for freely colonizing benthic species,
such as nidge larvae, and sampling results may not adequately represent the
overall benthic community, including resident Infauna components. In the
following discussion, major results of all benthic studies conducted to date
within the Fort Edward/Federal Dam reaches of the Upper Hudson are reviewed and
analyzed.
B.7-10
-------
During the summer of 1972, NYSDEC conducted a benthic survey at 12 stations
from Troy to an area upstream of Corinth (Simpson et al., 1974). Benthic
collections from Fort Edward to Troy (stations 5-12) Indicate that this stretch
of the Upper Hudson 1s dominated by midge larvae (Diptera: Chironomidae) and
ollgochaete worms (Naididae). An average of 14 taxa were recorded, Including the
dominant genera listed below.
Midge larvae (Diptera: Chironomidae)
Cricotopus
Pentaneura
Polypedllua
Rheotanytarsus
Dance Flies (Diptera: Empididae)
Hemerodroala
Aquatic earthworms (Oligochaetes)
Nais
Pristine!la
The midge larvae and oligochaetes collectively account for greater than 85
percent of the total benthic Invertebrates sampled. In most collections, midge
larvae were approximately two to three times more abundant than the oligochaetes.
During the summer of 1977 (Bode, 1979), three stations (UHUD- 4A, 6 and 11)
within the Federal Dam/Fort Edward reaches were sampled. These stations were
located in the vicinities of Fort Edward, Fort Miller and Waterford. Once again,
the midge larvae (73.5 percent of total) and oligochaetes (22.4 percent of total)
dominated the benthic invertebrate community. An average of 21 taxa were
recorded. In addition to midge larvae and oligochaetes, four species of
Ephemeroptera (mayflies), fourteen species of Trichoptera (caddlsflies) and one
species of Plecoptera (stoneflles) were collected within the Fort Edward to
Waterford region during 1977. The more Intensive 1972 study at seven stations
within the same region recorded only two species of mayflies, three species of
caddlsflies and no stoneflles. The Increased representation of these groups
during the 1977 study 1s of Interest, since mayflies, stoneflles and caddlsflies
are widely recognized as being far less tolerant to poor water quality (low
B. 7-11
-------
dissolved oxygen, high BOD, etc.) than either midge larvae or ollgochaetes
(Wilson, 1984). In addition, there was an Increase 1n species richness from an
average of 14 taxa 1n 1972 to 21 taxa In 1977.
The only study specifically designed to sample benthlc Infauna was
conducted 1n 1988 by NYSDEC (Preddlce and Karcher, 1990). The field sampling
Included quantitative collections of benthlc macrolnvertebrates by ponar grab 1n
order to ascertain 1f significant community changes were associated with bridge
maintenance at Fort Miller. Targeted samples were taken from soft sediment
areas; ponar sampling difficulties precluded collecting samples from gravelly
areas. Both these events bias the samples toward a preponderance of soft bottom
(sediment) dwelling Infauna. Preddlce and Karcher (1990) found an average of 13
taxa 1n the abundant groups listed below.
Group and Taxa (Percent of Total)
Midge larva (29.9 percent)
Trlbelos
Prod ad i us
Chlronoaus
Polypedilua
OUcochaetes (22.9 percent)
Tublficldae
Lunbrlcuius
Unnodrilus
Round worms Nematoda (17.6 percent)
Fingernail Clams (13.4 percent)
Sphaerlun
Pisidlum
Side Swimmers (6.7 percent)
Gamarus
Hyalella
Caddisfly Larvae (2.9 percent)
Phylocentropus
B.7-12
-------
Although midge larvae and ollgochaetes were still the most abundant groups,
there was an increased representation of nematodes, clams and amphipods compared
to the 1972 and 1977 studies (Simpson et a/., 1974; Bode, 1979). These
differences are somewhat expected, given the way the samples were secure^ The
1972 and 1977 studies used multiplate samplers suspended in the water column and
the 1988 study (Preddice and Karcher, 1990) utilized a ponar grab sampler which,
unlike the multiplate samplers, is extremely effective for sampling soft bottom
benthic infauna. Similar to the 1977 study, four genera of mayfly larvae were
recorded. In addition, the four genera of caddisflies found are known to Inhabit
silt or mud substrates (Bode, 1991, per. comm.).
A recent multiplate survey of benthic Invertebrates (data provided by R.
Bode, NYSOEC) within the Federal Dam/Fort Edward reaches of the river was
conducted as part of a RIBS (Rotating Intensive Biological Survey) study of
i
drainage basin 11 during 1987-1988.
The 1987 and 1988 surveys sampled three stations in the vicinity of Fort
Edward, Schuylerville and Waterford, approximately the same locations as the 1972
and 1977 surveys. Midge larvae dominated the benthic community and accounted for
72 percent of all organisms sampled. The most striking dissimilarity, however,
was in the relative percent contribution of the ollgochaetes. The ollgochaetes
comprised one to two percent of the total benthic community compared to an
average contribution of 25 percent recorded during the 1972 and 1977 surveys.
An average of 21 taxa were recorded within the Fort Edward to Waterford region
during the 1987-1988 RIBS program. The following is a list of dominant genera.
Nidge larvae (Diptera: Chlronomidae)
Corynoneura
Crlcotopus
Dlcrotendipes
Hudsonlmyla
Orthocladius
Polypedllun
Rheotanytarsus
Synorthocladlus
Tanytarsus
Thienemannlella
B. 7-13
-------
Caddisfly larvae
Cheumatopsyche
Chi aim
Hydropsyche
Neureclipsis
Mayfly larvae
Baetis
Stenonena
Flatwonns (Platyhelminthes)
Undetermined
Analogous to the 1977 study, the RIBS 1987-1988 biological samples revealed
that caddlsfHes (10 species), mayflies (8 species) and stoneflles (1 species)
are present within the Fort Edward to Waterford region of the Upper Hudson. The
continued presence of these pollution Intolerant (sensitive) groups suggests that
water quality Improvements have occurred 1n the Upper Hudson since 1972.
Fish
A number of studies have attempted to categorize the fisheries within the
40-mile stretch of the Upper Hudson between Federal 0am and Fort Edward. (Refer
to B.1.4 for a more detailed review of relevant fisheries surveys.) Historical
surveys by Greeley and Bishop (1933) and recent surveys (Makarewlcz, 1983;
Malcolm P1rn1e, 1984b; Green, 1985) Indicate rather diverse fish fauna. The vast
majority of species are year-round residents. Although some anadromous species
such as American shad, alewlfe, blueback herring and striped bass may be present
1n the Upper Hudson, the construction of the Federal 0am and Champlaln Canal has
essentially blocked major upstream spawning migrations.
The diversity of freshwater residents In the Upper Hudson 1s Indicative of
the varied habitats that occur 1n this section of the Hudson. The wide variation
1n habitats expands spatial heterogeneity and results 1n a quite complex fishery
resource. For example, Makarewlcz (1983) surveyed nine different habitats within
the Fort Edward to Federal Dam section of the Upper Hudson.
B.7-14
-------
Malcolm Pirnle (1984b) found that relative fisheries abundance varied aaong
areas sampled. For instance, far fewer fish (13 Individuals) were recorded In
the natural river channel compared to areas adjacent to outlets of streams (637
Individuals). Approximately 69 percent of all fish collected (Hakarewicz, 1983),
subsequently summarized by (Malcolm Pirnie, 1984b), were indicative of various
slow-flowing, shallow water zones and stream mouths.
All fish species collected by Makarewicz (1983) were categorized Into four
functional groups (Malcolm P1rn1e, 1984b), Including game fish, panfish, forage
fish and demersal fish. Listed below are the numerically dominant members of
these four functional groups that were collected within the Federal Oaa/Fort
Edward reaches.
Sroup Pwlnant Members
Game Fish Largemouth bass
Smallmouth bass
Panfish Bluegill
Pumpkinseed
Rock bass
Yellow perch
Forage Golden shiner
Spotfin shiner
Spottall shiner
Demersal Black bullhead
Brown bullhead
Common carp
White sucker
Collectively, the above 13 species account for 85 percent of the total
number of resident freshwater fish collected during the Makarewicz (1983) survey
and represent a variety of habitat preferences. For example, bullhead,
pumpkinseed, bluegill and yellow perch prefer relatively protected slow-flowing,
shallow habitats, whereas small mouth bass and spottall shiner seem to prefer more
moderate flowing waters and/or the outlets of small to large streams (Scott and
Grossman, 1973; Malcolm Pirnie, 1984b).
B.7-15
-------
All the major qualitative studies reviewed Indicate that the fish species
historically present In the Upper Hudson continue to reside 1n the Fort Edward
to Troy reaches.
Summary of Aquatic Ecosystem
A number of studies (see reviews by Wetzel, 1975 and Mann, 1975) have
concluded that various autochthonous primary production inputs (phytoplankton,
perlphyton and macrophytes) make Important contributions to aquatic food webs.
Unfortunately, the paucity of data 1n the Upper Hudson makes it Impossible to
determine the relative contribution of the various primary producers.
Furthermore, contributions of allochthonous carbon from upstream areas and the
surrounding watershed are not known, but may be more refractory in nature and of
reduced or limited nutritional value compared to the autochthonous sources.
Whatever the relative sources of organic carbon, data reviewed for the Upper
Hudson Indicate that the system 1s capable of supporting diverse fish fauna.
Many species of fish living 1n river systems can exploit a variety of food
resources, including Invertebrates, detritus and other fish (We1nste1n, 1977,
Moran and Limburg, 1986 and Gladden et al., 1988). Although no fish studies in
the Federal Dam/Fort Edward region of the Upper Hudson have Included routine
analyses of stomach contents, many resident species seem to have diverse and
opportunistic feeding habits (Malcolm Plrnie, 1984b). The exploitation of various
resources by fish populations may lead to a pattern of trophic partitioning, as
exemplified by the dietary preferences (Moran and Umburg, 1986 and Gladden et
a7., 1988) listed below.
Zooplankton
Emerald shiner
Tessellated darter
Spottall shiner
Benthic/Detrltal
Common carp
Goldfish
Eastern silvery minnow
Golden shiner
B.7-16
-------
White catfish
Brown bul1 head
Bluegill
White sucker
Epibenthic Invertebrates V
Yel1ow perch
Pumpkinseed
Redbreast sunfish
Fish/Macroinvertebrates
Largemouth bass
In addition to differences in trophic strategy, the year class strength of
some species may be linked to the timing or availability of food and the foraging
success of larval and juvenile fish. Thus, the temporal variability in fish
numerical abundance and spawning success may be linked to spatial and temporal
trends of trophic resources, although studies of seasonal populations would be
needed to confirm this possibility.
The apparent Improvement In the Upper Hudson fisheries (Shupp 1975, 1987)
appears to correspond with upgrades of sewage treatment plants and more stringent
regulations concerning Industrial discharges. The Improvement in the fish
community seems also to be reflected in the invertebrate populations.
Comparisons of recent RIBS (1987-1988) data on benthlc invertebrates to
studies In 1977 and 1972 (Bode, 1979; Simpson et *1., 1974) suggest a general
water quality improvement in this section of the Hudson. The water quality
improvement is indicated by the trends 1n the average number of intolerant
(sensitive) species such as mayflies, stoneflies and caddisflies and by the
average species richness (total number of species or taxa) 1n multiplate samples
from 1972-1988.
The number of mayflies, stoneflies and caddisflies species (EPT, as defined
below) and species richness are two of a variety of parameters currently in use
by the NYSDEC in order to determine overall stream water quality. Bode et ซ7.
(1991) state:
B. 717
-------
"...EPT denotes the total number of species of mayflies (Ephemeroptera),
stoneflies (Plecoptera), and caddisflies (Trichoptera) found in an average
100-organism subsample. These are considered to be mostly clean-water
organisms, and their presence generally 1s correlated with good water
quality (Lent, 1987). Expected ranges from most streams in New York State
are: greater than 10, non-impacted; 6-10, slightly Impacted; 2-5,
moderately impacted; and 0-1, severely Impacted."
It must be emphasized that the reference to "good water quality" above refers
generally to standard parameters such as dissolved oxygen, biological oxygen
demand, etc., and does not necessarily reflect the possible Influences of PCBs.
A preliminary analysis of the average EPT and average species richness
within the Fort Edward to Federal Dam reaches of the Upper Hudson is presented
below.
EPT and Species Richness Sunmary
1972
1977
1987
1988
EPT
0.2
6.0
7.4
8.8
Species
Richness
14.0
21.0
20.0
22.0
Overal1
Assessment*
1
2
2
2
1 - severely to moderately Impacted
2 - slightly impacted
Since 1972, there has been an increased representation of the more
pollution intolerant (sensitive) groups and an almost two-fold Increase 1n
species richness from 1972 to 1977-1988. In addition, the organic waste tolerant
Naid ol1gochaetes (Bode et a7., 1991) have declined from a relative abundance of
approximately 25 percent (1972 and 1977) to 1-2 percent (1987-1988).
B.7-18
-------
B.7.3 PCB Exposure Assessment
This ecological exposure assessment Identifies probable routes of PCB
exposures, quantifies such exposures where possible and presents a conceptual
trophic pathway or food chain approach to Indicate possible PCB transfer pathways
within the ecosystem. No attempt 1s made to define the aquatic and terrestrial
food web 1n detail. The "New York State Barge Canal Environmental Reportm
(Malcolm P1m1e, 1984b) contains a summary of available environmental/habitat and
species studies relevant to the site, but the Information therein 1s not adequate
to describe the food web 1n detail. Furthermore, data available specific to PCBs
are inadequate to evaluate species, population and community health and dynamics,
which are necessary components of an ecosystem approach. The available data are
used here 1n a simplified ecological framework to develop an initial assessment
of PCB exposure levels (measured and extrapolated) and possible risks for
selected Indicator species that are representative of the aquatic food web.
B.7.3.1 Exposure Pathways
Both exposure and uptake of PCBs by organisms 1n the ecosystem depend on:
ambient (time and space-dependent) PCB levels 1n water, sediments and air;
species growth and feeding habits; PCB storage; kinetic transfer rates; and
metabolism. Because the PCB contamination In the Upper Hudson 1s largely within
the river, aquatic species, such as benthlc Invertebrates, fish and piscivorous
(fish-eating) wildlife, are more susceptible to PCB exposure and uptake than
terrestrial herbivores and omnlvores. Volatilization of PCBs from the river may
also contribute to elevated PCB levels In plants, but this pathway Is likely to
be less significant than aquatic exposures. (Because PCB levels 1n plants near
the river are also not well-documented, such an analysis 1s precluded at this
time.) Similarly, because of the currently very low PCB concentrations 1n water,
direct Ingestion of contaminated river water by upland species 1s not likely to
lead to PCB uptake to the degree expected as a consequence of consumption of fish
from the river. Thus, aquatic organisms and predators of fish from the river are
the focus of this ecological exposure evaluation.
B.7-19
-------
Two primary modes of PCB exposure are examined: 1) direct contact and
absorption with PCBs in water and sediments; and 2) dietary Ingestion of PCBs and
potential food chain transfers. Because PCBs are lipophilic, they tend to
accumulate 1n body tissues, especially body lipid (fat), with the frequent result
that PCB levels 1n body tissues increase with higher trophic level. Therefore,
species at the top of the food chain may accumulate high PCB levels, even though
they may not be exposed directly to PCBs in the water and sediments. (A possible
exception may be species associated with the benthic environment, which are
exposed to very high PCB concentrations 1n sediments and interstitial water.)
A conceptual food chain of PCB exposure, uptake and transfer pathways In
the Upper Hudson can be described as a producer/consumer system, consisting of:
primary producers, such as aquatic macrophytes, filamentous algae,
phytoplankton and periphyton;
primary consumers, such as micro- and macrozooplankton, benthic
invertebrates and various aquatic insects;
secondary consumers, including forage or planktivorous fish, such as
shiners and several demersal species;
carnivorous (game) fish, such as largemouth bass and northern pike;
and
fish eating (piscivorous) birds and mammals at the top or near-top
trophic levels considered here.
Organisms associated with the benthos throughout their life history Include
certain Invertebrate primary consumers (some insect larvae, oligochaete worms,
snails and mussels) and a variety of demersal fish (catfish, bullhead and carp),
collectively categorized as secondary consumers. Other secondary consumers,
Including various forage fish (emerald shiner, spottall shiner, tesselated
darter) and panfish (pumpklnseed, rock bass and yellow perch), feed primarily on
zooplankton and epibenthic Invertebrates. Carnivorous (game) fish, such as
largemouth bass and northern pike, typically feed on aquatic insects and inverte-
brates as juveniles and on other fish and macroinvertebrates as adults. While
the different feeding behavior of fish may result in different exposure risks to
B.7-20
-------
PCBs in the Upper Hudson, PCB uptake 1n food has not been reported and cannot be
assessed at this time.
Birds and mammals are the top trophic level considered here. Wildlife in
this trophic level include the herring gull, heron, osprey, mink and river otter,
all of which consume various species of omnivorous and carnivorous fish.
Omnivorous birds, such as ducks, geese, and various shoreblrds, may be exposed
to PCBs In the river when they feed on several species of submerged and emergent
plants, plankton and aquatic Insects.
Other terrestrial exposure pathways are not part of this analysis, as
additional field studies would be required to permit quantification of
terrestrial exposure/uptake pathways.
B.7.3.2 Identification of Indicator Species
The table below Identifies Indicator species, covering various trophic
levels 1n the food chain. They were selected, based on their probability of
exposure, significance In the food web and the availability of applicable
exposure concentrations, I.e. PCB monitoring data.
Indicator Species
Component
Species
Exposure Route
Benthlc Insect
Chlronomid Larvae
Sediment, Water
Benthlc F1sh
Brown Bullhead
Sediment, Water,
Benthos (diet)
Carnivorous (Game) Fish
Largemouth Bass
Sediment, Water, Fish
(diet)
Piscivorous Bird
Herring Gull
F1sh (diet)
Piscivorous Mammal
Mink
Fish (diet)
The aquatic Indicator species (chlronomid, brown bullhead and largeaouth
bass) were chosen, because s1te-spec1f1c monitoring data are available for these
species.
B.7-21
-------
No site-specific data adequate to assess PCBs In non-aquatic species were
found. Therefore, non-aquatic species were selected based on the availability
of published information regarding PCB uptake and response. Non-aquatic
Indicator species chosen 1n this preliminary assessment may not necessarily
represent the preferred species from an ecological perspective, but their
selection was constrained by available data related to PCBs. For example, the
heron, osprey or even bald eagle would probably be preferred to the herring gull
as a fish-eating indicator bird species, because their diets more predominantly
come from fish than the diet of the herring gull. The herring gull 1s an
omnivore; 1t eats fish, but fish 1s, perhaps, a less significant fraction of its
overall diet. PCB toxicity data are, however, available for the herring gull,
whereas no data were found 1n the literature for heron, osprey or eagle.
Furthermore, there 1s no site-specific information documenting that either
ospreys or bald eagles are significant inhabitants of the site, while herring
gulls are common.
M1nk are often considered quite sensitive mammalian species that obtain a
major portion of their diet from fish (Newell et a7. 1987). As discussed
earlier, mink are known to Inhabit the Upper Hudson site (Malcolm P1m1e, 1984b),
although their population has not been reported. Furthermore, PCB toxicity in
mink has been widely reported.
Pending continued evaluation of the Upper Hudson ecosystem related to PCB
contamination, other Indicator species may be selected and monitored 1n future
phases.
B.7.3.3 Exposure Quantification
Table B.7-1 summarizes measured and estimated PCB exposure levels for the
Indicator species in the Upper Hudson, including an indication of the level of
confidence or reliability In these values. Reliability 1s considered high for
exposures based on measured values with extensive spatial and temporal coverage
(PCBs in water and fish). Reliability is considered moderate for exposures based
B.7-22
-------
on monitoring data that nay lack adequate temporal or spatial coverage but for
which a large data set 1s available (sediments, macrolnvertebrates). Reliability
1s considered low for estimated exposures based on assumed dietary Intake levels.
As Indicated 1n Table B.7-1, PCB concentrations In body tissues have been
measured for the chlronomid and fish Indicator species. There Is no direct
monitoring data for dietary Intake by any of the Indicator species and dietary
intake (where given In the table) 1s estimated, based on possible feeding rates,
as discussed below. The lack of dietary PCB uptake adds emphasis to the previous
discussion that despite all of the available PCB monitoring data, gaps in the
data exist, which hinder a definitive evaluation of PCB exposures and transfers
for species which have not been monitored, e.g., birds and mammals.
Ambient Water and Sediment Exposures
Ambient water Is a significant route of exposure for all aquatic dwelling
organisms. A strong correlation exists between PCB levels measured in fish and
those 1n water (see B.4). Whether this relationship 1s due to direct partition-
ing through gill and body surfaces, contact with sediments, ingestion of
contaminated food or a combination of these factors is not determined.
The measurements of PCBs in the water column in 1986-89 provide a
reasonable baseline for this Investigation. Upper 95 percent confidence bounds
on mean PCB concentrations in the Upper Hudson for this period range from 0.06
i*g/7 (Fort Edward) to 0.034 |ปg/7 (Waterford), as explained 1n Section B.3.
Summer low-flow PCB concentrations could provide the best indicators of possible
ecological harm, because there Is less flushing during low flows and water
temperatures are generally higher, leading to Increased biological activity.
These concentrations, however, do not differ from full-year values to such an
extent as would affect the risk analysis.
Ambient PCB levels in river sediments are less well-deflhed in time and
space and are far more locally variable than those 1n water. Recent sediment
data, preferred for evaluating baseline conditions, are relatively sparse. PCB
B.7-23
-------
levels measured by GE in a recent study Indicate median PCB levels on the order
of <10 mg/kg to >100 mg/kg at the 12 locations sampled, targeted at previously
defined hot spot locations. Earlier data collected In 1977 by NYSDEC yielded
median PCB levels within river reaches on the order of 5 to <30 mg/kg, with hot
spots commonly having >100 mg/kg average PCB levels. The 1984 Thompson Island
Pool survey results Indicate that approximately 80 percent of the samples had PCB
levels less than 100 mg/kg. The highest concentrations measured exceed 1,000
mg/kg 1n the Thompson Island Pool. Taking all surface sediment samples
(including all grab and top core section samples - 569 samples) from the 1984
Thompson Island Pool survey, the 95-percent upper confidence bound on the mean
Is 66.2 mg/kg. This estimate provides a reasonable maximum exposure level for
sediment-bound PCBs. Subsequent analyses 1n Phase 2 and 3 may focus attention
on evaluating sediment exposures as a function of river reach and areas of
particularly high PCB concentrations.
Chlronomld Larvae
The results of the NYSDOH multiplate monitoring (see B.3.4 for a more
detailed discussion of the monitoring results) Indicate that chironomlds (midges)
were the most abundant macrolnvertebrate component of the 1985 samples,
comprising up to 86 percent of the total macrolnvertebrate population at Fort
Miller and Waterford (Simpson et al., 1986). As larvae and pupae, these
organisms often live in burrows or flocculent tubes created from bits of sediment
held together by a secretion of their silk glands. Uptake of PCBs for these life
stages 1s likely to reflect this direct association with contaminated sediments
and Interstitial water.
The 96-hour b1omon1toring studies conducted by NYSDOH 1n 1985 Indicated
that PCB concentrations 1n the tissue of laboratory-reared chlronomld larvae that
were exposed to ambient water 1n the Thompson Island Pool ranged from 5 to >7
mg/kg (dry weight) (NYSDOH, 1986). Corresponding PCB concentrations in ambient
water during these periods ranged from 0.06 to 0.13 |ig/7 In July and from 0.03
to 0.07 |ig/7 for September. Although the NYSDOH study was short-term, I.e. 96
B.7-24
-------
hour exposures, PCB accumulation appeared to reach a quasi-steady state level
during the course of the study.
The NYSDOH study also Investigated PCB uptake in laboratory-beared
chlronomld larvae exposed to sediments 1n the Thompson Island Pool. These
chlronomld larvae had greater than 100 mg/kg (dry weight) of PCBs after 96 hours
of exposure, an apparent reflection of the higher PCB concentrations 1n the
sediments and Interstitial water.
*
b
In light of the very sparse database of Information from which these short-
term exposure estimates were drawn, the baseline PCB levels in chlronomld are
uncertain. Nevertheless, the limited data suggest that considerable
bloaccumulatlon takes place.
Fish *
Exposure levels of PCBs 1n fish are needed to evaluate:
potential adverse effects on the health of fish caused by the PCB
levels 1n fish tissue, and;
potential adverse effects on fish-eating wildlife as a result of.
consumption of PCB-conta1n1ng fish.
As discussed 1n B.4, PCB levels 1n fish have generally displayed an
exponential decline from 1975 to 1988, although the PCB levels fluctuate frtm
year to year and average concentrations 1n 1988 were generally greater than those
observed In 1987. PCB levels In fish for the three most recent years for which
samples are available (1986-1988) are summarized below.
B.7-25
-------
Ifcpar-Bowd Naan KB Concant rations* In Hah at Thraa Uppar Kikan locations
Caa/fcfl Met wight)
Vaar
KM 153-155
KM 175
BM 190-195
(Thcapson Island Pool)
1MB
U
PICSO
1MB
BB
WCSD
IMS
BB
PKSO
1986
3.3
7.9
13.9
7.0
12.6
48.7
1987
3.1
2.3
4.3
16.9
5.9
1988
5.8
2.7
3.6
13.7
5.5
11.7
20.3
8.1
*Valuaa ara 95-parcant uppar confidanca bowda on tha aaan tokan froa Tab I as B.3-16, >.3-17, and 1.3-18.
Abbreviations:
KM Klvar Nllo
1MB tarjaaouth Bass (fillats)
BB Brown Bulthaad (fillats)
MCSO PtMpkinsaad (yearling saaplas -- haad and viseara raaovad)
Although PCB concentrations in fish have been monitored for over 15
species, pumpkinseed, largemouth bass and brown bullhead samples represent the
largest and most recent data available. These three species also comprise a
large proportion of the overall Upper Hudson fish fauna. Therefore, PCB
concentrations listed above for these three species are considered representative
of recent baseline conditions 1n the Upper Hudson.
PCB concentrations in these species vary with River Mile, with the highest
PCB levels 1n the Thompson Island Pool samples, lowest concentrations 1n those
samples caught near Federal Dam and intermediate PCB levels for the samples near
Stillwater. Recent (1986-1988) 95-percent upper confidence bound total PCB
concentrations 1n fish for Thompson Island Pool samples range from >5 ppm
(pumpkinseed) to <50 ppm (brown bullhead). Representative upper-bound PCB levels
near Federal Dam range from >2 ppm (brown bullhead) to <6 ppm (largemouth bass);
PCB levels for these three species near Stillwater are on the order of 5 to 10
ppm.
Because many toxicity studies report adverse affects of PCBs to birds and
mammals In terms of dietary exposure levels, assessing potential risks to fish
predators would be based Ideally on whole-fish PCB concentrations, because fish-
eating wildlife consume whole fish. PCB levels were measured, however, 1n fish
B.7-26
-------
fillet (muscle) tissue. Filleting and skinning the fish samples remove a large
proportion of Hpid-contalnlng tissue where PCBs tend to accumulate. Thus, PCB
levels 1n whole-fish samples from these species are likely to be somewhat greater
than the levels reported above. (Too few whole-fish samples were reported to
provide an adjustment.) N11ป1 and Oliver (1989, 1983) found that whole-fish PCB
concentrations exhibited up to five-fold higher PCB concentrations than PCB
levels In muscle in four salmonid species studied. Using their results, and
ignoring possible species differences, whole-fish PCB body burdens nay be
approximately five times greater than the levels reported above.
Although it is possible to adjust the fillet PCB concentrations to whole-
fish values, based on literature studies such as those cited above, this
adjustment was not considered necessary for several reasons. No site and
species-specific data to make the adjustment was found. A factor of five
adjustment as suggested by the N11m1 and Oliver studies would not affect this
preliminary assessment. No site-specific data on the type or amount of fish
actually consumed by fish-eating wildlife in the Upper Hudson Site vicinity is
available; approximate dietary intake 1s estimated from published sources where
needed. Thus, uncertain foraging and food preferences and unknown fish Intake
were considered much larger uncertainties than possible differences between
fillet versus whole-fish PCB levels. Finally, the data currently available are
Inadequate to provide quantitative risk estimates for fish or fish-eating
wildlife, precluding the need for such an adjustment.
Estimated Fish Dietary Intake
Brown bullhead are exposed to PCBs by ingestion of sediment matter, dietary
Intake, including various species of cladocerans, copepods, chironomids, algae
and fish, ingestion of ambient water and direct absorption. For some dietary
components of brown bullhead, PCB concentrations may be surmised from the results
of the NYSDOH long-term multlplate monitoring. These samples reflect PCB
concentrations in algae, zooplankton and macroinvertebrates as well as organic
detrltal matter and inorganic sediment components. Multlplate samples taken from
the Upper Hudson at Fort Miller have shown little change over the period of
B.7-27
-------
measurement (1976-86), with typical values on the order of 10 - 60 |ig/g (mg/kg)
on a dry-weight basis. As mentioned above, chironomid larvae tissues contained
approximately five to greater than seven mg/kg PCBs when exposed to ambient
water. Larvae exposed to sediments showed approximately a ten-fold Increase over
these levels. Although their major dietary component consists of several species
of zooplankton and aquatic Insects, including chlronomids, bullhead also feed
heavily on both dead and live fish. Studies on adult black bullhead Indicate
that fish may represent from 1 to over 50 percent (by volume) of the total
dietary Intake (Repsys et a7. 1976; Applegate and Mullan, 1966). Using the 1986-
1988 fish monitoring data described above, average dietary PCB concentrations in
the thrde prominent fish species in the Upper Hudson that are possibly consumed
by brown bullhead range from approximately >2 to <50 mg/kg.
Largemouth bass are carnivorous game fish. As juveniles, largemouth bass
feed predominantly on various species of cladocerans or water fleas and other
zooplankton. In contrast, adults are opportunistic; top carnivores feed
primarily on fish, but also consume large Insects, crayfish and other benthic
Invertebrates (PfHeger, 1978).
Stomach contents of bullhead and bass have not been analyzed to determine
food preferences and PCB levels in their food. Such studies may prove useful 1n
the future to determine mechanisms and rates of PCB uptake in fish. Results of
multlplate, short-term chironomid sampling and fish sampling all Indicate PCBs
in these species to be on the order of several to >10 mg/kg. Thus, average PCBs
in fish diets 1n the Upper Hudson may contain on the order of several to >10
mg/kg of PCBs.
Herring Gull
F1sh-eat1ng birds are exposed to PCBs in the fish they consume and, to a
lesser extent, through Ingestion of water. Herring gulls are essentially
opportunistic and may feed on a variety of fish, shellfish, Insects and carrion.
Hhittow and Rahn (1984) report that their diet consists of approximately 50
percent fish. Given that the recent PCB levels in the Upper Hudson fish fillet
B.7-28
-------
samples (summarized above) are approximately >2 to <50 mg/kg, a weighted
concentration of PCBs 1n gull diets 1n this area, assuming their other dietary
sources are free of PCBs, may be approximately 1-25 mg/kg.
The PCB dose, expressed as the average dally PCB Intake per un dy
weight, depends both on PCB levels In the food, the quantity of food Ingested and
body weight. Herring gulls weigh from 0.5 to 1.3 kg and consume approximately
20 percent of their body we1ght(BW)/day. Assuming a range of dietary concentra-
tions of 5 to 50 mg/kg, a body weight of 1 kg and a da^ly Intake rate of
approximately 0.1 kg f1sh/day (50 percent fish), the dally dietary PCB exposures
for herring gulls Is estimated to range from 0.1 to 2.5 ng/kgw-d. Actual
exposures would vary depending on PCB concentrations 1n fish typically eaten, the
percentage of fish In their diet and other dietary intakes containing PCBs.
Birds 1n the site vicinity have not been monitored for PCBs. Therefore,
the relationship between the amount of PCBs Ingested and the tissue levels in
birds (say gulls) remains unknown. One approach to estimate PCB body burdens,
which Ignores the mechanism and rates of PCB transfers from food to body tissues,
relies on the use of bloaccumulatlon factors (BAFs) or the ratio of PCBs 1n bird
tissues to the PCB concentrations 1n their diet. Braun and Norstrum (1989) of
the Canadian Wildlife Service have determined from field studies on herring gulls -
in the Great Lakes that these birds accumulate approximately 93 times the PCB
concentration 1n their diet (BAF tlMM - 93). This BAF 1s based on PCB levels of
approximately 0.5 mg/kg measured 1n alewife (fish), a major part of the diet of
the herring gull 1n their study. Braun and Norstrum determined that the eggs of
the herring gulls studied accumulated approximately 32 times the PCB concentra-
tion In whole fish (BAF,,, - 32). Adopting these BAF values as approximations and
assuming as discussed above that the weighted-average gull diet contains
approximately 1-25 mg/kg PCBs, the resulting PCB levels in herring gulls would
be approximately 93 - 2,325 mg/kg (whole body) and 32 - 800 mg/kg (egg). Without
field sampling to confirm PCB levels in fish-eating birds 1n the vicinity, these
estimates are very uncertain.
B.7-29
-------
Rink
According to Mnscombe et a7. (1982) and Aulerlch (1973), the mink's diet
1s comprised of approximately 30 to 50 percent fish. Other dietary components
Include crayfish, amphibians and other small mammals. Assuming mink consume 50
percent of their diet as fish from the Upper Hudson, they could Ingest a weighted
average of 1 - 25 mg/kg 1n their diet. Newell et a7. (1987) Indicate the average
body weight and dally food consumption for mink are 1 kg and 0.15 kg/d,
respectively. Based on these assumptions, the mink dally Intake or dietary dose
of PCBs could be on the order of 0.15 - 3.8 mg/kg-day. Because PCB toxicity to
mink has been studied based on the PCBs 1n the mink diet and not on the PCB
levels in mink body tissues, possible levels of PCBs In mink are not estimated.
B.7.4 Toxicity Assessment
B.7.4.1 Types of Toxicity
The toxicity of PCBs to aquatic and terrestrial organisms can vary
considerably depending on congener and Aroclor composition. Large differences
1n factors such as percent chlorine, solubility, congener structure, organism
sensitivity and species-specific sensitivity contribute to the overall complexity
In evaluating PCB toxicity. In spite of the congener- and organism-specific
toxicity response, 1t Is generally true that PCBs are largely chronic toxicants
In the environment, rather .than acutely toxic. Ambient concentrations of PCBs
1n the environment are frequently not high enough to pose acute toxic effects,
such as death, but exhibit a cumulative effect whereby toxicity Increases as the
length of PCB exposure Increases. Furthermore, PCB toxicity generally Increases
as the degree of PCB chlorlnatlon increases. Exceptions to this pattern Include
highly chlorinated congeners, which may be physically hindered from accumulating
in tissues, and congeners with a low chlorine content, which are more bio-
available and, thus, are more readily metabolized. (The toxicity of these less
highly chlorinated congeners may be due to the phenolic metabolites produced
during metabolism.)
B.7-30
-------
A further complexity In the determination of PCB toxicity arises as a
result of the differing toxlcologlcal characteristics among congeners with very
similar chlorine content. For example, 20 of the 209 possible PCB congeners
exhibit coplanar properties due to non-ortho substitution In the blphenyl ring
(Tanabe, 1988). Coplanar molecules have an approximately flat structure whereby
the two benzene rings of the PCB molecules lie 1n the same plane. Examples
Include 3,3*,4,4',5-pentachloroblphenyl and 3,3',4,4',5,5'-hexach1orob1pheny1,
compounds that are structurally similar to 2,3,7,8-tetrachlorodibenzo-p-d1oxin
(TCDD) and 2,3,4,7,8-pentachlorodlbenzofuran. Although coplanar PCBs are
constituents of Aroclor mixtures In minute quantities, consideration of their
dioxin-llke responses is warranted. Studies have indicated that coplanar
congeners are 10 to 1,000 times more toxic than similar non-coplanar PCB
congeners (Eisler, 1986). Such unique responses have initiated the development
of Toxic Equivalency Factors (TEFs), a process that consists of normalizing
enzyme induction potencies against known TCDD-induced responses (Safe, 1990).
This TEF approach 1s gaining acceptance for assessing biological risks. Its use,
however, depends upon reliable congener-specific PCB analyses, which are largely
unavailable for the Hudson.
In addition to coplanar toxicity research, other studies implicate the
contaminant dibenzofuran as the principal toxic agent 1n PCB mixtures.
Dibenzofurans are co-products of most Aroclor mixtures and are also produced as
PCB metabolites or as photochemical derivatives (Kalmaz and Kalmaz, 1979).
Toxic effects resulting from chronic exposures of PCBs include enzyme
inhibition and induction, 1mmunotox1c1ty, liver disorders, tissue lesions,
reproductive Impairment, reduced growth and, 1n some cases, death. The varying
species tolerances to Aroclor mixtures and specific congeners can result from the
presence or absence of detoxifying mechanisms. Such mechanisms Include the
cytochrome P-450 mixed function oxidase (NF0) system or the 3-methylcholanthrene-
1nduc1b1e form (3-MC) of cytochrome P-450 referred to as aryl hydrocarbon
hydroxylase (AHH) or cytochrome P-448 (Slpes and Gandolfl, 1986). These
microsomal systems are responsible for the Initial biotransformation of a wide
variety of xenoblotlcs, including several PCB congeners. Induction of 3-MC
B.7-31
-------
activity, which Includes AHH and ethoxyresorum O-deethylase (EROD) Induction, 1s
usually associated with the transformations of coplanar congeners, whereas non-3-
MC and HFO activity 1s considerably less specific.
The detoxifying process catalyzed by enzymatic oxidative reactions Is
essential to the organism's ability to metabolize congeners to more soluble
products, which are excreted 1n urine and bile. Studies Indicate that congeners
having ortho-pos1t1oned chlorines may be easily metabolized and rapidly released,
whereas congeners with chlorines In the para position are typically less soluble
and are subsequently retained by various tissues (USEPA, 1980).
Microsomal enzyme Induction has been used extensively as toxicity endpolnts
for several aquatic and terrestrial species, Including fish (Hucklns et a7.,
1988), birds (Miranda et ซ7., 1987) and various mammals (Tanabe, 1988). Although
Induction of MFO activity 1s often correlated with PCB exposure, quantification
of adverse effects related to Induction Is limited. Consequently, enzyme
Induction does not Implicitly constitute Irreparable harm.
PCBs are most often considered to be chronic toxicants, a characteristic
that 1s associated, at least 1n part, with their tendency to bloconcentrate or
bloaccumulate. Biological uptake of PCBs and accumulation In body tissues are
Influenced by several factors, such as direct exposure In water and sediment, PCB
levels In food, organism 11p1d composition, and depuration (loss/release) rates.
Both assimilation and depuration of PCBs tend to be related to chlorine content
of PCBs.
The following sections present toxicity Information, bloconcentratlon and
bloaccumulatlon factors and proposed guidelines and criteria for PCBs In water,
sediments and wildlife. To the extent possible, bloconcentratlon factors (BCFs)
are used for laboratory-derived ratios of PCBs In organisms to PCBS 1n water.
Bloaccumulatlon factors (BAFs) refer to values based on field measurements. (See
glossary for a definition of BCF versus BAF.) Organisms for which toxicity and
uptake data are available include plants, plankton, aquatic Invertebrates, fish,
birds and a limited number of terrestrial and sem1-aquat1c mammals.
B.7-32
-------
B.7.4.2 Toxicity Literature Review
Plants
Reduced growth, reproductive alterations and inhibited photosynthesis are
some of the documented responses exhibited by plants when exposed to PCBs.
Exposure may be due to foliar contact from aerial deposition, vapor sorption of
volatilized PCBs and root contact or uptake with contaminated soils. The latter
Is far less detrimental to plant survival (Pal et al., 19&i). Direct contact
with PCBs 1n water 1s likely to be an Important exposure pathway for aquatic
plants.
Pal et <7. (1980) Investigated the effects of PCBs on carrots, sugarbeets,
corn and tomatoes. Weber et al. (1979) found that soybeans exhibited decreased
plant height and fresh weight when exposed to Aroclor 1254 concentrations of 1
to 1,000 ppm. Duckweed, an aquatic plant, suffered decreased colony formation
when exposed to 5' mg/7 Aroclor 1242 1n water and complete growth Inhibition at
100 mg/7 (Mahantl, 1975).
Buckley and Tofflemlre (1983) reported that plants In the vicinity of the
Upper Hudson River have accumulated PCBs 1n the leaves and, to a lesser extent,,
in the stems, with little translocation to plant fruits (see Section B.3.5).
Several studies have examined PCB uptake from contaminated soils (Moza et <7.,
1976; Iwata et a7., 1974). Iwata et al. (1974) reported that translocation of
PCBs from carrot roots to other plant tissues (e.g., leaves) was minimal: 97
percent of the PCBs were retained 1n the outer peel and only 3 percent
translocated Into other plant tissues. Studies by Sawhney and Hankln (1984) have
Indicated that uptake from PCB-amended soils by beets, turnips and beans was
ultimately influenced by PCB solubility and volatility. Sawhney and Hankln found
that these plants accumulated Aroclors with the following preferences: Aroclor
1248 > Aroclor 1254 > Aroclor 1260.
B.7-33
-------
Pal et a7. (1980) reported BAFs, defined as the ratio of PCB concentration
1n plant tissue to PCB concentrations 1n the soil, ranging from 0.002 to 0.96 for
several crop plants. In contrast, reported accumulation factors for aquatic
plants ranged from 289 to 814 (Moza et a7., 1976).
Planktonlc Species
The range at which adverse effects are exhibited by zooplankton and algae
1s quite wide, varying from exposures of 1.3 to over 2,000 iปg/7 (USEPA, 1980).
When exposed to Aroclor 1254, various species of green algae exhibit adverse
reactions, such as reduced carbon fixation and reduced growth rates at
concentrations as low as 0.1 to 10 yg/7 (USEPA, 1980). Exposure to Aroclor 1242
at 30 |ig/7 resulted In no significant effects to the amphlpod Hyalella azteca,
whereas exposures of 100 yg/7 resulted 1n complete mortality (Borgmann et al.,
1990). Daphnia magna, a common aquatic organism used In establishing toxicity '
benchmarks, 1s very sensitive to PCB exposure. Lethal exposures for this species
range from 1.2 pg/7 for Aroclor 1248 to 2.5 |tg/7 for Aroclor 1254 (Elsler, 1986).
Table B.7-2 summarizes these effects.
As with most aquatic species, zooplankton and algae accumulate PCBs in
concentrations that greatly exceed ambient water column levels. BCFs for various
species of phytoplankton have been reported 1n the range of 1,000 to 100,000
(USEPA, 1980). Elsler (1986) reports a BCF value of 47,000 for Daphnia Magna.
Aquatic MacroInvertebrates and Insects
Although very few reports of the effects of PCBs on aquatic Insects and
macrolnvertebrates exist, many of these species appear to exhibit sensitivities
similar to those of Daphnia magna. The evaluation of exposure and toxic effects
for aquatic Insects is complicated by llfecycle stages, whereby insects are
submerged as larvae, but winged upon emergence and, thereby, removed from a more
contaminated medium, e.g., water. This difference 1n morphology and physiology
contributes to the difficulty 1n predicting exposures and subsequent effects.
In contrast, aquatic invertebrates such as oligochaete worms, mussels, snails and
B.7-34
-------
crayfish are less mobile benthlc organisms, lacking multiple life stages with
limited morphological change. These aquatic Invertebrates may be continuously
exposed to contaminated water and sediments, and, as a result, exhibit more
predictable body burdens than organisms exposed to PCBs for only a portion of
their llfecycle.
Elsler (1986) reports that chlronomlds (midges) are sensitive to Aroclor
1254 at concentrations of 0.5 to 1.2 |ig/7. Exposure to Aroclor 1254 at 1.5 jig/J
resulted In complete Inhibition of mosquito emergence (Sanders and Chandler,
1972). The 7-day LCM values, defined as the PCB concentration In water
resulting 1n 50 percent mortality, for crayfish exposed to Aroclor 1242 and 1254
are 30 jig/7 and 100 |ig/7, respectively (Mayer, et i7., 1977). These authors
report that grass shrimp exposed to Aroclor 1254 exhibited a 7-day LCW value of
3 ng/7.
Chlronomid larvae monitored during NYSDOH's short-term blomonltorlng study
1n the Upper Hudson exhibited BAFs on the order of 10,000 to 200,000 (Simpson
ซ7., 1986).
Fish
Much of the available literature on PCB toxicity 1n aquatic environments
addresses the effects of PCBs on various fish species. Common responses 1n fish
exposed to PCBs Include enzyme Induction and inhibition, reproductive Impairment,
lesions, tumors, fin rot and, at high enough concentrations, mortality (USEPA,
1980). Mayer et a7. (1977) report long term LCW values for trout, blueglll and
catfish ranging from 3 to 177 |ig/7 (Table B.7-2).
Studies of several freshwater species In Lake Michigan report that PCB
concentrations of 0.01 |ig/7 In water and 1 pg/g In food caused significant
reduction 1n lake trout survival (Wlllford et a7., 1981). Other studies have
shown that exposures of 0.7 and 1.5 |ig/7 cause mortality In brook trout and
largemouth bass, respectively (USEPA, 1980). Although very few studies address
the Issues of PCB toxicity associated with tissue residues, a PCB concentration
B.7-35
-------
of 0.4 mg/kg (ppm) 1n whole-body fish tissues has been observed to cause
reproductive Impairment 1n rainbow trout (Elsler, 1986). Trout are particularly
sensitive to PCBs and other species may be more tolerant to PCB exposures.
Laboratory-derived BCFs for several species of freshwater fish are on the
order of 10,000 to 100,000 (USEPA, 1980; Mayer et a7., 1977). BAFs obtained from
field surveys have ranged from 120,000 for rainbow trout to over 1,600,000 for
lake trout (USEPA, 1980). As explained 1n Section B.4, the slope of the
regression between PCBs 1n fish fillets (expressed on a lipid basis) and PCBs 1n
the water column Indicates that BAFs for Hudson River samples are on the order
of 10*. Recent studies by Jones et <7. (1989) Indicate that caged fathead
minnows studied 1n the Upper Hudson selectively accumulated tetra- and penta-
chlorinated blphenyls, suggesting a congener-specific bloaccumulatlon preference.
Birds
Aquatic and terrestrial birds have exhibited behavioral changes,
reproductive Impairment, enzyme Induction, reduced growth and mortality when
exposed to PCBs. Reported acute LDM values, defined as the dose resulting 1n
50 percent mortality, for various birds range from 604 to over 6,000 mg/kg
(concentration 1n the diet) for northern bobwhlte and from 1,975 to 3,182 mg/kg
for mallards (Elsler, 1986). It 1s evident from these values that birds may
exhibit a high degree of Initial resistance to PCB toxicity. Based on a six-week
study of white leghorn hens, Brltton and Huston (1973) reported that dietary
levels of Aroclor 1242 at 20, 40, and 80 mg/kg yielded PCB concentrations 1n egg
yolks of 6.2, 5.4, and 5.6 mg/kg, respectively, and resulted 1n a significant
decrease 1n hatchabillty and growth. Newell et ซ7. (1987), citing the Brltton
and Huston (1973) study, Indicate a No Observed Effect Level (NOEL) of 0.224
mg/kg-day as a dally intake dose. [It Is unclear from the Newell et el. (1987)
report whether they are referring to the same Brltton and Huston study, because
Newell et a7. report that the Brltton and Huston study was for nine weeks, using
Aroclor 1248. The cited Brltton and Huston study was based on six-week feedings
of Aroclor 1242.] Elsler (1986) reports that 40 mg/kg 1n the diet of doves
caused changes 1n mating behavior.
B.7-36
-------
Although several studies have Investigated the effects of PCB levels on
eggshell thickness in species such as cormorants (Koeman et al., 1973), screech
owls (Anne et a1.t 1980) and western grebes (Mndvall and Low, 1980), conclusive
evidence correlating PCB levels with eggshell thinning does not exist. It Is
possible, however, that eggshell thinning may be ascribed to the synergistic
effects of PCBs, organochlorlne Insecticides and their metabolites (Llndvall and
Low, 1980).
Birds may store relatively high concentrations of PCBs in fatty tissues,
muscle, liver and brain. It appears that depuration 1s generally slow, but
significant reductions 1n PCB body burdens have been observed as a result of egg
production. Screech owls exposed to a diet containing 3 ppm of Aroclor 1248 laid
eggs containing 3.9 to 17.8 ppm (Anne et a1.t 1980). Other dietary exposures of
PCBs-from 10 to 20 ppm have resulted 1n PCB concentrations 1n eggs of 14 to 16
ppm (Elsler, 1986).
Mammals
M1nk have been found to be very sensitive to PCBs. Aroclor 1254 fed to
mink at concentrations of 0.64 to 2 mg/kg has led to reproductive failure
(Platonow and Karstad, 1973; Aulerlch and Ringer, 1977). Long term LDW values'
of 0.1 mg/kg for hexachloroblphenyl, 8.6 mg/kg for Aroclor 1242 and 6.7 mg/kg for
Aroclor 1254 have also been reported (Elsler, 1986; Ringer, 1983).
Other sublethal effects exhibited by mammals Include lethargy, weight loss,
liver disorders, enzyme Induction and Inhibition and hormonal effects. Studies
on rabbits, monkeys, mice and rats Indicate that PCBs are also teratogenic and
carcinogenic (USEPA, 1980). Other studies have shown that PCBs may even alter
sleep and hibernation patterns 1n white-footed mice (Sanders and Klrkpatrick,
1977) and raccoons (Montz et al., 1982).
6
B.7-37
-------
B.7.4.3 Proposed Criteria and Guidelines
There exist no promulgated standards for PCBs in surface water, sediments,
and wildlife. A number of federal and state agencies, Including USEPA, NOAA,
USFWS, and NYSDEC, have developed criteria and guidelines for assessing
environmental thresholds for PCBs as summarized 1n Table B.7-3.
Ambient Water Quality Criteria
The USEPA has established a freshwater Ambient Hater Quality Criterion
(AUQC) for the protection of aquatic life Including algae, zooplankton, aquatic
Insects, Invertebrates, fish and piscivorous wildlife. The 24-hour average AUQC
for PCBs 1n ambient (surface) water Is 0.014 |ig/7 (USEPA, 1980). As set forth
1n USEPA guidance, this criterion 1s deemed to provide a conservative threshold
level of PCBs 1n freshwater aquatic environments to protect against long-term
exposure and bloaccumulatlon.
Mink, Identified as a species very sensitive to PCBs, are at the apex of
the aquatic food chain. The AWQC for PCBs 1s based, at least 1n part, on the
assumption that mink represent a highly sensitive wildlife species In aquatic
environments. Surface water PCB concentrations less than the AUQC are thought
to protect other, less sensitive species In the aquatic environment. The AUQC
was established using a dietary threshold of 0.64 mg/kg linked to reproductive
failure 1n mink (Platanow and Karstad, 1973). Additionally, as fish are a major
component of the mink diet, PCB levels In fish were taken as the major aquatic
exposure pathway for PCB uptake In mink. Adopting BCFs, which the guidelines use
to relate PCB levels In water to those In fish, the AUQC guideline 1s given by
(USEPA, 1980) as:
AUQC - 0.64 mo/ko - 0.014 |ปg/7
45,000 mg/kg / mg/7
B.7-38
-------
where the BCF (45,000 mg/kg / mg/7) Is a geometric mean of 3 BCFs for rainbow and
brook trout (USEPA, 1980).
The USFWS has reviewed the applicability of the AWQC guideline and
concluded that the criterion 1s environmentally protective, but also concludes
that 1t should be changed to reflect a aaximja of 0.014 jig/7 rather than a 24-
hour average (Eisler, 1986). The New York State Ambient Water Quality 6u1dance
Criterion 1s established at 0.001 |tg/7, derived from an acute (96-hour) LCM
(NYSDEC, 1985).
Sediment Quality Guidelines
The development of criteria or guidelines for PCBs 1n sediments 1s under
study by the USEPA Science Advisory Board. Several approaches are being
considered Including the Apparent Effects Threshold (AET), co-occurrence analysis
(C0A) and Equilibrium Partitioning (EP) methods.
The AET approach, developed from data derived In Puget Sound, Washington,
combines chemical concentration data In sediments and a biological indicator of
Injury, such as sediment bioassays and altered benthic Infauna abundance, to
determine the contaminant concentrations above which significant biological
response can be expected (PTI, 1988). Long and Morgan (1990) summarize the AET
values for several Puget Sound studies and report AET values for San Francisco
Bay, California. A range of AET values for PCBs, from 0.05 to 3.1 mg/kg (ppa
dry weight), were reported for saltwater environments. Although the AET approach
has been used primarily in saltwater systems, a similar approach nay be useful
In the Upper Hudson.
The Co-occurrence (C0A) method, like the AET, relies on field-collected
data matching chemical concentrations with observed biological effects. CQA
bloassay results for Hudson-Raritan Bay Indicated biological responses to PCBs
as low as 0.64 mg/kg. CQA results for freshwater environments Indicated effects
ranging from 0.13 mg/kg to 1,141 mg/kg (Long and Morgan, 1990).
B.7-39
-------
The EP nethod Is based on theoretical partitioning of PCBs on sediments
Into the Interstitial pore space of the sediment. A pore water PCB concentration
equal to the AWQC is used to set a target PCB threshold 1n sediment. Because
sediment-water partitioning depends on chemical and sediment properties, no
single sediment value can be calculated. For Aroclor 1254, the EP value is 0.42
ppm (dry weight) for sediment with a one percent total organic carbon fraction
(Long and Morgan, 1990).
As part of the National Status and Trends program, NOAA has recently
reviewed studies that examine the AET, EP and other methods, such as spiked
sediment bioassays (Long and Morgan, 1990). NOAA reviewed over 150 reports or
studies on contaminated sediments around the United States. From these, Long and
Morgan Identified 34 that contained adequate and reliable data from which to
synthesize sediment guideline values. Using these, NOAA has determined an
Effects Range (ER). This range 1s determined by the 10th percentile (the low ER,
or ER-L) and median (ER-M) chemical concentration thresholds for the 34 studies.
The findings of NOAA's investigation Indicate that PCB concentrations as low as
0.050 ppm (ER-L) with a median of 0.4 ppm (ER-M) have been shown to be associated
with biological response. It must be emphasized that NOAA states clearly that
the sediment thresholds are Intended only as guidance values possibly to be used
to Identify the need for site-specific monitoring studies, and do not represent
NOAA standards (Long and Morgan, 1990).
Guidelines for PCBs 1n Fish
Elsler (1986) presents a synoptic review of PCB toxicity 1n wildlife and
presents proposed dietary and body tissue guidelines of the USFWS for PCBs In
fish. The USFWS reported whole-body and egg-res1due guidelines for rainbow trout
of 0.4 mg/kg and 0.33 mg/kg, respectively. These tissue residues have been
associated with reproductive impairment, decreased hatch success and fry
deformities (Elsler, 1986). As mentioned previously, trout are particularly
sensitive to PCBs and other hazardous organic compounds, such that the levels
proposed by the USFWS are deemed to be conservative and protective of other, less
sensitive fish.
B.7-40
-------
As part of the Niagara River Biota Contamination Project, NYSDEC Division
of F1sh and Wildlife developed fish flesh criteria for piscivorous wildlife
(Newell et <7. 1987). These criteria, which are based on a review of mammalian
and avian laboratory studies, establish threshold PCB levels 1n fish set to
protect f1sh-eat1ng wildlife. Thus, they are thresholds for fish consumers,
rather than values set to protect the health of fish. Mink, because of their
largely fish diet and demonstrated sensitivity to PCBs, are the piscivorous
wildlife species used to establish the criterion. The noncardnogenlc fish flesh
criterion for the protection of piscivorous wildlife established by NYSDEC Is
0.13 mg/kg. This criterion was determined by using an adjustment factor of 0.2
to the dietary Intake threshold (0.64 mg/kg) causing reproductive impairment In
the study by Platanow and Karstad (1973). The adjustment factor Is used to
convert this Lowest Observed Effect Level (LOEL) to a NOEL.
Guidelines for PCBs In Birds
The most sensitive body tissue indicators of potential risks to aquatic and
terrestrial birds have been found to be PCB accumulation in brain and egg
tissues. Stickel et <7. (1984) reported that Aroclor 1254 concentrations In
brain tissue ranging from 349 to 763 ppm (wet weight) resulted In mortality in
cowblrds, grackles, starlings and red-w1nged blackbirds. Birds that survived the
same dietary dosage of 1,500 ppm had brain PCB residues of 54 to 301 ppm. This
study concluded that brain PCB levels greater than 301 ppm are highly correlated
with death. Based on these findings as well as other studies correlating brain
PCB residues of 76 to 180 ppm with mortality 1n cormorants, the USFWS has
recommended a threshold brain PCB concentration of 54 mg/kg (Eisler, 1986).
Although brain PCB residues are good Indicators of possible lethality,
residues of PCBs in eggs provide a more sensitive endpolnt for determining lethal
and sublethal effects as well as population effects as a consequence of decreased
hatch success. Based on the findings of Brltton and Huston (1973) discussed
earlier, the USFWS has recommended a whole egg PCB concentration of 0.4 ppm (wet
weight) as a threshold value (Kublak, 1991, pers. comm.).
B.7-41
-------
Guidelines for PCBs 1n Manuals
Guidelines for the protection of mammals are most conservatively
represented by dietary studies on mink. Such proposed guidelines Include the
findings of Platonow and Karstad (1973) whereby a mink diet of 0.64 mg/kg Aroclor
1254 led to reproductive Impairment. Ringer (1983) determined a NOEL 1n mink,
on the basis of body weight (BW) Intake, of 0.1 mg/kgM-d (approximately 0.67 ppm
in the diet) from a study where mink were given a dietary dose of 0.225 mg/kg*-
d, which resulted 1n significant reproductive Impairment. The USFWS recommended
dietary tolerance level for mink Is 1.54 |ig/kgw-d (Elsler, 1986).
B.7.5 Risk Characterization
This Interim ecological risk characterization 1s limited to comparisons of
PCB exposure levels 1n selected indicator species with: 1) proposed criteria and
guidelines for ecological protection; and 2) toxicity endpolnts or thresholds.
B.7.5.1 Ambient Nater
As discussed earlier, upper 95 percent confidence bounds on recent (1986-
89) mean concentrations of PCBs 1n the water column range from 0.034 (Waterford)
to 0.06 pg/7 (Fort Edward). These ambient levels are approximately two to five
times greater than the USEPA AUQC of 0.014 |ig/7; they are from 30 to 60 times
greater than the more conservative New York State water quality criterion of
0.001 |ig/7.
S1te-spec1f1c reports or monitoring studies of adverse effects to aquatic
life 1n the Upper Hudson resulting from PCB exposures do not exist. A review of
the toxicity studies summarized In Table B.7-2 reveals that the baseline PCB
concentrations In water are below levels that have been measured to cause aquatic
toxicity. Thus, although the water column PCB levels may be elevated somewhat
above AUQC values, toxlcologlcal data In the literature do not corroborate an
Imminent harm to algal, macrolnvertebrate, Insect and fish species from direct
contact with PCBs 1n water.
B.7-42
-------
B.7.5.2 Sediment
Concentrations of PCBs in sediments of the Upper Hudson, ranging from <1
ppm to >1,000 ppm are generally above various proposed sediment guldelin^levels
discussed previously. As reported in NOAA's National Status and Trends report,
with "very few exceptions" biological effects, such as reduced benthic
macroInvertebrate diversity or reduced chironomld and mayfly survival, were
observed in sediments with PCBs above 0.37 ppm (Long and Morgan, 1990).
'j
Long and Morgan attribute a "moderate" degree of - confidence 1n the
threshold effects values for PCBs In sediment. However, species richness/diver-
sity and species-specific biological responses to PCBs listed in the Long and
Morgan study are difficult to correlate with ecological risks in the Upper
Hudson. For example, Long and Morgan (1990) report that 0~7 ppb PCBs 1n
Mississippi River sediments caused reduced chironomld survival. Concentrations
of PCBs 1n the sediments of the Upper Hudson exceed these levels by orders of
magnitude, but, as discussed previously, chironomld species appear to be
increasing 1n abundance In the river.
Likewise, the PCB threshold concentration In sediment, according to the
equilibrium partitioning (EP) approach, 1s on the order of 2 ppm, using the
Aroclor 1254 partition coefficient and approximately five percent total organic
carbon, based on volatile solids content of Hudson River sediment. Sediments in
the Upper Hudson contain PCBs well above this threshold. Again, the AHQC is a
water column concentration based on bioaccumulation and protection of fish-eating
wildlife. Yet the connection between pore-water PCB concentrations and food-
chain uptake and transfers 1s largely uncertain. Thus, it is not clear what
ecological Impacts are occurring because of PCBs In the sedldfents.
B.7.S.3 F1sh
The USFWS has recommended a guideline of 0.4 ppm for PCBs 1n fish tissues
for the protection of fish. Recent PCB levels for fish (fillets) In the Upper
Hudson range from 2.3 to 48.7 ppm for brown bullhead and 3.1 to 12.6 for
B.7-43
-------
largemouth bass. The levels of PCBs 1n both Indicator fish species are over 10
tines the recommended USFWS guidelines.
The USFWS guideline 1s based on studies of rainbow trout 1n which whole
body residues of 0.4 ppm resulted 1n significant fry mortality and deformities
(USEPA, 1980; Elsler, 1986). Although trout are very sensitive to PCBs, they are
not a major species 1n the Upper Hudson site vicinity. It 1s unclear if present
PCB levels 1n the Upper Hudson fish species represent a significant detrimental
impact on the Indigenous population.
K1m et <7.(1989) conducted the only known study of the possible pathologi-
cal response of Upper Hudson fish to contamination the river. In their study,
brown bullhead were sampled near Griffin Island (River Mile 190) and Stillwater
(River Nile 175); these fish were designated the experimental fish from
contaminated stretches of the river. F1sh sampled near Corinth (River Hlle 200) 4
were designated as control samples. PCB levels 1n the contaminated fish ranged
from 16.3 to 102 mg/kg, with a mean of 38.3 mg/kg. PCB levels 1n the control
fish ranged from 0.38 mg/kg to 1.4 mg/kg, with a mean of 0.61 mg/kg. No
significant differences 1n gross abnormalities, such as outward physical
characteristics, were found 1n the experimental versus control fish. In
contrast, hlstopathologlcal (cell pathology) results indicated a statistically
significant Increase in bile-duct hyperplasia (abnormal cell growth) 1n the
contaminated fish (78 Incidences) compared to the control fish (13 Incidences).
Kim et al, (1989) note:
"The most significant finding was the high frequency of bile-duct
hyperplasia, accompanied by hepatic and renal hemosiderosis among
brown bullhead from the contaminated section of the Hudson River.
Since the hepatobiliary system plays a major role in removing toxic
materials from the blood after absorption from the gastrointestinal
tract and is the site where biotransformation and excretion of
xenoblotlcs 1s taking place, the observed bile-duct abnormalities
could serve as an Indicator of chemical contamination of an aquatic
environment."
B.7-44
-------
No single chemical could be Identified as the causal agent In their study,
although the authors note that studies by Norback and Weltman (1985) have shown
bile-duct hyperplasia to be a notable response of rats exposed to PCBs In their
diets during laboratory experiments. Further Indication that the observed
pathological differences were possibly Induced by toxic organlcs was the measured
Increase 1n cytochrome P-450 In the livers of two contaminated fish compared with
the levels in the control fish (K1m et al., 1989).
While the K1m et al. (1989) results Indicate that toxic organic contami-
nants, Including PCBs, hexachlorobenzne, and octachlorostyrene found in the fish,
are possible agents responsible for the observed pathological abnormalities, they
conclude that "the specific chemicals or metabolites responsible for the observed
bile-duct hyperplasia remain to be Investigated" (Kim et <7., 1989).
B.7.S.4 Fish-Eating Birds
Birds In the Upper Hudson area have not been monitored for PCBs. Weighted
concentrations of PCBs 1n the fish diet of herring gulls could be on the order
of 1 to 25 ppm, based on TAMS/Grad1ent estimates. The USFWS proposed dietary
guideline 1s 3.0 mg/kg for the protection of birds. This threshold concentration
Is based on feeding studies 1n screech owls where a diet of 3 ppm resulted in
high PCB residues 1n owl eggs. Several other studies cited previously have
reported that diets of 20 to 80 ppm 1n chickens caused significant reproductive
impairment. Elsler (1986) reports LDW values for mallard ducks of several
thousand ppm.
TAMS/Grad1ent estimates of PCBs 1n the fish diets of herring gulls are 1n
the same order of magnitude as the USFWS guidelines and lower than those levels
causing reproductive Impairment, as suggested by other studies. Additional
monitoring data will be required 1f PCB uptake 1n birds Is to be assessed.
Transfers of PCBs to developing eggs have been Implicated as sensitive
toxlcologlcal endpolnts 1n several species of birds. Kublak (pers. comm., 1991)
reports that PCB residues of 0.4 ppm In eggs of chickens decreased hatch success.
B.7-45
-------
E1sler (1986) Indicates that 16 ppm 1n turtle doves caused delayed development.
No site-specific information 1s available on the PCBs in eggs of local birds.
BAF values relating PCBs In the diet to PCBs 1n herring gull eggs were taken from
the literature and used to estimate PCB levels as high as 32 - 800 ppm. This
estimate, however, 1s highly uncertain. Furthermore, no Information Is available
on the effects of PCBs In the eggs of herring gulls.
B.7.S.5 Manuals
Like the bird pathway, possible dietary exposure levels of PCBs for fish-
eating mammals have not been monitored. Assuming 50 percent of their diet
consists of fish from the Upper Hudson, mink diets may contain on the order of
1-25 ppm PCBs. This dietary PCB concentration exceeds the 0.64 ppm level shown
to cause reproductive failure (Planatow and Karstad, 1973). These estimated
Intake levels also exceed the NYSDEC fish flesh criteria of 0.13 ppm (Newell, et
a7. 1987).
B.7.5.6 Summary
Measured baseline (1986-89) PCB concentrations in surface water of the
Upper Hudson exceed the USEPA ambient water quality criterion by as much as a
factor of five. Sediments, for which there are no clearly defined PCB criteria,
are contaminated with PCBs far above the <1 ppm guidelines outlined by NOAA as
levels Indicating biological effects. Levels of PCBs 1n Upper Hudson fish exceed
the USFWS recommended guidelines for trout by approximately an order of
magnitude. Finally, estimated PCB levels In the diets of fish eating birds and
mammals at the site appear to be on the same order or somewhat higher than
dietary levels recommended by USFWS and NYSDEC.
Aside from the comparison with published toxicity values and measured PCB
concentrations 1n Hudson River sediments, water and biota, no quantitative
ecological risks attributable directly to PCBs can be adequately presented based
on available site data. Future phases of the reassessment will address data
limitations and better define ecological risks due to PCBs in the River.
B.7-46
-------
SYNOPSIS
PHASE 1 FEASIBILITY STUDY
(Sections CI through G7)
The purpose of the Feasibility Study (FS) for the Hudson River PCB site is to identify
and evaluate alternatives for mitigating PCB contamination and controlling its effects on public
health or the environment By selecting from the range of technologies available for site clean-
up, a response can be formulated that is technically feasible, protects public health and the
environment, is cost-effective and is consistent with applicable or relevant and appropriate
requirements (ARARs). This Part provides the results of initial steps taken to date as part of
the comprehensive Feasibility Study for the project Three aspects of the FS are discussed: (1)
remedial objectives and response actions; (2) potential clean-up technologies; and (3) an initial
screening of technologies.
National legislation has established the purposes of remedial actions and the process by
which remediation alternatives are developed and evaluated (CI). Remedial objectives and
potential, general response actions are explained (C2).
Section 121(d) of SARA and the NCP require that CERCLA remedial actions comply
with all federal ARARs. State requirements must also be attained, if they are legally enforceable
and consistently enforced statewide. A listing of the specific federal and state requirements (C3)
cover three kinds of ARARs: chemical-specific (govern the extent of site cleanup); location-
specific (pertain to existing site features); and action-specific (pertain to proposed site remedies
and govern implementation of the selected site remedy).
The technologies and processes identified (C.4) are: containment, natural PCB
biodegradation in sediments, removal, disposal and treatment technologies. The treatment
technologies include physical, chemical, thermal and biological treatment Innovative
technologies (C.5) are also reviewed
An initial screening of technologies is performed (C.6). Although no particular
technology has been eliminated from further consideration in Phases 2 and 3, preliminary
judgments concerning remedial options are presented. Since the decision to remediate the Upper
Hudson's PCB-contaminated sediments will be significantly influenced by the feasibility and
costs of the technologies, recommendations for treatability studies are also presented (C 7).
Part C initiates the feasibility study process, which will continue through Phase 2 as
additional site characterization work is performed and will be finalized in Phase 3
-------
PAGE INTENTIONALLY LEFT BLANK
-------
PART C
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
CONTENTS
Page
C. PHASE 1 FEASIBILITY STUDY
C.l Introduction C.l-1
C.2 Remedial Objectives and Response Actions C.2-I
C.3 Potentially Applicable or Relevant and Appropriate
Requirements (ARARs) C.3-1
C.3.1 Definition of ARARs C.3-1
C.3.1.1 Applicable Requirements C.3-2
C.3.1.2 Relevant and Appropriate Requirements C.3-2
C.3.1.3 Other Requirements To Be Considered C.3-2
C.3.2 Development of ARARs C.3-3
C.3.2.1 Chemical-Specific ARARs C.3-3
C.3.2.2 Location-Specific ARARs C.3-3
C.3.2.3 Action-Specific ARARs C.3-4
C.3.3 Statutes and Regulations C.3-4
C.3.3.1 Federal Statutes and Regulations C.3-4
Toxic Substances Control Act (TSCA) C.3-4
15 USC 2601, 40 CFR 761
Resource Conservation and Recovery Act C.3-6
(RCRA) - 42 USC 6901, 40 CFR 260
1
-------
PART C
CONTENTS
(continued)
*
Paoe V
Comprehensive Environmental Response, C.3-7
Compensation and Liability Act (CERCLA) '
42 USC 9601, 40 CFR 372
Clean A1r Act - 42 USC 7401 C.3-7
C.3.3.2 New York State Statutes and Regulations C.3-8
New York Environmental Conservation Law C.3-8
Article 11, Title 5
New York Water Classification and Quality C.3-8
Standards
Groundwater Classifications C.3-9
New York State Constitution Article XV, C.3-9
Section 1-4
New York Environmental Conservation Law C.3-10
Article 21, Title 5
New York Environmental Conservation Law ' C.3-10
Article 24, Freshwater Wetlands
New York Environmental Conservation Law C.3-11
Article 25, Tidal Wetlands
C.4 Technology and Process Identification C.4-1
C.4.1 Containment C.4-1
C.4.2 Natural PCB Biodegradation in Sediments C.4-2
C.4.2.1 Aroclor Patterns C.4-2
C.4.2.2 Aerobic Biodegradation of PCBs C.4-3
C.4.2.3 Anaerobic Dechlorination C.4-5
1i
-------
PART C
CONTENTS
(continued)
Page
C.4.3 Removal Technologies C.4-7
C.4.4 Treatment Technologies C.4-8
C.4.4.1 Physical and Chemical Treatment Technologies C.4-8
Evaluation C.4-8
KOHPEG C.4-11
B.E.S.T. Solvent Extraction Process C.4-13
Low Energy Extraction Process (LEEP) C.4-14
Propane Extraction Process C.4-15
C.4.4.2 Thermal Treatment Technologies C.4-16
Introduction C.4-16
Low Temperature Desorption C.4-17
Rotary Kiln Incineration C.4-18
Multiple Hearth Incineration C.4-19
Fluid Bed Incineration C.4-19
Circulating Fluid Bed Incineration C.4-20
Conveyor Furnace C.4-21
Electric Pyrolyzer C.4-22
Contract Thermal Destruction C.4-23
Status of Thermal Destruction Technologies C.4-24
iii
-------
PART C
CONTENTS
(continued)
Paoe
C.4.4.3 Biological Treatment Technologies
C.4-25
Bioremedlatlon
C.4-25
In Situ Approach
C.4-27
Land-Based Approach
C.4-29
Bloreactor Approach
C.4-30
C.4.5 Disposal Technologies
C.4-32
C.5
Innovative Treatment Technologies (USEPA SITE Program)
C.5-1
C.6
Initial Screening of Technologies
C .6-1
C.7
Treatability Studies
TABLES
Tables Located in Volume-1 (2 of 2)
C.7-1
C.2-1
Remedial Technologies and Process Options
C.3-1
Potential Chemical-Specific ARARs and Criteria,
Guidance
Advisories and
C.3-2
Potential Location-Specific ARARs and Criteria,
Guidance
Advisories and
C.3-3
Potential Action-Specific ARARs
C.4-1
Physical/Chemical Technologies Reviewed by NUS (1984) and NPI (1985)
C.4-2
Initial Screening of Physical/Chemical Treatment Processes
1V
-------
PART C
CONTENTS
(continued)
C.4-3 Bench- and Pilot-Scale Tests of Physical /Chemical Sediment Treatment
Technologies
FIGURES
Figures Located 1n Volume 1 (2 of 2)
C. 1-1 Overview of the FS Process
C.4-1 PCB Content and Composition of Core 18-6
C.4-2 KOHPEG Process Flow Diagram
C.4-3 B.E.S.T. Process
C.4-4 LEEP-Low Energy Extraction Process
C.4-5 Propane Extraction Process
C.6-1 Response Actions and Associated Generic Technologies Retained for
Further Analyses
v
-------
PART C
CONTENTS
(continued)
PAGE INTENTIONALLY LEFT BLANK
vi
-------
C. PHASE 1 FEASIBILITY STUDY
C.l Introduction
Remedial actions, as defined by 300.5 of the National Contingency Plan
(NCP), are those responses to releases that are consistent with a permanent
remedy to protect against or minimize release of hazardous substances, pollutants
or contaminants so they do not migrate to cause substantial danger to current or
future human health and welfare or the environment.
In formulating a remedy, CERCLA requires USEPA to emphasize risk reduction
through destruction or treatment of hazardous waste. Section 121 of SARA
establishes a statutory preference for remedies that permanently and significant-
ly reduce the mobility, toxicity or volume of hazardous waste over remedies that
do not use such treatment. Section 121 also requires USEPA to select a remedy
that is protective of human health and the environment, is cost-effective and
utilizes permanent solutions and alternative treatment technologies to the
maximum extent practicable. Furthermore, Section 121 requires that, upon
completion, remedies must attain applicable or relevant and appropriate
requirements (ARARs), unless specified waivers are granted.
Section 300.430 of the NCP, 1n conjunction with USEPA guidance on
conducting a Feasibility Study (FS), sets forth the development and evaluation
process for remedial alternatives (USEPA, 1988). This process consists of the
following steps:
Identify the nature and extent of contamination and threat presented
by the release (300.430[d][2]);
Identify general response objectives for site remediation
(300.430[e][2][1J);
Identify and evaluate remedial technologies potentially applicable
to wastes and site conditions (300.430[e][2][11]);
C.l-1
-------
Develop alternatives to achieve site-specific response objectives
(300.430[e][2][111]);
Conduct Initial screening of alternatives (300.430[e][7]); and
Conduct detailed analysis of alternatives (300.430[e][9]). V
As an Initial step, both CERCLA and the NCP require Identification of the
nature and extent of site contamination. The nature and. distribution of
contamination and the threat posed by the release of contaminants.from the Upper
River has been presented 1n Part B. Phase 2 will continue to define the nature
and extent of contamination. This part Identifies potential clean-up technolo-
gies and presents an Initial screening of these technologies.
Figure C.l-1, an overview of the FS process as described above, highlights
the Phase 1 FS activities reported here.
C.l-2
-------
C.2 Remedial Objectives and Response Actions
Remedial action objectives are developed In order to set goals for
protecting human health and the environment early 1n the alternative development
process. The goals should be as specific as possible, but should not Unduly
limit the range of alternatives that can be developed.
In the original 1984 FS, the contaminant of Interest was PCBs, which remain
the specific contaminant of Interest for this reassessme*.t. The media of
Interest are the Upper Hudson River sediments.
It 1s apparent from the preliminary assessments that the impact of PCBs on
aquatic life and consumers of aquatic life will drive the clean-up of the site.
Target levels for clean-up will need to be determined once these assessments are
complete.
General response actions for remediation of the Hudson River PCB site
include the following:
1) No Action;
2) Containment;
3) In situ treatment;
4) Complete or partial removal with on-site or off-site disposal; and
5) Removal with on-site or off-site treatment and disposal.
Table C.2-1 lists general response actions for the Hudson River site and
identifies the technologies screened for each response action. A no action
alternative will be considered throughout each phase of the FS. Under the no
action alternative, contaminated river sediments would be left in place without
treatment or containment. Institutional controls such as fishing bans, site
access restrictions and monitoring may be continued.
C.2-1
-------
Remedial technologies and processes associated with the general response
actions are discussed In the following sections. A preliminary assessment of
each technology's applicability and state of development 1s also provided 1n
order to focus succeeding FS phases and the Phase 2 field Investigations.
C.2-2
-------
C.3 Potentially Applicable or Relevant and Appropriate Requirements (ARARs)
C.3.1 Definition of ARARs
Section 121(d) of SARA and the NCP require that CERCLA remedial actions
comply with all federal ARARs. State requirements must also be attained under
Section 121(d)(2)(c) of SARA, if they are legally enforceable and consistently
enforced statewide. ARARs are used to determine the appropriate extent of site
clean-up, Identify and formulate remedial action alternatives and govern the
Implementation and operation of the selected action. According to SARA,
requirements may be waived by USEPA, provided protection of human health and the
environment 1s still assured, under the following six specific conditions:
The selected remedial action 1s an Interim remedy;
Compliance with such requirements will result 1n greater risk to
human health and the environment than alternative options;
Compliance with such requirements 1s technically Impracticable from
an engineering perspective;
The selected remedial action will provide a standard of performance
equivalent to other approaches required under applicable
regulations;
The requirement Is a state requirement that has been Inconsistently
applied; or
Attainment of the ARAR would entail extremely high costs relative to
the added degree of reduction of risk afforded by the standard
(I.e., fund balancing).
To consider ARARs and, more Importantly, to Incorporate consideration of
ARARs 1n the FS and remedial response processes, the NCP and SARA have defined
both applicable requirements and relevant and appropriate requirements, as
described below.
C.3-1
-------
C.3.1.1 Applicable Requirements
Applicable requirements are those federal and state requirements that would
be legally applicable, either directly or as Incorporated by a federally
authorized state program, 1f response actions were not taken pursuant to Sections
104 or 106 of CERCLA.
Requirements that are applicable to and have jurisdiction over given
situations are considered "applicable requirements." An example of an applicable
requirement would be HCLs for a site that exhibits groundwater contamination
entering a public water supply.
C.3.1.2 Relevant and Appropriate Requirements
Relevant and appropriate requirements are those federal and state
requirements that, while not legally "applicable," can be applied 1f, 1n the
decision-maker's best professional judgment, site circumstances are sufficiently
similar to those situations that are jur1sd1ct1onally covered and use of the
requirement makes good sense. During the FS process, relevant and appropriate
requirements are Intended to have the same weight and consideration as applicable
requirements.
The term "relevant" was Included so that a requirement Initially screened
as nonappllcable because of jurisdictional restrictions would be reconsidered
and, 1f appropriate, Included as an ARAR for the site. For example, MCLs would
be nonappllcable, but relevant and appropriate for a site that exhibited
groundwater contamination 1n a potential (as opposed to an actual) drinking water
source.
C.3.1.3 Other Requirements To Be Considered
Other requirements to be considered are federal and state nonregulatory
requirements e.g., guidance documents or criteria.
C.3-2
-------
Nonpromulgated advisories or guidance documents do not have the status of
ARARs. However, where there are no specific ARARs for a chemical or situation
or where such ARARs are not sufficient to be protective, guidance or advisories
should be Identified and used to ensure that a remedy Is protective.
C.3.2 Development of ARARs
Under the description of ARARs set forth 1n the NCP and SARA, many federal
and state environmental requirements must be considered. These requirements
Include ARARs that are:
chemical-specific (I.e., govern the extent of site clean-up)
location-specific (i.e. pertain to existing site features)
action-specific (I.e., pertain to proposed site remedies and govern
Implementation of the selected site remedy)
C.3.2.1 Chemical-Specific ARARs
Chemical-specific ARARs govern the extent of site clean-up and provide
either actual clean-up levels or a basis for calculating such levels. For
example, surface water criteria and standards, as well as air standards, provide
necessary clean-up goals for the Hudson River PCB contaminant.
Chemical-specific ARARs are also used to Indicate acceptable levels of
discharge to determine treatment and disposal requirements and to assess the
effectiveness of remedial alternatives. Table C.3-1 lists and summarizes
potential chemical-specific ARARs. Chemical-specific ARARs will apply to every
alternative developed In later phases.
C.3.2.2 Location-Specific ARARs
Location-specific ARARs pertain to natural site features such as wetlands
and floodplalns, as well as manmade features, Including existing landfills,
disposal areas and historic buildings. Location-specific ARARs generally place
C.3-3
-------
restrictions on the concentration of hazardous substances or the conduct of
activities, because of the site's . particular characteristics or location.
Consideration of. such ARARs provides a basis for assessing existing site
conditions and subsequently aids in defining potential location-specific ARAR^
Potential location-specific ARARs are presented In Table C.3-2.
C.3.2.3 Act1on-Spedf1c ARARS
Action-specific ARARS are usually technology- or activlty-ba^fed 1 Imitations
that control actions at CERCLA sites. After remedial alternatives are developed,
action-specific ARARs pertaining to proposed site remedies provide a basis for
assessing the feasibility and effectiveness of the remedies. For example,
action-specific ARARs may include hazardous waste transportation and handling
requirements, air and water emissions standards and Iandf1l11ng and.treatment
requirements of TSCA and RCRA. Potential action-sped fie ARARs are presented in
Table C.3-3.
C.3.3 Statutes and Regulations
More detailed descriptions of some of the federal and state statutes
referenced in Tables C.3-1, C.3-2 and C.3-3 are presented below.
C.3.3.1 Federal Statutes and Regulations
Toxic Substances Control Act (TSCA) 15 USC 2601, 40 CFR 761
TSCA provides USEPA with authority to require testing of both new and
existing chemical substances entering the environment and to regulate them where
necessary. TSCA requirements do not apply to PCBs at concentrations less than
50 ppm; PCBs can not be diluted, however, to escape TSCA requirements.
TSCA establishes prohibitions and requirements for the manufacturing,
processing, distribution 1n commerce, use, disposal, storage and marking of PCBs.
Section 2605 includes provisions for incineration, disposal, storage for
C.3-4
-------
disposal, chemical waste landfills, decontamination, clean-up policy,
recordkeeping and reporting for PCBs.
Subpart D of 40 CFR 761 contains the following applicable provisions
regarding PCBs:
40 CFR 761.60(a)(5) states that "all dredged material and municipal
sewage treatment sludge that contain equal or greater than 50 ppm of
PCBs shall be disposed of:
(I) 1n an Incinerator pursuant to section 40 CFR 761.70,
(II) 1n a chemical waste landfill described 1n 40 CFR 761.75, or
(III) by a disposal method approved by the Agency's Regional
Administrator 1n the region 1n which the PCBs are located."
Applications for disposal by methods other than those specified In
subparagraphs (1) and (11) above may be made 1n writing and sent
directly to the Regional Administrator. Application procedures and
contents of the application are addressed 1n 40 CFR 761.60 (5)(111).
40 CFR 761.70 covers the Incineration of PCBs. Incinerators for the
burning of PCBs must be approved by the appropriate USEPA Regional
Administrator or the Director, Exposure Evaluation Division,
pursuant to section 40 CFR 761.70 (d), which lists application
requirements.
40 CFR 761.75 applies to facilities used to dispose of PCBs. In
general, a chemical waste landfill for PCBs must be approved by the
Agency Regional Administrator. The landfill must meet technical
requirements, which Include, but are not limited to, the following:
soil consistency surrounding the landfill, flood protection,
topography and appropriate record maintenance. 40 CFR 761.75 (b).
40 CFR 761.60(a)(6) provides that PCB articles with concentration
levels equal to or greater than 50 ppm must be stored prior to
disposal 1n compliance with section 40 CFR 761.65.
40 CFR 761.65 states that PCB articles must be removed from storage
within one year from the time they were placed 1n storage. In
addition, the regulation lists storage facility requirements and
container requirements. An exemption from this regulation exists
where PCBs are stored for 10 days or less at a transfer facility.
This section may be applicable should dredged materials be stored
before Incineration or landfill disposal.
USEPA has proposed amendments to the storage and disposal rules for PCBs
1n 40 CFR 761, subpart D. The tentative changes were published at 55 Fed. Reg.
C.3-5
-------
46,470 (1990). The purpose of this proposed rule 1s to set out criteria and
procedures for revoking and suspending PCB storage and disposal approvals.
Current (JSEPA rules require persons operating PCB disposal facilities to obtain
an approval for such activities. These rules do not, however, prescribe when or
how such approvals may be suspended or revoked.
Resource Conservation and Recovery Act (RCRA) - 42 USC 6901, 40 CFR 260
RCRA 1s an applicable statute because It establishes a cradle to grave
regulatory program for present hazardous waste activities. RCRA mandates the
national policy on how hazardous waste will be stored, treated and disposed of
to alleviate any potential threat to human health and the environment. PCBs
alone are not a RCRA hazardous waste; they are, however, subject to land disposal
restrictions, as discussed below, to the extent that the waste would otherwise
be considered hazardous under RCRA.
40 CFR 264 lists standards applicable for an owner or operator of hazardous
waste treatment, storage and disposal facilities. This part Includes, but 1s not
limited to, general facility standards, releases from the facility contingency
plan and emergency procedures to the generator of hazardous waste, landfills and
Incinerators.
40 CFR 268 Imposes land disposal restrictions. The purpose 1s to restrict
hazardous wastes from land disposal and to define limited circumstances 1n which
a hazardous waste may be disposed. Subparts Include a schedule for land disposal
prohibitions and establishment of treatment standards, prohibitions on land
disposal, treatment standards and prohibitions on storage.
40 CFR 270 establishes the USEPA-adm1n1stered permit programs, e.g.,
hazardous waste permit programs. This. regulation subparts set forth permit
application conditions and procedures. Specifically, Subpart F part 270.62
addresses hazardous waste Incinerator permits.
C.3-6
-------
Comprehensive Environmental Response, Compensation and Liability Act (CERCLA) 42
USC 9601, 40 CFR 372
Under CERCLA 101 (14) a "hazardous substance" 1s any substance USEPA has
designated under specified sections of the Clean Water Act, Clean Air Act or
TSCA, and any "hazardous waste" under RCRA. A list of such hazardous substances,
which Includes PCBs, may be found at 40 CFR part 302.
At 40 CFR 372 are set forth requirements for the submission of Information
relating to the release of toxic chemicals under section 313 of Title III of
SARA. The Information collected under this part 1s intended to Inform the
general public and communities surrounding regulated facilities about releases
of toxic chemicals. "Release" 1s defined as any spilling, leaking, pumping,
pouring, emitting, emptying, discharging, Injecting, escaping, leaching, dumping
or disposing into the environment (Including the abandonment or discarding of
barrels, containers and other closed receptacles) of any toxic chemical.
Notification requirements are set forth within this section, as well a toxic
chemical listing, which Includes PCBs. The section 1s applicable, since there
1s the possibility that PCBs could be released Into the surrounding area 1n which
the remedial action 1s occurring.
Clean A1r Act - 42 USC 7401
This Act promotes the protection and enhancement of the quality of the
nation's air resources through research and development and air pollution control
and prevention. For example,, if the remedial action Includes Incineration of
sediments, then the Act may be relevant and appropriate.
42 USC 7412 establishes national emission standards for hazardous air
pollutants. The term "hazardous air pollutant" means an air pollutant that may
reasonably be anticipated to result In an Increase 1n mortality or an Increase
1n serious Irreversible, or Incapacitating reversible Illness. In addition, the
regulation provides operational standards, designs, equipment and work standards.
PCBs have not previously been listed as a hazardous air pollutant. However,
C.3-7
-------
under Title III of the Clean Air Act Amendments of 1990, PCBs (Aroclors) are
Included on the list of toxic air pollutants for which USEPA will have to
establish standards.
C.3.3.2 New York State Statutes and Regulations
New York Environmental Conservation Law Article 11, Title 5
This statute 1s applicable, because 1t regulates the disposal of pollutants,
into the Hudson River, but 1t does not specifically address PCB levels.
The statute prohibits the disposal of "deleterious or poisonous substances"
Into any public or private waters which affect the welfare of wildlife, waterfowl
and fish Inhabiting those waters. The statute does not define the term
"deleterious or poisonous substance" nor does It define, "disposal." This
statute also provides that "oil, acid, sludge, cinders or ashes from a vessel of
any type shall not be thrown, dumped or allowed by any person to run Into the
waters of the Hudson or Hohawk rivers." Finally, the statute prohibits the
disposal of earth, soil, refuse or other solid substances into streams or
tributaries Inhabited by trout. The statute does not establish the maximum
acceptable levels for PCBs.
New York Water Classification and Quality Standards
This statute is applicable as 1t specifically lists the permissible ambient
water quality standard for PCBs and provides water classif1 cations for the Upper
and Lower Hudson.
The statute provides water classifications for the protection and propaga-
tion of fish, shellfish and wildlife, and for recreation 1n and on water, and
takes into account the use and value of water supplies for: public, agricultur-
al, industrial and other purposes, Including navigation. The classification of
water depends on permissible usage. Differences in water classifications are
determined according to what discharge is made to the water and where the water
discharge flows.
C.3-8
-------
For each classification, a best usage standard of water 1s established.
Within each classification, the standards for determining water quality, which
Include toxic waste and deleterious substances are established. The standards
for determining water quality must be consistent with the best usage of the
water. In addition to these general standards, parts 702.3-702.4 create special
classifications and standards for the Upper and Lower Hudson.
Ambient water quality standards are listed 1n Appendix 31; the established
standards for PCBs applicable to the classification of the Upper and Lower Hudson
River are as follows:
PCBs - Human Life and Other Usage 0.01 ppb
PCBs - Aquatic Life and Other Usage 0.001 ppb
Groundwater Classifications
This statute 1s applicable as 1t 1s specific to PCB concentrations In
groundwater. It provides groundwater quality standards and water classes. The
maximum level for PCBs 1s 0.1 ppb.
New York State Constitution Article XV, Section 1-4
Article XV provides that the state has sole ownership and management of the
canal system. The state has the power of granting revocable permits for the
occupancy or use of such barge canal lands or structures. The state nay lease
or may transfer the barge canal and terminals and facilities to the federal
government. The lease or transfer to the federal government may be for the
purposes of operation, Improvement and Inclusion 1n the national system of Inland
waterways, Including flood control, conservation and utilization of water
resources. This arguably would Include dredging for remediation purposes.
C.3-9
-------
The above article from the New York Constitution refers to the Canal Law
Article XV (1938), which provides that the acquisition of land and waters
necessary for the improvement, maintenance or repair of the canal system shall
be through appropriations by the New York Superintendent of Public Works. ^
New York Environmental Conservation Law Article 21, Title 5
The Tr1-State compact attempts to control pollution around the Lower Hudson
through the Interstate Sanitation Commission. The compact relevant and
appropriate.
This statute was created to control future pollution and to abate existing
pollution in the waters of the densely populated Tri-state district bordering New
York, Connecticut and New Jersey. In general, the states agree that no sewage
or other polluting matters will be discharged or will be permitted to flow into
the waters of the district. Each state signatory pledges to act in full
cooperation through the Interstate Sanitation Commission, which has jurisdiction
to enforce the provisions of the compact. The compact specifically Includes the
Lower Hudson River in the district. The compact does not further define "sewage
or other polluting matters," nor does 1t establish maximum concentration levels
for PCBs.
New York Environmental Conservation Law Article 24, Freshwater Wetlands
The purpose of this statute 1s to declare the public policy of New York
State regarding the preservation, protection and conservation of freshwater
wetlands.
Regulated activities, among others, are: dredging, draining excavation and
removal of sand, soil, mud, shells, gravel and other aggregate from any
freshwater wetland. Dredging or filling of navigable waters of the state must
be done pursuant to this article and any other applicable law. The statute does
not mb.re explicitly define navigable waters.
C.3-10
-------
New York Environmental Conservation Law Article 25, Tidal Wetlands
The regulation protects the state's Interest 1n tidal wetlands,
regulations cover activities such as dredging, draining and excavation of
wetlands.
C .3-11
-------
PAGE INTENTIONALLY LEFT BLANK
C.3-12
-------
C.4 Technology and Process Identification
Technologies are presented here 1n the following order: 1n situ
technologies Including containment and treatment; dredging and excavation
methods; sediment and treatment techniques, Including physical, chemical, thermal
and biological; and disposal alternatives. Several emerging sediment treatment
systems are discussed at C.5.
C.4.1 Containment
In situ control and containment measures are Intended to reduce dispersion
and leaching of contaminated sediments to other areas of a water body. Methods
Include retaining dikes and berms and capping. Retaining dikes and berms Include
earthen embankments, bulkheads and sheet pile walls. The structures can be
constructed perpendicular to the direction of stream flow to prevent suspended
particulate matter from flowing downstream or parallel to a river bank.
A wide variety of materials can be used to cap contaminated sediments In
order to minimize leaching and prevent their erosive transport. Cover materials
Include Inert materials such as silt, clay, sand, cement, or a geotextlle;
alternatively, active materials can be applied to the surface or mixed with the
sediment 1n an attempt to limit mobility. Issues such as the capping material's
susceptibility to scour and resuspenslon, ability to withstand leaching, and
effect on the ecosystem must be studied 1n order to determine suitability for a
given site. Capping may Increase anaerobic activity and has been used as a
component of bloremedlatIon ^programs.
Containment using capping was the remedial action chosen for the Hudson
River remnant deposit sites. A manufactured soil cover material (claymax) Is
being used along with two feet of soil cover at these deposit sites. Work has
now largely been completed on that project and post construction monitoring will
begin during this year (1991).
C.4-1
-------
C.4.2 Natural PCB Biodegradation In Sediments
C.4.2.1 Aroclor Patterns
4*
Much study by GE and other researchers has focused on the Issue of
naturally occurring, biologically mediated PCB transformations In Hudson River
sediments. Laboratory controlled experiments have been conducted In an attempt
to confirm dechlorination or biodegradation processes. Monitoring data have also
been evaluated specifically with respect to addressing the question of whether
biological transformations are responsible for the altered appearance of the PCBs
found 1n the river versus the original PCBs (Aroclor mixtures) discharged to the
river.
In reviewing the data from Hudson River sediment sample analyses, Brown,
Jr et a/. (1984) have observed that the concentration distribution of PCB
congeners In sediment varies from that 1n the Aroclor mixtures (predominantly
1242) believed to have been originally deposited there. Uncertainty 1n the
amount of different Aroclors that actually entered the river, differential
partitioning and transport of different PCB congeners and biodegradation are all
contributing factors to this observation.
In 1984, Brown, Jr. et al. analyzed sediment samples taken from eight
surface locations and two sediment cores in River Reach 8; an additional surface
sample was collected from Reach 6. They reported that a number of different PCB
congener distributions could be distinguished In the gas chromatograms generated
from the analysis of these samples. The three primary patterns observed were:
Pattern A, which looked very much like Aroclor 1242 but showed
somewhat Increased portions of penta- and hexachlorinated PCBs,
somewhat diminished portions of tetra- and trlcblorobiphenyls and
markedly diminished proportions of mono- and dlchlorobiphenyls;
Pattern B, which showed a reduction 1n levels of penta-, hexa- and
some tetrachlorobiphenyls and more mono- and dechlorinated congeners
as compared to Pattern A; and
C.4-2
-------
Pattern C, which was distinguished from pattern B by less reduction
1n the levels of penta-, hexa- and tetrachlorobiphenyls, but greater
reduction 1n the levels of tri- and some dlchloroblphenyls, and mich
greater concentrations of mono- and some dechlorinated congeners.
C.4.2.2 Aerobic Blodegradatlon of PCBs
Brown, Jr. et al. (1984) attributed Pattern A to aerobic blodegradatlon of
Aroclor 1242 for several reasons. Pattern A was typically associated with
surface deposit. Aerobic microbial degradation of mono- and dechlorinated PCB
congeners had been reported 1n laboratory experiments (Furukawa, 1982; Shlarls
and Sayler, 1982; Kong and Sayler, 1983; Safe, 1984). Bacteria capable of
degrading some tr1- and tetra-chlorinated PCB congeners were detected In every
Reach 8 sediment sample collected. The PCB congener distributions In Hudson
River surface sediment samples looked more similar to that produced from Aroclor
blodegradatlon by the most commonly encountered microbial populations than by
mammalian mixed function oxidases (MFOs), by anaerobes or by water extraction.
The microbial activity observed by Brown, Jr. et *1. (1984) and others (Furukawa,
1982; Safe, 1984; and Bedard, 1990) generally decreased with the number of
chlorine atoms on the molecule. Thus, an environmental sample containing a
biodegraded Aroclor mixture would be expected to show a pattern of PCBs like that
of Pattern A, In which the lesser chlorinated congeners are reduced 1n
concentration relative to the more chlorinated congeners. Non-b1o1og1cal1y
mediated process, such as selective dissolution or partitioning of congeners, are
other mechanisms that might also explain the Pattern A congener distribution.
More recent work supports the conclusion that blodegradatlon may account
for at least some of the observed patterns of PCB concentration distributions In
surface sediments. Bedard et a7. (1986, 1987a and 1987b), Sing et el. (1988),
and Unterman et ซ7. (1988) have shown 1n laboratory studies that there are at
least 25 different naturally occurring microbial strains capable of degrading one
or more PCB congeners. The congener specificities exhibited by these strains
suggests that the Initial aerobic microbial attack on PCBs is mediated by one of
C.4-3
-------
two types of enzymes, a 2,3- or a 3,4-d1oxygenase (Abramowicz, 1990; Bedard,
1990).
While Pattern A may be the result of aerobic PCB blodegradatlon 1n surf^e
sediments, the rate at which blodegradatlon 1s occurring 1n Hudson River
sediments Is extremely difficult to determine. Attempting to extrapolate a rate
from existing concentrations 1s not possible, since the Initial concentrations
of PCB congeners in surface sediments are unknown and the present concentrations
have undoubtedly been affected by sediment scouring, sediment deposition and
sediment-water partitioning, as well as blodegradatlon. Laboratory measured
blodegradatlon rates can not be extrapolated to the river, because of differences
1n temperature and nutrient concentrations. No time series data, which might
allow estimation of rates In situ, have been systematically collected from the
river. Furthermore, 1t Is likely that different blodegradatlon >ates are
occurring 1n different areas of the river, as a consequence of the presence of
different microbial populations and nutrient concentrations at different
locations.
Brown, Or. et aJ. (1984) have suggested, based on the levels of Aroclor
1221 and 1016 reported 1n fish between 1977 and 1981 (Sloan et al. 1983), that
the mono- and dlchlorlnated congeners are being biologically degraded from
surface sediments with a half-life of one to two years. They note, however, that
the rate of Aroclor 1221 and 1016 disappearance reflects the disappearance of
congeners that can be degraded by a 2,3-d1oxygenase; the degradation of other
mono- and dlchloro congeners 1s likely slower. This estimate 1s subject to
several sources of error: the small number of time points (five) from which 1t
was made; the variability Inherent 1n biological data; and Insufficiently
detailed data to rule out factors other than blodegradatlon, e.g. evaporative
loss of lesser chlorinated congeners from the water column.
Even 1f an aerobic blodegradatlon rate could be determined with some
accuracy, such a rate, by definition, would be characteristic of only the upper
most oxygenated layers of sediment. The rate 1n the remainder of the sediment
C.4-4
-------
column would be essentially zero. Such depth limitations to aerobic b1ฉdegrada-
tion may partially explain the fact that most of the surface sediment samples
discussed by Brown, Jr. et al. (1984, 1987b) actually showed patterns that
resembled a mix of surface and subsurface sediment rather than Pattern A
exclusively; these samples typically had higher levels of mono- and dl-
chlorinated congeners than found in Aroclor 1242.
C.4.2.3 Anaerobic Dechlorination
Brown, Jr. et a7. (1984, 1987a, 1987b, 1988) have suggested that anaerobic
dechlorination produces Patterns B and C in the subsurface. The evidence they
present to support their hypothesis is four-fold. First, Patterns B and C were
reportedly associated with subsurface hot spots. Second, sediments exhibiting
these patterns showed changes in homologue distribution from that of Aroclor 1242
corresponding to one-third of the chlorine lost. Third, subsurface sediments
tended to have PCB congener distributions, I.e. a greater percentages of mono-
and dlchloroblphenyls, which were difficult to explain on the basis of selective
extraction of the lesser chlorinated congeners upstream with redepositlon
downstream. Fourth, there was a selective loss of aeta and para, as opposed to
ortho chlorines 1n Patterns B and C, suggestive of a biological specificity. The
hypothesis of Brown, Jr. et a7. 1s supported by the work of Quensen et a7. (1988,
1990), who have shown that a bacterial Inoculum cultured from Hudson River
sediments can dechlorinate PCBs when incubated under an Nj-CO, atmosphere, at
25*C.
The ability of microorganisms from Hudson River sediments to dechlorinate
PCB mixtures reductlvely has been confirmed by researchers at GE's Biological
Sciences Laboratory, the University of Michigan, New York University Medical
Center and Wadsworth Center for Laboratories and Research (Abramowicz et a7.,
1989; Vogel et a7., 1989; Alder et a7., 1990; Rhee and Bush, 1990). Abramowicz
et a7. (1989) have shown the sequential, mlcroblally mediated removal of
chlorines from 2,3,4,3'4,-pentachloroblphenyl to form tetra-, tr1-, d1-, and
C.4-5
-------
nonochlorob1phenyls. To date, however, no anaerobic species demonstrating PCB
dechlorination have been Isolated (see review by Abramowlcz, 1990).
While It seems plausible that reductive dechlorination Is occurring In the
subsurface sediments of the Upper Hudson, the average rate at which this
degradation takes place Is difficult to determine and may be fairly slow. The
rate may be expected to slow further as the more reactive congeners are
dechlorlnated and the more recalcitrant congeners accumulate (Brown, Jr. et a?.,
1984). Chen et al. (1988) found no evidence of anaerobic PCB congener
biodegradatlon In sediments, unless they were Incubated with the addition of
cultured bacteria. Rhee et al. (1990) detected no decrease 1n Individual PCB
congener concentrations after anaerobic Incubation of Moreau sediments In situ
for seven months. Brown, Jr. et al. (1984) have calculated very rough half-lives
of ten years for the penta- and tetra- chlorinated congeners and two to three
years for the reducible tetra-, tr1- and dechlorlnated congeners, based on a
single core available (core 18-6) for analysis.
Figure C.4-1 shows the weight percentages of congeners with one, two,
three, four or five chlorines reported for core 18-6 as a function of depth.
Since the core was collected 1n 1977 and sediments at 19 Inches were tentatively
dated at 1952, this core represents sediments deposited over the course of
approximately 25 years, with progressively older sediments at greater depths.
Between the surface and eight Inches, Brown, Jr. et al. characterize the PCB
pattern in the core as a mixture of Patterns A and B. Thus, the changes in
congener percentages over this section of the core may reflect a mixture of
aerobic and anaerobic processes. Between a depth of 8 and 19 Inches, however,
the core 1s characterized by Pattern B alone. If the ratio of various Aroclor
mixtures Input to the river and the sedimentation rate were relatively constant
over this time period and 1f reductive dechlorination 1s producing Pattern B,
this section of the core should show progressively smaller percentages of the
more chlorinated congeners with contaminant Increases in the percentages of
terminal dechlorination products. The weight percentages of all congeners,
however, actually remain relatively constant through this section. This finding
C.4-6
-------
suggests either that the input of Aroclor mixtures was not constant, that the
source of Pattern B is not reductive dechlorination or that limited dechlorina-
tion occurred over the time period represented by this section of the core.
C.4.3 Removal Technologies
Removal technologies Include dredging and excavation. Excavation
techniques are not discussed further In this report.
Dredging system alternatives have been evaluated extensively by the US Army
Corps of Engineers, in general, and by USEPA and numerous consultants for
specific Superfund Projects. Dredging systems identified 1n the literature fall
Into the hydraulic, mechanical and specialty-type categories, with each category
serving particular applications. Fugitive sediment releases from dredging
equipment will be evaluated in subsequent phases, utilizing published reports.
The cutterhead hydraulic pipeline dredge 1s among the most commonly used
dredging systems. By combining mechanical cutting action with hydraulic suction,
this dredge has the capability of efficient excavation and removal of material
to disposal sites without rehandllng. Material would be drawn-up by the dredge
and pumped at an appropriate solids concentration through a floating or submerged
pipeline to shore and to ultimate disposal. The system would Include booster
pumps and tugs and would necessitate a water treatment pU spable of handling
large quantities of water.
The clamshell mechanical dredging system consists of barge-mounted cranes
outfitted with suitable clamshell buckets. Excavated material is placed In scows
or hopper barges for transport to the disposal site. In case of the Hudson
River project, the barge contents would probably be slurried for removal by a
hydraulic pump-out system located on shore. Where circumstances permit, bottom
dump scows can be used In concert with a mechanical system and would discharge
dredged material at sub-aqueous disposal sites.
C.4-7
-------
Results of recent field studies conducted by the US Army Corps of Engineers
Waterways Experiment Station Indicated that the cutterhead dredge was the most
successful 1n limiting sediment resuspenslon Into the water column, followed by
the hopper and clamshell dredges. Modifications such as overflow prevention or
use of an enclosed bucket may Improve resuspenslon characteristics of the hopper
and clamshell dredges. Specialty dredges were also tested (the modified dustpan
and matchbox dredges) and compared with the cutterhead. No reduction 1n sediment
resuspenslon was found with use of the specialty dredges.
Historically, contaminated sediments were removed from the river during
NYSDOT's routine channel maintenance dredging. As the river's PCB problem became
better understood, remedial alternatives, Including bank-to-bank dredging of the
river, full-scale dredging of the 40 PCB hot spots 1n the river and reduced-scale
dredging of the most contaminated hot spots, were considered. Due to limited
funding under the Clean Water Act, a reduced-scale dredging program had been
recommended by the USEPA and the NYSOEC in earlier studies. The NYSDEC currently
has an Action Plan for site remediation that Incudes dredging and encapsulation
of river sediments at an upland site 1n proximity to the river (Site 10).
C.4.4 Treatment Technologies
C.4.4.1 Physical and Chemical Treatment Technologies
Evaluation
To date, Incineration and disposal In landfills are the most widely
practiced and permitted methods for management of PCB-contam1nated soils and
sediment. However, other technologies have now emerged and are considered
technically and economically feasible alternatives to Incineration and
landfllUng In certain circumstances. In this section, a range of physical and
chemical treatment technologies and their general applicability to the Hudson
River site and level of development are presented.
C.4-8
-------
Technologies screened In previous studies of the Hudson River, Including
the NUS Feasibility Study Report (1984), the Engineering Report by Malcolm
P1ra1e, Inc. (MPI, 1985) as part of the Hudson River PCB Reclamation/Demonstra-
tion Project and a Research Triangle Institute Report (1987) are reexamined
herein. The Ebasco Feasibility Study for the New Bedford Harbor site (1990) has
also been reviewed, as have emerging technologies demonstrated through the USEPA
Superfund Innovative Technology Evaluation (SITE) Program.
The technologies discussed Include: 1) those that were retained in the
previous studies for further test and evaluation and have been developed to at
least a pilot-scale operation; 2) those not retained 1n previous studies, but
which have undergone significant development since the previous studies; and 3)
those new technologies that have recently emerged and which have been demonstrat-
ed on a pilot-scale 1n the USEPA SITE program.
Table C.4-1 lists the treatment technologies reviewed and screened In the
NUS and MPI studies. The majority of technologies reviewed at the time of
publication of the two reports were In the early stage of development; little
Information was known about their environmental effects and costs. The preferred
physical/chemical technologies 1n the NUS study were KOHPEG and wet air
oxidation. Results of the MPI study Indicated that none of the technologies
evaluated had been demonstrated to treat PCB-contamlnated waste effectively and
that further research was required before a preferred method could be selected
for a project of the magnitude of the Hudson River project.
In 1987, the Research Triangle Institute (RTI), under USEPA sponsorship,
conducted a study to Identify emerging technologies and their level of
development. From the Initial screening of 64 treatment technologies, eleven
processes, of which ten were physical/chemical processes and one was biological,
were selected for further technological assessment. Table C.4-2 lists the ten
physical/chemical processes.
C.4-9
-------
According to RTI, a significant number of solvent washing/extraction
processes had emerged since the NUS Feasibility Study was published 1n 1984. The
solvent washing/extraction processes are physical treatment technologies which
do not destroy PCBs, but transform them from a solid matrix to a solvent matri^
PCBs extracted to the solvent matrix can be further separated from the solvent,
forming a concentrated PCB stream, which then requires final disposal. In
comparison to thermal systems, solvent extraction may be easier to permit, since
it is not subject to the same regulatory restraints as thermal treatment. It may
also be economically feasible for treating a wider range of PCE Pastes and can
be less costly with respect to energy usage. RTI recommended that three
technologies be retained for thorough testing and evaluation, of which two were
physical/chemical. These were the Basic Extraction Sludge Treatment (B.E.S.T.)
process and the UV/llltrason1cs Technology process. None of these technologies
has been applied on a large scale.
In 1990, Ebasco conducted bench/p1lot-scale treatment tests on seven
selected PCB treatment technologies (Table C.4-3). Five were physical/chemical
processes, one was a dewaterlng process and the other was bloremedlatlon. The
two physical/chemical technologies retained for the development of remedial
alternatives were the B.E.S.T. process and solidification/stabilization with the
use of Portland cement. According to Ebasco, solvent extraction technology
(B.E.S.T. process) was not yet fully developed for PCB waste treatment
applications.. Incineration was still considered the only Best Demonstrated
Available Technology (BOAT) for PCB wastes.
Physical and chemical technologies chosen in previous studies for further
analysis are summarized below.
PROCESSES NUS H9841 RTI (19871 EBASCO (1990)
KOHPEG X
Wet A1r Oxidation X
B.E.S.T. Process X X
UV/lfttrason1cs Technology X
Sol1avf1cat1on/Stab1l1zat1on X
C.4-10
-------
To be retained here, a technology must have been preliminarily tested for
Its environmental compatibility, Its cost effectiveness and technical capability.
In some cases, however, where technologies have undergone preliminary testing,
they have not been developed further. For example, the wet air oxidation
process, a preferred technology 1n the NUS Feasibility Study Report, ha^been
found to be a non-viable application for effective destruction of PCBs 1n a
slurry matrix (J. R. Nicholson, personal communication, March 1991). The
UV/Ultrasonlcs technology, rated as a preferred process in the RTI Report, has
not been commercially developed (E. Pedzy, personal communication, February
1991). Many process developers Identified In previous reports have left the
market.
The physical and chemical treatment technologies that were retained for
further analysis In previous studies and that have undergone further development
since those studies Include:
*
KOHPEG Process (Galson Remediation Corporation)
8.E.S.T. Process (Resources Conservation Company)
Technologies that were not retained for analysis in previous studies, but
which have gained significant advancement in technical development Include:
Low Energy Extraction Process (ART International, Inc.)
Propane Extraction Process (CF Systems, Inc.)
These four technologies are discussed below.
KOHPEG
The KOHPEG process, developed by Galson Remediation Corporation (GRC) of
Syracuse, NY, Is a chemical treatment technology using potassium hydroxide (KOH)
In a solution of mixed polyethyleneglycol (PEG) and a phase-transfer catalyst,
C.4-11
-------
dimethyl sulfoxide (DMSO), to dechlorlnate PCBs. The dechlorination reaction
takes place 1n the liquid phase. Therefore, 1t 1s necessary to extract the
contaminants from the soil surface Into the reagent phase, where they can react.
The end products are glycol ether and potassium chloride, which are water
soluble, low toxicity materials. GRC commercialized the process and has marketed
1t under the name of APEG-PLUS since 1986; Its Initial unit Is a mobile
decontamination facility designed for treating contaminated soil and sludges at
a capacity of 20-200 tons per day. The material flow sequences of the system are
shown In Figure C.4-2.
GRC conducted a bench-scale test on New Bedford Harbor sediments. The
results Indicate that PCB removal efficiencies of above 99 percent were achieved
for samples containing PCB concentrations of 440 to 7,300 ppm. However, the
recovery of reagent and sediment-solids was low, 75 percent for DMSO and 43
percent for sol Ids. This ultimately proved to be a materials handling problem.
Later, a separate bench-scale simulation test conducted by GRC on PCB-contaminat-
ed soil reported that the recovery of both reagent and solids had Improved to 81
percent for DMSO and 102 percent for solids (Galson, 1991). Costs for treating
New Bedford Harbor sediments using the APEG-PLUS system were estimated to be $98
and $120 per ton based on 500,000 and 50,000 cubic yards of sediment treated,
respectively.
The APEG-PLUS system 1s primarily used for treating soils. Modifications
In the material handling equipment would be required for treating the Hudson
River sediments. The moisture content of the sediment may affect the heating
requirements and reactor cycle time of the treatment process. The affected
operation parameters should be determined through pilot tests using actual site
samples. Toxicity of the technique's reaction products and their long-term
effect on environmental conditions remain to be confirmed.
C.4-12
-------
B.E.S.T. Solvent Extraction Process
The B.E.S.T. process utilizes tr1ethyl amine (TEA) as a solvent. TEA Is
completely soluble with water at temperatures below 65ฐF. Above this temperature,
TEA and water are only partially mlsclble. This inverse m1sc1b111ty property of
the solvent 1s employed to separate PCB-contamlnated sediment Into PCB/oll,
water, and solids fractions. A block diagram for the B.E.S.T. process Is
presented 1n Figure C.4-3. The extraction of the contaminated material Is
conducted at a temperature of approximately 40ฐF. At this temperature, the TEA
freely mixes with the water and the PCB/oll fraction of the sediment matrix.
Therefore, the extract solution contains most of the water in the feed and
PCB/oll fractions. The extract solution Is then heated to temperatures above
130ฐF. At this elevated temperature, the water separates from the TEA/PCB/ofl
fraction. The TEA solvent 1s recovered for reuse from the separated phases via
steam stripping. The PCB/o1l fraction 1s disposed of by Incineration or chemical
dechlorination processes at a permitted facility.
The Resources Conservation Company (RCC) conducted a bench-scale study of
Its B.E.S.T. process on New Bedford Harbor sediment (Ebasco, 1990). PCB removal
efficiencies of above 99 percent were achieved with three extraction stages froa
Initial PCB concentrations of 5,800 and 420 ppm. The PCB concentrations In
treated sediment residues were 130 and 11 ppm, respectively, for the high and low
Initial concentrations. A test with an Initial PCB concentration of 11,000 ppa
resulted 1n a residue containing 16 ppm PCB after six extraction stages.
Recently, RCC has completed bench-scale treatability testing with PCB-
contamlnated soil samples from natural gas pipeline compressor stations. The PCBs
1n the feed were 1n the range of 500-2,000 ppm. Removal efficiencies of above
99.8 percent were obtained, leaving residual PCB In the treated soil at less than
2 ppm (Welmer, 1991). Similar PCB extraction efficiencies using the B.E.S.T.
process were obtained In other tests. Various PCB levels 1n the treated residue
were reported. The factors that affect the extraction efficiencies and the
residual PCB 1n treated matrices are believed to be the Initial concentration of
C.4-13
-------
PCB, stages of extraction and matrix characteristics. Those factors are site-
specific determinants, which should be verified by bench or pilot-scale
treatability tests.
RCC has recently completed a B.E.S.T. pilot plant configured for processing
liquid sludge (sediment) or contaminated soils. The pilot plant 1s configured
with a washer/dryer vessel for extracting and drying soils. The washer/dryer Is
a horizontal, cylindrical vessel that has a rotating shaft with mixing paddles
attached. Performance of the washer/dryer 1s key to successfully treating
sediments with the B.E.S.T. process.
Low Energy Extraction Process (LEEP)
The Low Energy Extraction Process (LEEP) was developed by a New York
University research team. Study of Its application 1n removing PCBs from soil,
sediment and sludge was funded by USEPA. LEEP technology was being developed and
commercialized by Remediation Technology, Inc., which 1s now Applied Remediation
Technology (ART) International, Inc.
The process 1s a solvent extraction technology based on the combined use
of hydrophlllc (water mlsdble) and hydrophobic (water Immiscible) solvents.
Contaminants (PCBs) are leached from solid material with acetone (a hydrophlllc
leaching solvent) and then concentrated 1n kerosene (a hydrophobic extractant)
by 11qu1d-l1qu1d extraction. While the acetone solvent 1s recycled Internally,
the kerosene containing PCBs 1s removed from the process for final destruction.
Decontaminated solids and water are returned to the environment. A schematic
diagram of the primary steps of LEEP 1s shown 1n Figure C.4-4.
LEEP was accepted Into the SITE Emerging Technologies Program 1n 1989. ART
International, Inc. has conducted bench-scale process simulation (a treatability
study) on Waukegan, Illinois harbor sediment. The experiment resulted 1n
selection of acetone as a leaching solvent by virtue of Its high leaching effi-
ciency and good settling characteristics for solid fines. Kerosene was chosen as
C.4-14
-------
the 11qu1d-11qu1d extractant for Its good affinity to PCBs, availability, cost,
relatively low toxicity and 1mm1sc1b1l1ty in the mixture of water and acetone.
The treated sediments contained 42 percent by weight of water and initial PCB
concentration 1n the sample was 33,600 ppm on a dry weight basis. The study
showed 99.9 percent PCB removal. A pilot plant with a nominal throughput of 200
Ib/hr of dry solids is under construction. Design and engineering of a trailer-
mounted commercial unit Is scheduled to be completed In 1992.
The LEEP technology uses basic unit operations and Is constructed with
commercially available equipment. Selection of equipment for application to the
Hudson River sediment requires specific engineering analyses. Concurrent Hudson
River studies may lead to a better definition of equipment 1n use. A major
disadvantage of LEEP Is that the system 1s not yet available for full-scale
applications. Also, feed stream particle size limitation may limit process
applicability. Further development and demonstration at a large scale are
needed.
Propane Extraction Process
C.F. Systems, a Morrison Knudsen Company In Woburn, Massachusetts, 1s the
developer of the liquefied propane extraction technology. The process* uses
propane at ambient temperatures and at pressures over 200 ps1 to extract PCBs
along with other oily organics from a sediment-water slurry. C.F. Systems has
been operating a commercial-scale unit to treat petroleum refinery sludge at a
capacity of 100 barrels per day. This unit was selected to demonstrate their
pilot-scale system using New Bedford Harbor sediments. The unit 1s trailer
mounted and Is designed to handle pumpable soils, sludge or sediments.
The basic operating steps, shown in Figure C.4-5, Include extraction, phase
separation and solvent recovery. A mixture of liquefied propane and butane Is
used as the extracting medium. Pumpable (slurried) solid waste is fed Into the
top of an extractor. Then the solvent, a propane/butane mix, Is condensed by
compression and allowed to flow upward through the same extractor. In the
C.4-15
-------
extractor the solvent makes non-reactive contact with waste materials, dissolving
the waste's organlcs. Following this extraction procedure, the residual mixture
of clean water or water/sol Ids can be removed from the base of the extractor.
The mixture of solvent and organlcs leaves the top of the extractor and passes
to a separator through a valve, which partially reduces pressure. The reduction
of pressure causes the solvent to vaporize out of the top of the separator. It
is then collected and recycled through the compressor as fresh solvent. The
organlcs left behind are drawn from the separator.
PCB removal efficiencies of 90 percent were achieved for New Bedford Harbor
sediments containing PCBs ranging from 350 to 2,500 ppm with up to ten passes or
recycles through the treatment unit. The propane extraction process was not
retained as a viable technology for that particular site because of problems with
materials handling, system operating parameters, extraction efficiencies and low
throughput rates observed during the pilot demonstration. Material treated with
the propane extraction process must be pumpable. The Hudson River sediments
would have to be prepared 1n a slurry form that can move through the system
without clogging. Formation of settled particles or the presence of abrasive
particles In the sediment would require special handling equipment 1n order to
keep the material flowing through the system.
C.4.4.2 Thermal Treatment Technologies
Introduction
Thermal treatment 1s the application of heat to a substance to reduce or
eliminate Its toxicity. The thermal level can be extreme, e.g., using a plasma-
arc for destruction, which generates temperatures In the thousands of degrees,
or 1t can be relatively low, e.g., passing the contaminated substance through a
dryer operating at a temperature of 400ฐF.
i
C.4-16
-------
The regulatory framework for the thermal destruction of PCBs 1s the Toxic
Substances Control Act (TSCA). Incineration under TSCA requires that PCBs be
maintained at a temperature of 2,200ฐF for at least two seconds. Additional
Incineration requirements Include a minimum combustion efficiency of the
burner(s), PCB destruction efficiency, acid gas scrubbing, and .control and
recording requirements. Residual materials from thermal treatment processes must
contain no more than 2 ppm PCBs.
Remedial thermal systems can be permanent, mobile or portable. Mobile
systems are brought to a site and then removed at the conclusion of the clean-up.
They normally Include all of the equipment and subsystems necessary for operation
of the facility, such as electric power generation equipment, a fuel supply and
equipment to collect and dispose of wastewater. Transportable equipment differs
from mobile equipment 1n that 1t requires a significant Installation effort.
This equipment Is provided 1n modular components and must be assembled before
use. A process water supply 1s needed on-site. Wastewater discharge would als,o
be disposed of on-site, although water or wastewater discharge treatment
facilities may be required. Transportable systems are designed so that they can
be dismantled, removed and re-Installed at another site.
The anticipated load for this project, as well as Its geographical
distribution, would probably necessitate Installation of one or more on-site
thermal systems.
Low Temperature Desorptlon
Low temperature desorptlon Is the application of low level heat (400ฐF to
800ฐF) to a material 1n a primary chamber. Organic materials are released In the
primary chamber and are directed to a secondary chamber. In the secondary
chamber, the organlcs are heated to 2,200ฐF for destruction and are then passed
through an air emissions control system to remove acid gases and particulate
matter.
C.4-17
-------
The low temperatures at which the system operates allow use of conventional
materials 1n the primary chamber and reduce the amount of supplemental fuel
required. By varying the amount of air admitted to the primary chamber, the
amount of volatlles released can be controlled. With less air (approaching a
pyrolysis condition), the release of organics from soils will increase.
PCBs bind tightly to soils. A temperature of 400ฐF to 800ฐF is not
sufficient to release the contaminant from soil, unless air flow 1s severely
restricted. In a restricted oxygen atmosphere (pyrolysis), however, the PCBs
will be encouraged to form other compounds, such as dioxins, which are considered
more toxic than PCBs. Consequently, thermal desorption 1s not a recommended
thermal treatment technology for PCB-contaminated sediments.
Rotary Kiln Incineration
The rotary kiln Incinerator is a horizontal cylinder, lined with refractory
material, which turns about Its horizontal axis. Waste, no greater than two
Inches mean particle size, is deposited 1n the kiln at one end and 1s reduced to
an ash by the time It reaches the opposite end of the kiln. Kiln rotational
speed 1s 1n the range of three-quarters to two revolutions per minute.
A source of heat is required to bring the kiln to operating temperature and
to maintain its temperature during incineration of the waste feed. Supplemental
fuel is normally Injected Into the kiln through a conventional burner. The kiln
will dry and burn solids and will volatilize organic material, Including PCBs.
All organics will generally not be incinerated 1n the kiln and a high temperature
must be maintained in an afterburner at a specific residence time for destruc-
tion. To meet regulatory requirements, the afterburner 1s designed for a
residence time of two seconds at a temperature of 2,200ฐF.
A rotary kiln system used for the Incineration of toxic waste would Include
the kiln, provisions for feeding, supplemental fuel Injection, an afterburner and
an ash collection system. The gas discharge from the afterburner is directed to
C.4-18
-------
an air emissions control system. An Induced draft fan 1s provided within the
emissions control system to draw gases from the kiln through the equipment line
and then to discharge them via a stack to the atmosphere. The fan (or another
prime mover) Is sized to maintain a negative pressure throughout the system so
that gas leakage 1s always into, not out of, the kiln system.
Multiple Hearth Incineration
The multiple hearth furnace Is a vertical cylindrical structure, lined with
refractory material. It Is designed for sludges or other wastes that require
drying. Haste Is dropped on one of a series of six to ten horizontal hearths.
A center shaft rotates within the incinerator, wiping waste across one hearth.
At the edge of the hearth waste drops to a lower hearth.
Waste loses most of Its moisture on the top hearths, bums or loses
volatlles at the center of the furnace, and burns to a sterile ash at the lower
hearths of the incinerator. The multiple hearth furnace does not sustain
temperatures in excess of approximately 1,500ฐF at Its outlet. When It Is used
for PCB destruction, an afterburner is required to obtain the 2,200ฐF temperature
specified in USEPA's rules.
Off-gas from the Incinerator is directed toward an air emissions control
system, which normally Includes a quench section, a Venturi scrubber and a tray
tower for control of particulate and acid gases. The multiple hearth system 1s
a very large system, with many pieces of equipment and many movable components.
Because of Its size and Its complexity, 1t 1s not adaptable for transportable use
and 1s not recommended for the Hudson River site.
Fluid Bed Incineration
The fluid bed furnace Is a cylindrical, refractory-Hned shell with a
supporting structure above Its bottom surface to hold a sand bed (fluidlzed bed).
The structure has a series of tuyeres, which allow the passage of air upward into
C.4-19
-------
the bed while tending to prevent the passage of sand. Air 1s Introduced Into a
wlndbox and then through the tuyeres Into the sand bed. A high degree of
turbulence Is created 1n the sand bed by the passage of the air stream, which
creates a motion on the top of the bed that has the appearance of a fluid.
Fluidlzed bed systems are relatively large and operate at temperatures less
than 1,600ฐF. They are not adaptable to transportable systems and they do not
devielop sufficiently high temperatures to destroy PCBs. Their use for the Hudson
River site 1s not recommended.
Circulating Fluid Bed Incineration
In the circulating bed concept, a high air/gas velocity (from 15 to 20
feet/second) 1s Introduced Into a fluid bed. This high velocity, which 1s ten
times the velocity 1n a conventional fluid bed furnace, elutriates both the bed
and the combustible waste. Circulating material rises through a reaction zone
to the top of a combustion chamber and passes through a hot cyclone. Hot gas
passes through the cyclone, while the majority of solids drop to the bottom of
the cyclone and are re-injected into the bed of the furnace. The hot flue gases
pass to a gas cooler and then to a baghouse for removal of particulate.
Feed 1s sized to less than a one-half inch average particle size and is
introduced into the leg between the cyclone and the bed of the reactor. Waste
1s fed to the system through a feeding bin. A metering screw conveys waste from
the bin to the feed leg. The waste feed rate Is automatically adjusted to
maintain a pre-set oxygen concentration 1n the flue gas. L1me can be added to
the waste feed through a lime metering system to neutralize acId-generating
constituents of the waste, such as PCBs.
The design operating temperature for circulating systems 1s normally
1,600ฐF, although the system can withstand temperatures up to 2,000*F on a
continuous basis. A combustion air fan provides air to the bed for fluidlzation
and oxidation. Furnace draft is maintained by an induced draft fan downstream
C.4-20
-------
of the cyclone. Flue gas exiting the cyclone passes through a conventional
exhaust gas treatment system, which removes particulate and other undesirable
constituents from the gas stream.
Retention time of material within the system 1s controlled by monitoring
the discharge from the ash cooler. The cyclone bottom ash discharges to the
reactor, but this ash flow can also be removed from the system through a water-
cooled ash conveyor. By Increasing the speed of this conveyor, additional
material 1s removed from the furnace system and the residence time within the
system Is likewise controlled. For Instance, by lowering the conveyor discharge
rate, less material would be discharged from the system and the solids retention
time would be Increased.
A major feature of the circulating fluid bed system 1s Its ability to
control the residence time of wastes to over SO seconds. Generally, destruction
of organlcs, Including PCBs, has been found to occur at temperatures below,
1,600ฐF 1n this system. This lower temperature translates to lower supplemental
fuel requirements, less refractory maintenance, less severe bed eutectics, etc.
Conveyor Furnace
The conveyor furnace 1s essentially a conveyor belt system passing through
a long chamber lined with refractory material. An induced draft fan maintains
a negative pressure throughout the system. Combustion air 1s Introduced at the
discharge end of the belt. A1r will pick up heat from the hot burning tfaste as
waste and air travel countercurrent to each other. Supplemental heat 1s provided
by electric Infrared heating elements or by conventional fossil fuel burners
within the furnace above the belt. Cooling air is injected Into the Incinerator
chamber to prevent local hot spots 1n the Immediate vicinity of the heat-
ers/burners and 1s used as secondary combustion air within the furnace. The
furnace is designed to provide and maintain a temperature of 1,600ฐF above the
travelling conveyor.
C.4-21
-------
The conveyor belt 1s woven wire mesh made of high temperature alloy steel
that will withstand the 1,300ฐF to 1,600ฐF temperatures encountered within the
furnace. The refractory material used Is ceramic felt rather than brick. The
furnace does not have a high capacity for holding heat and can be started frtm
a cold condition relatively quickly, 1n one to two hours. Soil or other wastes
are fed by gravity on to the belt and are Immediately leveled to a depth of two
to three Inches. The waste must be sized to no more than a two-Inch effective
diameter. The belt speed and travel time are chosen to provide burnout of the
waste with minimal agitation. This feature results 1n a relatively low level of
particulate emissions.
This furnace system has been adapted for the treatment of soils contaminat-
ed with trace organlcs. The soils are heated 1n the basic unit to release their
organic contamination. The organlcs are directed, through the exiting gas
stream, to an external afterburner where they are fired at a sufficient
temperature and a specific residence time for destruction. Depending on the
furnace manufacturer, electric power, natural gas, propane or fuel oil Is used
as supplemental fuel for the system.
Electric Pyrolyzer
Waste-bearing materials must be processed to a maximum two-Inch average
particle s1ze*pr1or to Introduction Into this system. Soil can have a moisture
content of up to 25 percent before drying Is necessary. Solid waste 1s dropped
by gravity through the reactor, or pyrolyzer, while liquid feed Is Injected Into
the system.
The electrl^ pyrolyzer process promotes the release of organlcs from the
surface of soil or other material. As waste 1s dropped Into the pyrolyzer, 1t
passes through a high temperature zone where the majority of the organlcs
volatilize. The soil, or other solid waste, drops to the bottom of the
pyrolyzer, which 1s maintained at a high enough temperature to keep the soil and
other*1norgan1cs in a molten state.
C.4-22
-------
Supplemental electrodes within the melt assure that Its temperature trill
be maintained at a relatively high level and will be uniform throughout the melt.
Any metals present will be found in their elemental form or as a salt, they will
be removed from the melt on a continuous basis from an appropriately placed tap.
Other taps are located at other levels of the reactor wall to provide a meafts for
discharge of slag and other materials generated by the process. The tapped
materials fall Into a water bath where they cool Immediately. The cooled
residual has the appearance of dark glass. Any organlcs that did not volatilize
as the soil dropped through the reactor will be destroyed within the melt.
Even though the supplemental electrodes generate a temperature of
approximately 4,000ฐF, the overall melt 1s maintained at a much lower tempera-
ture. Chemistry of the melt can be controlled by additives such as lime, salts
or other compounds. By adjusting melt composition, neutralization of acid gas
components of the waste occurs and properties of the by-produet slag can be
controlled. 1
Off-gas from the reactor passes through a cyclone where the majority of
particulate that may be elutriated Into the gas stream is removed. A baghouse
removes the balance of particulate matter. A wet scrubber placed downstream of
the baghouse will remove any halogenated (acid) gases that were not neutralized
within the reactor.
Contract Thermal Destruction
There are several commercial Incineration facilities that dispose of PCB-
contaminated waste on a contract basis. All of these facilities utilize rotary
kiln systems for the disposal of wastes. One facility also has a rotary reactor,
which has been developed for the treatment of contaminated soils.
The rotary reactor 1s a hollow, three compartment horizontal cylinder that
rotates at 10 to 30 revolutions per minute. It acts as a horizontal fluid bed
reactor utilizing various inert medium for the bed, e.g., sand. Solid and
C.4-23
-------
seal-solid wastes, 1n addition to the Inert medium, are mechanically lifted on
Internal radial fins and cascade through the combustion gases In the combustion
zone of the reactor. This cascading action provides mass and heat transfer. The
recycling feature of this system provides relatively high residence time for the
solid material and allows operation at lower temperatures than with more
conventional Incineration equipment. The normal operating temperature for this
reactor 1s approximately 1,600ฐF. Lime can be added to the unit to neutralize
acid gases generated from the burning of halogenated organics, such as PCBs.
Status of Thermal Destruction Technologies
Several of the above described systems have been Identified as not
recommended for further analysis. Of those remaining, several have not been
developed further than pilot-scale. The screening of thermal destruction
technologies follows.
Low temperature desorption. Additional analyses not recommended.
Incineration
Rotary kiln. Recommended for consideration. Has been applied
to the destruction of PCBs in sediments, most recently the
clean-up of Waukegan Harbor outside of Chicago.
Multiple hearth. Additional analyses not recommended.
Fluid bed incinerator. Additional analyses not recommended.
Circulating fluid bed incinerator. Recommended for consider-
ation. Full-scale systems are operating 1n California and
Alaska on contaminated soils.
Conveyor furnace. Recommended for consideration. Full-scale
systems have been used for the clean-up of contaminated soils.
Five conveyor systems utilizing electric power are on the
market, and a number of systems utilizing fossil fuel are also
operating.
Electric Pyrolyzer. Recommended for consideration. This equipment
has been developed at pilot-scale and mounted on two trailers. It
1s available for demonstrations.
C.4-24
-------
Contract disposal. Not recommended for further consideration.
C.4.4.3 Biological Treatment Technologies
Bloremediation
Bloremediation Is a technique In which the physical, chemical and
biological conditions of a contaminated medium are manipulated to accelerate the
natural blodegradatlon and mineralization processes. (B1odegradat1on 1s a
process whereby microorganisms alter the structure of a chemical. Mineralization
1s the complete blodegradatlon of a chemical to carbon dioxide, water and simple
inorganic compounds.) In nature, both partial blodegradatlon and complete
mineralization take place; the processes, however, are frequently slow.
Bloremediation has been used in the treatment of sewage for a number of
years. It has been used fairly successfully under some conditions to treat
petroleum products, creosote and pesticide contamination. PCBs, however, pose
greater challenges to bloremediation than many other types of contamination.
Research 1s necessary before effective full-scale, biological treatment Is
available for these compounds.
Paramount to successful PCB bloremediation is the identification of a
microbial population capable of degrading a large number of different PCB
congeners. Abramowlcz (1990) has recently reviewed the PCB degrading capability
of aerobic and anaerobic strains. As summarized 1n that review, more than 25
aerobic strains demonstrating varying degrees of PCB-degrading competence and
specificity have been isolated. Some strains, such as Psuedoaonts sp. LB400 and
A1cellgenes eutrophus H850, have shown the ability to degrade a large number of
congeners, Including several penta-, hexa- and heptachlorinated congeners. Other
species have shown PCB-degrading capabilities complementary to LB400 and H8S0,
suggesting that treatment by two or more strains could yield even more
degradation than treatment with one of these strains alone. Aerobic biodegrada-
C.4-25
-------
tion, however, is generally limited to the less chlorinated PCB congeners. To
date, no aerobic strain has shown the ability to degrade Aroclor 1260.
Those organisms that have exhibited PCB-degrading capabilities have not
shown the ability to mineralize PCBs. Rather, the microbes degrade PCBs to
chlorobenzoates (in some cases chloroacetophenones) and five carbon aliphatic
compounds. Bioremedlatlon of highly chlorinated congeners and the products of
aerobic PCB biodegradatlon will, therefore, require a consortium of microbes that
can degrade these compounds as well. DETOX Industries Indicate that they have
identified a consortium of microbes that will mineralize PCBs, Including the more
highly chlorinated congeners of Aroclor 1260 (Philip Balls, DETOX Industries,
personal communication). As the organisms are proposed for commercial
bioremedlatlon, the details regarding this consortium are proprietary and the
information can not be verified.
Anaerobic organisms have shown the ability to dechlorinate reductively
heavily chlorinated PCB congeners. Dechlorination does not change the total
molar concentration of PCBs, since the products are less chlorinated blphenyls.
Dechlorination, however, does yield products that can be degraded by aerobes.
Thus, sequential anaerobic/aerobic treatment may enable treatment of more
chlorinated PCB mixtures.
Theoretically, an alternative to the sequential treatment of PCBs with
several strains is the genetic engineering of a wide range of blodegradative
capabilities Into a single or reduced number of coexisting organisms. Toward
this end, Mondello (1989) has cloned genes encoding PCB degradatlve enzymes from
several PCB-degrading strains. Enzyme activity 1n recombinant E. coll containing
these genes from Psuedoaonas sp. LB400 was nearly as great as in the donor
strain. More research is still required to engineer recombinants with the
ability to degrade a wide variety of congeners.
In addition to the identification of PCB-degrading microbes, successful
bioremedlatlon will require identification of the environmental factors
C.4-26
-------
controlling blodegradatlon. Research sponsored by General Electric 1s ongoing
to define the environmental conditions most conducive to PCB blodegradatlon. The
results of their research to date Indicate that optimum aerobic microbial
activity requires: 1) microbial growth on blphenyl or chloroblphenyl (Bedard,
1990; Mondello, 1989; NcDermott, 1989); 2) temperatures elevated above those that
would be characteristic of Hudson River sediments (NcDermott et ซ7., 1989); 3)
aeration (McDermott et al., 1989); and 4) sufficient PCB bioavailability. The
Inherent Insolubility of PCBs and high concentrations of natural organic matter
concentrations can reduce bioavailability (Brooks, 1989; Harkness and Bergeron,
1990). Optimum anaerobic activity for Hudson River strains or consortia appears
to require: 1) the absence of Inhibitors, such as sulfate (Tledje et a1.t 1989);
2) elevated PCB concentrations, I.e. greater than 50 ppm (Tledje et al.f 1987);
3) the presence of certain Inorganic nutrients (Abramowlcz et al., 1989); 4) a
supplemental carbon source (Tledje et al., 1989; N1es et al., 1990; Alder et al.t
1990); and 5) temperatures elevated above those that would be characteristic of
river sediments (Tledje et al.f 1989). GE will be conducting s1te-spec1f1c tests
this summer to Identify the Important environmental variables affecting
b1oremed1at1on of Hudson River sediments.
Once an acceptable microbial consortium and proper environmental variables
have been Identified, one of three different engineering approaches to
b1oremed1at1on can be taken: an in situ approach, a land-based approach or a
bioreactor approach.
In Situ Approach
For In situ treatment of Hudson River sediments, the contaminated sediments
would be left 1n place. This approach obviously limits the amount of control
that can be exercised over environmental variables during b1oremed1at1on and can
pose significant engineering difficulties 1n the uniform Introduction and mixing
of microbes or nutrients that may be required. Additionally, mixing would
require a containment system to prevent suspension and transport of contaminated
C.4-27
-------
sediments. Regular monitoring of sediment conditions and PCB concentration would
be necessary to assess remedial progress.
McDermott et a/. (1989) have conducted some small-scale field tests
situ soil bloremediatlon at a former racing drag strip 1n Glens Falls, New York.
The top 15 cm of soil at this site was contaminated with 50 to 500 ppm of Aroclor
1242. A test area was prepared by rototllUng the top 20 cm of soil to
homogenize vertically the PCB contamination. Two test plots were marked off and
covered with a transparent tent to protect the site from the elewnts. One plot
was dosed three times a week with LB400 (2 x 10* cells/ml); the control plot was
dosed three times a week with buffer. One-half of each test plot was mixed by
rototllUng prior to each dosing. After 20 weeks, approximately 25 percent of
the PCBs had been blodegraded In the top three cm of the LB400 dosed, unmixed
soil. The amount of blodegradatlon 1n the lower 17 cm of this soil was not
reported. Approximately ten percent of the PCBs had been blodegraded 1n the
LB400 dosed, mixed soil. No biodegradatlon was observed 1n either the mixed or
unmixed portions of the control plot. McDermott et al. (1989) suggest that these
results might be improved, if soil and moisture conditions were less extreme.
(Temperatures Inside the tent sometimes exceeded 50ฐC, rapidly drying the soil
and desiccating the bacteria.) Genetic engineering of a bacteria that could
withstand environmental extremes and/or dosing with a bacterial culture showing
PCB degradatlve capacities complementary to LB400 might also Improve results
(McDermott et al, 1989).
No full-scale, In situ bloremediatlon of PCB-contam1nated sediments has
been conducted to date. At Sheboygan River and Harbor Superfund Site, however,
a type of limited In situ bloremediatlon not involving the Introduction or mixing
of microbes and nutrients 1s planned for some of the PCB-contam1nated sediments.
These sediments will be capped, with the expectation that this capping will
prevent transport of the sediments and Increase anaerobic conditions, thereby
enhancing anaerobic dechlorination already believed to be 1n progress. Capping
will take place only 1n areas that will not be disturbed by dredging. PCB-
contaminated sediments that are not capped will be removed. While relatively
C.4-28
-------
easily accomplished from an engineering standpoint, this bloremedlal solution may
provide little reduction 1n total molar concentration of PCBs. Anaerobic
processes are likely limited to reductive dechlorination without cleavage of the
blphenyl ring and certain congeners may prove resistant to dechlorination by the
indigenous microbes.
Land-Based Approach
Two types of land-based biological treatment approaches can be used for
bloremedlatlon of sediments: composting and land farming. Implementation of
either treatment for the Hudson River site would require dredging and may require
dewaterlng of the sediments. Since anaerobic conditions would be difficult to
maintain 1n both composting and landfarmlng, these systems would be appropriate
only if bloremedlatlon can be carried out aeroblcally. Additionally, some other
type of treatment system would be necessary for the water resulting from
dewaterlng. ,
In composting, sediments would be placed 1n large piles. A typical compost
pile may be six to eight feet In height and contain from 4,000 to 10,000 cubic
yards of sediment. Sediments would be placed on top of a prepared clqr or
plastic Uner with a leachate collection system In compliance with RCRA minimum
technology requirements. Oxygen would be supplied to the material through a
piping system. Installation of a watering system to maintain appropriate
moisture levels and deliver nutrients and possibly microbes to the piles would
likely also be necessary. In cool weather, steam or heated water may be
delivered to the piles to maintain elevated temperatures as well as to supply
moisture and nutrients. Collected leachate can be recirculated. The final
component of each pile would be a cover to minimize particulate emissions and
surface run-off. Depending upon the permeability of the sediments, they may
require mixing with a bulking agent prior to piling to develop a permeability
that will allow penetration of nutrients, water, air and/or microbes. Regular
monitoring of both pile conditions and changes In PCB concentration would be
necessary to assess remedial progress.
C.4-29
-------
In land farming, sediments would be spread over liners and leachate
collection systems 1n 9 to l8-1nch layers or lifts, for treatment. Although land
farms require more space for treatment than compost piles, a number of lifts can
be land-farmed sequentially over the same surface area. A sprinkler system would
be used to deliver microbes., nutrients and moisture to the land farm. Oxygen can
be supplied through periodic tilling. As land farms are generally not covered,
there 1s the potential for loss to the air of more volatile congeners, especially
during Initial spreading and tilling.
Compost piles and land farms have been used at a number of sites, with
varying degrees of success to treat soil contaminated with solvents, petroleum
products or creosote. These contaminants, however, are generally more readily
blodegraded than PCBs. Furthermore, some of the treatment success attributed to
these approaches is probably a result of compound loss through volatilization or
air stripping rather than actual compound destruction. No full-scale PCB
bloremediatlon projects using these techniques have been completed to date; few
vendors have experience treating any compounds 1n soil or sediment volumes
greater than 10,000 cubic yards.
Bloreactor Approach
In a bloreactor approach to bloremediatlon of the Hudson River sediments,
sediments would be dredged and then biologically treated 1n a container or
reactor. Because sediments are completely contained, a bloreactor approach
offers greater control over environmental variables than either an In situ or
land-based system. Furthermore, this approach is the only one of the three that
is conducive to the sequential anaerobic/aerobic treatment that may be required
to treat more heavily chlorinated PCBs. Remediation 1n a bloreactor, however,
may provide little advantage over alternative remedial techniques 1n terms of
sediment handling requirements and destruction efficiency.
Several commercial vendors claim experience with bloreactor technology, but
few vendors with demonstrated pilot-scale (or larger scale) experience in
C.4-30
-------
treating PCBs 1n bloreactors were Identified. DETOX Industries has completed a
pilot demonstration project 1n which 500 pounds of PCB-contanlnated mixed-wastes,
Including sludge, soil, electrical capacitor oil and water, were placed 1n an
open-air reactor. Initial concentrations of PCBs (some Aroclor 1260 as well as
lesser chlorinated mixtures) were as high as 2,000 ppm in sludge and 44,000 ppm
1n the liquid phase. Concentrations were aeroblcally reduced to 4 ppm overall
within 16 months (Joe Dally, DETOX Industries, personal communication). ENSR has
completed pilot-scale tests at the French Limited Superfund Site. Heavy sludge
contaminated with 1,600 ppm levels of lesser chlorinated PCBs (Aroclors 1232,
1242 and 1248) were aeroblcally treated 1n 2,000 gallon tanks. At the time the
tests were terminated, total PCB concentration had been reduced to 66 ppm.
Additional reduction 1n concentration may have been achievable, If the tests had
been carried out for longer periods of time (Richard Woodward, ENSR, personal
communication). The effect of volatilization on the performance of open air
bloreactors 1s also a consideration for Its use.
Using the results of treatability tests carried out at the University of
Michigan, Blasland and Bouck plan to conduct pilot-scale tests on PCB-contamlnat-
ed sediments from Sheboygan River and Harbor. It 1s expected that treatment will
Include anaerobic/aerobic cycles. These tests are scheduled to take place within
one year.
Additional evaluation of the applicability of bloremedlatlon for treatment
of Hudson River sediments should Include:
1) Laboratory demonstration that a consortium of microbes can degrade
all of the PCB congeners present In Hudson River sediments as well
as potentially undesirable blodegradatlon products;
2) Identification of optimum environmental conditions for consortium
activity;
3) Estimation, based on laboratory experiments, of maximum PCB
blodegradatlon rates achievable and the time required to biological-
ly treat Hudson River sediments;
C.4-31
-------
4) Demonstration that biodegradatlon results comparable to those
obtained 1n the laboratory, can be obtained on a larger scale 1n the
field, i.e., need for pilot-scale tests;
5) Comparative analysis of 1n-r1ver b1oremediat1on results and land-
based system results; and
6) The effect of PCB concentrations on bloremedlal techniques.
C.4.S Disposal Technologies
Disposal technologies Include upland disposal 1n a lined landfill, off-site
disposal 1n a permitted disposal facility and confined aquatic disposal. There
are nt> off-site, permitted facilities 1n the project area.
The use of subaqueous depressions or borrow pits can provide confinement
of contaminated material 1n an open-water disposal scenario. The US Army Corps
of Engineers has conducted research on subaqueous confinement and has utilized
the technology 1n the New York Harbor area. For the Upper Hudson, this
technology will not be considered further, since this part of the river has a
relatively steep gradient.
Disposal 1n a secure, lined landfill provides for long-term storage of
contaminated material and necessitates siting, design, construction, operation,
closure, post-closure monitoring and maintenance. The facility must meet
regulatory requirements, Including groundwater standards, for landfills In which
PCB-contam1nated materials will be stored. The landfill may be capped with an
Impermeable cover after each dredging season. A roughing and storage pond, surge
pond, water treatment plant, pump station, leachate collection system, stormwater
drainage system and chemical feed system are components of the design.
Construction site electrical services, fencing and seeding are required. A
secure landfill provides a high degree of Isolation of the contaminated material
with a low probability of subsequent discharge. The technology for landfill
design 1s routine and 1s readily available.
C.4-32
-------
In 1982, a location 1n the Town of Fort Edward, known as Site 10, was
chosen for the continued disposal of contaminated river sediments. In 1983f
approvals for Site 10 were revoked by the New York State Supreme Court on
grounds, among others, of violating local zoning regulations. The decision was
upheld by New York's Court of Appeals 1n 1985. Several alternate disposal sites
were subsequently analyzed. A site known as Site G, also located 1n the Town of
Fort Edward, was selected as a preferred alternative for a reduced-scale disposal
effort. Subsequently, a 1987 amendment to the New York State Siting Law allowed
the continued study of disposal Site 10. Site 10 has now been chosen by the
NYSDEC Project Sponsor Group as the preferred disposal site for approximately
three million cubic yards of river sediments, remnant deposits and PCBs
previously removed from the Hudson (NYSDEC, 1989).
C.4-33
-------
PAGE INTENTIONALLY LEFT BLANK
C.4-34
-------
C.S Innovative Treatment Technologies (USEPA SITE Progran}
Innovative technologies that have at least preliminarily demonstrated a
capacity to separate PCBs from solids, sludge or sediments Include:
V
AOSTRA TACIUK Process System So11Tech Inc.
Englewood, CO 80112
Solidification/Stabilization USEPA Risk Reduction
Processes Engineering Lab (RREL)
Cincinnati,.OH
The Soil Tech system has previously been named the Alberta 011 Sands
Technology and Research Authority (AOSTRA) - Tacluck process. SollTech, Inc.
holds the exclusive US license to apply this technology and 1s owned by Canonle
Environmental Services Corp. and LIMATAC Industrial Processes.. The SollTech
concept 1s a continuous pyrolysls system using a rotary drum heater with an Inner
core and two subatmospherlc processing chambers. The first chamber operates at
variable temperatures to about 600 ฐF and volatilizes water and light hydrocar-
bons from Injected wastes; The water and organlcs are collected and condensed.
The second chamber operates at temperatures of 1,000 to 1,100ฐF. In this chamber,
heavier hydrocarbons are evaporated and partially pyrolyzed to yield lighter gas
fractions and some coke. The vapor from this chamber 1s also collected and
condensed. For feed soils or sludges containing PCBs, the PCBs are recovered as
condensate 1n an oil phase.
Solids residues pass out of the second or light temperature chamber Into
the outer shell of the drum. Gas or oil-fired burners at the end of the outer
shell provide primary process heat with the Injection of combustion air. The
solids are heated 1n this zone to about 1,300 to 1,400ฐF and some hot solids are
recycled back Into the second chamber to create effective heat transfer. Clean
sol Ids are output from the unit. Other components of the major process equipment
are a flue gas train for removing particulate matter and acid gases, a preheat
vapor train and a retort vapor train, which condenses h1gh-bo1l1ng vapors and
separates vapors from liquids.
C .5-1
-------
The process has been developed at production levels and has been applied
to treating PCB-contamlnated soil at Wide Beach, New York and PCB-contam1nated
sediment at Waukegan Harbor, Illinois. SollTech's pilot demonstration unit has
a nominal capacity of five tons per hour; a commercial transportable unit has
a capacity of ten tons per hour. The latter 1s currently being used 1n treating
21,000 tons of PCB-contam1nated soils at the Wide Beach, NY Superfund site. The
SoilTech system Is retained here for further analyses and bench-scale testing.
A group of chemical fixation technologies that Immobilize contaminants
within the waste have emerged through the USEPA SITE Program. These technologies
Involve mixing waste material with settling agents .to enhance the physical
properties of the waste. Numerous commercial settling agents have been tested.
These agents either eliminate free water from the waste or alter the chemical
form of the contaminants to make them resistant to leaching.
A bench-scale study of solidification/stabilization as a treatment
technology for New Bedford Harbor sediments was conducted by the US Army Corp of
Engineers (1989). Composite sediment samples were processed with various dosages
of settling agent formulations, Including Portland cement, Portland cement with
Flrmax proprietary additive and a Silicate Technology Corporation (STC)
proprietary additive. Batch leaching tests showed that the leachablllty of PCBs
was reduced by factors of 10 to 100. Costs for treating New Bedford Harbor
sediments using the tested agents have been estimated at $100 per ton. While the
solidification/stabilization approaches offer potential low cost treatment
options, data on the long-term aging effects of the stabilized/solidified matrix
should be developed further.
Recently, USEPA's Risk Reduction Engineering Laboratory (RREL) 1n
Cincinnati Initiated a project with RMC Environmental of West Plains, Missouri
to conduct controlled experiments on PCB-contamlnated soils. The experiments were
conducted to investigate declining concentrations of PCB over time, which were
observed at contaminated sites that were stabilized through the addition of 11me
and other alkaline materials. The study has been recently published and will be
reviewed.
C.5-2
-------
C.6 Initial Screening of Technologies
While no particular technology has been removed from further consideration
for subsequent phases, 1t 1s possible at this Initial screening stage to render
some judgments concerning remedial options, based upon their applicability and
current level of development. Data upon which the Initial screening was based
were obtained from numerous sources, Including reports for other Superfund sites,
USEPA's technology assessment documents and direct communications with equipment
manufacturers. Results of this Initial screening effort are Illustrated as Figure
C.6-1.
Several technologies associated with particular response actions were not
screened 1n this preliminary reassessment. These Include methods to excavate
remnant deposits, If necessary. In addition, technologies applicable to treating
water resulting from sediment dewaterlng operations have not yet been evaluated.
A wide range of well-proven, commercially viable technologies are available to
treat effluent from dewaterlng operations. These will be evaluated In subsequent
phases.
Mechanical, hydraulic and specialty dredging systems or conventional
excavation methods are available to remove contaminated sediments. While
hydraulic systems have been preferred at other Superfund sites and have been
shown to minimize sediment resuspenslon during removal operations, these systems
result 1n the need to handle significant quantities of by-product water and tend
to be most cost-effective for dredging relatively large quantities of sediment.
Thus, the three generic dredging systems have been retained for further
assessment, when additional Information on materials characteristics and
quantities will become available.
Should a decision be made to remediate the Hudson River site by removing
some or all Its contaminated bottom materials, It would be necessary either to
landfill the removed materials or to treat those materials and landfill the
treated residuals. Physical, chemical, biological and thermal processes or
technologies are available to treat PCB-contamlnated solids. Of the large range
of treatment alternatives available, those considered to be either commercially
C.6-1
-------
available or sufficiently well developed for further consideration are
illustrated on Figure C.6-1. The decision to remove treatment technologies from
further consideration this time was based on data contained in reports for other
Superfund sites and on personal communications with system vendors.
As an alternative to removal, technologies applicable to remediating the
Hudson River site without removing its contaminated sediments have also been
considered. In situ technologies include those that contain the bottom sediments
as well as those that treat them. Containment systems, such as capping and
retaining structures, are methods for controlling sediment resuspenslon under the
range of expected river hydraulic conditions. In situ treatment methods,
primarily chemical and biological technologies, are also undergoing testing and
have been retained for further assessment. As illustrated on Figure C.6-1
generic in situ remedial technologies will continue to be evaluated and will be
considered applicable to either partial or complete remediation of the river's
contaminated sediments.
A final category of remedial technologies considered here 1s that which
involves disposal of either untreated contaminated soils or the residuals from
treatment of contaminated soils. Off-site disposal at comnercial landfills has
been discounted at this time because of the lack of nearby permitted landfills
and the relatively large quantities of materials that would have to be hauled a
considerable distance to any commercial facilities. On-site or 1n-river disposal
and upland disposal options are being retained for evaluation in the next phase.
Upland disposal, which would Involve obtaining approval to construct a proximate
landfill specifically for the river's contaminated sediments, 1s the disposal
technology being pursued by the NYSDEC Project Sponsor Group at this time.
C.6-2
-------
C.7 Treatability Studies
The decision to remediate the Hudson's contaminated sediments will be
significantly Influenced by the feasibility and costs of remedial technologies.
Treatability studies can be an effective technique to evaluate remedial
technologies, because such studies reduce equipment performance uncertainties and
lead to Improved estimates of treatment system cost. Since treatability studies
have not yet been performed on Hudson sediments, even a limited program would be
expected to provide considerable, useful Information.
Bench-scale treatability studies are particularly appropriate where
emerging chemical/physical technologies are being evaluated. Bench-scale tests
are usually performed 1n a laboratory and require relatively small quantities of
the contaminated material to achieve their objective. These tests are conducted
to determine the effectiveness of a particular processes chemistry and to test
a wide range of process operating variables. Bench-scale work can also be used
to set parameters for full-scale tests, should a particular technology warrant
further consideration.
Four physical/chemical technologies, KOHPEG, B.E.S.T., LEEP and Propane
Extraction, are being brought forward for further consideration on the basis of
preliminary screening. Bench-scale work to establish basic operating parameters
and costs for these would be appropriate 1n subsequent project phases. The
developers of selected technologies will be contacted to establish time frames,
costs, material quantities, and general goals for a treatability program. In
general, between two and five months would be required for the bench-scale work
(depending on the technology) and 1t would be necessary to obtain about ten
pounds of bottom material for each test.
Bench-scale analyses related to thermal treatment systems are not
recommended at this time. It will be necessary, at some point, to establish
parameters such as the temperature at which PCBs are released from the river
sediments, the sediment ash fusion temperature as well as the composition of the
bottom material 1n terms of total chlorides, heavy metals, and total organics.
Since thermal systems have been extensively applied to remediating contaminated
C.7-1
-------
soil, there 1s less need to establish their general feasibility at this time.
In addition, cost estimates for thermal systems are available from numerous other
projects and can, on a preliminary basis, be adapted to the Hudson River
C.7-2
-------
REFERENCES
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
Abramowicz, D.A. 1990. Aerobic and anaerobic biodegradation of PCBs: a review.
Critical Reviews in Biotechnology, 10(3):241-251.
Abramowicz, D.A., M.J. Brennan and H.M. Van Dort. 1989a. Microbial
dechlorination of PCBs: I. Aroclor mixtures. In Research and Development Program
for the Destruction of PCBs, Eighth Progress Report. Schenectady: General
Electric Company Corporate Research and Development Center.
Abramowicz, D.A., H.M. Van Dort and M.J. Brennan. 1989b. Microbial
dechlorination of PCBs: II. Single congeners. In Research and Development
Program for the Destruction of PCBs, Eighth Progress Report. Schenectady:
General Electric Company Corporate Research and Development Center.
Alder, A.C., M. Haggblom and L.Y. Young. 1990. Reductive dechlorination of PCBs
in sediments from the Hudson River and New Bedford Harbor. In Research and
Development Program for the Destruction of PCBs, Ninth Progress Report.
Schenectady: General Electric Company Corporate Research and Development Center.
A1ford-Stevens, A.I. 1986. Analyzing PCBs. Environmental Science t Technology
20(12):11941199.
Allan, J.D. 1976. Life history patterns in zooplankton. Am. Nat. 110:165-180.
Alonso, C.V., Neibling, U.H. and Foster, G.R. 1981. Estimating sediment
transport capacity in watershed modeling. TRANSACTIONS of the ASAE 24(5):
1211-1220, 1226.
Ambrose, R. A. Jr., T. A. Wool, J. P. Connolly and R. W. Schanz. 1988. "WASP4,
A Hydrodynamic and Water Quality Model; Model Theory, User's Manual, and
Programmer's Guide." EPA/600/3-87/039. USEPA Environmental Research Laboratory,
Athens, GA
American Conference of Governmental Industrial Hygienists (ACGIH). 1986.
Documentation of the Threshold Limit Values and Biological Exposure Indices.
Fifth edition.
Andrle, R.F. and J.R. Carroll. 1988. The Atlas of Breeding Birds in New York
State. Cornell University Press. Ithaca, New York.
Anne, M., R. McLane, and D. Hughes. 1980. Reproductive success of screech owls
fed aroclor 1248. Arch. Environm. Contam. Toxicol. 9: 661-665.
R-l
-------
Anonymous. 1960. "Hudson River Investigation." State of New York Conservation
Department. Typed report 3 pp.
Apicella, G.A. and T.F. Zimmie. 1978. Sediment and PCB transport model of the
Hudson River, pp. 645-653. In: Proc. 26P Ann. Hydraulics Division Specialty
Conference, ASCE., August 9-11, 1978. University of Maryland, College Park,
Applegate, R.L. and J.W. Mull an. 1966. Food of the black bullhead (Ictalurus
melas) in a new reservoir. Proc. Southeast. Assoc. Game Fish Com. 20: 288-292
Armstrong, R.W. and R.J. Sloan. 1980a. "PCB Patterns in Hudson River Fish, I.
resident/freshwater Species." Proc. Hudson River Environ. Soc., Hyde Park, NY.
Armstrong, R.W. and R.J. Sloan. 1980b. "Trends in levels of several known
chemical contaminants in fish from New York State waters." Technical Report 80-2,
NYSDEC.
Arulanandan, K., Gillogley, E., and Tully, R. 1980. "Development of a
quantitative method to predict critical shear stress and rate of erosion of
natural undisturbed cohesive soils." Report GL-80-5, US Army Engineers,
Waterways Experiment Station, Vicksburg, Miss.
ATSDR (Agency for Toxic Substances and Disease Registry). 1987. "Toxicological
Profile for Selected PCBs (Aroclor-1260, -1254, -1248, -1242, -1232, -1221, and -
1016)."
Aulerich, R.J. and R.K. Ringer. 1977. Current status of PCB toxicity, including
reproduction to mink. Arch. Environm. contain. Toxicol. 6: 279-292.
Baes, C.F., R.D. Sharp, A.L. Sjoreen, and R.W. Shor. 1984. "A Review and
Analysis of Parameters for Assessing Transport of Environmentally Released
Radionuclides through Agriculture." Oak Ridge National Laboratory, US Department
of Energy, ORNL-5786.
Bahn, A.K., Grover, P., Rosenwaike, I., 0'Leary, K., and Stellman, J. 1977. PCBs
and melanoma. N. Engl. J. Med., 296, 108. As cited in Silberhorn et al., 1990.
Bahn, A.K., Rosenwaike, I;, Herrmann, N., Grover, P., Stellman, J., and 0'Leary,
K. 1976. Melanoma after exposure to PCBs. N. Engl. J. Med., 295, 450. As
cited in Silberhorn et al., 1990.
Barnes, C. R. 1987. Polychlorinated biphenyl transport rates in the Upper Hudson
River, New York, 1977-1983. Northeastern Environmental Science 6(1): 31-36.
Barnthouse, L.W., R.J. Klauda, D.S. Vaughan and R.L. Kendall, editors. 1988.
Science, law and Hudson River Power Plants: a case study in environmental impact
assessment. American Fisheries Society Monograph 4, 347pp.
Bartek, M.J., J.A. La Budde, and H.I. Maibach. 1972. Skin permeability in vivo:
comparison in rat, rabbit, pig and man. J.Invest. Dermatol. 58: 114-123.
R-2
-------
Bason, W.H. 1981. Ecology and early life history of striped bass, Morone
saxatilis, in the Delaware estuary. M.S. Thesis, Cornell University, Ithaca,
N.Y. 122 p.
Bath, D., Beebe, C.A., Dew, C.D., Reider, R.H. and J.H. Hecht. 1976. "A List
of Common and Scientific Names of Fishes Collected from the Hudson River." Paper
No. 33, Hudson River Ecology 4th Symposium, March 28-30, 1976.
Bedard, D.L. 1990. Bacterial transformation of polychlorinated biphenyls. D.
Kamely, A. Chakrabarty and G.S. Omen [eds.] In Biotechnology and Biodegradation,
Advances in Applied Biotechnology, Vol. 4. The Woodlands, TX: Portfolio Pub.
Co.
Bedard, D.L., R.E. Wagner, M.J< Brennan, M.L. Haberl and J.F. Brown. 1987a.
Extensive degradation of aroclors and environmentally transformed polychlorinated
biphenyls by Alcaligenes eutropus H850. Applied and Environmental Microbiology,
53(5): 1094-1102.
Bedard, D.L., M.L. Haberl, R.J. May, M.J. Brennan. 1987b. Evidence for novel
Jiechanisms of polychlorinated biphenyl metabolism in Alcaligenes eutropus H85Q.
Applied and Environmental Microbiology, 53(5): 1103-1112.
Bedard, D.L., R. Unterman, L.H. Bopp, M.J. Brennan, M.L. Haberl and C. Johnson.
1986. Rapid assay for screening and characterizing microorganisms for the
ability to degrade polychlorinated biphenyls." Applied and Environmental
Microbiology 51(4): 761-768.
Beebe, A. and I.R. Savidge. 1988. Historical perspective on fish species
composition and distribution in the Hudson River estuary. American Fisheries
Society Monograph 4: 25-36.
Bek, T.A. 1972. Feeding of intertldal gammarids. Vestn. Mosk. Univ. Ser. 6.
Biol. Pochvoved 27: 106-107.
Belton, T., Robert Roundy, and Neil We1nste1n. 1986. Urban fishermen: managing
the risks of toxic exposure." Environment 28(9): 19-37.
Berger, H. personal communication. Waterford Water Works, Waterford New York.
Bertazzl, P.A., Riboldi, L., Pesatorl, A., Radice, I., and Zocchetti, C. 1987.
Cancer mortality of capacitor manufacturing workers. Am. J. Ind. Med. 11, 165.
As cited in Silberhorn et al., 1990.
Bode, R.W. 1979. 0EC Tech. Memorandum (unpublished). A Comparison of tke
1972/73 and 1977 Results, Hudson River.
Bode, R.W., M.A. Novak and L.E. Abele. 1991. Quality Assurance Work Plan for
Biological Stream Monitoring in New York State. NYS0EC, Feb. 26, 1991, 79 pp.
R-3
-------
Bode, R.W., M.A. Novak and L.E. Abele. 1990. Biological Stream Assessment
Eighteenmile Creek Niagara County, N.Y. Stream Biomonitoring Unit, Bureau of
Monitoring and Assessment, Division of Water, NYS Dept. of Environmental
Conservation Albany, N.Y. 25pp. NYS Dept. of Health.
Bode, R.W., M.A. Novak, J.P. Fagnani and D.M. DeNicola. 1986. "The Benthic
Macroinvertebrates From Troy to Albany, New York." Final Report to Hudson River
Foundation
Boesch, D.F. and D.J. Diaz. 1974. New records of peracarid crustaceans from
oligohaline warters of the Chesapeake Bay. Ches. Sci. 15: 56-59.
Bopp, R.F. 1983. Revised parameters for modelling the transport of PCB
component across an air-water interface, Jnl. Geophys. Res. 88: 2521-2529.
Bopp, R.F. 1979. The Geochemistry of Polychlorinated Biphenyls in the Hudson
River. Ph.D. Dissertation, Columbia University, New York, New York.
Bopp, R.F. and H.J. Simpson. 1989. Contamination of the Hudson River, the
sediment record, pp. 401-416 In Contaminated Marine Sediments - Assessment and
Remediation. National Academy Press, Washington, DC.
Bopp, R.F. and H.J. Simpson. 1988. "Sources and Trends of Persistent
Contaminants in the New York Harbor Complex." Final Report to the Hudson River
Foundation, 54 pp.
Bopp, R.F. and H.J. Simpson. 1985. "Persistent Chlorinated Hydrocarbon
Contaminants in the New York Harbor Complex." Final Report to the Hudson River
Foundation, 54 pp.
Bopp, R.F. and H.J. Simpson. 1984. "Persistent Chlorinated Hydrocarbon Contami-
nants in the New York Harbor Complex." Final Report to the Hudson River
Foundation, 152 pp.
Bopp, R.F., H.J. Simpson, and B.L. Deck. 1985. "Release of Polychlorinated
Biphenyls from Contaminated Hudson River Sediments." Final Report NYS C00708 to
NYSDEC, Albany, NY, Lamont Doherty Geological Observatory of Columbus University,
Palisades, NY.
Bopp, R.F., H.J. Simpson, B.L. Deck and N. Kostyk. 1984. The persistence of PCB
components in the sediments of the lower Hudson, Northeastern Environ Sc1 3: 180-
184.
Bopp, R.F., H.J. Simpson, C.R. Olson, R.M. Trier and N. Kostyk. 1982.
Chlorinated hydrocarbons and radionuclide chronologies in sediments in the Hudson
River and estuary, Environ. Sci. Techno J. 16: 666-672.
Bopp, R.F., H.J. Simpson, C.R. Olsen, R.M. Trier and N. Kostyk. 1981.
Polychlorinated biphenyls 1n the sediments of the tidal Hudson River, New York.
Environ. Sci, Technol. 15: 210-216.
R-4
-------
Borah, O.K. and P.K. Bordoloi. 1991. "Stream bank erosion and bed evolution
model." Manuscript under'revision for possible publication In the Journal of
Hydraulic Engineering, ASCE.
Borah, D.K. and P.K. Bordoloi. 1989a. Nonuniform sediment transport model.
TRANSACTIONS of the ASAE 32(5): 1631-1636.
Borah, D.K. and P.K. Bordoloi. 1989b. Stream bank erosion and bed evolution
model. In Proceedings of the International Symposium on Sediment Transport
Modeling, New Orleans, LA, August 14-18, ASCE, pp. 612-617.
Borah, O.K., C.V. Alonso, and Prasad, S.N. 1982a. Routing graded sediments In
streams: formulations. Journal of the Hydraulic Division, ASCE 108(12):
1486-1503.
Borah, D.K., C.V. Alonso, and Prasad, S.N. 1982b. Routing graded sediments 1n
streams: applications. Journal of the Hydraulic Division, ASCE 108(12):
1504-1517.
Borgmann, U., W.P. Norwood, and K.M. Ralph. 1990. Chronic toxicity and
bloaccumulatlon of 2,5,2',5'- and 3,4,3',4'- tetrachlorobiphenyl and aroclor 1242
1n the amphipod Hyalella azteca. Arch. Environm. Contam. Toxicol. 19: 558-564.
Boyce Thompson Institute. 1977. An Atlas of the Biological Resources of the
Hudson Estuary. Boyce Thompson Institute of Plant Research, Estuarlne Study
Group, Yonkers, N.Y.
Boyle, R.H. 1970. Poison roams our coastal seas. Sports Illustrated 33(26 Oct.
1970): 70-74.
Boyle, R.H. 1969. The Hudson River. W.W. Norton & Co., Inc., N.Y., 304pp.
Bras, R. L. 1990. Hydrology, an Introduction to Hydrologic Science. Addison-
Wesley, Reading, MA
Brilliant, L.B., G.V. Amburg, J. Isblster, H. Humprey, K. Wilcox, J. Eyster, A.
W. Bloomer, and H. Price. 1978. Breast milk monitoring to measure Michigan's
contamination with polybrominated biphenyls. Lancet 2: 643-686.
Brltton, W.M. and T.M. Huston. 1973. Influence of polychlorinated bi phenyls in
the laying hen. Poultry Science. 52: 1620-1624.
Brooks, R. 1991. Personal communication, Washington County Planning Department.
Brooks, R.E. 1989. Aroclor 1248 and drag strip PCB biodegradation. In Research
and Development Program for the Destruction of PCBs, Eighth Progress Report.
Schenectady: General Electric Company Corporate Research and Development Center.
Brown, D.P. 1987. Mortality of workers exposed to polychlorinated biphenyls -
an update." Arch. Environ. Health, 42, 333. As cited 1n Silberhorn et al.,
1990.
R-5
-------
Brown, D.P. and M. Jones. 1981. Mortality and industrial hygiene study of
workers exposed to polychlorinated biphenyls. Arch. Environ. Health, 36, 120.
As cited in Silberhorn et al., 1990.
Brown Jr., J.F., D.L. Bedard, M.J. Brennan, J.C. Carnahan, H. Feng and R.E.
Wagner. 1987a. Polychlorinated biphenyl dechlorination in aquatic sediments.
Science 236: 709-712.
Brown Jr., J.F., R.E. Wagner, H. Feng, D.L. Bedard, M.J. Brennan, J.C. Carnahan,
and R.J. May. 1987b. Environmental dechlorination of PCBs. Environmental
Toxicology and Chemistry 6: 579-593.
Brown Jr., J.F., R.E. Wagner, and D.L. Bedard. 1988a. PCB dechlorination in
Hudson River sediment. Science 240: 1675-1676.
Brown Jr., J.F., R.E. Wagner, D.L. Bedard, M.J. Brennan, J.C. Carnahan, and R.J.
May. 1984. PCB transformations in Upper Hudson sediments. Northeastern
Environmental Science 3(3/4): 166-178.
Brown Jr., J.F., R.E. Wagner, D,. L. Bedard, M.J. Brennan, J.C. Carnahan, R.J.
Mary and T.J. Tofflemire. 1985. "PCB Dechlorination in Upper Hudson Sediments."
Paper presented to the Division of Environmental Chemistry, American Chemical
Society, April 1985.
Brown, M.P. and M.B. Werner. 1985. "Distribution of PCBs in the Thompson Island
Pool of the Hudson River, PCB Hot Spot Confirmation Report." NYSDEC, Albany, New
York.
Brown, M.P. and M.B. Werner. 1983. "Recent Trends in the Distribution of
Polychlorinated Biphenyls in the Hudson River System." Report to NYSDEC, Albany,
NY, cited in NUS, 1984.
Brown, M.P., M.B. Werner, C.R. Carusone, and M. Klein. 1988b. "Distribution of
PCBs in the Thompson Island Pool of the Hudson River: Final Report of the Hudson
River PCB Reclamation Demonstration Project Sediment Survey." Division of Water,
New York State Department of Environmental Conservation, Albany, NY.
Brown, M.P., M.B. Werner, R.J. Sloan and K.W. Simpson. 1985. Polychlorinated
biphenyls in the Hudson River, recent trends in the distribution of PCBs in
water, sediment and fish. Environmental Science and Technology 19(8): 656-661.
Brunner, C. 1990. "On-Site Incineration, Updated." Initially Presented at the
International Conference on Incineration and Radioactive Wastes, San Francisco,
CA. 1988.
Brunner, C. 1989. Handbook of Hazardous Waste Incineration. TAB Professional
and Reference Books. Blue Ridge Summit, PA.
Brunner, C. 1988. Site Cleanup by Incineration. Hazardous Materials Control
Research Institute. Silver Spring, Maryland.
R-6
-------
Brunner, S., E. Hornung, H. Santl, E. Wolff, 0.6. Piringer, J. Altschuh, and R.
Bruggemann. 1990. Henry's law constants for polychlorinated biphenyls:
experimental determination and structure-property relationships. Environmental
Science and Technology 24: 1751-1754.
Buckley, E.H. 1987. Phytochemical effects of environmental compounds. In Recent
Advances in Phytochenistry (Eds. J. A. Saunders and L. Kosak-Channing and E.EConn)
Plenum Press, New York.
Buckley E.H. 1983. Decline of background PCB concentrations in vegetation in
New York State. N.E. Environ. Sci. 2(3/4): 181-187
Buckley, E.H. 1982. Accumulation of airborne polychlorinated biphenyls in
foliage. Science 216: 520-22.
Buckley E.H. and T.J. Tofflemire. 1983. Uptake of airborne PCBs by terrestrial
plants in the tailwater of a dam. Proc. Nat. Conf. on Environ. Eng. 1983 ASCE
Specialty Conf. July 6-8, 1983, Boulder, CO, pp. 662-669.
Bufflngton, B. 1991. New York State threatened and endangered species list,
provided in letter from Burrell Buffington, dated January, 18, 1991, of the
Wildlife Resources Center, Significant Habitat Unit, NYSDEC, Albany, NY.
Burke, M. 1991. Personal communication. NYSDOH, Environmental Protection
Division, NY.
Burkholder, P.R. and R. Bere. 1933. VIII plankton studies in some lakes of the
upper Hudson watershed. In A Biological Survey of the Upper Huson Watershed, pp.
239-263. N.Y. State Conservation Dept., Supp. to 22nd Annual Report for 1932,
A1bany.
Bush, B. 1991. Personal communication March 5, 1991. New York State Department
of Health. Albany, New York.
Bush, B. and F.C. Lo. 1973. Thin-layer chromotography for quantitative
polychlorinated b1phenyl analysis. J. Chromatogr. 77: 377-388.
Bush, B. and L.A. Shane. 1989. Accumulation of polychlorobiphenyl congeners and
p,p'-DDE at environmental concentrations by corn and beans. Ecotoxicology and
Environmental Safety. 17: 38-46.
Bush, B., R.F. Seegal, and E. Fitzgerald. 1990. Human monitoring of PCB by
urine analysis. In Dioxin, Bayieutt, Germany.
Bush, B., R.W. Streeter and R.J. Sloan. 1989. Polychlorobiphenyl (PCB)
congeners in striped bass (Morone saxztilis) from marine and estuarine waters of
New York State determined by capillary gas chromatography. Arch. Environ
Contan. Toxicol. 19: 49-61.
R-7
-------
Bush, B., J. Snow, S. Connor, and R. Koblintz. 1985. Polychlorinated biphenyl
congeners (PCBs), p,p-DDE and hexachlorobenzene in human milk in three areas of
upstate New York. Arch. Environ. Contain. Toxicol. 14: 443-450.
Bush, G., L.A. Shane, L.R. Wilson, E.L. Barnard, and D. Barnes. 1986. Uptake
of polychlorobiphenyl congeners by purple loosestrife (Lythrum salicaria) on t^e
banks of the Hudson River. Arch. Environ. Contam. Toxicol. 15: 285-290.
Cairus, T. G. M. Doose, J. E. Froberg, R. A. Jacobson, and E. G. Siegward.
1986. Analytical Chemistry of PCBs. In PCBs and the Environment, Volume I (ed:
J. S. Waid), CRC Press, Inc., Boca Raton, Florida, pp. 1-45.
Cantelmo, F.R. 1978. The Ecology of Sublittoral Heiofauna in ^Shallow Marine
Embayment. Doctoral dissertation. The City University of New York, New York.
137pp.
Cantelmo, F.R. and C.H. Wahtola, Or. 1989. "Predicted impacts to the aquatic
biota in the area of the proposed waterfront development site, Stevens Institute,
Hoboken, N.J." Report for Hartz Mountain Industries, 28 pp.
Cantelmo, R.R., L. Vale, T. Kelly and B. Stewart. 1985. Importance of underpier
areas in the lower Hudson River for striped bass. Coastal Zone 84(2): 706-715.
Carpenter, B.H. 1987. "PCB Sediment Decontamination Processes - Selection for
Test and Evaluation." Research Triangle Institute.
Chen, M., C.S. Hong, B. Bush and G.-Y. Rhee. 1988. Anaerobic biodegradation of
polychlorinated bi phenyls by bacteria from Hudson River sediments. Ecotoxicology
and Environmental Safety 16: 95-105.
Chervin, M.B. 1977. The Assimilation of Particulate Organic Carbon by
Estuarine and Coastal Copepods. Doctoral dissertation. The City University of
New York, N.Y.
Clark, J.R. and S.E. Smith. 1971. Migratory fish studies of the Hudson Estuary,
p. 292-319. In Hudson River ecology: proceedings of a symposium: October, 1969
(Eds. G.P. Howells and G.J. Lauer) New York Univ. Institute of Environmental
Medicine. 473p.
Clark, Milton. Personal communication, March 8, 1991. USEPA, Region V.
Chicago, IL.
Clark, S.L. 1990. "Lurking on the Bottom: Heavy Metals in the Hudson-Raritan
Estuary." Environmental Defense Fund, New York.
Clarkeson and Clough Associates. 1970. "Pilot Investigation on Silt Deposit
Behind Fort Edward Dam." For Niagara Mohawk Power Corporation.
Claussen, J.H., General Electric Company. 1991a. Letter to Douglas J. Tomchuk
(USEPA Region II), dated March 8, 1991, regarding Hudson River PCB Site,
containing transmittal of GE 1990 sediment samples from selected locations.
R-8
-------
Claussen, J.H., General Electric Company. 1991b. Letter to Douglas J. Tomchuk
(USEPA Region II), dated March 29, 1991, regarding Hudson River PCB Site
Reassessment RI/FS Phase 1 Work Plan Comments.
Clough, Harbour & Associates. 1984. "Final Report: Wastewater Facilities Plan
for Washington County Sewer District No. 2." NYSDEC, Division of Construction
Management, Albany, NY.
Cochran, W.6. 1977. Sampling Techniques. 3rd Edition. John Wiley and Sons,
Inc., NY.
Code of Federal Regulations (CFR). 1990. 29 CFR 1910.1000.
Cohn, T. A. 1988. Adjusted maximum likelihood estimation of the moments of
lognormal populations for type I censored samples. US Geol. Surv. Open File Rep.
88-350.
Cooper, J.C., F.R. Cantelmo and C.E. Newton. 1988. Overview of the Hudson River
estuary. Amer. Fish. Soc. Monograph 4: 11-24.
Cordle, F., R. Locke, and J. Springer. 1982. Risk assessment in a federal
regulatory agency: an assessment of risk associated with human consumption of
some species of fish contaminated with polychlorlnated biphenyls (PCBs).
Environ. Hlth. Persp. 45: 171-182.
Cornell University, Department of Agricultural Economics. 1990. Dairy Fara
Business Summary. Northern Hudson Region, 1989.
Cowardin, L.M., V. Carter, F.C. Golet, and E.T. LaRoe. 1979. Classification of
wetlands and deep-water habitats of the United States, US Fish and Wildlife
Service. Washington D.C. FWS/0BS-79/31.
Cummins, K.W. 1979. The natural stream ecosystem. In The Ecology of Regulated
Streams (Eds. J.V. Ward and J.A. Stanford), pp. 7-24.
Cummins, K.W. 1974. Structure and function of stream ecosystems. Bioscience
24: 631-641.
Cummins, K.W. and M.J. Klug. 1979. Feeding ecology of stream invertebrates.
Ann. Rev. Ecol. Syst. 10:
Cummins, K.W., R.R. Costa, R.E. Rowe, G.A. Moshiri, R.M. ScanIon, and R.K.
Zajdel. 1969. Ecological energetics of a natural population of the predaceous
zooplankton Leotodora kindtii Focke (Cladocera). Oikos 20: 189-223.
Cushing, D.H. 1975. Marine ecology and fisheries. Cambridge University Press,
Cambridge. >
Dames and Moore. 1970. "Foundation Investigation and Design Consultations,
Proposed Hydroelectric Dam, Fort Edward New York." Report to Niagara Mohawk
Power Corporation, July 17, 1970.
R-9
-------
Darmer, K.I. 1987. "Overview of Hudson River Hydrology." Hudson River
Foundation, New York.
Davidorf, F.H. and Knupp, J.A. 1979. Epidemiology of ocular melanoma, incidence
and geographical relationship in Ohio (1967-1977). Ohio State Ned. J., 75, 561.
As cited in Silberhorn et al., 1990.
Deck, B.L., 1981, Nutrient Element Distributions in the Hudson Estuary. Ph.D.
Dissertation, Columbia University, New York, New York.
Dergosits, J. 1991. Personal communication. NYSDEC. Albany, NY.
Dethier, B.E. 1966. Precipitation in New York State, Cornell University
Agricultural Experiment Station, Bulletin 1009, 78 pp.
Dey, W.P. 1981. Mortality and growth of young-of-the-year striped bass 1n the
Hudson River estuary. Trans. Am. fish. Soc. 110(1): 151-157.
Dickhut, R.M., A.W. Andren, and D.E. Armstrong. 1986. Aqueous solubilities of six
polychlorinated biphenyl congeners at four temperatures. Environmental Science
and Technology 20: 807-810.
Dourson, Michael. Personal communication, March 6, 1991. USEPA, Environmental
Criteria Assessment Office, Office of Research and Development. Cincinnati, OH.
Dourson, M. and M. Clark. 1990. Fish consumption advisories: toward a unified,
scientifically credible aproach. Reg. Toxicol. Pharmacol 12: 161-178.
Draper, N.R. and H. Smith. 1966. Applied Regression Analysis. John Wiley &
Sons, Inc. NY
Dunne, T. and L. B. Leopold. 1978. Hater in Environmental Planning. S. H.
Freeman, San Francisco.
Ebasco Services Inc. 1990. "Draft Final Feasibility Study of Remedial Alterna-
tives for the Estuary and Lower Harbor/Bay, New Bedford Harbor, Massachusetts."
Volume II.
Eisler, R. 1986. Polychlorinated biphenyl hazards to fish, wildlife, and
invertebrates: a synoptic review. US Fish Uildl. Serv. Biol. Rep. 85(1.7).
Emmett, E.A., Maroni, M., Jefferys, J., Schmith, J., Levin, B.K., and Alveres,
A. 1988. "Studies of transformer repair workers exposed to PCBs: II. Results
of clinical laboratory investigations." Am. J. Ind. Ned. 14(1): 47-62.
Energy and Environmental Analysts, Inc. 1988. "Hudson River Center Aquatic
Environmental Study." Final Report prepared for New York City Public Development
Corporation.
EPA. 1988. "WASP4, a hydrodynamic and water quality model - model theory,
users' manual and programmer's guide." US Department of Commerce, National
R-10
-------
Technical Information Service, Springfield, VA.
Fabrizio, M.C., R.J. Sloan and O.F. O'Brien. 1991. Striped Bass Stocks and
Concentrations of Polychlorinated Biphenyl. (accepted for public, by Trans. AFS
1/8/91).
Farrell, M.A. 1933. VI Pollution Studies of the Upper Hudson Watershed. In
A Biological Survey of the Upper Hudson Watershed, pp. 208-215. N.Y. State
Conser. Dept., Supp. to 22nd Annual Report for 1932, Albany.
FDA (Food and Drug Administration). March, 1987. "Action Levels for Poisonous
or Deleterious Substances in Human Food and Animal Feed." Center for Food Safety
and Applied Nutrition, Industry Programs Branch.
Federal Register. May 22, 1989. Vol 54: 22062.
Federal Register. June 29, 1979. Vol 44(127): 38330-38340.
Feeney, S. 1991. Personal communication, Albany County Planning Department.
Fein, G.G., Jacobson, J.L., Jacobson, S.W., Schwartz, P.M., and Dowler, J.K.
1984. Prenatal exposure to Polychlorinated biphenyls: Effects on birth size and
gestational age. J. Pediatrics 105: 315-320.
Feldman, R. 1991. Personal communication, Biological Sciences, SUNY
Binghampton, Binghampton, NY.
FEMA. 1984. "Flood Insurance Study, Village of Fort Edward, New York." Federal
Emergency Management Agency, Community #361536.
FEMA. 1982. " Flood Insurance Study, Town of Fort Edward, New York." Federal
Emergency Management Agency, Community #360885.
FEMA. 1980. " Flood Insurance Study, Town of Fort Edward, New York." Federal
Emergency Management Agency, Community #360885.
Fenton, L. 1991. Personal communication, Saratoga County Planning Department.
Finch, R.G. 1925. The Story of the New York State Canals. Historical and
Commercial Information. Albany, NY.
Firda, G.D., R. Lumla and P. M. Burkฃ 1989. Water Resources Data New York,
Water Year 1988, Volume I. Eastern New York Excluding Long Island. USGS Water-
Data Report NLY-88-1. US Geological Survey, Albany, NY
Fischbeln, A. 1985. Liver function tests 1n workers with occupational exposure
to polychlorinated biphenyls (PCBs): comparison with Yusho and Yu-Cheng.
Environ. Health Perspect. 60: 145-150.
Fisher, S.G. 1977. Organic matter processing by a stream-segment ecosystem:
Fort River, Massachusetts, USA Int. Rev. Ges. HydrobioJ. 63: 701-727.
R-ll
-------
Fisher, S.G. and 6.E. Likens. 1973. Energy flow in Bear Brook, New Hampshire:
an integrative approach to stream ecosystem metabolism. Ecol. Monogr. 43:
421-439.
Fitzgerald, E.F., Weinstein, A.L., Youngblood, L.G., Standfast, S.J., and Melius,
O.M. 1989. Health effects three years after potential exposure to the toxic
contaminants of an electrical transformer fire." Arch. Environ. Health 44(4):
214-221.
Fitzgerald, E.F., Standfast, S.J., Youngblood, L.G., Melius, J.M., and Janerlch,
D.T. 1986. Assessing the health effects of potential exposure to PCBs, dioxins,
and furans from electrical transformer fires: the Binghamton Stat Office Building
medical surveillance program." Arch. Environ. Health 41(6): 368-376.
Flynn, G. L. 1990. Physiochemical determinants of skin absorption. In:
Principles of Route-to-Route Extrapolation for Risk Assessment (Eds: T.R.
Gerrity and C. J. Henry), Elsevier, New York, pp. 93-127.
Frakes, R.A. 1990. "Health-Based Water Quality Criteria for
2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCOD)." Maine Department of Human Services,
Bureau of Health.
Francis, H. 1991. Personal Communication to P. Chen from H. Francis, National
Lime Association.
Fredrick, S.W., R.L. Heffner, A.T. Packard, P.M. Eldridge, J.C. Eldridge, G.J.
Schumacher, K.L. Eichorn, J.H. Currie, J.N. Richards and O.C. Broody. 1976.
Notes on phytoplankton distribution in the Hudson River estuary. In Hudson River
Ecology, 4th Symp., Hudson River Environ. Soc., Bronx, N.Y.
Fries, G.F. 1982. Potential polychlorinated biphenyl residues in animal
products from application of contaminated sewage sludge to land. J. Environ.
Qual. 11(1): 14-19.
Fries, G.F. and G.S. Marrow. 1975. Retention and excretion of 2,3,7,8-
tetrachlorodibenzo-p-d1oxin in rats. J. Agric. Fd. Chen. 23: 265-269.
Fries, G.F., G.S. Marrow, Jr., and C.H. Gordon. 1973. Long-term studies of
residue retention and excretion by cows fed a polychlorinated biphenyl (Aroclor
1254). J. Agr. Food Chen. 21(1): 117-121.
Furukawa, K. 1982. Microbial degradation of polychlorinated blphenyls. In
Biodegradation and Detoxification of Environmental Pollutants (Ed. A.M.
Chakarabarty), Boca Raton: CRC Press.
Gahagan & Bryant Associates. 1982.. "Probing Report - Thompson Island Pool."
New York State Department of Environmental Conservation, Albany, NY.
Gahagan and Bryant Associates. 1982. PCB Dredging, Borings & Soundings Location
Survey for Thompson Island Pool (Map).
R-12
-------
Gailanl, J. et a/. 1991. The Transport of Suspended Solids in the Fox River
(manuscript draft).
Galloway, J.N., S.J. Eisemrich and B.C. Scott. 1980. "Toxic Substances in
Atmospheric Deposition: A Review and Assessment." National Atmospheric
Deposition Program, NC-141.
Galson Remediation Corporation. 1991. "Company History and Process
Description."
Galson Remediation Corporation. 1991. "Galson's APEG-PLUS Treatment System,
Equipment and Job Description."
Garvey, E.A. 1990. The Geochemistry of Inorganic Carbon in the Hudson Estuary.
Ph.D., Dissertation, Columbia University, New York, New York.
Gilbert, C.R. 1980. Genus Umbra, mudminnows. In Atlas of North American
Freshwater Fishes. N.C. State Museum of Natural History, Raleigh, pp. 129-130.
Gilbert, R. 0. 1987. Statistical Methods for Environmental Pollution
Monitoring. Van Nostrand Reinhold, New York.
Gladden, J., F.R. Cantelmo, J.M. Croom and R. Shapot. 1988. Evaluation of the
Hudson River ecosystem in relation to the dynamics of fish populations. American
Fisheries Society Monograph 4: 37-52.
Gladen, B.C., Rogan, W.J., Hardy, P., Thullen, J., Tingelstad, J., Tully, N.
1988. Development after exposure to polychlorinated biphenyls and
dichlorodiphenyl dichloroethene transplacentally and through human milk. J.
Pediatr. 113(6): 991-995.
Gray, H. 1978. Anatomy. Bounty Books, New York.
Greeley, J.R. 1937. Fishes of the area with annotated list. In "A Biological
Survey of the Lower Hudson Watershed," pp. 45-103. N. Y. State Conservation
Department, Supp. to 26th Annual Report for 1936, Albany.
Greeley, J.R. 1935. Annotated list of fishes occurring in the watershed. In
"A Biological Survey of the Mohawk-Hudson Watershed," pp. 88-101. N. Y. State
Conservation Department, Supp. to 24th Annual Report for 1934, Albany.
Greeley, J.R. and S.C. Bishop. 1933. II Fishes of the Upper Hudson Watershed
with Annotated List. In "A Biological Survey of the Upper Hudson Watershed," pp.
64-101. N.Y. State Conservation Dept., Supp. to 22nd Annual Report for 1932,
Albany.
Green, D. 1985. "Initial Report: Hudson River Sampling." (Internal Report to
DEC, June 10-12, 1985)
R13
-------
Gustavsson, P., Hogstedt, C., and Rappe, C. 1986. Short-term mortality and
cancer incidence in capacitor manufacturing workers exposed to polychlorinated
biphenyls (PCBs). Am. J. Ind. Med., 10, 341. As cited in Silberhorn et al.,
1990.
Hammerstrom, K.A. November 26, 1990. Dermal absorption of 2,3,7,8-TCDD, PCBi,
and PAHs contained in soil. Memorandum to L. Woodruff, Region X. Exposure
Assessment Group, Office of Health and Environmental Assessment, Office of
Research Development, US Environmental Protection Agency: Washington, DC.
Hammond, D.E. 1975. Dissolved Gases and Kinetic Processes in the Hudson River
Estuary. Ph.D. Dissertation, Columbia University, New York, New York.
Harger, J.R.E. 1979. A model for the determination of the action level for the
removal of curene contaminated soil. Memorandum to P.S. Cole, Executive
Director. Toxic Substance Control Commission, Department of Management and
Budget, State of Michigan: Lansing, MI.
Harza Engineering. 1990. "Ft. Edward Dam PCB Remnant Deposit Containment
Environmental Monitoring Program Baseline Studies, Report of 1989 Results,
August-December 1989." Report to General Electric Company, Fairfield, Conn.,
February 1990.
Hawker, D.W. and D.W. Connell. 1988. Octanol-water partition coefficients of
polychlorinated biphenyl congeners. Environ. Sci. Techno7. 22: 382-387.
Heinle, D.R. 1974. An alternative grazing hypothesis of the Patuxent estuary.
Ches. Sci. 15: 146-150.
Helgers, K. 1991. Personal communication, Rensselaer County Planning Office.
Helsel, D. R. and T. A. Cohn. 1988. Estimation of descriptive statistics for
multiply censored water quality data. Uater Resources Research 24(12): 1997-
2004.
Hester, F.E. and J.S. Dendy. 1962. A multiple-plate sampler for aquatic
macroinvertebrates. Trans. Am. Fish. Soc. 91: 420-421.
Hetling, L.J. 1976. Trends in wastewater loading, 1900 to 1976. Paper 14. In:
Hudson River Ecology, 4th Symposium Hudson River Environmental Society, Bronx,
NY.
Hetling, L.J., E.G. Horn, and T.J. Tofflemire. 1978. "Summary of Hudson River
PCB study results." NYSDEC, Tech. Paper No. 51, Albany, NY 88pp.
Hetling, L.J., T.J. Tofflemire, E.G. Horn, R. Thomas and R. Mt. Pleasant. 1979.
The Hudson River PCB Problem: Management Alternatives, pp.630-50. In Health
Effects of Halogenated Aromatic Hydrocarbons (Eds. UJ Nicholson and AJ Morre)
Annals NY Academy of Science, Vol.230.
R-14
-------
Horn, E. G. and R. J. Sloan. 1984. "PCB 1n Hudson River Striped Bass, 1984."
NYSDEC, Albany, NY
Horn, E.G., R.J. Sloan and L.C. Skinner. 1986. Insights from Contaminated Fish
in New York. Trans. 51st N.A. Wildlife & Nat. Res. Conf. Vol. 51, pp. 384-391
HSDB. (1990). Hazardous Substances Database: Polychlorinated biphenyls.
Huckins, J.H., T.R. Schwartz, J.D. Petty, and L.M. Smith. 1988. Determination,
fate, and potential significance of PCBs in fish and sediment samples with
emphasis on selected AHH-inducing congeners. Chemosphere. 17: 1995-2016.
Hulslzer, E.E. 1976. Zooplankton of lower Narragansett Bay. Ches. Sci. 17:
260-270.
Humphrey, H. 1986. The human population An ultimate receptor for aquatic
contaminants. Hydrobiologia 149: 75-80.
Hurst, B. personal communication. US EPA Superfund Technical Assistance
Hotline.
Hydroqual, Inc. 1991. Task 7.1: Assessment of Pollutant Loadings to NY/NJ
Harbor. USEPA, Region II, Marine and Wetlands Protection Branch.
Hydroscience. 1979. Analysis of the fate of PCB's in the ecosystem of the
Hudson River estuary. Prepared by Hydroscience, Inc., Uestwood, NY for NYS
Department of Environmental Conservation, Albany, NY.
Hynes, H.B.N. 1970. The Ecology of Running Waters. Liverpool Press, Liverpool,
555pp.
Ikeda, M., Kuratsune, M., Nakamura, Y., and Hirohata, T. 1986. A cohort study
on mortality of Yusho patients - a preliminary report. Fukuoka Acta Med., 78,
297. As cited in Silberhorn et al., 1990.
Iwata, Y., F.A. Gunther, and W.E. Westlake. 1974. Uptake of PCB (aroclor 1254)
from soil by carrots under field conditions. Bull. Environm. Contam. Toxicol.
11: 523-528.
Jensen, D.J., R.A. Hummel, N.H. Mahle, C.VI. Kocher, and H.S. Higgens. 1981. A
residue study of beef cattle consuming 2,3,7,8-tetrachlorodibenzo-p-diox1n
(TCDD). J. Agricul. Food Chew. 29: 265-268.
Jones. P.A. and R.J. Sloan. 1989. An in situ exposure vessel for
bloaccumulatlon studies with juvenile fish. Environmental Toxicology and
Chemistry 8: 151-155.
Jones, P.A., R.J. Sloan and M.P. Brown. 1989. PCB congeners to monitor with
caged juvenile fish in the upper Hudson River. Environmental Toxicology and
Chemistry 8: 793-803.
R15
-------
Kalmaz, E.V. and 6.D. Kalmaz. 1979. Transport, distribution and toxic effects
of polychlorinated biphenyls in ecosystems: review. Ecological Modelling. 6:
223-251.
Kenaga, E.E. 1980. Correlation of bioconcentration factors of chemicals in
aquatic and terrestrial organisms with their physical and chemical properties.
Environ. Sci. Technol. 14(5): 553-556.
Kim, J.C.S., E.S. Chao, M.P. Brown and R. Sloan. 1989. Pathology of brown
bullhead Ictalurus nebulosus from highly contaminated and relatively clean
sections of the Hudson River. Bull. Environ. Contam. Toxicol. 43: 144-150.
Kimbrough, R.D. 1987. Human health effetcs of polychlorinated biphenyls (PCBs)
and polybrominated biphenyls (PBBs). Ann. Rev. Pharmacol. Toxicol. 27: 87-111.
Kimbrough, R.D., Squire, R.A., Under, R.E., Strandberg, J.D., Montali, R.J., and
Burse, V.M. (1975). Induction of liver tumors in Sherman strain female rats by
polychlorinated biphenyl Aroclor 1260. J. Natl. Cancer Inst. 55: 1453-1459.
Kiviat, E. 1973. A freshwater tidal marsh on the Hudson Tivoli North Bay. In
Hudson River Ecology, 3rd symposium. Hudson River Environmental Society, New
Paltz, N.Y.
Klyshtorin. L.B. and A.A. Yarzkombek. 1975. Some aspects of the physiology of
the striped bass, Morone saxatilis. J. Ichthyol. 15(6): 985-989.
Koeman, J.H., H.C.W. Van Velzen-Blad, R. De Vries, and J.G. Vos. 1973. Effects
of PCB and DDE in cormorants and evaluation of PCB residues from an experimental
study. J. Reprod., Suppl. 19: 353-364.
Komura, S. and Simmons, D.B. 1967. River-bed degradation below dams. Journal
of the Hydraulic Division, ASCE 93(4): 1-13.
Kong, H.-L. and G.S. Sayler. 1983. Degradation and total mineralization of
monohalogenated biphenyls in natural sediment and mixed bacterial culture.
Applied and Environmental Microbiology 46: 666-672.
Krouse, J.S. 1968. Effects of dissolved oxygen, temperature and salinity on
survival of young striped bass, Roccus saxatilis (Ualbaum). M.S. Thesis, Univ.
Main, Orono, ME. 61p.
Kubiak, T.J. February 21, 1991. Personal communication. US Fish and Wildlife
Service, East Lansing, Michigan.
Kuratsune, M., Nakamura, Y., Ikeda, M., and Hirohata, T. 1987. Analysis of
death seen among patients with Yusho - a preliminary report. Chemosphere 16,
2085. As cited in Silberhorn et al., 1990.
Lane, G.A. 1970. "An Initial Fisheries Survey of the Hudson River from Lock No.
1 to Fort Edward." New York DEC, Bureau of Fisheries, Albany, New York. Typed
report. 18pp.
R16
-------
Lara, R. and W. Ernst. 1989. Interaction between polychlorinated biphenyls and
marine humic substances. Determination of association coefficients. Cheaosphere
19: 1655-1664.
Lawler, Matsuky and Skeller. 1979. "Upper Hudson River PCB Transport Modeling
Study." Report to NYSDEC, Dec. 1979. LMS, Pearl River, NY
Lawler, Matusky and Skeller. 1978. "Upper Hudson River PCB No Action
Alternative Study." Report to NYSDEC. LMS, Pearl River, NY
Lawler, Matusky & Skelly Engineers (LMS). 1983a. "1982-1983 Uestway Winter
Sampling Program, Vol. I, Trawl Data." Prepared for New York State Department
of Transportation.
Lawler, Matusky & Skelly Engineers (LMS). 1980. "Biological and Water Quality
Data Collected in the Hudson River Near the Proposed Westway Project During
1979-1980. Vol. I." Prepared for the New York State Department of
Transportation and System Design Concepts, Inc.
Lawton, R.W., Ross, M.R., and Feingold, J. 1986. Spirometric findings in
capacitor workers occupationally exposed to polychlorinated biphenyls (PCBs).
J. Occup. Med. 28(6): 453-456.
Lepow, M.L., L. Bruckman, M. Gillette, S. Markowitz, R. Rabino, and J. Kapish.
1975. Investigations into sources of lead in the environment of urban children.
Environ. Res. 10: 415-426.
Lettenmaier, D.P., E. R. Hooper, C. Wagoner and K. B. Faris. 1991. Trends in
stream quality in the continental United States, 1978-1987. Uater Resources
Research 27(3): 327-339.
Limburg, K.E. 1984. Environmental impact assessment of the PCB problem : A
review. Northeastern Environmental Science 3(3/4): 122-136
Limburg, K.E. 1986. PCBs in the Hudson. In the Hudson River Ecosystem.
Springer-Verlag, New York, pp. 83-130.
Limburg, K.E., M.A. Moran and W.H. McDowell. 1986. The Hudson River Ecosystem.
Springer-Verlag, New York, 331 pp.
Lindvall, M.L. and J.B. Low. 1980. Effects of DDE, TDE, and PCBs on shell
thickness of western grebe eggs, bear river migratory bird refuge, Utah: 1973-74.
Pesticides Monitoring Journal. 14: 108-111.
Linscombe, G., N. Kinler, and R. Aulerich. 1982. Mink. In Wild animals of
North America: biology, management, and economics (Eds. J. Chapman and G.
Feldhamer), John Hopkins Univ. Press, Baltimore, Maryland, pp. 629-643.
Liss, G.M. 1990. "Mortality and Cancer Morbidity Among Transformer
Manufacturing Workers." Ferranti-Packard Transformers LTD, St. Catharines, Health
Studies Service, Ontario Ministry of Labour. Draft.
R-17
-------
Little, W.C. and Mayer, P.G. 1972. "The role of sediment gradation on channel
armoring." School of Civil Engineering, Georgia Institute of Technology, Atlanta,
GA.
Long, E.R. and L.G. Morgan. 1990. The potential for biological effects of
sediment-sorbed contaminants tested in the national status and trends program.
NOAA Technical Memorandum NOS OMA 52.
Lowe, 0.1., P.R. Parrish, O.M. Patrick, Or., and 0. Forester. 1972. Effects of
the polychlorinated biphenyl aroclor 1254 on the american oyster, Crassostrea
virginlca. Mar. Biol. 17: 209-214.
MacLeod, W.O., L.S. Ramos, A.O. Friedman, D.G. Burrows, P.G. Prohaska, D.L.
Fisher and D.U. Brown. 1981. Analysis of Residual Chlorinated Hydrocarbons,
Aromatic Hydrocarbons and Related Compounds in Selected Sinks, Sources and Biota
of the New York Bight, NOAA Technical Memorandum OMPA-6, National Oceanic and
Atmospheric Administration, C.U.S. Dept. of Commerce.
Mahanti, H.K. 1975. A study on the effects of polychlorinated biphenyl (aroclor
1242) on an aquatic plant Spirodela oligorhiza (Kurz) Bull. Environ. Contan.
Toxicol. 14: 558.
Makarewicz, O.C. 1983. "Champlain Canal Fisheries Survey New York State Barge
Canal, Data Report to Malcolm Pirnie, Inc." 242 pp.
Malcolm Pirnie. 1976. "Preliminary Appraisal, Sediment Transport Relations,
Upper Hudson River." Engineering Report to NYDEC, Oune, 1976.
Malcolm Pirnie. 1975. "Investigation of Conditions Associated with the Removal
of Fort Edward Oam, Fort Edward, New York." Engineering Report to NYOEC, Feb.
1975.
Malcolm Pirnie, Inc. 1986. "Draft Supplement To: Final Environmental Impact
Statement on the Hudson River PCB Reclamation Demonstration Project, October
1982." Report to USEPA Region II and NYSDEC, Albany, NY.
Malcolm Pirnie Inc. 1985. "Hudson River PCB Reclamation/Demonstration Project,
Evaluation of Alternative PCB Sites/Disposal Technologies."
Malcolm Pirnie, Inc. 1984a. "Westway Fisheries Study, December 1983 to April
1984, New Oersey Marine Sciences Consortium Data Presentation." Prepared for US
Army Corps of Engineers, New York District.
Malcolm Pirnie, Inc. 1984a. "Study of Impacts of Hudson River Flow Regulations,
Draft Report." Hudson River - Black River Regulating District. Albany, NY.
Malcolm Pirnie, Inc. 1984b. "Westway Fisheries Study Data Evaluation Workshop
April 16-18, 1984. Final Summary Report." Prepared for US Army Corps of
Engineers, New York District.
R-18
-------
Malcolm Pirnie, Inc. 1984b. "New York State Barge Canal Environmental Report,
Maintenance Dredging Program, 1985-1995, Volumes 1 and 2." NYSDOT. Albany, NY.
Malcolm Pirnie, Inc. 1984c. "1983-1984 Westway Fisheries Study. Summary
Report." Prepared for US Army Corps of Engineers, New York District.
Malcolm Pirnie, Inc. 1983. "Data Presentation. 1982-83 Mitigation Studies,
Upriver Studies, and Special Studies, LMS: and 1982-1983 Hudson River Fishery
Habitat Study, Bayonne to Piermont, NJDEP." Prepared for the US Army Corps of
Engineers, New York district, NY.
Malcolm Pirnie, Inc. 1982. "Hudson River Estuary Fish Habitat Study." Prepared
for the US Army Corps of Engineers, New York district, NY.
Mann, K.H. 1975. Patterns of energy flow. In River Ecology (Ed. B.A. Whitton),
pp. 248-263.
Mansueti, R.J. 1961. Age, growth, and movements of the striped bass, Roccus
saxatilis, taken in size selective fishing gear in Maryland. Chesapeake Sci.
1-2(2): 9-36.
Mayer, F.L., P.M. Mehrle, and H.O. Sanders. 1977. Residue dynamics and
biological effects of polychlorinated biphenyls in aquatic organisms. Arch.
Environ. Contam. Toxicol. 5: 501-511.
McCreary, S. T. 1988. "Managing PCBs in the Hudson/Raritan Estuary and the New
York Bight System." The Negotiated Single Text of the New York Academy of
Sciences, New York Bight Initiative.
McDermott, J.M., R. Unterman, M.J. Brennan, R.E. Brooks, D.P. Mobley, C.C.
Schwartz and D.K. Dietrich. Two Strategies for PCB Soil Remediation:
Biodegradation and Surfactant Extraction. Environmental Progress 8(1): 46-51.
McFadden, J.T., Texas Instruments, Lawler, Matusky & Skelly. 1978. "Influence
of the proposed Cornwall pumped storage project and steam electric generating
plants on the Hudson River estuary with emphasis on striped bass and other fish
populations, revised." Report to Consolidated Edison Company of New York, Inc.
McGill, R., J. Tukey and W.A. Larsen. 1978. Variation of box plots. American
Statistician 32: 16.
McGovern, B. 1991. Personal Communication to P. Chen from B. McGovern, C. F.
Systems.
Mclntire, C.D. and J.A. Colby. 1978. A hierarchical model of lotic ecosystems.
Ecol. Honogr. 48: 167-190.
McLellan, T., R. Havis, D. Hayes, G. Raymond. 1989. "Field Studies of Sediment
Resuspension Characteristics of Selected Dredges." U. S. Army Corps of Engineers
Waterways Experiment Station, Vicksburg, Mississippi.
R-19
-------
Metcalf & Eddy. 1990. "Final Report, Waterford Drinking Water Supply
Evaluation. (Hudson River PCB Remnant Site Project - Task 3)." Report prepared
for NYSDEC, Albany, NY
Miranda, C.L., M.C. Henderson, J.L. Wang, H.S. Nakaue, and O.R. Buhler. 1987.
Effects of polychlorinated biphenyls on porphyrin synthesis and cytochrome P-4S0-
dependent monooxygenases in small intestine and liver of japanese quail.
Toxicol. Environ. Health. 20: 27-35.
Mondello, F.J. 1989. Cloning and Expression in Escherichia coli of Pseudononas
Strain LB400 Genes Encoding Polychlorinated Biphenyl Degradation. Journal of
Bacteriology 171(3): 1725-1732.
Mondello, F.J. 1987. Molecular Cloning of the Genes for ' Bacterial PCB
Biodegradation. In "Research and Development Program for the. Destruction of
PCBs, Sixth Progress Report." Schenectady: General Electric Company Corporate
Research and Development Center.
Moran, M.A., and K.E. Limburg. 1986. The Hudson River Ecosystem. In The Hudson
River Ecosystem. Springer-Verlag, New York, pp. 6-39.
Montz, W.E., W.C. Card, and R.L. Kirkpatrick. 1982. Effects of polychlorinated
biphenyls and nutritional restriction on barbituate-induced sleeping times and
selected blood characteristics in raccoons (Procyon lotor). Bull. Environ.
Contain. Toxicol. 28: 578-583.
Moza, P.N., I. Weisgerber, and W. Klein. 1976. Fate of 2,2'-d1chlorobiphenyl-
"C in carrots, sugar beets and soil under outdoor conditions. J. Agric. Food
Chew. 24: 881-885.
Mueller, J.A. T.A. Gerrish and M.C. Casey. 1982. Contaminant Inputs to the
Hudson Raritan Estuary, NOAA Technical Memorandum OMPA-21, National Oceanic and
Atmospheric Administration, US Department of Commerce.
Muenscher, W.C. 1933. VII aquatic vegetation of the upper Hudson watershed.
In "A Biological Survey of the Upper Hudson Watershed," pp. 216-238. N.Y. State
Conservation Dept., Supp. to 22nd Annual Report for 1932, Albany.
Mukerjee, Dubdas (personal communication). 1991. USEPA, Environmental Criteria
Assessment Office, Office of Research and Development. Cincinnati, OH.
Nadeau, R. J. and R. P. Davis. 1974. "Investigation of Polychlorinated
Biphenyls 1n the Hudson River: Hudson Falls-Fort Edward Area." USEPA Region II
Report..
National Academy of Sciences (NAS). 1980. Drinking Hater and Health.
Washington DC: National Academy Press.
National Institute for Occupational Safety and Health (NIOSH). 1977. "Criteria
for a Recommended Standard Occupational Exposure to Polychlorinated Biphenyls
(PCBs)." Cincinnati, OH. NIOSH-77-225.
R-20
-------
National Research Council. 1989. "Contaminated Marine Sediments - Assessment
and Remediation." Committee on Contaminated Marine Sediments, Marine Board.
National Academy Press. Washington, D.C.
Nau-Ritter, G.M., C.F. Wurster, and R.G. Rowland. 1982. Polychlorinated
biphenyl (PCB) desorbed from clay particles inhibit photosynthesis by natural
phytoplankton communities. Environmental Pollution. 28: 177-182.
New Jersey Marine Sciences Consortium. 1988. "Fisheries Study of the Proposed
Lincoln Harbor Yacht Club, Weehawken, N.J." Data Reports submitted to Stephen
Sloan Marine Corporation.
New Jersey Marine Sciences Consortium. 1988. "The Hudson-Raritan: State of the
Estuary. Appendix (Volume 1, Part 2) of Water Quality of New Jersey Coastal
Waters."
New Jersey Marine Sciences Consortium. 1987. "The Hudson-Rariton: State of the
Estuary. Summary (Volume 1, Part 1) of Water Quality of New Jersey Coastal
Waters."
New Jersey Marine Sciences Consortium. 1984. "Summary Report, Westway Fisheries
Study. 7 vols." For: Army Corps of Engineers, NYD.
Newell, A.J., D.W. Johnson, and L.K. Allen. 1987. "Niagara River Biota
Contamination Project: Fish Flesh Criteria for Piscivorous Wildlife." DEC
Technical Report 87-3, Bureau of Environmental Protection, Div. of F1sh and
Wildlife. New York.
Nicholson, J. R. 1991. Personal communication to P. Chen from J. R. Nicholson,
Zimpro-Passavant.
Niering, W.A. 1985. The Audobon Society Nature Guides. Wetlands. Alfred A.
Knopf, Inc. New York.
Nies, (.., P.J. Anid and T.M. Vogel. 1990. Sequential anaerobic-aerobic
biodegradation of PCBs. In Research and Development Program for the Destruction
of PCBs, Ninth Progress Report. Schenectady: General Electric Company Corporate
Research and Development Center.
Niimi, A.J. 1983. Biological and toxicological effects of environmental
contaminants in fish and their eggs. Can. J. Fish. Aquat. Sci. 40: 306-312.
Niimi, A.J. and B.G. Oliver. 1989. Distribution of polychlorinated biphenyl
congeners and other halocarbons in whole fish and muscle amoung lake Ontario
salmonids. Environ. Sci. Technol. 23: 83-88.
N11m1, A.J. and B.G. Oliver. 1983. Biological half-lives of polychlorinated
biphenyl (PCB) congeners 1n whole fish and muscle of rainbow trout (Saloo
gairdneri). Can. J. Fish. Aquat. Sci. 40: 1388-1394
R21
-------
NJDEP. 1991. Discharge Monitoring Data for Dischargers to the Hudson River.
Computer Printout.
Norback, D.H. and Weltman, R.H. 1985. Polychlorlnated biphenyl Induction of
hepatocellular carcinoma in the Sprague-Dawley rat. Environ. Health Perspect.
60: 97-105.
Novak, M.A., R.W. Bode and L.E. Abele. 1987. Rapid Biological Stream
Assessment, Battenkill River, Shushan to CI arks Hills, N.Y. 19 pp.
Novak, M.A., A.A. Reilly, B. Bush, and L. Shane. 1990. In-situ determination
of PCB congener-specific first order absorption/desorption rate using Chironomus
tentans larvae (Insecta: Diptera: Chironomidae). Wat. Res. 24(3): 321-327.
Novak, M.A., A.A. Reilly and S.J. Jackling. 1988. Long-term monitoring of
polychlorinated biphenyls in the Hudson River (New York) using caddisfly and
other macroinvertebrates. Arch. Environ. Contain. Toxicol. 17: 699-711.
NUS Corporation. 1984. "Feasibility Study: Hudson River PCBs Site, New York.
Volume 1." Pittsburgh, PA
NYCDEP. 1987. Bureau of Wastewater Treatment, Water Quality Monitoring Unit.
"New York Harbor Water Quality Survey, 1987."
NYCDEP. 1991. Bureau of Wastewater Treatment, Water Quality Monitoring Unit.
Sediment Harbor Survey Data, 1984-1989. Computer Files.
NYSDEC (New York State Department of Environmental Conservation). 1991. Bureau
of Wastewater Facilities Operation. SPDES Permits in response to TAMS' Freedom
of Information Act Request, letter dated 3/21/91.
NYSDEC (New York State Department of Environmental Conservation). 1991. Bureau
of Spill Prevention and Response. Spills Information Data Base, PCB Spills in
the Hudson River, 4/1/86 - 2/28/91. Computer Files.
NYSDEC (New York State Department of Environmental Conservation). 1991. Bureau
of Wastewater Facilities Operation, Source Surveillance Section. PCB Aroclors
Data in Upper Hudson River Basin, DMRs for period 1/1/85 - 12/31/90. Computer
Printout.
NYSDEC (New York State Department of Environmental Conservation). 1990.
Division of Water. Amendments to 6NYCRR Parts 940-941 Classifications and
Standards of Quality and Purity for Surface Waters in the Upper Hudson River
Drainage Basin. Albany, NY.
NYSDEC (New York State Department of Environmental Conservation). 1990.
Division of Hazardous Waste Remediation. "Inactive Hazardous Waste Disposal
Sites in New York State. Volumes 4 and 5." Albany, NY.
R-22
-------
NYSDEC (New York State Department of Environmental Conservation). 1990.
Division of Water. 1990. New York State Water Quality. Section 305(b) of the
Federal Clean Water Act Amendments of 1977. Bureau of Monitoring and Assessment.
Albany, NY.
NYSDEC (New York State Department of Environmental Conservation). 1990.
Division of Water. "Water Quality Standards and Guidance Values, Technical and
Operational Guidance Series 1.1.1 (TOGS)." Albany, NY.
NYSDEC (New York State Department of Environmental Conservation). 1990. "New
York Statewide Angler Survey, 1988." New York Department of Environmental
Conservation, Division of Fish and Wildlife, Bureau of Fisheries. Albany, New
York.
NYSDEC (New York State Department of Environmental Conservation). 1990. "Hudson
River Estuary Management Plan and Generic Environmental Impact Statement, Draft."
NYSDEC (New York State Department of Environmental Conservation). 1990.
"Waterford Drinking Water Supply Evaluation: Hudson River PCB Remnant Site
Project - Task 3."
NYSDEC (New York State Department of Environmental Conservation). 1990.
Oivision of Water. "Biennial Report, Rotating Intensive Basin Studies (RIBS),
Water Quality Assessment Program, 1987-88." Albany, NY.
NYSDEC (New York State Department of Environmental Conservation). 1989. "New
York State Air Guide: Guidelines for the Control of Toxic Ambient Air
Contaminants." Division of Air Resources.
NYSDEC (New York State Department of Environmental Conservation). 1988.
Division of Water. "Priority Water Problems List, 1988." Albany, NY.
NYSDEC (New York State Department of Environmental Conservation). 1987.
Endangered, threatened, and special concern species of New York state. Official
List. Albany, New York.
NYSDEC (New York State Department of Environmental Conservation). 1986.
Division of Water. Water Quality Regulations, Surface Water and Groundwater
Classifications and Standards, New York State Codes, Rules and Regulations, Title
6, Chapter X, Parts 700-705.
NYSDEC (New York State Department of Environmental Conservation). 1985. "New
York State (DEC) Water Quality Criteria." Bureau of Environmental Protection.
New York.
NYSDEC (New York State Department of Environmental Conservation). 1982-4. "New
York State Toxics Air Monitoring Report." Albany, NY.
NYSDEC (New York State Department of Environmental Conservation). 1979. A
comparison of the 1972/73 and 1977 results, Hudson River. Tech. memorandum
(unpublished).
R-23
-------
NYSDOH (New York State Department of Health). 1991. Bureau of Public Water
Supply Protection. "Public Water Systems - Hudson River Source, Revised."
NYSDOH (New York State Department of Health). 1986. "Biomonitoring of PCBs in
the Hudson River: Final Report." Albany, New York.
NYSDOT. 1991. "Dredging - Region 1 DOT, Champlain Canal: Fort Edward to
Waterford (1971-1990)." Albany, NY.
NYSDOT. 1987. "Recreational Hap and Guide of the New York State Canals."
Albany, NY.
NYSDOT. 1981. "New York State Barge Canal System Fact Sheets." Waterways
Maintenance Subdivision. Albany, NY.
O'Brien and Gere. 1978. "PCB Analysis Final Report, July, 1978."
O'Connor, J.M. and J.C. Pizza. 1984. Eco-kinet1c model for the accumulation of
PCB in marine fishes.- Pages 285-303. In Management of Bottom Sediments
Containing Toxic Substances. (Ed. T.R. Patin). US Army Corps Engin., Wat. Res.
Supp. Cen.
Oak Ridge National Laboratory. 1989. "The Installation Restoration Program
Toxicology Guide, Volumes III and IV." Oak Ridge National Laboratory: Oak Ridge,
TN.
Odum, H.T. 1957. Trophic structure and productivity of Silver Springs, Florida.
Ecol. Monogr. 27: 55-112.
Oliver, B.G. and A.J. Niimi. 1988. Trophodynamic analysis of polychlorinated
biphenyl congeners and other chlorinated hydrocarbons in the Lake Ontario
ecosystem. Environ. Sci. Tech. 22: 388-397.
Olsen, C.R. 1979. Radionuclides, Sedimentation and the Accumulation of
Pollutants in the Hudson Estuary. Ph.D. Dissertation, Columbia University, New
York, New York.
Osman, A.M., and Thorne, C.R. 1988. Rlverbank stability analysis. I: theory.
Journal of Hydraulic Engineering, ASCE 114(2): 134-150.
Pal, D., J.B. Weber, and M.R. Overcash. 1980. Fate of polychlorinated blphenyls
(PCBs) 1n soil-plant systems. Residue Reviews 74: 45-98.
Parrish, P.R. 1974. Aroclor 1254, DDT, DDD, and dleldren: accumulation and loss
by amerlcan oysters (Crassostrea virginica) exposed continuously for 56 weeks.
Proc. Nat. Shellfish Assoc. 64: 7
Pate, V.S.L. 1933. IV studies on fish food in selected areas. In "A
Biological Survey of the
Upper Hudson Watershed." pp. 130-156. N.Y. State Conservation Dept., Supp. to
22nd Annual Report for 1932, Albany.
R-24
-------
PATHCO. 1990. "Pathology Working Group - 2,3,7,8-Tetrachlorodibenzo-p-Dioxin
1n Sprague Dawley Rats." Report prepared by Robert M. Sauer, V.M.D., Pathology
Working Group Chairperson, and submitted to the Maine Science Advisory Panel,
March 13, 1990.
Pedzy, E. 1991. Personal communication to P. Chen from E. Pedzy. Ozonicf
Technology, Inc.
Pflieger, W.L. 1978. "The Fishes of Missouri." Missouri Department of
Conservation. Columbia, Missouri.
Plafkin, J.L., M.T. Barbour, K.D. Porter, S.K. Gross and R.M. Hughes. 1989.
"Rapid Bioassessment Protocols for Use in Streams and Rivers: Benthic
Macroinvertebrates and Fish." US EPA/444/4-89/001.
Platonow, N. and C. Karstad. 1973. Dietary effects of polychorinated biphenyls
on mink. Can. J. Comp. Med. 37: 391-400.
Poiger, H. and C. Schlatter. 1980. Influence of solvents and adsorbents on
dermal and intestinal absorption of TCDD. Fd. Cosnet. Toxicol. 18: 477-481.
Porter, J.W. (January 27, 1989). Memorandum to regional administrators, region
I-X, regarding interim final guidance on soil ingestion rates. USEPA, Office of
Solid Waste and Emergency Response.
Prandle, D. 1981. Salinity intrusions in estuaries. J. Phys. Oceanography, II
1311-1324.
Preddice, T.L. and R.W. Karcher. 1990. "Bridge Maintenance and its Effects on
Aquatic Life in the Champlain Canal at Fort Miller, New York." N.Y. DEC, Div.
of Fish and Wildlife, Bureau of Environmental Protection, Tech. Report 90-2
(BEP), 60 pp.
Pritchard, D.W. 1969. Dispersion and flushing of pollutants in estuaries. Jnl.
Hydraul. Div., Am. Soc. Civil Eng. 95: NoHYl, 115-124.
Pritchard, D.W. 1955. Estuarine circulation patterns. Proc. As. Soc. Civ. Eng.
81: 717/1 - 717/11
Que Hee, S.S., B. Peace, C.S. Clark, J.R. Boyle, R.L. Bornschein, and P.B.
Hammond. 1985. Evolution of efficient methods to sample lead sources, such as
house dust and hand dust, in the homes of children. Environ. Res. 38: 77-95.
Quensen, J.F., S.A. Boyd, and J.M. Tieje. 1990. Dechlorination of four
commercial
polychlorinated biphenyl mixtures (aroclors) by anaerobic microorganisms from
sediments. Applied and Environmental Microbiology 56: 2360-2369.
Quensen, J.F., J.M. Tieje and S.A. Boyd. 1988. Reductive dechlorination of
polychlorinated biphenyls by anaerobic microorganisms from sediment. Science
242: 752-754.
R-25
-------
Rand, G.N. and S.R. Petrocelli (eds.). 1985. Fundamentals of Aquatic
Toxicology. Hemisphere Publishing Company. New York.
Raney, E.C. 1954. The striped bass in New York waters. New York State
Conservationist (Feb.-Mar.): 14-16.
Raney, E.C. 1952. The life history of the striped bass, Roccus saxatiliV
(Walbaum). Bull Brigharo Oceanog. Coll. 14(1): 5-97.
Rensselaer County Planning Office. 1990. Rensselaer County Master Plan.
Repsys, A.J., R.L. Applegate, and D.C. Hales. 1976. Food and food selectivity
of the black bullhead, Ictalurus melas, in Lake Poinsett, South Dakota. J. Fish.
Res. Board Can. 33: 768-775.
Research Triangle Institute. 1987. See Carpenter.
Rhee, G.-Y. and B. Bush. 1990. Dechlorination and biodegradation of chlorinated
biphenyls in anaerobic sediments. In Biological Remediation of Contaminated
Sediments with Special Emphasis on the Great Lakes, Report of a Workshop (Eds.
C.T. Jafvert and J.E. Rogers). Manitowoc, WI: 17-19 July, 1990.
Rhee, G.-Y., B. Bush, M.P. Brown, M. Kane and L. Shane. 1989. Anaerobic
biodegradation of polychlorinated biphenyls in Hudson River sediments in clay
encapsulation. Water Research 23(8): 957-964.
RIBS Section of 305(b) 1987-1988 Rotating Intensive Basin Study (RIBS) Report
(Preliminary data, NYSDEC Report).
Ringer, R. 1983. Toxicology of PCBs in mink and ferrets. In PCBs: Human and
Environmental Hazards (Eds. F. D'ltri and M. Kamrin). Butterworth Pub., Woburn,
Ma.
Ristich, S.S., M. Crandall, and J. Fortier. 1977. Benthic and epibenthic
macroinvertebrates of the Hudson River. I. Distribution, natural history and
community structure. Est. Coastal Mar. Sci. 5:255-266.
Roels, H.A., J.P. Buchet, R.R. Lauwerys, P. Bruaux, F. Claeys-Thoreau, A.
Lafontaine, and G. Verduyn. 1980. Exposure to lead by the oral and pulmonary
routes of children living in the vicinity of a primary lead smelter. Environ.
Res. 22: 81-94.
Roesch, S. E., L. J. Clark and M. M. Bray. 1979. "User's Manual for the Dynamic
(Potomac) Estuary Model." US Environmental Protection Agency, Annapolis, MD.
EPA-903/9-79-001.
Rogan, W.J., B.C. Gladen, J.D. McKinney, N. Carreras, P. Hardy, J. Thullen. J.
Tingelstad, and M. Tully. 1986. Polychlorinated biphenyls (PCBs) and
dichlorodiphenyl dichlroethene (DDE) in human milk: effects of maternal factors
and previous lactation. Am. J. Public Health 76(2): 172-177.
R-26
-------
Rogan, W.J., B.C. Gladen, J.D. McKinney, N. Carreras, P. Hardy, J. Thullen, J.
Unglestad, M. Tully. 1986. Neonatal effects of transplacental exposure to PCBs
and DDE. J. Pediatr. 109(2): 335-341.
Rogers, Bob. 1991. Personal communication, USGS Albany, NY
Rogers, Bob. 1991. Personal communication to Jonathan Butcher from Bob Rogers,
USGS Albany, NY
Russell-Hunter, W.D. 1970. Aquatic Productivity. Macmillan Pub. Co., Inc.,
N.Y. 306pp.
Ryan, Dan. 1991. Personal communication to Michael Spera from Dan Ryan, NYSDOT
Region 1, Albany, NY
Ryan, D. 1991. Personal communication. NYSDOT, Region 1, Waterways Maintenance
Subdivision. Albany, NY.
Safe, S. 1990. Polychlorinated biphenyls (PCBs), dibenzo-p-diox1ns (PCDDs),
dibenzofurans (PCDFs), and related compounds: environmental and mechanistic
considerations which support the development of toxic equivalency factors (TEFs).
Critical Reviews in Toxicology. 21: 51-88.
Safe, H.S. 1984. Microbial degradation of polychlorinated blphenyls.
Microbiol. Ser. 13: 361-369.
Sanders, J.E. 1989. PCB-pollution problem in the Upper Hudson River: Froa
environmental disaster to 'environmental gridlock.' Northeastern Env. Sci. 8(1):
1-86.
Sanders, H.O. and J.H. Chandler. 1972. Biological magnification of a
polychlorinated biphenyl (aroclor 1254) from water by aquatic invertebrates.
Bull. Environ. Contam. Toxicol. 7: 257-263.
Sanders, O.T. and R.T. Kirkpatrick. 1977. Reproductive characteristics and
cotlcoid levels of female white-footed mice fed ad libitum and restricted diets
containing a polychlorinated biphenyl. Environ. Res. 13: 358-363.
Saratoga County Planning Board. 1978. "Land Use and Development Plan, Saratoga
County, New York."
Saratoga Economic Development Corporation. 1990. "Industrial Directory."
Saunders, G.W. 1969. Some aspects of feeding 1n zooplankton. In
"Eutrophlcatlon: Causes, Consequences, Correctives." Nat. Acad, of Sci., Wash.,
D.C. pp. 556-573.
Sawhney, B.L. and L. Hankln. 1984. Plant contamination from amended soils. J.
Food Prot. 47: 232-236.
R-27
-------
Sayler, G.S., M. Shon, and R.R. Colwell. 1977. Growth of an estuarine
Pseudomonas sp. on polychlorinated biphenyl. Microbial Ecology. 3: 241-255.
Schaeffer, E.H., Greim H., and Goessner W. 1984. Pathology of chronic
polychlorinated biphenyl (PCB) feeding in rats. Toxicol. Appl. Pharmacol. 75:
278-288.
Schaffer, S.A. 1978. Concentrations of Organic Carbon and Protein in the Hudson
Estuary Near Indian Point. Master's thesis. New York University, New York.
Schaum, J. 1984. "Risk Analysis of TCDD Contaminated Soil." Office of Health
and Environmental Assessment, USEPA:Washington, D.C. EPA-600/8-84-031.
Scheuplein, R. J. 1977. Permeability of skin. In: Handbook of Physiology.
Section 9 Reactions to Environmental Agents (Eds: D. H. K. Lee, H. L. Falk, S.
D. Murphy, and S. R. Geiger), American Physiological Society, Bethesda, MO, pp.
299-322.
Schierbaum, D., D. Benson, L.W. DeGraff, and D.P. Foley. 1959. Waterfowl
banding in New York. New York Fish and Game Journal. 6: No. 1.
Schroeder, R. A. and C. R. Barnes. 1983a. "Trends in Polychlorinated Biphenyl
Concentrations in Hudson River Water Five Years after Elimination of Point
Sources." USGS Water-Resources Investigations Report 83-4206. USGS, Albany, NY
Schroeder, R. A. and C. R. Barnes. 1983b. "Polychlorinated Biphenyl Concentra-
tions in Hudson River Water and Treated Drinking Water at Waterford, New York."
USGS Water-Resources Investigations Report 83-4188. USGS, Albany, NY
Schulz, D.E., G. Petrick, and J.C. Duinker. 1989. Complete characterization of
polychlorinated biphenyl congeners 1n commercial Aroclor and Clophen mixtures by
multidimensional gas chromatography-electron capture detection. Environmental
Science i Technology 23: 852-857.
Schwartz, P.M., S.W. Jacobson, G. Fein, J.L. Jacobson, and H.A. Price. 1983.
Lake Michigan fish consumption as a source of polychlorinated biphenyls In human
cord serum, maternal serum, and milk. Am. J. Public Health 73(3): 293-296.
Scott, R.C., P.H. Dugard, J.D. Ramsey, and C. Rhodes. 1987. In vitro absorption
of some o-phthalate diesters through human and rat skin. Environ. Health
Perspec. 74: 223-227.
Sedell, J.R., R.J. Triska, J.D. Hall, N.H. Anderson and J.H. Lyford. 1973.
Sources and fates organic inputs in coniferous forest streams. Contrib. Con.
B1ome No. 66, IBP, Oregon State Univ., CorvalUs pp. 1-23.
Shane, L.A., and B. Bush. 1989. Accumulation of polychlorobiphenyl congeners
and p,p'-DDE at environmental concentrations by corn and beans. Ecotox. and
Environ. Safety 17: 38-46.
R-28
-------
Shen, T.T. . 1982. Air quality assessment for land disposal of wastes.
Environmental Management 6: 4.
Sheppard, D.J. 1976. "Valuation of the Hudson River Fisheries Resources: Past,
Present and Future." A Report Prepared by J. Douglas Sheppard, Bur. of
Fisheries, N.Y. DEC, April 1976, 51 pp.
Shiaris, H.P. and Sayler, G.S. 1982. Biotransformation of PCB by natural
assembledges of freshwater microorganisms. Environmental Science and Technology
16(6): 367-369.
Shubat, P. 1991. Letter to Milt Clark, USEPA, Chicago, II. dated January 28,
1991. Minnesota Department of Health Division of Environmental Health.
Minneapolis NM.
Shupp, B.D. 1987. Transcript of Proceedings at the Washington County Office
Building, Fort Edward, New York, State of New York Industrial Hazardous Waste
Facility Siting Board and the Department of Environmental Conservation, DEC
Project No. UPA 50-86-0024, June 30, 1987-July 1, 1987.
Shupp, B.D. 1976. Prepared Testimony of Bruce D. Shupp, Chief, Bur. of
Fisheries, NYS DEC, State of New York, DEC and Industrial Hazardous Waste
Facility Siting Board. In the Matter of the Application of the PCB Project Group
for Approvals of the Hudson River PCB Reclamation/Demonstration Project.
Shupp, B.D. 1975. "Report of PCB fish sample collections, Upper Hudson River."
New York DEC, Bureau of Fisheries. Mimeo report 33pp.
Silberhorn, E.M., Glavert, H.P., and Robertson, L.W. 1990. Carcinogenicity of
polyhalogenated biphenyls: PCBs and PBBs. Crit. Rev. Tox. 20(6): 439-496.
Simpson, K.W. 1986. "Final Report Blomonitoring of PCBs 1n the Hudson River
I. Results of Long-Term Monitoring Using Caddisfly (Insecta: Trichoptera:
Hydropsychidae) Larvae and Multlplate Residues II. Development of Field Protocol
for Monitoring PCB Uptake bu Caged Live Chlronomus tentans (Insecta: Dlptera:
Chironomidae) Larvae During Dredging Operations." Report prepared by M.A. Novak
and A.A. Reilly, N.Y.S. Department of Health 99pp. plus appendices.
Simpson, K.W. 1976. "A Water Quality Evaluation of the Hudson River, Based on
the Collection and Analysis of Macrolnvertebrate Communities." Paper No. 24, 47
pp. Hudson River Ecology, 4th Symposium, 1976.
Simpson, K.W. and R.W. Bode. 1985. "Rapid Biological Stream Assessment Hossic
River from Adams, Massachusetts to Mouth." (DEC Tech. Report dated November 6,
1985)
Simpson, K.W., T.B. Lyons and S.P. Allen. 1974. "Macroinvertebrate Survey of
the Upper Hudson River, New York, 1972." State DOH, Envir. Rep. No. 2, October
1974, 52 pp.
R-29
-------
Simpson, K.W., M.A. Novak, and A.A. Rei11y- 1986. "Final Report, Biomonitoring
of PCBs 1n the Hudson River." NYSDOH.
Simpson, K.W., R.W. Bode, J.P. Fagnani and D.M. DeNicola. 1984. "The Freshwater
Macrobenthos of the Main Channel, Hudson River, Part B, Biology, Taxonomy and
Distribution of Resident Macrobenthic Species." Final Rep. to Hudson River
Foundation, 203 pp.
Simpson, K.W., J.P. Fagnani, D.M. DeNicola and R.W. Bode. 1985a. "The
Freshwater Macrobenthos of the Main Channel, Hudson River. Part A General Study
Description and Results, Including a Discussion of Organism-Substrate
Relationships." Final Report to Hudson River Foundation, 69 pp.
Simpson, K.U., J.P. Fagnani, D.M. DeNicola and R.W. Bode. 1985b. Widespread
distribution of some estuarine crustaceans (Cyathura polita, Chiridotea almyra,
Almyracuma proximoculi) in the limnetic zone of the lower Hudson River, New York.
Estuaries 8(4): 373-380.
Simpson, K.W., J.P. Fagnani, R.W. Bode, D.M. DeNicola and L.W. Abele. 1986.
Organism-substrate relationships in the main channel of the Lower Hudson River.
J.N. An. Benthol. Soc. 5(1): 41-57.
Singh, S. F.K. Higson, L.M. Nadim, D.T. Gibson. 1988. Oxidation of
Polychlorinated Biphenyls by Psuedomonas Putida LB400. In Land Disposal,
Remedial Action, Incineration and Treatment of Hazardous Waste, Proceedings of
the Thirteenth Annual Research Symposium. Cincinnati, Ohio: Hay 6-8, 1987.
USEPA-600/9-88/021. Cincinnati: Hazardous Waste Engineering Research
Laboratory, USEPA.
Sinks, T., G. Stede, A.B. Smith, R. Rinsky, K. Watskins. 1990. "Risk Factors
Associated with Excess Mortality Among Polychlorinated Biphenyl Exposed Workers."
Interim Report HETA 89-116. Westinghouse Electric, Bloomington, Indiana, Health
Hazard Evaluation and Technical Assistance Branch, National Institute for
Occupational Safety and Health, Cincinnati, Ohio.
Sipes, I.G. and A.J. Gandolfi. 1986. Biotransformation of toxicants. In
Casarett and DoulVs Toxicology: the Basic Science of Poisons (Eds. C.d.
Klaassen, M.O. Amdur, and J. Doull). Macmillan Publishing Co. New York.
Sirois, D.L. and S.W. Fredrick. 1978. Phytoplankton and primary production on
the Hudson River estuary. Estuar. and Coast. Har. Sci. 7: 413-423.
Sloan, R.J. 1991. Personal communication. NYSDEC.
Sloan, R.J. 1989. " Results of the 1988 Hudson River Fish Sampling for PCB
Analyses." (internal report by Sloan dated October 3, 1989)
Sloan, R.J. and R.W. Armstrong. 1988. PCB patterns in Hudson River fish II.
Migrant/marine species. In Fisheries Research in the Hudson River (Ed. C.L.
Smith), pp. 325 -350State Univ., N.Y. Press, Albany.
R-30
-------
Sloan, R.J. and R.W. Armstrong. 1980. PCB patterns in Hudson River fish, II
Migrant/marine species. Proc. Hudson River Environ. Soc. Hyde Park, NY.
Sloan, R.J. and E.G. Horn. 1986. "Contaminants in Hudson River Striped Bass:
1978-1985." Technical Report 86-2 (BEP), NYSDEC, Div. of Fish and Wildlife,
Albany, NY
Sloan, R. and E.G. Horn. 1985. "PCB in Striped Bass from the Marine District
of New York in 1984." Bur. of Envir. Prot., Div. of Fish and Wildlife, NYS DEC,
February 14, 1985.
Sloan, R.J. and K. Jock. 1990. "Chemical Contaminants in Fish from the St.
Lawrence River Drainage on Lands of the Mohawk Nation at Akwesasne and Near the
General Motors Corporation/Central Foundry Division Massena, New York Plant."
N.Y.S. DEC Tech. Rep. 90-1(BEP), Div. of Fish and Wildlife. 96pp.
Sloan, R. and R. Karcher. 1984. On the origin of high cadmium concentrations
1n the Hudson River blue crab (Callinectes sapidus) rathbun. Northeastern
Environmental Science 3(3/4): 221-231.
Sloan, R., M. Brown, R. Brandt and C. Barnes. 1984. Hudson River PCB
relationships between resident fish, water and sediment. Northeastern
Environmental Sci. Vol. 3 (3/4: 138-152.)
Sloan, R.J., D. Stang and E.A. O'Connell. 1988. "Ten Years of Monitoring PCB
in Hudson River Striped Bass." Technical Report 88-2 (BEP), NYSDEC, Div. of Fish
and Wildlife, Albany, NY
Sloan, R.J., E.G. Horn, L.C. Skinner and W.J. Woodworth. 1987. "PCB in Hudson
River Striped Bass, Update 1986." Technical Report 86-4, NYSDEC, Albany, NY
Sloan, R.J., K.W. Simpson, R.A. Schroeder and C.R. Barnes. 1983. Temporal
trends toward stability of Hudson River PCB contamination. Bull. Environ.
Contaa. Toxicol. 31: 377-365.
Sloan, R. J., E. G. Horn, B. Young, C. Zawackl and A. Fortl. 1986. PCB in Striped
Bass from the Marine District of New York, 1985. Technical Report 86-1, NYSDEC,
Albany, NY.
Sloan, R.J., B. Young, V. Vecchio, K. McKown and E. O'Connell. 1988. "PCB
Concentrations in the Striped Bass from the Marine District of New York State."
Tech. Rep. 88-1 (BEP), Div. of Fish and Wildlife, NYSDEC.
Smith, C.L. and T.R. Lake. 1990. Documentation of the Hudson River fish fauna.
Amer. Mus. Novitates (2981): 17
Sofaer, A.D. 1976. Interim Opinion and Order in the matter of alleged
violations of the Environmental Conservation Law of the State of New York by
General Electric Co., Respondent. NYSDEC File No. 2833. February 9, 1976.
Albany, NY.
R-31
-------
Spagnoll, J.J. and L.C. Skinner. 1977. PCB's in fish from selected waters of
New York state. Pesticides Monitoring Journal 11(2): 69-87.
Squibb, K.S., J.M. O'Connor, and T.J. Kneip. 1991. M Module 3.1: Toxics
Characterization, Draft Report." NY/NJ Harbor Estuary Program. New York
University Medical Center, Institute of Environmental Medicine, New York. *
V
Stachiw, N.C., M.E. Zabik, A.M. Booren, and M.F. Zabik. 1988. TCDO Residue
Reduction Through Cooking/Processing of Restructured Carp Fillets. J. Agric.
Food Chen. 36: 848-852.
Stedfast, D.A. 1980. "Cross Sections of the Hudson River Estuary from Troy to
New York City, New York." USGS, Water Resources Investigations 80-24, 76 pp.
Steiner, W. 1991. ART, Inc. "Remedial Technologies: Innovative and
Permittable." Low Energy Extraction Process Seminar.
Stephens, G.C. 1967. Dissolved organic material as a nutritional source for
marine and estuarine invertebrates. Aner. Assoc. for the Adv. Sci. Special
Public. 83: 367-373.
Stickel, W.H., L.F. Stickel, R.A. Dyrland, and D.L. Hughes. 1984. Aroclor 1254
residues in birds: lethal levels and loss rates. Arch. Environ. Contam. Toxicol.
13: 7-13.
Stommel, H. 1953. Computation of pollution in a vertically mixed estuary.
Sewage and Ind. Hastes 25: 1065-1071.
Storm, P.C., and R.L. Heffner. 1976. "A Comparison of Phytoplankton Abundance,
Chlorophyll A and Water Quality Factors in the Hudson River and its Tributaries."
Paper No. 17, Hudson River Ecology, 4th Symposium on Hudson River Ecology, March
28-30, 1976.
Storti, M. 1991. Personal communication to S. Miller from M. Storti, Capital
District Regional Planning Commission.
Strayer, D. 1987. Ecology and zoogeography of the freshwater mollusks of the
Hudson River Basin. Narcological Review 20: 1-68.
Sutton, A. 1985. The Audobon Society Nature Guides. Eastern Forests. Alfred
A. Knopf, Inc. New York.
TAMS Consultants, Inc. 1990. "Hudson River Deepening Reconnaissance Study for
USAC0E, New York District." Project Files. NY.
Tanabe, S. 1988. PCB problems in the future: foresight from the current
knowledge. Environ. Pollut. 50: 5-28.
Terrell, C.R. and P.B. Perfetti. 1989. "Water Quality Indicators Guide: Surface
Waters. US Dept. of Agric." Soil Conservation Ser. SCS-TP-161, 129 pp.
R-32
-------
Thomann, R.V. 1989. Bloaccumulation model of organic chemical distribution in
aquatic food chains. Environ. Sci. Tech. 23: 699-707.
Thomann, R.V., J.A. Mueller, R.P. Winfield and Chi-Rong Huang. 1989.
"Mathematical Model of the Long-term Behavior of PCBs in the Hudson River
Estuary." Report prepared for The Hudson River Foundation, June 1989. Grant
Nos. 007/87A/030 and 011/88A/030.
Thome, C.R., and Osman, A.M. 1988. Riverbank stability analysis. II:
applications. Journal of Hydraulic Engineering, ASCE 114(2): 151-172.
Thurow, F. 1974. Zur Starke des Dorschjahrganges 1972 in der Westlichen Ostee.
Berichte der Deutschen Wissenshaftlichen Kommission fuer Meereforschung 23:
129-136.
TI (Texas Instruments Inc.) 1976a. "A Synthesis of Available Data Pertaining
to Major Physicochemical Variables within the Hudson River Estuary Emphasizing
the Period from 1972 Through 1975." Report to Consolidated Edison Company of New
York.
TI (Texas Instruments Inc.) 1976b. "Hudson River Ecological Study in the Area
of Indian Point. Thermal Effects Report." Report to Consolidated Edison Company
of New York.
TI (Texas Instruments Inc.) 1976c. "Liberty State Park Ecological Study."
Final Report prepared for Port of Authority of New York and New Jersey.
TI (Texas Instruments Inc.) 1975. "Benthic Landfill Studies, Cornwall Final
Report." Report to Consolidated Edison Company of New York.
TI (Texas Instrument Inc.) 1977. 1974 year-class report for the multiplant
impact study of the Hudson River estuary. Vol. II. Appendices. Prepared by
Texas Instruments for Consolidated Edison Co. of NY, Inc.
Tiedje, J.M., S.A. Boyd and J.F. Quensen. 1987. Reductive dechlorination of
PCBs in anaerobic microbial communities. In "Research and Development Program
for the Destruction of PCBs, Sixth Progress Report." Schenectady: General
Electric Company Corporate Research and Development Center.
Tiedje, J.M., S.A. Boyd, J.F. Quensen and J. Schimel. 1989. Reductive
dechlorination of PCBs by anaerobic microorganisms. In "Research and Development
Program for the Destruction of PCBs, Eighth Progress Report." Schenectady:
General Electric Company Corporate Research and Development Center.
Tofflemlre, T. J. 1980. "PCB 1n Sediments and Water and Their Transport."
Draft Report to NYSDEC, Albany, NY; cited in NUS, 1984.
Tofflemlre, T.J., and S.O. Quinn. 1979. "Estimates of PCB losses to air in the
Upper Hudson River Mapping and Sediment Relationships." VS "PCB 1n the Upper
Hudson River: Mapping and Sediment Relationships."? NYSDEC Technical Paper No.
56, Albany, NY. 140pp.
R-33
-------
Tofflemire, T.J., T.T. Shen, and E.H. Buckley. 1981. "Volatilization of PCBs
from Sediment and Water: Experimental and Field Data." NYSDEC Tech. Paper No.
63, Albany, New York 37pp.
Tofflemire, T.J., S.O. Quinn, and P.R. Hague. 1979. "PCBs in the Hudson River,
Mapping, Sediment Sampling, and Data Analysis." NYSDEC Technical Report No. 57.
Travis, C.C., and A.D. Arms. 1988. Bioconcentration of organics in beef, milk,
and vegetation. Environ. Scl. Techno!. 22(3): 271-274.
Turk, J.T. and D. E. Troutman. 1981. "Polychlorinated biphenyl transport in the
Hudson River, New York." USGS Water-Resources Investigations Report 81-9. USGS,
Albany, NY.
US Army Corps of Engineers, NY District. 1984. "Final Supplemental
Environmental Impact Statement, Westside Highway Project, Vol. II-Fisheries
Portion." 217pp.
US Army Corps of Engineers. 1977. "HEC-6 Computer Program: Scour and Deposition
in Rivers and Reservoirs, Users Manual." Hydrologic Engineering Center, Davis,
CA
USEPA. 1991. Region II, Toxics Substance Branch, Environmental Services
Division. Personal Communication on Toxic Release Inventory.
USEPA. 1991. "Research to Assess the Disappearance of PCB Resulting from
Treatment of Contaminated Materials with Quicklime." RREL, Cincinnati, OH.
USEPA. 1990. "Risk Assessment Guidance for Superfund. Volume I: Human Health
Evaluation Manual, Supplemental Guidance 'Standard Exposure Factors' Final
Draft." December 4, 1990. OSWER Directive: 9285.6-03.
USEPA. 1990. Region II. Section 304(1) of the Clean Water Act, Priority Water
Problems List. New York.
USEPA. 1990a. "Guidance on Remedial Actions for Superfund Sites with PCB
Contamination." Office of Emergency and Remedial Response. OSWER Directive No.
9355.4-01.
USEPA. 1990b. "Polychlorinated biphenyls (PCBs)." Integrated Risk Information
Service (IRIS).
USEPA. 1989. Washington, DC. EPA/600/8-89/043.
USEPA. 1989a. Exposure Factors Handbook. Exposure Assessment Group. Office
of Health and Environmental Assessment. US Pnvlronmental Protection Agency.
Washington D.C.
USEPA. 1989b. "Risk Assessment Guidance for Superfund. Volume I: Human Health
Evaluation Manual (Part A)." Office of Emergency and Remedial Response,
Washington, DC. Interim final, December 1989. EPA/540/1-89/002.
R-34
\
-------
USEPA. 1989c. "Dermal Absorption of Dioxlns and PCBs from Soil." Office of
Health and Environmental Assessment, USEPA: Washington, DC. Draft.
USEPA. 1989d. "Risk Assessment Guidance for Superfund. Volume II: Environmental
Evaluation Manual." Office of Emergency and Remedial Response, Washington, DC.
Interim final, March 1989.
USEPA. 1988. "Guidance for Conducting Remedial Investigations and Feasibility
Studies under CERCLA." EPA 540/G-89/004. Office of Emergency and Remedial
Response, Washington, D.C.
USEPA. 1988. "Superfund Exposure Assessment Manual." Office of Remedial
Response, USEPA: Washington, DC. EPA/540/1-88/001.
USEPA. 1988a. "Assessing Human Health Risks from Chemically Contaminated Fish
and Shellfish: A Guidance Manual." Office of Marine and Estuarine Protection,
EPA-503/8-89-002. Washington, D.C.
USEPA. 1988b. "Analysis of Risks From Consumption of Qulncy Bay F1sh and
Shellfish."
USEPA. 1988c. "Drinking Water Criteria Document for Polychlorinated B1 phenyls.
Final." Environmental Criteria and Assessment Office.
USEPA. 1987. Region II/ New York Department of Environmental Conservation.
January, 1987. Draft Joint Supplement to the Final Envlronmnetal Impact
Statement on the Hudson River PCB Reclamation Demonstration Project and
Appendices.
USEPA. 1986. "Development of Advisory Levels for Polychlorinated Biphenyls
(PCBs) Cleanup." Office of Research and Development, Office of Health and
Environmental Assessment, Exposure Assessment Group: Washington, DC. EPA/600/6-
86/002.
USEPA. 1985. "Handbook-Remedial Action at Waste Disposal Sites." EPA 625/6-
85/006. Office of Emergency and Remedial Response, Washington, D.C.
USEPA. 1984. "Health Effects Assessment for Polychlorinated Biphenyls."
Environmental Criteria and Assessment Office: Cincinnati, OH. ECA0-CIN-H004.
USEPA. 1983. "Environmental Transport and Transformation of Polychlorinated
Biphenyls." USEPA 560/5-83-025. Office of Pesticides and Toxic Substances,
Washington D.C.
USEPA. 1981. "Environmental Impact Statement on the Hudson River PCB.
Reclamation Demonstration Project." Draft May 1981.
USEPA. 1980. "Ambient Water Quality Criteria Document for Polychlorinated
Biphenyls." Office of Health and Environmental Assessment, Cincinnati, OH. EPA
440/5-80-068.
R-35
-------
USFWS. February 22, 1991. Letter to Gradient Corporation regarding federally
listed threatened and endangered species for New York state.
USGS (US Geological Survey). 1989. "Water Resources Data Report, New York,
Water Year 1988, Volume 1."
USGS (US Geological Survey). 1987. Water Resources Data - New York, Volume 1,
Water Year 1986, US Department of the Interior, Geological Survey.
USGS (US Geological Survey). 1985. Water Resources Data - New York, Volume 1,
Water Year 1984, US Depth of the Interior, Geological Survey.
USGS (US Geological Survey). 1982. Guidelines for determining flood flow
frequency. Hater Resources Council Bull. 17B, Washington, DC
USGS (US Geological Survey). 1982. Water Resources Data - New York, Volume 1,
Water Year 1981, US Department of Interior, Geological Survey.
USGS (US Geological Survey). 1970. "Surface Water Supply of the United States,
1961-1965, Part 1." Geological Survey Water Supply Paper 1902, US Government
Printing Office, Washington, D.C. 924.
Umbreit T.H., E.J. Hesse, and M.A. Gallo. 1986. Bioavailability of dioxin in
soil from a 2,4,5-T manufacturing site. Science 232: 497-499.
Unterman, R. 1990. Aerobic biodegradation of PCBs. In Biological Remediation
of Contaminated Sediments with Special Emphasis on the Great Lakes, Report of a
Workshop
(Eds. C.T. Jafvert and J.E. Rogers). Manitowoc, WI: 17-19 July, 1990.
Unterman, R., D.L. Bedard, M.J. Brennan, L.H. Bopp, F.J. Mondello, R.E. Brooks,
D.P. Mobley, J.B. McDermott, C.C. Schwartz and D.K. Dietrich. 1988. Biological
approaches for polychlorinated biphenyl degradation. Basic Life Sci. 45: 253-69.
Urabe, H., Koda, H., and Asahi, M. 1979. Present state of Yusho patients. Ann.
N.Y. Acad. Sci. 320, 273. As cited in Silberhorn et al., 1990.
Vanoni, V.A. (Editor). 1975. Sedimentation engineering. ASCE H & R (54) ASCE,
345 E 47 St., New York, NY.
Veith, G.D., D.L. DeFoe, and B.J. Bergstedt. 1979. Measuring and estimating the
bioconcentration factor of chemicals in fish. J. Fish. Res. Board Can. 36: 1040-
1048.
Vladykov, V.D. and D.H. Wallace. 1952. Studies of the striped bass, Roccus
saxatilis, (Walbaum), with special reference to the Chesapeake Bay region during
1936-1938. Bull. Brigham Oceanogr. Coll. 14(1): 132-177.
Vladykov, V.D. and D.H. Wallace. 1938. Is the striped bass (Roccus saxatilis)
of the Chesapeake Bay a migratory fish? Trans. Amer. Fish. Soc. 67: 67-86.
R-36
-------
Vogel, T.M., L. Niles and P.J. Anid. 1989. Long term anaerobic-aerobic
degradation of PCBs. In "Research and Development Program for the Destruction
of PCBs, Eighth Progress Report." Schenectady: General Electric Company
Corporate Research and Development Center.
Vorum, M. 1991. Personal Communication to P. Chen from M. Vorum, SoilTech, Inc.
Wagner, L.A. 1982. "Drainage Areas of New York State Streams, by River Basis-A
Stream Gazetteer." Water Resources Investigations Open File Report 81-1055, US
Dept. of the Interior, Geological Survey.
Warren, S.D., R.F. Bopp, and H.J. Simpson. 1987. "Violation of PCBs from
Contaminated Sediments and Water." Submitted to N.Y.S. Department of Environment
Conservation, Contract NYS C001263.
Washington County. 1976. "Recommended Land Use Pattern Map, Washington County,
New York."
Webb, R.G. and A.C. McCall. 1973. Quantitative PCB standards for electron
capture gas chromatography. Journal of Chromatographic Science 11-: 366-373.
Weber, J.B. and E. Mrozek, Jr. 1979. Polychlorinated biphenyls: phytotoxicity,
absorption and translocation by plants, and inactivation by activated carbon.
Bull. Environ. Contain. Toxicol. 23: 412-417.
Webster, R.C. and P.K. Noonan. 1980. Relevance of animal models for
percutaneous absorption. Int. J. Pharn.
Webster, W. 1987. Memorandum dated June 23, 1987 to J. VanHoesen, entitled
"Kingsbury Landfill Remediation Air Sampling, April 2- June 16, 1987."
Weimer, L. D. 1991. Resources Conservation Co. "B.E.S.T. Solvent Extraction,
Treatment of PCB Contaminated Soils from Natural Gas Pipeline Sites." Haztech
International Conference, Houston, Texas.
Weimer, L. D. 1989. Resources Conservation Co. "The B.E.S.T. Solvent
Extraction Process, Applications with Hazardous Sludges, Soils and Sediments."
Third International Conference. New Frontiers for Hazardous Waste Management,
Pittsburgh, PA.
Weimer, L. D. 1990. Resources Conservation Co. "B.E.S.T. - Solvent Extraction
Treatment of PCB Contaminated Soil." Case Study Demonstration.
Weinsteln, L.H. (ed.). 1977. An Atlas of the Biologic Resources of the Hudson
River Estuary. Boyce Thompson Institute of Plant Research. Yonkers, N.Y. 104
pp.
Westin, D.T. and B.A. Rogers. 1978. "Synopsis of Biological Data on Striped
Bass, Morone saxatilis (Walbaum), 1972." Univ. Rhode Island Mar. Tech. Rep. 67.
154 pp.
R-37
-------
Weston Environmental Consultants. 1978. "Migration of PCBs from Landfills and
Dredge Spoil Sites in the Hudson River Valley, New York Final Report."
Prepared for New York State Department of Environmental Conservation, Albany, NY.
Wetzel, R.G. 1975. Limnology. W.B. Saunders Co., Phil., 743 pp.
%
Whitton, B.A. 1975. River Ecology. Blackwell Sci. Public., Oxford 725pp. V
Whittow, 6. and H. Rahn. 1984. Seabird Energetics. Plenium Press. New York.
328pp.
Wilk, S.J., W.W. Morse, D.W. Ralph and T.R. Asarovitz. 1977. "Fish and
Associated Environmental Data Collected in New York Bight, June,!974-Junel973."
NOAA tech rpt., NMFS SSRF-716.
Will ford, W.A., R.A. Bergstedt, W.H. Berlin, et al. 1981. Chlorinated
hydrocarbons as a factor in the reproduction and survival of lake trout in Lake
Michigan: introduction and summary. Tech. Paper US Fish and Wildlife Service.
105:1-7
Wolff, M. 1983. Occupationally derived chemicals in breast milk. Am. J.
Indust. Med. 4: 259-281.
Woodyard, J. P. 1990. PCB detoxification technologies: A critical assessment.
Environmental Progress 9(2).
Wool, T. A. 1990. "WASP4 Model Enhancement and System Update Documentation."
USEPA Environmental Research Laboratory, Athens, GA
Yakushiji, T., I. Watanabe, K. Kuwabara, S. Yoshida, K. Koyama, I. Hara, and N.
Kunita. 1978. Long-term studies of the excretion of PCBs through mother's milk
of an occupationally exposed worker. Arch. Enviro. Contaa. Toxicol. 7: 493-504.
Young, B.H. 1981. "A Study of Striped Bass in the Marine District of NY III,
April 1980-March 1981, Segment 2." For: Anadromous Fish Act (NYS-DEC).
Young, B.H. 1980. "A Study of Striped Bass in the Marine District of NY III,
April 1979-March 1980, Segment 1." For: Anadromous Fish Act (NYS-DEC).
Zabik, M.E., C. Merrill, and M.J. Zabik. 1982. PCBs and other xenobiotics in
raw and cooked carp. Bull. Env. Contain. Toxicol. 28: 720-725.
Zack, T.A. and Musch, D.C. 1979. "Mortality of PCB workers at the Monsanto
Plant 1n Sanget, Illinois." Monsanto Internal Report, St. Louis. As cited in
Silberhorn et al., 1990.
Zembrzuski, T.J. and B. Dunn. 1979. Techniques for Estimating Magnitude and
Frequency of Floods on Rural Unregulated Streams in New York State, Excluding
Long Island. US Geological Survey, Water Resources Investigations 77-83, 66 pp.
R-38
-------
Zimmie, T. F. 1985, "Assessment of Erodibility of Sediments In the Thompson
Island Pool of the Hudson River." Report to NYSDEC, published as Appendix B of
M. P. Brown et a?., 1988, Distribution of PCBs in the Thompson Island Pool of the
Hudson River, Final Report of the Hudson River PCB Reclamation Demonstration
Project Sediment Survey. EPA Grant C361167-01
ZVI Blank. 1990. ART, Inc. "Low Energy Extraction Process - LEEP - A New
Technology to Decontaminate Soils, Sediments, and Sludges." Haztech Internation-
al 90, Houston, Texas.
R-39
-------
PAGE INTENTIONALLY BLANK
R-40
-------
GLOSSARY
PHASE 1 REPORT
INTERIM CHARACTERIZATION AND EVALUATION
HUDSON RIVER PCB REASSESSMENT RI/FS
Adjusted Mean: Used here to refer to an estimate of the mean adjusted to account
for bias 1n the sampling process.
Absorb: A chemical/physical bonding which 1s normally not reversible.
Adsorb: A chemical/physical surface adhesion which is normally reversible.
Adsorption Coefficient: The ratio of the amount of a chemical adsorbed per unit
weight of organic carbon soil or sediment to the concentration of the chemical
in solution at equilibrium, (see K* K#e.)
Allocthonous: Inputs of organic carbon from the surrounding terrestrial
watershed and/or those carbon inputs generated outside the riverine system.
Anadromous: Fish that ascend rivers from the sea for spawning.
ARAR: Applicable or Relevant and Appropriate Requirements.
Arithmetic Mean (Average!: For a set of data, the total sum of all sample values
divided by the number of samples.
Aroclor: A trade name applied to mixtures of PCBs. When used to refer to a
particular PCB mixture, 1t is usually followed by a numerical code which
indicates the percentage of chlorine in that mixture.
Arvl Hydrocarbon Hydroxylase (AHH1: A type of enzyme produced by the liver.
Concentration or activity of AHH in the liver may Increase in response to
exposure to specific PCB congeners or other toxic compounds.
Autocthonous: Inputs of organic carbon that are generated within the riverine
system.
Average Lifetime Dally Dose (ALDD): Exposure expressed as mass of a substance
contacted per unit body weight per unit time (mg/kg-d), averaged out over a
lifetime.
BAF: B1oaccumulat1on factor (BAF) 1s the ratio of the concentration of a
chemical 1n an organism living 1n contaminated water and consuming contaminated
food, to the concentration of the same chemical in the surrounding water. See
G-l
-------
also bloconcentration factor (BCF).
BCF: Bioconcentration factor (BCF) is the uptake and tissue accumulation of a
chemical from water. No other sources such as sediments and food are included.
Benthos: Animals associated with the aquatic substrata.
Bioaccumulation: The uptake and tissue accumulation of a chemical by biota from
water and food sources.
Bioconcentration: The partitioning of a chemical between water and biota (often
fish). Due to the chemical composition of water and biota, many organic
chemicals have a greater tendency to partition into (I.e. concentrate in) biota
rather than to remain in water.
Biodearadation: The microbially mediated transformation of a chemical.
BOD: Biological oxygen demand.
Cancer Potency Factor (CPF): An index of the cancer-causing ability of a
compound, based on the slope of the dose-response curve. The potency factor is
used to estimate an upper-bound probability of an individual developing cancer
as a result of a lifetime of exposure to a particular level of a potential
carcinogen.
Capillary Column Gas Chromatograph: An instrument used to separate, detect, and
quantify chemicals. Separation of the chemicals employs a relatively small
diameter column. This instrument can separate many chemicals a packed column
chromatograph can not separate. See packed column chromatograph.
Carcinogen: A chemical or physical agent (such as radiation) capable of inducing
cancer.
Carcinogenic: Capable of producing or inciting cancer.
Catadromous: Fish that live in fresh water and go to the sea to spawn, e.g.,
eels.
CERCLA: Comprehensive Environmental, Response and Compensation Liability Act.
cfs: cubic feet per second; english unit measurement of river discharge.
Chloohen: PCB mixture manufactured in Germany.
cm/hr:' centimeters per hour.
Confidence Interval: An interval defining the probable range of a random
variable with the given confidence. For example, a 95 percent confidence
interval about the sample mean is the range within which the true mean can be
G-2
-------
expected to fall 95 percent of the time in repeated trials.
Confidence Limit: The value defining the end of a confidence interval.
Congener: A generic term used to refer to any possible PCB molecule. (See also
isomer and homologue)
Coplanar PCB: A subset of PCB congeners whose physical configuration is such
that the biphenyl molecules lie in approximately the same plane in space,
resulting in a relatively flat molecular structure.
Correlation Coefficient: Measure of the degree to which two variables are
linearly related; the range is 0 (no correlation) to 1 (perfect correlation).
Demersal: Bottom dwelling.
Detection Limit: Technically, the lowest concentration of a chemical that can
be observed using a particular analytical method. Frequently used to refer to
the lowest concentration of a chemical that can be routinely quantified using a
particular analytical method.
Dissolved Oxvoen I percent saturation): For comparative purposes, dissolved
oxygen concentration in mg/1 1s often expressed in terms of the percentage
saturation of the sample (X saturation - 100 G/G') where G 1s the observed
concentration of the gas and G' is its solubility in water of the appropriate in
situ temperature and salinity. Values may be greater than 100% (supersaturated)
or less than 100% (undersaturated). Supersaturated conditions are generally
caused by algae blooms, whereas undersaturated conditions are due to respiration
by plants and animals, sediment oxygen demand and oxidation of organic
particulates within the water column.
Dose-Response Curve: The line or curve that depicts the relationship between the
exposure dose of a compound and the toxic response of the exposed population;
usually taken from a graph showing laboratory experimental results where the dose
is presented on the X-ax1s, and the response presented on the y-axis.
Drv weight: The weight of a substance (e.g. fish tissue or sediment) after all
water has been removed from it.
ฃh: Measure of the state of oxidation or reduction of a system in millivolts
(mv). It 1s basically a measure of the ability of the system to supply or use
up electrons. The Eh scale ranges from +500 mv (extremely oxidizing) to -500 mv
(extremely reducing).
Error Bars? Graphical designation of a confidence interval.
Eurvhaline: Organisms that can tolerate a wide range of salinities.
G-3
-------
Exposure Pathway: The course a chemical or physical agent takes from a source
to an exposed individual. An exposure pathway describes a unique mechanism by
which an individual or population is exposed to chemicals or physical agents at
or from a site. Each exposure pathway includes a source or release from a
source, an exposure point, and an exposure route. If the exposure occurs at a
point different from the exact source, then a transport/exposure medium {e.g.,
air) also is involved.
F-Test: A statistical test of significance of the ratio of variances of random
variables.
FDA: Food and Drug Administration.
Flow-Duration Curve: A plot showing river discharges and their corresponding
exceedance probabilities.
Gas Chromatography: An analytical technique used to separate, detect and
quantify chemicals.
Geometric Mean (Average): For a set of data, the anti-log arithmetic mean of the
natural logarithms of the data.
Hazard Quotient: The ratio of a single substance exposure dose over a specified
time period (e.g., chronic) to a reference dose for that substance derived from
a similar exposure period.
Hemosiderosis: A pathological condition marked by the deposition of hemosiderin
in the tissues as a result of the breakdown of red blood cells.
Henry's Law Constant: Provides a measure of the extent of chemical partitioning
between air and water at equilibrium. The higher the Henry's Law constant, the
more likely a chemical is to volatilize than to remain in the water.
Hepatic Carcinoma: A malignant new growth in the liver made up of epithelial
cells.
Hepatic Neoplastic Nodules: A small rounded mass of tissue in the liver that is
in the form of a swelling, knot, or protuberance, which is a new and abnormal
growth, and in which the growth is progressive but is not malignant [i.e., does
not spread to other parts of the body).
Hepatocellular Carcinoma: A malignant new growth in the liver made up of
epithelial cells of hepatic origin, and tending to infiltrate the surrounding
tissues and to be malignant (I.e., spreads from one area to another in the body).
Hepatocellular Neoplasms: Any new and abnormal growth in the liver; specifically
a new growth of tissue in which the growth is uncontrolled and progressive, and
may or may not be malignant (i.e., may or may not spread to other parts of the
body).
G-4
-------
Hlstopathologv: Study of cellular pathology.
Homoloaues: Molecules making up a series in which each successive group of the
series is characterized by one more atom, or group of atoms, than the preceding
group. For example, mono- through decachlorinated biphenyls make up a homologous
series in which each successive homologue group contains compounds with one more
chlorine atom than the preceding group. Members of this series are referred to
as homologues.
Hyperplasia: An abnormal increase in the cells of a tissue.
Interstitial: Referring to the sediment pore space.
In Vitro: In an artificial environment outside of the living organism, as in a
test tube.
Instantaneous Discharge: A measurement of river flow at a specific point (e.g.,
hourly) in time.
IRIS: Integrated Risk Information System. IRIS is an on-line database created
by the EPA and maintained on the National Library of Medicine's Toxicology Data
Network system. The database contains EPA health risk and regulatory information
on some 400 chemicals, with both carcinogenic and non-carcinogenic risk
assessment data for oral and inhalation routes of exposure. These data include
Reference Doses (RfD), indicators of non-carcinogenic risks, and Unit Risks,
indicators of carcinogenic risks. The regulatory information relates to
environmental statutes such as the Clean A1r Act, Clean Water Act, and SUPERFUND
legislation. IRIS is further supplemented with EPA Drinking Water Health
Advisories, substance identification, chemical and physical properties, acute
toxicity, and aquatic toxicity.
Isomers: Molecules containing the same number and types of atoms but differing
1n arrangement. For example, 2,3-dichlorob1pheny1 and 2,4-d1chlorob1phenyl are
isomers of each other.
Provides a soil or sediment-specific measure of the extent of chemical
partitioning between soil or sediment and water, unadjusted for dependence upon
organic carbon. The higher the K,, the more likely a chemical is to bind to soil
or sediment than to remain in water. See adsorption coefficient and partition
coefficient.
j^: Provides a measure of the extent of chemical partitioning between organic
carbon and water at equilibrium. The higher the K.., the more likely a chemical
is to bind to soil or sediment than to remain in water.
K^: Provides a measure of the extent of chemical partitioning between water and
octanol at equilibrium. The greater the K^, the more likely a chemical is to
partition to octanol than to remain In water. Octanol is used as a surrogate for
lipids (fats) and K^, can be used to predict bioconcentration in aquatic
6-5
-------
organisms.
Lipid: One of a class of compounds that contain long chain aliphatic
hydrocarbons and their deviatives such as acids (fatty acids), amino alcohols and
aldehydes. The presence of the long aliphatic chain confers distinct solubility
properties and has led to the traditional definition of lipids as substances
which are insoluble in water but soluble in ether, chloroform and benzene.
Linear Regression: A method of fitting a linear relationship between variables,
in which one variable (the dependent variable) is estimated from one or more
other variables (independent variables). The best fit relationship is usually
accomplished by the method of least squares.
Linearized Multistage (LMS) Model: A mathematical model used to extrapolate the
risk of cancer from high dosed observed in animal studies to the lower doses
usually experienced by humans from environmental exposures. A multistage
extrapolation model is based on the theory that several distinct changes are
necessary to transform a normal cell into a malignant cell, and that human cancer
can arise from a single transformed cell. The presence of a linear term in the
model insures near linearity for the confidence limit of the extrapolation.
Log Transformation: A data transformation made by taking the logarithms of the
sample data, often used to stabilize, or minimize, variance in environmental
data.
Loo Pearson Type III Distribution: An extreme-value probability distribution
proposed for flood frequency analysis by the U.S. Water Resources Council.
Lotic: Refers to any system of flowing water such as streams and rivers.
Lowest-Observed-Adverse-Effect-Level fLOAEL^: In dose-response experiments, this
is the lowest exposure level at which there are statistically or biologically
significant increases in frequency or severity of adverse effects in an exposed
population.
m*/s: cubic meters per second; metric unit measurement of river discharge.
Macrophvte: Macroscopic forms of vegetation.
Mann-Kendall Test: A non-parametric trend test used here to determine
statistically significant declines in PCB concentrations in various media; a
useful test in cases of missing data and non-detects.
Mass Spectrometry: An analytical technique used to identify chemical structures
and to determine chemical concentrations. It is often used in combination with
gas chromatography.
G-6
-------
Maximum Tolerated Dose fMTPl: The highest dose that causes no more than t 10
percent weight decrement in the exposed animal population, and does not produce
mortality, clinical signs of toxicity, or pathologic lesions (other than those
related to cancer) that would be predicted to shorten the animals* natural
lifespan.
Maximum Contaminant Level (MCLl: MCLs are the maximum allowable concentrations
of contaminants in drinking water as established by federal and state agencies
responsible for regulating public water systems. These are legally enforceable
standards. MCLs are set to protect the public from acute or chronic health
effects or an "unacceptable" cancer risk. They must also take into account
technological feasibility and economic impact.
Median: For a set of data, defined as the midpoint in the data set, e.g., half
of the sample values are lower and half higher than the median.
mo/cm2: milligrams per square centimeter.
a dose rate of exposure to contaminants expressed in milligrams of
contaminant per kilogram of body weight per day.
Multiple Repression: Linear regression with multiple dependent variables.
no/a: nanograms per gram (parts per billion), representing one-trillionth (10"*)
of a gram of chemical per gram of media (e.g. sediment).
NIOSH: National Institute of Safety and Health.
No-Observed-Adverse-Effect-Level (NOAEH: In dose-response experiments, a
chemical exposure level at which there are no statistically or biologically
significant adverse effects between a population exposed to the chemical versus
a control population; some effects may be produced at this level, but they are
not considered to be adverse, nor precursors to specific adverse effects.
NOAA: National Oceanic and Atmospheric Association.
Non-detects: Samples in which a particular chemical can not be measured at a
concentration exceeding the detection limit.
Non-Parametric Test: A statistical test which does not depend on assumptions
regarding the nature of the probability distribution defining a given data set
(e.g., no assumption of normality, lognormality, etc. is made about the data).
NPL: National Priority List.
NYSDEC: New York State Department of Environmental Conservation.
NYSDOH: New York State Department of Health.
6-7
N
-------
NYSDOT: New York State Department of Transportation.
OSWER: Office of Solid Waste and Emergency Response within the USEPA.
Packed Column Analysis: The detection and quantitation of chemicals using a
packed column gas chromatograph.
Packed Column Gas Chromatoaraph: An instrument used to separate, detect, and
quantify chemicals. Separation of the chemicals employs a relatively large
diameter column which may be incapable of separating some very similar chemicals.
Chemicals which can not be separated are difficult to accurately quantitate. See
capillary column chromatograph.
Partition Coefficient: A constant describing the equilibrium distribution of a
chemical between two materials. This constant is the ratio of the chemical
concentration in one material (in environmental applications, usually a
hydrophobic material) to the concentration of the same chemical in a second
material (in environmental applications, usually water). and Koe are two
examples.
PCB Metabolite: Following absorption of PCBs, the body metabolizes them.
Metabolism of PCBs is dependent upon on the number and position of chlorine
atoms, with lesser chlorinated isomers metabolized more readily than more
chlorinated isomers. PCB metabolites are the products of this metabolism, and
tend to be more water soluble (and thus more easily excreted) than the parent
compound.
PCB: A polychlorinated biphenyl (PCB) is a member of the chemical class of
chlorinated, aromatic hydrocarbons. A PCB consists of two connected rings of six
carbon atoms each (this ring structure is called a biphenyl), to which one or
more chlorine atoms are attached at any of 10 available sites.
PCDF: Polychlorinated dibenzofuran, a chemical related to TC00.
Percentile: Value at which a given percent of the probability mass of a variable
is exceeded. Commonly reported percentiles are the 50th percentile, or median,
and the 95th percentile, which is a value exceeded by 5 percent of the sample
data.
Peri&hvton: Microfloral growth upon the aquatic sediments.
CU: Measure of the acidity or alkalinity of a substance based.on a 0-14 scale
and Is a function of the free hydrogen ions or protons. A pH of 7.0 is
considered neutral and values below pH 7.0 are rcsre acidic and values above pH
7.0 are more basic.
Phvtoplankton: Free-floating raieroseopie algae.
Picocurie: One-trillionth of a Curie.
G-8
V
-------
Piscivorous: F1sh-eat1ng.
PI anktlvorous: Plankton-eating.
ppb: Parts per billion, equivalent to jig/7 in water at low solute
concentrations.
ppin: Parts per million, equivalent to mg/7 in water at low solute
concentrations.
Primary Contact Recreation: Recreational activities where the human body may
come in direct contact with raw water to the point of complete body submergence,
such as swimming, diving, etc.
RCRA: Resource Conservation and Recovery Act. RCRA (1976) was passed by
Congress 1n response to the potential problems posed by disposal of wastes
generated by chemical and other industrial processes. RCRA promotes continuous
management from point of generation to final disposal of hazardous wastes. The
program works through requirements for hazardous waste generators, transporters,
and treatment, storage, and disposal facilities. In 1984, RCRA was amended by
the Hazardous and Solid Waste Amendments that required EPA to focus on permitting
land disposal facilities and eventually phasing out land disposal of some wastes.
Reference Dose (RfD): A benchmark for the dally dose to which humans, Including
sensitive populations (such as children or pregnant women), may be subjected
without an appreciable risk of adverse non-carcinogenic health effects during a
lifetime of exposure.
Residual Variance: Portion of total variation in a data set due to chance or
error.
RI/FS: Remedial Investigation/Feasibility Study. The RI/FS 1s the framework for
determining appropriate remedial actions at Superfund sites. Remedial
investigations are conducted to characterize the contamination at the site and
to obtain information needed to identify, evaluate, and select cleanup
alternatives. The feasibility study includes an analysis of remedial action
alternatives based on National Contingency Plan evaluation criteria.
RIBS: Rotating Intensive Basin Studies, NYSDEC monitoring program for the
assessment of ambient water quality.
ROD: Record of Decision, typically for a Superfund site.
Salt Front: The furthest upstream point at which the influence of seawater can
be measured in the freshwater, typically about 500 ppm of salinity.
SARA: Superfund Amendments and Reauthorization Act of 1986. "
G-9
-------
Secondary Contact Recreation: Recreational activitifis where contact with water
is minimal and where ingestion of water is not probable, such as fishing,
boating, etc.
Shear Stress ft): Force per unit area exerted by a moving fluid on its substrate
-irt the plane of contact, used in this report in reference to the stress Imparted
by water flowing over river bottom sediments.
SITE: Superfund Innovative Technology Evaluation.
Skewed Distribution: A probability distribution which 1s not symmetric about the
mean, e.g. the lognormal distribution.
Sorb: The process of either adsorbing and/or absorbing a chemical species onto
a solid medium. (See definitions of absorb and adsorb.)
SPDES: State Pollutant Discharge Elimination System, a permit program adopted
pursuant to Article 17, Title 8 of the Environmental Conservation Law and Section
402 of the Clean Water Act that imposes discharge limitations on point sources.
Spearman Rank Correlation: A non-parametric measure of correlation between two
variables, in which rank orders rather than actual values are compared.
Stage-Discharge Curve: Plot showing the relationship between river height
(stage) and flow (discharge).
Suggested No Adverse Response Level: Values suggested by the National Academy
of Sciences that represent exposure concentrations of contaminants in drinking
water or air that are expected to be safe for human exposure. Generally, an
acceptable time period of exposure (e.g., 24-hour, 7-day or long-term exposure)
is also indicated, and different exposure concentrations may be suggested for
each period of exposure.
Taxa: A major taxonomic category or subdivision such as class, order or family.
TCDD: Tetrachlorodibenzodioxin, of which the 2,3,7,8-TCDD Isomer 1s considered
to be the most toxic.
Time Weighted Average - Threshold Limit Values (TWA-TLV): The average
concentration for a normal 8-hour work-day and a 40-hour workweek, to which
nearly all workers may be repeatedly exposed, day after day, without adverse
effect.
Toxic Eouivalency Factor (TEF1: An approach used to characterize the toxicity
of a complex mixture of toxic compounds when adequate toxicity information is
available only for some of the components of the mixture. In general, the
toxicity of one component of the mixture is known. This compound's potency is
set to 1.0, and the toxicity, of other components ft estimated by comparison
against this base compound. Ideally, use of the TEF approach accounts for the
G-10
-------
differing toxicologic*! properties of the various components of the mixture, and
yields a more reel supporttd estimate .of the potential
risks from exposure to t^% . ;5; ..7- . //ฆ.
Trophic: Refers to food lev.el, e.$r primary producer, consumer, etc. ฆ
ue/J: micrograms per liter, equivalent .to PP& ,for water at low* sdfiite
concentrations. x
uq/q: micrograms per gram (parts per million), representing one-millionth (10"*)
of a gram of chemical per gram of media (e.g. sediment).
ug/m3: micrograms per cubic meter, equivalent to ppi (parts per trillion) for
water at low solute concentrations.
.' ..
USEPA: United States Environmental Protection Agency.
USFWS: United States Fish and Wildlife Service.
USSS: United States Geological Survey.
Water Year: The US6S Water Year is defined as running from Octdberf^ of ^the
previous calendar year to September 30 of th# current calendar ฃea
6-n
------- |