United States
Environmental Protection
Agency
Office of Water
4301
EPA-820-B-95-009
March 1995
Great Lakes Water
Quality Initiative
Technical Support
Document for
Wildlife Criteria
-------
EPA-820-B-95-009
March 1995
Great Lakes
Water Quality Initiative
Technical Support Document
for Wildlife Criteria
U.S. Environmental Protection Agency
Region 5, Libiary (PL-12.)
77 Wost Jackson Cc^icvard, 12th Floor
Chicago, IL G0604-3590
Office of Science and Technology
Office of Water
United States Environmental Protection Agency
Washington, D.C. 20460
-------
DISCLAIMER
This document has been reviewed by the Health and Ecological Criteria Division,
Office of Science and Technology, U.S. Environmental Protection Agency, and
approved for publication as a support document for the Great Lakes Water
Quality Initiative. Mention of trade names and commercial products does not
constitute endorsement of their use.
ACKNOWLEDGEMENTS
Portions of this document were prepared under EPA Contract No. 68-C3-0332
by Abt Associates and by ICF Incorporated. The Office of Water would like to
express appreciation to Keith Sappington and Mike Wise of Abt Associates and
to Margaret McVey of ICF Incorporated for their technical assistance to this
project.
AVAILABILITY NOTICE
This document is available for a fee upon written request or telephone call to:
National Technical Information Center (NTIS)
U.S. Department of Commerce
5285 Port Royal Road
Springfield, VA 22161
(800) 553-6847
(703) 487-4650
NTIS Document Number: PB95-187332
or
Education Resources Information Center/Clearinghouse for Science,
Mathematics, and Environmental Education (ERIC/CSMEE)
1200 Chambers Road, Room 310
Columbus, OH 43212
(800) 276-0462
(614) 292-6717
ERIC Number: D053
-------
TABLE OF CONTENTS
Section Page
I. INTRODUCTION 1
II. DEFINITIONS 2
III. CALCULATION OF TIER I WILDLIFE CRITERIA AND TIER II WILDLIFE
VALUES 4
III.A. Derivation of the Equation 4
III.B. Derivation of the Final Tier I Wildlife Criterion 5
III.C. Derivation of a Tier II Wildlife Value 5
'IV. DETERMINATION OF THE REPRESENTATIVE SPECIES 6
IV.A. Selection of Mammalian Species 7
IV.B. Selection of Avian Species 7
V. DETERMINATION OF THE MOST APPROPRIATE TEST DOSE FOR USE IN
CALCULATING WILDLIFE VALUES 10
V.A. Study Duration Requirements for the Mammalian Study from which a
Test Dose is Derived 11
V.B. Study Duration Requirements for the Avian Study from which a Test
Dose is Derived 12
V.C. Dose Conversions for Calculating the Test Dose 12
V.C.1. Mammals 13
V.C.2. Birds 16
VI. THEORY AND DETERMINATION OF APPROPRIATE UNCERTAINTY
FACTORS FOR CALCULATION OF WILDLIFE VALUES 17
VI.A. The Interspecies Uncertainty Factor (UFA) 18
VI.A.1. Purpose and Recommended Range 18
VI.A.2. Theoretical Basis: Toxicokinetic and Toxicodynamic
Differences 18
VI.A.3. Allometric Scaling 18
VI.A.4. Empirical Basis: Variability in Acute Sensitivity 20
VI.A.5. Empirical Basis: Variability in Chronic Sensitivity 21
VI.A.6. Guidance on Selecting the Interspecies Uncertainty
Factor (UFA) when Deriving Tier I Wildlife Criteria 22
VI.A.7. Guidance on Selecting the Interspecies Uncertainty
Factor (UFA) when Deriving Tier II Wildlife Values 23
VLB. The Subchronic-to-Chronic Uncertainty Factor (UFS) 24
VI.B.1. Purpose and Recommended Range 24
VI.B.2. Technical Basis 24
VLB.3. Guidance on Selecting the Subchronic-to-Chronic
Uncertainty Factor 25
11
-------
TABLE OF CONTENTS (Continued)
Section Page
VI.C. The LOAEL-TO-NOAEL Uncertainty Factor (UFL) 26
VI.C.1. Purpose and Recommended Range 26
VI.C.2. Technical Basis 27
VI.C.3. Guidance on Selecting the LOAEL-to-NOAEL
Uncertainty Factor 27
VI.D. The Intraspecies Uncertainty Factor (UF() 28
VI.D.1. Purpose and Recommended Value 28
Vl.D.2. Technical Basis 29
VII. DETERMINATION OF EXPOSURE PARAMETER VALUES 30
VILA Chemical-specific Bioaccumulation Factors and Biomagnification
Factors for DDT and metabolites; Mercury; 2,3,7,8-TCDD; and PCBs . . 30
VII.B Species-specific Trophic Levels 31
VU.B.1. Mink (Mustela vison) 32
VII.B.2 River Otter (Lutra canadensis) 33
VII.B.3 Belted Kingfisher (Ceryle alcyon) 33
VII.B.4 Herring Gull (Larus argentatus) 34
VII.B.5. Bald Eagle (Haliaeetus leucocephalus) 34
VII.B.6 Summary 35
VII.C Species-specific Values for Body Weights and Water Ingestion Rates . . 35
VII.C.1 Body Weights 36
VII.C.2 Drinking Water 36
VII.C.3 Food Ingestion Rates 36
VII.C.4 Free-living Metabolic Rates 37
VII.D. Exposure Parameter Values for the Representative Wildlife Species .... 38
VII.D.1 Mink (Mustela vison) 38
VII.D.2 River Otter (Lutra canadensis) 39
VII.D.3 Belted Kingfisher (Ceryle alcyon) 41
VII.D.4. Herring Gull (Larus argentatus) 42
VII.D.5 Bald Eagle (Haliaeetus leucocephalus) 44
VII.D.6 Summary of Exposure Parameter Values for the
Representative Wildlife Species 47
VIII. REFERENCES 48
ill
-------
SUMMARY OF TABLES
Table Number and Title
Table 1.
Page
Exposure Parameter Values and Trophic Level of Prey for Species
Potentially at Risk from Bioaccumulative Contaminants in the Great
Lakes (U.S. EPA, 1995a) 9
Table 2. Laboratory Mammal Body Weights and Food Ingestion Rates 14
Table 3. Body Weights of Farm-reared or "Ranch" Mink 15
«*-.
Table 4. Body Weight and Food Ingestion Rates for Common Avian Laboratory
Species 16
Table 5. Food Ingestion Rates of Growing White Leghorn Female Chickens
(Diet Consists of 9% Water; Medway and Kare 1959) 17
Table 6. Composition of Diet and Average Trophic Level of Aquatic Prey for
Representative Wildlife Species 35
Table 7. Percent Composition of Diet by Wet Weight for the Representative
Species 35
Table 8. Body Weights of Mink Populations 38
Table 9. Body Weights of River Otter Populations 40
Table 10. Body Weights of Belted Kingfisher Populations 41
Table 11. Body Weights of Herring Gull Populations. 43
Table 12. Body Weights of Bald Eagle Populations . 44
Table 13. Exposure Parameter Values for the Five Representative Wildlife
Species 47
iv
-------
GREAT LAKES WATER QUALITY INITIATIVE TECHNICAL SUPPORT
DOCUMENT FOR WILDLIFE CRITERIA
Note: This Technical Support document contains background material and material intended to
clarify portions of the regulation. It does not establish any additional regulatory requirements.
I. INTRODUCTION
The purpose of this document is to provide technical information and the rationale
for the procedure to derive chemical-specific water quality criteria to protect wildlife
species. For the purposes of this document, "wildlife" are defined as non-domesticated
species in the taxonomic classes Aves and Mammalia (birds and mammals).
Because the waters of the Great Lakes System not only support numerous human
activities and habitat for aquatic organisms, but also sustain viable mammalian and avian
wildlife populations, specific water quality criteria are derived to ensure the quality of the
waters in the System are adequate to support these populations. The water quality criteria
for wildlife are surface water concentrations of toxicants that will cause no significant
reduction in the viability or usefulness (in a commercial or recreational sense) of a
population of exposed animals utilizing waters of the Great Lakes System as a drinking
and/or foraging source over several generations. For the purpose of the Great Lakes Water
Quality Initiative (GLWQI) regulation, this concentration is called the Great Lakes Wildlife
Criterion (GLWC).
This document contains a number of sections. In Section II a number of terms are
defined which are used in Appendix D to Part 132 of the final GLWQI guidance, the
GLWQI Methodology for the Development of Wildlife Criteria, as well as in this document.
Section III presents the derivation of the equation used to determine wildlife values, and a
description of the methods to be used to derive either Tier I wildlife criteria or Tier II
wildlife values. Section IV describes U.S. EPA's intent in selecting representative wildlife
species as well as summarizes the analyses carried out by U.S. EPA to determine the
appropriate representative species for the final GLWQI regulation. Section V presents the
minimum toxicity data requirements for the derivation of wildlife values and describes the
scientific judgements required to select the most appropriate toxicity study to use in
deriving wildlife values. Section VI discusses the theory for each of the uncertainty factors
which are considered in deriving wildlife values, summarizes analyses carried out by U.S.
EPA and others to support the recommended ranges for the uncertainty factors, and
provides guidance for the selection of values for each of the uncertainty factors. Section
VII describes the approach used to determine the trophic levels, body weights, and food
and water ingestion rates for each of the representative species as well as describes
general methods which can be used to estimate body weights and food and water
ingestion rates for other wildlife species.
Great Lakes States and Tribes are required to adopt methodologies consistent with
the method set out in appendix D to the GLWQI final to derive wildlife criteria only for the
GLWQI Technical Support Document for Wildlife Criteria Page 1
-------
bioaccumulative contaminants of concern (BCCs) as defined in Part 132.2. This is
consistent with recommendations from the U.S. EPA's Science Advisory Board (U.S. EPA,
1992cand 1994) which endorsed the initial emphasis of the wildlife risk assessment
program being on the direct effect of bioaccumulative chemicals on wildlife.
II. DEFINITIONS
Acceptable endpolnts. For the purpose of wildlife criteria derivation, acceptable subchronic
and chronic endpoints are those which affect reproductive or developmental success,
organismal viability or growth, or any other endpoint which is, or is directly related to,
parameters that influence population dynamics.
Assessment endpoint. An explicit expression of the environmental value that is to be
protected. (As defined in U.S. EPA, 1992b.)
Measurement Endpoint. A measurable ecological characteristic that is related to the valued
characteristic chosen as the assessment endpoint. Measurement endpoints are often
expressed as the statistical or arithmetic summaries that comprise the measurement. (As
defined in U.S. EPA, 1992b.)
Chronic effect. An adverse effect that is measured by assessing an acceptable endpoint,
and results from continual exposure over several generations, or at least over a significant
part of the test species' projected life span or life stage.
Subchronic effect. An adverse effect, measured by assessing an acceptable endpoint,
resulting from continual exposure for a period of time less than that deemed necessary for
a chronic test.
Lowest-observed-adverse-effect-level (LOAEL). The lowest tested dose or concentration of
a substance which resulted in an observed adverse effect in exposed test organisms when
all higher doses or concentrations resulted in the same or more severe effects.
Bounded LOAEL. A LOAEL, as defined above, that is obtained from a study where a
NOAEL for the same endpoint is also determined, such that the dose-response threshold is
bracketed or "bounded" by both the NOAEL and LOAEL.
Unbounded LOAEL. A LOAEL, as defined above, that is obtained from a study where the
corresponding NOAEL for the same endpoint cannot be determined because all of the dose
levels in the test were shown to cause adverse effects relative to the controls. Thus, only
the upper bound of the dose-response threshold can be determined from the unbounded
LOAEL.
No-observed-adverse-effect-level (NOAEL). The highest tested dose or concentration of a
substance which resulted in no observed adverse effect in exposed test organisms where
higher doses or concentrations resulted in an adverse effect.
Page 2 . GLWQI Technical Support Document for Wildlife Criteria
-------
Bounded NOAEL. A NOAEL, as defined above, that is obtained from a study where a
corresponding LOAEL for the same endpoint is also determined, such that the dose-
response threshold is bracketed or "bounded" by both the available NOAEL and LOAEL.
Unbounded NOAEL A NOAEL, as defined above, that is obtained from a study where the
corresponding LOAEL for the same endpoint cannot be determined because none of the
dose levels in the test were shown to cause adverse effects relative to the controls. Thus,
only the lower bound of the dose-response threshold can be determined from the
unbounded NOAEL.
Trophic Level. A functional classification of taxa within a community that is based on
feeding relationships (e.g., aquatic green plants comprise the first trophic level and
herbivores comprise the second). (As defined in U.S. EPA, 1992b.)
Bioaccumulation. The net accumulation of a substance by an organism as a result of
uptake from all environmental sources. (As defined in Appendix B to Part 132 of the final
GLWQI guidance, GLWQI Methodology for Deriving Bioaccumulation Factors.)
Bioaccumulation Factor (BAF). The ratio (in L/kg) of a substance's concentration in tissue
of an aquatic organism to its concentration in the ambient water, in situations where both
the organism and its food are exposed and the ratio does not change substantially over
time. (As defined in Appendix B to Part 132 of the final GLWQI guidance, GLWQI
Methodology for Deriving Bioaccumulation Factors.)
Biomagnification. The increase in tissue concentration of poorly depurated materials in
organisms along a series of predator-prey associations, primarily through the mechanism of
dietary accumulation. (As defined in Appendix B to Part 132 of the final GLWQI
guidance, GL WQI Methodology for Deriving Bioaccumulation Factors.)
Biomagnification Factor (BMF). The ratio of a substance's concentration in the tissue of an
animal that consumes aquatic organisms to its concentration in the aquatic organisms
which it consumes (unitless). The BMF is used in the wildlife methodology to determine
the concentration of a contaminant in gulls, which consume fish, and which, in turn, are
consumed by eagles.
Representative Species. Wildlife species representative of avian and mammalian species
resident in the Great Lakes basin which are likely to experience the highest exposures to
bioaccumulative contaminants through the aquatic food web.
Population. An aggregate of individuals of a species within a specified location in space
and time. (As defined in U.S. EPA, 1992b.)
Wildlife Value, species-specific. The value derived from applying Equation 5 (below) using
exposure parameters for a representative species.
Wildlife Value, taxonomic class-specific. The value derived for a given taxonomic class by
taking the geometric mean of wildlife values for the representative species within a given
taxonomic class.
GLWQI Technical Support Document for Wildlife Criteria Page 3
-------
Great Lakes Wildlife Criterion. The lower of the two taxonomic class-specific wildlife
values (one for birds and one for mammals).
III. CALCULATION OF TIER I WILDLIFE CRITERIA AND TIER II WILDLIFE VALUES
III.A. Derivation of the Equation
The equation used to calculate Wildlife Values (WV) has both an effect and an
exposure component. The effect component is defined as the Test Dose (TD) which is
either a LOAEL or NOAEL for milligrams of substance per kilogram body weight per day
(mg/kg-d). The exposure routes considered in this derivation are food and water ingestion,
and because the intake level is dependent on organism size, it is scaled to body weight.
The total toxicant intake through thes£ exposure routes is determined and then set equal to
the TD as follows:
Toxicant intake through drinking water = (WV x W)/Wt (1)
Toxicant intake through food = [WV x Z. (FTlj x BAFTU]/Wt (2)
Where:
WV = Species-specific wildlife value in milligrams of substance per liter
(mg/L).
W = Average daily volume of water consumed in liters per day (L/d) by the
representative species.
FTU = Average daily amount of food consumed from trophic level i in
kilograms per day (kg/d) by the representative species.
BAF-nj = Bioaccumulation factor for wildlife food in trophic level i in liters per
kilogram (L/kg). Developed using guidelines for wildlife presented in
Apendix B of Part 132 of the final GLWQI guidance, GLWQI
Methodology for Deriving Bioaccumulation Factors. For consumption
of piscivorous birds by other birds, the BAF is derived by multiplying
the trophic level three BAF for fish by a BMF for biomagnification of
the chemical from fish to birds that consume these fish.
Wt = Average weight in kilograms (kg) for the representative species.
Equations one and two are combined to yield Equation three.
TD > (WV x W)/Wt + [WV x Z (FTU x BAFTU]/Wt (3)
Where:
TD = Test Dose in milligrams of substance per kilogram body weight per
day (mg/kg-d) for the test species (either a NOAEL or LOAEL derived
from mammalian or avian toxicity studies).
Factoring and rearranging produces:
Page 4 . GLWQI Technical Support Document for Wildlife Criteria
-------
WV < - - (4)
W + Z IF™ x BAFTLi]
To account for differences in toxicity among species and uncertainties in LOAEL to
NOAEL extrapolations and subchronic to chronic extrapolations, the TO is divided by three
uncertainty factors, UFA, UFS, and UFL.
Where:
UFA = Uncertainty Factor for extrapolating toxicity data across species
(unitless). A species-specific uncertainty factor shall be selected for
each representative species.
UFS = Uncertainty Factor for extrapolating from subchronic to chronic
exposures (unitless).
UFL = Uncertainty Factor for LOAEL to NOAEL extrapolations (unitless).
The final equation for the WV is:
xWt
WV = UF» x UFS x UFL (5)
W + Z [FTlj x BAFTU]
III.B. Derivation of the Final Tier I Wildlife Criterion
Under the final methodology, the wildlife values specific for each taxonomic class
are derived by taking the geometric mean of the wildlife values across all of the
representative species within each taxonomic class. The equation to do this is presented
below (Equation 6).
WV (taxonomic class) = e Exp [Z In WV,,.,,,. ,„,„.,,, /n]
(6)
Where:
n = The number of representative species in a given taxonomic class for
which species-specific wildlife values were calculated.
As required in the final methodology, the Great Lakes Wildlife Criterion is then set equal to
the lower of the two taxonomic class-specific wildlife values.
III.C. Derivation of a Tier II Wildlife Value
The equation to derive a Tier II wildlife value is the same as that presented above to
derive the taxonomic class-specific Tier I wildlife values which are then used to determine
the Tier I wildlife criterion. There are three differences in the derivation of a Tier I wildlife
criterion and a Tier II wildlife value. The first is that for a Tier I wildlife criterion, a
taxonomic class-specific wildlife value is derived for both taxonomic classes (i.e., Aves and
Mammalia) while a Tier II wildlife value may be determined when a taxonomic class-
GLWQI Technical Support Document for Wildlife Criteria Page 5
-------
specific wildlife value is available for only one taxonomic class. The second difference is in
the study duration requirements for both birds and mammals and these are described in
Sections V.A and V.B below. The third difference is in the intent of the interspecies
uncertainty factor (UFA) and its associated value which is described in more detail in
Section VI.A.7 below.
It is important to note that, based on the duration of the study from which the test.
dose was derived and the value selected for the interspecies uncertainty factor (UFA)
selected, it is possible to derive a Tier II wildlife value with as much scientific validity as a
Tier I wildlife criterion. The only difference which could exist would be the assurance that
the taxonomic class for which a wildlife value was not calculated would be protected by
the Tier II wildlife value derived.
If a State or Tribe uses the methodology described in this document to derive a Tier
II wildlife value for non-bioaccumulative chemical, it may be appropriate to select different
representative species which are better examples of wildlife species with the greater
exposure for a given chemical.
IV. DETERMINATION OF THE REPRESENTATIVE SPECIES
The wildlife criteria derived using the approach discussed in this document are
intended to protect all avian and mammalian wildlife species in the Great Lakes basin.
While it would be desirable also to consider reptiles and amphibians, toxicity data and
approaches for assessing exposures of these two classes of vertebrates are not yet
available.
To estimate water quality criteria to protect wildlife in the Great Lakes basin, the
decision was made to establish an approach to ensure protection of those species likely to
face the highest exposure levels to bioaccumulative chemicals as a consequence of their
behaviors and dietary habits. Therefore, U.S. EPA identified species representative of
avian and mammalian species resident in the Great Lakes basin which are likely to
experience the highest exposures through the aquatic food web (i.e., piscivorous birds and
mammals, as discussed below). Within each taxonomic class, the comparative toxicity
data available for a given chemical is frequently very limited. This makes it difficult to
determine the relative toxicologies! sensitivities of the numerous avian and mammalian
wildlife species for which no data are available. The representative species concept
intends to ensure that a wildlife criterion derived using this method would be protective of
the wide distribution of species resident in the Great Lakes basin for which no toxicity data
is available. While it is possible that a less exposed wildlife species (i.e., a non-piscivorous
bird or mammal) may be more lexicologically sensitive than a representative species, that
species will have a much lower exposure and therefore should still be protected by a
criterion derived to protect the identified representative species. U.S. EPA also did not
attempt to select species based on their toxicological sensitivities because such
sensitivities will likely vary according to the chemical and its mode of action. The analysis
described in this section was performed to determine which of the fish-eating avian and
mammalian species of the Great Lakes basin are likely to experience the highest exposure
to contaminants in the basin through aquatic food chains.
Page 6 GLWQI Technical Support Document for Wildlife Criteria
-------
In general, smaller endotherms (i.e., warm-blooded animals, e.g., birds and
mammals) have higher food ingestion rates relative to their body mass than do larger
endotherms. This is because smaller animals generally have a larger surface area to
volume ratio, and lose proportionately more energy to their environment as heat than do
larger animals. This suggests that small animals would be exposed to a larger quantity of
contaminants relative to their body size than larger animals. However, small piscivores
(i.e., fish eating animals) are generally size-limited predators, and feed on smaller fish in
lower trophic-levels than do larger piscivores. Because the concentration of
bioaccumulative pollutants usually is lower in lower trophic level organisms, it is not clear
that smaller animals will experience higher exposures than larger animals. Therefore, to
Identify species likely to experience the highest exposure levels, both relative food
ingestion rates and the trophic level of the prey must be considered.
The identification of species of birds and mammals that are likely to experience the
highest exposures to contaminants in aquatic food chains in the Great Lakes basin was
"based on both animal size and diet of piscivores in the region. In addition, a literature
review was conducted to identify those species for which populations in the Great Lakes
basin have already been adversely impacted by contaminants in aquatic systems at some
time. Those species that have already experienced adverse impacts are likely to be more
at risk in general than those species that have not yet been adversely affected by
contaminants in the Great Lakes. A large part of the higher risk may result from higher
exposure levels. The review and selection of representative mammalian species and
representative avian species are summarized in Sections IV.A and IV.B below. The full
analyses are presented in U.S. EPA (1995a).
1V.A. Selection of Mammalian Species
There are only three mammal species in the Great Lakes region and much of North
America that are likely to be exposed to contaminants through aquatic prey: mink (Mustela
vison], river otter (Lutra canadensis), and raccoon (Procyon fotor). Raccoons tend to be
omnivorous, consuming both plant materials and animal matter; they rarely capture large
fish (U.S. EPA, 1993d). In contrast, mink and otter can capture large fish, and river otter
feed almost exclusively on aquatic prey (U.S. EPA, 1993c). These dietary habits suggest
that the mink and otter are likely to be more exposed than raccoons. This suggestion is
supported by comparative studies of contaminant levels in mammals. Studies of mercury
contamination in furbearing mammals by Sheffy and Amant (1982) and by Wren et al.
(1980) indicate that raccoons tend to exhibit lower concentrations of mercury in their
tissues than both mink and otter captured in the same areas. Based on this information,
the mink and otter were selected as representative of mammalian species most likely to
experience the greatest exposures to bioaccumulative contaminants through the aquatic
food web in the Great Lakes basin.
IV.B. Selection of Avian Species
In contrast to the situation with mammals, there are numerous species of fish-eating
birds in the Great Lakes basin. Thus, selection of species likely to be most exposed
required additional considerations. First, all species of birds that consume aquatic prey
that could inhabit the Great Lakes region were identified on the basis of the overlap of their
GLWQI Technical Support Document for Wildlife Criteria Page 7
-------
habitat requirements and range with the Great Lakes, as indicated in the National
Geographic Field Guide to the Birds of North America (NGS, 1987). In addition, a literature
review was conducted to identify those avian species for which documented reports of
adverse effects in the field have been attributed to toxic chemicals in their food. These
species included bald eagles, osprey, herring gulls, ring-billed gulls, black-crowned night
herons, common terns, Forster's tern, Caspian tern, and the double-crested cormorant
(Colborn, 1991; Environment Canada, 1991; Gilbertson et al., 1991; Peakall, 1988;see
U.S. EPA, 1995a). Most of these species are colonial nesters. For all species with
documented adverse impacts in the past, a literature search was conducted, and body
weights, food ingestion rates, dietary composition, and likely trophic level of prey were
estimated. Of the impacted species, the ones most likely to suffer the highest exposures
to contaminants that bioaccumulate in aquatic food chains were judged to be the bald
eagle, herring gull, and common tern (U.S. EPA, 1995a).
In addition, several species for which adverse impacts have not been documented,
but which appear to have diets and food ingestion rates similar to species that have been
affected, were considered further. These included the belted kingfisher, common (and
American) merganser, and the red-breasted merganser. These species in general are
solitary nesters, and the lack of documented adverse effects may simply reflect the fact
that these difficult-to-study species have not been studied in the Great Lakes basin. Of
these, the species likely to suffer the highest exposure levels was judged to be the belted
kingfisher (U.S. EPA, 1995a). Table 1 summarizes the exposure parameters for all of the
wildlife species considered above. Supporting documentation for the values are in U.S.
EPA(1995a,b).
The avian species selected as those likely to suffer the highest exposure levels to
bioaccumulative contaminants in aquatic food chains were the bald eagle, herring gull,
common tern, and belted kingfisher. Because the common tern and belted kingfisher are
similar in size, and can be very similar in diet, the belted kingfisher was selected to
represent the two species and to provide an example of an unstudied species that may be
at risk.
It is important to remember that these species were selected as likely to be most
exposed on the basis of food ingestion rates and diet. This selection did not cover some
aspects of foraging behavior that can affect exposure. For example, the double-crested
cormorant may capture more benthic fish than the other bird species because of its diving
abilities. If the benthic fish have higher body burdens of contaminants than other fish
because most bioaccumulative contaminants are in the sediments, the cormorants could be
more exposed than birds feeding on higher trophic level fish that are not directly associated
with the sediments. Although considered inappropriate for developing wildlife criteria
applicable to the Great Lakes basin as a whole, this type of information may be worth
considering for local areas with specific and well-defined sites of contaminated sediments.
Page 8 GLWQI Technical Support Document for Wildlife Criteria
-------
Table 1. Exposure Parameter Values and Trophic Level of Prey for Species Potentially at
Risk from Bioaccumulative Contaminants in the Great Lakes (U.S. EPA, 1995a).
Species
mink
otter
kingfisher
osprey
bald eagle
herring gull
ring-billed gull
black-crowned night
heron
common tern
Forster's tern
Caspian tern
double-crested
cormorant
American merganser
(common merganser)
red-breasted
merganser
Adult Body Weight
(kg)
male: 1 .0
female: 0.55
average: 0.80
male: 8.1
female: °S.7
average: 7.4
both: 0.15
male: 1.4
female: 1.8
male: 4.1
female: 5.2
average: 4.6
male: 1 .23
female: 1 .00
average: 1.1
male: 0.566
female: 0.471
both: 0.850
both: 0.120
both: 0.158
both: 0.661
male: 1 .82
female: 1.54
both: 1 .27
male: 1.135
female: 0.908
Ingestion Rate
(kg/kg-d)
male: 0.22
female: 0.24
average: 0.23
male: 0.16
female: 0.17
average: 0.17
both: 0.45
both: 0.19
both: 0.12
male: 0.25
female: 0.26
male: 0.31
female: 0.33
both: 0.22
both: 0.49
both: 0.45
both: 0.30
male: 0.18
female: 0.19
both: 0.24
male: 0.25
female: 0.27
Average Trophic Level
of Aquatic Prey:
Percent of Diet
(wet weight)
Terrestrial: 10%
Aquatic TL 3: 90%
TL 3: 80%
TL 4: 20%
TL3: 100%
TL3: 100%
Fish: 92%
(TL3: 80%)
(TL4: 20%)
Birds: 8%
(herring gull: 70%)
(non-aquatic: 30%)
Fish: 90%
TL 3: 80%
TL 4: 20%
Terrestrial: 10%
Terrestrial: 25%
Aquatic TL 3: 75%
Aquatic: 30-80%
TL 2: 20%
TL 3: 80%
Terrestrial: 20-70%
TL3: 100%
TL3: 100%
TL3: 100%
TL3: 100%
TL3: 100%
TL2: 10%
TL 3: 90%
GLWQI Technical Support Document for Wildlife Criteria
Page 9
-------
V. DETERMINATION OF THE MOST APPROPRIATE TEST DOSE FOR USE IN
CALCULATING WILDLIFE VALUES
The selection of the TD for use in the calculation of individual species-specific WVs
for each taxonomic class requires the application of best professional judgement in
evaluating the results of all available studies.
Because the intent of wildlife criteria is to protect populations rather than
individuals, the measurement endpoints assessed in the study from which a TD is derived
are defined as a set of frank effects which could reasonably be expected to have
implications at the population level. Therefore, for the purposes of wildlife criteria
derivation (and as defined in the Definitions section of this document), acceptable
endpoints are those which affect reproductive or developmental success, organismal
viability or growth, or any other endpoint which is, or is directly related to, a parameter
that influences population dynamics. In evaluating the studies which assess acceptable
endpoints, preference should be given to studies which assess effects on developmental or
reproductive endpoints because, in general, these are more important endpoints in ensuring
that a population's productivity is maintained.
Another restriction placed on the study from which a TD is obtained is that it
provide a defensible, chemical-specific, dose-response curve in which cause and effect are
clearly established. In order to ensure this evaluation criterion is met, the methodology
requires that any study used to obtain a TD be available in the peer-reviewed literature.
The duration of the study must also be considered in evaluating the available data.
In terms of the duration of the study from which the TD is derived, the use of acute data
with the application of an acute-to-chronic conversion ratio for wildlife criteria derivation
was considered. However, the empirical relationship between acute endpoints and chronic
endpoints is too uncertain to use in the derivation of wildlife values for application in this
regulatory context.
Because this methodology is derived specifically for bioaccumulative pollutants, the
duration of the study is required to be subchronic or chronic. General definitions of both
subchronic effect and chronic effect are provided in the Definitions section of this
document. More detailed definitions of subchronic and chronic, relating them to the
lifespan of the test species are provided in Appendix A: Uncertainty Factors to the GLWQI
Technical Support Document for Human Health Criteria and Values. This guidance
indicates that, typically, continuous or repeated exposure for a period of approximately
10% of the lifespan of the test species is considered subchronic and exposure for
approximately 50% of the lifespan of the test species is considered chronic. For the
evaluation of studies to use in the determination of a TD for derivation of wildlife criteria,
the definitions of subchronic and chronic effects are not more quantitative because of the
wide variety of species, with their associated variety in life-spans, which may be tested, in
addition to the variety of chemical characteristics and associated modes of toxic action.
Minimum durations for the study from which a test dose is derived are discussed in
Sections V.A. and V.B. below. Whether a given study which meets these minimum
duration requirements is interpreted as subchronic or chronic is dependent on both
characteristics of the chemical as well as lifespan and life-stage of the test species. The
Page 10 GLWQI Technical Support Document for Wildlife Criteria
-------
minimum duration requirements are established mainly to ensure that the TD is not based
on a study with an insufficient length of exposure which could underestimate the potency
a compound.
If more than one study within a given taxonomic class meets the above
qualifications (i.e., the study evaluates an acceptable endpoint over a subchronic or chronic
duration for the test species and provides a defensible, chemical-specific, dose-response
curve) additional evaluation is required. If studies of equal quality which assess equally
significant endpoints are available for both "wildlife" species as well as traditional
laboratory animals, the preference for obtaining the TD is for the study with the "wildlife"
species. This is because obtaining a TD from study which tested a "wildlife" species may
minimize the uncertainties associated with interspecies extrapolations, depending on both
the test species and the representative species. In addition, many traditional laboratory
species (and especially rats and mice) are bred from a fairly homogeneous gene-pool. Use
of a TD derived from a "wildlife" species is thought to provide a more realistic
representation of the dose-response relationship which may occur in the natural
environment.
When evaluating studies for a given taxonomic class which meet the minimum
requirements to derive a TD for calculation of taxonomic class-specific wildlife values, field
studies are preferred over laboratory studies. It is recognized that there are very few field
studies which meet the qualifications outlined above, but when available, this data would
be most appropriate for derivation of a criterion because it provides an estimate of impacts
within a natural ecosystem.
Another consideration in evaluating available laboratory studies for determination of
the appropriate TD is the route of exposure used in the study. Studies involving exposure
routes other than oral may be considered only when an equivalent oral daily dose can be
estimated and technically justified. The need for oral exposures relates to the nature of the
chemicals of concern, which are bioaccumulative. As a consequence, exposures to the
representative species will be predominantly through the food chain and involve uptake
across the gastrointestinal tract with first-pass metabolism by the liver. Because the
toxicokinetics of bioaccumulative chemicals, and the resulting delivery of the chemical to
the site of action, is critically related to this exposure route, it is imperative that use of a
toxicity study involving a different route of exposure be carefully critiqued and an approach
to convert a non-oral dose to an oral dose be presented. Use of an oral route of exposure
in the study from which the TD is determined should reduce uncertainty in extrapolating
toxicity results from the laboratory to the field.
V.A. Study Duration Requirements for the Mammalian Study from which a Test Dose is
Derived
For mammals, the study from which the TD is derived should be of 90 days or
greater in length. The 90-day study duration for mammals is consistent with the minimum
requirements established in the 1980 Human Health National Guidelines (U.S. EPA, 1980)
and in Appendix C to 40 CFR Part 132 of the final GLWQI guidance, GLWQI Methodology
for Development of Human Health Criteria and Values. In the development of human
health criteria, the 90-day duration is considered to be the minimal time-span for
GLWQI Technical Support Document for Wildlife Criteria Page 11
-------
subchronic effects to emerge based on the life-span of a rodent. Although the test species
used in the study from which a TD is derived may have a very different life-span than a
rodent, it is reasonable to use the minimum 90-day study duration for mammalian wildlife,
keeping in mind the minimum duration requirements are established to ensure that the
potency of a compound is not underestimated.
In order to derive a taxonomic class-specific wildlife value for a Tier II wildlife value,
the minimum study duration for mammals is 28 days rather than 90 days.
V.B. Study Duration Requirements for the Avian Study from which a Test Dose is
Derived
For avian species, the study from which the TD is derived should be of 70 days or
greater in length. However, for tests which assess effects of a contaminant on growth or
mortality endpoints to chicks only exposed post-hatch, 28 days of exposure may be
-considered adequate for determination of a TD.
The 70-day study duration for birds is derived from the Hazard Evaluation Division:
Standard Evaluation Procedure - Avian Reproduction Test (U.S. EPA, 1986). This
Standard Evaluation Procedure explains the procedures used to evaluate effects data
submitted to the Office of Pesticide Programs and is available to other program offices for
the evaluation of studies and scientific data. This document specifies procedures specific
for reproduction tests in birds and requires that the test chemical be administered for at
least ten weeks prior to the onset of egg laying. It also specifies a number of specific
reproductive parameters which may be assessed in a reproductive study. These are
presented here to indicate some acceptable endpoints for the determination of an
appropriate TD for wildlife value derivation, but this should not be seen as a comprehensive
list. These effects are: the number of eggs laid per hen; the percentage of eggs cracked;
the percentage of viable embryos of the egg set; the percentage of live three-week
embryos; the percentage of normal hatchlings of live three-week embryos; the number of
14-day-old survivors per hen and the percentage of 14-day-old survivors of normal
hatchlings.
In order to derive an avian class-specific wildlife value for a Tier II wildlife value, the
minimum study duration is 28 days, regardless of the endpoint or life stage of the test
species.
V.C. Dose Conversions for Calculating the Test Dose
As indicated in Section III, the wildlife criterion is estimated on the basis of a
NOAEL, or the highest ingested dose at which no significant adverse effects occurred. The
dose is expressed in terms of mg contaminant per kg body weight per day (i.e., mg/kg-day,
wet-weight). Many reports on the results of toxicity tests describe exposure levels for the
test animals only in terms of the concentration of the contaminant in their diet. Two
common expressions of dietary concentration are mg of contaminant per kilogram of diet
(mg/kg diet), or parts per million in the diet (ppm diet). These two values are equivalent
when ppm is determined on the basis of weight.
Page 12 GLWQI Technical Support Document for Wildlife Criteria
-------
To convert a toxicity value expressed as ppm in the diet {or drinking water) to
ingested doses expressed as mg/kg-day to calculate a wildlife value, the following
equations can be used:
D (mg/kg-day) = C in diet (mg/kg-diet) x F (kg/day) / Wt (kg), or
D (mg/kg-day) = C in water (mg/L-water) x W (kg/day) / Wt (kg),
where
D = dose in mg of contaminant per kg of body weight per day,
C = contaminant concentration as mg/kg in the diet or mg/L in water,
F = food ingestion rate in kg of diet/kg body weight per day,
W = water ingestion rate (L/day), and
Wt = body weight in kg wet weight.
Thus, to convert doses expressed as ppm in the diet or in drinking water to an intake of
contaminant expressed as mg/kg body weight per day, the body weight and food or
drinking water ingestion rates are required. One should always use the body weights and
feeding or drinking rates reported for the experimental animals in the toxicity test. If these
values are not reported by the investigators, one must estimate the values, if possible,
from other sources of information. Information sources and methods for estimating food
and water ingestion rates are described below for mammals (Section V.C.1) and birds
(Section V.C.2).
V.C.1. Mammals
Common mammalian laboratory species include various strains of rats, mice,
hamsters, guinea pigs, and for chemicals of particular concern for human health, rhesus
monkeys. U.S. EPA has developed values that can be used as default exposure parameter
values for these (and a few other) laboratory mammals in the absence of information in the
toxicity test report. U.S. EPA recommends consulting the Recommendations for and
Documentation of Biological Values for Use in Risk Assessment (U.S. EPA, 1988) to
estimate body weights and food and water ingestion rates for the strain, sex, and ages
covered during the exposure period that are appropriate for the animals exposed to the test
chemical in the toxicity test.
U.S. EPA's (1988) Recommendations do not, however, cover all animals for which
body weights and ingestion rates may be needed to perform dose conversions for wildlife
toxicity tests. First, U.S. EPA has not developed recommended exposure parameter values
for all strains of small mammal commonly used in toxicity tests. Second, not all
investigators report the strain of the animals in their test. To estimate exposure
parameters for unspecified strains, a review of the strains for which data are presented in
U.S. EPA (1988) is appropriate. Table 2 presents body weights and food ingestion rates
for mice, some strains of rats, hamsters, and rhesus monkeys from U.S. EPA (1988).
Given the variation in body size among different strains of laboratory animals, it can be
difficult to determine which body weight/food ingestion rate values might be most
appropriate for an animal of unspecified weight in the toxicity test under consideration. On
the other hand, food ingestion rates normalized to body weight are much more uniform
across strains. The final column in Table 2 illustrates this point by presenting food
GLWQI Technical Support Document for Wildlife Criteria Page 13
-------
Table 2. Laboratory Mammal Body Weights and Food Ingestion Rates.
Species
Mice
Rats
Hamster
Rhesus
Monkey
Strain:
sex
BAF1 mice: males
(mature) females
B6C3F1 mice: males
(mature) females
BAF1 mice: males
(chronic) females
B6C3F1 mice: males
(chronic) females
(a) Body
Weight (kg)
(U.S. EPA.
1988)
0.035
0.030
0.040
0.035
0.0261
0.0222
\»
0.0373
0.0353
(b) Food Ingestion
Rate (kg/day)
(U.S. EPA, 1988)
0.0061
0.0055
0.0067
0.0061
0.0050
0.0045
0.0064
0.0061
Mouse Food Ingestion Rate (mature)
Sprague-Dawley: males
(mature) females
Wistar: males
(mature) females
Fisher: males
(mature) females
Sprague-Dawley: males
(chronic) females
Wistar: males
(chronic) females
Fisher: males
(chronic) females
0.60
0.35
0.50
0.32
0.40
0.25
0.523
0.338
0.462
0.297
0.380
0.229
0.040
0.028
0.035
0.026
0.031
0.022
0.036
0.027
0.034
0.025
0.030
0.021
Rat Food Ingestion Rate (mature or chronic)
Golden Syrian males
(mature) females
Golden Syrian males
(chronic) females
0.090
0.096
0.082
0.088
0.014
0.015
0.013
0.014
Hamster Food Ingestion Rate (mature or chronic)
Rhesus monkey males
(mature) females
Rhesus monkey males
(chronic) females
12
9
10.9
8.0
0.46
0.37
0.43
0.33
Female Rhesus Monkey Food Ingestion Rate
(c) Food
Ingestion Rate
(kg/kg-day)
(c) = (b)/(a)
0.17
0.18
0.17
0.17
0.19
0.20
0.17
0.17
0.17
0.067
0.080
0.070
0.081
0.078
0.088
0.069
0.080
0.074
0.084
0.079
0.092
0.080
0.16
0.16
0.16
0.16
0.16
0.026
0.041
0.039
0.041
0.041
Page 14
GLWQI Technical Support Document for Wildlife Criteria
-------
ingestion rates expressed as kg food ingested/kg body weight per day (i.e., kg/kg-day).
is easier to determine an appropriate value to use when the food ingestion rate is
expressed this way. Also, it is easy to convert toxicity values expressed as ppm in the
diet to doses expressed as mg/kg-day using food ingestion rates normalized to body
weight:
It
D (mg/kg-day) = C (mg/kg diet) x F (kg diet/kg-day)
where
D = dose in mg of contaminant per kg of body weight per day,
C = contaminant concentration as ppm or mg/kg in the diet, and
F = food ingestion rate in kg of diet/kg body weight per day.
When the toxicity test animals are mink (Mustela vison), they are assumed to be
farm-bred or "ranch" mink unless otherwise stated. These tend to weigh more than free-
living mink (U.S.EPA, 1993d). On the basis of the body weight data provided in Table 3,
the female mink used in toxicity tests were assumed to weigh 1 kg and the males were
assumed to weigh 1.8 kg.
Table 3. Body Weights of Farm-reared or "Ranch" Mink.
males
females
average
males
females
average
Body Weight (kg)
1.734 ± 350SD(N=4)
0.974 ± 202SD(N = 12)
1.35
1.822 ± 0.095 SE(N = 6)
0.872 ± 0.036 SE (N = 6)
1.35
Reference/Location
Hornshaw et al., 1 983/Michigan
Bleavins and Aulerich,
1981 /Michigan
Estimates of food ingestion rates of captive mink range from about 12 percent to 16
percent of the adult body weight per day (Aulerich et al., 1973; Bleavins and Aulerich,
1981). Thus, mink used in toxicity tests were assumed to consume about 15 percent of
the adult body weight per day (Aulerich et al., 1973; Newell et al., 1987). For a female
mink weighing 1 kg, this would equal 150 grams of food per day.
For other mammalian species used in toxicity tests that are not covered in U.S.
EPA's (1988) Recommendations, the open literature can be used to identify body weights.
Food and water ingestion rates can be estimated from the appropriate allometric equations
presented in U.S. EPA (1993c), which include the allometric equations from Caulder and
Braun (1983) and Nagy (1987) referenced in Appendix D to Part 132 of the Final GLWQI
guidance. It is important to note the water content of the diet; dry laboratory chows tend
to contain 10 percent water (Altman and Dittmer, 1972), whereas fresh fish and meat tend
to contain approximately 75 percent water (U.S. EPA, 1993c).
GLWQI Technical Support Document for Wildlife Criteria
Page 15
-------
V.C.2.
Birds
The most common avian laboratory species include the mallard (Anas
pfatyrhynchas). Japanese quail (Coturnix japonica), domestic chickens (often white
leghorn) (Callus domesticus), and ring-necked pheasants (Phasianus colchicus). U.S. EPA
has not previously developed recommended exposure parameter values to use as default
values for these birds in the absence of information in the toxicity test report. In this
section, therefore, some guidance is provided on how to identify exposure parameter
values for birds tested in the laboratory. In addition, the exposure parameter values that
U.S. EPA has identified for the common avian laboratory species are provided. Table 4
presents body weights and food ingestion rates for adult mallards, Japanese quail,
domestic chickens, and ring-necked pheasants.
Table 4. Body Weight and Food Ingestion Rates for Common Avian Laboratory Species.
Species
mallard
ring-necked
pheasant
Japanese quail
chicken
Body Weight
(kg)
1
females: 0.95
males: 1 .3
average: 1.1
0.12
2.0
Reference
Delnicki and
Reinecke,
1986; U.S.
EPA 1993c,d
Nelson and
Martin, 1953
Davison et al.,
1 976; Altman
and Dittmer,
1972
Scon et al.,
1976
Food Ingestion
Rate (kg/kg-day)'
0.060 (lab chow)
0.060 (seeds)
0.056
0.058
0.090 (seeds)
0.067 (seeds)
Reference
Nagy, 1987"
Nagy, 1987"
Nagy, 1987"
Medway and
Kare, 1959
b Estimated from Nagy's (1987) allometric equation for estimating dry food ingestion rates for free-living non-
passerine birds by assuming 10% water content of the laboratory feed (Altman and Dittmer, 1972).
If body weights or food or water ingestion rates are not reported for the toxicity
test, data might be obtained from other sources; however, data from the toxicity test itself
should always be used to the extent possible. It is important to identify the age of the bird
over the experimental period. Orfe can usually assume that adult body weights do not
change (unless changes are reported by the investigators). Toxicity tests that begin with
young birds require information on the size of the bird at the beginning and at the end of
the test, if the rate of body weight increase is roughly linear. For growth rates that are
nonlinear over the ages tested, a time-weighted average body weight would be most
appropriate. Table 5 presents data on how body weight and food ingestion rates change
with age for white leghorn chicks.
Page 16
GLWQI Technical Support Document for Wildlife Criteria
-------
Table 5. Food Ingestion Rates of Growing White Leghorn Female Chickens (Diet Consists
of 9% Water; Medway and Kare 1959).
Age
(weeks)
1
2
3
4
6
8
16
32
N
08
96
66
74
50
40
20
20
Body Weight
(g)
Mean ± SO
61.9 ±9.3
102.8 ± 14.7
153.5 ± 18.9
250.2 ± 32.4
384.4 ± 45.6
578.3 ± 39.0
1,293.4 ± 138.2
2,035.2 ± 199.6
Food Ingestion Rate
(g/chick-day)
Mean ± SO
10.4 ± 1.2
16.9 ± 2.0
20.5 ± 1.8
32.8 ± 4.8
38.8 ± 7.4
49.5 ± 6.0
75.5 ± 20.7
136.2 ± 27.9
Food Ingestion
(g/g-day)
0.17
0.16
0.13
0.13
0.10
0.086
0.058
0.067
Body weights of the adults of many wild species of birds can be found in Dunning
(1984), but most of these species are not used in toxicity tests. Information on additional
species, if required, would require a search through the existing literature.
If measured values for food ingestion rates under captive conditions are not
available in the open literature, one can estimate food ingestion rates on the basis of free
living metabolic rate (FMR) or existence metabolic rate (EMR) (depending on the species)
and information on the caloric content of the diet. For birds whose normal activities are
significantly curtailed by captivity, (e.g., birds of prey, seabirds, passerines), EMR would be
more appropriate than FMR. For birds that normally do not fly often, for example, most
gallinaceous birds, FMR and EMR may not be significantly different, and either could be
used. If necessary, EMR or FMR can be estimated on the basis of body weight from an
allometric equation for the appropriate group of birds, as described in U.S. EPA (1993c,
Chapter 4). When estimating food ingestion rates, it is important to note the water
content of the diet. For example, seeds and dry laboratory chows generally contain only 9
to 10 percent water (U.S. EPA, 1993c, Table 4-2; Altman and Dittmer, 1972), while diets
comprised of fresh meat can be 75 percent water (U.S. EPA, 1993c, Table 4-1).
VI. THEORY AND DETERMINATION OF APPROPRIATE UNCERTAINTY FACTORS FOR
CALCULATION OF WILDLIFE VALUES
In applying Equation 5 above, uncertainty factors are applied to adjust the TO for
interspecies differences in toxicological sensitivity (UFA), subchronic to chronic
extrapolations (UFC), and LOAEL to NOAEL extrapolations (UFL). The purpose of this
section is to present the scientific basis for the ranges of the uncertainty factor values
GLWQI Technical Support Document for Wildlife Criteria
Page 17
-------
recommended by U.S. EPA for use in deriving wildlife values. There is also discussion of
an uncertainty factor to adjust for protection of individuals within a given population (UF,)
as discussed in Appendix F, Procedure 1 .A: Requirements for Site-specific Modifications
to Criteria and Values to 40 CFR Part 132 of the final GLWQI guidance.
VI.A. The Interspecies Uncertainty Factor (UFA)
VI.A.1. Purpose and Recommended Range
In the derivation of wildlife values, an interspecies uncertainty factor ranging from 1
to 100 is applied to the NOAEL to account for uncertainties when extrapolating toxic
effects across species (Appendix D of 40 CFR Part 132). This factor is typically used
when adequate toxicity data are lacking for one or more of the representative species. The
magnitude of the uncertainty factor chosen is based on the physicochemical, toxicokinetic,
and toxicodynamic properties of the chemical of concern as well as the amount and quality
of data available. The interspecies uncertainty factor is not intended to account for
differences in ingested dose between the test and representative species since these
factors (body weight, food consumption and water consumption rates) are incorporated
into the derivation of the wildlife value;
VI.A.2. Theoretical Basis: Toxicokinetic and Toxicodynamic Differences
The toxicological basis of observed variability in species sensitivity to a given
toxicant can be grouped into two broad categories: toxicological differences that are
related to variability in toxicokinetics and toxicological differences that are related to
variability in toxicodynamics. Toxicokinetics refers to processes that determine the
delivery of a toxicant to the site of action. Such processes include absorption of the
toxicant, its distribution to the target tissues, its metabolism, and its elimination from the
organism. Toxicodynamics refers to factors that determine the magnitude of response of
target tissue to a given degree of exposure at the site of toxic response. Toxicodynamic
processes are specific to the mechanism of action of the chemical and can include factors
such as binding of the toxicant (or metabolite) to enzymes, DNA sequences or other
receptors that closely mediate the toxic response at the site of action. It is the combined
effect of both toxicokinetics and toxicodynamics that determines the overall sensitivity of a
species to a toxicant. Additional discussion of toxicokinetic and toxicodynamic differences
across species is provided in U.S. EPA, 1995c.
VI.A.3. Allometric Scaling
The current state of the science indicates that some of the variability in sensitivity
across species can be related to some rather simple and general quantitative patterns of
anatomy and physiology of different sizes of mammalian species. It has been shown that
the absolute rates of such processes, including basal metabolic rate, cardiac output, renal
clearance, oxygen consumption, food consumption, water consumption, etc., tend to vary
across species according to allometric scaling factors that can be expressed as a non-linear
function of body weight (e.g., body weight374 as endorsed in U.S. EPA, 1992a). The
general form of these allometric equations is:
Page 18 GLWQI Technical Support Document for Wildlife Criteria
-------
Y = a W*> (7)
where b is the power of body weight (W) to which attribute Y maintains a constant
proportionality, a. A key point in this discussion of allometry is that the value of the
allometric scaling factor, b, has been shown to be relatively constant for certain attributes
relevant to toxicokinetics (Travis et al., 1990 and U.S. EPA, 1992a).
The scaling of physiological processes that relate to toxicokinetics can be drawn
together in the concept of "physiological time". The concept of physiological time
proposes that quantitative differences across mammalian species in physiological
processes can be related to similar anatomical, physiological, and biochemical machinery
operating at different rates in different size species. Thus, small species are considered to
have faster "physiological clocks," whereby various processes stay in proportion to each
other but are relatively sped up compared to larger species. This is consistent with the
observation that life spans of mammals scale roughly by W1'4. The relevance of the
physiological time concept to toxicokinetic differences across species is that, in general,
smaller species would be expected to process equal mg/kg body weight doses of toxicants
at a faster rate than larger species, owing to faster metabolic, distribution, and elimination
rates (when expressed on a per body weight basis) compared to larger species. This
would imply, based solely on these generalized trends in toxicokinetic-related processes,
that species of larger body mass would be more sensitive to the same mg/kg-body weight
adjusted dose compared to smaller species, owing to slower metabolic, distribution, and
elimination rates per gram body weight1. Clearly, chemical-specific and species-specific
factors other than the generalized trends in anatomical and physiological differences
represented by allometric scaling contribute to sensitivity differences across species.
Therefore, when applicable data on toxicokinetic, toxicodynamic, and sensitivity
differences between species are available, these data should be used to refine projections
of allometric scaling when necessary.
The previous theoretical considerations of allometric scaling of toxicokinetically-
related physiological processes, in addition to limited empirical evidence, summarized in
U.S. EPA, 1992a, has led to the explicit use of allometric scaling in the estimation of
toxicologically equivalent doses in U.S. EPA's human health cancer methodology and its
implicit use in the human health noncancer methodology. When test results are expressed
on a per body weight basis (i.e., mg/kg-d), the following equation can be used to assess an
allometric scaling factor that can be used to address some of the general toxicokinetic
differences that may occur between species. (The basis of this equation is discussed in
detail in U.S. EPA, 1992a.)
1 It should be noted that toxicity data expressed in terms of concentrations of contaminant in
environmental media (e.g., ppm in food) inherently adjust for differences predicted by allometry
since food and water consumption rates also scale in approximate proportion to W3'4.
GLWQI Technical Support Document for Wildlife Criteria Page 19
-------
Wt —
TD = TD x (--I?)4 (8)
V
where:
7D/, = test dose scaled for the given representative species (mg/kg-d),
TDT = test dose for the test species {mg/kg-d),
WTT = body weight of test species (kg), and
WTa = body weight of the given representative species (kg).
When applied to the derivation of human health criteria, this equation would yield an
allometric scaling factor of 7 when extrapolating results from a mouse to a human (i.e., the
toxicokinetically equivalent dose would be 7-fold lower in humans than mice on a mg/kg-d
basis). When applied to wildlife criteria, smaller factors typically would be derived, since
the body weight differences among test and representative species are usually smaller
compared to the differences between small laboratory mammals and human body weights.
The largest allometric scaling factors that could be expected for the GLWQI wildlife criteria
would involve potential extrapolation between the mouse and the river otter. In this
example, a factor of 4 would account for allometric scaling from a mouse to a river otter
(i.e., the toxicokinetically equivalent dose would be 4-fold lower in the otter than the TD in
the mouse), assuming body weights of 0.03 kg and 7.4 kg, respectively.
It is very important to convey that allometric scaling according to body mass is
based on broad differences in physiology between species that can mostly be related to
general toxicokinetic processes. Allometric scaling does not necessarily account for all
toxicokinetic and toxicodynamic differences that occur between species, owing to the
many complexities of toxicokinetic and toxicodynamic processes that influence an
organism's response to a toxicant. Significant departures from the predictions of allometric
scaling have been noted for some individual chemicals, some up to two orders of
magnitude in either direction. In addition, allometric scaling between typical test species
and Great Lakes representative species, which have relatively similar body weights, may
result in small allometric adjustment factors that are well within the error and uncertainty
associated with the allometric scaling equations. Therefore, allometry should be used only
as one part in the overall process of determining an interspecies UF. It should not
supersede the use of available chemical-specific information on differences in sensitivity,
toxicokinetics, and toxicodynamics across species.
VI.A.4. Empirical Basts: Variability in Acute Sensitivity
An analysis of the variability in acute sensitivity of birds and mammals (summarized
in U.S. EPA, 1995c), was carried out because the available acute toxicity data represent
the largest repository of information currently available for evaluating interspecies
variability in sensitivity to toxicants.
Three analyses of avian acute toxicity data were carried out by analyzing the results
of three studies (Schafer and Brunton, 1979; Hudson et al. 1984; and Hill et al. 1975) in
which toxicity data for each study were collected from the same laboratory using
Page 20 GLWQI Technical Support Document for Wildlife Criteria
-------
standardized experimental protocols. Each study tested a number of chemicals (21-48
chemicals in each study) and a number of species (4-13 species per chemical). Within
each study and for each chemical, the variability in species sensitivity relative to the most
sensitive species tested was illustrated by calculating "LD50 or LC50 ratios." These are the
ratios of each species' LD60 or LCSO to the lowest LD50 or LC50 observed for that chemical.
These analyses revealed that in 88 to 95 percent of the cases, the LD50 or LC50 ratios were
less than a factor of 10, with a slightly smaller range for the dietary LC50s.
To determine the variability in acute sensitivity of mammals, acute oral LD50s were
obtained and analyzed from the Registry of Toxic Effects of Chemical Substances (RTECS)
database (NIOSH, 1993). Acute data for 58 chemicals with 5-8 species tested per
chemical were analyzed in the same manner as described for birds. For all chemicals,
greater than 50% of the LD50s were within a factor of four of each other and 90% were
within a factor of 20. A factor of 100 accounted for 96% of the LD50 ratios.
The analyses of variability in acute sensitivity provide support for the 1 to 100
recommended range for the interspecies uncertainty factor for both mammals and birds.
However, in considering the results of the analyses of acute data presented above, several
caveats apply. First, the number and type of avian and mammalian species tested in these
analyses is fairly limited relative to the species that exist in nature. Also, the chemicals
represented are heavily weighted by certain classes of pesticides with a similar mode of
action. To the extent that chemicals and mechanisms of toxic action differ from those
represented in this acute database, uncertainty exists in the applicability of these
comparisons to other chemicals. Finally, these analyses are based on acute tests using
high doses; therefore, the extent to which they reflect interspecies variability in responses
from much lower, subchronic and chronic exposures is subject to uncertainty.
VI.A.5. Empirical Basis: Variability in Chronic Sensitivity
For this analysis, dietary, chronic and subchronic toxicity data were assembled from
174 separate toxicity studies on birds and mammals for four chemicals (cadmium, DDT
and metabolites, dieldrin, and mercury). Specifics about the studies included in the
database and the analyses that were carried put are provided in U.S. EPA, 1995c. In
keeping with the wildlife criteria methodology, the endpoints assessed in the studies
included in this analysis were reproductive, developmental, mortality, and growth. For this
analysis, comparisons of chronic sensitivity were only made within the most specific
endpoint, within specified exposure duration categories and when the exposure was to
analogous life stages for each species. Chronic interspecies sensitivity ratios (like the LD50
and LC50 ratios in the acute analyses) were computed within each chemical-exposure-
duration-life stage category.
A total of 122 interspecies comparisons of NOAELs for birds and mammals could be
made within the various chemical-duration-endpoint-life stage categories. Because the
database available for comparisons of species sensitivity is much smaller for chronic
effects than it was for acute effects, the avian and mammalian data were combined in
these analyses. Results from these analyses were expressed using three different
approaches to gain the most information possible from the comparisons. These
approaches are described in detail in U.S. EPA, 1995c. The results from the most
GLWQI Technical Support Document for Wildlife Criteria Page 21
-------
comprehensive analysis indicate an interspecies sensitivity ratio of 100 encompasses 84
percent of the ratios determined from this data set. There was also an indication of
overall greater variability in the chronic sensitivity compared to the ranges observed in
acute sensitivity, although other factors such as variability in chronic test designs likely
contributed somewhat to the observed variability in chronic sensitivity.
VI.A.6. Guidance on Selecting the Interspecies Uncertainty Factor (UFA) when
Deriving Tier I Wildlife Criteria
The information on species sensitivity differences briefly summarized above and
discussed in more detail in U.S. EPA (1995c), provides support for the 1-100
recommended range for the value of the interspecies uncertainty factor (UFA). However,
the actual selection of an interspecies UP for application to a particular situation must be
made on a case-by-case basis and requires the use of best professional judgment. In
determining the appropriate UFA to apply in a given situation, consideration should be given
•to the physicochemical, toxicokinetic, and toxicodynamic properties of the chemical of
concern and the amount and quality of available data. Generally, smaller values for UFA
{closer to 1) will be applied when toxicity data is available for a larger number of species
within a given taxonomic class. Consideration should also be given to the taxonomic and
physiological diversity of test species available and their relationship to the representative
species. Comparative toxicity data on compounds which operate by the same mechanism
of action should also be considered when determining the appropriate value for UFA.
Allometric scaling of doses (expressed as mg/kg-d) can also be considered in the
process of selecting an interspecies uncertainty factor based on toxicokinetic differences
using equation 8 above. However, it should be recognized that allometric scaling may not
accurately reflect the toxicokinetics of all chemicals nor encompass all the toxicodynamic
differences among species. For example, based solely on predictions of allometric scaling,
one would expect similar sensitivities between two taxonomically-related species of similar
body weights, such as for starlings, Sturnus vulgaris (75g) and red-winged blackbirds,
Agelaius phoeniceus (55g) which are both passerine species. However, when tested with
acute, oral doses to 21 organophosphate and 9 carbamate pesticides, the red-winged
blackbird was shown to be consistently more sensitive than the starling (Schafer and
Brunton, 1979). For certain pesticides, acute oral LD50s for red-winged blackbirds (mg/kg
body weight) were more than 10-fold lower than LD50s for starlings (e.g., coumaphos,
bufencarb, carbofuran, mobam) and were at least 100-fold lower for diazinon,
dichlofenthion, and trichloronat. Therefore, in determining an interspecies UF,
allometrically derived TDs should be considered only as one component in the UFA
selection process and should be used in conjunction with chemical class-specific
information on sensitivity, toxicokinetics, and toxicodynamics across species. This is
consistent with the guidance provided in the U.S. EPA Science Advisory Board
commentary (U.S. EPA, 1994) which stated that allometric relationships should not be the
sole basis for selecting an interspecies UF. To see examples of how values for UFAs are
determined, please refer to the GLWQl Criteria Documents for the Protection of Wildlife:
DDT; Mercury; 2,3,7,8-TCDD;PCBs (U.S. EPA, 1995d).
To illustrate an approach to estimating the contribution of allometry to the value of
UFA, an example based on the avian PCB wildlife value is provided here. In this example.
Page 22 GLWQl Technical Support Document for Wildlife Criteria
-------
the test species was a 1 kg pheasant and the three representative species were herring gull
(1.1 kg), bald eagle (4.6 kg), and kingfisher (0.1 5 kg). The allometric adjustment of the TD
between a pheasant and herring gull (1.0 kg vs. 1.1 kg) is negligible. The allometric
adjustment factor of a TD between a pheasant and bald eagle would be:
Wt .. Q3S 1 0.25
) - (-L) •* - 0.68 0)
Wtbald eagle *'°
The UFA(b(lW taal.| would be the inverse of this value (1 / 0.68), or 1 .5. The allometric
adjustment factor of a TD between a pheasant and kingfisher would be:
°-2S 1 0.25
) = (-L-) = 1.6 (10)
;
The UFA ,kinofilher) would be the inverse of this value (1 / 1 .6), or 0.65. These allometric
adjustment factors were not applied in the derivation of the PCB avian wildlife values
because their small magnitudes are well within the bounds of uncertainty and error
associated with the allometric relationship.
As discussed in the PCB criterion document (in U.S. EPA, 1995d), toxicity data
across a number of species were also examined, and based on best professional
judgement, a UFA of 3 was used to extrapolate from the pheasant to the eagle, to the
herring gull and to the kingfisher. Because the body masses of the test species and
representative species used in calculating the four GLWQI criteria (U.S. EPA, 1995d) are so
similar, the largest allometric scaling factors that would be calculated are about two or
less. Therefore, the use of allometric scaling in the derivation of these criteria is of limited
utility. However, if the differences in body mass were large (e.g., extrapolating from a
mouse to a moose), the use of allometry would provide much more significant insights.
VI.A.7. Guidance on Selecting the Interspecies Uncertainty Factor (UFA) when
Deriving Tier II Wildlife Values
When a Tier II wildlife value is derived, the value for UFA may range from 1 to 1000.
The larger range is provided to allow for extrapolation from the available data for one
taxonomic class to a level thought to be protective of the taxonomic class for which the
needed toxicity data was not available. However, at the discretion of the State or Tribe,
the UFA selected can be based solely on protection of the taxonomic class for which data
was available (i.e., a UFA between 1 and 100). In this case, a Tier II taxonomic class-
specific wildlife value would be identical to a Tier I taxonomic class-specific wildlife value.
In implementing this type of Tier II value, the State or Tribe should discuss the likelihood
that a Tier II value of this type will or will not be protective of the other taxonomic class.
GLWQI Technical Support Document for Wildlife Criteria Page 23
-------
VI.B. The Subchronic-to-Chronic Uncertainty Factor (UFS)
VI.8.1. Purpose and Recommended Range
Wildlife criteria that are derived according to the methodology presented in this rule
(see Appendix D to 40 CFR Part 132 of the Final GLWQI guidance, GL WQI Methodology
for the Development of Wildlife Criteria} are designed to protect wildlife from long-term
exposures to toxic substances in the diet. However, in some situations, the NOAEL
chosen to derive a wildlife value may be obtained from a study that tested organisms for a
less than chronic (i.e., subchronic) exposure duration. In these cases, a subchronic-to-
chronic uncertainty factor, UFS/ ranging between 1 and 10 is applied to the NOAEL to
account for the possibility of greater toxicity of the substance to the test organisms had
they actually been exposed over a chronic duration. The concept of a subchronic-to-
chronic UF has previously been endorsed by U.S. EPA in the Federal Register for deriving
human health criteria (U.S. EPA, 1985, and U.S. EPA, 1980). Since the conceptual basis
of the UFS is common to both wildlife and human health criteria, the reader is also referred
to the discussion of the subchronic-to-chronic UF presented in Appendix A to the GLWQI
Technical Support Document Methodologies for Human Health Criteria and Values.
However, it should be noted that in contrast to the human health criteria methodology, the
UFS for wildlife criteria is not meant to be used to compensate for deficiencies in the study
design not related to exposure duration.
VI.B.2. Technical Basis
It is well-recognized that subchronic toxicity studies may be conducted for exposure
durations of insufficient length to measure adverse effects that would be observed in
chronic tests of longer durations (National Academy of Sciences, 1977). Conceptually,
there are several reasons why subchronic toxicity tests may be inadequate for detecting
chronic adverse effects. First, exposure durations associated with subchronic studies may
be too short to quantify adverse effects that result from long-term (chronic) accumulation
of a toxicant at the site of action. Chronic exposure durations are particularly important for
substances that require long time periods to reach equilibrium in the target tissues. '
Second, subchronic studies may not quantify adverse effects that result from long-term
(chronic) failure of physiological compensation mechanisms that can mask adverse effects
over shorter exposure durations associated with subchronic tests. Third, subchronic
studies may not record latent adverse effects that can be manifested much later in an
organism's lifespan. Finally, subchronic studies may simply fail to expose a particularly
sensitive life stage of an organism that would be included in a chronic test.
Experimental support for the use of a subchronic-to-chronic uncertainty factor can
be found in two published studies: Weil and McCollister (1963) and McNamara (1976). In
the study by Weil and McCollister (1963), ratios of short-term (subchronic) to long-term
(chronic) NOAELs were determined from tests on laboratory rats. Tests were performed
with 33 different chemicals, including agricultural chemicals, food additives, chemicals
used in water treatment processes, and chemical ingredients of food packaging materials.
In these studies, subchronic tests typically lasted 90 days while chronic tests lasted
approximately 2 years. The NOAELs used in the subchronic-to-chronic comparisons were
determined from the most sensitive of a suite of whole organism, tissue-specific, and
Page 24 GLWQI Technical Support Document for Wildlife Criteria
-------
biochemical endpoints for each chemical. For comparisons of subchronic and chronic
NOAELs, Weil and McCollister determined that 50% of the 33 subchronic-to-chronic ratios
had values of 2 or less. In addition, they reported that 97% of the ratios were 9 or below.
McNamara (1976) analyzed 82 studies (encompassing 126 compounds) along with
data on the 33 compounds analyzed by Weil and McCollister (1963). The 41 additional
comparisons performed by McNamara were taken mostly from studies of rats, with a much
smaller number taken from studies of dogs and monkeys. In contrast to the analysis by
Weil and McCollister (1963), McNamara's comparisons were based on tests performed by
a wider array of researchers employing study designs of varying quality. Despite this
difference, results from the two studies appear to agree reasonably well. McNamara
reported that of the additional studies analyzed, 34 of the 41 subchronic-to-chronic ratios,
83% of the cases, were 1.0 or less. Some short-term effects were apparently reversible in
the long-term, thus yielding ratios less than 1.0. McNamara further reported that 98% of
the ratios for the 41 additional comparisons he made were less than 3, and all of the ratios
were less than 7.
In an attempt to fill some gaps in the knowledge of subchronic-to-chronic
extrapolations for avian species, additional analyses of avian chronic and subchronic
toxicity data were evaluated for four chemicals (cadmium, DDT, dieldrin, and mercury),
two of which are chemicals for which wildlife criteria were derived as part of the GLWQI
(U.S. EPA, 1995d). As a starting point, an analysis of subchronic-to-chronic extrapolations
were performed on mortality and growth endpoints for birds. Ratios of NOAELs from
short-term (28-89 days), intermediate (90-180 days), and long-term (greater than 180
days) exposure.durations were made within common chemical, species, endpoint and life
stage categories. Further details of the analysis are provided in U.S. EPA (1995c). Based
on the more comprehensive analysis of the data, results indicate about half of the long-
term NOAELs were within a factor of 10 of the short-term NOAELs and 90% were within a
factor of 20. As expected, smaller ratios were observed for intermediate-to-long-term
NOAEL comparisons, where 50% of the ratios were within a factor of 1 and 90% were
within a factor of 5. The somewhat higher variability observed in this analysis compared
to those of Weil and McCollister (1963) and McNamara (1976) is likely due in part to
greater variability in study designs evaluated in this analysis.
VI.B.3. Guidance on Selecting the Subchronic-to-Chronic Uncertainty Factor
While the preceding section provides support for the conceptual basis of a
subchronic-to-chronic UF and suggest that a range of 1-10 is reasonable, two important
issues remain that relate to when and how the UFS is applied for deriving wildlife criteria.
First and foremost is making a determination of when the application of a UFS is necessary.
This involves defining what exposure conditions constitute a reasonable definition of
"chronic". As reviewed in Appendix A: Uncertainty Factors to the GLWQI Technical
Support Document for Human Health Criteria and Values, a quantitative definition of
chronic in terms of duration remains elusive. There have been various attempts to define
chronic exposure in terms of a set duration of time, exposure across different life stages,
or exposure to a given percentage of the expected lifespan of an organism with no clear
standard emerging.
GLWQI Technical Support Document for Wildlife Criteria Page 25
-------
Qualitatively, chronic exposure can be defined as an exposure period of sufficient
length to reveal most adverse effects that occur, or would be expected to occur, over the
entire lifetime of an organism. However, the actual exposure duration required to detect
chronic effects in a given test will depend on several factors, including the toxicokinetic
properties of the substance, mechanism of toxic action, lifespan of the organism,
indications of possible latent effects, whether critical life stages of the organism were
exposed, etc. As discussed in Appendix A: Uncertainty Factors to the GLWQI Technical
Support Document for Human Health Criteria and Values, a reasonable working definition
of a minimum chronic exposure duration was defined as at least 50% of the lifespan of the
organism. For rats and mice, this would correspond to about 52 and 45 weeks,
respectively. For most bird species used in toxicology studies, this duration is also
reasonable, given that their life expectancy in the wild (which is a function of their
potential lifespan and other factors such as predation, disease, etc. that contribute to
increased mortality rates) is in the range of 2 years. However, this working definition of a
minimum chronic exposure duration should not be viewed as inflexible, and is subject to
modification depending on chemical, species, and test-specific considerations as discussed
above.
Once the determination has been made to apply a subchronic-to-chronic uncertainty
factor, actual selection of the UFS will require scientific judgement through the
consideration of several factors, including length of the exposure, toxicokinetic properties
of the chemical (e.g., bioaccumulation, metabolism, etc.), indications of latent effects,
whether important life stages were tested, mechanism of toxic action, the lifespan of the
organism, etc. Unless data indicate otherwise, larger values of the UFS will generally be
required for extrapolations from studies that are shorter relative to the lifespan of the
organism compared to studies of durations that are longer relative to the lifespan of the
organism. Thus, the concept of a "sliding scale" for selecting the UFS based on the
duration of exposure, as discussed in Appendix A: Uncertainty Factors to the GLWQI
Technical Support Document for Human Health Criteria and Values, is appropriate unless
other pertinent data indicate otherwise. Selection of the UFS should include consideration
of the amount of time required for the chemical to reach equilibrium in the tissues. All else
being equal, chemicals that require longer time periods to reach steady-state will require a
larger UFS compared to chemicals that reach steady-state relatively quickly. To see
examples of how values for UFS are determined, please refer to GLWQI Document for the
Protection of Wildlife; DDT, Mercury, 2,3,7,8-TCDD, andPCBs (U.S. EPA, 1995d).
VI.C. The LOAEL-TO-NOAEL Uncertainty Factor (UFL)
VI.C.I. Purpose and Recommended Range
In some circumstances, a wildlife value may be calculated from a study that only
provides a LOAEL without a corresponding NOAEL. This situation occurs when all the
treatments of the chosen study caused a significant adverse effect relative to the control
treatment. In these cases, U.S. EPA recommends dividing the available LOAEL (which is
actually an unbounded LOAEL as defined in Section II) by a LOAEL-to-NOAEL uncertainty
factor, UFL, to reduce it to a dose range corresponding to the expected NOAEL. U.S.
EPA's recommended range for the UFL is from 1 to 10. Use of a LOAEL-to-NOAEL
Page 26 GLWQI Technical Support Document for Wildlife Criteria
-------
uncertainty factor has been endorsed by U.S. EPA in the Federal Register for Water Quality
Criteria Documents (U.S. EPA, 1980) and in the National Drinking Water Regulations (U.S.
EPA. 1985).
VI.C.2. Technical Basis
Dourson and Stara (1983) conducted an analysis of chronic and subchronic rat data
presented by Weil and McCollister (1963) that was designed to evaluate extrapolations
from the LOAEL to the NOAEL. In their study. Dourson and Stara (1983) compared
bounded LOAELs to their corresponding NOAELs for 33 different chemicals, including
agricultural chemicals, food additives, chemicals used in water treatment processes, and
chemical ingredients of food packaging materials. Of the 52 LOAEL-to-NOAEL
comparisons made, 96% of the LOAEL-to-NOAEL ratios were 5 or less and all ratios were
10 or less.
An analysis similar to that conducted by Dourson and Stara (1983) was performed
using additional chronic toxicity data for a variety of avian and mammalian species based
on four chemicals (cadmium, DDT, dieldrin, and mercury), two of which are chemicals for
which GLWQI wildlife criteria exist (more detail is provided in U.S. EPA, 1995c). Ratios of
bounded LOAELs to their corresponding NOAELs were determined for a variety of growth,
reproductive, developmental, and mortality endpoints. Of the 275 LOAEL-to-NOAEL ratios,
more than half were less than or equal to 3, and 97% were less than or equal to 10. The
ratios are slightly higher in this analysis because many of the studies evaluated used wider
dose-spacings than did the studies in Dourson and Stara's analysis.
It should be recognized that the range of LOAEL-to-NOAEL ratios derived from these
two analyses (i.e., ratios of "bounded" LOAELs to NOAELs) are determined by the dose-
spacing of the experiments from which they were derived. Since a LOAEL-to-NOAEL
uncertainty factor is applied in situations where the available LOAEL is unbounded (for
which a corresponding NOAEL could not be determined), it is implicitly assumed that the
unbounded LOAEL is reasonably close to the dose-response threshold for the endpoint
being evaluated. To the extent that an available unbounded LOAEL reflects a response
level that is substantially greater than the dose-response threshold, additional uncertainty is
introduced in the extrapolation down to the NOAEL. Subsequently, the magnitude of the
LOAEL-to-NOAEL ratio required would increase and could conceivably extend beyond the 1
to 10 range derived from previous two analyses, depending on the slope of the dose-
response curve for the endpoint being evaluated.
VI.C.3. Guidance on Selecting the LOAEL-to-NOAEL Uncertainty Factor
The previous analyses provide general support for the recommended range of 1 to
10 for the UFL. In cases where a NOAEL cannot be quantified and only an unbounded
LOAELs is available, determination of the appropriate value for the UFL must be done on a
chemical-specific and test-specific basis with the use of best professional judgement. In
selecting the value of the UFL, both the magnitude of the response observed at the
unbounded LOAEL and the characteristics of the dose-response curve for the endpoint
which the study evaluated should be considered.
GLWQI Technical Support Document for Wildlife Criteria Page 27
-------
In considering the magnitude of the response at the unbounded LOAEL, a smaller
value for the UFL (closer to 1) could be used for unbounded LOAELs that are judged to be
at or near the dose-response threshold for the particular endpoint being evaluated. Since a
value greater than one is applied for UFL in cases where the dose-response threshold is
unknown, estimates of the magnitude of the response associated with dose-response
thresholds from other studies for the same endpoint-perhaps for a similar chemical or
species-may be useful in evaluating the proximity of the unbounded LOAEL to a threshold
response level.
In evaluating the characteristics of the dose-response curve, particular attention
should be paid to the steepness of the slope. All else being equal, unbounded LOAELs
from dose-response curves with steeper slopes would require smaller UFL's compared to
unbounded LOAELs derived from shallow slopes. However, this recommendation may not
be applicable for a shallow dose-response curve if it is judged that an unbounded LOAEL is
approaching a NOAEL threshold. To see examples of how values for UFL are determined,
please refer to GLWQI Document for the Protection of Wildlife; DDT, Mercury, 2,3,7,5-
TCDD, andPCBs (U.S. EPA, 1995d).
It should also be noted that consideration of the severity of effects (e.g., more
severe liver cell necrosis vs. less severe fatty infiltration of the liver) is incorporated into
the selection of the LOAEL-to-NOAEL uncertainty factor for human health criteria, where
protection of individuals is desired (see Appendix A: Uncertainty Factors to the GLWQI
Technical Support Document for Human Health Criteria and Values). However,
consideration of severity of effects is not to be considered when determining the value of
UFL for deriving wildlife values since a more narrowly defined set of frank effects (e.g.,
growth, reproductive and developmental impairment) is used in the context of protecting
populations.
VI.D. The Intraspecies Uncertainty Factor (UF,)
VI.D.I. Purpose and Recommended Value
In situations where protection of sensitive individuals within a population of birds or
mammals is required, U.S. EPA recommends an intraspecies uncertainty factor of 10 be
applied to the NOAEL on a site-specific basis. Such protection is required for the
protection of endangered species and discussed in Appendix F, Procedure 1 .A:
Requirements for Site-specific Modifications to Criteria and Values to 40 CFR Part 132 of
the final GLWQI guidance. States and Tribes that choose to protect sensitive individuals
for other reasons, may also use this intraspecies uncertainty factor. This uncertainty factor
is intended to reduce the chosen NOAEL, which reflects the response of average
individuals within a tested population, to a dose level that is protective of the more
sensitive individuals of the population. As discussed in Appendix F, Procedure 1 .A:
Requirements for Site-specific Modifications to Criteria and Values to 40 CFR Part 132, the
intraspecies uncertainty factor is intended for application to a NOAEL based on endpoints
closely related to effects on populations (e.g., reproductive, growth, developmental, or
lethal effects) rather than more subtle effects on individuals (e.g., biochemical responses,
behavioral changes). The intraspecies uncertainty factor for wildlife is analogous to that
Page 28 GLWQI Technical Support Document for Wildlife Criteria
-------
endorsed by U.S. EPA and the National Academy of Sciences (NAS, 1980) for deriving
human health criteria.
VI.D.2. Technical Basis
Conceptually, the basis of intraspecies variability in sensitivity can be attributed to
individual differences in toxicokinetics (factors affecting the magnitude and duration of
toxicant exposure at target site) and toxicodynamics (factors affecting the degree of
response at the target site to a given exposure level). It is the combined effect of
individual toxicokinetic and toxicodynamic differences that determines the overall variability
in susceptibility of individuals of a species to a toxicant. Sources of inter-individual
variability in toxicokinetics and toxicodynamics include, among other factors, genetic
heterogeneity, nutritional condition, changes in physiology with age, differences related to
the sex of the organism, and disease!" Examples of inter-individual variability in
toxicokinetics and toxicodynamics that pertain to human subjects are reviewed in detail in
U.S. EPA (1995c) and in Appendix A: Uncertainty Factors to the GLWQl Technical
Support Document for Human Health Criteria and Values.
Experimental support for an intraspecies uncertainty factor for wildlife criteria is
mostly limited to acute toxicity data extrapolations, where characterization of the dose-
response curve has been relatively standardized. In a re-analysis of rat LD50 data presented
by Weil (1972), Dourson and Stara (1983) analyzed 490 probit, log-dose slopes in an
attempt to provide scientific support for the intraspecies uncertainty factor used to derive
noncancer human health criteria. Although their analysis was originally intended to support
an intraspecies UF of 10 for humans, its basis on rat data also make it applicable to
mammalian wildlife criteria. Dourson and Stara (1983) report that approximately 92% of
the probit, log-dose slopes were 3 or greater, which corresponds to intraspecies
uncertainty factors of approximately 10 or less when extrapolating three probits below a
median lethal response (e.g., from an LD50 to an LD013). However, Dourson and Stara
caution that their results may be under-protective for humans by acknowledging that the
population on which their results are based (i.e., laboratory rats), may be more
homogenous in their responses compared to those of the human population. Similarly, it
seems reasonable to expect responses of laboratory rats to be more homogenous than
responses of natural populations of wildlife, which wildlife criteria are designed to protect.
Intraspecies variability in sensitivity was also investigated for birds using a similar
approach as described by Dourson and Stara (1983). Analysis of intraspecies variability in
acute sensitivity was performed on dietary LC50 data reported by Hill et al. (1975) for four
species of birds (bobwhite quail, Japanese quail, ring necked pheasant, and mallard) and
88 chemicals (mostly organophosphate, organochlorine pesticides and PCBs). Based on
the reported log concentration-probit slopes for each species-chemical combination,
extrapolations were made from the median lethal concentration (50% lethality) to a
concentration corresponding to a response rate for the most sensitive individuals (e.g., a
1 % lethality or LC,). Of the 248 extrapolations made, about 75% of the estimated LC,
values are within a factor of four of the median lethal response and 95% are within a
factor of 10.
GLWQl Technical Support Document for Wildlife Criteria Page 29
-------
In an attempt to evaluate intraspecies variability in sensitivity from longer exposure
durations that are more applicable to wildlife criteria, recent data were evaluated that
characterize the dose-response-time surface from 28-day dietary tests on bobwhite quail
(see Shirazi et at., 1994). In their study, Shirazi et al. exposed adult and juvenile quail to
seven chemicals (two organochlorines, a carbamate, an organophosphate, and three
rodenticides) in the diet for 28 days. Based on their model and reported parameter values,
comparisons of reported 28-day LC50 values to estimated 28-day LC, values were
performed. Of the 13 LC50 to LC, ratios calculated, nine were less than 10. The two
highest LC50 to LC, ratios (from two rodenticides} were near 60.
The previous experimental analyses of intraspecies variability in sensitivity provide
limited, indirect experimental support for an intraspecies UF and indicate that the majority
of extrapolations to the more sensitive individuals are within a factor of 10 of the median
response. However, several caveats to the empirical analysis described above should be
noted. First, because quantitative analyses of the chronic dose-response curves are rare,
much of the experimental evidence is based on acute exposures at high doses, all of which
are for measures of mortality. Therefore, the relationship between intraspecies variability
observed in these data to variability in other chronic responses (such as reproductive and
developmental effects) at lower doses is unclear. Second, comparisons to extrapolated
low response rates (e.g., LC013 or LC,) are inherently uncertain, and no attempt has been
made in this evaluation to characterize this uncertainty. Finally, intraspecies variability in
sensitivity has been characterized for relatively few species, and those are somewhat
taxonomically removed from the GLWQI representative wildlife species. Therefore, these
analyses suggest that using an intraspecies uncertainty factor of 10 is reasonable and is
not likely to be unduly conservative.
VII. DETERMINATION OF EXPOSURE PARAMETER VALUES
Exposure parameters for deriving a wildlife value are of two types: (1) chemical-
specific BAFs and (2) wildlife species-specific weights, and food and water ingestion rates.
Section VILA, describes the chemical-specific exposure parameters. Section VII.B.
describes the species-specific trophic levels. Section VII.C. describes how to identify or
estimate values for the exposure parameters, and Section VII.D. describes the derivation of
the exposure parameter values for the representative wildlife species used to derive the
wildlife criteria.
VILA Chemical-specific Bioaccumulation Factors and Biomagnification Factors for DDT
and metabolites; Mercury; 2,3,7,8-TCDD; and PCBs
A BAF is necessary to estimate the concentration of the chemical in the wildlife
food source based on its concentration in the water source. Appendix G to the GLWQI
Technical Support Document for the Procedure to Determine Bioaccumulation Factors
presents the baseline BAFs from which BAFs for the derivation of wildlife values can be
calculated. The procedure to derive these BAFs is specified in Appendix B to Part 132 of
the final GLWQI guidance, Great Lakes Water Quality Initiative Methodology for Deriving
Bioaccumulation Factors.
Page 30 GLWQI Technical Support Document for Wildlife Criteria
-------
In addition to BAFs for aquatic prey of wildlife species, the eagle also consumes
piscivorous birds. To estimate the concentration of contaminant in piscivorous birds (i.e.,
gulls) which are in turn consumed by eagles, a BMP is needed. The BMFs used for the
derivation of wildlife criteria for the four chemicals for which wildlife criteria exist are
presented in Appendix K to GLWQI Technical Support Document for the Procedure to
Determine Bioaccumulation Factors, entitled Determination of BAFs for DDT and
Metabolites and Biomagnification Factors for the Derivation of Wildlife Criteria. The BMFs
are the ratio of the concentration of a contaminant in the gulls to the concentration in their
prey fish.
VII. B Species-specific Trophic Levels
To determine trophic levels for the aquatic prey of the representative wildlife
species, the dietary habits of both the wildlife species and their prey were investigated.
The concept of numerical average trophic level was used to account for diets that
represent prey from different trophic levels in a food web. For example, if half of an
animal's diet came from trophic level 3 and half from trophic level 4, the numerical average
trophic level for that animal's prey would be 3.5. Mearns et al. (1981) introduced the
concept of numerical average trophic levels for marine predators that feed at more than
one level, and Sanger (1987) has applied this concept to seabirds that feed at more than
one trophic level (Hobson, 1990). Trophic levels were estimated first for the aquatic prey
and then for the wildlife species that feed on them, as described below.
Trophic levels for aquatic prey species were determined from a literature search on
their dietary habits and the dietary habits of their prey. For aquatic species that change
their dietary habits as they grow in size (e.g., trout, perch), the species was categorized
into two to five size-classes to reflect the change in dietary habits. For a single dietary
study (i.e., one species at one location over a specified time frame), the weighted average
trophic level for the species (by size class) is calculated as:
ATL = [(TL^., * %Vwl) + (TL^.2 * %Vw2) ... (TLprey.n * %Vprev.n)] + 1
where
ATL = weighted average trophic level for the fish species of interest;
TLpf«y-n = trophic level for prey n; and
%Vpr.y.n = percent volume of prey n in the diet.
For aquatic species for which dietary studies from different locations were available, U.S.
EPA developed an estimate of the range of trophic levels that might be exhibited by the
species under different environmental conditions (U.S. EPA, 1995b).
To determine feeding habits of the piscivorous wildlife selected for evaluation, data
in Volume II of U.S. EPA's (1993d) Wildlife Exposure Factors Handbook were reviewed.
For the values in the Handbook cited from secondary sources, the original literature was
retrieved and reviewed. To identify the size of prey captured by the wildlife species,
GLWQI Technical Support Document for Wildlife Criteria Page 31
-------
important for determining the trophic level of fish, all wildlife dietary studies that had been
used to develop the Handbook were reviewed for information on prey size.
To estimate an average trophic level for a wildlife species in a given location, a
similar approach was used to that described above for aquatic prey, except that the
terrestrial and wetland components of the wildlife species' diet were separated from the
aquatic components. Dietary studies from the Great Lakes region were preferred, but data
from other locations were used to help define the wildlife species' range in trophic level of
prey nationwide and the potential for variation within the Great Lakes region. The analyses
are presented in detail in U.S. EPA's (1995a,b) Trophic Level and Exposure Analyses for
Selected Piscivorous Species, Volumes I and III, Draft (Volume II looks at wildlife from a
nationwide, instead of a Great Lakes, perspective).
As indicated above, the results of the analysis were numerical average trophic levels
for each wildlife species. As indicated in Section VILA., BAFs were developed for discrete
integer trophic levels, i.e., for trophic levels 3 and 4, not for trophic level 3.2 for example.
The non-integer average trophic level estimates were used to determine what proportion,
on average, of trophic level 3 and trophic level 4 would be consumed by the wildlife
species. For example, an average trophic level of 3.2 of the prey of an otter implies that
the otter consumes on average 80 percent trophic level 3 and 20 percent trophic level 4
prey.
VII.B.I. Mink (Mustela vis on)
The diet of mink consists primarily of prey linked to aquatic ecosystems, including
fish, crayfish, frogs, muskrat, and waterfowl. Mink also consume terrestrial prey that are
not associated with aquatic food chains, for example, shrews, mice, and voles (U.S. EPA
1993d). Most of the prey that mink consume from aquatic and wetland ecosystems
nationwide represent trophic level 2 (muskrats, waterfowl in winter) to trophic level 3
(frogs, most fish in the size range captured, waterfowl in summer), with an average trophic
level around 2.5 (including crayfish estimated to be trophic level 2.4), depending on prey
availability. Alexander (1977) estimated dietary composition of mink in Michigan on the
basis of percent wet weight of stomach contents from mink collected year round (U.S.
EPA, 1995a). From this study, the average trophic level of the prey of mink was estimated
to be 2.9 (for the size of fish captured by mink) and the proportion of the diet taken from
strictly aquatic food webs ranged from 75 to 90 percent, depending on habitat. Sealander
(1943) also studied mink in Michigan but during the winter only, and found a higher
proportion of their diet to consist of prey from strictly terrestrial ecosystems (e.g., 20 to 30
percent) and a lower proportion to consist of strictly aquatic prey (about 10 percent),
probably because of the winter ice cover. Other mink populations seem to obtain less of
their food from aquatic and more from terrestrial sources (e.g., rabbits, voles, ground
squirrels) year-round and during the summer than do mink in Michigan. However, most
other studies have measured dietary composition by percent frequency of occurrence,
which may not reflect percent biomass very well.
One difficulty for estimating aquatic trophic levels for mink is that the prey of most
mink populations include a relatively high proportion of wetland animals, such as muskrat,
waterfowl, and amphibians (U.S. EPA, 1995a, b). Ideally, one would estimate separate
Page 32 GLWQI Technical Support Document for Wildlife Criteria
-------
bioaccumulation factors for contaminants in water relative to tissue concentrations in
herbivorous muskrat, insectivorous waterfowl, and cold-blooded amphibians. In the
absence of bioaccumulation data to support this level of detail, however, it is reasonable to
assume that mink in the Great Lakes region are similar to mink in Michigan. It is assumed
that mink obtain about 10 percent of their prey from terrestrial food webs and 90 percent
from aquatic food webs on average for significant portions of the year. It is also
reasonable to assume that their aquatic prey average trophic level 3 on the basis of
Alexander's (1977) study.
VII.B.2 River Otter (Lutra canadensis]
In general, otters feed primarily on trophic level 3 fish, although they do catch some
higher trophic level fish, including medium-sized piscivorous fish such as northern pike,
walleye, and large trout. Knudsen and Hale (1968) estimated the dietary composition of
otters in the Great Lakes region of Wisconsin, Michigan, and Minnesota on the basis of
percent wet volume of the stomach contents of otters trapped by hunters. Based on this
study, U.S. EPA estimated the average trophic level for the aquatic prey of these otters to
be 2.9 and the average trophic level of the wetland prey (amphibians) to be trophic level
3.2 (U.S. EPA, 1995a). Lagler and Ostenson (1942) also estimated dietary composition of
otters trapped in Michigan based on percent wet volume of stomach contents. Otters
feeding in trout waters consumed on average trophic level 3.0 prey, while otters feeding in
non-trout waters consumed on average trophic level 3.4 prey. Thus, it would be
reasonable to assume an average trophic level of 3.2 for the aquatic prey of otter in the
Great Lakes region. Depending on their availability, otters may consume waterbirds, but
the amount of birds consumed is a minimal fraction of the diet and need not be quantified.
VII.B.3 Belted Kingfisher (Ceryle alcyon)
Belted kingfishers feed predominantly in shallow waters, capturing unusually large
fish for their small body size. Although no feeding studies from the Great Lakes were
identified, studies of belted kingfishers from almost any habitat feed exclusively on aquatic
prey representing, on average, trophic level 2.7 to 3.0 (U.S. EPA, 1995a). Studies of
belted kingfishers feeding on rivers and streams in Michigan indicate an average aquatic
trophic level of 2.7 to 2.9 accounting for 70 to 99 percent of the diet, depending on the
study. The remaining prey are primarily frogs. U.S. EPA (1995a,b) illustrated that many
kingfisher populations are likely to feed on average at trophic level 3. Therefore, assuming
100 percent of the kingfisher's diet is trophic level 3 would be both reasonable and
protective for populations in the Great Lakes basin.
As indicated earlier, the common tern is similar to the belted kingfisher in body
weight. Courtney and Blokpoel (1980) found that common terns in one of two study
locations in Lake Erie consumed almost 100 percent smelt and alewife in the spring. U.S.
EPA (1995a) estimated that common terns feed on prey that average trophic level 2.9 in
Lake Erie. Therefore, the trophic level assumed for the kingfisher is adequate to cover the
feeding trophic level of the common tern as well.
GLWQI Technical Support Document for Wildlife Criteria Page 33
-------
VII.B.4 Herring Gull (Larus argentatus)
Fish usually considered trophic level 3 (e.g., small freshwater drum, alewife, smelt)
comprise a large proportion of herring gull diets in the Great Lakes. Fox et al. (1990) and
Ewins et al. (unpublished) found that during the breeding season, herring gulls feed
primarily on alewife, and to a lesser extent on rainbow smelt. Based on these studies, U.S.
EPA (1995a) estimated the average trophic level of the aquatic prey on which herring gulls
feed in the Great Lakes to be 3.2.
Herring gull populations may feed on terrestrial organisms for some period of time.
Ewins et al. (unpublished) found small mammals to account for between 1 and 98 percent
of the diet for a given year, season, and site in the Great Lakes region. A diet consisting
primarily of small herbivorous mammals (Trophic Level 2 in a terrestrial system) would
effectively remove the gulls from exposure to aquatic food chains. On average, terrestrial
sources may account for 10 percent of the herring gulls diet (U.S. EPA, 1995a,b).
VII.B.5. Bald Eagle (Haliaeetus leucocephalus)
Bald eagles prey on a variety of dead and living aquatic and terrestrial prey. They
are opportunistic, and take whatever prey are readily available, although they appear to
prefer live fish when possible. Thus, the trophic level at which bald eagles feed is likely to
change with location, season, and year. One study of bald eagle dietary habits in the
Great Lakes has been published (Kozie and Anderson, 1991; from Kozie, 1986) which
assessed eagles nesting on islands and along the shore of Lake Superior in Wisconsin.
This population fed both on fish from the Great Lakes and a wide diversity of birds. Data
in U.S. EPA (1995a) indicate that the average trophic level of the fish component of the
bald eagles' diet in Lake Superior is 3.2. Although bald eagles are large birds and can
capture and carry large trophic level 4 fish, most of the time they capture the more
abundant smaller fish (U.S. EPA, 1995a,b).
The herring gull comprises a portion of the bald eagles' diet on Lake Superior. In
the Great Lakes, herring gulls feed primarily on alewife and smelt (see previous section).
Most of the observations of the diet of bald eagles nesting on Lake Superior were from
prey remains under the nest, which tend to overestimate the proportion of birds in the diet
because the fish remains degrade faster than the bird remains. Data on both prey remains
under the nest and observations of prey delivered to the nest on one island allow one to
estimate the bias inherent in examining the prey remains only. As described in U.S. EPA
(1995a), an adjustment factor was developed to estimate the relative number of birds and
fish delivered to a nest on the basis of remains under the nest alone.
Using the adjustment factor described above, U.S. EPA estimated that on average,
92 percent of the diet of bald eagles in Lake Superior was comprised of fish, and 8 percent
was comprised of birds or mammals. Of the birds/mammals component of the diet, herring
gulls comprised 70 percent (on a wet-weight basis) of the biomass. Thus, herring gulls
comprised 5.6 percent of the total bald eagle diet. For one bald eagle pair nesting near a
gull colony, the proportion of herring gulls in the diet was higher, 12.5 percent of the total
diet.
Page 34 GLWQI Technical Support Document for Wildlife Criteria
-------
VII.B.6
Summary
Tables 6 and 7 provide a summary of the information on dietary composition and
trophic level of the prey for the five representative wildlife species.
Table 6. Composition of Diet and Average Trophic Level of Aquatic Prey for
Representative Wildlife Species.
Species
mink
river otter
belted kingfisher
herring gull
bald eagle
% Aquatic :
% Terrestrial
Prey
90 : 10
100 : 0
100 : 0
90 : 10
92 : 8
Average
Trophic Level of
Aquatic Prey
3.0
3.2
3.0
3.2
3.2
Of the Aquatic
Prey,
% TL3 : % TL4
100%
80% : 20%
100%
80% : 20%
80% : 20%
Of the Terrestrial
Prey, %
Piscivorous Birds
0 %
NA
NA
0 %
70 %
Table 7. Percent Composition of Diet by Wet Weight for the Representative Species.
Species
mink
river otter
belted
kingfisher
herring gull
bald eagle
% TL 3 Fish
90%
80%
100%
0.90 xO.8
x 100 =
72%
0.92 x 0.8 x
100 = 74%
% TL 4 Fish
0%
20%
0%
0.90 x 0.2
x 100 =
18%
0.92 x 0.2 x
100 = 18%
% Piscivorous
Prey
0%
0%
0%
0%
0.08 x 0.7
X 100 =
5.6%
% Non-
piscivorous Prey
10%
0%
0%
10%
0.08 x 0.3
x 100 = 2.4%
Total
%
100
%
100
%
100
%
100
%
100
%
VII.C Species-specific Values for Body Weights and Water Ingestion Rates
As indicated in Section III, a body weight and food and water ingestion rates are
required to estimate a wildlife value for each representative species in order to derive the
wildlife criterion. This section explains how values can be identified for these exposure
GLWQI Technical Support Document for Wildlife Criteria
Page 35
-------
parameters for wildlife in general. Section VII.D describes the selection of specific
exposure parameter values for the five representative species for the Great Lakes basin.
VI1.C.1 Body Weights
Adult body weights can vary by location, and often between sexes. In developing
wildlife criteria, the body weights appropriate for the area or region in question should be
used.
There are several sources that list body weight information for birds and mammals
that might be of interest. For several species of both birds and mammals, a review of body
weight data can be found in U.S. EPA's (1993c,d) Wildlife Exposure Factors Handbook.
Dunning (1984) has compiled the body weights of 686 bird species from various sources.
Body weights for mammalian species are included in Chapman and Feldhamer's (1982)
Wild Mammals of North America, and in the Special Publication series of the American
Society of Mammalogy. For species for which body weights cannot be found in these
sources, a search through the open literature would be necessary.
V11.C.2 Drinking Water
Drinking water rates are quite variable among species and even within a species.
The degree to which an animal must drink free water (e.g., from a pool) depends on the
moisture content of its food as well as the animal's physiology (e.g., water economy).
Many species in mesic habitats satisfy the bulk of their water requirements with the
moisture in their food. In these habitats, drinking free water becomes more important
during the higher temperatures of summer than at other times of the year.
There are few measures of the rate at which wildlife species drink free water. More
information is available on the total water flux of wildlife species (e.g., Bartholomew and
Cade, 1963; Nagy and Peterson, 1988), of which drinking free water is only one
component. Calder and Braun (1983) reviewed available data on the drinking rates of
wildlife species and developed allometric equations that can be used to estimate drinking
water rates from body weights for birds and for mammals:
Birds: W (L/day) = 0.059Wt°-87 (kg), and
Mammals: W (L/day) = 0.099 Wt090 (kg).
where
W = rate of ingesting drinking water expressed as liters per day, and
Wt = the average weight in kilograms of the population or species.
VII.C.3 Food Ingestion Rates
Food ingestion rates are less variable within animals of the same age and species
than are water ingestion rates. Food ingestion rates have been estimated for a few wildlife
Page 36 GLWQI Technical Support Document for Wildlife Criteria
-------
species in the field (see U.S. EPA, 1993c). For the many wildlife species for which free-
living food ingestion rates have not been measured or estimated, Chapter 4 of U.S. EPA's
(1993c,d) Wildlife Exposure Factors Handbook explains how to estimate the rates.
Food ingestion rates can be calculated from the energetic needs of an animal, the
composition and caloric content of its diet, and the efficiency with which the energy in the
diet is assimilated. Basically, the first step is to identify or to estimate the average free-
living metabolic rate (FMR) for the species (see Section VII.C.4). The FMR establishes the
metabolizable energy (ME) needs of the species in the wild. ME is equal to the gross
energy (GE) in the diet multiplied by the assimilation efficiency (AE) for the species of that
diet:
ME (kcal/g diet) = GE (kcal/g diet) x AE.
The Wildlife Exposure Factors Handbook provides tables from which the GE of various prey
can be estimated and from which the AE of different groups of animals consuming
different types of prey can be estimated. Also note the water content of diet. Seeds tend
to contain little water (9-10% water) whereas the tissues of fish, birds, and mammals tend
consist of 75 percent water (U.S. EPA 1993c).
Examples of how food ingestion rates are estimated from body weight and dietary
composition can be reviewed in Section VII.D, where the derivations of food ingestion
rates for the five representative species selected for the GLWQI are described.
VII.C.4 Free-living Metabolic Rates
The free-living metabolic rates of several wildlife species have been estimated using
doubly-labeled water measurements of C02 production. Using the data developed by the
few laboratories that use the technique well, Nagy (1987) developed allometric equations
relating free-living metabolic rate (FMR) to body weight for birds and mammals. Nagy
(1987) includes four allometric equations for birds that include a reasonable number and
taxonomic breadth of avian species:
All birds: FMR (kcal/day) = 2.601 Wt0640 (g),
Passerine birds: FMR (kcal/day) = 2.123Wt°-749 (g),
Seabirds: FMR (kcal/day) = 1.916 Wt0-704 (g), and
Non-passerines: FMR (kcal/day) = 1.146 Wt0-749 (g).
Passerines have higher metabolic rates for their body size than do other birds. For small
birds, the difference between the passerine and non-passerine and between the passerine
and all-birds allometric equations are relatively large. For large birds, all of these equations
converge and provide similar estimates.
Nagy (1987) similarly provides several allometric equations for mammals. There are
four equations that are likely to be useful for North American mammals:
GLWQI Technical Support Document for Wildlife Criteria Page 37
-------
Placental mammals: FMR (kcal/day) = 0.800 Wt°-813 (g).
Herbivores: FMR (kcal/day) = 1.419 Wt°-727 (g).
Non-herbivores: FMR (kcal/day) = 0.6167 Wt°-8S2 (g), and
Rodents: FMR (kcal/day) = 2.514Wtas°7 (g).
Nagy (1987) and U.S. EPA (1993c)can be consulted to calculate the 95 percent
confidence limits on an estimate of FMR for any single body weight estimate.
VII.D. Exposure Parameter Values for the Representative Wildlife Species
This section describes the sources of information for and derivation of exposure
parameter values for the five representative wildlife species.
VII.D.1 Mink (Mustela vison)
Body weight. Mink body weights vary greatly throughout the species' range (adult
males reaching 1.4 kg in the east and 2.3 kg in the west, according to Harding, 1934,
cited in Linscombe et al., 1982). Males weigh markedly more than females (in some
populations, almost twice as much). U.S. EPA's (1993c) Wildlife Exposure Factors
Handbook had limited data on body weights of wild mink. From those data, body weights
for mink from Montana during the summer are listed in Table 8. Thus, an average body
weight for both sexes of 0.8 kg would be a reasonable value to assume for wild mink.
Table 8. Body Weights of Mink Populations.
Age/Sex/Season
adult/male/summer
adult/female/summer
average
Body Weight (kg)
1.04(N = 5)
0.550 (N = 25)
0.78
Reference/Location
Mitchell 1961/
Montana
Water ingestion rate. No measured values for mink drinking water ingestion rates
were identified. The moisture content of their diet of fish may satisfy much of their daily
water requirements. In the absence of specific information, however, Calder and Braun's
(1983) allometric equation for estimating drinking water ingestion rates predicts that mink
averaging 0.8 kg drink 0.081 L/day:
Water Ingestion (L/day) = 0.099 Wt090 (kg) = 0.099 (0.8)°90 = 0.081 L/day.
Food ingestion rate. No measured values for free-living mink food ingestion rates
were found. Nagy's (1987) allometric equation for estimating free-living metabolic rate for
non-herbivorous mammals predicts that mink weighing 0.80 kg require 200 kcal per day:
Page 38 GLWQI Technical Support Document for Wildlife Criteria
-------
FMR (kcal/day) = 0.6167 Wt0862 (g) = 0.6167 (800)°-862 = 196kcal/day.
Assuming that the fish consumed by mink provide 1.2 kcal gross energy (GE)/g wet weight
and that the assimilation efficiency (AE) of mink consuming fish is 91 percent (U.S. EPA,
1993c, Tables 4-1 and 4-3), the metabolizable energy (ME) in fish for mink would be 1.1
kcal/g wet weight:
ME (kcal/g fish) = GE x AE = 1.2 (kcal/g) x 0.91 = 1.09 (kcal/g fish)
Assuming that the GE of birds and mammals consumed by mink is 1.8 kcal/g wet weight
and that the AE of mink consuming birds and mammals is 84 percent (U.S. EPA, 1993c),
the ME in birds and mammals for mink would be 1.5 kcal/g wet weight:
ME (kcal/g birds-mammals) ="<3E x AE = 1.8 (kcal/g) x 0.84 = 1.51 (kcal/g birds-
mammals)
Thus, if mink consumed only fish, they would require 180 grams of fish daily (196 kcal/day
divided by 1.09 kcal/g fish). If mink consumed only birds and mammals, they would
require 130 grams of birds or mammals daily (196 kcal/day divided by 1.51 kcal/g birds or
mammals). Given the assumption that mink consume 90 percent fish and 10 percent birds
and mammals on a wet-weight basis, however, the total grams of each is estimated as
follows:
If
Y = grams of birds or mammals consumed,
9 Y = grams of fish consumed,
then
Y (g) x 1.51 (kcal/g birds-mammals) + 9Y (g) x 1.09 (kcal/g fish) = 200
kcal,
1.51 Y (kcal) + 9.81 Y (kcal) = 200 kcal,
11.3Y = 200,
and
Y = 17.7 grams of birds-mammals consumed, and
9 Y = 159 grams of fish consumed.
As indicated in Table 5, all of the fish consumed are assumed to represent trophic level 3.
VII.D.2 River Otter (Lutra canadensis)
Body weight. Sexual dimorphism in size is seen among all subspecies of river otters
(Toweill and Tabor, 1982), and adult males (5 to 10 kg) outweigh females (4 to 7 kg) by
approximately 17 percent (Melquist and Hornocker, 1983; see Table 9). No body weights
are available for otters in the vicinity of the Great Lakes. Given the large sample size and
distributional data provided for the Alabama otter studied by Lauhachinda (1978), these
data were used to estimate an average river otter body weight of 7.4 kg. If river otters
tend to be larger at more northerly latitudes (Bergman's rule), this value may underestimate
otter body weights in the Great Lakes region.
GLWQI Technical Support Document for Wildlife Criteria Page 39
-------
Table 9. Body Weights of River Otter Populations.
Age/Sex
adult female
adult male
average
adult female
adult male
average
adult female
adult male
average
Body Weight (kg)
6.7 ± 1.0SD (N = 71) (range 4.74 -8. 72)
8.1 ± 1.2SD (N = 153) (range 5.84- 10.4)
7.40
7.9 ± 0.2SE(N = 6)
9.2 ± 0.6 SE (N=4)
8.55
7.0 (N = 100)
8.3 (N = 138)
7.65
Reference/
Location
Lauhachinda
1978/Alabama,
Georgia
Melquist &
Hornocker
1983/ldaho
Wilson 19597
North Carolina
Water ingestion rate. No measured values for river otter drinking water ingestion
rates were identified. The moisture content of their diet of fish may satisfy much of their
daily water requirements. In the absence of specific information, however, Calder and
Braun's (1983) allometric equation for estimating drinking water ingestion rates predicts
that otter averaging 7.4 kg drink 0.60 L/day:
Water Ingestion (L/day) = 0.099 Wt090 (kg) = 0.099 (7.4)090 = 0.60 L/day.
Food ingestion rate. No measured values for free-living river otter food ingestion
rates were found. Nagy's (1987) allometric equation for estimating free-living metabolic
rate for non-herbivorous mammals predicts that otters weighing 7.4 kg require 1,335 kcal
per day:
FMR (kcal/day) = 0.6167 Wt0-862 (g) = 0.6167 (7400)°862 = 1,335 kcal/day.
Assuming that the GE of the prey (i.e., fish) of the otter be 1.2 kcal/g wet weight and the
AE of otter consuming fish is 91 percent (U.S. EPA, 1993c), the ME in fish for otters
would be 1.1 kcal/g wet weight:
ME (kcal/g fish) = GE x AE = 1.2 kcal/g fish x 0.91 =1.09 kcal/g fish.
Thus, to satisfy the daily requirement for 1,335 kcal ME, the food ingestion rate (F) of an
otter would need to be 1,220 g of fish per day:
F (kg/day) = (1,335 kcal/day)/(1.09 kcal/g wet weight fish) = 1,220g/day.
As indicated in Table 6, the fish consumed by otter are assumed to consist of 80 percent
trophic level 3 fish and 20 percent trophic level 4 fish. Thus, the trophic-level-specific fish
ingestion rates for river otters can be estimated as:
Page 40
GLWQI Technical Support Document for Wildlife Criteria
-------
F (kg/day) trophic level 3 fish = 1.22 x 0.80 = 0.976 kg TL3 fish/day, and
F (kg/day) trophic level 4 fish = 1.22 x 0.20 = 0.244 kg TL4 fish/day.
VII.D.3 Belted Kingfisher (Ceryle alcyon)
Body weight. The sexes are similar in size, and a value of 0.148, rounded to 0.15
kg would be a representative weight for this species (see Table 10).
Table 10. Body Weights of Belted Kingfisher Populations.
Age/Sex
adults/both sexes
adults/both sexes
adults/both sexes
adults/both sexes
Body Weight (kg)
0.148 ± 0.0208 SD {N = 29)
0.136 ± 0.156 SE(N = 5)
0.150 (N = 98)
0.158 ± 0.115 SE(N = 11)
Reference/Location
Powdermill Nature Center (unpub.)
cited in Dunning,
1 984/Pennsylvania
Brooks & Davis, 1 987/Pennsylvania
Alexander, 1 977/Michigan
Brooks & Davis, 19877 Ohio
Water ingestion rate. No measured values for kingfisher drinking water ingestion
rates were identified. The moisture content of their diet of fish may satisfy much of their
daily water requirements. In the absence of specific information, however, Calder and
Braun's (1983) allometric equation for estimating drinking water ingestion rates predicts
that kingfishers averaging 0.15 kg drink 0.017 L/day:
Water Ingestion (L/day) = 0.059 Wt087 (kg) = 0.059 (0.148)067 = 0.017 L/day.
Food ingestion rate. No measured values for free-living kingfisher food ingestion
rates were found. Nagy's (1987) allometric equation for estimating free-living metabolic
rate for all birds predicts that kingfishers weighing 0.15 kg require 64 kcal per day, or 430
kcal/kg kingfisher-day:
FMR (kcal/day) = 2.601 Wt0-640 (g) = 2.601 (148)0640 = 63.7 kcal/day, or
FMR (kcal/kg-day) = (63.7 kcal/day)/(0.148 kg) = 430 kcal/kg-day.
This might overestimate the food ingestion rate for kingfishers because they are not in the
order Passeriformes, and therefore the non-passerine allometric equation would be more
appropriate:
FMR (kcal/day) = 1.146Wt°-749 (g) = 1.146 (148)°-749 = 48.4 kcal/day, or
FMR (kcal/kg-day) = (49.4 kcal/day)/(0.148 kg) = 334 kcal/kg-day.
However, the exposure parameter values for the belted kingfisher are assumed to be
representative of the likely exposures of common terns as well as belted kingfishers.
GLWQI Technical Support Document for Wildlife Criteria
Page 41
-------
Nagy's (1987) allometric equation for estimating the free-living metabolic rate for seabirds
predicts that common terns weighing 0.12 kg (LeCroy and LeCroy, 1974; Dunning, 1984)
require 56 kcal/day, or 467 kcal/kg-day:
FMR (kcal/day) = 1.916 Wta704 (g) = 1.916 (120)°704 = 55.7 kcal/day, or
FMR (kcal/kg-day) = (55.7 kcal/day)/(0.120 kg) = 467 kcal/kg-day.
Thus, the daily energy requirement of 63.7 kcal/day predicted from Nagy's (1987)
allometric equation for all birds is used for the belted kingfisher and is intended to represent
the caloric needs of the common tern relative to its body weight.
Assuming that the fish consumed by the kingfisher (or common tern) provide 1.2
kcal GE/g wet weight and that the AE of kingfishers consuming fish is 79 percent (U.S.
EPA, 1993c),the ME in fish for belted kingfishers would be 0.948 kcal/g wet weight:
ME (kcal/g fish) = GE x AE = 1.2 kcal/g fish x 0.79 = 0.948 kcal/g fish.
Thus, to satisfy the daily requirement for 63.7 kcal ME, the fish ingestion rate (F) of a
belted kingfisher would need to be 67.2 g of fish per day:
F (kg/day) = (63.7 kcal/day)/(0.948 kcal/g wet weight fish) = 67.2g/day.
As indicated in Table 6, it is assumed that all of the 67.2 g of fish consumed by the
kingfisher are trophic level 3 fish.
VII.D.4. Herring Gull (Laws argent a tus)
Body weight. Adult herring gulls generally weigh between 1.0 and 1.2 kg, with the
males being slightly heavier than the females (Table 11). An average weight for both
sexes would be 1.1 kg. The data indicating slightly lower body weights from Lake Huron
in Table 11 is not used for the GLWQI because of the small number of individuals of each
sex in that sample.
Water ingestion rate. No measured values were found for the water ingestion rates
of herring gulls. The moisture content of their diet may satisfy much of their daily water
requirements. In the absence of specific information, however, Calder and Braun's (1983)
allometric equation for estimating drinking water ingestion rates predicts that herring gulls
averaging 1.1 kg in weight drink 0.063 L/day:
Water Ingestion (L/day) = 0.059 Wt°-fl7 (kg) = 0.059 (1.1)067 = 0.063 L/day.
Page 42 GLWQI Technical Support Document for Wildlife Criteria
-------
Table 11. Body Weights of Herring Gull Populations.
Age/Sex
adult/female
adult/male
average
adult/female
adult/male
average
adult/female
adult/male
average
Body Weight (kg)
1.00 ± 0.090 SD (N = 78)
(range = 0.83 - 1.27)
1.23 ± 0.107 SD (N = 180)
(range = 1.01 - 1.62)
1.12
0.920 ± 0.057 SD(N = 10)
1.05 ± 0.058 SD (N = 7)
0.99
1.04 (N = 139; range = 0.717 - 1.39)
1 .23 (N = 220; range = 0.755 - 1 .50)
1.14
Reference/Location
Threlfall and Jewer, 1 978/
Newfoundland, Canada
Norstrom et al., 1 9867
Lake Huron
Belopolskii, 1 957 cited in Dunning,
1984/Barent Sea (arctic)
Food ingestion rate. No measured values for free-living herring gull food ingestion
rates were found. Nagy's (1987) allometric equation for estimating free-living metabolic
rate for seabirds predicts that herring gulls weighing 1.1 kg require 265 kcal per day:
FMR (kcal/day) = 1.916 Wt°-704 (g) = 1.916 (1100)°-704 = 265 kcal/day.
Norstrom et al. (1986) have estimated an annual energy budget for free-living female
herring gulls that breed in the Great Lakes. Between September and March, the non-
breeding season, they estimate that adult females require 250 to 260 kcal/day. Following
a dip in energy requirements when the male feeds the female during courtship, Norstrom et
al. (1986) estimated that the female's needs increase to peak at 280 kcal/day for egg
production, then fall to approximately 210 kcal/day during incubation. The average over
this period is likely to be similar to the estimate for the non-breeding season. Thus, the
estimate of 265 kcal/day from Nagy's (1987) allometric equation for seabirds is considered
a reasonable estimate of the ME requirement of free-living herring gulls
Assuming that the GE of fish consumed by herring gulls is 1.2 kcal/g wet weight
and that the AE of gulls consuming fish is 79 percent (U.S. EPA, 1993c),the ME in fish for
gulls can be estimated:
ME (kcal/g fish) = GE x AE = 1.2 (kcal/g) x 0.79 = 0.948 (kcal/g fish).
Assuming that the birds and mammals consumed by herring gulls provide 1.8 kcal GE/g
wet weight and that the AE of seabirds is the same as the AE of birds of prey consuming
birds and mammals (i.e., 78 percent; U.S. EPA, 1993c), the ME in birds and mammals for
the herring gull would be 1.40 kcal/g wet weight:
ME (kcal/g birds) = GE x AE = 1.8 (kcal/g) x 0.78 = 1.404 (kcal/g birds).
Thus, if herring gulls consumed only fish, they would require 280 grams of fish daily (265
kcal/day divided by 0.948 kcal/g fish). If herring gulls consumed only birds and small
GLWQI Technical Support Document for Wildlife Criteria
Page 43
-------
mammals, they would require 189 grams of birds/mammals daily (265 kcal/day divided by
1.404 kcal/g birds). Given the assumption that herring gulls consume 90 percent fish and
10 percent birds and mammals on a wet-weight basis (Table 6), the total grams of each is
estimated as follows:
If
then
kcal.
and
Y = grams of birds-mammals consumed, and
9 Y = grams of fish consumed
Y (g) x 1.404 (kcal/g birds-mammals) + 9 Y (g) x 0.948 (kcal/g fish) = 265
1.404Y (kcal) + 8.532 Y (kcal) = 265 kcal,
9.936Y = 265,
*»-
Y = 26.7 grams of terrestrial birds-mammals consumed, and
9 Y = 240 grams of fish consumed.
As indicated in Table 6, the fish consumed by herring gulls are assumed to consist of 80
percent trophic level 3 fish and 20 percent trophic level 4 fish. Thus, of the 240 grams of
fish consumed daily, 80 percent, or 192 grams, are of trophic level 3 fish and 20 percent,
or 48 grams, are of trophic level 4 fish.
VII.D.5
Bald Eagle (Ha/laeetus feucocepha/us)
Body weight. As for other birds of prey, female bald eagles are about 20 percent
heavier than the males. Given the data provided in Table 12, 4.6 kg would be a
reasonable value to use for the average weight of bald eagles.
Table 12. Body Weights of Bald Eagle Populations.
Age/Sex
adult/female
adult/male
average
juvenile/female (to 3 yrs old)
juvenile/male (to 3 yrs old)
average
Body Weight (kg)
5.2 (N=37)
4.1 (N = 35)
4.65
5.1 (range = 4.3
4.0 (range = 3.5
4.55
- 5.8)1
- 4.6)^
Reference/Location
Synder & Wiley, 1 976/
North America
Imler & Kalmbach, 1 9557
Alaska
- N = 18 for both sexes combined.
Water ingestion rate. No measured values for bald eagle drinking water ingestion
rates were identified. The moisture content of their diet may satisfy much of their daily
water requirements. In the absence of specific information, however, Calder and Braun's
(1983) allometric equation for estimating drinking water ingestion rates predicts that bald
eagles averaging 4.6 kg in weight drink 0.16 L/day:
Water Ingestion (L/day) = 0.059 Wt°-67 (kg) = 0.059 (4.6)°-67 = 0.16 L/day.
Page 44
GLWQI Technical Support Document for Wildlife Criteria
-------
Food ingestion rate. Several investigators have estimated the rate of food
consumption by bald eagles, but only some normalized their estimates to the body weight
of the birds or reported the body weights so that others could normalize the ingestion
rates. Two separate research teams provided similar estimates of food ingestion rates for
free-flying bald eagles.
Stalmaster and Gessaman (1984) observed bald eagles taking pre-weighed salmon
provided at artificial feeding stations in Washington State. Although the eagles may have
fed elsewhere on occasion, Stalmaster and Gessaman (1984) felt that the feeding stations
provided most of the birds' intake. Adults were observed to take 552 grams of fish per
bird per day, while juveniles and subadults (up to three years old) were observed to take
410 and 459 grams of fish, respectively, per bird per day. Assuming that adult eagles
weigh 4.5 kg on average, Stalmaster and Gessaman's (1984) observations indicate that
the adult bald eagles ingest 12.3 percent of their body weight in fish daily. Assuming that
the GE in fish consumed by bald eagles is 1.2 kcal/g wet weight and that the AE of birds
of prey consuming fish is 79 percent (U.S. EPA, 1993c, Tables 4-1 and 4-3), the ME of the
fish is estimated to be 0.948 kcal/g fish:
ME (kcal/g fish) = GE x AE = 1.2 (kcal/g) x 0.79 = 0.948 (kcal/g fish).
Using an ME of 0.948 kcal/g fish, the total ME requirement for an adult bald eagle on a
daily basis would be 523 kcal/bird per day (i.e., 552 g fish/bird-day x 0.948 kcal/g fish).
This energy "requirement" is based on the amount of fish the bald eagles were observed to
eat, which may have been more than they "needed" over the short term.
Craig et al. (1988) estimated gross energy consumption rates of bald eagles in
Connecticut based the observed time spent feeding and the model of Stalmaster and
Gessaman (1984). Assuming a body weight of 4.5 kg, they concluded that adult bald
eagles require 448 kcal ME per day and that for adults and that subadults require 499 kcal
ME per day.
Given the data presented above, it would be reasonable to assume that bald eagles
require 500 kcal ME per bird per day.
As indicated in Table 6, bald eagles also include birds and mammals in their diet.
Assuming that the birds consumed by bald eagles in the Lake Superior study provide 1.9
kcal GE/g wet weight (value for gulls and terns in Table 4-1 in U.S. EPA, 1993c) and that
the AE of birds of prey consuming birds is 78 percent (Table 4-3 in U.S. EPA, 1993c),the
ME in birds for the bald eagle would be 1.48 kcal/g birds:
ME (kcal/g birds) = GE x AE = 1.9 {kcal/g) x 0.78 = 1.482 (kcal/g birds).
If bald eagles consumed only fish, to obtain 500 kcal of ME daily, they would need
to ingest 527 grams of fish daily (i.e., 500 kcal/bird-day divided by 0.948 kcal/g fish). If
bald eagles consumed only birds, to obtain 500 kcal of ME daily, they would need to ingest
338 grams of birds daily (500 kcal/day divided by 1.48 kcal/g birds). Given the
assumption that bald eagles consume 92 percent fish and 8 percent birds on a wet-weight
basis (see Section VII.B), the total grams of each type of prey is estimated as follows:
GLWQI Technical Support Document for Wildlife Criteria Page 45
-------
If
Y = grams of birds consumed, and
11.5 Y = grams of fish consumed (i.e., 92/8 = 11.5),
then
Y (g) x 1.482 {kcal/g birds) + 11.5 Y (g) x 0.948 (kcal/g fish) = 500 kcal,
1.482Y(kcal) + 10.902 Y (kcal) = 500 kcal,
12.38Y = 500,
and
Y = 40.4 grams of birds consumed, and
11.5 Y = 464 grams of fish consumed.
Of the 40.4 grams of birds consumed, 70 percent, or 28.3 grams are comprised of herring
gulls, and the remaining 12.1 grams are comprised of other non-piscivorous birds. Of the
464 grams of fish consumed, 80 percent, or 371 grams, are of trophic level 3 fish, and 20
percent, or 92.8 grams, are of trophic level 4 fish.
For the sensitivity analyses carried out in the Great Lakes Water Quality Initiative
Criteria Documents for the Protection of Wildlife: DDT; Mercury; 2,3,7,8-TCDD;andPCBs
(U.S. EPA, 1995d), different values for the amount of gulls and fish in the bald eagles diet
were considered. Kozie (1986) provided data for one pair of bald eagles nesting near a
herring gull colony. For this pair, 13.75 percent of the diet consisted of birds and 91
percent of the wet mass of birds captured consisted of herring gulls (Exhibit 6-8 in U.S.
EPA 1995a). Thus, 12.5 percent of the diet of this pair consisted of herring gulls on a wet
weight basis. In contrast, considering all eight pairs studied on lake Superior, only 5.8
percent of the diet consisted of herring gulls on average (i.e., 8 percent of diet consists of
birds x 70 percent of birds are comprised of herring gulls). To conduct the sensitivity
analysis for the bald eagle related to its dietary composition, it was assumed that at one
extreme the composition of the bald eagle's diet was the same as for the pair near the gull
colony. The quantities of fish, gulls, and non-aquatic birds that would be needed to
provide 500 kcal of ME are calculated as follows.
Still assuming that the ME in birds for the bald eagle is 1.48 kcal/g bird, and that
the ME in fish for the bald eagles is 0.948 kcal.g fish, then assuming that bald eagles
consume 86.3 percent fish and 13.8 percent birds on a wet-weight basis (see above), the
total grams of each type of prey is estimated as follows:
If
Y = grams of birds consumed, and
6.27 = grams of fish consumed (i.e., 86.3/13.8 = 6.27)
then
Y (g) x 1.482 (kcal.g bird) + 6.27 Y (g) x 0.948 (kcal.g fish)
= 500 kcal.
1.482 Y (kcal) + 5.944 (kcal) = 500 kcal,
7.425 Y = 500 kcal,
and
Y = 67.33 grams of birds consumed, and
6.27 Y = 422.2 grams of fish consumed.
Page 46 GLWQI Technical Support Document for Wildlife Criteria
-------
Of the 67.3 grams of birds consumed, 91 percent, or 61.3 grams are comprised of herring
gulls, and the remaining 6.0 grams are comprised of other non-piscivorous birds. Of the
422 grams of fish consumed, 80 percent, or 338 grams, are of trophic level 3 fish, and 20
percent, or 84.5 grams, are of trophic level 4 fish.
Alternatively, for the sensitivity analyses, calculations assuming a 100 percent fish
diet (with 80% trophic level 3 and 20% trophic level 4) were also performed. In that both
of these assumptions (i.e., gull ingestion and no gull ingestion) assume no terrestrial
component to the eagle diet, they are likely to provide slight overestimates of contaminant
exposures (see Table 13 below).
VII.D.6 Summary of Exposure Parameter Values for the Representative Wildlife
Species
Table 13 summarizes the exposure parameter values for the five representative
wildlife species. This table also is presented as Table D-2 in Appendix D to Part 132.
Table 13. Exposure Parameter Values for the Five Representative Wildlife Species.
Species
Mink
Otter
Kingfisher
Herring gull
Bald Eagle
Adult
Body
Weight
(kg)
0.80
7.4
0.15
1.1
4.6
Water
Ingestion Rate
(L/day)
0.081
0.60
0.017
0.063
0.16
Food Ingestion
Rate of Prey in
Each Trophic Level
(kg/day)
TL3: 0.159
Other: 0.0177
TL3: 0.976
TL4: 0.244
TL3: 0.0672
TL3: 0.192
TL4: 0.0480
Other: 0.0267
TL3: 0.371
TL4: 0.0928
PB: 0.0283
Other: 0.0121
Trophic Level of
Prey
as
Percent of Diet
TL3: 90%
Other: 10%
TL3: 80%
TL4: 20%
TL3: 100%
Fish: 90%
TL3: 80%
TL4: 20%
Other: 10%
Fish: 92%
TL3: 80%
TL4: 20%
Birds: 8%
PB: 70%
non-aquatic:
30%
Note: TL3 = trophic level three fish; TL4
= non-aquatic-based birds and mammals
trophic level four fish; PB = piscivorous birds; Other
GLWQI Technical Support Document for Wildlife Criteria
Page 47
-------
Vlll. REFERENCES
Alexander, G.R. 1977. Food of vertebrate predators on trout waters in north central lower
Michigan. Michigan Academician 10: 181-195.
Altman, P.L. and D.S. Dittmer, Editors. 1972. Biology Data Book, Second Edition,
Volumes I - III. Federation of American Societies for Experimental Biology, Bethesda, MD;
pp. 195-215,1450-1457.
Aulerich, R.J., R.K. Ringer and S. Iwamoto. 1973. Reproductive failure and mortality in
mink fed on great lakes fish. J. Reprod. Pert. Suppl. 19: 365-376.
Bartholomew, G.A. and T.J. Cade. 1963. The water economy of land birds. Auk 80:
504-539.
Belopol'skii, L.O. 1957. (Ecology of sea colony birds of the Barents Sea.) Translated by:
Israel Program for Scientific Translations, Jerusalem.
Bleavins, M.R. and R.J. Aulerich. 1981. Feed consumption and food passage in mink
(Mustafa vison) and European ferrets (Mustela putorius furo). Lab. Animal Sci. 31:
268-269.
Brooks, R.P., and W.J. Davis. 1987. Habitat selection by breeding belted kingfishers
(Ceryle alcyon). Am. Midi. Nat. 117: 63-70.
Calder III, W. A. and E. J. Braun. 1983. Scaling of Osmotic Regulation in Mammals and
Birds. American Journal of Physiology. 244:601-606.
Chapman, J.A. and G.A. Feldhammer, eds.. Wild Mammals of North America. Baltimore,
MD: Johns Hopkins University Press; pp. 329-643.
Colborn, T.1. 1991. Epidemiology of Great Lakes bald eagles. J. Environ. Health. Toxicol.
4: 395-453.
Courtney, P.A. and H. Blokpoel. 1980. Food and indicators of food availability for
common terns on the lower Great Lakes. Can. J. Zool. 58: 1318-1323.
Craig, R.J., E.S. Mitchell, and J.E. Mitchell. 1988. Time and energy budgets of bald
eagles wintering along the Connecticut River. J. Field Ornithol. 59: 22-32.
Davison. K.L., K.A. Engebretson, and J.H. Cox. 1976. p,p'-DDT and p,p'-DDE effects on
egg production, eggshell thickness, and reproduction of Japanese quail. Bull. Environ.
Contam. Toxicol. 15: 265-270.
Delnicki, D. and K.J. Reinecke. 1986. Mid-winter food use and body weights of mallards
and wood ducks in Mississippi. J. Wild). Manage. 50:43-51.
Page 48 GLWQI Technical Support Document for Wildlife Criteria
-------
Dourson, M.L. and J.F. Stara. 1983. Regulatory history and experimental support of
uncertainty (safety) factors. Regulatory Toxicology and Pharmacology. 3:224-238.
Dunning, J.B. 1984. Body Weights of 686 North American Birds. Monograph #1.
Western Bird Banding Association.
Environment Canada. 1991. Toxic Chemicals in the Great Lakes and Associated Effects.
Vol. II: Effects. Canada: Department of Fisheries and Oceans, Health and Welfare Canada.
Ewins, P.J., D.V. Weseloh, J.H. Groom, et al. 1991. The diet of herring gulls (Larus
argentatus) during winter and early spring on the lower Great Lakes, (unpublished
manuscript)
Fox, G.A. L.J. Allan, D.V. Weseloh,-st al. 1990. The diet of herring gulls during the
nesting period in Canadian waters of the Great Lakes. Can. J. Zool. 68: 1075-1085.
Gilbertson, ML, T.J. Kubiak, J.P. Ludwig, et al. 1991. Great Lakes embryo mortality,
edema, and deformities syndrome (GLEMEDS) in colonial fish-eating birds: similarity to
chick edema disease. J. Toxicol. Environ. Health 33: 455-520.
Harding, A.R. 1934. Mink trapping. A.R. Harding, Columbus, OH; 171 pp.
Hill, E.F., R.G. Heath, J. W. Spann, and J.D. Williams. 1975. Lethal dietary toxicities of
environmental pollutants to birds. U.S. Fish and Wildlife Service, Special Scientific Report
No. 191. Washingotn, D.C.
Hobson, K.A. 1990. Stable isotope analysis of marbled murrelets: evidence for freshwater
feeding and determination of trophic level. Condor 92: 897-903.
Hornshaw, T.C., R.J. Aulerich, and H.E. Johnson. 1983. Feeding great lakes fish to mink:
Effects on mink and accumulation and elimination of PCBs by mink. J. Toxicol. Environ.
Health. 11: 933-946.
Hudson, R.H., R.K. Tucker, and M.A. Haegele. 1984. Handbook of toxicity to wildlife,
second edition. U.S. Fish and Wildlife Service, Resource Publication No. 153.
Washington, D.C.
Imler, R.H., and E.R. Kalmbach. 1955. The bald eagle and its economic status. US Fish
Wildl. Ser. Circ. 30.
Knudsen, G.J. and J.B. Hale. 1968. Food habits of otters in the Great Lakes region. J.
Wildl. Manage. 32: 89-93.
Kozie, K.D. and R.K. Anderson 1991. Productivity, diet, and environmental contaminants
in bald eagles nesting near the Wisconsin shoreline of Lake Superior. Arch. Environ.
Contam. Toxicol. 20: 41-48.
GLWQI Technical Support Document for Wildlife Criteria Page 49
-------
Lagler, K.F. and B.T. Ostenson. 1942. Early spring food of the otter in Michigan. J.
Wildl. Manage. 6: 244-254.
Lauhachinda, V. 1978. Life history of the river otter in Alabama with emphasis on food
habits. Ph.D. dissertation. Auburn, AL: University of Alabama.
LeCroy. M., and S. LeCroy. 1974. Growth and fledging in the common tern (Sterna
hirundo). Bird-Banding 45: 326-340.
Linscombe, G., N. Kinler, and R.J. Aulerich 1982. Mink. In: Chapman, J. A.;
Feldhammer, G. A., eds., Wild Mammals of North America. Baltimore, MD: Johns Hopkins
University Press; pp. 329-643.
t
McNamara, B.P., 1976. Concepts in health evaluation of commercial and industrial
chemicals, in Mehlman, M.A. et al. (eds), Advances in Modern Toxicology, Volume 1,
Part 1: New Concepts in Safety Evaluation. John Wiley & Sons, New York.
Mearns, A.J. D.R. Young, R.J. Olson, et al. 1981. Trophic structure of the
cesium-potassium ratio in pelagic ecosystems. Calif. Coop. Ocean. Fish. Invest. Rep. 22:
99-110.
Medway, W. and Kare, M.R. 1959. Water metabolism of the growing domestic fowl with
special reference to water balance. Poultry Sci. 38: 631-637.
Melquist, W.E., and M.G. Hornocker. 1983. Ecology of river otters in west central Idaho.
In: Kirkpatrick, R. L, ed.. Wildlife Monographs; Vol. 83, Bethesda, MD: The Wildlife
Society; pp. 60.
Nagy, K.A. 1987. Field metabolic rate and food requirement scaling in mammals and
birds. Ecol. Monogr. 57: 111-128.
Nagy, K.A. and C.C. Peterson. 1988. Scaling of water flux rate in animals. Berkeley, CA:
University of California Press.
National Academy of Sciences. 1980. Drinking Water and Health. Vol.2. National
Academy Press. Washington, D.C.
National Academy of Sciences. 1977. Drinking Water and Health. Vol.1. National
Academy Press. Washington, D.C.
National Geographic Society (NGS). 1987. Field Guide to the Birds of North America.
Second edition. National Geographic Society, Washington, DC.
Nelson, N.L. and A.C. Martin. 1953. Gamebird weights. J. Wildl. Manage. 17: 36-42.
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, Division of
Environmental Contaminants. Tech. Rep. 87-3.
Page 50 GLWQI Technical Support Document for Wildlife Criteria
-------
National Institute of Occupational Safety and Health (NIOSH). 1993. Registry of Toxic
Effects of Chemical Substances (RTECS), A Comprehensive Guide. U.S. Department of
Health and Human Services, Public Health Service, Centers for Disease Control and
Prevention, NIOSH Publ. No. 93-120. (data extracted from RTECS database in November,
1994).
Norstrom, R.J., T.P. Clark, J.P. Kearney, et al. 1986. Herring gull energy requirements
and body constituents in the Great Lakes. Ardea 74: 1-23.
Peakall, D. B. 1988. Known effects of pollutants on fish-eating birds in the Great Lakes of
North America. In: Schmidtke, N. W., ed. Toxic Contamination in Large Lakes. Vol. I:
Chronic Effects of Toxic Contaminants in Large Lakes, Chelsea, Ml: Lewis Publishers, Inc.;
pp. 39-54.
Sanger. G.A. 1987. Trophic levels and trophic relationships of seabirds in the Gulf of
Alaska. In: Croxall, J.P., ed., Seabirds: Feeding Ecology and Role in Marine Ecosystems.
New Rochelle, NY: Cambridge University Press; pp. 229-258.
Scott. M.L., M.C. Nesheim, and R.J. Young. 1976. Nutrition of the Chicken. Second
Edition. Department of Poultry Science and Division of Nutritional Sciences, Cornell
University. M.L. Scott and Associates, Ithaca, NY.
Sealander, J.A. 1943. Winter food habits of mink in southern Michigan. J. Wildl.
Manage. 7: 411-417.
Schafer, E.W., Jr., and R.B. Brunton. 1979. Indicator bird species for toxicity
determinations: is the technique usable in test method development?. In: Vertebrate Pest
Control and Management Materials, ASTM STP 680, J.R. Beck, Ed., American Society for
Testing and Materials, pp. 157-168.
Sheffy, T.B. and J.R. St. Amant. 1982. Mercury burdens in furbearers in Wisconsin. J.
Wildl. Manage. 46: 1117-1120.
Shirazi, M.A., R.A. Bennett and R.K. Ringer. 1994. An interpretation of toxicity response
of Bobwhite Quail with respect to duration of exposure. Archives of Environmental
Contamination and Toxicology 26:417-424.
Snyder, N. F., and J.W. Wiley. 1976. Sexual size dimorphism in hawks and owls of North
America. Orni. Monogr. 20.
Stalmaster, M.V., and J.A. Gessaman. 1984. Ecological energetics and foraging behavior
of overwintering bald eagles. Ecol. Monogr. 54: 407-428.
Threlfall, W., and D.D. Jewer. 1978. Notes on the standard body measurements of two
populations of herring gulls (Larus argentatus). Auk 95: 749-753.
GLWQI Technical Support Document for Wildlife Criteria Page 51
-------
Toweill, D.E., and J.E. Tabor. 1982. River otter. In: Chapman, J. A.; Feldhammer, G.
A., eds.. Wild Mammals of North America. Baltimore, MD: Johns Hopkins University
Press; pp. 688-703.
Travis. C.C., R.K. White and B.C. Ward. 1990. Interspecies extrapolation of
pharmocokinetics. J. Theor. Biol. 142:285-304.
•y U.S. Environmental Protection Agency. 1995a. Trophic Level and Exposure Analyses for
Selected Piscivorous Birds and Mammals. Volume I: Analyses of Species for the Great
Lakes. Draft. Office of Water, Washington. DC. Available in the GLWQI docket.
^U.S. Environmental Protection Agency. 1995b. Trophic Level and Exposure Analyses for
Selected Piscivorous Birds and Mammals. Volume III: Appendices. Office of Water,
Washington, DC. Available in the GLWQI docket.
v U.S. Environmental Protection Agency. 1995c. Technical Basis for Recommended Ragnes
of Uncertainty Factors used in Deriving Wildlife Criteria for the Great Lakes Water Quality
Initiative. Draft. Washington, DC. Available in the GLWQI docket.
U.S. Environmental Protection Agency. 1995d. Great Lakes Water Quality Initiative
Criteria Documents for the Protection of Wildlife; DDT; Mercury; 2,3,7,8-TCDD, PCBs.
Washington, DC. EPA-820-B-95-008.
U.S. Environmental Protection Agency. 1994. Advisory on the Development of a National
Wildlife Criteria Program. EPA-SAB-EPEC-ADV-94-001.
U.S. Environmental Protection Agency. 1993a. Great Water Quality Initiative Criteria
Documents for the Protection of Wildlife Proposed: DDT, Mercury, 2.3.7.8-TCDD, PCBs.
Washington, DC: Office of Water, Office of Science and Technology; EPA/822/R-93/007.
U.S. Environmental Protection Agency. 1993b. Wildlife Criteria Portions of the Proposed
Water Quality Guidance for the Great Lakes System. Washington, DC: Office of Water,
Office of Science and Technology; EPA/822/R-93/006.
U.S. Environmental Protection Agency. 1993c. Wildlife Exposure Factors Handbook.
Volume I. Washington, DC: Office of Research and Development. EPA/600/R-93/187a.
U.S. Environmental Protection Agency. 1993d. Wildlife Exposure Factors Handbook.
Volume II. Appendix: Literature Review Database. Washington, DC: Office pf Research
and Development. EPA/600/R-93/187b.
U.S. Environmental Protection Agency. 1992a. Draft Report: A cross-species scaling
factor for cacinogenic risk assessment based on equivalence of mg/kg3/4/day: Notice.
Federal Register. Vol. 57, No. 109, pp. 24152-24173. Friday, June 5.
U.S. Environmental Protection Agency. 1992b. Framework for Ecological Risk
Assessment. Washington, D.C. EPA/630/R-92/001.
Page 52 GLWQI Technical Support Document for Wildlife Criteria
-------
U.S. Environmental Protection Agency. 1992c. An SAB Report: Evaluation of the
Guidance for the Grat Lakes Water Quality Initiative. Washington, DC. EPA-SAB-
EPEC/DWC-93-005
U.S. Environmental Protection Agency. 1988. Recommendations for and documentation
of biological values for use in risk assessment. NTIS-PB88-179874.
U.S. Environmental Protection Agency. 1986. Hazard Evaluation Division: Standard
Evaluation Procedure - Avian Reproduction Test. EPA 540/9-86-139.
U.S. Environmental Protection Agency. 1985. Section V.C. Evaluation of health effects
and determination of RMCLs pp. 46944-46950m National Primary Drinking Water
Regulations; Synthetic Organic Chemicals; Inorganic Chemicals and Microorganisms. 50
Federal Register 46936-47022. Wednesday, November 13, 1985.
U.S. Environmental Protection Agency. 1980. Appendix C. Guidelines and methodology
used in the preparation of health effect assessment chapters of the consent decree water
criteria documents pp. 79347-79357in Water Quality Criteria Documents; Availability.
45 Federal Register 79318-79378. Friday, November 28, 1980.
Weil, C.S. 1972. Statistics versus safety factors and scientific judgement in the
evaluation of safety for man. Toxicology and Applied Pharmacology. 21:454-463.
Weil, C.S. and D.D. McCollister. 1963. Relationship between short- and long-term
feeding studies in designing an effective toxicity test. Agricultural and Food Chemistry.
11(6):486-491.
Wilson, K.A. 1959. The otter in North Carolina. Proc. Southeast. Assoc. Fish and Game
Comm. 13: 267-277.
Wren, C., H. MacCrimmon, R. Frank, and P. Suda. 1980. Total and methyl mercury levels
in wildlife mammals from the precambrian shield of south central Ontairo, Canada. Bull.
Environ. Contam. Toxicol. 25: 100-105.
GLWQI Technical Support Document for Wildlife Criteria Page 53
------- |