p
QNV H11V3H NV1AIHH 3AllVaVdlAIOO
¥dl
-------
COMPARATIVE HUMAN HEALTH AND WILDLIFE RISK
ASSESSMENT: BUFFALO RIVER, NEW YORK,
AREA OF CONCERN
by
Judy L. Crane
EVS Consultants
Seattle, Washington 98119
Project Officer
Marc Tuchman
Great Lakes National Program Office
U.S. Environmental Protection Agency
Chicago, Illinois 60604-3590
-------
DISCLAIMER
The information in this document has been funded by the U.S. Environmental Protection Agency. It
has been subjected to the Agency's peer and administrative review, and it has been approved for
publication as an EPA document. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use by the U.S. Environmental Protection Agency.
-------
TABLE OF CONTENTS
Page
DISCLAIMER ii
LIST OF TABLES v
LIST OF FIGURES vii
ACKNOWLEDGMENTS viii
1.0 EXECUTIVE SUMMARY 1-1
1.1 OVERVIEW 1-1
1.2 STUDY AREA 1-1
1.3 EXPOSURE ASSESSMENT 1-2
1.3.1 Modeled Data 1-2
1.3.2 Human Exposure 1-2
1.3.3 Wildlife Exposure 1-3
1.4 RISK ASSESSMENT 1-4
1.4.1 Determination of Human Health Risk 1-4
1.4.1.1 Consumption of Contaminated Fish 1-4
1.4.1.2 Ingestion of Contaminated Surface Water 1-6
1.4.2 Determination of Wildlife Risk 1-6
1.5 UNCERTAINTIES 1-6
2.0 INTRODUCTION 2-1
2.1 BACKGROUND 2-1
2.2 BASELINE HUMAN HEALTH RISK ASSESSMENT 2-2
2.3 BASELINE WILDLIFE RISK ASSESSMENT 2-2
2.4 COMPARATIVE RISK ASSESSMENT APPROACH 2-4
3.0 COMPARATIVE RISK ASSESSMENT FRAMEWORK 3-1
3.1 OVERVIEW 3-1
3.2 MASS BALANCE MODELING 3-5
3.2.1 Selection of Remedial Action Scenarios 3-5
3.2.2 Mass Balance Modeling Framework 3-6
3.2.2.1 SUNY Modeling Effort 3-6
3.2.2.2 Limitations of Modeling Effort 3-9
4.0 EXPOSURE ASSESSMENT 4-1
4.1 EXPOSURE PATHWAYS 4-1
4.1.1 Human Exposure Pathways 4-1
4.1.2 Wildlife Exposure Pathways 4-2
4.2 MODELED DATA USED IN THE EXPOSURE ASSESSMENT 4-5
4.2.1 Carp Data 4-5
4.2.2 Surface Water Data 4-11
4.3 EXPOSURE ASSESSMENT FOR HUMAN HEALTH 4-12
4.3.1 General Determination of Chemical Intakes 4-12
4.3.2 Ingestion of Contaminated Fish 4-14
4.3.3 Ingestion of Surface Water While Swimming 4-17
in
-------
TABLE OF CONTENTS
Page
5.0 TOXICITY/HAZARD ASSESSMENT 5-1
5.1 HUMAN HEALTH TOXICITY VALUES 5-1
5.2 WILDLIFE HAZARD ASSESSMENT 5-2
5.3 LIMITATIONS 5-2
6.0 COMPARATIVE RISK CHARACTERIZATION: HUMAN HEALTH 6-1
6.1 PURPOSE OF THE RISK CHARACTERIZATION STEP 6-1
6.2 QUANTIFYING RISKS 6-1
6.2.1 Determination of Noncarcinogenic Risks 6-1
6.2.2 Determination of Carcinogenic Effects 6-2
6.3 COMPARATIVE HUMAN HEALTH RISKS IN THE BUFFALO RIVER 6-2
6.3.1 Consumption of Contaminated Fish 6-2
6.3.2 Consumption of Contaminated Surface Water 6-5
6.3.3 Additive Risks 6-6
7.0 COMPARATIVE RISK CHARACTERIZATION: WILDLIFE 7-1
7.1 INTRODUCTION 7-1
7.2 COMPARATIVE RISKS TO MINK 7-1
8.0 CHARACTERIZATION OF QUALITATIVE UNCERTAINTIES 8-1
8.1 INTRODUCTION 8-1
8.2 QUALITATIVE LIST OF UNCERTAINTIES: HUMAN HEALTH 8-1
8.2.1 Exposure Assessment 8-1
8.2.2 Toxicity Values 8-2
8.2.3 Risk Characterization 8-3
8.3 QUALITATIVE LIST OF UNCERTAINTIES: WILDLIFE 8-3
8.4 SUMMARY 8-4
REFERENCES 9-1
APPENDIX A: EXECUTIVE SUMMARY OF THE BASELINE HUMAN HEALTH RISK
ASSESSMENT A-1
APPENDIX B: WILDLIFE FOUND IN THE BUFFALO RIVER AOC B-1
APPENDIX C: HUMAN HEALTH TOXICITY ASSESSMENT INFORMATION C-1
APPENDIX D: TOXICITY PROFILES D-1
IV
-------
LIST OF TABLES
Table Page
1.1 Amount of Carp Assumed to be Consumed per Person per Day from the Buffalo River
for each Exposure Scenario 1 -3
1.2 Noncarcinogenic Risks Associated with Consuming Whole Carp Under Various
Remediation and Consumption Scenarios 1-5
1.3 Carcinogenic Risks Associated with Consuming Whole Carp Under Various Remediation
and Consumption Scenarios 1 -5
4.1 Complete Exposure Pathways in the Buffalo River AOC 4-1
4.2 Age Class Data on Carp Collected from the Buffalo River 4-6
4.3 Modeled PCB Concentrations in Carp for all Upstream Scenarios (DePinto et al., 1994) 4-7
4.4 Modeled PCB Concentrations in Carp for all Downstream Scenarios (DePinto et al., 1994)4-8
4.5 Summary of Modeled PCB Concentrations in Buffalo River Carp for 10-year and 30-year
Scenarios 4-11
4.6 Modeled Upstream Water Column Concentrations for Various Remedial Alternatives 4-14
4.7 Modeled Downstream Water Column Concentrations for Various Remedial Alternatives 4-15
4.8 Generic Equation for Calculating Chemical Intakes (USEPA, 1989a) 4-16
4.9 Equation Used to Estimate Contaminant Intakes Due to Ingestion of Fish 4-17
4.10 Parameters Used in Estimating Contaminant Intakes Due to Consumption of Fish
from the Buffalo River AOC 4-18
4.11 PCB Intake Rates Resulting from the Typical Consumption of Carp under Different
Remediation Scenarios 4-19
4.12 PCB Intake Rates Resulting from the Reasonable Maximum and Subsistence
Consumption of Carp under Different Remediation Scenarios 4-20
4.13 Equation used to Estimate Contaminant Intake Rates due to Ingestion of Surface Water
While Swimming 4-21
4.14 Parameters Used for Computing Ingestion of Surface Water While Swimming 4-22
4.1 5 Exposure Intake Rates Associated with Ingesting Contaminated Surface Water While
Swimming 4-22
5.1 EPA Weight-of-Evidence Classification System for Carcinogenicity (USEPA, 1989a) 5-1
5.2 Human Health Risk Toxicity Data for Chemicals of Interest in the Buffalo River 5-3
-------
LIST OF TABLES
Table Page
6.1 Hazard Quotients for Noncarcinogenic Risks Associated with Consuming Whole Carp
Under Various Remediation and Consumption Scenarios 6-3
6.2 Carcinogenic Risks Associated with Consuming Whole Carp Under Various
Remediation and Consumption Scenarios 6-3
6.3 Noncarcinogenic and Carcinogenic Risks Associated with Ingesting Contaminated Surface
Water While Swimming in the Buffalo River: Environmental Dredging Scenario 6-6
7.1 Comparative Risks to Mink Resulting from the Consumption of Contaminated Carp for
Various Remediation Alternatives 7-2
VI
-------
LIST OF FIGURES
Figure Page
2.1 Map of Buffalo River Area of Concern (NYSDEC, 1989) 2-3
3.1 Comparative risk assessment in the risk management process 3-2
3.2 Components of the mass balance modeling study 3-7
4.1 Location of Times Beach Confined Disposal Facility 4-4
4.2 Modeled PCB concentrations in carp: Upstream scenario 4-9
4.3 Modeled PCB concentrations in carp: Downstream scenario 4-10
4.4 Water column segmentation 4-13
VII
-------
ACKNOWLEDGMENTS
EVS Consultants gratefully acknowledges the U.S. EPA's Great Lakes National Program Office (GLNPO)
in Chicago, IL and the National Oceanographic and Atmospheric Administration (NOAA) in Seattle, WA
for their support during the course of this work. Alyce Fritz was the NOAA Project Officer, whereas
Marc Tuchman was the EPA Project Officer. Modeled fish and water column data were obtained from
Joseph DePinto (State University of New York-Buffalo). This risk assessment was prepared by Judy
Crane as part of the EPA's Assessment and Remediation of Contaminated Sediments (ARCS) program.
In-house review of this report was provided by Robert Dexter. Blair Luscombe, Angela Crampton, Steve
Coleman, Vickie Duff, and Gail Binder assisted with report production. Sandra Salazar was the EVS
Project Manager for this work.
VIII
-------
CHAPTER 1
EXECUTIVE SUMMARY
1.1 OVERVIEW
The Assessment and Remediation of Contaminated Sediments (ARCS) program, a 5-year study and
demonstration project relating to the control and removal of contaminated sediments from the Great
Lakes, is being coordinated and conducted by the U.S. Environmental Protection Agency's (EPA) Great
Lakes National Program Office (GLNPO). As part of the ARCS program, baseline human health risk
assessments have been performed at five Areas of Concern (AOCs) in the Great Lakes region (Crane,
1992a,b; 1993a,b; 1994). In addition, baseline aquatic (Passino-Reader, et al., 1995) and wildlife
(Mann-Klager, 1993) risk assessments have been prepared for the Buffalo River, NY.
In this report, exposure and risk assessment guidelines, developed for the EPA Superfund program,
have been applied to determine the comparative human health risks associated with direct and indirect
exposures to contaminated sediments in the lower Buffalo River under selected remedial alternatives.
These risks were estimated for noncarcinogenic (e.g., reproductive toxicity, teratogenicity, liver toxicity)
and carcinogenic (i.e., probability of an individual developing cancer over a lifetime) effects, based on
currently available data. In addition, noncarcinogenic risks to mink, resulting from the ingestion of PCB-
contaminated forage fish (carp), were estimated to give an indication of ecological risks to a piscivorous
species.
1.2 STUDY AREA
The Buffalo River AOC is located in Buffalo, NY in western New York State. The river flows from the
east and discharges into Lake Erie near the head of the Niagara River. The study area has a history of
water quality problems due primarily to point sources of contaminants (i.e., industrial and municipal
discharges). The extent of contamination in the Buffalo River led to the International Joint
Commission's (IJC) decision to designate this region as a Great Lakes AOC. In response, the New York
State Department of Environmental Conservation (NYSDEC) has completed one phase of a remedial
action plan (RAP) to identify and implement pollution abatement measures for the Buffalo River AOC
(NYSDEC, 1989).
High concentrations of heavy metals, polychlorinated biphenyls (PCBs), polynuclear aromatic
hydrocarbons (PAHs), and pesticides have been measured in different compartments of the Buffalo
River (e.g., sediments, water column, and fish). Fish advisories have been issued against consuming
carp from the Buffalo River because of excessive concentrations of PCBs. The transport of these
contaminants into Lake Erie is also of concern. However, it was beyond the scope of this risk
1-1
-------
assessment to estimate human health and wildlife risks to people using the nearshore areas of Lake
Erie.
1.3 EXPOSURE ASSESSMENT
1.3.1 Modeled Data
In this comparative risk assessment, a modeling exercise was carried out in which estimates of water
column and fish contaminant concentrations were made for different remedial alternatives developed
by the ARCS Risk Assessment and Modeling (RAM) work group. Remediation was based on upstream
and downstream reaches of the AOC and included the following scenarios: no action; no action, no
external loading; environmental dredging; no navigational dredging above Hamburg Cove; Hamburg
Cove scenario with no external loading. Water column concentrations were modeled for five chemicals
[polychlorinated biphenyls (PCBs), benzo(a)anthracene, benzo(a)pyrene, copper, and lead] under five
different remediation scenarios. PCBs were the only chemical of interest modeled in the fish
bioaccumulation study for carp. Carp were selected because they are a benthic feeder and have a high
lipid content. Thus, carp may readily accumulate sediment-derived contaminants through the ingestion
and assimilation of contaminated food.
1.3.2 Human Exposure
Water-contact and noncontact recreational activities are limited along the Buffalo River. No public
bathing facilities exist along the river, and fish consumption advisories have warned the public not to
eat carp from the river. However, there is anecdotal evidence that these activities occur, even near
industrial outfalls. This assessment focused on two complete pathways by which residents near the
lower Buffalo River could be exposed to sediment-derived contaminants: (1) consumption of
contaminated carp and (2) ingestion of surface water while swimming. A third complete pathway of
dermal exposure to surface water was assumed to be insignificant in comparison to the risk resulting
from the ingestion of contaminated surface water. This assumption was made because contaminants
are more easily transported across the gut than the skin. Data for other exposure pathways were
determined to be incomplete (e.g., ingestion of sediments).
Noncarcinogenic and carcinogenic risks to humans were estimated for typical, reasonable maximum,
and subsistence (fish only) exposure scenarios. Typical (i.e., average) exposures were assumed to
occur over a period of 9 years, whereas reasonable maximum (i.e., the maximum exposure that is
reasonably expected to occur at a site) and subsistence (i.e., reliance on fish as a major source of
protein) exposures were assumed to occur over a period of 30 years (USEPA, 1989a). These exposure
durations were extrapolated over a period of 70 years for estimating carcinogenic risks. Exposure
intakes were determined for each chemical and added for each exposure pathway.
1-2
-------
TABLE 1.1. AMOUNT OF CARP ASSUMED TO BE CONSUMED PER PERSON PER DAY FROM THE
BUFFALO RIVER FOR EACH EXPOSURE SCENARIO
Exposure Scenario
Typical
Reasonable Maximum
Subsistence
Ingestion
Rate*
(Q/day)
19.2
54
132
X Fl**
0.10
0.25
0.70
Amount of Buffalo R.
Carp Consumed per Day
(g/day)
1.92
13.5
92.4
* Sources: Typical (West et al., 1989); Reasonable Maximum and Subsistence (USEPA,
1991a)
** Fl = Fraction of fish (i.e., carp) estimated to be ingested from the Buffalo River (study
assumption).
For each of the exposure scenarios, different consumption patterns of fish were assumed to take place
(Table 1.1). These consumption patterns were based on recommended values given in U.S. EPA
Superf und guidance (USEPA, 1989a,b; 1991 a), on published literature values, or on study assumptions.
Based on an average meal of fish (i.e., 150 g or 0.33 Ib), the amount of Buffalo River fish consumed
for each exposure scenario (Table 1.1) can also be converted to meals per year using the following
equation:
Ingestion Rate (meals/yr) = [Ingestion Rate (g/day)l x Fl x (meal/150 g) x (365 days/yr)
Where Fl is the fraction of contaminated fish estimated to be ingested from the Buffalo River. The
number of meals of Buffalo River fish consumed over a year-long period for typical, reasonable
maximum, and subsistence exposures corresponded to approximately 4.5, 33, and 225 meals,
respectively.
1.3.3 Wildlife Exposure
The comparative wildlife risk assessment focused on one exposure pathway: the consumption of
contaminated forage fish by mink. Mink were considered due to their occurrence at Tifft Nature
Preserve in the AOC and by their inclusion in the baseline wildlife risk assessment (Mann-Klager, 1993).
Exposure intakes were not calculated for mink because a simple hazard quotient method was used to
screen for risks. A No-Observed-Adverse-Effect-Level (NOAEL) was obtained from the literature for
mink exposed to total PCBs through feeding studies. A NOAEL concentration of 0.069 (JQ/Q wet weight
was used in the draft baseline wildlife risk assessment (Heaton et al., 1991 cited in Mann-Klager,
1993). This same NOAEL concentration was used in this report.
1-3
-------
1.4 RISK ASSESSMENT
1.4.1 Determination of Human Health Risk
Noncarcinogenic effects were evaluated by comparing an exposure level over a specified time period
with a reference dose (RfD)1 derived from a similar exposure period [otherwise known as a hazard
quotient (HQ)]. Thus, HQ = exposure level/RfD. An HQ value of less than 1 indicates that exposures
are not likely to be associated with adverse noncarcinogenic effects. HQ values between 1 and 10 may
be of concern, particularly when additional significant risk factors are present (e.g., other contaminants
are present at concentrations of concern! (USEPA, 1989a). HQ values for multiple substances and/or
multiple exposure pathways were summed to yield an overall Hazard Index (HI). The His were
interpreted in the same fashion as the HQs. Summing the HQs did not account for any synergistic or
antagonistic effects that may occur among substances.
Carcinogenic risks were estimated as the incremental probability of an individual developing cancer over
a lifetime as a result of exposure to potential carcinogens. This risk was computed using average
lifetime exposure values that were multiplied by the oral slope factor2 for a particular chemical. The
resulting carcinogenic risk estimate generally represented an upper-bound estimate, because slope
factors are usually based on upper 95th percentile confidence limits. Carcinogenic effects were
summed for all chemicals in an exposure pathway. This summation of carcinogenic risks assumed that
intakes of individual substances were small, that there were no synergistic or antagonistic chemical
interactions, and that all chemicals caused cancer. The EPA believes it is prudent public health policy
to consider actions to mitigate or minimize exposures to contaminants when estimated, upper-bound
excess lifetime cancer risks exceed the 10"6 to 108 range, and when noncarcinogenic health risks are
estimated to be significant (USEPA, 1988a).
1.4.1.1 Consumption of Contaminated Fish
The consumption of PCB-contaminated carp resulted in significant noncarcinogenic risks (i.e.. Hazard
Quotient > 1) and carcinogenic risks li.e., risk greater than one person in a million (10"8)1 for all
remedial alternatives and fish consumption scenarios (Tables 1.2 and 1.3). The degree of risk increased
as local residents of the Area of Concern consumed more locally caught carp. For noncarcinogenic
risks, the Hazard Quotient values ranged from 1.7 to 4.2 for typical exposures, 5.6 to 19 for reasonable
maximum exposures, and 39 to 130 for subsistence exposures (Table 1.2). For carcinogenic risks, the
The RfD provides an estimate of the daily contaminant exposure that is not likely to cause harmful
effects during either a portion of a person's life or their entire lifetime (USEPA, I989a).
Slope factors are estimated through the use of mathematical extrapolation models for estimating
the largest possible linear slope (within 95% confidence limits) at low extrapolated doses that is
consistent with the data (USEPA, 1989a).
1-4
-------
TABLE 1.2. NONCARCINOGENIC RISKS ASSOCIATED WITH CONSUMING WHOLE CARP UNDER
VARIOUS REMEDIATION AND CONSUMPTION SCENARIOS
Remediation Scenario
UPSTREAM
No Action
No Action, No Load
Environmental Dredging .
Hamburg Cove
Hamburg Cove, No Load
DOWNSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
Noncarcinogenic Risk
Typical
-
2.5
2.4
1.7
2.1
1.8
4.2
4.1
1.7
4.2
4.1
RME
13
11
10
7.9
5.6
19
18
11
19
18
Subsistence
89
76
70
54
39
130
130
76
130
120
TABLE 1.3. CARCINOGENIC RISKS ASSOCIATED WITH CONSUMING WHOLE CARP UNDER
VARIOUS REMEDIATION AND CONSUMPTION SCENARIOS
Remediation Scenario
UPSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
DOWNSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
Lifetime Cancer Risk
Typical
4.9E-05
4.7E-05
3.4E-05
4.2E-05
3.6E-05
8.3E-05
8.1E-05
3.4E-05
8.3E-05
8.1E-05
RME
8.5E-04
7.3E-04
6.7E-04
5.2E-04
3.7E-04
1.3E-03
1.2E-03
7.3E-04
1 .3E-03
1.2E-03
Subsistence
5.8E-03
5.0E-03
4.6E-03
3.6E-03
2.5E-03
8.8E-03
8.4E-03
5.0E-03
8.8E-03
7.9E-03
1-5
-------
risks ranged from 3.4 x 10"B to 8.3 x 10"6 for typical exposures, 3.7 x 10"4 to 1.3 x 103 for reasonable
maximum exposures, and from 2.5 x 10~3 to 8.8 x 103 for subsistence exposures (Table 1.3). A
greater degree of risk was observed in the downstream remediation scenarios than the upstream ones.
1.4.1.2 Ingestion of Contaminated Surface Water
The noncarcinogenic and carcinogenic risks resulting from the ingestion of surface water while
swimming under typical and reasonable maximum exposure scenarios were estimated to be far below
levels of concern for the environmental dredging scenario. These risks were estimated based on a
modeled data set for copper, lead, benzo(a)anthracene, benzo(a)pyrene, and PCBs. Noncarcinogenic
risks ranged from 0.000006 to 0.00001 for typical and reasonable maximum exposures, respectively.
Lifetime cancer risks were on the order of 10"10 for both scenarios. Lower risks could be expected for
the other remedial alternatives which had equal or lower contaminant concentrations.
Based on these estimated risk values, an assumption was made that insignificant risks would also result
from dermal exposure to surface water while swimming. This assumption was made because the direct
intake of contaminants into the gut generally results in a greater intake of contaminants than the
absorption of contaminants (with varying capacity for penetration) through the skin.
1.4.2 Determination of Wildlife Risk
Wildlife risk was assessed for a representative piscivorous species, mink. A simple hazard quotient
method was used to compare the modeled PCB concentration in carp, a representative forage fish, to
the target forage concentration (i.e., the NOAEL value). The NOAEL was exceeded by 19 to 46 times
for the various remedial alternatives. The estimated degrees of exceedance were very protective
because an assumption was made that local mink populations consumed 100% of their diet from
contaminated carp.
1.5 UNCERTAINTIES
Several assumptions and estimated values were used in this comparative risk assessment that
contributed to the overall level of uncertainty associated with the noncarcinogenic and carcinogenic risk
estimates. One of the greatest sources of uncertainty in this risk assessment arose from the use of
modeled data. Another large source of uncertainty was using modeled data for whole, raw carp in the
exposure assessment for human health; PCB concentrations could be greatly reduced in fish by
trimming off the fat and skin and cooking the fish. As with most environmental risk assessments, the
uncertainty of the risk estimates probably ranges over an order of magnitude or greater. On a
comparison basis, this risk assessment exercise was useful in judging the adequacy of different remedial
alternatives in reducing risk to human health and wildlife.
1-6
-------
CHAPTER 2
INTRODUCTION
2.1 BACKGROUND
The 1987 amendments to the Clean Water Act, in Section 118(c)(3), authorized the U.S. Environmental
Protection Agency's (EPA) Great Lakes National Program Office (GLNPO) to coordinate and conduct
a 5-year study and demonstration project relating to the control and removal of contaminated sediments
from recommended areas in the Great Lakes region (USEPA, 1991b). To achieve this task, GLNPO
initiated the Assessment and Remediation of Contaminated Sediments (ARCS) program. A number of
agencies were involved in conducting toxicology, chemistry, risk assessment, modeling, and
engineering studies at the following Areas of Concern (AOCs): Ashtabula River, OH; Buffalo River, NY;
Grand Calumet River/Indiana Harbor Canal, IN; Saginaw River, Ml; and Sheboygan River, Wl. The
ARCS program was extended by one year by the Great Lakes Critical Programs Act of 1990.
This report will present the results of a comparative risk assessment for the Buffalo River AOC.
Comparative risk assessments are used to estimate and compare the risks that may be associated with
various remedial alternatives (including the "no action" alternative). A mass balance modeling approach
is used to estimate how contaminant concentrations in the sediment, water column, and biota will vary
with remedial alternative. In turn, the exposure and subsequent risk to humans and biota from
contaminants are estimated.
The Buffalo River AOC has a history of water quality problems due primarily to point sources of
contaminants (i.e., industrial and municipal discharges). The extent of contamination in this area led
to the International Joint Commission's (UP decision to designate this region as a Great Lakes AOC.
In response, the New York State Department of Environmental Conservation (NYSDEC) has completed
one phase of a remedial action plan (RAP) to identify and implement pollution abatement measures for
the Buffalo River AOC (NYSDEC, 1989).
High concentrations of heavy metals, polychlorinated biphenyls (PCBs), polynuclear aromatic
hydrocarbons (PAHs), and pesticides have been measured in different compartments of the Buffalo
River (e.g., sediments, water column, and fish). Fish advisories have been issued against consuming
carp from the Buffalo River because of high concentrations of PCBs. The benthic macroinvertebrate
community has been adversely affected as the fauna are dominated by pollutant-tolerant oligochaetes
and some invertebrates frequently display abnormal mouthparts (Diggins and Stewart, 1993). These
observations generated concern that organisms higher up in the food chain may be exposed to
unhealthy doses of contaminants.
2-1
-------
2.2 BASELINE HUMAN HEALTH RISK ASSESSMENT
As part of the ARCS Risk Assessment and Modeling (RAM) work group activities, a baseline human
health risk assessment was conducted for an area adjacent to the lower Buffalo River (Figure 2.1). The
purpose of this risk assessment was to determine current risk conditions to human health. Baseline risk
estimates were determined for both noncarcinogenic effects (chronic or subchronic) and carcinogenic
effects (i.e., the probability of an individual developing cancer over a lifetime) resulting from direct and
indirect exposures to sediment-related contaminants (Crane, 1993a). These risk estimates were made
using conservative assumptions about exposure scenarios when complete data were not available.
Thus, the risk estimates were designed to be overprotective of human health. The baseline risk
assessment focused on two complete pathways by which residents of the lower Buffalo River could
be exposed to sediment-derived contaminants: (1) consumption of contaminated carp and
spottail/emerald shiners, and (2) ingestion of surface water while swimming. Carp data were used in
the exposure assessment, despite the fish advisories, because these fish are benthic feeders and
accumulate contaminants more readily than pelagic species of fish. Thus, the consumption of carp
provided a conservative estimate of risk to anglers and their families.
The results of the baseline human health risk assessment are summarized in Appendix A. The ingestion
of surface water did not result in a significant risk to swimmers. However, the consumption of carp
represented a noncarcinogenic risk to subsistence anglers that consumed over 225 meals of fish per
year. The consumption of either carp or spottail/emerald shiners represented a cancer risk (i.e., greater
than one in a million) to anglers and their families that consumed 4.5, 33, or 225 meals of fish per year.
As given in Appendix A, only a proportion of these meals of fish were composed of contaminated fish
from the Buffalo River.
2.3 BASELINE WILDLIFE RISK ASSESSMENT
The draft baseline wildlife risk assessment for the Buffalo River AOC (Mann-Klager, 1993) focused on
one exposure pathway: the consumption of contaminated forage food by piscivorous wildlife. Common
terns (Sterna hirundo), a colonial waterbird, and mink (Mustela vison) were selected for the wildlife risk
assessment. Common terns were considered due to the availability of data for this species in the upper
Niagara River and their documented contaminant sensitivity and usefulness as bioindicators [Hays and
Risebrough, 1972; Fasala et al., 1987; and Karwowski, 1991 (references cited in Mann-Klager, 1993)1.
In addition, Gilbertson (1974) concluded that the organochlorines found in common terns were ingested
within the vicinity of the breeding colony. Karwowski (1991) observed that there was a large
percentage of tern chicks that died while pipping (i.e., trying to break open the shell of their egg during
hatching) in the upper Niagara River. Mink were considered due to their occurrence at Tifft Nature
Preserve. Although the population density of mink at Tifft has not been determined, sightings are not
uncommon (Landsittel, personal communication cited in Mann-Klager, 1993). Foley et al. (1988)
concluded that PCB and DDE concentrations in minks were influenced by local sources of the
contaminants.
2-2
-------
Lake Ontario
Buffalo River
Area of Concern Map
Figure 2.1. Map of Buffalo River Area of Concern (NYSDEC, 1989).
2-3
-------
The selection of these species was supported by the Biological Effects Subcommittee to the
International Joint Commissions's (UC) Science Advisory Board. This subcommittee is evaluating the
use of the bald eagle (Haliaeetus leucocephalus], mink or river otter (Lutra canadensis), colonial
waterbirds, and lake trout (Salvelinus namaycush) as bioindicators of Great Lakes water quality (Kubiak
and Best, 1991). Bald eagles were not considered in the draft baseline risk assessment as their
occurrence in the area is presently limited to occasional transients. Historic eagle breeding occurred
at the northern end of Grand Island on the Niagara River. River otters were also not selected because
they are presently rare in the Niagara River area (Newell et al., 1987).
Wildlife risks were estimated for noncarcinogenic (e.g., reproductive toxicity) effects using the simple
hazard assessment model developed by Kubiak and Best (1991). The model uses the lowest observed
adverse effect levels (LOAEL) and no observed adverse effect levels (NOAEL) of contaminants to assess
wildlife health. Preliminary results from the baseline risk assessment indicated the possibility of
detrimental effects to common terns and mink. However, a number of assumptions were made in the
baseline risk assessment which increased the level of uncertainty with the risk estimates.
2.4 COMPARATIVE RISK ASSESSMENT APPROACH
Another objective of the ARCS RAM work group was to use a mass balance approach at two AOCs
(i.e., Buffalo River, NY and Saginaw River, Ml) to address management questions concerning the
remediation of contaminated sediments. These two AOCs were chosen based upon anticipated impacts
from sediments, lack of other on-going activities (such as Superfund remedial activities), and lack of
complicating factors (such as complicated groundwater/surface water interactions, multiple sources of
contaminant inputs, etc.). The mass balance approach involved an evaluation of the sources, transport,
and fate of contaminants in the system. This process follows the law of conservation of mass and
requires that the quantities of contaminants entering the system, less quantities stored, transformed,
or degraded in the system, must equal the quantities leaving the system.
ARCS RAM participants developed and applied Level 1 (preliminary) models for understanding and
predicting the transport and fate of contaminated sediments and the bioaccumulation of persistent
sediment contaminants in the affected AOC. The RAM work group used a comparative risk assessment
approach (described in Chapter 3) to integrate the results from the baseline risk assessment, field, and
mass balance modeling studies to provide estimates of the potential impact of remedial actions on
human health and wildlife. Thus the risk, relative to the baseline risk, that would result from the
implementation of various sediment remedial alternatives could be evaluated.
The framework for the comparative risk assessment approach is discussed in the following chapter.
Subsequent chapters provide the exposure assessment, toxicity assessment, risk characterization, and
uncertainty sections of the comparative risk assessment for human health and wildlife at the Buffalo
River AOC.
2-4
-------
CHAPTER 3
COMPARATIVE RISK ASSESSMENT FRAMEWORK
3.1 OVERVIEW
The comparative risk assessment framework used by the ARCS RAM work group is given in Figure 3.1.
The framework (USEPA, 1993a) was developed to: 1) identify existing risks to human health and
ecological receptors at sites with contaminated sediments, 2} estimate the potential impact of various
sediment remedial alternatives on contaminant concentrations in various media and their associated
risks, and 3) compare existing and potential future risks to aid in the selection of sediment remedial
alternatives. This framework can be applied to other contaminated sediment sites and should be
viewed as a demonstration of the steps one would take to perform a comparative risk assessment.
The main components of the framework are listed below. The components are based on guidance
given in USEPA (1988b, 1989a, 1993a) and on discussions generated from the ARCS RAM work
group.
• Characterization of the Area of Concern including:
collection of existing data to determine extent of contamination in the key
sources and media of interest
collection of background data
examination of QA/QC measures
determination of potential contaminants of concern
identification of contaminant sources, especially information related to release
potential
characterization of the environmental setting that may affect the fate, transport,
and persistence of contaminants.
• Evaluation of available data including:
assessment of data quality
identification of data gaps.
• Implementation of field studies including:
collection of contaminant data in the media of interest to fill data gaps for risk
assessment and modeling components
3-1
-------
Site Characterization
Mass Balance
Modeling Studies
Evaluation
o< Available Data
Field Studies
Baseline Risk Assessment
(human health, wildlife,
aquatic organisms)
Human Health
COMPARATIVE RISK ASSESSMENT
Wildlife
en
B C No
Action
REMEDIAL
ALTERNATIVE
(A
A B C No
Action
REMEDIAL
ALTERNATIVE
Aquatic Organisms
co
A B C No '
Action '
REMEDIAL !
ALTERNATIVE ,
Selection and
Implementation of
Remedial Alternative
Figure 3.1.
Comparative risk assessment in the risk management process (USEPA, 1993a).
3-2
-------
considerations for sampling including:
• sample size
• sampling locations
• types of samples (e.g., grab or composite, whole fish or fillet)
• temporal and meteorological factors
• field screening analyses
• time and cost of sampling
• QA/QC procedures.
Evaluation of baseline risks to human health, wildlife, and aquatic organisms including:
identification of receptors of concern
determination of exposure pathways and exposure intake concentrations
identification of toxicity values for contaminants of concern
characterization of risks
estimation of uncertainties associated with risk estimates.
Implementation of mass balance modeling studies including:
identification of management questions, such as:
• How long will it take for contamination to be reduced to an acceptable
level through natural processes (e.g., sedimentation)?
• What will happen if the sediments are dredged to a particular depth?
• Will the sediments become recontaminated following remediation?
selection of contaminants of concern
selection of sediment remedial alternatives
development of screening models
determination of appropriate scales
implementation of modeling framework including:
• hydrodynamic modeling
• sediment transport modeling
• contaminant transport and fate modeling
• food chain modeling
calibration/verification of models.
3-3
-------
Evaluation of comparative risks to human health, wildlife, and aquatic life including:
adaptation of the baseline risk assessment framework
• site characterization
• exposure assessment
• toxicity/hazard assessment
• risk characterization
integration of the modeling results in the exposure assessment
derivation of risk values for selected remedial alternatives
comparison of risk values
evaluation of uncertainty.
Selection and implementation of remedial alternative(s).
A multidisciplinary team is needed to carry out the aforementioned components of the framework. It
is very important before work commences on a comparative risk assessment that the team agrees on
the management questions to be addressed, the level of effort necessary to complete work, and the
timelines for completion of tasks. Regular meetings are essential to evaluate progress and to reassess
priorities. Interested stakeholders should be informed of progress/results of the comparative risk
assessment through an effective risk communication effort.
The results of the comparative risk assessment are evaluated by the risk manager. The risk manager
must also take into consideration social, economic, and political factors which would influence the
selection of a remedial option. A final remedial action plan can then be developed and implemented.
For additional information on the comparative risk assessment framework, refer to the "Risk
Assessment and Modeling Overview Document" (USEPA, 1993a) completed for the ARCS program.
Specific details on the methodology used for the ARCS baseline human health risk assessments are
given by Crane (1992a,b; 1993a,b; 1994). These assessments followed exposure and risk assessment
guidelines established by the U.S. EPA for use at Superfund sites (USEPA, 1988b; 1989a,b; 1991 a).
The following sections describe the activities of the modelers from the RAM work group for the Buffalo
River AOC. This modeling effort was done as a demonstration project. Therefore, the same approach,
level of effort, and data needs pertinent to the Buffalo River AOC may not apply to other contaminated
sites.
3-4
-------
3.2 MASS BALANCE MODELING
3.2.1 Selection of Remedial Action Scenarios
An important component of the comparative risk assessment was the use of modeled data to estimate
changes in contaminant concentrations in sediment, water, and fish tissue based on different
remediation scenarios. The RAM work group selected the following remedial action scenarios for the
Buffalo River AOC, of which the first three were the most important (DePinto et al., 1994).
1. No Action Scenario. This scenario focused on the system response over time under
existing external loadings and continued navigational dredging. No additional actions
on the river were simulated.
2. Hamburg Cove Scenario. This scenario examined the impacts of discontinuing
navigational dredging above Hamburg Cove (about halfway from the river mouth to the
AOC upstream boundary). This scenario permitted this portion of the river to fill in with
"clean" sediments from upstream. The potential for flooding existed as a result of this
option.
3. Environmental Dredging Scenario. This scenario examined the impact of nearshore
dredging along the entire river within the AOC. This option would remove several "hot
spots" along the banks.
In order to determine the importance of resuspension on water column contamination, Scenarios 1 and
2 were also evaluated with no external loadings. All other factors were kept the same. Thus, any
contamination of the water column would be strictly from sediment resuspension.
4. No Action - No Loading Scenario.
5. Hamburg Cove - No Loading Scenario.
Two additional scenarios were modeled by DePinto's group (DePinto et al., 1994) to aid in the
interpretation of the other remedial action scenarios.
6. Zero Initial Conditions Scenario. Similar to Scenario 3, the initial conditions in the top
two layers of sediments (depositional and erosional) were set to zero contaminant
concentrations. This effectively nullified currently contaminated sediments as a source
of water column contamination. Thus, the sole impact would be from external loading.
7. Flow Switching Scenario. Two years of actual flow data were switched with each
other to evaluate the effect on cumulative export and concentrations in the no action
3-5
-------
scenario. The flows from 1978-79, which contained several high flow events, were
switched with those from 1970-71, which had no events, and vice versa. The results
showed the importance of the sequence of flow events in altering the final results.
3.2.2 Mass Balance Modeling Framework
The application of the mass balance modeling approach involved the quantification of the sources,
transport, and fate of contaminants (Figure 3.2). Specific detail on the components of mass balance
models can be found in the ARCS "Risk Assessment and Modeling Overview Document" (USEPA,
1993a). The typical steps in a mass balance modeling study are to:
• predict water and sediment transport
• use the predicted transport, along with estimates of contaminant loadings from point
and nonpoint sources, to estimate the changes in chemical concentrations in water and
sediments
• use the predicted contaminant concentrations in water and sediments to estimate the
transfer of contaminants through the food chain and their accumulation in fish.
A mass balance modeling framework (using a modified version of WASP4/TOXI4) was developed for
use as a management tool for the Buffalo River AOC (DePinto et al., 1994). Contaminant exposure and
benthic food chain bioaccumulation were analyzed based on the response to various remedial
alternatives. The model also simulated the export from the Buffalo River to Lake Erie resulting from
these remediation actions.
Joseph DePinto of the State University of New York-Buffalo (SUNY-Buffalo) led most of the modeling
work for the Buffalo River comparative risk assessment (Atkinson et al., 1994; DePinto et al., 1994).
A different application of sediment and contaminant transport was done by Joseph Gailani and
coworkers (Gailani et al., 1994) to predict the transport of fine-grained sediments and associated
contaminants in the Buffalo River. Gailani's team was especially interested in the transport of these
sediments from the river to Lake Erie. The SUNY-Buffalo studies are briefly described below. Each of
these studies made use of data collected from field sampling programs conducted for the ARCS
program. Refer to the modeler's reports for a description of their data needs and sources of data. The
only modeled data used in this comparative risk assessment came from the SUNY-Buffalo modelers.
3.2.2.1 SUNY Modeling Effort
The modeling work by DePinto's group involved several components. One component was to develop
estimates of mass loading rates for the following contaminants: total PCBs, chlordane, dieldrin, DDT,
benzo(a)anthracene, benzo(b)fluoranthene, benzo(k)fluoranthene, benzo(a)pyrene, chrysene, lead and
3-6
-------
Contaminant Transport
Figure 3.2. Components of the mass balance modeling study (USEPA, 1993a).
3-7
-------
copper (Atkinson et al., 1994). The aforementioned contaminants were chosen based on fish
advisories, concerns cited in the Remedial Action Plan, and results obtained from Toxicity Identification
Evaluation work (USEPA, 1994). Total suspended solids (TSS) loading was also calculated. The annual
loading calculations indicated relatively small loadings for most of the contaminants of interest.
The major source for all the contaminants of interest was found to be the upstream tributary flows
(Atkinson et al., 1994). Upstream loadings were calculated on the basis of average daily flows and TSS,
along with measured contaminant concentrations. Groundwater and combined sewer overflow (CSO)
loadings were estimated on the basis of separate model calculations and industrial loadings taken from
the Buffalo River Remedial Action Plan (NYSDEC, 1989). Results were presented for use in water
quality mass balance models which could be used to simulate the time history of contaminant
concentrations in the water column, sediments, and biota of the river as a function of source inputs.
This information was used to evaluate system response to selected remedial action scenarios.
Another component of this work was to calculate several parameters needed to develop and apply
general water quality and contaminant transport models to the river (Atkinson et al., 1994). These
parameters included primary distribution (partition) coefficients for each of the contaminants of interest,
as well as data for a number of conventional parameters. Annual and monthly average flows were
presented and data were provided for specifying upstream and downstream boundary conditions.
Additional modeling was done by DePinto et al. (1994) to develop water quality mass balance and
bioaccumulation models for the Buffalo River by simulating a time-history of contaminant concentrations
in the water column, sediments, and biota of the river as a function of source inputs. The overall
modeling framework followed Figure 3.2, and a modified version of WASP4/TOXI4 was used for the
contaminant mass balance model. The contaminants that were modeled included total PCBs,
benzo(a)anthracene, benzo(a)pyrene, copper, and lead. The parameterization, segmentation, and
calibration of the models are described in detail in DePinto et al. (1994).
The following conclusions were made by DePinto et al. (1994):
• On days of average or low flow, resuspension of contaminated sediment was not a
significant factor in water column concentrations.
. • Sediment remediation will not have a significant impact on reducing water column
contaminant exposure. Environmental or full dredging of bottom sediments will not
alleviate water column concerns for the five chemicals examined. Also, the potential
to exacerbate the water column problem still exists with these dredging options by
exposing higher contaminated sediments in deeper layers.
3-8
-------
• Sediment remediation will be a potentially important action for reducing direct sediment
exposure, especially in "hot spots." Environmental dredging of nearshore "hot spots"
could be beneficial to the benthic community and corresponding food web.
• The contaminant body burdens of bottom-dwelling and bottom-feeding organisms, such
as carp, will improve in response to sediment remediation actions. On a river-wide
basis, environmental dredging in the nearshore depositional areas lead to the largest
reduction in PCB concentrations in carp. However, the cessation of navigational
dredging above Hamburg Cove proved to be the best alternative for that portion of the
modeled river.
3.2.2.2 Limitations of Modeling Effort
Specific limitations and data needs noted by the SUNY modelers (Atkinson et al., 1994; DePinto et al.,
1994) included:
• Year-round TSS data were not available, especially during high flow events, for
calculating loading estimates.
• Uniform sample collection and analytical protocols were not applied by all groups
involved with data collection.
• Finer resolution of vertical profiles in sediment cores and a more uniform distribution of
the horizontal location of sediment cores are needed to improve quantification of initial
conditions for model runs.
• A more accurate description of erosional/depositional areas of the river would enhance
the model simulation.
• Sediment transport could be characterized more accurately if deposition rates and other
physical and chemical properties of resuspended and upstream sediments were
measured as a function of flow.
• Profiles of sediment concentration data before and after dredging would be useful to
examine the effect of sediment sloughing.
The water column and carp data from the modeling effort of DePinto et al. (1994) were used in the
exposure assessment described in the following chapter. The uncertainty associated with these data
cannot be quantitatively calculated but will be qualitatively assessed in Chapter 8.
3-9
-------
CHAPTER 4
EXPOSURE ASSESSMENT
4.1 EXPOSURE PATHWAYS
In this exposure assessment, the magnitude, frequency, duration, and route of direct and indirect
exposures of people to sediment-derived contaminants from the Buffalo River AOC will be determined
for different remediation scenarios. In addition, the exposure to mink residing within the AOC will be
evaluated.
4.1.1 Human Exposure Pathways
Human exposure to contaminants in the Buffalo River AOC can potentially occur via three pathways:
dermal contact, inhalation, and ingestion. Dermal contact involves direct contact of the skin with either
contaminated sediments, riverplain soils, or overlying water. Inhalation of airborne vapors or dust may
introduce chemicals of potential concern into the respiratory system. Ingestion of contaminants through
the consumption of contaminated soils, sediment, or food (e.g., fish) is potentially significant because
of the direct transfer of contaminants across the gut.
The potential pathways by which people may be exposed to contaminants from the Buffalo River AOC
were given in the baseline risk assessment (Crane, 1993a). Although four exposure pathways were
considered complete (i.e., exposure could occur through each of those routes) in the baseline risk
assessment (Table 4.1), not all of those exposure pathways may result in substantial human health
risks. In addition, for humans inhaling airborne contaminants, it would be difficult to separate out the
contributions of contaminants from the river (if any) and those from industrial, municipal, and
background sources. Thus, although the inhalation exposure pathway may be complete, the currently
available data set for atmospheric contaminants in the Buffalo River AOC are inadequate to
quantitatively assess the risks to human health.
TABLE 4.1. COMPLETE EXPOSURE PATHWAYS IN THE BUFFALO RIVER AOC
• Ingestion of Contaminated Fish
• Ingestion of Surface Water while Swimming or Playing in the Water
• Dermal Contact with Water while Boating, Fishing, Swimming, Water Skiing, etc.
• Inhalation of Airborne Contaminants
4-1
-------
In the baseline risk assessment, the only substantial risk resulted from the consumption of fish (Crane,
1993a). For this comparative risk assessment, it was recognized that contaminant concentrations in
the water might change enough under different remediation scenarios to pose a potential risk to people
swimming in the river. Thus, an exposure assessment was conducted for swimmers ingesting small
amounts of surface water under the environmental dredging scenario; this scenario was selected
because it resulted in the highest estimated surface water contaminant concentrations. If this risk was
determined to be substantial, then the ingestion of surface water pathway would be examined for all
remediation scenarios. However, the risk was insignificant for both carcinogenic and noncarcinogenic
effects (see Chapter 6).
Dermal exposure to surface water was judged to result in an insignificant risk based on the low
frequency with which these exposures would take place and because the direct intake of contaminants
into the gut is usually greater than the absorption of contaminants (with varying capacities to penetrate)
across the skin interface. Thus, risks from dermal exposure should be less than those from ingesting
surface water.
The only complete exposure pathways considered for this risk assessment were the consumption of
fish (for all remedial action scenarios) and ingestion of surface water while swimming (for the
environmental dredging scenario). Noncarcinogenic and carcinogenic risks were determined for both
typical (i.e., average) and reasonable maximum exposures (i.e., the maximum exposure that is
reasonably expected to occur at a site). In addition, risks were calculated for subsistence anglers that
relied on the consumption of fish for their main source of protein. The subsistence exposure scenario
was chosen because of economic problems in the area which might contribute to an
underemployed/unemployed person consuming large amounts of locally caught fish.
4.1.2 Wildlife Exposure Pathways
At least twelve wildlife species in the Great Lakes basin have experienced reproductive or other
problems and/or population decreases since the 1960s that have been associated with chemical
contaminants (Government of Canada, 1991 cited in Fox, 1993). The list includes two mammals, nine
species of birds, and one reptile. All are long-lived, fish-eating species. A number of studies in the
Great Lakes area have been conducted to study impairments to wildlife using such biomarkers as
induction of mixed function oxidases, alterations in heme biosynthesis, retinol homeostasis, thyroid
function, DMA integrity, and various manifestations of reproductive and developmental toxicity (Fox,
1993).
Several species of biota have been used as biomonitors of general ecosystem health in the Great Lakes.
In 1975, the Ontario Ministry of the Environment adopted the use of young-of-the-year spottail shiners
(Notropis hudsonius) as biomonitors for the nearshore waters of the Great Lakes (Suns et al., 1993).
Contaminant residue data from spottail shiner surveys have been used to identify areas of concern for
contaminant inputs and trend assessment over time. Results of annual collections of these fish indicate
4-2
-------
that, at most sites sampled, PCB residues continued to decline during the 1980s. Shiners are an
important forage fish in the Great Lakes, and contaminant concentrations in the fish could
cause/contribute to detrimental effects in other biota, such as waterfowl.
Colonial fish-eating birds have been used to study the impact of chronic exposure to complex mixtures
of hydrophobic organic contaminants (HOCs) within the Great Lakes ecosystem (Fox et al., 1991 a).
Fox et al. (1991 a) suggest that double-crested cormorants (a totally piscivorous species) be used as
a biomarker due to their long life, abundance, and mostly ground nesting. The occurrence of bill
malformations in double-crested cormorant chicks have been documented from colonies in Green Bay
and elsewhere in the Great Lakes and in reference areas off the Great Lakes, in the years 1979 through
1987 (Fox et al., 1991b). Although the severity of waterfowl impairment effects have generally
decreased between the early 1970s and late 1980s, these studies confirm the continued presence of
sufficient amounts of PCBs and other HOCs in forage fish to cause physiological impairments in these
birds over much of the Great Lakes basin (Fox, 1993). Tillitt et al. (1992 cited in Fox, 1993) suggest
that PCBs are the major contaminant influencing cormorant reproductive success in the Great Lakes.
Fox and coworkers (1991b) suggest that monitoring reproductive outcomes in fish-eating birds is a
cost-effective and sensitive method of detecting biologically significant concentrations of developmental
toxins in Great Lakes fish.
In terms of estimating risks to piscivorous wildlife at the Buffalo River AOC, the baseline risk
assessment looked at two species: mink and common terns (Mann-Klager, 1993). Mann-Klager
assumed that the majority of contaminant uptake by these species would be from the consumption of
fish. For the comparative risk assessment, the ingestion of fish pathway was the only exposure
pathway examined. Mink was the only piscivorous wildlife species considered here, because carp was
the only forage fish for which contaminant concentrations were modeled for different remedial action
scenarios. Common terns were not included in this exposure assessment because they would not be
consuming carp. Instead, they are shoreline feeders that would be consuming small fish such as
minnows and shiners.
The majority of habitat available for wildlife utilization within or near the AOC is at Tifft Nature Preserve
and Times Beach confined disposal facility (CDF) (Figure 4.1). The Times Beach CDF was constructed
by the Buffalo District U.S. Army Corps of Engineers in 1971 for the containment of dredged materials
from the Federal navigation channel. The CDF is approximately 19 ha in size and is located west of
the mouth of the Buffalo River. The site received dredged material from 1972 to 1976. In response
to a request by the Buffalo Ornithological Society, the site was designated a native preserve and left
only partially filled (Stafford et al., 1991 cited in Mann-Klager, 1993).
Little wildlife habitat is available along the river, outer harbor, and Erie basin due to shoreline
development. Included in Appendix B is a list of endangered, threatened, and special concern species
of New York State observed near the Buffalo River AOC as well as wildlife species observed at the
Times Beach CDF and Tifft Nature Preserve.
4-3
-------
Figure 4.1. Location of Times Beach Confined Disposal Facility (NYSDEC, 1 989).
4-4
-------
4.2 MODELED DATA USED IN THE EXPOSURE ASSESSMENT
4.2.1 Carp Data
Three different age classes of carp (i.e., young, middle, and old) were collected from the Buffalo River
by the Great Lakes Laboratory, State University College at Buffalo. The fish were collected specifically
for the ARCS program. Each age class included three composite samples of five whole fish. The
samples were received on November 13, 1991 by Battelle, Pacific Northwest Division, and were
analyzed for eighty individual PCB congeners and eight chlorinated pesticides. Total PCB
concentrations are given in Table 4.2. The samples were analyzed using Battelle Standard Operating
Procedures (SOP) MSL-042 and MSL-044. The data underwent a QA/QC review by Lockheed
Engineering and Sciences Company (Lockheed-ESC) under a contract with the EPA Environmental
Monitoring Systems Laboratory in Las Vegas, NV. The collection and analysis of these carp had to
comply with a detailed QA/QC plan, and these data have been approved for use by the ARCS program
(USEPA, 1994). The results were reported as ng/g on a dry weight basis and were converted to wet
weight for use in the exposure assessment.
The carp data were used by DePinto et al. (1994) to calibrate a bioaccumulation model which had
originally been developed for Green Bay, Lake Michigan (Connolly et al., 1992). The model describes
the major features of the predator-prey relationship, seasonal movement of fish, and species
bioenergetic parameters, as well as the dependency of the transfer of PCBs across the gill and gut of
the animals on water column and sediment PCB concentrations (Connolly et al., 1992). It used the
estimates of water column and sediment concentrations from the physical-chemical mass balance model
to compute the time-variable concentration of PCBs in benthic biota of the Buffalo River. From the
latter, the food chain accumulation of PCBs in carp could be estimated for each remedial alternative.
Refer to DePinto et al. (1994) for a more detailed explanation of the assumptions used for the various
parameters in the model.
Ten year predictive runs were chosen for model simulations of each scenario. Navigational dredging
was simulated every two years when estimating sediment and water column concentrations.
Bioaccumulation modeling was done for the following scenarios:
1. no action
2. Hamburg Cove
3. environmental dredging
4. no action/no load
5. Hamburg Cove/no load.
The scenarios were run using modeled water column and sediment data obtained for selected segments
of the river located upstream and downstream of the AOC boundary. Bioaccumulation results for each
scenario are listed in Tables 4.3 and 4.4. The values shown represent mean PCB concentrations
4-5
-------
TABLE 4.2. AGE CLASS DATA ON CARP COLLECTED FROM THE BUFFALO RIVER*
Age Class
Age,
Years
Wet Weight
(kg)
PCB Concentration
\jjQlg wet weight)
YOUNG AGE CLASS
BRF Y W-1
BRF Y W-2
BRF Y W-3
Mean
4.2
4.0
4.6
4.3
0.944
0.972
0.927
0.948
1.89
1.8
2.2
1.96
MIDDLE AGE CLASS
BRF M W-1
BRF M W-2
BRF M W-3
Mean
6.4
6.0
5.4
5.9
1.633
1.667
1.61
1.637
2.76
2.34
3.7
2.93
OLD AGE CLASS
BRF 0 W-1
BRF 0 W-2
BRF 0 W-3
Mean
10.0
10.8
10.0
10.3
4.552
4.257
4.45
4.42
5.9
3.1
3.4
4.13
* Each age class has three groups associated with it, each group containing five fish. Values shown
are mean values for the five fish (Irvine et al., 1992 cited in DePinto et al., 1994).
of the young, middle, and old age classes of carp used in the model (DePinto et al., 1994). The
upstream scenario data from Table 4.3 are plotted in Figure 4.2, whereas the downstream scenario data
from Table 4.4 are plotted in Figure 4.3.
In the upstream scenario (Figure 4.2), environmental dredging resulted in the most rapid decline of PCBs
in carp tissue until stabilizing at approximately 1 mg/kg after Day 1280. The Hamburg Cove, no load
scenario resulted in the greatest overall decline of PCBs to 0.2 mg/kg in carp at Day 3660. The no
action scenario resulted in the highest PCB concentrations in carp, ending at 1.1 mg/kg on Day 3660.
4-6
-------
TABLE 4.3. MODELED PCB CONCENTRATIONS IN CARP FOR ALL UPSTREAM SCENARIOS
(DEPINTO ET AL., 1994)
Day
10
190
370
550
730
920
1100
1280
1470
1650
1830
2020
2200
2380
2560
2750
2930
3110
3300
3480
3660
Average
S.D.*
Modeled PCB Concentrations (mg/kg wet weight) in Carp
No Action
2.82
2.74
2.89
2.76
2.90
2.62
2.41
2.11
2.05
1.83
1.81
1.64
1.65
1.49
1.43
1.26
1.24
1.16
1.20
1.11
1.12
1.9
0.65
No Action,
No Load
2.82
2.71
2.84
2.70
2.82
2.53
2.31
1.99
1.89
1.66
1.63
.45
.44
.29
.21
.04
.01
0.934
0.964
0.891
0.895
1.8
0.72
Environmental
Dredging
2.80
2.24
2.01
1.65
1.52
1.26
1.15
1.07
1.12
1.04
1.08
1.02
1.06
1.00
0.994
0.916
0.927
0.914
0.982
0.950
0.983
1.3
0.49
Hamburg
Cove
2.82
2.74
2.89
2.76
2.90
2.61
2.39
1.96
1.78
1.49
1.41
1.19
1.15
0.977
0.908
0.769
0.738
0.617
0.587
0.502
0.479
1.6
0.89
Hamburg Cove,
No Load
2.82
2.71
2.84
2.70
2.82
2.52
2.29
1.82
1.60
1.30
1.18
0.957
0.894
0.722
0.635
0.500
0.454
0.347
0.307
0.236
0.208
1.4
0.98
*S.D. = standard deviation
Note that at Day 3660, the final range of PCB concentrations for all scenarios was fairly narrow
(0.2- 1.1 mg/kg).
For the downstream scenario (Figure 4.3), environmental dredging resulted in similar results to the
upstream scenario. However, the other remediation actions all resulted in an almost two-fold increase
in PCB concentrations from Day 10 to Day 730. After Day 730, the modeled PCB concentrations in
carp declined for the other four remediation actions, ending slightly greater than the environmental
dredging scenario at Day 3660. Thus, environmental dredging resulted in the most rapid decrease of
PCBs in fish tissue, but the end result at Day 3660 was about the same for all remediation actions.
4-7
-------
TABLE 4.4. MODELED PCB CONCENTRATIONS IN CARP FOR ALL DOWNSTREAM SCENARIOS
(DEPINTO ET AL., 1994)
Day
10
190
370
550
730
920
1100
1280
1470
1650
1830
2020
2200
2380
2560
2750
2930
3110
3300
3480
3660
Average
S.D.*
Modeled PCB Concentrations (mg/kg wet weight) in Carp
No Action
2.86
3.62
4.45
4.74
5.35
5.07
4.80
4.11
3.87
3.40
3.31
2.94
2.91
2.58
2.42
2.09
2.02
1.78
1.77
1.59
1.59
3.2
1.2
No Action,
No Load
2.86
3.60
4.41
4.69
5.28
5.00
4.71
4.00
3.74
3.27
3.15
2.78
2.74
2.41
2.23
1.90
1.82
1.59
1.56
1.39
1.39
3.1
1.2
Environmental
Dredging
2.80
2.24
2.01
1.65
1.52
1.27
1.18
1.11
1.17
1.10
1.15
1.09
1.14
1.08
1.09
1.01
1.03
1.01
1.08
1.06
1.13
1.3
0.46
Hamburg
Cove
2.86
3.62
4.45
4.74
5.35
5.07
4.80
4.11
3.87
3.40
3.31
2.94
2.91
2.58
2.42
2.09
2.02
1.78
1.76
1.58
1.58
3.2
1.2
Hamburg Cove,
No Load
2.86
3.60
4.41
4.69
5.28
5.00
4.71
4.00
3.74
3.27
3.15
2.78
2.74
2.41
2.23
1.90
1.82
1.59
1.56
1.39
1.38
3.1
1.2
'S.D. = standard deviation
A summary table of modeled PCB concentrations in carp for a ten-year predictive run are given in Table
4.5. These results were also extrapolated over a 30-year period for use in the exposure assessment.
The 30-year mean was derived using the PCB concentration at Day 3660 as the representative
concentration for the next 20 years. Thus, a weighted mean could be derived for a 30-year period by
using the following equation:
30-year mean = (10-vear mean) + 2(Dav 3660 concentration)
3
4-8
-------
-p.
cb
to
c
ro
o
Q.
cp_
CD
Q.
-D
o
00
o
o
3
o
CD
O'
3
CO
o
0)
c
•o
CA
r^
^
CB
Q>
3
co
o
CD
01
Concentration (mg/kg wet weight)
z
o
5'
O ;-»
O Ul -• Wl M
04. ' 1 1
-
10
1 1 1 1
10
CJl CO
1
1 1
M
f*
Q. >
O
m
<
190
370 4-
550
730
920
1100
1280
1470
„ 1650
D
0)
" 1830
g | 3480
? 3660
CD
- ]
:
- i
-------
to
c
•
GJ
o
Q.
SL
CD
Q.
TJ
O
CO
o
o
3
o
CO
o
CO
o
Q>
o
o
I
CO
(D
Q>
U
O
CD
D)
Concentration (mg/kg wet weight)
CO
Ul
o
o o
2s
(Q
n i
o 01
S 3
CD
CT
D
r- O I
8 O <"
Ol «? -3
al I
fl
D)
0
10
190
370
550
730
920
1100
1280
1470
1650
1830
2020
2200
2380
2560
2750
2930
3110
3300
3480
3660
-------
TABLE 4.5. SUMMARY OF MODELED PCB CONCENTRATIONS IN BUFFALO RIVER CARP FOR 10-
YEAR AND 30-YEAR SCENARIOS
Remediation Scenario
UPSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
DOWNSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
Modeled PCB Concentration (mg/kg wet weight)
10 Year
Mean
1.9
1.8
1.3
1.6
1.4
3.2
3.1
1.3
3.2
3.1
S.D.
0.65
0.72
0.49
0.89
0.98
1.2
1.2
0.46
1.2
1.2
30 Year
Mean
1.4
1.2
1.1
0.85
0.61
2.1
2.0
1.2
2.1
1.9
S.D.
0.53
0.59
0.32
0.74
0.81
1.0
1.1
0.28
1.0
1.1
This calculation assumed that the PCB concentrations in carp reached a steady state by Day 3660 for
each remediation scenario.
4.2.2 Surface Water Data
Water quality data from the Buffalo River were obtained for the following contaminants: total PCBs,
o-chlordane, /-chlordane, dieldrin, DDT, benzo(a)anthracene, benzo(b)fluoranthene,
benzo(k)fluoranthene, benzo(a)pyrene, chrysene, lead, and copper. Samples were collected from six
sites during two primary sampling periods, each covering a few weeks during the fall of 1990 and
spring of 1992. Specific sampling dates were October 18, 22, 27, 31, and November 5, 9, 13, 1990
and April 4, 18, 22, 1992 (Atkinson et al., 1994). Detailed descriptions of analytical techniques and
results are under preparation by researchers at Buffalo State College (as cited in Atkinson et al., 1994).
For the contaminant transport model, the initial water column conditions were taken as the average
concentrations from six different sites collected on October 18, 1990. The model was run for total
PCBs, benzo(a)anthracene, benzo(a)pyrene, copper, and lead. The modeled data used in each
remediation scenario came from selected segments of the river. The modeled upstream water column
4-11
-------
data were derived from Segment 9, whereas the downstream data were derived from Segment 25
(Figure 4.4) (DePinto et al., 1994).
The modeled water column concentrations are given in Tables 4.6 and 4.7. The mean contaminant
concentrations for the no action, Hamburg Cove, environmental dredging, zero initial conditions, and
flow switching scenarios were quite similar. Contaminant concentrations decreased by over an order
of magnitude when upstream loads were eliminated under the no action and Hamburg Cove scenarios.
For the human health exposure assessment, only the data from the downstream environmental dredging
scenario were used to estimate the risk from consuming water while swimming. Since the risk posed
by this scenario was insignificant (see Chapter 6), it would thus follow that the risk would also be
insignificant for the other remedial alternative scenarios.
4.3 EXPOSURE ASSESSMENT FOR HUMAN HEALTH
4.3.1 General Determination of Chemical Intakes
Once the complete exposure pathways were identified and modeled contaminant concentrations for fish
and surface water were obtained, the human health exposure assessment was conducted. The period
of exposure was assumed to take place after the remedial alternative was completed. This approach
was used so that the increase or reduction in risk, compared to the no action alternative, could be
determined. A similar assumption was made for the wildlife risk assessment.
Exposures were normalized for time and body weight to determine chemical "intakes," expressed in
units of mg chemical/kg body weight-day. For the ingestion of contaminated fish and water, intakes
represent the amount of chemical available for absorption in the gut. The general equation for
calculating chemical intakes is given in Table 4.8. Several variables were used to determine intakes,
including specific information about the exposed population and the period over which the exposure
was averaged. Noncarcinogenic effects were averaged over the same time period as the exposure
duration [i.e., 9 years for typical exposures and 30 years for reasonable maximum (RME) and
subsistence exposures]. Carcinogenic effects were averaged over a lifetime (i.e., 70 years). Intake
variable values were selected so that the combination of all values resulted in an estimate of either the
typical, reasonable maximum, or subsistence exposure intakes.
Modeled water column and fish data that were averaged over a ten-year period were used as the
contaminant concentrations for a typical duration scenario of nine years. This averaging was done to
take into account the movement of people into the Buffalo River AOC one year after the remedial
alternative was implemented.
Chemical intakes were not calculated for the wildlife exposure assessment as a different approach was
used to estimate risks (see Chapter 7).
4-12
-------
28
Buffalo River
Water Column Segmentation
1000 (m)
6
Figure 4.4. Water column segmentation.
4-13
-------
TABLE 4.6. MODELED UPSTREAM WATER COLUMN CONCENTRATIONS FOR VARIOUS REMEDIAL
ALTERNATIVES
Remediation Scenario
No Action
Mean
S.D.
Variance
Hamburg Cove
Mean
S.D.
Variance
Environmental Dredging
Mean
S.D.
Variance
No Action - No Load
Mean
S.D.
Variance
Hamburg Cove - No Load
Mean
S.D.
Variance
Zero Initial Conditions
Mean
S.D.
Variance
Flow Switching
Mean
S.D.
Variance
PCBs
(ng/L)
1.8E+00
4.8E-01
2.3E-04
1.8E+00
4.7E-01
2.2E-04
1 .8E+00
4.8E-01
2.3E-04
2.3E-01
1.1E-01
1.2E-05
2.2E-01
1.1E-01
1.3E-05
1 .8E+00
4.8E-01
2.3E-04
1 .8E+00
4.7E-01
2.2E-04
Benzo(a)-
anthracene
(ng/L)
1.1E+01
8.5E+00
7.2E-02
1.1E+01
8.4E+00
7.0E-02
1.1E+01
8.5E+00
7.2E-02
1.5E-01
1.7E-01
2.9E-05
4.6E-02
5.0E-02
2.5E-06
1.1E+01
8.5E+00
7.2E-02
1.1E+01
8.4E+00
7.0E-02
Benzo(a)-
pyrene
(ng/L)
4.2E+01
5.3E+01
2.8E+00
4.2E+01
5.3E+01
2.8E+00
4.2E+01
5.3E+01
2.8E+00
1.6E-01
1.6E-01
2.5E-05
7.2E-02
8.8E-02
7.7E-06
(no data)
(no data)
Copper
(ug/L)
2.6E+00
1 .3E+00
1 .7E+OO
2.6E+00
1.3E+00
1 .7E+00
2.6E+00
1.3E+00
1.7E+00
1.5E-01
9.4E-02
8.8E-03
1.5E-01
9.4E-02
8.9E-03
(no data)
(no data)
Lead
(ug/L)
1.9E+00
9.8E-01
9.5E-01
1.9E+00
9.8E-01
9.5E-01
1 .9E+00
9.8E-01
9.5E-01
1.3E-01
7.9E-02
6.3E-03
1.3E-01
8.0E-02
6.3E-03
(no data)
(no data)
4.3.2 Ingestion of Contaminated Fish
The equation used to estimate intakes of contaminants due to the ingestion of contaminated fish is
provided in Table 4.9. The parameter values used in that equation are given in Table 4.10. Parameter
values were obtained mostly from recommended EPA sources. The exposure parameters used in the
typical fishing scenario were assumed to be applicable to the general angling population of Buffalo,
whereas the reasonable maximum exposure scenario applied to recreational anglers and their families.
The subsistence exposure scenario was chosen for a sensitive subpopulation of people who would be
4-14
-------
TABLE 4.7. MODELED DOWNSTREAM WATER COLUMN CONCENTRATIONS FOR VARIOUS
REMEDIAL ALTERNATIVES
Remediation Scenario
No Action
Mean
S.D.
Variance
Hamburg Cove
Mean
S.D.
Variance
Environmental Dredging
Mean
S.D.
Variance
No Action - No Load
Mean
S.D.
Variance
Hamburg Cove - No Load
Mean
S.D.
Variance
Zero Initial Conditions
Mean
S.D.
Variance
Flow Switching
Mean
S.D.
Variance
PCBs
(ng/L)
2.0E+00
3.6E-01
1.3E-04
1 .9E+00
3.4E-01
1.1E-04
2.0E+00
3.7E-01
1.3E-04
7.8E-01
3.5E-01
1.2E-04
7.6E-01
3.6E-01
1.3E-04
1 .9E+00
3.5E-01
1 .2E-04
2.0E+00
3.5E-01
1.2E-04
Benzo(a)-
anthracene
(ng/L)
9.9E+00
9.6E+00
9.2E-02
9.8E+00
9.4E+00
8.8E-02
9.9E+00
9.6E+00
9.2E-02
3.8E-01
6.0E-01
3.6E-04
2.4E-01
2.9E-01
8.2E-05
9.8E+00
9.6E+00
9.1E-02
1.0E+01
9.4E+00
8.8E-02
Benzo(a)-
pyrene
(ng/L)
2.0E+01
* 2.3E+01
5.4E-01
2.0E+01
2.3E+01
5.4E-01
2.0E+01
2.3E+01
5.4E-01
4.2E-01
6.3E-01
4.0E-04
3.0E-01
4.3E-01
1.9E-04
(no data)
(no data)
Copper
(ug/L)
2.6E+00
1 .2E+00
1.5E+00
2.6E+00
1.2E+00
1.5E+00
2.6E+00
1.2E+00
1 .5E+00
4.7E-01
2.5E-01
6.4E-02
4.7E-01
2.6E-01
6.5E-02
(no data)
(no data)
Lead
(ug/L)
2.3E+00
1 .OE+00
1.1E+00
2.3E+00
1. OE+00
1.1E+00
2.3E+00
1. OE+00
1.1E+00
4.0E-01
2.1E-01
4.5E-02
4.0E-01
2.1E-01
4.6E-02
(no data)
(no data)
relying on locally caught fish for a large proportion of their diet. The ingestion rates used for each of
those scenarios are listed in Table 4.10; the rationale for selecting these values has been discussed in
detail in Crane (1993a). An assumption was made that the ingestion rate included both "clean" and
contaminated fish. Only fish consumed from the Buffalo River were assumed to be contaminated. In
addition, only modeled PCB data for uncooked, whole carp were available for use in the exposure
assessment. Thus, the contaminant intakes may overestimate risk since PCB concentrations can be
reduced in fish by trimming off the fat and cooking the fish (Zabik et al., 1979).
4-15
-------
TABLE 4.8. GENERIC EQUATION FOR CALCULATING CHEMICAL INTAKES (USEPA, 1989a)
Intake =
C X CR X EFD
BWXAT
where:
Intake
CR
EFD
BW
AT
Intake = the amount of chemical at the exchange boundary (mg/kg body weight-
day)
Chemical-Related Variables
Chemical Concentration = the average concentration contacted over the exposure
period (e.g., mg/L)
Variables that Describe the Exposed Population
Contact Rate = the amount of contaminated medium contacted per unit time or
event (e.g., L/day)
Exposure Frequency and Duration = how long and how often exposure occurs.
Often calculated using two terms, EF and ED, where
EF = exposure frequency (days/year)
ED = exposure duration (years)
Body Weight = the average body weight over the exposure period (kg)
Assessment-Determined Variables
Averaging Time = period over which exposure is averaged (days)
Because there was no quantitative information available on the fraction of fish ingested from the Buffalo
River (i.e., Fl), conservative estimates were made. Based on an average meal of fish (150 g or 0.33 Ib),
the amount of Buffalo River fish consumed for each exposure scenario could also be converted to meals
per year using the following equation:
Ingestion Rate (meals/yr) = [Ingestion Rate (g/day)] x Fl x (meal/150 g) x (365 days/yr)
The number of meals of Buffalo River fish consumed over a year-long period for typical, reasonable
maximum, and subsistence exposures corresponded to approximately 4.5, 33, and 225 meals,
respectively.
4-16
-------
TABLE 4.9. EQUATION USED TO ESTIMATE CONTAMINANT INTAKES DUE TO INGESTION OF
FISH
Intake =
CXIRXFI XEFXED
BWXAT
where:
Intake
C
IR
Fl
EF
ED
BW
AT
Intake Rate (mg/kg-day)
Contaminant Concentration in Fish (mg/kg)
Ingestion Rate (kg/day)
Fraction of Fish Ingested from Contaminated Area (unitless)
Exposure Frequency (days/yr)
Exposure Duration (yr)
Body Weight (kg)
Averaging Time (days)
Chemical intake rates for carp are given in Tables 4.11 and 4.12. Both noncarcinogenic and
carcinogenic intake rates were calculated for typical, reasonable maximum, and subsistence exposures.
4.3.3 Ingestion of Surface Water While Swimming
Ingestion of surface water occurs naturally during swimming. The equation used in computing this
exposure is provided in Table 4.13, and the corresponding exposure are given in Table 4.14. Where
possible, site-specific exposure values were selected following consultation with local residents and
agencies. Where values were taken from the literature, the sources of the values are provided. The
typical exposure scenario assumed that someone went swimming three days per year, whereas six days
per year was the frequency of swimming under reasonable maximum exposure conditions. The
estimated intake rates are listed in Table 4.15.
4-17
-------
TABLE 4.10. PARAMETERS USED IN ESTIMATING CONTAMINANT INTAKES DUE TO
CONSUMPTION OF FISH FROM THE BUFFALO RIVER AOC
Var.
IR
Fl
EF
ED
BW
AT
Units
kg/day
-
day/yr
yrs
kg
days
Value
Used
0.0192
0.054
0.132
0.1
0.25
0.7
350
9
30
70
3285
10950
25550
Comment
Typical: West et al. (1989)
Reasonable Maximum Exposure (RME): USEPA
(1991a)
Subsistence: used the 95th percentile daily intakes
averaged over 3 days for consumers of fin fish [Pao
et al. (1982) cited in USEPA (1989a)l
Typical: study assumption
RME: study assumption
Subsistence: study assumption
USEPA (1991a)
Typical: USEPA (1989a)
RME and Subsistence: USEPA (1989a)
50th percentile average for adult men and women
(USEPA, 1989b)
9 yrs x 365 days/yr (typical noncarcinogenic risk)
30 yrs x 365 days/yr (RME and subsistence
noncarcinogenic risk)
70 yrs x 365 days/yr (carcinogenic risk)
4-18
-------
TABLE 4.11. PCB INTAKE RATES RESULTING FROM THE TYPICAL CONSUMPTION OF CARP
UNDER DIFFERENT REMEDIATION SCENARIOS
Remediation Scenario
UPSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
DOWNSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
Mean Modeled
PCB Cone.*
wet wt. (mg/kg)
1.9
1.8
1.3
1.6
1.4
3.2
3.1
1.3
3.2
3.1
Noncarcinogenic
Intake
(mg/kg-day)
5.0E-05
4.7E-05
3.4E-05
4.2E-05
3.7E-05
8.4E-05
8.2E-05
3.4E-05
8.4E-05
8.2E-05
Carcinogenic
Intake
(mg/kg-day)
6.4E-06
6.1E-06
4.4E-06
5.4E-06
4.7E-06
1.1E-05
1 .OE-05
4.4E-06
1.1E-05
1 .OE-05
* Based on a 10-year modeling duration
4-19
-------
TABLE 4.12. PCB INTAKE RATES RESULTING FROM THE REASONABLE MAXIMUM AND SUBSISTENCE CONSUMPTION OF CARP UNDER
DIFFERENT REMEDIATION SCENARIOS
Remediation Scenario
UPSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
DOWNSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
Mean Modeled
PCB Cone.*
wet wt. (mg/kg)
1.4
1.2
1.1
0.85
0.61
2.1
2.0
1.2
2.1
1.9
Noncarcinogenic Intake
(mg/kg-day)
RME
2.6E-04
2.2E-04
2.0E-04
1 .6E-04
1.1E-04
3.9E-04
3.7E-04
2.2E-04
3.9E-04
3.5E-04
Subsistence
1 .8E-03
1 .5E-03
1 .4E-03
1.1E-03
7.7E-04
2.7E-03
2.5E-03
1 .5E-03
2.7E-03
2.4E-03
Carcinogenic Intake
(mg/kg-dayl
RME
1.1E-04
9.5E-05
8.7E-05
6.7E-05
4.8E-05
1 .7E-04
1 .6E-04
9.5E-05
1 .7E-04
1 .5E-04
Subsistence
7.6E-04
6.5E-04
6.0E-04
4.6E-04
3.3E-04
1.1E-03
1.1E-03
6.5E-04
1.1E-03
1 .OE-03
* Based on a 30-year modeling duration
RME = reasonable maximum exposure
4-20
-------
TABLE 4.13. EQUATION USED TO ESTIMATE CONTAMINANT INTAKE RATES DUE TO INGESTION
OF SURFACE WATER WHILE SWIMMING
Intake =
CWxCRxETxEFxED
BWxAT
where:
Intake
CW
CR
ET
EF
ED
BW
AT
Lifetime Average Intake Rate (mg/kg/day)
Chemical Concentration in Water (mg/L)
Ingestion Rate (L/hour)
Exposure Time (hours/day)
Exposure Frequency (days/year)
Exposure Duration (years)
Body Weight (kg)
Period of Exposure (days)
4-21
-------
TABLE 4.14. PARAMETERS USED FOR COMPUTING INGESTION OF SURFACE WATER WHILE
SWIMMING
Variable
CR
ET
EF
ED
BW
AT
Units
L/hr
hr/day
day/yr
yrs
kg
days
Exposure Scenario
Typical, RME
Typical, RME
Typical
RME
Typical
RME
Typical, RME
Typical
RME
Typical, RME
Value
Used
0.05
0.5
3
6
9
30
70
3285
10950
25550
Reference
USEPA (1989a)
Study assumption
Study Assumption
Study Assumption
USEPA (1989a)
USEPA (1989a)
50th percentile average for
adult men and women (USEPA,
1989b)
9 yrs x 365 days/yr
(noncarcinogenic risk)
30 yrs x 365 days/yr
(noncarcinogenic risk)
70 yrs x 365 days/yr
(carcinogenic risk)
TABLE 4.15. EXPOSURE INTAKE RATES ASSOCIATED WITH INGESTING CONTAMINATED
SURFACE WATER WHILE SWIMMING
Chemical
METALS
Copper
Lead
ORGANIC COMPOUNDS
Benzo(a)anthracene
Benzo(a)pyrene
PCBs
Mean
Water Cone.
(mo/U
2.6E-03
2.3E-03
9.9E-06
2.0E-05
2.0E-06
Noncarcinoaenic Intake
(mg/kg-day)
Typical RME
7.5E-09 1.5E-08
6.7E-09 1.3E-08
2.9E-11 5.8E-11
5.8E-11 1.2E-10
5.9E-12 1.2E-11
Carcinogenic Intake
(mg/kg-day)
Tvcical RME
9.7E-10 6.5E-09
8.6E-10 5.7E-09
3.7E-12 2.5E-11
7.5E-12 5.0E-11
7.5E-13 5.0E-12
4-22
-------
CHAPTER 5
TOXICITY/HAZARD ASSESSMENT
5.1 HUMAN HEALTH TOXICITY VALUES
Two types of toxicity values were used in combination with the chemical intake rates to calculate
noncarcinogenic and carcinogenic health risks to humans. One toxicity value, the reference dose (Rf D),
provides an estimate of the daily contaminant exposure that is not likely to cause harmful effects during
either a portion of a persons' life or his/her entire lifetime. The RfD is the toxicity value used in
evaluating noncarcinogenic effects. The other toxicity value, the slope factor, is used in risk
assessments to estimate an upper-bound lifetime probability of an individual developing cancer as a
result of exposure to a particular concentration of a potential carcinogen. In addition, the EPA weight-
of-evidence classification scheme indicates the strength of evidence that the contaminant is a human
carcinogen (Table 5.1). Slope factors are typically calculated for potential carcinogens in Classes A,
B1, and B2, as well as for Class C on a case-by-case basis. A more detailed description of these
toxicity values, summarized from "Risk Assessment Guidance for Superfund. Volume 1. Human Health
Evaluation Manual (Part A)" (USEPA, 1989a), is given in Appendix C.
TABLE 5.1. EPA WEIGHT-OF-EVIDENCE CLASSIFICATION SYSTEM FOR CARCINOGENICITY
(USEPA, 1989a|
Group Description
A Human carcinogen
B1 or Probable human carcinogen
B2
B1 indicates that limited human data are available
B2 indicates sufficient evidence in animals and inadequate or no
evidence in humans
C Possible human carcinogen
D Not classifiable as to human carcinogenicity
E Evidence of noncarcinogenicity for humans
Chronic oral RfD values and oral slope factors were used for the fish ingestion and surface water
ingestion pathways examined in this risk assessment. Toxicity values, which had undergone an EPA
review process, were obtained from the EPA's Integrated Risk Information System (IRIS) data base.
For chemicals lacking a "verified value," interim toxicity values were obtained from the Health Effects
5-1
-------
Assessment Summary Tables (HEAST), if available (USEPA, 1989c). Table 5.2 lists the toxicity data
used for the chemicals of interest. Although RfDs are provided for known carcinogens, it does not
imply that these doses are protective against carcinogenicity. This table also includes the form in which
the chemical was administered to the test animal or patient (e.g., drinking water, diet, or gavage) for
determination of the oral RfD. The endpoints of concern for evaluating noncarcinogenic risks are listed
in Appendix C. Toxicity profiles of the chemicals of interest are listed in Appendix D.
5.2 WILDLIFE HAZARD ASSESSMENT
A No Observed Adverse Effect Level (NOAEL) was obtained from the literature for mink exposed to
total PCBs through feeding studies. Mann-Klager (1993) used a NOAEL value of 0.069 fjQlg wet weight
for the baseline wildlife risk assessment. This value was used in the comparative risk assessment.
5.3 LIMITATIONS
This risk assessment was limited by the current availability of toxicity information for the select group
of chemicals examined in the modeling exercise. Toxicity values were not available for lead because
age, health, nutritional state, body burden, and exposure duration influence the absorption, release, and
excretion of lead. These factors make it difficult to estimate noncarcinogenic and carcinogenic toxicity
values for lead. Another limitation was that some toxicity values were only available for a certain form
of chemical. For example, the RfD value for PCBs applies only to Aroclor 1254, whereas the oral slope
factor applies to Aroclor 1260. At the present time, toxicity values are not available for all PCB
Aroclors.
5-2
-------
TABLE 5.2. HUMAN HEALTH RISK TOXICITY DATA FOR CHEMICALS OF INTEREST IN THE BUFFALO RIVER
Chemical
METALS
Copper
Lead
PAHs
Benzo(a)anthracene
Benzo(a)pyrene
CHLORINATED HYDROCARBONS
PCBs
Oral RfD
(mg/kg/day)
1 .3E-03
2.0E-05
Form
diet
Source
b
a
Carcinogenic
Weight-of-
Evidence Class
D
B2
B2
B2
B2
Source
a
a
a
a
a
Oral Slope
Factor
1 /(mg/kg/day)
1.15E + 01
7.30E + 00
7.70E + 00
Source
c
a
a
Sources:
a: IRIS (current as of 12/09/94)
b: USEPA (1989c)
c: Interim guidance, relative to benzo(a)pyrene, suggested by OERR (USEPA, 1989d)
Blank spaces denote a lack of information for the chemical of interest.
5-3
-------
CHAPTER 6
COMPARATIVE RISK CHARACTERIZATION: HUMAN HEALTH
6.1 PURPOSE OF THE RISK CHARACTERIZATION STEP
The purpose of the risk characterization step is to combine the exposure and toxicity estimates into an
integrated expression of human health risk. This section presents the calculated potential human health
risks associated with the consumption of contaminated fish and surface water from the Buffalo River
AOC under alternative remediation scenarios. It is important to recognize that these calculated risk
estimates are not intended to be used as actual values. Risk assessment is a regulatory process that
provides risk managers with quantitative estimates that are to be used for comparative purposes only.
These risk estimates must be interpreted in the context of all the uncertainties associated with each
step in the process. Some of the major uncertainties in this risk assessment are addressed in Chapter
8.
Two means of expressing the carcinogenic and noncarcinogenic risks of adverse health effects are
presented in this chapter. First, chemical-specific risks were estimated for each exposure pathway.
Secondly, chemical specific risks were added to estimate a cumulative pathway-specific risk.
6.2 QUANTIFYING RISKS
6.2.1 Determination of Noncarcinogenic Risks
Noncarcinogenic effects are evaluated by comparing an exposure level over a specified time period with
a RfD derived from a similar exposure period [otherwise known as the hazard quotient (HQ)]. Thus,
HQ = exposure level (or intake)/RfD.
Hazard quotients are expressed to one significant figure in a nonprobabilistic way. In this risk
assessment, HQ values were expressed to two significant figures for each chemical; this was done to
reduce rounding errors when HQ values were summed for each pathway. An HQ value of less than
1 indicates that exposures are not likely to be associated with adverse noncarcinogenic effects (e.g.,
reproductive toxicity, teratogenicity, or liver toxicity). As the HQ approaches or exceeds 10, the
likelihood of adverse effects is increased to the point where action to reduce human exposure should
be considered. Owing to the uncertainties involved with these estimates, HQ values between 1 and
10 may be of concern, particularly when additional significant risk factors are present (e.g., other
contaminants are present at concentrations of concern). However, the level of concern does not
6-1
-------
increase linearly as the RfD is approached or exceeded because RfDs do not have equal accuracy or
precision; nor are RfDs based on the same severity of toxic effects (USEPA, 1989a).
In assessing health risks, all HQ values are representative of long-term exposures (i.e., exposures
assumed to occur over a period of 9 or 30 years). The sum of more than one HQ value for multiple
substances and/or multiple exposure pathways is the Hazard Index (HI). Adding the HQs does not
account for any synergistic or antagonistic effects that may occur among chemicals. For this risk
assessment, no attempt was made to distinguish between risk endpoints (e.g., target organs and
related effects) when calculating the HI. Thus, this expression of total risk may be extremely
conservative; it would be better to refine the HI to specific endpoints for HQ values greater than one.
Additional limitations of HQ values and the segregation of hazard indexes have been described
elsewhere (USEPA, 1989a).
6.2.2 Determination of Carcinogenic Effects
Unlike noncarcinogenic effects, carcinogenic substances are thought to pose some degree of risk at all
exposure levels. These effects are estimated as the incremental probability of an individual developing
cancer over a lifetime as a result of exposure to the potential carcinogen. This risk is computed using
average lifetime exposure values that are multiplied by the oral slope factor for a particular chemical.
Slope factors are used to convert estimated daily intakes averaged over a lifetime of exposure directly
to the incremental risk of an individual developing cancer. The resulting carcinogenic risk estimate is
generally an upper-bound estimate, because slope factors are usually based on upper 95th percentile
confidence limits. The EPA believes it is prudent public health policy to consider actions to mitigate
or minimize exposures to contaminants when estimated excess lifetime cancer risks exceed the 10"B
to 10~6 range, and when noncarcinogenic health risks are estimated to be significant (USEPA, 1988a).
Carcinogenic effects were summed for all chemicals in an exposure pathway. This summation of
carcinogenic risks assumed that intakes of individual substances were small, that there were no
synergistic or antagonistic chemical interactions, and that all chemicals produced the same effect (i.e.,
cancer). The limitations to this approach are discussed in detail elsewhere (USEPA, 1989a).
6.3 COMPARATIVE HUMAN HEALTH RISKS IN THE BUFFALO RIVER
6.3.1 Consumption of Contaminated Fish
The consumption of PCB-contaminated carp resulted in substantial noncarcinogenic risks, with the HQ
> 1, and carcinogenic risks [i.e., greater than one person in a million (10"8)1 for all remedial alternatives
and fish consumption scenarios (Tables 6.1 and 6.2). The degree of risk increased as local residents
of the Area of Concern consumed more locally caught carp. For noncarcinogenic risks, the HQ values
ranged from 1.7 to 4.2 for typical exposures, 5.6 to 19 for reasonable maximum exposures, and 39
to 130 for subsistence exposures. For carcinogenic risks, the risks ranged from 3.4x106 to 8.3 x 10B
6-2
-------
TABLE 6.1. HAZARD QUOTIENTS FOR NONCARCINOGENIC RISKS ASSOCIATED WITH
CONSUMING WHOLE CARP UNDER VARIOUS REMEDIATION AND CONSUMPTION
SCENARIOS
Remediation Scenario
UPSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
DOWNSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
Noncarcinogenic Risk
Typical
2.5
2.4
1.7
2.1
1.8
4.2
4.1
1.7
4.2
4.1
RME
13
11
10
7.9
5.6
19
18
11
19
18
Subsistence
89
76
70
54
39
130
130
76
130
120
TABLE 6.2. CARCINOGENIC RISKS ASSOCIATED WITH CONSUMING WHOLE CARP UNDER
VARIOUS REMEDIATION AND CONSUMPTION SCENARIOS
Remediation Scenario
UPSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
DOWNSTREAM
No Action
No Action, No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove, No Load
Lifetime Cancer Risk
Typical
4.9E-05
4.7E-05
3.4E-05
4.2E-05
3.6E-05
8.3E-05
8.1E-05
3.4E-05
8.3E-05
8.1E-05
RME
8.5E-04
7.3E-04
6.7E-04
5.2E-04
3.7E-04
1 .3E-03
1 .2E-03
7.3E-04
1 .3E-03
1 .2E-03
Subsistence
5.8E-03
5.0E-03
4.6E-03
3.6E-03
2.5E-03
8.8E-03
8.4E-03
5.0E-03
8.8E-03
7.9E-03
6-3
-------
for typical exposures, 3.7 x 10~4 to 1.3 x 10~3 for reasonable maximum exposures, and from 2.5 x 103
to 8.8 x 10"3 for subsistence exposures. A greater degree of risk was observed in the downstream
remediation scenarios than the upstream ones.
For the downstream remediation scenarios, some reduction in noncarcinogenic and carcinogenic risk
was observed with environmental dredging, whereas the risk for the other four remediation scenarios
was about the same. Slightly more variability was observed in the upstream scenarios; the lowest risk
generally occurred for no upstream loading of the Hamburg Cove scenario. However, all of the risk
estimates were still above levels of concern. On a comparison basis, this risk assessment exercise was
useful in judging the adequacy of different remedial alternatives in reducing risk. For example,
environmental dredging resulted in the quickest way by which risks to public health could be reduced
during the first 3.5 years (i.e., 1280 days) of the upstream and downstream scenarios. However, the
final modeled output at Day 3660 did not vary much among remediation actions for downstream and
upstream scenarios. A risk manager would need to evaluate the costs and benefits from these results
before deciding which remedial action, if any, to implement.
PCBs were the only chemical included in the exposure and risk assessment because of modeling
limitations. There is a possibility that people who ingest, inhale, or have dermal contact with certain
PCB mixtures may have a greater chance of incurring liver cancer; however, this statement is based
on suggestive evidence rather than on verified data. Studies with three strains of rats and two strains
of mice have verified the carcinogenicity of PCBs through the occurrence of hepatocellular carcinomas.
This evidence was used to classify PCBs as a probable human carcinogen. Monkeys that ingested
0.005-0.08 mg/kg-day doses of Aroclor 1254 showed the following noncarcinogenic effects: ocular
exudate, prominence and inflammation of the eyelid Meibomian glands, and distortion in nail bed
formation. Similar changes have been documented in humans for accidental oral ingestion of PCBs
(IRIS data base retrieval for PCBs, 1994).
This risk assessment assumed that all of the human health risk was attributable to the direct and
indirect (e.g., food chain transfer) exposure of fish to contaminants in the sediments. Since carp are
mostly benthic feeders that generally reside in a localized area, they were used as an indicator of local
contamination problems. In addition, carp have a high lipid content which may readily accumulate
contaminants through the ingestion and assimilation of contaminated food and possibly through the
consumption of contaminated sediment while feeding.
These risk estimates are likely to be overly conservative because they are based on the consumption
of whole fish, rather than fillets. The modeling exercise could not be conducted on fillets, and at the
present time, extrapolations of contaminant concentrations from whole fish to fillets cannot be
accurately made. In addition, these risk estimates were based on raw fish. At the present time,
contaminant concentrations in raw fish cannot be accurately extrapolated to concentrations in cooked
products. For the past 20 years, Mary Zabik and coworkers from Michigan State University have been
investigating whether cooking methods can reduce pesticide and PCB residues in meat and fish (Smith
et al., 1973; Stachiw et al., 1988; Zabik, 1974, 1990; Zabik et al., 1979, 1982). However, their
6-4
-------
results have not been consistent between and within species of fish. In one instance, different cooking
methods did not result in significant changes in the concentrations of PCBs, DDE, or DDT in cooked
carp fillets (Zabik et al., 1982). In another case, cooking resulted in reductions of TCDD in
restructured, deboned carp fillets (Stachiw et al., 1988).
The Michigan Department of Public Health and Michigan State University have conducted a joint
investigation to further assess how cooking techniques may alter the concentrations of contaminants
in fish (H. Humphrey, Michigan Department of Public Health, personal communication, 1994). Studies
have been performed on a variety of sport fish, including chinook salmon, carp, and walleye, in the
Great Lakes area for skin-on and skin-off fillets. Preliminary results indicate that:
• removal of skin produced up to a 50% reduction in contaminant concentrations in
uncooked fillets due to removal of the fat layer below the skin
• cooking produced a 30 to 50% reduction in organic contaminant concentration for
chemicals such as total PCBs, DDT, TCDD, and several pesticides
• the choice of cooking method made no significant difference in the reduction of
contaminant concentrations in the tissues.
The resultsu of the Michigan study will be useful for future human health risk assessments for
determining better estimates of contaminant concentrations in cooked fish. For the present time,
anglers can use the following cooking techniques to reduce their risk to contaminants: (1) trim fatty
areas, (2) puncture or remove skin before cooking so that fats drain away, or (3) deep-fry trimmed
fillets in vegetable oil and discard the oil.
6.3.2 Consumption of Contaminated Surface Water
The noncarcinogenic and carcinogenic risks resulting from the ingestion of surface water while
swimming under typical and reasonable maximum exposure scenarios were estimated to be far below
levels of concern for the environmental dredging scenario (Table 6.3). These risks were estimated
based on a modeled data set for copper, lead, benzo(a)anthracene, benzo(a)pyrene, and PCBs. Hazard
Indices for noncarcinogenic risks ranged from 0.000006 to 0.00001 for typical and reasonable
maximum exposures, respectively. Lifetime cancer risks were on the order of 10"10 for both scenarios.
Based on these estimated risks, an assumption was made that insignificant risks would also result from
dermal exposure to surface water while swimming. This assumption was made because the direct
intake of contaminants into the gut generally results in a greater intake of contaminants than the
absorption of contaminants (with varying capacity for penetration) through the skin. This assumption
has been supported by dermal exposure estimates at more contaminated sites in the Great Lakes region
(Crane, 1994) which have shown negligible carcinogenic risk.
6-5
-------
TABLE 6.3. NONCARCINOGENIC AND CARCINOGENIC RISKS ASSOCIATED WITH INGESTING
CONTAMINATED SURFACE WATER WHILE SWIMMING IN THE BUFFALO RIVER:
ENVIRONMENTAL DREDGING SCENARIO
Chemical
METALS
Copper
Lead
ORGANIC COMPOUNDS
Benzo(a)anthracene
Benzo(a)pyrene
PCBs
Mean
Water Cone.
(mg/L)
2.6E-03
2.3E-03
9.9E-06
2.0E-05
2.0E-06
Hazard Index
(Intake/RfD)
Typical RME
5.8E-06 1.2E-05
2.9E-07 5.9E-07
Lifetime Cancer Risk
(lntake*Slope Factor)
Typical RME
4.3E-11 2.9E-10
5.5E-11 3.6E-10
5.8E-12 3.9E-11
CUMULATIVE RISK 0.000006 0.00001 1E-10 7E-10
Blank spaces denote chemicals lacking toxicity data.
6.3.3 Additive Risks
Risks may be added among exposure pathways to yield an overall estimate of risk to the human
population. For the Buffalo River AOC, the risk associated with consuming fish far outweighed that
of surface water ingestion. Therefore, the additive risks correspond to the fish consumption risks.
6-6
-------
CHAPTER 7
COMPARATIVE RISK CHARACTERIZATION: WILDLIFE
7.1 INTRODUCTION
A simple hazard assessment model, developed by Kubiak and Best (1991), was used to assess
comparative risks to mink in the Buffalo River AOC. The same model was used in the baseline wildlife
risk assessment (Mann-Klager, 1993). The model basically uses a quotient method by which the forage
contaminant concentration is compared to either the lowest observed adverse effect level (LOAEL) or
no observed adverse effect level (NOAEL) for contaminants of concern. To obtain a reasonable amount
of protection, this ratio should not exceed one. Two other techniques which incorporate either
sensitive lifestage information or a biomagnification factor are discussed by Kubiak and Best (1991);
these latter techniques were not relevant to this hazard assessment.
The model provides a way to determine the exceedance over NOAEL by using the derived dietary
NOAEL/LOAEL as the target environmental forage concentration (Kubiak and Best, 1991). This target
forage concentration is divided into the environmental contaminant concentration measured in the
locally caught forage to determine the exceedance over NOAEL.
Exceedance over NOAEL = Measured Forage eqn. 1
Target Forage
7.2 COMPARATIVE RISKS TO MINK
Mink was used as a representative piscivorous species inhabiting the Buffalo River AOC. Thus, any
adverse effects observed in mink may signal a hazard to other piscivorous species.
The amount by which potential mink forage exceeded the NOAEL was determined using equation 1 and
the 10-year modeled carp data from Table 4.5. The only contaminant of concern included in this
assessment was total PCBs. The NOAEL was exceeded by 19 to 46 times for the various remedial
alternatives (Table 7.1). Environmental dredging resulted in the least risk to mink, whereas the other
downstream remediation scenarios were almost identical in risk. In the upstream remediation scenarios,
there was little difference between the environmental dredging and Hamburg Cove - no loading
scenarios. A greater exceedance of the NOAEL was observed for the other upstream remediation
7-1
-------
TABLE 7.1. COMPARATIVE RISKS TO MINK RESULTING FROM THE CONSUMPTION OF
CONTAMINATED CARP FOR VARIOUS REMEDIATION ALTERNATIVES
Remediation Scenario
UPSTREAM
No Action
No Action — No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove - No Load
DOWNSTREAM
No Action
No Action - No Load
Environmental Dredging
Hamburg Cove
Hamburg Cove - No Load
Exceedance over
NOAEL*
28
26
19
23
20
46
45
19
46
45
* NOAEL for mink exposed to PCBs = 0.069 ug/g wet weight
scenarios. The exceedance values in Table 7.1 are overly conservative because an assumption was
made that local mink populations consumed 100% of their diet from contaminated carp.
PCBs are known to cause reproductive and behavioral impairments in mammals. Mink have been found
to be one of the most sensitive mammals to PCBs (Eisler, 1986). The status and health of the mink
population in the Buffalo River system is not known. The results of this comparative risk assessment
warrant additional work to gain a better idea of the local diet of mink in the Buffalo River AOC. In
addition, a more detailed ecological risk assessment should be done to assess the exposure and
potential risks of all contaminants of concern to mink and other indicator species.
7-2
-------
CHAPTER 8
CHARACTERIZATION OF QUALITATIVE UNCERTAINTIES
8.1 INTRODUCTION
A number of assumptions and estimated values were used in the comparative risk assessments that
contributed to the level of uncertainty about the risk estimates. For most environmental risk
assessments, the uncertainty of the risk estimates ranges over an order of magnitude or greater
(USEPA, 1989a). A qualitative listing of the uncertainties associated with each step in the risk
assessment process will be given below to determine the impact of these uncertainties on the final risk
assessment results.
8.2 QUALITATIVE LIST OF UNCERTAINTIES: HUMAN HEALTH
8.2.1 Exposure Assessment
The exposure assessment was based on modeled data. Although the models were verified, uncertainty
still exists about the assumptions used to derive the contaminant concentrations in fish and water under
different remedial alternatives. Specific limitations of the modeling effort were given in Section 3.2.2.3.
The modeling effort was hampered by the limited data set used to verify the models. A medium to high
level of uncertainty was probably associated with using these data in the exposure assessment.
Additional uncertainties associated with the data and assumptions used in the human health exposure
assessment are given below.
• An adequate assessment of complete and incomplete exposure pathways was made.
There is a low uncertainty that some exposure pathways were either not identified or
else were incorrectly classified as a complete or incomplete exposure pathway.
• The exclusion of the complete exposure pathways of dermal exposure to surface water
from the exposure assessment was justifiable because of the low probability that these
pathways would result in substantial human health risks. There is low uncertainty
associated with this assumption.
• The complete exposure pathways chosen for the exposure assessment represent the
primary pathways by which people in the Buffalo River are exposed to contaminants.
The pathways chosen were based primarily on observed activities and on available data.
Thus the level of uncertainty is low.
• Modeled data were only available for a 10-year duration. The data were extrapolated
over 30 years for use in the reasonable maximum and subsistence exposure scenarios.
The uncertainties associated with doing this, beyond those resulting from the model
8-1
-------
itself, are low to medium because the contaminant concentrations from the model
leveled off after 10 years.
• Carp was the only species of fish for which modeled PCB data were available. A
medium to high level of uncertainty was associated with using carp data because other
species of fish (e.g., walleye and perch) would more likely be consumed by people.
• The number of chemicals included in the modeled data were limited. Other chemicals
(e.g., dieldrin, chlordane) not included may contribute to risk. However, PCBs caused
the majority of carcinogenic risk in the baseline risk assessment (Crane, 1993a), and
the uncertainty associated with not including other chemicals is probably low. PCBs
also contributed to most of the noncarcinogenic risk in this risk assessment, and the
uncertainty of not including other chemicals is probably low as well.
• The fish consumption exposure assessment was based on raw, whole carp. In reality,
most fish would be cut into fillets and cooked. This would reduce the contaminant
burden in the cooked fish. The uncertainty associated with using raw, whole carp is
medium to high.
• The selection of exposure frequencies and durations, body weight, life expectancy, and
population characteristics were appropriate. The values for body weight, life
expectancy, and exposure frequency were based on EPA guidance (USEPA, 1989a,b;
1991 a) and have a low to moderate level of uncertainty associated with them. Similar
levels of uncertainty may be attributed to professional judgements about the fraction
of fish ingested from contaminated sources and the number of times someone would
go swimming in the Buffalo River.
• The exposure assessment only estimated contaminant intakes after a remedial
alternative was implemented and not prior to implementation. This was used so that
the effectiveness of the remedial alternatives could be judged more easily. As such,
this risk assessment provided a paper exercise to assist risk managers with making
decisions concerning remediation. If one of the remedial alternatives presented in this
report were ever implemented, it would be useful to collect follow-up data on
contaminant concentrations in fish to compare with the modeled predictions.
8.2.2 Toxicity Values
The oral RfDs and oral slope factors used in this risk assessment were either verified values obtained
from IRIS or interim values obtained from other sources. RfDs and slope factors are subject to change
as result of new information and updates of the IRIS data base. In addition, chemicals will be added
to IRIS in the future to expand the data base. Thus, this risk assessment is "dated" to the toxicity
values available at the time it was prepared. Listed below are the uncertainties associated with using
these toxicity values.
• RfD values and oral slope factors have uncertainty associated with them. Uncertainty
and modifying factors are incorporated into the calculation of RfDs (see Appendix C)
and take into consideration factors such as extrapolating data from long-term animal
studies to humans. In general, RfD values have an uncertainty range of about one order
of magnitude. Since oral slope factors represent an estimate of an upper-bound lifetime
8-2
-------
probability of an individual developing cancer, these values are already conservative.
Thus, the amount of uncertainty associated with oral slope factor values is minimal.
• Toxicity values were not available for all of the chemicals detected in the Buffalo River.
A risk characterization could not be done for lead because the U.S. EPA's Carcinogen
Assessment Group recommends that a numerical estimate not be used for an oral slope
factor. This recommendation was made because quantifying lead's cancer risk involves
many uncertainties, some of which may be unique to lead. The uncertainty of not
being able to include lead in the risk assessment is unknown.
8.2.3 Risk Characterization
The uncertainties associated with the risk characterization step are listed below.
• Exposure intakes and toxicity values will remain the same over the exposure duration.
This assumes that human activities and contaminant levels will remain the same over
the exposure duration, and that toxicity values will not be updated. A moderate to high
level of uncertainty is probably associated with this assumption since toxicity values
are frequently updated in the IRIS data base as new information becomes available.
The level of uncertainty will probably increase with longer exposure durations.
• Health risks are additive for both noncarcinogenic and carcinogenic effects. The
uncertainty associated with this assumption is unknown for the ingestion of surface
water pathway. The toxicity exhibited by a mixture of chemicals may involve
synergistic and antagonistic effects. However, no guidelines are available to judge the
complex interactions a mixture of contaminants may possess in terms of its potential
toxicity to humans. At the present time, standard risk assessment guidance assumes
that health risks are additive.
• The risk characterization only included substances for which data were available. The
potential contribution of other substances expected to be present is probably low.
8.3 QUALITATIVE LIST OF UNCERTAINTIES: WILDLIFE
The comparative risk assessment for wildlife was limited to mink and used the ecological risk
assessment techniques common to a screening assessment. As such, the wildlife assessment was
designed to be very protective. A similar range of uncertainties, as for human health, can be identified
for:
• accuracy of modeled carp data
• selection of exposure pathways
• choice of carp, as opposed to other fish species, for use in the exposure assessment
• exclusion of other chemicals in the exposure assessment.
In addition, some uncertainties are specific to the wildlife risk assessment including:
8-3
-------
• Selection of a NOAEL from the literature. The uncertainty associated with this NOAEL
is probably low to moderate.
• Exclusion of other species (e.g.. common terns) from the wildlife risk assessment due
to the lack of modeled contaminant data for other species of fish. A high degree of
uncertainty is associated with this assumption.
• Exclusion of specific information about habitat utilization in the AOC. The uncertainty
associated with this lack of information is medium to high.
8.4 SUMMARY
Based on the current information available, a complete description of the level of uncertainty associated
with all of the assumptions and data used in this risk assessment cannot be made. This comparative
human health and wildlife risk assessment was based on data and assumptions that, in reality,
represent a snapshot in time. One of the greatest sources of uncertainty in this risk assessment arose
from the use of modeled data. The overall uncertainty of these risk estimates varies by over an order
of magnitude.
This risk assessment was useful for illustrating the process by which a comparative risk assessment
could be conducted. This same methodology could be applied to other contaminated sediment sites.
8-4
-------
REFERENCES
Atkinson, J.F., T. Bajak, M. Morgante, S. Marshall, and J.V. DePinto. 1994. Model Data
Requirements and Mass Loading Estimates for the Buffalo River Mass Balance Study
(ARCS/RAM Program). Final Report. Prepared for U.S. Environmental Protection Agency, Great
Lakes National Program Office, Chicago, IL by State University of New York at Buffalo, Buffalo,
NY.
ATSDR (Agency for Toxic Substances and Disease Registry). 1988. Toxicological Profile for Lead.
Draft. Prepared by Technical Resources, Inc. Oak Ridge National Laboratory, Oak Ridge, TN.
Borneff, J. and H. Kunte. 1965. Carcinogenic Substances in Water and Soil. Part XVII. Concerning
the Origin and Estimation of the Polycyclic Aromatic Hydrocarbons in Water. Arch. Hyg.
(Berlin). 149:226-243.
Bryan, A.M., W.B. Stone, and P.G. Olafsson. 1987. Disposition of Toxic PCB Congeners in Snapping
Turtle Eggs: Expressed as Toxic Equivalents of TCDD. Bull. Environ. Contam. Toxicol. 39:791-
796.
Connolly, J.P., T.F. Parkerton, J.D. Quadrini, S.T. Taylor, and A.J. Thumann. 1992. Development and
Application of a Model of PCBs in the Green Bay, Lake Michigan Walleye and Brown Trout and
Their Food Webs. Report to U.S. EPA Office of Research and Development, ERL-Duluth, Large
Lakes Research Station, Grosse He, Ml. Cooperative Agreement CR-815396.
Crane, J.L. 1992a. Baseline Human Health Risk Assessment: Ashtabula River, Ohio, Area of Concern.
EPA-905-R92-007. U.S. Environmental Protection Agency, Environmental Research Laboratory,
Athens, GA.
Crane, J.L. 1992b. Baseline Human Health Risk Assessment: Saginaw River, Michigan, Area of
Concern. U.S. EPA-905-R92-008. U.S. Environmental Protection Agency, Environmental
Research Laboratory, Athens, GA.
Crane, J.L. 1993a. Baseline Human Health Risk Assessment for the Buffalo River, New York, Area
of Concern. EPA-905-R93-008. U.S. Environmental Protection Agency, Great Lakes National
Program Office, Chicago, IL.
Crane, J.L. 1993b. A Baseline Assessment of Human Health Risks Resulting from PCB Contamination
at the Sheboygan River, Wisconsin, Area of Concern. EPA-905-R93-001. U.S. Environmental
Protection Agency, Environmental Research Laboratory, Athens, GA.
9-1
-------
Crane, J.L. 1 994. Baseline Human Health Risk Assessment: Grand Calumet River/Indiana Harbor Canal,
Indiana, Area of Concern. EPA-905-R94-025. U.S. Environmental Protection Agency,
Environmental Research Laboratory, Athens, GA.
DePinto, J.V., M. Morgante, J. Zaraszczak, T. Bajak, and J.F. Atkinson. 1994 (Draft Report).
Application of Mass Balance Modeling to Assess Remediation Options for the Buffalo River
(ARCS/RAM Program). Prepared for U.S. Environmental Protection Agency, Great Lakes
National Program Office, Chicago, IL by State University of New York at Buffalo, Buffalo, NY.
Diggins, T.P. and K.M. Stewart. 1993. Deformities of Aquatic Larval Midges (Chironomidae: Diptera)
in the Sediments of the Buffalo River, New York. J. Great Lakes Res. 19:648-659.
Durfee, R.L., G. Contos, F.C. Whitmore, J.D. Borden, E.E. Hackman, and R.A. Westin. 1976. PCBs
in the United States: Industrial Use and Environmental Distributions. Office of Toxic
Substances, U.S. Environmental Protection Agency, Washington, DC.
Eadie, B.J., W.R. Faust, P.P. Landrum et al. 1983. Bioconcentrations of PAH by some Benthic
Organisms of the Great Lakes, pp. 437-449. In: Polynuclear Aromatic Hydrocarbons:
Formation, Metabolism and Measurement. M. Cooke and A.J. Dennis (eds.). Battelle Press,
Columbus, OH.
Eisler, R. 1986. Polychlorinated Biphenyl Hazards to Fish, Wildlife and Invertebrates: A Synoptic
Review. U.S. Fish and Wildlife Services Biol. Rep. 85(1.7). Patuxent Wildlife Research Center,
Laurel, MD. 72 pp.
Eisler, R. 1987. Polycyclic Aromatic Hydrocarbon Hazards to Fish, Wildlife and Invertebrates: A
Synoptic Review. U.S. Fish and Wildlife Services Biol. Rep. 85(1.14). Patuxent Wildlife
Research Center, Laurel, MD. 134 pp.
Foley, R.E., S.J. Jackling, R.J. Sloan, and M.K. Brown. 1988. Organochlorine and Mercury Residues
in Wild Mink and Otter: Comparison with Fish. Environ. Toxicol. Chem. 7:363-374.
Fox, G.A. 1993. What Have Biomarkers Told us About the Effects of Contaminants on the Health of
Fish-eating Birds in the Great Lakes? The Theory and a Literature Review. J. Great Lakes Res.
19:722-736.
Fox, G.A., D.V. Weseloh, T.J. Kubiak, and T.C. Erdman. 1991a. Reproductive Outcomes in Colonial
Fish-eating Birds: A Biomarker for Developmental Toxicants in Great Lakes Food Chains. I.
Historical and Ecotoxicological Perspectives. J. Great Lakes Res. 17:153-157.
9-2
-------
Fox, G.A., B. Collins, E. Hayakawa, D.V. Weseloh, J.P. Ludwig, T.J. Kubiak, and T.C. Erdman. 1991 b.
Reproductive Outcomes in Colonial Fish-eating Birds: A Biomarker for Developmental Toxicants
in Great Lakes Food Chains. II. Spatial Variation in the Occurrence and Prevalence of Bill
Defects in Young Double-crested Cormorants in the Great Lakes, 1979-1987. J. Great Lakes
Res. 17:158-167.
Gailani, J.Z., W. Lick, M.K. Pickens, C.K. Ziegler, and D.D. Endicott. 1994 (Draft Report). Sediment
and Contaminant Transport in the Buffalo River. Prepared for U.S. Environmental Protection
Agency, Great Lakes National Program Office, Chicago, IL by U.S. EPA Large Lakes Research
Station.
Gilbertson, M. 1974. Seasonal Changes in Organochlorine Compounds and Mercury in Common Terns
of Hamilton Harbour, Ontario. Bull. Environ. Cont. Toxicol. 12:726-732.
Golub, M.S., J.M. Donald, and J.A. Reyes. 1991. Reproductive Toxicity of Commercial PCB Mixtures:
LOAELS and NOAELS from Animal Studies. Environ. Health Perspect. 94:245-253.
Grimmer, G., H. Bohnke, and H. Borwitzky. 1978. Profile Analysis of Polycyclic Aromatic
Hydrocarbons in Sewage Sludge by Gas Chromatography. Fresenius Z. Anal. Chem. 289:91-
95.
Hesse, J.L. 1976. Polychlorinated Biphenyl Usage and Sources of Loss to the Environment. In:
National Conference on Polychlorinated Biphenyls. J.L. Buckly et al. (eds). QV 633 N277c
1975. U.S. Environmental Protection Agency, Washington, DC.
Hodges, L. 1977. Environmental Pollution. Holt, Rinehart and Winston, New York, NY.
IRIS (Integrated Risk Information System). 1992. U.S. Environmental Protection Agency, Duluth, MN.
IRIS (Integrated Risk Information System). 1993. U.S. Environmental Protection Agency, Duluth, MN.
IRIS (Integrated Risk Information System). 1994. U.S. Environmental Protection Agency, Duluth, MN.
Karwowski, K. 1991. Biomonitoring and Assessment of Environmental Contaminants in Fish-eating
Birds of the Upper Niagra River. U.S. Fish and Wildlife Service, Cortland, NY.
Kimbrough, R.D. and A.A. Jensen. 1989. Topics in Environmental Health: Halogenated Biphenyls,
Terphenyls, Naphthalenes, Dibenzodioxins and Related Products. 2nd Edition. Elsevier, North
Holland, NY.
9-3
-------
Korach, K.S., P. Sarver, K. Chae, J.A. McLachlan, and J.D. McKinney. 1988. Estrogen Receptor-
Binding Activity of Polychlorinated Hydroxybiphenyls: Conformationally Restricted Structural
Probes. Molecular Pharmacology. 33:120-126.
Kubiak, TJ. and D.A. Best. 1991. Wildlife Risks Associated with Passage of Contaminated
Anadromous Fish at Federal Energy Regulatory Commission Licensed Dams in Michigan. U.S.
Fish and Wildlife Service, Contaminants Program, Division of Ecological Services, East Lansing,
Ml.
Kubiak, T.J., H.J. Harris, L.M. Smith, T.R. Schwartz, D.L. Stalling, J.A. Trick, L. Sileo, D.E. Docherty,
and T.C. Erdman. 1989. Microcontaminants and Reproductive Impairment of the Forster's
Tern on Green Bay, Lake Michigan--1983. Arch. Environ. Contam. Toxicol. 18:706-727.
MacKenzie, MJ. and J.V. Hunter. 1979. Sources and Fates of Aromatic Compounds in Urban
Stormwater Runoff. Environ. Sci. Technol. 13:179-183.
Mann-Klager, D.P. 1993. Baseline Wildlife Risk Assessment of the Buffalo River, New York, Area of
Concern. Draft. Department of Interior, Fish and Wildlife Service, Cortland, NY.
May, T.W. and G.L. McKinney. 1981. Cadmium, Lead, Mercury, Arsenic and Selenium Concentrations
in Freshwater Fish, 1976-1977-National Pesticide Monitoring Program. Pesticides Monitoring
J. 15:14-38.
McConnell, E.E. 1980. pp. 109-150. In: Topics in Environmental health: Halogenated Biphenyls,
Terphenyls, Naphthalenes, Dibenzodioxins and Related Products. R.D. Kimbrough (ed.).
Elsevier, North Holland, NY.
Miller, G.T. 1979. Living in the Environment. Wadsworth Publishing Company, Belmont, CA.
NAS (National Academy of Sciences). 1991. Seafood Safety. Committee on Evaluation of the Safety
of Fishing Products, National Academy Press, Washington, DC.
Neff, J.M. 1985. Polycyclic Aromatic Hydrocarbons. In: Fundamentals of Aquatic Toxicology. G.M.
Rand and S.R. Petrocelli (eds.). Hemisphere Publishing Corporation, Washington, DC.
New York State Department of Health. 1989. Health Advisories: Chemicals in Sportfish or Game. New
York State Department of Health, Division of Environmental Health Assessment, Albany, NY.
Newell, A.J., D.W. Johnson, and L.K. Allen. 1987. Niagara River Biota Contamination Project: Fish
Flesh Criteria for Piscivorous Wildlife. New York State Department of Environmental
9-4
-------
Conservation, Division of Fish and Wildlife, Bureau of Environmental Protection. Technical
Report 87-3.
Nicholls, T.P., R. Perry, and J.N. Lester. 1979. The Influence of Heat Treatment on the Metallic and
Polycyclic Aromatic Hydrocarbon Content of Sewage Sludge. Sci. Total Environ. 12:137-150.
Niimi, A.J. 1987. Biological Half-Lives of Chemicals in fishes. Rev. Environ. Contam. Toxicol. 99:1-
46.
Nimmo, D.R. 1985. Pesticides. In: Fundamentals of Aquatic Toxicology. G.M. Rand and S.R.
petrocelli (eds.). Hemisphere Publishing Corporation, Washington, DC.
NOAA (National Oceanic and Atmospheric Administration). 1987. National Status and Trends Program
for Marine Environmental Quality-Progress Report: A Summary of Selected Data on Chemical
Contaminants in Tissues Collected During 1984, 1985 and 1986. NOAA Technical
Memorandum NOS OMA 38. U.S. Department of Commerce, Rockville, MD.
NOAA (National Oceanic and Atmospheric Administration). 1989. National Status and Trends Program
for Marine Environmental Quality-Progress Report: A Summary of Selected Data on Chemical
Contamination from the First Three Years (1986-1988) of the Mussel Watch Project. NOAA
Technical Memorandum NOS OMA 49. U.S. Department of Commerce, Rockville, MD.
Norstrom, R.J. 1988. In: Hazards, Decontamination and Replacement of Polychlorinated Biphenyls.
J.P. Crine (ed.). Plenum Publishing Corporation, New York.
NYSDEC. 1989. Buffalo River Remedial Action Plan. New York State Department of Environmental
Conservation, Albany, NY.
Oliver, B.C. and A.J. Niimi. 1988. Trophodynamic Analysis of Polychlorinated Biphenyl Congeners and
other Chlorinated Hydrocarbons in the Lake Ontario Ecosystem. Environ. Sci. Technol.
22:388-397.
Passino-Reader, D., P.L. Hudson, and J.P. Hickey. 1995. Baseline Risk Assessment for Aquatic Life
for the Buffalo River, New York, Area of Concern. EPA-905-R95-001. National Biological
Survey, Ann Arbor, Ml.
Poland, A. and J.C. Knutson. 1982. 2,3,7,8-Tetrachlorodibenzo-p-dioxin and Related Halogenated
Aromatic Hydrocarbons: Examination of the Mechanism of Toxicity. Annu. Rev. Pharmacol.
Toxicol. 22:517-554.
9-5
-------
Rand, G.M. and S.R. Petrocelli. 1985. Introduction, pp. 1 -30 m Fundamentals of Aquatic Toxicology:
Methods and Applications. Hemisphere Publ. Co., New York.
RTI (Research Triangle Institute). 1993. National Listing of State Fish and Shellfish Consumption
Advisories and Bans. (Current as of July 22, 1993.) Prepared for Office of Science and
Technology. U.S. Environmental Protection Agency. Research Triangle Park, IMC.
Safe, S. 1985. CRC Critical Reviews in Toxicology. Polychlorinated Biphenyls (PCBs) and
Polybrominated Biphenyls (PBBs): Biochemistry, Toxicology and Mechanisms of Action. CRC
Press, Cleveland, OH.
Safe, S. 1990. Polychlorinated Biphenyls (PCBs), Dibenzo-p-dioxins (PCDDs), Dibenzofurans (PCDFs),
and Related Compounds: Environmental and Mechanistic Considerations Which Support the
Development of Toxic Equivalency Factors (TEFs). Critical Rev. Toxicol. 21:51-88.
Safe, S., B. Astroff, M. Harris, T. Zacharewski, R. Dickerson, M. Romkes, and L. Biegel. 1991.
2,3,7,8-Tetrachlorodibenzo-p-dioxin and Related Compounds as Antiestrogens: Characterization
and Mechanism of Action. Pharmacol. Toxicol. 69:400-409.
Schmitt, C.J. and W.G. Brumbaugh. 1990. National Contaminant Biomonitoring Program:
Concentrations of Arsenic, Cadmium, Copper, Lead, Mercury, Selenium, and Zinc in U.S.
Freshwater Fish, 1978-1984. Arch. Environ. Contam. Toxicol. 19:731-747.
Schmitt, C.J., J.L. Zajicek, and P.H. Peterman. 1990. National Contaminant Biomonitoring Program:
Residues of Organochlorine Chemicals in U.S. Freshwater Fish, 1976-1984. Arch. Environ.
Contam. Toxicol. 19:748-781.
Schwartz, T.R., D.L. Stalling, and C.L. Rice. 1987. Are Polychlorinated Biphenyl Residues Adequately
Described by Aroclor Mixture Equivalents? Isomer-Specific Principal Components Analysis of
Such Residues in Fish and Turtles. Environ. Sci. Technol. 21:72-76.
Shain, W., B. Bush, and R. Seegal. 1991. Neurotoxicity of Polychlorinated Biphenyls: Structure-
Activity Relationship of Individual Congeners. Toxicol. Appl. Pharmacol. 111:33-42.
Smith, L.M., T.R. Schwartz, and K. Feltz. 1990. Determination and Occurrence of AHH-active
Polychlorinated Biphenyls, 2,3,7,8-Tetrachloro-p-dioxin and 2,3,7,8-Tetrachlorodibenzofuran
in Lake Michigan Sediment and Biota, the Question of their Relative Toxicological Significance.
Chemosphere. 21:1063-1085.
9-6
-------
Smith, W.E., K. Funk, and M.E. Zabik. 1973. Effects of Cooking on Concentrations of PCB and DDT
Compounds in Chinook (Oncorhvnchus tshawvtscha) and Coho (0. kisutch) Salmon from Lake
Michigan. J. Fish. Res. Board Can. 30:702-706.
Stachiw, N., M.E. Zabik, A.M. Booren, and M.J. Zabik. 1988. Tetrachlorodibenzo-p-dioxin Residue
Reduction through Cooking/Processing of Restructured Carp Fillets. J. Agric. Food Chem.
36:848-852.
Suns, K.R., G.G. Hitchin, and D. Toner. 1993. Spatial and Temporal Trends of Organochlorine
Contaminants in Spottail Shiners from Selected Sites in the Great Lakes (1975-1990). J. Great
Lakes Res. 19:703-714.
Tilson, H.A., J.L. Jacobson, and W.J. Rogan. 1990. Polychlorinated Biphenyls and the Developing
Nervous Systems: Cross-species Comparisons. Neurotox. Teratol. 12:239-248.
U.S. DHHS (U.S. Department of Health and Human Services). 1990. Toxicological Profile for
Polycyclic Aromatic Hydrocarbons. TP-90-20. Agency for Toxic Substances and Disease
Registry, Public Health Service, Atlanta, GA.
U.S. EPA. 1979. The Health and Environmental Impacts of Lead and an Assessment of a need for
Limitations. EPA 560/2-79-001.
U.S. EPA. 1985a. Ambient Water Quality Criteria for Copper - 1984. EPA 44/5-84-031. Off ice of
Water Regulations and Standards, Criteria and Standards Division, Washington, DC.
U.S. EPA. 1985b. Ambient Water Quality Criteria for Lead - 1984. EPA 440/5-84-037. Office of
Water Regulations and Standards, Criteria and Standards Division, Washington, DC.
U.S. EPA. 1988a. Risk Management Recommendations for Dioxin Contamination at Midland,
Michigan. Final Report. EPA Region 5, Chicago, IL. EPA-905/4-88-008.
U.S. EPA. 1988b. Superfund Exposure Assessment Manual. Office of Remedial Response,
Washington, DC. EPA/540/1-88/001.
U.S. EPA. 1989a. Risk Assessment Guidance for Superfund: Human Health Evaluation Manual Part
A. Interim Final. OSWER Directive 9285.7-01 a.
U.S. EPA. 1989b. Exposure Factors Handbook. Office of Health and Environmental Assessment,
Washington, DC. EPA/600/8-89/043.
9-7
-------
U.S. EPA. 1989c. Health Effects Assessment Summary Tables. Fourth Quarter, FY 1989. OERR
9200.6-303-<89-4).
U.S. EPA. 1989d. Interim Policy for Estimating Carcinogenic Risks Associated with Exposures to
Polycyclic Aromatic Hydrocarbons (PAHs). OSWER Directive #9285.4-02. (Contained in
Memorandum from H.L. Longest and B. Diamond to Region Directors).
U.S. EPA. 1991 a. Risk Assessment Guidance for Superfund. Volume I: Human Health Evaluation
Manual. Supplemental Guidance: "Standard Default Exposure Factors." Interim Final (March
25, 1991). OSWER Directive 9285.6-03.
U.S. EPA. 1991b. ARCS: Assessment and Remediation of Contaminated Sediments. 1991 Work Plan.
Great Lakes National Program Office, Chicago, IL.
U.S. EPA. 1992a. National Study of Chemical Residues in Fish. Volume I. EPA-823/R-92-008a.
Office of Science and Technology, Washington, DC.
U.S. EPA. 1992b. National Study of Chemical Residues in Fish. Volume II. EPA-823/R-92-008b.
Office of Science and Technology, Washington, DC.
U.S. EPA. 1993a. Risk Assessment and Modeling Overview Document. U.S. Environmental
Protection Agency, Great Lakes National Program Office, Chicago, IL. EPA 905-R93-007.
U.S. EPA. 1993b. Guidance for Assessing Chemical Contaminant Data for Use in Fish Advisories.
Volume I. Fish Sampling and Analysis. Off ice of Water, Washington, DC. EPA 823-R-93-002.
U.S. EPA. 1993c. Workshop Report on Developmental Neurotoxic Effects Associated with Exposure
to PCBs. September 14-15, 1992, Research Triangle Park, NC. Risk Assessment Forum,
Washington, DC.
U.S. EPA. 1994. Assessment Guidance Document. EPA 905-B94-002. U.S. Environmental Protection
Agency, Great Lakes National Program Office, Chicago, IL.
Varanasi, U., W.L. Reichert, J.E. Stein, et at. 1985. Bioavailability and Biotransformation of Aromatic
Hydrocarbons in Benthic Organisms Exposed to Sediment from an Urban Estuary. Environ. Sci.
Technol. 19:836-841.
West, P.C., J.M. Fly, R. Marans, and F. Larkin. 1989. Michigan Sport Anglers Fish Consumption
Survey. University of Michigan School of Natural Resources, Natural Resource Sociology
Research Lab Technical Report #1, Ann Arbor, Ml.
9-8
-------
Wong, P.T.S., B.A. Silverberg, Y.K. Chau, and P.V. Hodson. 1978. Lead and the Aquatic Biota, pp.
279-342 in The Biogeochemistry of Lead in the Environment. Part B. Biological Effects. J.O.
Nriagu led.). Elsevier/North Holland Biomedical Press, Amsterdam.
Worthing, C.R. 1991. The Pesticide Manual: A World Compendium. 9th edition. British Crop
Protection Council, Croydon, England.
Zabik, M.E. 1974. Polychlorinated Biphenyl Levels in Raw and Cooked Chicken and Chicken Broth.
Poultry Sci. 53:1785-1790.
Zabik, M.E. 1990. Effect of Roasting, Hot-holding, or Microwave Heating on Polychlorinated Biphenyl
Levels in Turkey. School Food Service Res. Rev. 14:98-102.
Zabik, M.E., P. Hoojjat, and C.M. Weaver. 1979. Polychlorinated Biphenyls, Dieldrin and DDT in Lake
Trout Cooked by Broiling, Roasting or Microwave. Bull. Environm. Contam. Toxicol. 21:1 SB-
US.
Zabik, M.E., C. Merrill, and M.J. Zabik. 1982. PCBs and Other Xenobiotics in Raw and Cooked Carp.
Bull. Environm. Contam. Toxicol. 28:710-715.
9-9
-------
APPENDIX A
EXECUTIVE SUMMARY OF THE BASELINE HUMAN HEALTH RISK ASSESSMENT
A.I OVERVIEW
The Assessment and Remediation of Contaminated Sediments (ARCS) program, a 5-year study and
demonstration project relating to the control and removal of contaminated sediments from the Great
Lakes, is being coordinated and conducted by the U.S. Environmental Protection Agency's (EPA) Great
Lakes National Program Office (GLNPO). As part of the ARCS program, baseline human health risk
assessments have been performed at five Areas of Concern (AOCs) in the Great Lakes region. The
Buffalo River, located in western New York State, is one of these AOCs.
In this report, exposure and risk assessment guidelines, developed for the EPA Superfund program,
have been applied to determine the baseline human health risks associated with direct and indirect
exposures to contaminated sediments in the lower Buffalo River. These risks were estimated for
noncarcinogenic (e.g., reproductive toxicity, teratogenicity, liver toxicity) and carcinogenic (i.e.,
probability of an individual developing cancer over a lifetime) effects, based on currently available data.
The risk estimates were not extrapolated to potential future scenarios.
A.2 STUDY AREA
This baseline risk assessment covers an area adjacent to the lower Buffalo River as it passes through
Buffalo, NY before entering Lake Erie. This area has a history of water quality problems due primarily
to point sources of contaminants (i.e., industrial and municipal discharges). The extent of
contamination in the Buffalo River led to the International Joint Commission's (IJC) decision to
designate this region as a Great Lakes AOC. In response, the New York State Department of
Environmental Conservation (NYSDEC) has completed one phase of a remedial action plan (RAP) to
identify and implement pollution abatement measures for the Buffalo River AOC (NYSDEC, 1989).
High concentrations of heavy metals, polychlorinated biphenyls (PCBs), polynuclear aromatic
hydrocarbons (PAHs), and pesticides have been measured in different compartments of the Buffalo
River (e.g., sediments, water column, and fish). Fish advisories have been issued against consuming
carp from the Buffalo River because of excessive levels of PCBs. The transport of these contaminants
into Lake Erie is also of concern. However, it was beyond the scope of this risk assessment to estimate
human health risks to people using the nearshore areas of Lake Erie.
A-1
-------
A.3 EXPOSURE ASSESSMENT
Contact and noncontact recreational activities are limited along the Buffalo River. No public bathing
facilities exist along the river, and fish consumption advisories have warned the public not to eat carp
from the river. However, there is anecdotal evidence that these activities occur, even near industrial
outfalls. This assessment focused on two complete pathways by which residents of the lower Buffalo
River could be exposed to sediment-derived contaminants: (1) consumption of contaminated carp and
spottail/emerald shiners, and (2) ingestion of surface water while swimming. A third complete pathway
of dermal exposure to surface water was assumed to be insignificant in comparison to the risk resulting
from the ingestion of contaminated surface water. This assumption was made because contaminants
are more easily transported across the gut than the skin. Data for other exposure pathways were
determined to be incomplete (e.g., ingestion of sediments).
A limited data set of fish contaminant concentrations was available for use in the exposure assessment.
Carp from three different age classes (i.e., young, middle, and old), collected as part of the ARCS
program, were used. Carp generally represent the most contaminated fish in water bodies due to their
benthic feeding habits and high lipid content. Data from young-of-the-year spottail/emerald shiners
were used to represent another type of fish. Young-of-the-year fish are an important food source for
a variety of fish species consumed by humans. If young-of-the-year fish were the sole food source of
piscivores, they could be used as an indicator of chemical contaminants that may be present in fish
consumed by humans.
Since many species of fish travel between the river and Lake Erie, there is some uncertainty as to
where the fish accumulated their contaminant burden. For the purpose of this risk assessment, it was
assumed that fish collected from the mouth of the river accumulated most of their contaminant burden
from the lower Buffalo River.
Noncarcinogenic and carcinogenic risks were estimated for typical, reasonable maximum, and
subsistence (fish only) exposure scenarios. Typical (i.e., average) exposures were assumed to occur
over a period of 9 years, whereas reasonable maximum (i.e., the maximum exposure that is reasonably
expected to occur at a site) and subsistence (i.e., reliance on fish as a major source of protein)
exposures were assumed to occur over a period of 30 years (USEPA, 1989a). These exposure
durations were extrapolated over a period of 70 years for estimating carcinogenic risks. Exposure
intakes were determined for each chemical and added for each exposure pathway.
For each of the fish exposure scenarios, different consumption patterns of fish were assumed to take
place (Table A.1). These consumption patterns were based, in part, on recommended values given in
EPA Superfund guidance (USEPA, 1989a,b; 1991 a), on published literature values, or on study
assumptions. Based on an average meal of fish (i.e., 150 g or 0.33 Ib), the amount of Buffalo River
fish consumed for each exposure scenario (Table A.1) can also be converted to meals per year using
the following equation:
A-2
-------
TABLE A.1. AMOUNT OF FISH ASSUMED TO BE CONSUMED PER PERSON PER DAY FROM THE
BUFFALO RIVER FOR EACH EXPOSURE SCENARIO
Exposure Scenario
Typical
Reasonable Maximum
Subsistence
Ingestion
Rate*
(g/day)
19.2
54
132
X Fl**
0.10
0.25
0.70
Amount of Buffalo
R. Fish Consumed
(g/day)
1.92
13.5
92.4
* Sources: Typical (West et al., 1989); Reasonable Maximum and Subsistence (USEPA,
1991a)
** Fl = Fraction of fish (i.e., carp or spottail/emerald shiner) estimated to be ingested from the
Buffalo River (study assumption).
Ingestion Rate (meals/yr) = [Ingestion Rate (g/day)] x Fl x (meal/150 g) x (365 days/yr)
The number of meals of Buffalo River fish consumed over a year-long period for typical, reasonable
maximum, and subsistence exposures corresponded to approximately 4.5, 33, and 225 meals,
respectively.
A number of heavy metals and organic compounds were included in the exposure assessment. Toxicity
values for the chemicals detected in the media of interest were obtained from the EPA's Integrated Risk
Information System (IRIS) data base.
A.4 RISK ASSESSMENT
A.4.1 Determination of Risk
Noncarcinogenic effects were evaluated by comparing an exposure level over a specified time period
with a reference dose (RfD)1 derived from a similar exposure period [otherwise known as a hazard
quotient (HQ)]. Thus, HQ = exposure level/RfD. An HQ value of less than 1 indicates that exposures
are not likely to be associated with adverse noncarcinogenic effects. HQ values between 1 and 10 may
be of concern, particularly when additional significant risk factors are present (e.g., other contaminants
are present at concentrations of concern) (USEPA, 1989a). HQ values for multiple substances and/or
multiple exposure pathways were summed to yield an overall Hazard Index (HI). The His are interpreted
The RfD provides an estimate of the daily contaminant exposure that is not likely to
cause harmful effects during either a portion of a person's life or their entire lifetime
(USEPA, 1989a).
A-3
-------
in the same fashion as the HQs. Summing the HQs does not account for any synergistic or antagonistic
effects that may occur among substances.
Carcinogenic risks were estimated as the incremental probability of an individual developing cancer over
a lifetime as a result of exposure to potential carcinogens. This risk was computed using average
lifetime exposure values that were multiplied by the oral slope factor2 for a particular chemical. The
resulting carcinogenic risk estimate generally represented an upper-bound estimate, because slope
factors are usually based on upper 95th percentile confidence limits. Carcinogenic effects were
summed for all chemicals in an exposure pathway. This summation of carcinogenic risks assumed that
intakes of individual substances were small, that there were no synergistic or antagonistic chemical
interactions, and that all chemicals caused cancer. The EPA believes it is prudent public health policy
to consider actions to mitigate or minimize exposures to contaminants when estimated, upper-bound
excess lifetime cancer risks exceed the 10~6 to 108 range, and when noncarcinogenic health risks are
estimated to be significant (USEPA, 1988a).
A.4.2 Noncarcinogenic Risks
A summary of noncarcinogenic risks, as represented by the Hazard Indices, is given in Table A.2.
Noncarcinogenic risks were below levels of concern (i.e., HK1) for typical and reasonable maximum
exposure levels for the fish consumption and surface water ingestion pathways. An assumption was
made that dermal exposure to surface water while swimming would also be insignificant. The risk
levels were of concern (i.e., HI ranged from 2 to 4) for subsistence anglers and their families who
consumed carp from the Buffalo River. Most of the risk was attributable to dieldrin contamination.
Because some of the chemicals detected in the fish do not presently have EPA approved RfD values
(e.g., PCBs), this assessment may underestimate the noncarcinogenic risks from consuming fish from
the lower Buffalo River area. The noncarcinogenic risk reported here is an estimated risk based on
currently available data and toxicity information and should not be construed as an absolute risk.
A.4.3 Carcinogenic Risks
The estimated, upper-bound carcinogenic risks for all fish consumption exposure scenarios were at or
above levels of concern (i.e., 10~B to 10~6 range) (Table A.3). The carcinogenic risk increased with the
age class of carp, and the risk increased by about an order of magnitude for each exposure scenario
from typical to reasonable maximum to subsistence exposures. Spottail/emerald shiners presented less
risk to consumers by at least an order of magnitude, perhaps because of their young age and limited
time for accumulating contaminants.
Slope factors are estimated through the use of mathematical extrapolation models for
estimating the largest possible linear slope (within 95% confidence limits) at low
extrapolated doses that is consistent with the data (USEPA, 1989a).
A-4
-------
TABLE A.2. SUMMARY OF NONCARCINOGENIC RISKS ASSOCIATED WITH TWO EXPOSURE
PATHWAYS IN THE BUFFALO RIVER AOC*
Exposure
Pathway
Fish Consumption
Carp
Carp
Carp
Surface Water Ingestion
Age Class
Young
Middle
Old
-
Exposure Scenario
Typical RME Subsistence
0.04
0.05
0.08
0.002
0.3
0.4
0.6
0.005
2
2
4
-
Non-carcinogenic risks were averaged over the same period as the exposure duration.
TABLE A.3. SUMMARY OF CARCINOGENIC RISKS ASSOCIATED WITH TWO EXPOSURE
PATHWAYS IN THE BUFFALO RIVER AOC*
Exposure
Pathway
Fish Consumption
Carp
Carp
Carp
Spottail/Emerald Shiners
Surface Water Ingestion
Age Class
Young
Middle
Old
Young-of-the-Year
-
Exposure Scenario
Typical RME Subsistence
5E-05
8E-05
1E-04
4E-06
6E-08
1E-03
2E-03
3E-03
9E-05
4E-07
9E-03
1E-02
2E-02
6E-04
-
Carcinogenic risks were averaged over a period of 70 years (i.e., average lifetime of an
individual).
PCBs accounted for most of the carcinogenic risk from fish consumption. There is a possibility that
people who ingest, inhale, or have dermal contact with certain PCB mixtures may have a greater chance
of incurring liver cancer; however, this statement is based on suggestive evidence rather than on
verified data (IRIS database retrieval for PCBs, 1993).
The carcinogenic risk associated with ingesting surface water while swimming ranged from 6 x 10 8
to 4 x 107 for typical and reasonable maximum exposures, respectively. Because these risk estimates
were below levels of concern, it was also assumed that dermal exposure to surface water would also
result in an insignificant carcinogenic risk.
A-5
-------
A.4.4 Uncertainties
Several assumptions and estimated values were used in this baseline risk assessment that contributed
to the overall level of uncertainty associated with the noncarcinogenic and carcinogenic risk estimates.
As with most environmental risk assessments, the uncertainty of the risk estimates probably ranges
over an order of magnitude or greater. The uncertainties were addressed in a qualitative way for the
parameters and assumptions that appeared to contribute the greatest degree of uncertainty. One of
the greatest sources of uncertainty was the assumption that exposure intakes and toxicity values would
not change during the exposure duration. This assumption requires that human activities and
contaminant concentrations remain the same over the exposure duration, and that toxicity values would
not be updated.
A-6
-------
APPENDIX B
WILDLIFE FOUND IN THE BUFFALO RIVER AOC
The following tables provide a list of wildlife species found in the Buffalo River AOC.
TABLE B-1. ENDANGERED, THREATENED, AND SPECIAL CONCERN SPECIES OF NEW YORK
STATE OBSERVED NEAR THE BUFFALO RIVER AOC
Common Name
Least Bittern
Osprey
Northern Harrier
Cooper's Hawk
Red-shouldered Hawk
Peregrine Falcon*
Common Tern
Black Tern
Short-eared Owl
Common Nighthawk
Eastern Bluebird
Scientific Name
Ixobrychus exilis
Pandion haliaetus
Circus cyaneus
Accipiter cooperii
Buteo lineatus
Falco peregrinus
Sterna hirundo
Chlidonias niger
Asio flammeus
Chlordeiles minor
Sialia sialis
Status
Special Concern
Threatened
Threatened
Special Concern
Threatened
Endangered
Threatened
Special Concern
Special Concern
Special Concern
Special Concern
Also Federally Endangered
B-1
-------
TABLE B-2. MAMMMALS, REPTILES, AND AMPHIBIANS RECORDED AT TIMES BEACH CONFINED
DISPOSAL SITE, BUFFALO, NEW YORK (ANDRLE, 1 986 CITED IN MANN-KLAGER,
1993)
Common Name
Muskrat
Eastern Cottontail
Raccoon
White-footed Mouse
Meadow Vole
Red Fox
Weasel
Eastern Garter Snake
American Toad
Bullfrog
Scientific Name
Ondatra zibethicus
Sylvilagus floridanus
Procyon lotor
Peromyscus leucopus
Microtus pennsylvanicus
Vulpes fulva
Mustela sp.
Thamnophis sirtalis
Bufo americanus
Rana catesbeiana
B-2
-------
TABLE B-3. MAMMALS OBSERVED AT TIFFT NATURE PRESERVE, BUFFALO, NEW YORK
(LANDSITTEL, 1990 CITED IN MANN-KLAGER, 1993).
Common Name
Opossum
Masked shrew
Shorttail shrew
Longtail shrew
Starnose mole
Big brown bat
Raccoon
Shorttail weasel
Longtail weasel
Mink
Striped skunk
Coyote
Red fox
Gray fox
Woodchuck
Eastern chipmunk
Eastern gray squirrel
Beaver
White-footed mouse
Deer mouse
Meadow vole
Muskrat
Norway rat
House mouse
Meadow jumping mouse
Woodland jumping mouse
Eastern cottontail
Whitetail deer
Scientific Name
Didelphis marsupialis
Sorex cinereus
Blarina brevicauda
Sorex dispar
Condylura cristata
Eptesicus fuscus
Procyon lotor
Mustela erminea
Mustela frenata
Mustela vison
Mephitis mephitis
Canis latrans
Vulpis fulva
Urocyon cinereoargenteus
Marmota monax
Tamias striatus
Sciurus carolinensis
Castor canadensis
Peromyscus leucopus
Peromyscus maniculatus
Microtus pennsylvanicus
Ondatra zibethica
Rattus norvegicus
Mus musculus
Zapus hudsonius
Mapaeozapus insignis
Sylvilagus floridanus
Odocoileus virginianus
B-3
-------
APPENDIX C
HUMAN HEALTH TOXICITY ASSESSMENT INFORMATION
C.1 TOXICITY ASSESSMENT
The toxicity assessment step is an integral part of the human health baseline risk assessment. This
step includes four tasks: (1) gather qualitative and quantitative toxicity information for substances being
evaluated, (2) identify exposure periods for which toxicity values are necessary, (3) determine toxicity
values [i.e., reference doses (RfDs)] for noncarcinogenic effects, and (4) determine toxicity values (i.e.,
slope factors) for carcinogenic effects (USEPA, 1989a). fne EPA has performed the toxicity
assessment step for a limited number of chemicals and these assessments have undergone extensive
peer review. Therefore, the toxicity assessment step of this study involves primarily a compilation of
available toxicity data.
Once a "verified" toxicity value is agreed upon by the EPA's toxicologists, it is entered into the EPA's
Integrated Risk Information System (IRIS) data base; these values are updated as necessary. IRIS is
the primary source of toxicity information used in baseline risk assessments. The Health Effects
Assessment Summary Tables (HEAST) are the second most current source of toxicity information and
include both verified and interim RfD and slope factor values. Interim values are used for chemicals that
have not yet been approved by the EPA. Specific EPA workgroups, such as the Carcinogen Risk
Assessment Verification Endeavor (CRAVE) and Rf D Workgroups, are another source of interim toxicity
values. If toxicity values are not available in the aforementioned sources, then interim values from
other reports may be used.
This appendix includes brief descriptions of the most important toxicity values used to evaluate
noncarcinogenic and carcinogenic effects; these subsections were summarized from the EPA guidance
document: "Risk Assessment Guidance for Superfund. Volume 1. Human Health Evaluation Manual
(Part A)" (USEPA, 1989a).
C.1.1 Noncarcinogenic Chronic Toxicity
The RfD is the toxicity value used most often in evaluating noncarcinogenic effects. RfDs are based
on the assumption that thresholds exist for certain toxic effects (e.g., cellular necrosis) but may not
exist for other toxic effects (e.g., carcinogenicity). The RfD is defined as an estimate of the daily
exposure to the human population that is likely to be without an appreciable risk of deleterious effects
during either a portion of the lifetime (i.e., subchronic RfD or "RfD,") or during the lifetime (i.e., chronic
RfD or "RfD"). This toxicity value has an uncertainty range of about an order of magnitude and
C-1
-------
includes exposures to sensitive subgroups in the population. For each chemical, the RfD is calculated
from the following equation:
NOAEL or LOAEL
UFXMF
where:
NOAEL = No-Observed-Adverse-Effect-Level
LOAEL = Lowest-Observed-Adverse-Effect-Level
MF = Modifying Factor
U = Uncertainty Factor
The NOAEL and LOAEL are derived from dose-response experiments. The NOAEL represents the
highest exposure level tested at which no adverse effects occurred (including the critical toxic effect),
whereas the LOAEL represents the lowest exposure level at which significant adverse effects occurred.
Uncertainty factors usually consist of multiples of ten, with each factor representing a specific area of
uncertainty included in the extrapolation from available data. An uncertainty factor of ten is usually
used to account for variation in the general population so that sensitive subpopulations are protected.
An additional ten-fold factor is usually applied for each of the following extrapolations: from long-term
animal studies to humans, from a LOAEL to a NOAEL, and when subchronic studies are used to derive
a chronic RfD. A modifying factor (MF), ranging from >0to 10, is included as a qualitative assessment
of additional uncertainties; the default value for the MF is one.
For Aroclor 1254, an oral RFD of 2 x 10"6 mg/kg/day has been approved for the IRIS data base. An
uncertainty factor of 300 and a modifying factor of 1 were assigned to this RfD value. This was based
on clinical and immunologic studies with monkeys.
C.1.2 Carcinogenicity
Human carcinogenic risks are usually evaluated for a chemical by using its slope factor (formerly
designated as a cancer potency factor) and corresponding weight-of-evidence classification. These
variables were listed in Table 5.2 for the Buffalo River chemicals. Slope factors are estimated through
the use of mathematical extrapolation models, most commonly the linearized multistage model, for
estimating the largest possible linear slope (within 95% confidence limits), at low extrapolated doses,
that is consistent with the data. The slope factor is characterized as an upper-bound estimate so that
the true risk to humans, while not identifiable, is not likely to exceed the upper-bound estimate.
The weight of evidence classification for a particular chemical is determined by the EPA's Human Health
Assessment Group (HHAG). Chemicals are placed into one of five groups according to the weight of
evidence from epidemiological studies and animal studies. These groups are designated by the letters
A, B, C, D, and E which represent the level of carcinogenicity to humans (see Table 5.1). Quantitative
C-2
-------
carcinogenic risk assessments are performed for chemicals in Groups A and B, and on a case-by-case
basis for chemicals in Group C.
C.2 UNCERTAINTIES
A number of uncertainties are involved with using toxicity values for estimating noncarcinogenic and
carcinogenic risks. Some of these qualitative uncertainties are listed below:
• Using dose-response information from healthy animal or human populations to predict
effects that may occur in the general population, including susceptible subpopulations
(e.g., elderly, children),
• Using dose-response information from animal studies to predict effects that may occur
in human populations,
• Using NOAELs derived from short-term animal studies to predict effects that may occur
in humans during long-term exposures,
• Using dose-response information from effects observed at high doses to predict the
adverse health effects that may occur following exposure of humans to low levels of
the chemical in the environment, and
• Using a toxicity value derived from exposure to a particular chemical mixture (e.g.,
Aroclor 1260) to represent the level of carcinogenic toxicity for other similar chemical
mixtures (e.g., Aroclor 1242, 1248, and 1254).
C-3
-------
APPENDIX D
TOXICITY PROFILES
The following toxicity profiles were largely derived from those given in U.S. EPA (1993b) and the EPA's
IRIS data base. Although other organic chemicals (i.e., dieldrin, chlordane, ODD, DDE) were found to
contribute to baseline human health (Crane, 1993a) and wildlife (Mann-Klager, 1993) risks, these
chemicals were not included in the comparative risk assessments due to time and cost constraints.
D.1 POLYCHLORINATED BIPHENYLS (PCBs)
PCBs are base/neutral compounds that are formed by direct chlorination of the biphenyl ring. There
are 209 different PCB compounds, termed congeners, based on the possible chlorine substitution
patterns (C12H,0.NCIN, where N = 1-10). PCBs were manufactured by Monsanto under the trade name
Aroclor. Aroclors contain a mixture of congeners, and are named with numbers which indicate the
weight percent of chlorine in each mixture (e.g., Aroclor 1242 represents 42% chlorination of the
biphenyl ring).
PCBs have been widely used in industrial systems. Important properties of PCBs for industrial
applications include thermal stability, fire and oxidation resistance, and solubility in organic compounds
(Hodges, 1977). PCBs were used as insulating fluids in electrical transformers and capacitors, as
plasticizers, as lubricants, as fluids in vacuum pumps and compressors, and as heat transfer and
hydraulic fluids (Hodges, 1977; Nimmo, 1985). Although use of PCBs as a dielectric fluid in
transformers and capacitors was generally considered a closed-system application, the uses of PCBs,
especially during the 1960s, were broadly expanded to many open systems where losses to the
environment were likely. Heat transfer systems, hydraulic fluids in die cast machines, and uses in
speciality inks are examples of more open-ended applications that resulted in serious contamination in
fish near industrial discharge points (Hesse, 1976).
Although PCBs were once used extensively by industry, their production and use in the United States
was banned by the EPA in July 1979 (Miller, 1979). Prior to 1979, the disposal of PCBs and PCB-
containing equipment was not subject to Federal regulation. Prior to regulation, of the approximately
1.25 billion pounds purchased by U.S. industry, 750 million pounds (60 percent) were still in use in
capacitors and transformers, 55 million pounds (4 percent) had been destroyed by incineration or
degraded in the environment, and over 450 million pounds (36 percent) were either in landfills or dumps
or were available to biota via air, water, soil, and sediments (Durfee et al., 1976).
PCBs are extremely persistent in the environment and are bioaccumulated throughout the food chain
(Eisler, 1986; Worthing, 1991). There is evidence that PCB health risks increase with increased
chlorination because more highly chlorinated PCBs are retained more efficiently in fatty tissues (IRIS,
D-1
-------
19941. However, individual PCB congeners have widely varying potencies for producing a variety of
adverse biological effects including hepatoxicity, developmental toxicity, immunotoxicity, neurotoxicity,
and carcinogenicity. The non-ortho-substituted coplaner PCB congeners, and some of the mono-ortho-
substituted congeners, have been shown to exhibit "dioxin-like" effects (Golub et al., 1991; Kimbrough
and Jensen, 1989; McConnell, 1980; Poland and Knutson, 1982; Safe, 1985, 1990; Tilson et al.,
1990; USEPA 1993c). The neurotoxic effects of PCBs appear to be associated with some degree of
ortho-chlorine substitution. There is increasing evidence that many of the toxic effects of PCBs result
from alterations in hormonal function. However, because PCBs can act directly as hormonal agonists
or antagonists, PCB mixtures may have complex interactive effects in biological systems (Korach et al.,
1988; Safe et al., 1991; Shain et al., 1991; USEPA, 1993c). Because of the lack of sufficient
toxicological data, the EPA has not developed quantitative estimates of health risk for specific
congeners.
A recent summary of the National Contaminant Biomonitoring Program data from 1976 through 1984
indicated a significant downward trend in total PCBs in fish, although PCB residues in fish tissue
remained widespread (Schmitt et al., 1990). Total PCBs were detected at 91 percent of 374 sites
surveyed in the National Study of Chemical Residues in Fish (USEPA, 1992a, 1992b). Currently, PCB
contamination in fish and shellfish has resulted in the issuance of consumption advisories in 31 states
(RTI, 1993).
PCBs may be analyzed quantitatively as Aroclor equivalents or as individual congeners. Historically,
Aroclor analysis has been performed by most laboratories. This procedure can, however, result in
significant error in determining total PCB concentrations (Schwartz et al., 1987) and in assessing the
toxicologic significance of PCBs, because it is based on the assumption that the distribution of PCB
congeners in environmental samples and parent Aroclors is similar.
The distribution of PCB congeners in Aroclors may be altered considerably by physical, chemical, and
biological processes after release into the environment, particularly when the process of
biomagnification is involved (Norstrom, 1988; Oliver and Niimi, 1988; Smith et al., 1990). Recent
aquatic environmental studies indicate that many of the most potent, dioxin-like PCB congeners are
preferentially accumulated in higher organisms (Bryan et al., 1987; Kubiak et al., 1989; Oliver and
Niimi, 1988). This preferential accumulation probably results in a significant increase in the total toxic
potency of PCB residues as they move up the food chain. Consequently, the congener-specific analysis
of PCBs is required for more accurate determination of total PCB concentrations and for more rigorous
assessment of the toxicological effects of PCBs.
D.2 POLYCYCLIC AROMATIC HYDROCARBONS (PAHs)
PAHs are base/neutral organic compounds that have a fused ring structure of two or more benzene
rings. PAHs with two to five benzene rings are generally of greatest concern for environmental and
human health effects. These include benzo[a]pyrene and benzo[a]anthracene. Benzo[a]pyrene, one
D-2
-------
of the most widely occurring and potent PAHs, and several other PAHs have been classified by the EPA
as probable human carcinogens (B2) (IRIS, 1992). Although benzo[alpyrene is one of the most
toxicologically potent PAHs, it may represent only a small fraction of the total PAH concentration in fish
or shellfish tissue. Evidence for the carcinogenicity of PAHs in humans come primarily from
epidemiologic studies that have shown an increased mortality due to lung cancer in humans exposed
to PAH-containing coke oven emissions, roof-tar emissions, and cigarette smoke (U.S. DHHS, 1990).
However, it is not possible to conclude from this information that benzofalpyrene is the responsible
agent.
PAHs are ubiquitous in the environment and usually occur as complex mixtures with other toxic
chemicals. They are components of crude and refined petroleum products and of coal. They are also
produced by the incomplete combustion of organic materials. Many domestic and industrial activities
involve pyrosynthesis of PAHs, which may be released into the environment in airborne particulates or
in solid (ash) or liquid by-products of the pyrolytic process. Domestic activities that produce PAHs
include cigarette smoking, home heating with wood or fossil fuels, waste incineration, broiling and
smoking foods, and use of internal combustion engines. Industrial activities that produce PAHs include
coal coking; production of carbon blacks, creosote, and coal tar; petroleum refining; synfuel production
from coal; and use of Soderberg electrodes in aluminum smelters and ferrosilicum and iron works (Neff,
1985). Historic coal gasification sites have also been identified as significant sources of PAH
contamination (J. Hesse, Michigan Department of Public Health, personal communication, March,
1991).
Major sources of PAHs found in marine and fresh waters include biosynthesis (restricted to anoxic
sediments), spillage and seepage of fossil fuels, discharge of domestic and industrial wastes,
atmospheric deposition, and runoff (Neff, 1985). Urban storm water runoff contains PAHs from leaching
of asphalt roads, wearing of tires, deposition from automobile exhaust, and oiling of roadsides and
unpaved roadways with crankcase oil (MacKenzie and Hunter, 1979). Solid PAH-containing residues
from activated sludge treatment facilities have been disposed of in landfills or in the ocean (ocean
dumping was banned in 1989). Although liquid domestic sewage contains < 1 ug/L total PAH, the total
PAH content of industrial sewage is 5 to 15 ug/L (Borneff and Kunte, 1965) and that of sewage sludge
is 1 to 30 mg/kg (Grimmer et at., 1978; Nicholls et al., 1979).
PAHs can accumulate in aquatic organisms from water, sediments, and food. BCFs of PAHs in fish and
crustaceans have frequently been reported to be in the range of 100 to 2,000 (Eisler, 1987). In
general, bioconcentration was greater for the higher molecular weight PAHs than for the lower
molecular weight PAHs. Biotransformation by the mixed function oxidase system in the fish liver can
result in the formation of carcinogenic and mutagenic intermediates, and exposure to PAHs has been
linked to the development of tumors in fish (Eisler, 1987).
Sediment-associated PAHs can be accumulated by bottom-dwelling invertebrates and fish (Eisler,
1987). For example, Great Lakes sediments containing elevated levels of PAHs were reported by Eadie
D-3
-------
et al. (1983) to be the source of the body burdens of the compounds in bottom-dwelling invertebrates.
NAS (1991) reported that PAH contamination in bivalves has been found in all areas of the United
States. Bivalves are good bioaccumulators of some PAHs because they do not metabolize these
compounds as rapidly as fish. Varanasi et al. (1985) ranked benzo[a]pyrene metabolisms by aquatic
organisms as follows: fish > shrimp > amphipod crustaceans > clams. Half-lives for elimination of
PAHs in fish ranged from less than 2 days to 9 days (Niimi, 1987).
D.3 COPPER
Copper is commercially extracted from ores that typically contain several additional metals such as zinc,
cadmium, and molybdenum. Once extracted, the primary uses of copper are for heating equipment,
plumbing, and electrical equipment. Copper has also been widely used as an algicide and aquatic
herbicide (USEPA, 1985a).
Copper is an essential dietary element for plants and animals. In animals, copper makes up an essential
part of many enzymes and is important in hemoglobin formation (Rand and Petrocelli, 1985). Ingestion
of large doses may cause liver or kidney damage.
In aquatic systems, copper's toxic effects tend to decrease with increasing water hardness (USEPA,
1985a). Copper has a low potential to be bioaccumulated in freshwater organisms. Little information
exists relating copper contaminated soils to increased body burdens in mammals.
D.4 LEAD
Lead is derived primarily from the mining and processing of limestone and dolomite deposits, which are
often sources of lead, zinc, and copper (May and McKinney, 1981). It is also found as a minor
component of coal. Historically, lead has had a number of industrial uses, including use in paints, in
solder used in plumbing and food cans, and as a gasoline additive. As recently as the mid-1980s, the
primary source of lead in the environment was the combustion of gasoline; however, use of lead in U.S.
gasoline has fallen sharply in recent years. At present, lead is used primarily in batteries, electric cable
coverings, some exterior paints, ammunition, and sound barriers. Currently, the major points of entry
of lead into the environment are from mining and smelting operations, from fly ash resulting from coal
combustion, and from the leachates of landfills (May and McKinney, 1981).
In aquatic environments, water-borne lead is the most toxic form, with uptake modified by water
hardness (USEPA, 1985b). As water hardness, increases, precipitation also increases, making lead less
available and reducing toxic effects. Lead has been shown to bioaccumulate, with the organic forms,
such as tetraethyl lead, appearing to have the greatest potential for bioaccumulation in fish tissues.
Lead is primarily accumulated in the epidermis and intestine of fish, whereas very little is accumulated
in muscle. High concentrations of lead have been found in marine bivalves and finfish from both
estuarine and marine waters (NOAA, 1987, 1989). Although water-borne lead is concentrated by
D-4
-------
biota, it has not been shown to transfer through the food chain (USEPA, 1979). Lead concentrations
tend to decrease with increasing trophic level in both detritus-based and grazing aquatic food chains
(Wong et al., 1978). Lead concentrations in freshwater fish declined significantly from a geometric
mean concentration of 0.28 ppm in 1976 to 0.11 ppm in 1984. This trend has been attributed
primarily to reductions in lead content of U.S. gasoline (Schmitt and Brumbaugh, 1990).
In terrestrial environments, lead has been shown to cause toxicity in animals, including waterfowl. The
skeleton is the main long-term storage site for lead, with bone concentrations reflecting total chronic
exposure levels. The highest lead concentrations occur in bone, followed by the kidneys. The lowest
lead values are found in the liver, brain, and muscle tissues. Species and individuals respond differently
to lead exposure, with effects of organolead compounds more pronounced than inorganic lead
compounds. Younger developmental stages are more sensitive to lead than older life stages, and
effects are exacerbated by diets deficient in minerals, fats, and proteins.
Lead is particularly toxic to children and fetuses. Subtle neurobehavioral effects (e.g., fine motor
dysfunction, impaired concept formation, and altered behavior profile) occur in children exposed to lead
at concentrations that do not result in clinical encephalopathy (ATSDR, 1988). A great deal of
information on the health effects of lead has been obtained through decades of medical observation and
scientific research. This information has been assessed in the development of air and water quality
criteria by the Agency's Office of Health and Environmental Assessment (OHEA) in support of
regulatory decisionmaking by the Office of Air Quality Planning and Standards and by the Office of
Drinking Water (ODW). By comparison to most other environmental toxicants, the degree of
uncertainty about the health effects of lead is quite low. It appears that some of these effects,
particularly changes in the levels of certain blood enzymes and in aspects of children's neurobehavioral
development, may occur at blood lead levels so low as to be essentially without a threshold. The
Agency's Reference Dose Work Group discussed inorganic lead (and lead compounds) in 1985 and
considered it inappropriate to develop an RfD for inorganic lead (IRIS, 1994). Lead and its inorganic
compounds have been classified as probable human carcinogens (B2) by EPA (IRIS, 1992). However,
at this time, a quantitative estimate of carcinogenic risk from oral exposure is not available (IRIS, 1 994).
D-5
------- |