-------
August 21. 1996 DRAFT
Has contamination migrated from source areas and resulted in 'off-site"
impacts or the threat of impacts, in addition to on-site threats or impacts?
These questions should be answered using the site reports, maps (e.g, U.S. Geological
Survey, National Wetlands Inventory), available aerial photographs, communication with
appropriate agencies (e.g., U.S. Fish and Wildlife Service, National Oceanic and Atmospheric
Administration, State Natural Heritage Programs), and a site visit. Activities that should be
conducted during this site visit include:
Note the layout and topography of the site.
Note and describe any water bodies and wetlands.
* Identify and map evidence indicating contamination or potential contamination
(e.g., areas of no vegetation, runoff gullies to surface waters).
Describe existing aquatic, terrestrial, and wetland ecological habitat types (e.g.,
forest, old field), and estimate the area covered by these habitats.
Note any potentially sensitive environments (see Section 1.2.3 for examples of
sensitive environments).
Describe and, if possible, map soil and water types, land uses, and the
dominant vegetation species present.
Record any observations of animal species or sign of a species.
Mapping can be useful in establishing a "picture" of the site to assist in problem
formulation. The completed checklist (U.S. EPA, 1996c) will provide information regarding
habitats and species potentially or a»rtually present on site, potential contaminant migration
pathways, exposure pathways, and the potential for non-chemical stresses at the site.
After finishing the checklist, it might be possible to determine that present or future
ecological impacts are negligible because complete exposure pathways do not exist and could
not exist in the future. Many Superfund sites are located in highly industrialized areas where
there could be few if any ecological receptors or where site-related impacts might be
indistinguishable from non-site-related impacts (see Highlight Box 1-2). For such sites,
remediation to reduce ecological risks might not be needed. However, all sites should be
evaluated by qualified personnel to determine whether this conclusion is appropriate.
STEP 1, Page 3
-------
A'ue-ust 21. 1996
DRAFT
HIGHLIGHT BOX 1-2
Industrial or Urban Settings
Many hazardous waste sites exist
m currently or historically industrialized
or urbanized areas. In these instances, it
can be difficult to distinguish between
impacts related to contaminants from a
particular site and impacts related to
non-contaminant stressors or to
contaminants from other sites. However,
even in these cases, it could be
appropriate to take some remedial
actions based on ecological nsks. These
actions might be limited to source
removal or might be more extensive.
An ecological risk assessment can assist
the nsk manager in determining what
action, if any. is appropriate.
Other Superfund sites are located in less
disturbed areas with protected or sensitive
environments that could be at nsk of adverse
effects from contaminants from the site. State
and federal laws (e.g., the Clean. Water Act, the
Endangered Species Act) designate certain types
of environments as requiring protection. Other
types of habitats unique to certain areas also
could need special consideration in the nsk
assessment (see Section 1.2.3).
1.2.2 Contaminant Fate and Transport
During problem formulation, pathways
for migration of a contaminant (e.g., windblown
dust, surface water runoff, erosion) should be
identified. These pathways can exhibit a
decreasing gradient of contamination with
increasing distance from a site. There are
exceptions, however, because physical and
chemical charactenstics of the media also
influence contaminant distribution (e.g., the pattern of contaminated sediment deposition in
streams vanes depending on stream flow and bottom characteristics). For the screening-level
risk assessment, the highest contaminant concentrations measured on the site should be
documented for each medium.
1.2.3 Ecotoxicity and Potential Receptors
Understanding the toxac mechanism of a contaminant helps to evaluate the importance
of potential exposure pathways (see Section 1.2.4) and to focus the selection of assessment
endpoints (see Section 1.2,5). Some contaminants, for example, affect pnmarily vertebrate
animals by interfering with organ systems not found in invertebrates or plants (e.g., distal
tubules of vertebrate kidneys, vertebrate hormone systems). Other substances might affect
primarily certain insect groups (e.g., by interfenng with hormones needed for metamorphosis),
plants (e.g., some herbicides), or other groups of organisms. For substances that affect, for
example, reproduction of mammals at much lower environmental' exposure levels than they
affect oiher groups of organisms, the screening-level risk assessment can initially focus on
exposure pathways and nsks to mammals. Example Box 1-1 illustrates this point using the
PCB site example provided in Appendix A. A review of some of the more recent ecological
risk and toxiciry assessment literature can help identify likely effects of the more common
contaminants at Superfund sites.
STEP 1. Page 4
-------
August 21. 1996 DRAFT
EXAMPLE BOX 1-1
Ecotoxicity-PCB Site
Some PCBs are reproductive toxins in mammals. When ingested, they induce (i.e.,
increase concentrations and activity of) enzymes in the liver (Melancon and Lech, 1983). The
enzymes are not specific for PCBs and will enhance the degradation of steroid hormones
(Peakall, 1975). The observed impairment of reproduction in mammals exposed to PCBs might
be caused by PCB-induced reduction in circulating steroid hormones (Tanabe, 1988). Other
effects, such as liver pathology, are also evident at high exposure levels (Fuller and Hobson,
1986). Given this information, the screening ecological risk assessment should include potential
exposure pathways for mammals to PCBs (see Example Box 1-2).
An experienced biologist or ecologist can determine what plants, animals, and habitats
exist or can be expected to exist in the area of the Superfund site. Exhibit 1-1, adapted from
the Superfund Hazard Ranking System, is a partial list of types of sensitive environments that
could require protection or special consideration. Information obtained for the environmental
checklist (Section 1.2.1), existing information and maps, and aerial photographs should be
used to identify the presence of sensitive environments on or near a site that might be
threatened by contaminants from the site.
1.2.4 Complete Exposure Pathways
Evaluating potential exposure pathways is one of the primary tasks of the screening-
level ecological characterization of the site. For an exposure pathway to be complete, a
contaminant must be able to travel from the source to ecological receptors and to be taken up
by the receptors via one or more exposure routes. (Highlight Box I-3 defines exposure
pathway and exposure route.) Identifying complete exposure pathways prior to a quantitative
evaluation of toxicity allows the assessment to focus on only those contaminants that can
reach ecological receptors.
Different exposure routes are important for different groups of organisms. For
terrestrial animals, three basic exposure routes need to be evaluated: inhalation, ingestion,
ancl dermal absorption. For terrestrial plants, root absorption of contaminants in soils or leaf
absorption of contaminants evaporating from the soil are of concern at Superfund sites. For
aquatic animals, direct contact (of water or sediment with the gills or dermis) and ingestion of
food (and sometimes sediments) should be considered. For aquauc plants, direct contact with
water, and sometimes with air or sediments, is of primary concern.
STEP I, Page 5
-------
Aumist 2!. 1996 DRAFT
EXHIBIT 1-1
List of Sensitive Environments in the Hazard Ranking System*
CnucaJ habitat for FederaJ designated endangered or threatened species
Marine Sanctuary
NationaJ Park
Designated FederaJ Wilderness Area
.Areas identified under the Coastal Zone Management Act
Sensiuve areas identified under the National Estuary Program or Near CoasiaJ Waters Program
CnucaJ areas identified under the Clean Lakes Program
NationaJ Monument
National Seashore Recreational Area
NationaJ Lakeshore Recreational Area
Habitat known to be used by FederaJ designated or proposed endangered or threatened species
NationaJ Preserve
NauonaJ or State Wildlife Refuge
Unit of CoastaJ Barrier Resources System
CoastaJ Barrier (undeveloped)
FederaJ land designated for protection of natural ecosystems
Administratively Proposed FederaJ Wilderness Area
Spawning areas critical for the maintenance of fish/shellfish species within nver, lake, or
coastaJ tidaJ waters
Migratory pathways and feeding areas cnucaJ for maintenance of anadromous fish species within nver
reaches or areas in lakes or coastaJ tidaJ waters in which the fish spend extended periods of time
Terrestrial areas utilized for breeding by large or dense aggregations of animals
National nver reach "designated as RecreationaJ
Habitat known to be used by state designated endangered or threatened species
Habitat known to be used by species under review as to its Federal endangered or threatened status
CoastaJ Bamer (partially developed)
FederaJ designated Scenic or Wild River
State land designated for wildlife or game management
State-designated Scenic or Wild River
State-designated NamnJ Areas
Particular areas, relatively small in size, important to maintenance of unique bioQc communities
State-designated areas for protection or maintenance of aquanc life
Wedands6
* The categories UT listed in groups from those assigned higher factor values to those assigned
lower factor values in the Hazard Ranking Svsiem iHRS) for listing hazardous waste sues on Lhe National
Pnonues List (U.S. EPA, 1990b). See Federal Register. Vol. 55, pp. 51624 and 51648 for additional
information regirding definition!.
Under the HRS, wetlands ire rated separately from the sensitive environments on the basis of
size. See Federal Register. Vol. 55, pp. 51615 and 51662 for additional information.
STEP 1, Page 6
-------
August 21. 1996
DRAFT
The most likely exposure pathways and exposure routes also are related to the physical
and chemical properties of the contaminant (e.g., whether or not the contaminant is bound to
a matnx. such as organic carbon). Of the basic exposure routes identified above, more
information generally is available to quantify exposure levels for ingesuon by terrestrial
animals, and for direct contact with water or sediments by aquatic organisms, than for other
exposure routes and receptors. Although other exposure routes can be important, more
assumptions are needed to estimate exposure levels for those routes, and the results are less
certain. Professional judgment is needed to determine if evaluating those routes sufficiently
improves a risk assessment to warrant the effort
If an exposure pathway is not
complete for a specific contaminant (i.e.,
ecological receptors cannot be exposed to
the contaminant), that exposure pathway
does not need to be evaluated further. For
example, suppose a contaminant that impairs
reproduction in mammals occurs well below
the root zone of plants that occur or are
expected to occur on a site. Herbivorous
mammals would not be exposed to the
contaminant through their diets because
plants would not be contaminated.
Assuming that most soil macroinvertebrates
available for ingestion live in the root zone,
insectivorous mammals also would be
unlikely to be exposed. Burrowing
mammals would not be expected to come
into direct contact with the contaminated
soils. In this case, a complete exposure
route for this contaminant for surface-dwelling mammals would not exist, and the
contaminant would not pose a significant nsk to this group of organisms. Secondary
questions might include whether the contaminant is leaching from the soil to ground water
that discharges to surface water, thereby posing a risk to the aquatic environment or to
terrestrial mammals that drink the water or consume aquatic prey. Example Box 1-2
illustrates the process of identifying complete exposure pathways based on the PCB site
described in Appendix A.
1.2.5 Assessment and Measurement Endpoints
For the screening-level ecological risk assessment, assessment endpoints are any
adverse effects on ecological receptors, where receptors are plant and animal populations and
HIGHUGHT BOX 1-3
Exposure Pathway and
Exposure Route
Exposure Pathway: The pathway by
which a contaminant travels from a source
(e.g., drums, contaminated soils) to
receptors. A pathway can involve multiple
media (e.g., soil runoff to surface waters and
sedimentation, or volatilization to the
atmosphere).
Exposure Route: A point of contact/entry
of a contaminant from the environment into
an organism (e.g., inhalation, ingestion,
dermal absorption).
STEP 1, Page?
-------
August 2! 1996 DRAFT
EXAMPLE BOX 1-2
Complete Exposure Pathways for Mammals-PCB Site
Three possible exposure pathways for mammals were evaluated at the PCB Site:
inhalation, mgestion through the food chain, and incidental soil/sediment ingestion.
Inhalation. PCBs are not highly volatile, so the inhalation of PCBs by mammals
would be an essentially incomplete exposure pathway.
Ingestion through the food chain. PCBs tend to bioaccumulate and biomagnify in
food chains. PCBs in soils generally are not taken up by most plants, but are accumulated by
soil macromvertebrates. Thus, mammalian herbivores would not be exposed to PCBs in most of
their diet. In contrast, mammalian insectivores, such as shrews, could be exposed to PCBs in
most of their diet. For PCBs, the ingestion route for mammals would be essentially incomplete
for herbivores but complete for insectivores. For the PCB site, therefore, the ingestion exposure
route for a mammalian insectivore (e.g., shrew) would be a complete exposure pathway that
should be evaluated.
Incidental soil/sediment ingestion. Mammals can ingest some quantity of soils or
sediments incidentally, as they groom their fur or consume plants or animals from the soil.
Burrowing mammals are likely to ingest greater quantities of soils during grooming than non-
burrowing mammals, and mammals that consume plant roots or soil-dwelling macroinvertebrates
are likely to ingest greater quantities of soils than mammals that consume other foods. The
intake of PCBs from incidental ingestion of PCB-contaminated soils is difficult to estimate, but
for mammalian insectivores, it is likely to be far less than the intake of PCBs in the diet. For
herbivores, the incidental intake of PCBs in soils might be higher than the intake of PCBs in
their diet, but sull less than the intake of PCBs by mammals feeding on soil macroinvertebrates.
Thus, the exposure pathway for mammalian msecdvores remains the exposure pathway that
should be evaluated.
communities, habitats, and sensitive environments. Adverse effects on populations can be
inferred from measures related to unpaired reproduction, growth, and survival. Adverse
effects on communiues can be inferred from changes in commuiury structure or function.
Adverse effects on habitats can be inferred from changes in composition and characteristics
that reduce the habitats' ability to support the plant and animal populations and comrnumues.
Many of the screening ecotoxiciry values now available or likely to be available in the
future for the Superfund program (see Section 1.3) are based on generic assessment endpoints
(e.g., protection of aquatic communities from changes in structure or function) and are
assumed to be widely applicable to sites around the United States.
STEP 1, Page 8
-------
August 21. 1996
DRAFT
1.3 SCREENING-LEVEL ECOLOGICAL EFFECTS EVALUATION
The next step in the screening-level nsk assessment is the preliminary ecological
effects evaluation and the establishment of contaminant exposure levels that represent
conservative thresholds for adverse ecological effects. In this guidance, those conservative
thresholds are called screening ecotoxicity values. Physical stresses unrelated to contaminants
at the site are not the focus of the risk assessment (see Highlight Box 1-4) (although they can
be considered later when evaluating effects of remedial alternatives).
A literature search for studies that
quantify toxicity (i.e., exposure-response) is
necessary to evaluate the likelihood of toxic
effects in different groups of organisms.
Appendix C provides a basic introduction to
conducting a literature search, but an expert
should be consulted to minimize time and
costs. The toxicity profile should describe
the toxic mechanisms of action for the
exposure routes being evaluated and the
dose or environmental concentration that
causes a specified adverse effect.
For each complete exposure
pathway/route and contaminant, an
screening ecotoxicity value should be
developed. The U.S. EPA Office of
Emergency and Remedial Response has
developed screening ecotoxicity values
(called ecotox thresholds) specifically for
this nsk assessment guidance (U.S. EPA,
1996a). These values are for surface waters
and sediments and are based on direct
exposure routes only. Bioaccumulation and
biomagnification in food chains have not been accounted for. When screening ecotoxicity
values are not available, they should be developed from other sources.1 The following
HIGHLIGHT BOX 1-4
Non-Chemical Stressors
Ecosystems can be stressed by
physical, as well as by chemical, alterations
of their environment. For this reason,
EPA's (1992a) Framework for Ecological
Risk Assessment addresses "stressor-
response" evaluation to include all types of
stress instead of "dose-response" or
"exposure-response" evaluation, which
implies that the stressor must be a toxic
substance.
For Superfund sites, however,
CERCLA addresses risks from hazardous
substances released to the environment, not
risks from physical alterations of the
environment. This guidance document,
therefore, focuses on exposure-response
evaluations for toxic substances.
1 It is possible to conduct a screening risk assessment with limited information and conservative assumptions.
If site-specific information is too limited however, the nsk assessment is almost certain to move into Steps 3 through
7, which require field-collected data. The more complete the initial information, the better the decision that can be
made at this preliminary stage.
STEP 1, Page 9
-------
August
1996
DRAFT
subsections describe preferred data sources (Section 1.3.1), dose conversions (Section 1.3.2),
and analyzing uncertainty in the values (Section 1.3.3).
1.3.1 Preferred Toxicity Data
Screening ecotoxicity values should
represent a no-observed-adverse-effect-level
(NOAEL) for long-term (chronic) exposures
to a contaminant. Ecological effects of most
concern are those that can impact
populations (or higher levels of biological
organization). These include adverse effects
on development, reproduction, and
survivorship. Community-level effects also
can be of concern, but toxicity data on
community-level endpoints are limited and
might be difficult to extrapolate from one
community to another.
When reviewing the literature, one
should be aware of the limitations of
published information in characterizing
actual or probable hazards at a specific site.
U.S. EPA discourages reliance on secondary
references because study details relevant for
determining the applicability of findings to a
given site usually are not reported in
secondary sources. Only primary literature
that has been carefully reviewed by an
ecotoxicologist should be used to support a
decision. Several considerations and data
preferences are summarized in Highlight Box
I-5 and described more fully below.
HIGHLIGHT BOX 1-5
Data Hierarchy for Deriving
Screening Ecotoxicity Values
To develop a chronic NOAEL for a
screening ecotoxicity value from existing
literature, the following data hierarchy
minimizes extrapolations and uncertainties
in the value:
A NOAEL is preferred to a
LOAEL, which is preferred to an
LCjQ or an EC^.
Long-term (chronic) studies are
preferred to medium-term
(subchromc) studies, which are
preferred to short-term (acute)
studies.
If exposure at the site is by
ingesuon, dietary studies are
preferred to gavage studies, which
are preferred to non-ingestion routes
of exposure. Similarly, if exposure
at the site is dermal, dermal studies
are preferred to studies using other
exposure routes.
NOAELS and LOAELS For each contaminant for which a complete exposure
pathway/route exists, the literature should be reviewed for the lowest exposure level (e.g.,
concentration in water or in the diet, ingested dose) shown to produce adverse effects (e.g.,
reduced growth, impaired reproduction, increased mortality) in a potential receptor species.
This vaiue is called a lowest-observed-adverse-effect-level or LOAEL. For those
contaminants with documented adverse effects, one also should identity the highest exposure
level that is a NOAEL. A NOAEL is more appropriate than a LOAEL to use as an screening
ecotoxicity value to ensure that nsk is not underestimated (see Highlight Box 1-6). However,
STEP 1, Page 10
-------
August 21, 1996
DRAFT
NOAELs currently are not available for
many groups of organisms or many
chemicals. When a LOAEL value, but not a
NOAEL value, is available from the
literature, a standard practice is to multiply
the LOAEL by 0.1 and to use the product as
the screening ecotoxicity value. Support for
this practice comes from a data review
indicating that 96 percent of chemicals
included in the review had LOAEL/NOAEL
ratios of five or less, and that all were ten
or less (Dourson and Stara, 1983).
Exposure duration. Data from
studies of chronic exposure are preferable to
data from medium-term (subchronic), short-
term (acute), or single-exposure studies
because exposures at Superfund remedial
sites usually are long-term. Literature
reviews by McNamara (1971) and Weil and
McCollister (1963) indicate that chronic
NOAELs can be lower than subchronic (90-
day duration for rats) NOAELs by up to a factor
HIGHLIGHT BOX 1-6
NOAEL Preferred to LOAEL
Because the NOAEL and LOAEL
are identified by the hypothesis testing (i.e.,
by comparing the response level of a test
group to the response level of a control
group for a statistically significant
difference), the actual proportion of the test
animals showing the adverse response at an
identified LOAEL depends on sample size,
variability of the response, and the dose
interval. LOAELs can represent a 30
percent or higher effect level for the
minimum sample sizes recommended for
standard test protocols. For this reason,
EPA recommends that NOAELs. instead of
LOAELs, are used to determine a screening
exposure level that is unlikely to adversely
impact populations.
of ten.2
Exposure route. The exposure route used in the toxiciry study should be
comparable to the exposure route in the risk assessment. For example, data from studies
where exposure is by gavage generally are not preferred for estimating dietary concentrations
that could produce adverse effects, because the rate at which the substance is absorbed from
the gastrointestinal tract usually is greater following gavage than following dietary
administration. Similarly, intravenous injection of a substance results in "instantaneous
absorption" and does not allow the substance to first pass through the liver, as it would
2 The literature reviews of McNamara (1976) and Wed and McColuSter (1963) included both rodent
and non-rodent species. The duration of the subchronic exposure usually was 90 days, but ranged from
30 to 210 days. A wide variery of endpoints and criteria for adverse effects were included in these
reviews. Despite this variation in the original studies, their findings provide a general indication of the
ratio between subchronic to chronic NOAELs for effects other than cancer and reproductive effects. For
some chemicals, chronic dosing resulted in increased chemical tolerance. For over 50 percent of the
compounds tested, the chronic NOAEL was less than the 90-day NOAEL by a factor of 2 or less.
However, in a few cases, the chronic NOAEL was up to a factor of 10 less than the subchronic NOAEL
(U.S. EPA, 1993e).
STEP 1, Page 11
-------
August 21. !996 DRAFT
following dietary exposure. If it is necessary to attempt to extrapolate toxicity test results
from one route of exposure to another, the extrapolation should be performed or reviewed by
a toxicoloeist experienced in route-to-route extrapolations for the class of animals at issue.
Field versus laboratory. Most toxicity studies evaluate effects of a single
contaminant on a single species under controlled laboratory conditions. Results from these
studies might not be directly applicable to the field, where organisms typically are exposed to
more than one contaminant in environmental situations that are not comparable to a laboratory
setting and where genetic composition of the population can be more heterogeneous than that
of organisms bred for laboratory use. In addition, the bioavailabdity of a contaminant is
different at a site than in a laboratory toxicity test. In a field situation, organisms also will be
subject to other environmental variables, such as unusual weather conditions, infectious
diseases, and food shortages. These variables can have either positive or negative effects on
the organism's response to a toxic contaminant that only a site-specific field study would be
able to evaluate. Moreover, single-species toxicity tests seldom provide information regarding
toxicant-related changes in community interactions (e.g., behavioral changes in prey species
that make them more susceptible to predation).
1.3.2 Dose Conversions
For some data reported in the literature, conversions are necessary to allow the data to
be used for species other than those tested or for measures of exposure other than those
reported. Many doses in laboratory studies are reported in terms of concentration in the diet
(e.g., mg contarrunant/kg diet or ppm in the diet). Dietary concentrations can be converted to
dose (e.g., mg contaminant/kg body weight/day) for comparison with estimated contaminant
intake levels in the receptor species,
Ingesuon rates and body weights for a test species often are reported in a toxicity
study or can be obtained from other literature sources (e.g., U.S. EPA, I993a,b). For
extrapolations between animal species with different metabolic rates as well as dietary
composition, consult U.S. EPA (1992e, 1996b).
1.3.3 Uncertainty Assessment
Professional judgment is needed to determine the uncertainty associated with
information taken from the literature and any extrapolations used in developing an screening
ecotoxiciry value. The nsk assessor should be consistently conservative in selecting literature
values and describe the limitations of using those values in the context of a particular site.
Consideration of the study design, endpoints, and other factors are important in determining
the utility of loxiciry data in the screening-level nsk assessment All of these factors should
be addressed in an uncertainty analysis prior to the screening-level nsk calculation.
STEP I, Page 12
-------
August 21. 1996 DRAFT
1.4 SUMMARY
At the conclusion of the screening-level problem formulation and ecological effects
evaJuation, the following information should have been compiled:
Environmental setting and contaminants known or suspected to exist at the site
and the maximum concentrations present (for each medium);
Contaminant fate and transport mechanisms that might exist at the site;
The mechanisms of ecotoxicity associated with contaminants and likely
categories of receptors that could be affected;
The complete exposure pathways that might exist at the site from contaminant
sources to receptors that could be affected; and
Screening ecotoxicity values equivalent to chronic N'OAELs based on
conservative assumptions.
For the screening-level ecological nsk assessment, assessment endpoints will include
any likely adverse ecological effects on receptors for which exposure pathways are complete,
as determined from the information listed above. Measurement endpoints will be based on
the available literature regarding mechanisms of toxicity and will be used to establish the
screening ecotoxicity values. Those values will be used with estimated exposure levels to
screen for ecological risks, as described in Step 2.
STEP 1, Page 13
-------
August 21. 1996 DRAFT
STEP 2: SCREENING-LEVEL EXPOSURE ESTIMATE
AND RISK CALCULATION
OVERVIEW
The screening-level exposure estimate and risk calculation comprise the second
step in the ecological risk screening for a site. Risk is estimated by comparing
maximum documented exposure levels with the ecotoxicity screening values from
Step 1. At the conclusion of Step 2, it will be decided that either: (1) the screening-
level ecological risk assessment is adequate to determine that ecological threats are
negligible; or (2) the process should continue to a more detailed ecological risk
assessment (steps 3 through 7). If the process continues, the screening-level
assessment serves to identify exposure pathways and preliminary contaminants of
concern for the baseline risk assessment by eliminating those contaminants and
exposure pathways that pose negligible risks.
2.1 INTRODUCTION
This step includes estimating exposure levels and screening for ecological risks as the
last two phases of the screening-level ecological risk assessment. The process concludes with
a Scientific/Management Decision Point (SMDP) at which it is determined that: (1)
ecological threats are negligible; (2) the ecological risk assessment should continue to
determine whether a risk exists; or (3) impacts are likely and a more detailed ecological risk
assessment, incorporating more site-specific information, is needed.
Section 2.2 describes the screening-level exposure assessment, focusing on the
complete exposure pathways identified in Step 1. Section 2.3 describes the risk calculation
process, including estimating a hazard quotient, documenting the uncertainties in the quotient,
and summarizing the overall confidence in the screening-level ecological risk assessment.
Section 2.4 describes the SMDP that concludes Step 2.
2.2 SCREENING-LEVEL EXPOSURE ESTIMATES
To estimate exposures for the screening-level ecological risk calculation, on-site
contaminant levels and general information on the types of biological receptors that might be
exposed should be known from Step 1. Only complete exposure pathways should be
STEP 2, Page 1
-------
AUZVJSI 21. 1996
DRAFT
evaluated. For these, the highest measured or estimated on-site contaminant concentration for
each environmental medium should be used to estimate exposures. This should ensure that
potential ecological threats are not be missed.
2.2.1 Exposure Parameters
For parameters needed to estimate exposures for which sound site-specific information
is lacking or difficult to develop, conservative assumptions should be used at this screening
level. Examples of conservative assumptions are listed below and described in the following
paragraphs:
Area use factor - 100 percent (factor
related to home range and population
density, see Highlight Box 2-1);
Bioavailabiliry - 100 percent;
Life stage - most sensitive life stage;
Body weight and food ingestion rate
- minimum body weight to
maximum ingestion rate; and
Dietary composition - 100 percent of
diet consists of the most
contaminated dietary component.
Area use factor. For the
screening-level exposure estimate for
terrestrial animals, assume that the home range of one or more animals is entirely within the
contaminated area, and thus the animals are exposed 100 percent of the time. This is a
conservative assumption and, as an assumption, is only applicable to the screening-level phase
of the nsk assessment. Species- and site-specific home range information would be needed
later, in Step 6, to estimate more accurately the percentage of time in animal would use a
contaminated area. Also evaluate the possibility that some species might actually focus their
activities m contaminated areas of the site. For example, if contamination has reduced
emergent vegetation in a pond, the pond might be more heavily used for feeding by waterfowl
than uncontaminated ponds with little open water.
Bloavailabllity. For the screening-level exposure estimate, in the absence of site-
specific information, assume that the bioavailabiiity of contaminants at the site is 100 percent
For example, at the screening-level, lead would be assumed to be 100 percent bioavailable to
HIGHLIGHT BOX 2-1
Area Use Factor
An animal's area use factor can be
defined as the ratio of the area of
contamination (or the site area under
investigation) to its home range, breeding
range, or feeding/foraging range. To ensure
that ecological risks are not underestimated,
the smallest reasonable area use factor
should be assumed. This allows the
maximum number of animals to be exposed
to site contaminants and makes it more
likely that "hot spots" (i.e., areas of
unusually high contamination levels) will be
significant proportions of individual animal's
home ranges.
STEP 2, Page 2
-------
August :i. 1996 DRAFT
mammals. While some literature indicates that mammals absorb approximately 10 percent of
ingested lead, absorption efficiency can be higher, up to about 60 percent, because dietary
factors such as fasting, and calcium and phosphate content of the diet, can affect the
absorption rate (Friberg et al., 1986). Because few species have been tested for
bioavailability, and because Steps 3 through 6 provide an opportunity for this issue to be
addressed specifically, the most conservative assumption is appropriate for this step.
Life stage. For the screening level assessment, assume that the most sensitive life
stages are present. If an early life stage is the most sensitive, the population should be
assumed to include or to be in that life stage. For vertebrate populations, it is likely that most
of the population is not in the most sensitive life stage most of the time. However, for many
invertebrate species, the entire population can be at an early stage of development during
certain seasons.
Body weight and food ingestion rates. Estimates of body weight and food
ingestion rates of the receptor animals also should be made conservatively to avoid
understating nsk, although uncertainties in these factors are far less than the uncertainties
associated with the environmental contaminant concentrations. U.S. EPA's Wildlife Exposure
Factors Handbook (U.S. EPA, 1993a,b) is a good source or reference to sources of this
information.
Bioaccumulation. Bioaccumulation values obtained from a literature search can be
used to estimate contaminant accumulation and food chain transfer at a Superfund site at the
screening stage (Steps 1 and 2). Because many environmental factors influence the degree of
bioaccumulation, sometimes by several orders of magnitude, the most conservative (i.e.,
highest) bioaccumulation factor reported in the literature should be used in the absence of
site-specific information. Thus, the most conservative BCF values identified in the literature
almost always are used to estimate bioaccumulation in screening-level ecological nsk
assessment in Step 1.
Dietary composition. For species that feed on more than one type of food, the
screening-level assumption should be that the diet is composed entirely of whichever type of
food is most contaminated. For example, if some foods (e.g., insects) are likely to be more
contaminated than other foods (e.g., seeds and fruits) typical in the diet of a receptor species,
assume that the receptor species feeds exclusively on the more contaminated type of food.
Again, EPA's Wildlife Exposure Factors Handbook (U.S. EPA, 1993a,b) is a good source or
reference to sources of this information.
2.2.2 Uncertainty Assessment
Professional judgment is needed to determine the uncertainty associated with
information taken from the literature and any extrapolations used in developing a parameter to
STEP 2, Page 3
-------
August 2! 1996 DRAFT
estimate exposures. All assumptions used to estimate exposures should be stated, including
some description of the --gree of bias possible in each. Where literature values are used, an
indication of the range of values that could be considered appropriate also should be
indicated.
2.3 SCREENING-LEVEL RISK CALCULATION
A quantitative screening-level nsk can be estimated using the exposure estimates
developed according to Section 2.2 and the screening ecotoxicity values developed according
to Section 1.3. For the screening-level nsk calculation, the hazard quotient approach, which
compares point estimates of screening ecotoxicity values and exposure values, is adequate to
estimate nsk. As descnbed in Section 1.3, the screening ecotoxicity value should be
equivalent to a documented and/or best conservatively estimated chronic NOAEL. Thus, for
each contaminant and environmental medium, the hazard quotient can be expressed as the
ratio of a potential exposure level to the NOAEL:
*
HQ - Dose or HQ = EEC
NOAEL NOAEL
where:
HQ = hazard quotient;
Dose = estimated contaminant intake at the site (e.g., mg contaminant/kg body
weight per day);
EEC = estimated environmental concentration at the site (e.g., mg
contaminant/L water, mg contaminant/kg sod, mg contaminant/kg food);
and
NOAEL = no-observed-adverse-effects-level (in units that match the dose or EEC).
An HQ less than one (unity) indicates that the contaminant alone is unlikely to cause adverse
ecological effects. If multiple contaminants of potential ecological concern exist at the site, it
might be appropnate to sum the HQs for receptors that could be simultaneously exposed to
the contaminants that produce effects by the same toxic mechanism (U.S. EPA, 1986a). The
sum of the HQs is called a hazard index iHT);
-------
August 21. 1996
DRAFT
HIGHLIGHT BOX 2-2
Hazard Index (HI) Calculation
For contaminants that produce adverse
effects by the same toxic mechanism:
Hazard Index =
EEC,/NOA£L,
EECVNOAEL,
EEC/NOAEL,*
quotient. As certainty in the exposure
concentrations and the NOAEL increase,
there is greater confidence in the predictive
value of the hazard quotient model, and
unity (HQ = 1) becomes a pass/fail decision
point.
The screening-level risk calculation
is a conservative estimate to ensure that
potential ecological threats are not
overlooked. The calculation is used to
document a decision about whether or not
there is a negligible potential for ecological
impacts, based on the information available
at this stage. If the potential for ecological
impacts exists, this calculation can be used
to eliminate the negligible-risk combinations
of contaminants and exposure pathways
from further consideration.
If the screening-level risk assessment
indicates that adverse ecological effects are
possible at environmental concentrations
below standard quantitation limits, a "non
detect" based on those limits cannot be used
to support a 'no risk" decision. Instead, the
risk assessor and risk manager should
request appropriate detection limits or agree
to continue to Step 3 where exposure concentrations can be estimated from other information
(e.g., fate and transport model, estimated dilution or attenuation from the source or areas
where the contaminant was detected).
where:
EEC. =
NOAELi =
estimated environmental
concentration for the 1th
contaminant; and
NOAEL for the i* contaminant
(expressed either as a dose or
environmental concentration).
The EEC and the NOAEL are expressed in
the same units and represent the same
exposure period (e.g., chronic). Dose could
substitute for EEC throughout provided the
NOAEL is expressed as a dose.
2.4 SCIENTIFIC/MANAGEMENT DECISION POINT (SMDP)
At the end of Step 2, the risk assessor communicates the results of the preliminary
ecological nsk assessment to the nsk manager. The risk manager needs to decide whether the
information available is adequate to make a nsk management decision, and could require
technical advice from the ecological risk assessment team to reach a decision. There are only
three possible decisions at this point:
STEP 2, Page 5
-------
August :i 1996 DRAFT
(1) There is adequate information to conclude chat ecological risks are negligible
and therefore no need for remediation on the basis of ecological risk;
11 \ The information is not adequate to make a decision at this point, and the
ecological nsk assessment process wiil continue to Step 3; or
(3) The information indicates a potential for adverse ecological effects, and a more
thorough assessment is warranted.
Note that the SMDP made at the end of the screening-level nsk calculation will not
set a preliminary cleanup goal Screening ecotoxiciry values are derived to avoid
underestimating nsk. Requiring a cleanup based solely on these values would not be
technically defensible.
The nsk manager should document both the decision and the basis for it. If the risk
characterization supports the first decision (i.e., negligible risk), the ecological risk assessment
process ends here with appropriate documentation to support the decision. The documentation
should include all analyses and references used in the assessment, including a discussion of
the uncertainties associated with the HQ and HI estimates.
For assessments that proceed to Step 3, the screening-level analysis in Step 2 can
indicate and justify which contaminants and exposure pathways can be eliminated from
further assessment because they pose negligible risk.
U.S. EPA must be confident that the SMDP made after completion of this calculation
will protect the ecological components of the environment. The decision to continue beyond
the screening-level nsk calculation does not indicate whether remediation is necessary at the
site. That decision will be made in Step 8 of the process.
2.5 SUMMARY
At the conclusion of the exposure estimate and screening-level nsk calculation step,
the following information should have been compiled:
(1) Exposure estimates based on conservative assumptions and maximum
concentrations present; and
(2) Hazard quotients (or hazard indices) indicating winch, if any, contaminants and
exposure pathways might pose ecological threats.
STEP 2, Page 6
-------
August 21. 1996 , DRAFT
Based on the results of the screening-level ecological risk calculation, the lead nsk
manager and lead nsk assessor will determine whether or not contaminants from the site pose
an ecological threat. If there are sufficient data to determine that ecological threats are
negligible, the ecological nsk assessment will be complete at this step with a finding of no
ecological nsk. If the data indicate that there is (or might be) a nsk of adverse ecological
effects, the ecological nsk assessment will continue.
Conservative assumptions have been used for each step of the screening-level
ecological nsk assessment. Therefore, requinng a cleanup based solely on this information
would not be technically defensible. To end the assessment at this stage, the conclusion of
negligible ecological risk must be adequately documented and technically defensible. A lack
of information on the toxicity of a contaminant or on complete exposure routes will result in
a decision to continue with the ecological risk assessment process (Steps 3 through 7)-not a
decision to delay the ecological risk assessment until a later date when more information
might be available.
STEP 2, Page 7
-------
August 21. 1996 DRAFT
STEP 3: BASELINE RISK ASSESSMENT PROBLEM FORMULATION
OVERVIEW
Step 3 of the eight-step process inmates the problem formulation phase of the
baseline ecological nsk assessment. Step 3 refines the screening-level problem
formulation and expands on the ecological issues that are of concern at the particular
site. In the screening-level assessment, conservative assumptions were used where site-
specific information was lacking. In Step 3, the results of the screening assessment
and additional site-specific information are used to determine the scope and goais of
the baseline ecological risk assessment. Steps 3 through 7 are required only for sites
for which the screening-level assessment indicated a need for further ecological risk
evaluation.
Problem formulation at Step 3 includes several activities:
Refining preliminary contaminants of ecological concern;
Further characterizing ecological effects of contaminants;
Reviewing and refining contaminant fate and transport, complete exposure
pathways, and ecosystems potentially at risk;
Selecting assessment endpoints; and
Developing of a conceptual model with working hypotheses or questions that
the site investigation will address.
At the conclusion of Step 3, there is a SMDP, which consists of agreement on four
items: the assessment endpoints, the exposure pathways, the nsk questions, and a
conceptual model integrating these components. The products of Step 3 are used to
select measurement endpoints and to develop the ecological nsk assessment work plan
(WP) and sampling and analysis plan (SAP) for the site in Step 4. Steps 3 and 4 are,
effectively, the data quality objective (DQO) process for the baseline ecological nsk
assessment.
3.1 THE PROBLEM FORMULATION PROCESS
In Step 3, problem formulation establishes the goals, breadth, and focus of the baseline
ecological nsk assessment. It also establishes the assessment endpoints, or specific ecological
values to be protected (U.S. EPA, 1992a). Through Step 3, the questions and issues that need
STEP 3, Page 1
-------
August 21. 1996 DRAFT
to be addressed in the baseline ecological nsk assessment are defined based on potentially
complete exposure pathways and ecologicaJ effects. A conceptual model of the site is
developed that includes questions about the assessment endpoints and the relationship between
exposure and effects. Step 3 culminates in an SMDP, which is agreement between the nsk
manager and nsk assessor on the assessment endpoints, exposure pathways, and questions as
portrayed ;n the conceptual model of the site.
The conceptual model, which is completed in Step 4, also will describe the approach,
types of data, and analytical tools to be used for the analysis phase of the ecological nsk
assessment (Step 6). Those components of the conceptual model are formally descnbed in
the ecological nsk WP and SAP in Step 4 of this eight-step process. If there is not
agreement among the nsk manager, nsk assessor, and the other professionals involved with
the ecological nsk assessment on the initial conceptual model developed in Step 3, the final
conceptual model and field study design developed in Step 4 might not resolve the issues that
must be considered to manage nsks effectively.
The complexity of questions developed during problem formulation does not depend
on the size of a site or the magnitude of its contamination. Large areas of contamination can
provoke simple questions and, conversely, small sites with numerous contaminants can require
a complex senes of questions and assessment points. There is no rule that can be applied to
gauge the effort needed for an ecological nsk assessment based on site size or number of
contaminants; each site should be evaluated individually.
At the beginning of Step 3, some basic information should exist for the site. At a
minimum, information should be available from the site history, PA, SI, and Steps 1 and 2 of
this eight-step process. For large or complex sites, information might be available from
earlier site investigations.
It is important to be as complete as possible early in the process so that Steps 3
through 8 need not be repeated. Repeating the selection of assessment endpoints and/or the
questions and hypotheses concerning those endpoints is appropnaie only if new information
indicating new threats becomes available. The SMDP process should prevent having to return
to the problem formulation step because of changing opinions on the questions being asked.
RepcLtion of Step 3 should not be confused with the intentional uer.ng (or phasing) of
ecological site investigations at large or complex sites (sec Highlight Box 3-1). The process
of problem formulation at complex sites is the same as at more simple sites, but the number,
complexity, and/or level of resolution of the juesuons and hypotheses can be greater at
complex sues.
While problem formulation is conceptually simple, in practice it is generally a
complex and interactive process. Defining the ecological problems to be addressed during the
baseline nsk assessment involves identifying toxic mechanisms of the contaminants,
STEP 3, Page 2
-------
August 21. 1996
DRAFT
characterizing potential receptors, and
estimating exposure and potential
ecological effects. Problem formulation
also constitutes the DQO process for the
baseline ecological risk assessment (U.S.
EPA, 1993d. "
The remainder of this section
describes six activities to be conducted
prior to the SMDP for this step:
refining preliminary contaminants of
ecological concern (Section 3.2); a
literature search on the potential
ecological effects of the contaminants
(Section 3.3); qualitative evaluation of
complete exposure pathways and
ecosystems potentially at risk (Section
3.4); selecting assessment endpoints
(Section 3.5); and developing the
conceptual model and establishing risk
questions (Section 3.6).
3.2 REFINEMENT OF
PRELIMINARY
CONTAMINANTS OF
CONCERN
HIGHLIGHT BOX 3-1
Tiering an Ecological Risk
Assessment
The ecological nsk assessment at
Superfund sites is at least a two-tier process.
Steps 1 and 2 of this guidance serves as a first,
or screening, tier pnor to expending a larger
effort for a detailed, site-specific ecological risk
assessment. The baseline risk assessment serves
as a second tier. More than one tier could be
needed in the baseline nsk assessment for large
or complex sites where there is a need to
sequentially test interdependent hypotheses
developed during problem formulation (i.e.,
evaluating the results of one field assessment
before designing a subsequent field study based
on the results).
While tienng can be an effective way of
to manage site investigations, multiple sampling
phases typically require some resampling
matrices sampled during earlier tiers and
increased field mobilization costs. Thus, in
some cases, a tiered ecological risk assessment
might cost more than a non-tiered assessment.
The benefits of tienng should be weighed
against the costs.
The results of the screening-level
risk assessment (Steps 1 and 2) should
have indicated which contaminants found at the site can be eliminated from further
consideration and which should be evaluated further. It is important to realize that
contaminants that might pose an ecological nsk can be different from those that might pose a
human health nsk because of differing exposure pathways, sensitivities, and responses to
contaminants.
The initial.list of contaminants investigated in Steps 1 and 1 included ail contaminants
identified or suspected to be at the site. During Steps I and 2, it is likely that several of the
contaminants found at the site were eliminated from further assessment because the risk
screen indicated that they posed a negligible ecological risk. Because of the conservative
assumptions used during the nsk screen, some of the contaminants retained for Step 3 also
might pose negligible risk. At this stage, the risk assessor should review the assumptions
STEP 3, Page 3
-------
August 21. 1996 DRAFT
used (e.g., 100 percent bioavailability) against vaJues reported in the literature (e.g., only up
to 60 percent for a particular contaminant), and consider how the HQs would change if more
realistic conservative assumptions were used instead. For those contaminants for which the
HQs drop to near or below unity, the nsk assessor and nsk manager should discuss and agree
on which can be eliminated from further consideration at this time.
Sometimes, new information becomes available that indicates the initial assumptions
that screened some contaminants out in Step 2 are no longer valid (e.g., site contaminant
levels are higher than originally reported, joint-action toxicity was not considered in the
preliminary nsk calculation, etc.). In this case, contaminants can be placed back on the list of
contaminants to be investigated with that justification.
Note that a contaminant should not be eliminated from the list of contaminants to be
investigated only because toxicity information is lacking; instead, limited or missing toxicity
information must be addressed using best professional judgment and discussed as an
uncertain ry.
3.3 LITERATURE SEARCH ON KNOWN ECOLOGICAL EFFECTS
The literature search conducted for the screening-level risk assessment (Steps 1 and 2)
should be expanded to obtain the information needed for the more detailed problem
formulation phase of the baseline ecological nsk assessment If pre-developed screening
ecotoxicity values (e.g., ecotox thresholds developed by the Superfund Program; U.S. EPA,
1996a) were applied in Steps 1 and 2, then the literature search might need to be expanded to
identify NOAELs, LOAELs, exposure-response functions, and the mechanisms of toxic
responses. Appendix C presents a discussion of some of the factors important in conducting a
lerarurc search. Several U.S. EPA publications (e.g., U.S. EPA, 1995a,e,g,h) provide a
window to original toxicity Literature for contaminants often found at Superfund sites.
3.4 CONTAMINANT FATE AND TRANSPORT, ECOSYSTEMS POTENTIALLY AT
RISK, AND COMPLETE EXPOSURE PATHWAYS
A preliminary identification of contaminant fate and transport, ecosystems potentially
at nsk, and complete exposure pathways was conducted in the screening ecological nsk
assessment. In Step 3, the exposure pathways and the ecosystems associated with the
assessment endpoints that were retained by the screening nsk assessment are evaluated in
more detail. This effort typically involves compiling additional information on:
(1) The environmental fate and transport of the contaminants;
STEP 3, Page 4
-------
August 21. 1996
DRAFT
(2) The ecological setting of the site (including habitat, potential receptors, etc.);
and
(3) The magnitude and extent of contamination, including its spatiaJ and temporal
variability.
For individual contaminants, it is frequently possible to reduce the number of exposure
pathways that need to be evaluated to one or a few "critical exposure pathways" which (1)
reflect maximum exposures of receptors within the ecosystem or (2) constitute exposure
pathways to ecological receptors sensitive to the contaminant. The critical exposure pathways
influence the selection of assessment endpoints for a particular site. If multiple critical
exposure pathways exist, they each should be evaluated because it is often difficult to predict
which pathways could be responsible for the greatest ecological risk.
3.4:1 Contaminant Fate and Transport
Information on how the contaminants will or could be transported or transformed in
the environment physically, chemically, and biologically is used to identify the exposure
pathways that rrught lead to significant ecological effects (see Highlight Box 3-2).
Chemically, contaminants can undergo several processes in the environment:
Degradation3;
Complexation;
lonization;
Precipitation; and/or
Adsorption.
Physically, contaminants might move
through the environment by one. or more
means:
Volatilization;
Erosion;
Deposition (contaminant
sinks);
Weathering of parent material
with subsequent transport;
anoVor
HIGHUGHT BOX 3-2
Environmental Fate and Exposure
If a contaminant in an aquatic
ecosystem is highly lipophilic (i.e.,
essentially insoluble in water), it is likely to
partition primarily into sediments and not
into the water column. When sampling
sediments for contamination, characterizing
sediment grain-size and total organic carbon,
which can influence contaminant
partitioning, is important in evaluating
ecological exposure.
The product might be more toxic or less toxic than the parent compound.
STEP 3, Page 5
-------
August 21. 1996 DRAFT
Water transport:
in solution.
as suspended matenaJ in the water, and
bulk transport of solid material.
Several biological processes also affect contaminant fate and transport in the environment:
Bioaccumulauon;
Bi ode gradation;
Biological transformation ;
Food chain transfers; and/or
Excretion.
Additional information should be gathered on past as well as current mechanisms of
contaminant release from source areas at the site. The mechanisms of release along with the
chemical and physical form of a contaminant can affect its fare, transport, and potential for
reaching ecological receptors.
A contaminant flow diagram (or exposure pathway diagram) comprises a large part of
the conceptual model, as illustrated in Section 3.6. A contaminant flow diagram originates at
the primary contaminant source(s) and identifies primary release mechanisms and contaminant
transport pathways. The release and movement of the contaminants can create secondary
sources (e.g., contaminated sediments in a river; see Example Box 3-1), and even tertiary
sources.
The above information is used to evaluate: (1) where the contaminants are likely to
partition in the environment; and (2) the bioavailability of the contaminant (historically,
currently, or in the future).
3.4.2 Ecosystems Potentially at Risk
The ecosystems or habitats potentially at nsk depend on the ecological setting of a
site. An initial source of information on the ecological setting of a site is the data collected
during the preliminary site visit and characterization (Step 1), including the site ecological
checklist (Appendix B). The site description should provide answers to several questions
including:
What habitats (e.g., maple-beech hardwood forest, early successional fields) are
present17
The product might be more toxjc or less IOKJC than the parent compound.
STEP 3, Pige 6
-------
August :i. 1996 DRAFT
What types of water bodies are present, if any?
Do any other habitats listed in Exhibit 1-1 exist on or adjacent to the site?
EXAMPLE BOX 3-1
Exposure Pathway Model-OOT Site
An abandoned pesticide production facility had released DDT to soils through poor
handling practices dunng its operation. Due to erosion of contaminated soils, DDT migrated to
stream sediments. The contaminated sediments might affect benthic organisms through direct
contact or ingestion. Benthic organisms that have accumulated DDT can be consumed by fish.
and fish that have accumulated DDT can be consumed by piscivorous birds. This example
illustrates how contaminant transport is traced from a primary source to a secondary source and
from there through a food chain to an exposure point that can affect an assessment endpoint.
While all available information must be used, it is not cnucal that complete site
setting information be collected during thus phase of the risk assessment. However, it is
important that habitats at the site are not overlooked; hence, a site visit might be needed to
supplement the site visit conducted during the screening risk assessment. If a habitat that is
actually present on the site is omitted during the problem formulation phase, this step could
need to be repeated later when the habitat is found, resulting in delays and additional costs
for the nsk assessment.
Available information on ecological effects of contaminants (see Section 3.3) can help
focus the assessment on specific ecological resources that should be evaluated more
thorouanly, because some groups of organisms can be more sensiuve (more susceptible) than
others to a particular contaminant. For example, a species or group of species could be
physiologically sensitive to a particular contaminant (e.g., the contaminant might interfere
with the species' hormone systems); or, the species might not be able to metabolize and
detoxify the particular contaminant(s) (e.g., honey bees and grass snnmp cannot effectively
biodegrade PAHs, whereas fish generally can). Alternatively, an already stressed population
(e.g., due to habitat degradation) could be particularly sensitive to any added stresses.
Variation in sensitivity should not be confused with vanauon in exposure, which can
result from behavioral and dietary differences among species. For example, predators can be
exposed to higher levels of contaminants that biomagmfy in food chains than are herbivores.
A specialist predator could feed primarily on one prey type that is a primary receptor of the
contaminant. Some species might preferentially feed in a habitat where the contaminant tends
to accumulate. On the other hand, a species might change its behavior to avoid contaminated
STEP 3, Page 7
-------
August :i. 1996 DRAFT
areas. Both susceptibility to toxic effects of a contaminant and behaviors that affect exposure
levels can influence risks for particular groups of organisms.
3.4.3 Complete Exposure Pathways
The potentially complete exposure pathways identified in Steps 1 and 2 are described
in more detail in Step 3 on the basis of the refined contaminant fate and transport evaluations
(Section 3.3.1) and evaluation of potential ecological receptors (Section 3.3.2).
Some of the potentially complete exposure pathways identified in Steps I and 2 might
be ruled out from further consideration at this time. Sometimes, additional exposure
pathways might be identified, particularly those originating from secondary sources. Any data
gaps that result in questions about whether an exposure pathway is complete should be
identified, and the type of data needed to answer those questions should be described to assist
in developing the WP and SAP in Step 4.
During Step 3, the potential for food-chain exposures deserves particular attention.
Some contaminants are effectively transferred through food chains, while others are not. To
illustrate this point, copper and DDT are compared in Example Box 3-2.
EXAMPLE BOX 3-2
Potential for Food Chain Transfer-Copper and DOT Sites
Copper can be toxic in aquatic ecosystems or to terrestrial plants. However, it is an
essential nutnem for both plants and animals, and organisms can regulate internal copper
concentrations within limits. For this reason, copper tends not to accumulate in most organisms
or to biomagnify m food chains, and thus tends not to reach levels high enough to cause
adverse responses through food chain transfer 10 upper-trophic-level organisms. (Copper is
known to accumulate by several orders of magnitude in phyxoplankton and in filter-feeding
mollusks, however, and thus can pose a threat to organisms that feed on these components of
aquatic ecosystems; U.S. EPA, 1985a.) In contrast, DDT, a contaminant that accumulates in
fatty tissues, can biomagnify in many different types of food chains. Upper-trophic-level
species (such as predatory birds), therefore, are likely to be exposed to higher levels of DDT
through their prey than are lower-trophic-level species in the ecos;item.
3.5 SELECTION OF ASSESSMENT ENDPOINTS
\s noted in tie introduction to this guidance, an assessment endpoint is "an explicit
expression of the environmental value that is to be protected" (U.S. EPA, 1992a). There is an
STEP 3, Page 8
-------
August 21. 1996 DRAFT
important distinction between human health nsk assessment and ecological risk assessment.
In human health risk assessment, only one species is evaluated, and cancer and noncancer
systemic effects are the usual assessment endpoints. Ecological nsk assessment, on the other
hand, involves multiple species that are likely to be exposed to differing degrees and to
respond differently to the same contaminant. Nonetheless, it is not practical or possible to
directly evaluate risks to all of the individual components of the ecosystem at a site. Instead,
assessment endpoints focus the risk assessment on particular components of the ecosystem
that could be adversely affected by contaminants from the site.
The selection of assessment endpoints includes discussion between the nsk assessor
and the nsk manager concerning management policy goals and ecological values. Input from
the regional BTAG, PRPs, and other stakeholders associated with a site at this stage can help
the risk assessor to identify ecological assessment endpoints that the risk manager can clearly
defend when making decisions for the site. ECO Update Volume 3, Number 1, briefly
summarizes the process of selecting assessment endpoints (U.S. EPA, 1995b).
Individual assessment endpoints usually encompass a group of species or populations
with some common characteristics, such as a specific exposure route or contaminant
sensitivity. Sometimes, individual assessment endpoints are limited to one species (e.g., a
species known to be particularly sensitive to a site contaminant). Assessment endpoints also
can encompass the typical structure and function of biological communities or ecosystems
associated with a site.
Assessment endpoints for the baseline ecological risk assessment must be selected
based on the ecosystems, communities, and/or species at the site. The selection of assessment
endpoints depends on:
(1) The contaminants present and their concentrations;
(2) Mechanisms of toxicity of the contaminants to different groups of organisms;
(3) Potential sensitive or highly exposed receptor groups present and attributes of
their natural history; and
(4) Potential complete exposure pathways.
Thus, the process of selecting assessment endpoints can be, and often is, iterative and
interactive with other phases of problem formulation.
The nsk assessor must "think through" the contaminant mechanism(s) of ccotoxicity to
determine what receptors will or could be at nsk. This understanding must include how the
adverse effects of the contaminants might be expressed (e.g., eggshell thinning in birds), as
STEP 3, Page 9
-------
August :i. 1996 DRAFT
well as how the chemical and physical form of the contaminants influence bioavailability and
the type and magnitude of adverse response (e.g., inorganic versus organic mercury).
The nsk assessor also should determine if the contaminants adversely affect organisms
in direct contact with the contaminated media (i.e., direct exposure to water, sediment, soil) or
if the contaminants accumulate in food chains, resulting in adverse effects in organisms that
are not directly exposed or are minimally exposed to the original contaminated media
(indirect exposure). The risk assessor should decide if the nsk assessment should focus on
toxiciry resulting from direct or indirect exposures, or if both must be evaluated.
Broad assessment endpomts (e.g., protecting aquatic communities) are generally of less
value in problem formulation than specific assessment endpoints (e.g., maintaining aquatic
community composition and structure downstream of a site similar to that upstream of the
site). Specific assessment endpoints define the ecological value in sufficient detail to
identify the measures needed to answer specific questions or to test specific hypotheses.
Example Box 3-3 provides three examples of assessment endpomt selection based on the
hypothetical sites in Appendix A.
The formal identification of assessment endpoints is pan of the SMDP for this step.
Regardless of the level of effort to be expended on the subsequent phases of the nsk
assessment, the assessment endpoints identified are critical elements in the design of the
ecological nsk assessment and must be agreed upon as the focus of the risk assessment.
Once assessment endpoints have been selected, testable hypotheses and measurement
endpoints can be developed to determine whether or not a potential threat to the assessment
endpoints exists. Testable hypotheses and measurement endpoints cannot be developed
without agreement on the assessment endpoints among the risk manager, risk assessor, and
other involved professionals.
3.6 THE CONCEPTUAL MODEL AND TESTABLE HYPOTHESES
The site conceptual model establishes the complete exposure pathways that will be
evaluated in the ecological nsk assessment and the relationship of the measurement endpoints
to the assessment endpoints. In the conceptual model, the possible exposure pathways
depicted in the exposure pathway diagram must be linked directly to the assessment endpoints
identified in Section 3.5. Risk questions, testable hypotheses, measurement endpoints, and the
SAP are based on the conceptual model. Developing the conceptual model and risk questions
are descnbed in Sections 3.6.1 and 3.6.2, respectively. Selection of measurement endpoints,
completing the conceptual model, is descnbed in Step 4.
STEP 3. Page 10
-------
August :i. 1996 DRAFT
EXAMPLE BOX 3-3
Assessment Endpoint Selection
DDT Site
An assessment endpomt such as protection of the ecosystem from the effects of DDT"
would give little direction to the nsk assessment. However, "protection of piscivorous birds
from eggshell thinning due to DDT exposure" directs the nsk assessment toward the food-chain
transfer of DDT that results in eggshell thinning in a specific group of birds. This assessment
endpoint provides the foundation for identifying appropriate measures of effect and exposure
and ultimately the design of the site investigation for the site. It is not necessary that a specific
piscivorous species of bird be identified on site. It is necessary that the exposure pathway to a
piscivorous bird exists and that the presence of a piscivorous bird could be expected.
Copper Site
Copper can be acutely or chronically toxic to organisms in an aquatic community
through direct exposure of the organisms to copper in the water and sediments. Threats of
copper to higher trophic level organisms are unlikely to exceed threats to organisms at the base
of the food chain, because copper is an essential nutrient which is effectively regulated by most
organisms if the exposure is below immediately toxic levels. Aquatic plants (particularly
phytoplankton) and mollusks, however, are poor at regulating copper and might be sensitive
receptors or effective in transferring copper to the next trophic level. In addition, fish fry can
be very sensitive to eopper in water. Based on these receptors and the potential for both acute
and chronic toxictty, an appropriate general assessment endpoint for the system could be the
maintenance of the pond community composition. An operational definition of the assessment
endpoint would be pond fish and invertebrate community composition similar to that of other
ponds of similar size and characteristics in the area.
PCS Site
The primary ecological threat of PCBs in ecosystems is not through direct exposure and
acute toxicity. Instead, PCBs bioaccumulate in food chains and can diminish reproductive
s-ccess in some vertebrate species. PCBs have been implicated as a cause of reduced
reproductive success of piscivorous birds (e.g., cormorants, terns) in the Great Lakes (Colborn,
1991) and of mink along several waterways (Aulerich and Ringer, 1977; Foley et al., 1988).
Therefore, reduced reproductive success in high-trophic-level species exposed via their diet is a
more appropriate assessment endpoint than either toxicity to organisms via direct exposure to
PCBs in water, sediments, or soils, or reproductive impairment in lower-trophic-level species.
STEP 3, Page 11
-------
August 21. 1996 DRAFT
3.6.1 Conceptual Model
Based on the informauon obtained from Steps 1 and 2 of the ecological nsk
assessment process, knowledge of the contaminants present, the exposure pathway diagram,
and the assessment endpomts, an integrated conceptual model is developed (see Example Box
3-4). The conceptual model includes a contaminant fate and transport diagram that traces the
contaminants' movement from sources through the ecosystem to receptors that include the
assessment endpomts (see Example Box 3-5V Contaminant exposure pathways that do not
lead to a species or group of species associated with the proposed assessment endpomt
indicate that either:
(1) There is an incomplete exposure pathway to the receptors) associated with the
proposed assessment endpomt: or
(2) There are missing components or data necessary to demonstrate a complete
exposure pathway.
EXAMPLE BOX 3-4
Description of the Conceptual Model-DDT Site
One of the assessment endpoints selected for the DDT site (Appendix A) is the
protecuon of piscivorous birds. The site conceptual model includes the release of DDT
from the spill areas to the adjacent stream, followed by food chain accumulation of DDT
from the sediments and water through the lower trophic levels to forage fish in the stream.
The forage fish are the exposure point for piscivorous birds. Eggshell thinning was
selected as the measure of effect During the literature review of the ecological effects of
DDT. loxicity studies were found that reported reduced reproductive success (i.e., number
of young fledged) in birds that experienced eggshell thinning of 2C percent or more
(Anderson and Hickey, 1972; Dilworth et al., 1972). Based on those data, the nsk
assessor and nsk manager agreed that eggshell thinning of 20 percent or more would be
considered an adverse effect for piscivorous birds.
Another effect of chronic DDT exposure on some animals is to reduce their ability
to escape predanon. Thus, DDT can indirectly increase the mortality rate of these
organisms by making them more susceptible to predators (Cooke, 1971; Krebs et al.,
1974). This effect of DDT on prey also can have an indirect consequence for the
predators. If predators are more likely to capture the more contaminated prey, the
predators could be exposed to DDT at levels higher than represented in the avenge prey
population.
STEP 3, Page 12
-------
August 21. 1996
DRAI
EXAMPLE BOX 3-5
Conceptual Model Diagram - DDT Site
MEASUREMENT ENDPOINT
(DDT concentration in fish tissue,
exposure point for kingfisher)
SECONDARY RECEPTOR
(Fish)
IERIIARY RECEPIOR
(Kingfisher)
PRIMARY SOURCE
(Plant site)
SECONDARY SOURCE
(Surface drainage) I
TERTIARY SOURCE
(Stream sediments, exposure point
for fish and macroinvertcbralcs)
PRIMARY RECEI'IOR
(l)cnlhic macroinverlciuules.
ex|iosurc point, fish)
MEASUREMEN I ENDI'OIN I
(Benlhic macroinvcrtebralc
community structure)
STEP 3. |»age I
-------
August II. 1996 DRAFT
If case (1) is true, the proposed assessment endpoint should be Devaluated to determine if it
is an appropriate endpoint for the sue. If case (2) is true, then additional field data could be
needed to evaluate contaminant fate and transport at the site. Failure to identify a complete
exposure pathway that does exist at the site can result in incorrect conclusions or in extra
time and effort being expended on a supplementary investigation to the ecological nsk
assessment.
As indicated in Section 3.5, appropriate assessment endpomis differ from site to site,
and can be at one or more levels of biological organization. At any particular site, the
appropriate assessment endpomts might involve local populations of a particular species,
community level integrity, and/or habitat preservation. The site conceptual model must
encompass the level of biological organization appropriate for the assessment endpoints for
the site. The conceptual model can use assumptions that are generally representative of a
group of organisms or ecosystem components encompassed by the assessment endpoint in that
region of the country.
The intent of the model is not to describe exactly a particular species or site as much
as it is to be systematic, representative, and conservative where information is lacking (with
assumptions biased to be more likely to overestimate than to underestimate risk). For
example, it is not necessary or even recommended to develop new test protocols to use
species that exist at a site to test the toxicity of site media. Species used in standardized
laboratory toxicity tests (e.g., fathead minnows, Hyallela amphipods) usually are adequate
surrogates for species in their general taxa and habitat at the site.
3.6.2 Risk Questions
Ecological risk hypotheses for Superfund sites basically arc questions about the
relationships among assessment endpoints and their predicted responses when exposed to
contaminants. The nsk questions should be based on the assessment endpoints (Step 3);
provide a basis for the development of the study design (Step 4>; and provide a structure for
evaluating the results of the site investigation in the analysis phase Step 6) and during nsk
characterization (Step 7).
The most basic test hypothesis applicable to virtually all Superfund sites is that site-
related contaminants are causing adverse effects (or could cause adverse effects in the future)
on the assessment endpoints). To use the baseline ecological nsk assessment in the FS to
evaluate remedial alternatives, it is helpful if the specific contaminant(s) responsible can be
identified. Thus refined, the question becomes "is chemical X is causing (or could cause)
adverse effects on the assessment endpoint0" In general, there are four Lines of evidence that
can be used to answer this question:
STEP 3. Page 14
-------
August li. 1996
DRAFT
(1) Comparing estimated or measured
exposure levels to chemical X
with levels that are known from
the literature to be toxic to
receptors associated with the
assessment endpoints;
(2) Comparing laboratory bioassays
with media from the site with
boiassays with media from a
reference site;
(3) Comparing in situ toxicity tests at
v the site with in situ toxicity tests
in a reference body of water,
and
(4) Comparing observed effects in
the receptors associated with the
site with similar receptors at a
reference site.
HIGHLIGHT BOX 3-3
Definitions:
Null and Test Hypotheses
Null hypothesis: Usually a hypothesis of
no differences between two populations
formulated for the express purpose of being
rejected.
Test-(or alternative) hypothesis: An
operational statement of the investigator's
research hypothesis.
When appropriate, formal hypothesis
testing is preferred. However, it might not
be practical for some assessment endpoints
or be the only acceptable way to state
questions about those endpoints.
These lines of evidence are considered further in Step 4, as measurement endpoints are
selected and the site-specific study is designed.
3.7 SCIENTIFIC/MANAGEMENT DECISION POINT (SMDP)
At the conclusion of Step 3, there is a SMDP. The SMDP consists of agreement on
four items: the assessment endpoints, exposure pathways, conceptual model, and risk
questions. Without agreement between the nsk manager, risk assessor, and other involved
professionals on these items, measurement endpoints cannot be selected, and a site study
cannot be developed.
These items can be summarized with the assistance of the diagram of the conceptual
model. Example Box 3-5 shows the conceptual model for the DDT site example in
Appendix A.
STEP 3. Page 15
-------
August 21. 1996 DRAFT
3.8 SUMMARY
By combining information on: (1) the potential contaminants present; (2) the
ecological setting; (3) environmental fate and transport; and (4) the ecotoxicity of the
contaminants, an evaluation is made of what aspects of the ecosystem at the site could be at
nsk and what the adverse ecological response could^be. "Critical exposure pathways" are
based on: (1) exposure pathways to sensitive species populations, or communities; and (2)
exposure levels associated with predominant fate and transport mechanisms at a site.
Based on that information, the nsk assessor and risk manager agree on assessment
endpoints and specific questions or testable hypotheses that, together with the conceptual
model, form the basis for the site investigation. At this stage, site-specific information on
exposure pathways and/or the presence of specific species is likely to be incomplete. By
using the conceptual model developed in Step 3, measurement endpoints and a plan for filling
information gaps can be developed and written into the ecological WP and SAP as described
in Step 4.
STEP 3, Page 16
-------
August 21. 1996 _ DRAFT
STEP 4: STUDY DESIGN AND DATA QUALITY
OBJECTIVE PROCESS
OVERVIEW
The site conceptual model developed in Step 3, which includes exposure
pathways, assessment endpomts, and questions or hypotheses, is used to develop
measurement endpoints, the study design, and data quality objectives in this step.
The products of Step 4 are the ecological risk assessment work plan (WP) and
sampling and analysis plan (SAP), which describe the details of the site
investigation as well as the data analysis methods and data quality objectives
(DQOs). As part of the DQO process, the SAP specifies acceptable levels of
decision errors that will be used as the basis for establishing the quantity and
quality of data needed to support ecological risk management decisions.
The lead risk assessor and the lead risk manager should agree that the WP
and SAP describe a study that will provide the risk manager with the information
needed to fulfill the requirements of the baseline risk assessment and to
incorporate ecological considerations into the site remedial process. Once this
step is completed, most of the professional judgment needed for the ecological
risk assessment will have been incorporated into the design and details of the WP
and SAP This does not limit the need for qualified professionals in the
implementation of the investigation, data acquisition, or data interpretation.
However, there should be no fundamental changes in goals or approach to the
ecological risk assessment once the WP and SAP are finalized.
It is important to coordinate this step with the WP and SAP for the site
investigation, which is used to document the nature and extent of contamination
and to evaluate human health risks.
Step 4 of the ecological risk assessment establishes the measurement endpoints
(Section 4.1), study design (Section 4.2), and data quality objectives based on statistical
considerations (Section 4.3) for the site assessment that will accompany site-specific studies
for the remedial investigation. The site conceptual model developed in Step 3 is used to
identify which points or assumptions in the nsk assessment include the greatest degree of
conservatism or uncertainty. The field sampling then can be designed to address the nsk
model parameters that have important effects on the risk estimates (e.g., bioavailability and
toxicity of contaminants in the field, contaminant concentrations at exposure points).
STEP 4, Page 1
-------
August 2i. ;996 DRAFT
The products of Step 4 are the WP and SAP for the ecological component of the field
investigations < Section 4 4) Involvement of the BTAG in the preparation, review, and
approval of WPs and SAPs can help ensure that the ecological nsk assessment is well
focused, performed efficiently, and technically correct.
The WP and SAP should specify the site conceptual model developed in Step 3, and
the measurement endpomts developed in the beginning of Step 4. The WP describes:
Assessment endpomts;
Exposure pathways;
Questions and testable hypotheses;
The relationship of measurement endpomts to the assessment endpomts; and
Uncertainties and assumptions.
The SAP should describe:
Data needs;
Scientifically valid and sufficient study design and data analysis procedures;
Study methodology and protocols, including sampling techniques;
Data reduction and interpretation techniques, including statistical analyses; and
Quality assurance procedures and quality control techniques.
The SAP must include the data reduction and interpretation techniques, because it is necessary
to known how the data will be interpreted to specify the number of samples needed.
Pnor to formal agreement on the WP and SAP, the proposed field sampling plan is
verified in Step 5.
4.1 ESTABLISHING MEASUREMENT ENDPOINTS
As indicated m the Introduction, a measurement endpoint is defined as "a measurable
ecological characteristic that is related to the valued characteristic ;~osen as the assessment
endpoint' iL'.S. EPA, 1992a) and is a measure of biological effects e.g., mortality,
reproduction, growth). Measurement endpomts arc frequently numerical expressions of
observations (e.g., toxicity test results, community diversity measures) that can be compared
statistically to a control or reference site to detect adverse responses to a site contaminant.
The relationship between measurement and assessment endpomts must be clearly described
within the conceptual model and must be based on scientific evidence. This is critical
because the assessment and measurement endpomts usually are different endpomts (see the
Introduction).
STEP 4, Page 2
-------
August 21. 1996
DRAFT
Highlight Box 4-1
Importance of Distinguishing
Measurement from Assessment
Endpoints
If a measurement endpoint is
mistaken for an assessment endpoint, the
misperception can anse that Superfund is
basing a remediation on an arbitrary or
esoteric justification. For example,
protection of a few invertebrate and algal
species could be mistaken as the basis for a
remedial decision, when the actual basis for
the decision is the protection of the aquatic
community as a whole (including higher-
trophic-level game fish that depend on lower
trophic levels in the community), as
indicated by a few sensitive invertebrate and
algal species.
Typically, the number of
measurement endpoints that are potentially
appropriate for any given assessment
endpoint and circumstance are limited. The
most appropriate measurement endpoints for
an assessment endpoint depend on several
considerations, a primary one being how
many and which lines of evidence are
needed to support risk management
decisions at the site (see Section 3.6.2).
The risk assessor must consider the utility of
each type of data, the cost of collecting the
data, and the likely sensitivity of the risk
estimates to the data. Given the potential
ramifications of site actions, the site risk
manager might want to use more than one
line of evidence to identify site-specific
thresholds for effects.
There are some situations in which it
might only be necessary or possible to
compare estimated or measured contaminant exposure levels at a site to ecotoxicity values
derived from the literature. For example, for contaminants in surface waters for which there
are state water quality standards, exceedance of the standards indicates that remediation to
reduce contaminant concentrations in surface waters to below these levels couid be needed
whether impacts are occurring or not. For assessment endpoints for which impacts are
difficult to demonstrate in the field (e.g., because of high natural variability), and toxicity
tests are not possible (e.g., food chain accumulation is involved), comparing environmental
concentrations with a well supported ecotoxicity value might have to suffice.
A toxicity test on contaminated medja from the site can suffice if the risk manager and
risk assessor agree that laboratory toxicity tests with surrogate species will be taken as
indicative of likely effects on the assessment endpoint. For sites with complex mixtures of
contaminants without robust ecotoxicity values and high natural variability in potential
measures for the assessment endpoint, either laboratory or in situ toxjcity testing might be the
best technique for evaluating risks to the assessment endpoint. For inorganic substances in
soils or sediments, toxjcity testing often is needed to determine the degree to which a
contaminant is bioavailable at a particular sue. Laboratory toxicity tests can indicate the
potential for adverse impacts in the field, whale m situ toxicity testing with resident organisms
can provide evidence of actual impacts occurring in the field.
STEP 4, Page 3
-------
August I: .9% DRAFT
Sometimes more than one line of evidence is needed to reasonably demonstrate that
contaminants from a site are likely to cause adverse effects on the assessment endpoint. For
example, total recoverable copper in a surface water body could exceed aquatic ecotoxicity
values, but not cause adverse effects because the copper is only partially bioavailablc or
because the ecocoxicity value is too conservauve for the particular ecosystem. Additional
evidence from bioassays or community surveys could help resolve whether the copper is
actually having adverse effects (See Example Box 4-1). Alternatively, if stream community
surveys indicate impairment of community structure downstream of a site, comparing
contaminant concentrations with aquatic toxiciry values can help identify which contaminants,
if any, are most likely to be causing the effect. When some lines of evidence conflict with
others, professional judgment is needed to determine which data should be considered more
reliable or relevant to the questions.
EXAMPLE BOX 4-1
Lines of Evidence-Copper Site
Primary question: Are ambient copper levels in sediments in the pond causing
adverse effects in benthic organisms1
Possible lines of evidence phrased as test hypotheses:
(1) Docs mortality in early life stages of benthic aquatic insects in contact
with sediments from the site significantly exceed (p < 0.05) mortality
in the same kinds of organisms in contact with sediments from a
reference site?
(2) Does mortality in m situ ioxicity tests in sediments at the pond
significantly exceed (p < 0 05) mortality in in situ toxiciry tests in
sediments at a reference pond"1
i 3) .Are there significantly fewer (p <0.10) numbers of benthic aquatic
insect species present per m* of sediment at the pond near the seep
than at the opposite side of the pond?
Once there is agreement on which lines of evidence are requ'red to answer questions
concerning ihe assessment endpoint, the measurement endpoints by which the questions or
test hypotheses will be tested can be selected.
STEP 4, Page 4
-------
August 21. 1996
DRAFT
HIGHLIGHT BOX 4-2
Terminology and Definitions
In the field of ecotoxicology, there
historically have been multiple definitions for
some terms, including definitions for direct
effects, indirect effects, acute effects, chronic
effects, acute tests, and chronic tests. This
multiplicity of definitions has resulted in
misunderstandings and inaccurate communication
of study designs. {Definitions of these and other
terms, as they are used in this document, are
provided in the glossary. When consulting other
reference materials, the user should evaluate how
the authors are defining terms.
Each measurement endpoint
should represent the same exposure
pathway and toxic mechanism of action
as the assessment endpoint it
represents; otherwise, irrelevant
exposure pathways or toxic
mechanisms might be evaJuated. For
example, if a contaminant primarily
causes damage to vertebrate kidneys,
the use of daphnids (which do not have
kidneys) would be inappropriate.
Potential measurement
endpoints in toxicity tests or in field
studies should be evaluated according
to how well they can answer questions
about the assessment endpoint or
support or refute the hypotheses
developed for the conceptual model. Statistical Considerations, including sample size and
statistical power described in Section 4.3, also must be considered in selecting the
measurement endpoints. The following subsections describe additional considerations for
selecting measurement endpoints, including species/community/habitat (Section 4.1.1),
relationship to the contaminant(s) of concern (Section 4.1.2), and mechanisms of ecotoxicity
(Section 4.1.3).
4.1.1 Species/Community/Habitat Considerations
The function of a measurement endpoint is to represent an assessment endpoint for the
site. The measurement endpoint must allow clear inferences about potential changes in the
assessment endpoint. Whenever assessment and measurement endpoints are not the same
(which usually is the case), measurement endpoints should be selected to be inclusive of risks
to all of the species, populations, or groups included in the assessment endpoint that are not
directly measured. In other words, the measurement endpoir? should be representative of the
assessment endpoint for the site and not lead to an underestimate of risk to the assessment
endpoint. Example Box 4-2 illustrates this point for the DDT site in Appendix A.
In selecting a measurement endpoint, the species and life stage, population, or
community chosen should be the one(s) most susceptible to the contaminant for the
assessment endpoint in question. For species and populations, this selection is based on a
review of the species: (1) life history; (2) habitat utilization; (3) behavioral characteristics;
and (4) physiological parameters. Selection of measurement endpoints also should be based
STEP 4, Page 5
-------
August :;. ;996 DRAFT
EXAMPLE BOX 4-2
Selecting Measurement Endpoints-DDT Site
As described in Example Box 3-1. one of the assessment endpoints selected for the
DDT site is the protection of piscivorous birds. The belted kingfisher was selected as a
piscivorous bird with the smallest home range that couid utilize the area of the site, thereby
maxirmzing the calculated dose to a receptor. In this illustration, the kingfishers are used as the
most highly exposed of the piscivorous birds potentially present. Thus, one can conclude that,
if the nsk assessment shows no threat of eggshell thinning to the kingfisher, there should be
minimal or no threat to other piscivorous birds that might utilize the site. Thus, eggshell
thinning in belted kingfishers is an appropriate measurement endpoint for this site.
on which routes of exposure are likely. For communities, careful evaluation of the
contaminant fate and transport in the environment is essential.
4.1.2 Relationship of the Measurement Endpoints to the Contaminant of
Concern
Additional criteria to consider when selecting measurement endpoints are inherent
properties (such as the physiology or behavioral characteristics of the species) or life history
parameters that make a species useful in evaluating the effects of site-specific contaminants.
For example, Chironomus lemans (a species of midge that is used as a standard sediment
toxicity testing species in the larval stage) is considered more tolerant of metals contamination
than is C. npanus. a similar species (Klemm et al., 1990; Nebeker et al., 1984; Pascoe et al.,
1989). To assess the effects of exposure of benthic communities to metal-contaminated
sediment, C. npanus might be the better species to use as a toxicity test organism for many
aquatic systems to ensure that risks are not underestimated. In general, the most sensitive of
the measurement endpoints appropriate for infemng risks to the assessment endpoint should
be used-
Some species have been identified as being particularly sensitive to certain
contaminants For example, numerous studies have demonstrated that mink are among the
most sensitive of the tested mammalian species to the toxic effects of PCBs (U.S. EPA,
1995a). Species that rely on quick reactions or behavioral responses to avoid predators can
be particularly sensitive to contaminants affecting the central nervous system, such as
mercury. Thus, the sensitivity of the measurement endpoint relative to the assessment
endpoint should be considered for each contaminant of concern.
STEP 4, Page 6
-------
August 21. 1996 DRAFT
4.1.3 Mechanisms of Toxicity
A contaminant can exert adverse ecologicai effects in many ways. First, a
contaminant might affect an organism after exposure for a short penod of time (acute) or after
exposure over an extended penod of time (chronic). Second, the effect of a contaminant
could be lethal (killing the organism) or sublethaJ (causing adverse effects other than death,
such as reduced growth, behavioral changes, etc.). Sublethal effects can reduce an organism's
lifespan or reproductive success. For example, if a contaminant reduces the reaction speed of
a prey species, the prey can become more susceptible to predauon. Third, a contaminant
might act directly or indirectly on an organism. Direct effects include lethal or sublethaJ
effects of the chemical on the organism. Indirect effects occur when the contaminant
damages the food, habitat, predator-prey, or competition of the organism in its community.
Mechanisms of ecotoxicity and exposure pathways have already been considered
during problem formulation and identification of the assessment endpoints. However, toxicity
issues are revisited when selecting appropriate measurement endpoints to ensure that the same
toxic response that is of concern for the assessment endpoint is measured.
4.2 STUDY DESIGN
In Section 4.1, one or more lines of evidence that could be used to answer questions
or to test hypotheses concerning the assessment endpoint(s) were identified. This section
provides recommendations on how to design a field study for bioaccumulation and field
tissue residue studies (Section 4.2.1); population/community evaluations (Section 4.2.2); and
toxicity testing (Section 4.2.3). A thorough understanding of the strengths and limitations of
these types of field studies is necessary to properly design any investigation.
Typically, no one line of evidence can stand on its own. Analytic chemistry on co-
located samples and other lines of evidence are needed to support a conclusion. When
population/community evaluations are coupled with toxicity testing and media chemistry, the
procedure often is referred to as a triad approach (Chapman et ah, 1992; Long and Chapman,
1985). This method has proven effective in defining the area affected by contaminants in
sediments of several large bays and estuaries.
The development of exposure-response relationships is critical for evaluating nsk
management options; thus, for all three types of studies, sampling is applied to a
contamination gradient when possible as well as compared to reference data. Reference data
are baseline values or characteristics that should be representative of the sue in the absence of
contaminants released from the site. Reference data might be data collected from the site
before contamination occurred or new data collected from a reference site. The reference site
can be the least impacted (or unimpacted) area of the Superfund site or a nearby site that is
STEP 4, Page 7
-------
August 21. 1996 DRAFT
ecologically similar, but noc affected by the site's contaminants. For additional information
on selecune and usins reference information in Superfund ecological nsk assessments, see
ECO Update Volume 2. Number 1 (U.S. EPA, 1994e).
The following subsections present a starting point for selecting an appropriate study
design for the different rypes of biological sampling that might apply to the sue investigation.
4.2.1 Bloaccumulation and Field Tissue Residue Studies
Bioaccumulation and field tissue residue studies typically are conducted at sites where
contaminants are likely to accumulate in food chains. The studies help to evaluate
contaminant exposure levels associated with measures of effect for assessment endpoint
species.
The degree to which a contaminant is transferred through a food chain can be
evaluated in several ways. The most common cype of study reported in the Literature is a
contaminant bioaccumulation (uptake) study. As indicated in Section 2.2.1, the most
conservative BCF values identified in the literature almost always are used to estimate
bioaccumulation in a screening-level nsk assessment. Where the potential for overestimating
bioaccumulation by using conservative literature values to represent the site is substantial, a
site-specific tissue residue study might be advisable (see Example Box 4-3).
A tissue residue study generally is conducted on organisms that are in the exposure
pathway (i.e., food chain) associated with the assessment endpoint. Data seldom are available
EXAMPLE BOX 4-3
Bioaccumulation-ODT Site
Data from the literature suggest that DDT can bioaccumulate in aquatic food
chains as mucn as six orders of magnitude ' 10 ); however, in many systems, the actual
bioaccumulation of DDT from the environment is substantially lower than 106. Several
factors influence the actual accumulation of DDT in the environment. Because there is
considerable debate over the parameters of proposed theoretical bioaccumulation
models, it might be advisable to conduct a site-specific study.
to link tissue residue levels in the sampled organisms to adverse effects. Instead, literature
toxiciry studies usually are used to associate effects with an administered dose (or data that
can be converted to an administered dose) from a laboratory study. Thus, the purpose of a
field tissue residue study usually is to measure contaminant concentrations in foods consumed
STEP 4, Page
-------
August 21. 1996 DRAFT
by the species associated with the assessment endpoint. This measurement minimizes the
uncertainty associated with estimating a dose (or intake) to that species, particularly in
situations in which several media and trophjc levels are in the exposure pathway.
The concentration of a contaminant in the primary prey/food also should be linked to
an exposure concentration from a contaminated medium (e.g., soil, sediment, water), because
it is the medium, not the food chain, that will be remediated. Thus, contaminant
concentrations must be measured in environmental media at the same locations at which the
organisms are collected along contaminant gradients and at reference locations. Co-located
samples of the contaminated medium and organisms are needed to establish a correlation
between the tissue residue levels and contamination levels in the medium under evaluation;
these studies are most effective if conducted over a gradient of contaminant concentrations.
In addition, tissue residues from sessile organisms (e.g., rooted plants, clams) are easier to
attribute lt> specific contaminated areas than are tissue residues from mobile organisms (e.g.,
fish). Example Box 4-4 illustrates these concepts using the DDT site example in
Appendix A.
EXAMPLE BOX 4-4
Tissue Residue Studies-ODT Site
In the DDT site example, a forage fish (e.g., creek chub) will be collected at
several locations with known DDT concentrations in sediments. The forage fish will be
analyzed for body burdens of DDT, and the relationship between the DDT levels in the
sediments and the levels in the forage fish will be established. The forage fish DDT
concentrations can be used to evaluate the DDT threat to piscivorous birds feeding on
the forage fish at each location. Using the DDT concentrations measured in fish that
correspond to a LOAEL and NOAEL for adverse effects in birds, the corresponding
sediment contamination levels can be determined. These sediment DDT levels can then
be used to estimate a cleanup level that would reduce threats of eggshell thinning to
piscivorous birds.
Although it might seem obvious, it is important to confirm that the organisms
examined for tissue residue levels are in the exposure pathways of concern established by the
conceptual model. Food items targeted for collection should be those that are likely to
constitute a large portion of the diet of the species of concern (e.g., new growth on maple
trees as a food source for deer, rather than cattails) and/or represent pathways of maximum
exposure. If not, erroneous conclusions or study delays and added costs can result. Because
specific organisms often can only be captured in one season, the timing of the study can be
critical, and failure to plan accordingly can result in serious site management difficulties.
STEP 4, Page 9
-------
August 21. '^96 DRAFT
There are numerous factors that must be considered when selecting a species in which
to measure contaminant residue levels. Several investigators have discussed the "ideal"
characteristics of the species to be collected and analyzed. The recommendations by Phillips
(1977. 19"8) and Butler (1971) include that the species selected should be:
(1) Able to accumulate the chemical of concern without being adversely affected
by the levels encountered at the site;
(2) Sedentary (small home range) in order to be representative of the area of
collection;
(3) Abundant in the study area; and
(4) Of reasonable size to give adequate tissue for analysis (e.g., 10 grams for
organic analysis and 0.5 gram for metal analysis for many laboratories).
Additional considerations for some situations would be that the species is:
(5) Sufficiently long-lived to allow for sampling more than one age class; and
(6) Easy to sample and hardy enough to survive in the laboratory (allowing for the
organisms to eliminate some contaminants from their bodies prior to analysis, if
desired, and allowing for laboratory studies on the uptake of the contaminant).
It is usually not possible or necessary to find an organism that fulfills all of the above
requirements. The selection of an organism for tissue analysis should balance these
characteristics with the hypotheses being tested, knowledge of the contaminants' fate and
transport, and the practicality of using the particular species. In the following sections,
several of the factors mentioned above are described in greater detail.
Ability to accumulate the contaminant. The objecu\es of a tissue residue study
are (1) to measure bioavailabiliry directly; (2) to provide site-specific estimates of exposure to
higher-trophic-level organisms; and (3) to relate tissue residue leve.s to concentrations in
environmental media (e.g., in soil, sediment, or water). Sometimes ihese studies cdso can be
used to link tissue residue levels with observed effects in the organisms sampled. However,
in a "pure" accumulation study, the species selected for collection and tissue analysis should
be ones that can accumulate a contaminant(s) without being adversely affected by the levels
encountered in the environment. While it is difficult to evaluate whether or not a population
in the field is affected by accumulation of a contaminant, it is important to try. Exposure that
results in adverse responses might alter the animal's feeding rates or efficiency, diet, degree
of activity, or metabolic rate, and thereby influence the animal's daily intake or accumulation
of the contaminant and the estimated bioaccumulation factor (BAF). For example, if the rate
STEP 4, Page 10
-------
August :i. 1996 DRAFT
of bioaccumuJation of a contaminant in an organism decreases with increasing environmental
concentrations (e.g., its toxic effects reduce food consumption rates), using a BAF determined
at low environmental concentrations to estimate bioaccumulauon at high environmental
concentrations would overestimate risk. Conversely, if bioaccumulation increased with
increasing environmental concentrations (e.g., its toxic effects impair the organisms' ability to
excrete the contaminant), using a BAF determined at low environmental concentrations would
underestimate risks at higher environmental concentrations.
Consideration of the physiology and biochemistry of the species selected for residue
analysis also is important. Some species can metabolize certain organic contaminants) (e.g.,
fish can metabolize PAHs). If several different types of prey are consumed by a species of
concern, it would be more appropriate to analyze prey species that do not metabolize the
contaminant.
Home range. When selecting species for residue analyses, one should be confident
that the contaminant levels found in the organism depend on the contaminant levels in the
environmental media under evaluation. Otherwise, valid conclusions cannot be drawn about
ecological risks posed by contaminants at the site. The home range, particularly the foraging
areas within the home range, and movement patterns of a species are important in making this
determination. Organisms do not utilize the environment uniformly. For species that have
large home ranges or are migratory, it can be difficult to evaluate potential exposure to
contaminants at the site. Attribution of contaminant levels in an organism to contaminant
levels in the surrounding environment is easiest for animals with small home and foraging
ranges and limited movement patterns. Examples of organisms with small home ranges
include young-of-the-year fish, burrowing Crustacea (such as fiddler crabs or some crayfish),
and small mammals.
Species also should be selected for residue analysis to maximize the overlap between
the area of contamination and the species1 home range or feeding range. This provides a
conservative evaluation of potential exposure levels. The possibility that a species' preferred
foraging areas within a home range overlap the areas of maximum contamination also should
be considered.
Population size. A species selected for tissue residue analysis should be sufficiently
abundant at the site that adequate numbers (and sizes) of individuals can be collected to
support the tissue mass requirements for chemical analysis and to achieve the sample size
needed for statistical comparisons. The organisms actually collected should be not only of
the same species, but also of similar age or size to reduce data variability when BCFs are
being evaluated. The practicality of using a particular species is evaluated in Step 5.
Size/composites. When selecting species in which to measure tissue residue levels,
it is optimum to have animals large enough for individual chemical analysis, without having
STEP 4, Page 11
-------
August :i. 1996 DRAFT
to pool individuals. However, composite samples will be needed if individuals from the
species selected cannot yield sufficient tissue for the required analytical methods. Linking
contaminant levels in organisms to concentrations in enviror mental media is easier if
composites are made up of members of the same species, sex, size, and age, and therefore
exhibit similar accumulation characteristics. When deciding whether or not to pool samples,
it is important to consider what impact the loss of information on variability of contaminant
levels along these dimensions will have on data interpretation. The size, age, and sex of the
species collected should be representative of the range of prey consumed by the species of
concern.
Summary. Although it can be difficult to meet all of the suggested cntena for
selecting a species for tissue residue studies, an attempt should be made to meet as many
criteria as possible. No formula is available for ranking the factors in order of importance
within a particular site investigation because the ranking depends on the study objectives.
However, a key criterion is that the organism be sedentary or have a limited home range. It
is difficult to connect site contamination to organisms that migrate over great distances or that
have extremely large home ranges. Further information on factors that can influence
bioaccumulation is available from the literature (e.g., Butler, 1971; Phillips, 1977, 1978, U.S.
EPA, 1995d).
4.2.2 Population/Community Evaluations
Population/community evaluations, or biological field surveys, are potentially useful
for both contaminants that are toxic to organisms through direct exposure to the contaminated
medium and contaminants that bioaccumulate in food chains. In either case, careful
consideration must be given to the mechanism of contaminant effects. Since
population/community evaluations are "impact' evaluations, they typically are not predictive.
The release of the contaminant must already have occurred and exerted an effect in order for
the populationycommumry evaluation to be an effective tool for a risk assessment.
Population and community surveys evaluate the current status of an ecosystem, often
using several measures of population or community structure (e.g., standing biomass, species
richness) or function (e.g., feeding group analysis). The most commonly used measures
include number of species and abundance of organisms in an ecosystem. It is difficult to
detect changes in top predator populations affected by bioaccumulation of substances in their
food chain due to the mobility of top predators. In addition, some populations, most notably
insects, can develop a tolerance to contaminants (particularly pesticides); in these cases, a
populauonycommumry survey would be ineffective for evaluating existing impacts. While
population,community evaluations can be useful, the nsk assessor snould consider the level of
effort required as well as the difficulty in accounting for natural variability.
STEP 4. Page 12
-------
August 21. 1996 DRAFT
A variety of population/community evaluations have been used at Superfund sites.
Benthic macroinvertebrate surveys are the most commonly conducted population/community
evaluations. There are methods manuals (e.g., U.S. EPA 1989c, 1990a) and publications.that
describe the technical procedures for conducting these studies. In certain instances, fish
community evaluations have proven useful at Superfund sites. However, these investigations
typically are more labor-intensive and costly than a comparable macroinvertebrate study. In
addition, fish generally are not sensitive measures of the effects of sediment contamination,
because they usually are more mobile than benthic macroinvertebrates. Terrestrial plant
community evaluations have been used to a limited extent at Superfund sites. For those
surveys, it is important to include information about historical land use and physical habitat
disruption in the uncertainty analysis.
Additional information on designing field studies and on field study methods can be
found in ECO Update Volume 2, Number 3 (U.S. EPA, 1994d).
Although population- and community-level studies can be valuable, several factors can
confound the interpretation of the results. For example, many fish and small mammal
populations normally cycle in relation to population density, food availability, and other
factors. Vole populations have been known to reach thousands of individuals per acre and
then to decline to as low as tens of individuals per acre the following years without an
identifiable external stressor (Geller, 1979). It is important that the "noise of the system" be
evaluated so that the impacts attributed to chemical contamination at the site are not actually
the result of different, "natural" factors. Populations located relatively close to each other can
be affected independently: one might undergo a crash, while another is peaking. Physical
characteristics of a site can isolate populations so that one population level is not a good
indicator of another; for example, a paved highway can be as effective a barrier as a river,
and populations on either side can fluctuate independently. Failure to evaluate these issues
can result in erroneous conclusions. The level of effort required to resolve some of these
issues can make population/community evaluations impractical in some circumstances.
4.2.3 Toxicity Testing
The bioavailability and toxicity of site contaminants can be tested directly with
toxicity tests. As with other methods, it is critical that the media tested are in exposure
pathways relevant to the assessment endpoint. If the site conceptual model involves exposure
of benthic invertebrates to contaminated sediments, then a solid-phase toxicity test using
contaminated sediments (as opposed to a water-column exposure test) and an infaunal species
would be appropriate. As indicated earlier, the species tested and the responses measured
must be compatible with the mechanism of toxjcity. Some common site contaminants are not
toxic to most organisms at the same environmental concentrations that threaten top predators
because the contaminant biomagnifies in food chains (e.g., PCBs); toxicity tests using
STEP 4, Page 13
-------
August 21. 1996 DRAFT
contaminated media from the site would not be appropriate for evaluating this type of
ecological threat.
There are numerous U.S. EPA methods manuals and ASTM guides and procedures for
conducting to.xicity tests (see references in the Bibliography). While documented methods
exist for a wide variety of toxicity tests, particularly laboratory tests, the risk assessor must
evaluate what a particular toxicity test measures and, just as importantly, what it does not
measure. Questions to consider when selecting an appropriate toxicity test include:
(1) What is the mechanism of toxicity of the contammant(s)?
(2) What contaminated media are being evaluated (water, soil, sediment)?
(3) What toxicity test species are available to test the media being evaJuated?
(4) What life stage of the species should be tested?
(5) What should the duration of the toxicity test be0
(6) Should the test organisms be fed dunng the test0
(7) What endpoints should be measured?
There are a limited number of toxicity tests that are readily available for testing
environmental media. Many of the aquatic toxicity tests were developed for the regulation of
aqueous discharges to surface waters. These tests are useful, but one must consider the
original purpose of the test
New toxicity tests are being developed continually and can be of value in designing a
site ecological nsk assessment. However, when non-standard tests are used, complete
documentation of the specific test procedures is necessary to support use of the data.
In situ toxicity tests involve placing organisms in locations that might be affected by
site contaminants and in reference locations. Non-native species should not be used, because
of the nsk of their release into the environment in which they could adversely affect (e.g.,
prey on or outcompete) resident species. In situ tests might provide more realistic evidence
of existing adverse effects than laboratory toxicity tests; however, me investigator has little
control over many environmental parameters and the experimental organisms can be lost to
adverse weather or other events (e.g., human interference) at the site or reference location.
For additional information on using toxicity tests in ecological risk assessments, see
ECO Update Volume 2, Numbers 1 and 2 (U.S. EPA, I994b.c).
STEP 4, Page 14
-------
August 21. 1996 ^ DRAFT
4.3 DATA QUALITY OBJECTIVES AND STATISTICAL CONSIDERATIONS
The SAP indicates the number and location of samples to be taken, the number of
replicates for each sampling location, and the method for determining sampling locations. In
making these decisions, the investigator needs to consider, among other things, the DQOs and
statistical methods that will be used to analyze the data.
4.3.1 Data Quality Objectives
The DQO process represents a series of planning steps that can be employed
throughout the development of the WP and SAP to ensure that the type, quantity, and quality
of environmental data to be collected during the ecological investigation are adequate to
support the intended application. Problem formulation in Steps 3 and 4 is essentially the
DQO process. By employing problem formulation and the DQO process, the investigator is
able to define data requirements and error levels that are acceptable for the investigation prior
to the collection of data. This approach helps ensure that results are appropriate and
defensible for decision making. The specific goals of the general DQO process are to:
Clarify the study objective and define the most appropriate types of data to
collect;
Determine the most appropriate field conditions under which to collect the data;
and
Specify acceptable levels of decision errors that will be used as the basis for
establishing the quantity and quality of data needed to support risk management
decisions.
As the discussion of Steps 3 and 4 indicates, these goals are subsumed in the problem
formulation phase of an ecological risk assessment Several U.S. EPA publications provide
detailed descriptions of the DQO process (U.S. EPA, 1993c,d,f, I994f). Because many of the
steps of the DQO process are already covered during problem formulation, the DQO process
should be reviewed by the investigator and applied as needed.
4.3.2 Statistical Considerations
Sampling locations can be selected "randomly" to characterize an area or non-
randomly, as along a contaminant concentration gradient. The way in which sampling
locations are selected determines which statistical tests, if any, are appropriate for evaluating
test hypotheses.
STEP 4, Page 15
-------
August :i. !996 DRAFT
For toxicity tests that use a small number of test and control organisms or for which
the toxic response in highly variable, the response rate of test animals often must be relatively
high (e.2.. 30 to 50 percent) for the response to be considered a LOAEL (i.e., statistically
different' than control levels). If a LOAEL based on a 30 to 50 percent effect level is
unacceptable (e.g., a population is unlikely to sustain itself with an additional 30 to 50
percent mortality), then the power of the study design must be increased, usually by ,
increasing sample size, but sometimes by taking full advantage of all available information to
improve the power of the design (e.g., stratified sampling, special tests for trends, etc.). A
limitation on the use of toxicity values from the literature is that often, the investigator does
not discuss the statistical power of the study design, and hence does not indicate the
minimum statistically detectable effect level. Appendix D describes additional statistical
considerations, including a description of Type I and Type n error, statistical power,
statistical models, and power efficiency.
4.4 CONTENTS OF WORK PLAN AND SAMPLING AND ANALYSIS PLAN
The WP and SAP for the ecological investigation should be developed as part of the
initial RI sampling event if possible. If not, the WP and SAP can be developed as an
additional phase of the site investigation. In either case, the format of the WP and SAP
should be similar to that described by U.S. EPA (1988a, 1989b). Accordingly, these
documents should be consulted when developing the ecological investigation WP and SAP.
The WP and SAP are typically written as separate documents. When developed as
separate documents, the WP can be submitted to the risk manager for review prior to the
development of the SAP and any differences in approach can be resolved prior to the
development of the SAP For some smaller sites, however, it might be more practical to
combine the WP and SAP into a single document. If the WP and SAP are combined, the
investigators should discuss the overall objectives and approach of the WP/SAP document
with the risk manager pnor to its development to ensure agreement from all parties about the
approach being taken for the investigation.
The WP and SAP are briefly described in Sections 4.4.1 and 4.4.2, respectively. A
plan for testing the SAP before the sue WP and SAP are signed and the investigation begins
is described in Section 4.4.3.
4.4.1 Work Plan
The purpose of the WP is to document the decisions and evaluations made during
problem formulation and to identify additional investigative tasks needed to complete the
evaluation of risks to ecological resources. As presented in U.S. EPA (1988a), the WP
generally includes the following:
STEP 4, Page 16
-------
August 21. 1996 DRAFT
A general overview and background of the site including the site's physical
setting, ecology, and previous uses;
A summary and analysis of previous site investigations and conclusions;
A site conceptual model, including an identification of the potential exposure
pathways selected for analysis, the assessment endpomts and questions or
testable hypotheses, and the measurement endpoints selected for analysis;
The identification of additional site investigations needed to conduct the
ecological risk assessment; and
A description of assumptions used and the major sources of uncertainty in the
site conceptual model and existing information.
The general scope of the additional sampling activities also is presented in the WP. A
detailed description of the additional sampling activities is presented in the SAP along with an
anticipated schedule of the site activities.
4.4.2 Sampling and Analysis Plan
The SAP typically consists of two components: a field sampling plan (FSP) and a
quality assurance project plan (QAPP). The FSP provides guidance for all field work by
providing a detailed description of the sampling and data-gathering procedures to be used for
the project. Meanwhile, the QAPP provides a description of the steps required to achieve the
objectives dictated by the intended use of the data.
Field sampling plan. The FSP provides a detailed description of the samples
needed to meet the objectives and scope of the investigation outlined in the WP. The FSP for
the ecological assessment should be detailed enough that a sampling team unfamiliar with the
site would be able to gather all the samples and/or required field data based on the guidelines
presented in the document The FSP for the ecological investigation should include a
description of the following elements:
Sampling type and objectives;
Sampling location, timing, and frequency;
Sample designation;
Sampling equipment and procedures; and
Sample handling and analysis.
A detailed description of these elements for chemical analyses is provided in Appendix B of
U.S. EPA (1988a). Similar specifications should be developed for the biological sampling.
STEP 4, Page 17
-------
August I!. 1996
DRAFT
Quality assurance project plan. The objective of the quality assurance project
plan (QAPP^ is to provide a description of the policy, organization, functional activities, and
quahtv control protocols necessary for achieving the study objectives. Highlight Box 4-3
presents the elements typically contained in a QAPP.
L.S. EPA has prepared guidance on
the contents of a QAPP (U.S. EPA, 1987a,
1988a, 1989a). Formal quality assurance
and quality control (QA/QC) procedures
exist for some types of ecological
assessments, for example, for laboratory
toxicity tests on aquatic species. For
standardized laboratory tests, there are
formal QA/QC procedures that specify (1)
sampling and handling of hazardous wastes;
(2) sources and cultunng of test organisms;
(3) use of reference toxicants, controls, and
exposure replicates; (4) instrument
calibration; (5) record keeping; and (6) data
evaluation. For other types of ecological
assessments, however, QA/QC procedures
are less well defined (e.g., for biosurveys of
vegetation, terrestrial vertebrates). BTAG
members can provide-input on appropriate
QA/QC procedures based on their experience
HIGHLIGHT BOX 4-3
Elements of a QAPP
(1) Project description
(2) Designation of QA/QC
responsibilities
(3) Statistical tests arid data quality
objectives
(4) Sample collection and chain of
custody
(5) Sample analysis
(6) System controls and preventive
maintenance
(7) Record keeping
(8) Audits
(9) Corrective actions
(10) Quality control reports
with Superfund sites.
4.4.3 Field Verification of Sampling Plan and Contingency Plans
For biological sampling, uncontrolled variables can influence the availability of species
to be sampled, the efficiency of different types of sampling techniques, and the level of effort
required to achieve the sample sizes specified in the SAP. As a consequence, the nsk
assessor should develop a plan to test the sampling design before ±e WP and SAP arc signed
and the sue investigation begins. Otherwise, field sampling dunng the site investigation could
fail to meet the DQOs specified in the SAP. and the study cou'd fa:! :o meer its objectives.
Step 5 provides a description of the field verification process.
To the extent that potential field problems can be anticipated, contingency plans also
should be specified in the SAP An example of a contingency plan is provided-in Step 5
(Example Box 5- i i
STEP 4. Page 18
-------
August 21. 1996 ^ DRAFT
4.5 SCIENTIFIC/MANAGEMENT DECISION POINT (SMDP)
The completion of the ecological nsk assessment WP and SAP should coincide with
an SMDP Within this SMDP, the ecological nsk assessor and the ecological nsk manager
agree on: 11) selection of measurement endpomts; (2) selection of specific investigation
methodology; and (3) selection of data reduction and interpretation methods. The WP or SAP
also should specify how inferences will be drawn from the measurement to the assessment
endpomts.
4.6 SUMMARY
At the conclusion of Step 4, there will be an agreement on the contents of the WP and
SAP. AS noted earlier, these plans can be parts of a larger WP and SAP that are developed
to meet other remedial investigation needs, or they can be separate documents. When
possible, any field sampling efforts for the ecological risk assessment should overlap with
other site data collection efforts to reduce sampling costs and to prevent redundant sampling.
The WP and/or the SAP should specify the methods by which the collected data will
be analyzed. The plan(s) should include all food-chain-exposure-model parameters, data
reduction techniques, data interpretation methods, and statistical analyses.
STEP 4, Page 19
-------
August :i. 1996 DRAFT
STEP 5: FIELD VERIFICATION OF SAMPLING DESIGN
OVERVIEW
Before the WP and SAP is signed, it is important to verify that the field
sampling plan as specified in the WP and SAP is appropriate and implementable at
the site. During the field verification of the sampling plan, the testable hypotheses,
exposure pathway models, and measurement endpoints are evaluated for their
appropriateness and implementabiliry. The assessment endpoint(s), however, should
not be under evaluation in this step; the appropriateness of the assessment endpoint
should have been resolved in Step 3. If the assessment endpoint is changed at this
step, the risk assessor must return to Step 3, because the entire process leading to the
actual site investigation in Step 6 assumes the selection of the correct assessment
endpoints.
5.1 PURPOSE
The primary purpose of field verification of the sampling plan is to ensure that the
samples specified by the SAP actually can be collected. A species that will be associated
with a measurement endpoint and/or exposure point concentration should have been observed
at the preliminary site characterization or noted during previous site visits. During this step,
previously obtained information should be verified and the feasibility of sampling will need to
be checked by a site visit. Preliminary sampling will determine if the targeted species is
present andequally importantcollectable in sufficient numbers or total biomass to meet
data quality objectives. This preliminary field assessment also allows for final confirmation
of the habitats that exist on or near the site. Habitat maps are verified a final time, and
interpretations of aenal photographs can be checked.
Final decisions on reference areas also should be made in this step. The reference
areas should be chosen to isolate a particular variable at the site (e.g., chemical
contamination, rock cobble stream). Parameters to be evaluated for similarity include, but are
not limited to: slope, habitat, species potentially present, soil and sediment characteristics,
and for surface waters, flow rates, substrate type, water depth, temperature, turbidity, oxygen
levels, water hardness, pH, and other standard water quality parameters. If several on-site
habitats or habitat variables are being investigated, then several reference areas could be
required. Reference areas should be as free of site-related contaminants as practical.
STEP 5, Page 1
-------
August :i. !996 DRAFT
5.2 DETERMINING SAMPLING FEASIBILITY
\Vben sampling biota, it is difficult to predict what level of effort will be necessary to
obtain an adequate number of individuals of the required size. Some preliminary field
measurements often can help determine adequate sampling efforts to attain the sample sizes
specified in the SAP for statistical analyses. The WP and SAP should be signed and the site
investigation should be implemented immediately after verification of the sampling design to
limit effects of uncontrolled field variables. For example, evaluation of current small
mammal population density might indicate to the investigator that 400 trap-nights instead of
50 are necessary to collect the required number of small mammals in the specified time. If
there is a time lag between the field sampling venficauon and the actual site investigation, it
could be necessary to revenfy the field sampling to determine if conditions have changed.
Sarnpling methods for abiotic media also should be tested. There is a wide variety of
sampling devices and methods, and it is important to use the most appropriate, as the
following examples illustrate:
When sampling a stream's surface water, if the stream is only three inches
deep, collecting the water directly into 32-ounce bottles would not be practical.
Sampling the substrate in a stream might be desirable, but if the substrate is
bedrock, it might not be feasible or the intent of the sampling design.
An expo sure-response relationship between contamination and biota response is a key
component of establishing causality during the analysis phase of the baseline risk assessment
(Step 6). If extentof-contarmnation sampling is conducted in phases, abiotic exposure media
and biotic samples must be collected simultaneously because the interactions (both temporal
and spatial) between the matrix to be remediated and the biota are crucial to the development
of a field exposure-response relationship. Failure to collect one sample properly or to
coordinate samples temporally can significantly impact the interpretation of the data.
These and other problems associated with the practical implementation of sampling
should be resolved pnor to finalizing the SAP to the extent practicable. Assessing the
feasibility of the sampling plan before the site investigation begins saves costs in the long
term because it minimizes the chances of failing to meet data quality objectives during the
sue investigation.
Sampling locations need to be checked to make sure that they are appropriately
described and placed within the context of the sampling plan. Directions for a sediment
sample "to be laken 5 feet from the north side of stream A," could cause confusion if the
stream is only 4 feet wide, or if the sampler doesn't know if the sample should be taken in
STEP 5. Page 2
-------
August 21. 1996 DRAFT
the stream, or 5 feet away from the edge of the stream. All samples should be checked
against the intended use of the data to be obtained.
Contaminant migration pathways might have changed, either due to natural causes
(e.g., storms) or during site remediation activities (e.g., erosion channels might have been
filled or dug up to prevent further migration of contaminants). Channels of small or large
streams, brooks, or nvers might have moved; sites might have been flooded. All of the
assumptions of the migration and exposure pathways need to be venfied prior to the full site
investigation. If a contaminant gradient is necessary for the sampling plan, it is important to
verify that the gradient exists and that the range of contaminant concentrations is appropriate.
A gradient of contamination that causes no impacts at the highest concentration measured has
as little value as a gradient that kills everything at the lowest concentration measured; in
either case, the gradient would not provide useful exposure-response information. A gradient
verification requires chemical sampling, but field screening-level analyses might be effective.
All pathways for the migration of contaminants off site should be evaluated, such as
windblown dust, surface water runoff, and erosion. Along these pathways, a gradient of
decreasing contamination with increasing distance from the site might exist. Site-specific
ecological evaluations and risk assessments can be more useful to nsk managers if gradients
of contamination can be located and evaluated.
Example Boxes 5-1 and 5-2 describe the field verification of the sampling plan for the
copper and DDT sites illustrated in Appendix A. Note that the scope of the field verification
differs for the copper and DDT sites. For the DDT site, a modification to the study design
was necessary. For both sites, the issues were resolved and a sign-off was obtained at the
SMDP for this step.
Any change in measurement endpoints will require that exposure pathways to the new
measurement endpoint be checked. The new measurement endpoint must fit into the
established conceptual model. Changes to measurement endpoints might require revision of
the conceptual model and agreement to the changes at the SMDP It is highly desirable that
the agreed-upon conceptual model should be modified and approved by the same basic group
of individuals who developed it.
5.3 SCIENTIFIC/MANAGEMENT DECISION POINT (SMDP)
The SMDP for the field verification of the sampling plan is the signing of the
finaliied WP and SAP. Any changes to the investigation proposed in Step 4 must be made in
consultation with the nsk manager and the risk assessors. The nsk manager must understand
what changes have been made and why, and must ensure that the nsk management decisions
can be made from the information that the new study design can provide. The risk assessors
STEP 5, Page 3
-------
Ausrust 21. 1996 DRAFT
EXAMPLE BOX 5-1
Field Verification of Sampling Plan-Copper Site
Copper was released from a seep area of a landfill adjacent to a small pond: the
release and resulting elevated copper levels in the pond are of concern. The problem
formulation and conceptual model stated that the assessment endpomt was the maintenance
of a tvpicaJ pond community for the area, including the benthic invertebrates and fish.
Toxicity testing was selected to evaluate the potential toxicity of copper to aquatic
organisms. Three toxicity tests were selected: a 10-day solid-phase sediment toxicity test
(with the amphipod Hyalella azteca), and two water column tests (i.e., the 7-day growth
test wuh the green aJga Selenasirum capricornutum and the fathead minnow, Pimephales
promelas, 1-day larval growth test). The study design specified that sediment and water
for the toxicity tests would be collected at the leachate seeps known to be at the pond
edge, and at three additional equidistant locations transecting the pond (including the point
of maximum pond depth). The pond contains water year-round; however, the seep flow
depends on rainfall. Therefore, it is only necessary to verify that the leachate seep will be
active at the time of sampling.
must be involved to ensure that the assessment endpoints and testable hypotheses are still
being addressed.
In the worst cases, changes in the measurement endpoints could be necessary, with
corresponding changes to the risk hypotheses and sampling design. Any new measurement
endpoints must be evaluated according to their utility for inferring changes in the assessment
endpoints and their compatibility with the sue conceptual model (from Steps 3 and 4). Loss
of the relationship between measurement endpoints and the assessment endpoints, the
questions or :estable hypothesis, and the site conceptual model will result in a failure to meet
study objectives.
Despite one's best efforts to conduct a sound site assessment, unexpected
circumstances might still make it necessary for changes of the sampling design to be adopted
in the field during the sue investigation stage. In these instances, the changes should be
wntten and initialled by those agreeing to the change in consultation with the nsk assessor
and nsk manager
Once the finalized WP and SAP art approved and signed. Step 6 should begin.
STEP 5, Page 4
-------
August 21. 1996 DRAFT
EXAMPLE BOX 5-2
Field Verification of Sampling Plan-DDT Site
For the stream DDT site, the assessment endpoint was protection of piscivorous birds
from adverse reproductive effects. The conceptual model included the exposure pathway of
sediment to forage fish to the kingfisher. The measurement endpoint selected was tissue residue
levels in creek chub, which could be associated with contaminant levels in sediments. Existing
information on the stream contamination indicates that a gradient of contamination exists and
that five specific sampling locations should be sufficient to characterize the gradient to the point
where concentrations are unlikely to have adverse effects. The study design specified that
10 creek chub of the same size and sex be collected at each location. Each chub should be
approximately 20 grams, so that minimum sample mass requirements could be met without
relying on the use of composite samples for analysis. In addition, QA/QC protocol requires that
10 more fish be collected at one of the locations.
In this example, a site assessment is necessary to verify that a sufficient number of
creek chub of the specified size are present to meet the sampling requirements. Stream
conditions must be evaluated to determine what fish sampling technique will work at the
targeted locations. A field assessment was conducted, and several fish collection techniques
were used in order to determine which was the most effective for the site. Collected creek chub
and other fish were examined to determine the size range available and whether the sex of the
individuals could be determined.
The site assessment indicated that the creek chub might not be present in sufficient
numbers to provide the necessary biomass for chemical analyses. Based upon these findings, a
contingency plan was agreed to, which stated that both the creek chub and the longnosed dace
(Rhinichthys cataractae) would be collected. If the creek chub were collected at all locations in
sufficient numbers, then these samples would be analyzed and the dace would be released. If
sufficient creek chub could not be collected but sufficient longnosed dace could, the longnosed
dace would be analyzed and the creek chub released. If neither species could be collected at all
locations m sufficient numbers, then a mix of the two species would be used; however, for any
given sampling location only one species would be used to make the sample. In addition, at
one location, which preferably had high DDT levels in the sediment, sufficient numbers (20
grams) of both species would be collected to allow comparison (and calibration) of the
accumulation between the two species.
5.4 SUMMARY
In summary, field verification of the sampling plan is very important to ensuring that
the data quality objectives of the site investigation can be met. This step verifies that the
selected assessment endpoints, testable hypotheses, exposure pathway model, measurement
endpoints, and study design from Steps 3 and 4 are appropriate and implementable at the site.
STEP 5, Page 5
-------
August I!. 1996 _ DRAFT
By verifying the field sampling plan pnor to conducting the full site investigation, well-
considered alterations can be made to the study design and/or implementation if necessary.
These changes will ensure that the ecological nsk assessment meets the study objectives.
STEP 5, Page 6
-------
August 21. 1996 DRAFT
STEP 6: SITE INVESTIGATION AND ANALYSIS PHASE
OVERVIEW
Information collected during the site investigation is used in the analysis phase
of the,baseline ecological nsk assessment to characterize exposures and ecological
effects. The site investigation includes all of the field sampling and surveys that are
conducted as part of the ecological risk assessment. The site investigation and
analysis of exposure and effects should be,straightforward, following the work plan
(WP) and sampling and analysis plan (SAP) developed in Step 4 and tested in Step 5.
Exposure characterization relies heavily on data from the site investigation and
can involve fate and transport modeling. Much of the information for characterizing
potential ecological effects was gathered from the literature review during problem
formulation, but the site investigation might provide evidence of existing ecological
impacts and additional exposure-response information.
6.1 INTRODUCTION
The site investigation (Section 6.2) and analysis phase (Section 6.3) of the ecological
risk assessment should be straightforward. In Step 4, all issues related to the study design,
sample collection, data quality objectives, and procedures for data reduction and interpretation
should have been identified and resolved. However, as described in Step 5, there arc
circumstances that can anse during a site investigation that could require modifications to the
original study design. If any unforeseen events do require a change to the WP or SAP, all
changes must be agreed upon at the SMDP (Section 6.4). The results of Step 6 are used to
characterize ecological risks in Step 7.
6.2 SITE INVESTIGATION
The WP for the site investigation is based on the site conceptual model and should
specify the assessment endpoints, questions, and testable hypotheses. The SAP for the site
investigation should specify the relationship between measurement and assessment endpoints,
the necessary number, volume, and types of samples to be collected, and the sampling
techniques to be used. The SAP also should specify the data reduction and interpretation
techniques and the DQOs. The feasibility of the sampling design was tested in Step 5.
STEP 6, Page 1
-------
August 21. 1996 DRAFT
Therefore, the site investigation should be a direct implementation of the previously designed
study.
During the sue investigation, it is important to adhere to the DQOs and to any
requirements for co-located sampling. Failure to collect one sample properly or to coordinate
samples temporally can significantly affect interpretation of the data. Changing field
conditions (Section 6.2.1) and new information on the nature and extent of contamination
(Section 6.2.2) can require a change in the SAP
6.2.1 Changing Field Conditions
In instances where unexpected conditions arise in the field that make the collection of
specified samples impractical or not ideal, the ecological risk assessor should reevaluate the
feasibility of the sampling design as described in Step 5. Field efforts should not necessarily
be halted, but decisions to change sampling procedures or design must be agreed to by the
site manager and risk assessor or project-delegated equivalents.
Field modifications to study designs are not uncommon during field investigations.
When the WP and SAP provide a precise conceptual model and study design with specified
data analyses, informed modifications to the SAP can be made to comply with the objectives
of the study. As indicated in Step 4, contingency plans can be included in the original SAP
in anticipation of situations that might arise during the site investigation (see Example Box
6-1).
EXAMPLE BOX 6-1
Fish Sampling Contingency Plan-DDT Site
At the DDT site where creek chub are to be collected for DDT tissue residue analyses.
a contingency plan for the sice invesugauon was developed. An alternate species, the longnosed
dace, was specified with the expectauon that, at one or all locations, ihe creek chub might be
absent at the time of the site uwestigauon (see also Example Box 4-4). These contingencies are
prudent even when the venficaiion of the field sampling design described in Step 5 indicates that
the samples are obtainable.
6.2.2 Unexpected Nature or Extent of Contamination
It is not uncommon for an initial sampling phase of the RI to reveal that
contamination at levels of concern extend beyond areas initially established for characterizing
contamination and ecological effects at the site or that contaminant gradients are much steeper
STEP 6. Page 2
-------
August:!. 1996 DRAFT
than anticipated. If this contingency changes the opportunity for evaluating biological effects
along a contamination gradient, the ecological risk assessor and risk manager need to
determine whether additional sampling (e.g., further downstream from the site) is needed.
Thus, u is important for the ecological risk assessor to track information on the nature and
extent of contamination as RI sampling is conducted.
On occasion, new contaminants are identified dunng an RI. In this case, the risk
assessors and site manager will need to return to Step 1 to screen the new contaminants for
ecological nsk.
Immediate analysis of the data for each type of sampling and communication between
the risk assessors and nsk managers can help ensure that the site investigation is adequate to
achieve the study goals and objectives when field modifications are necessary. If a change to
the WP oY SAP is needed, the risk assessor and risk manager must agree on all changes (the
SMDP in Section 6.4).
6.3 ANALYSIS OF ECOLOGICAL EXPOSURES AND EFFECTS
The analysis phase of the ecological risk assessment consists of the technical
evaluation of data on existing and potential exposures (Section 6.3.1) and ecological effects
(Section 6.3.2) at the site. The-analysis is based on the information collected during Steps 1
through 5 and often includes additional assumptions or models to interpret the data in the
context of the site conceptual model. As illustrated in Exhibit 6-1, analysis of exposure and
effects is performed interactively, with the analysis of one informing the analysis of the other.
This step follows the data interpretation and analysis methods specified in the WP and SAP,
and therefore should be a straightforward process.
In the analysis phase, the site-specific data obtained during the site investigation
replace many of the assumptions that were made for the screening-level analysis in Steps 1
and 2. For the exposure and ecological effects characterizations, the uncertainties ass~riated
with the field measurements and with assumptions where site-specific data are not available
must be documented.
6.3.1 Characterizing Exposures
Exposure can be expressed as the co-occurrence or contact of the stressor with the
ecological components, both in time and space (U.S. EPA, 1992a). Thus, both the stressor
and the ecosystem must be characterized on similar temporal and spauai scales. The result of
the exposure analysis is an exposure profile which quantifies the magnitude and spatial and
temporal patterns of exposure as they relate to the assessment endpoints and questions
STEP 6, Page 3
-------
August I!. .996
DRAFT
EXHIBIT 6-1
Analysis Phase
WSK CHARACTEW2ATUX
PROBLEM FORMULATION
Chartctcnutkxi of Exposure
Cfiaractartzatlon of Ecological Effacta
Stmsaor
Characterization:
Olitrtoutton or
Ptttam of Chanoa
Ecoayatam
Charactartzation;
Btoifc
AbkMic
Evaluation
EtfactaOatt
Ecological
R««pona*
Analyvia
RISK CHARACTERCATTON
O
»
a
c_
7
*
o
5"
09
a
o
-------
August 21. 1996
DRAFT
developed dunng problem formulation. The exposure profile and a description of associated
uncertainties and assumptions serve as input to the risk characterization in Step 7.
Stressor characterization involves determining the stressor's distribution and pattern of
change. The analytic approach for characterizing ecological exposures should have been
established in the WP and SAP on the basis of the site conceptual model. For chemical
stressors at Superfund sites, usually a combination of fate and transport modeling and
sampling data from the site are used to predict the current and likely future nature and extent
of contamination at a site. A variety of modeling approaches are available for estimating
chemical fate and transport, and bioaccumulation; however, a description of the approaches
and their relative strengths and weaknesses is beyond the scope of this guidance. Additional
information on this topic can be found in
multiple governmental and academic
sources; however, these are areas of active
research and require consultation with the
BTAG to identify the most up-to-date
approaches.
When characterizing exposures, the
ecological context of the site established
during problem formulation is analyzed
further, both to understand potential effects
of the ecosystem on fate and transport of
chemicals in the environment and to
evaluate site-specific characteristics of
species or communities of concern. Any
site-specific information that can be used to
replace assumptions based on information
from the literature or from other sites is
incorporated into the description of the
ecological components of the site.
Remaining assumptions and uncertainties in
the exposure model (Highlight Box 6-1) should be documented.
6.3.2 Characterizing Ecological Effects
At this point, all evidence for existing and potential adverse effects on the assessment
endpoints is analyzed. The information from the literature review on ecological effects is
integrated with any evidence of existing impacts based on the site investigation. The methods
for analyzing site-specific data should have been specified in the WP and SAP, and thus
should be straightforward. Both exposure-response information and evidence that site
contaminants are causing or can cause adverse effects are evaluated.
HIGHUGHT BOX 6-1
Uncertainty in Exposure Models
The "accuracy" of an exposure
model depends on the accuracy of the input
parameter values and the validity of the
model's structure (i.e., the degree to which it
represents the actual relationships among
parameters at the site). Field measurements
related to model outputs or intermediate
model results can be used to help calibrate
and validate an exposure model for a
particular site. Such field measurements
should be specified in the WP and SAP
For example, studies of tissue residue levels
often are used to calibrate exposure and
food-chain models.
STEP 6, Page 5
-------
August 21. 1996 DRAFT
Exposure-response analysis. The exposure-response analysis for a Super-fund site
describes the relationship between the magnitude, frequency, or duration of a contaminant
stressor in an experimental or observational setting and the magnitude of response. In this
phase of the analysis, measurement endpoints are related to the assessment endpoints using
the logical structure provided by the conceptual model. Any extrapolations that are required
to relate measurement to assessment endpoints (e.g., between species, between response
levels, from laboratory to field) are explained. Finally, an exposure-response relationship is
described to the extent possible (e.g., by a regression equation), including the confidence
limits (quantitative or qualitative) associated with the relationship.
Under some circumstances, site-specific exposure-response information can be
obtained by evaluating existing ecological impacts along a contamination gradient at the site.
Various statistical regression techniques can be used to identify or describe the relationship
between e-xposure and response from the field data. In these cases, the potential for
confounding stressors that might correlate with the contamination gradient should be
considered (e.g., decreasing water temperature downstream of a site; reduced soil erosion
further from a site).
An exposure-response analysis is of particular importance to nsk managers who must
balance human health and ecological concerns against the feasibility and effectiveness of
remedial options. An exposure-response function can help a risk manager to specify the
trade-off between the degree of cleanup and likely benefits of the cleanup and to balance
ecological and financial costs and benefits of different remedial options, as discussed in
Step 8.
When exposure-response data are not available or cannot be developed, a threshold for
adverse effects can be developed instead, as in Step 2. For the baseline risk assessment,
however, site-specific information should be used instead of conservative assumptions
whenever possible.
Evidence of causality. At Superfund sues, it is important to evaluate the strength
of the causaJ association between site-related contaminants and effects on the measurement
and assessment endpoints. Demonstrating a correlation between a contaminant gradient and
ecological impacts at a site is a key component of establishing causality, but other evidence
can be used in the absence of such a demonstration. Moreover, an exposure-response
correlation at a site is not sufficient to demonstrate causality, but requires one or more types
of supporting evidence and analysis of potential confounding factors. Hill's (1965) cntena
for evaluating causal associations are outlined in the Framework (U.S. EPA, 1992a).
STEP 6. Page 6
-------
August 21. 1996 DRAFT
6.4 SCIENTIFIC/MANAGEMENT DECISION POINT (SMDP)
An SMDP dunng the site investigation and analysis phase is needed oniy if alterations
to the WP or SAP become necessary. In the worst cases, changes in measurement endpoints
could be required, with corresponding changes to the testable hypotheses and sampling
design. Any new measurement endpoints must be evaluated according to their utility for
inferring changes in the assessment endpoints and their compatibility with the site conceptual
model; otherwise, the study could fail to meet its objectives.
Proposed changes to the SAP must be made in consultation with the nsk manager and
the risk assessor. The risk manager must understand what changes have been made and why,
and must ensure that the risk management decisions can be made from the information that
the new study design can provide. The risk assessor must be involved to ensure that the
assessment endpoints and study questions or testable hypotheses are still being addressed.
6.5 SUMMARY
The site investigation step of the ecological risk assessment should be a
straightforward implementation of the study designed in Step 4. In instances where
unexpected conditions arise in the field that indicate a need to change the study design, the
ecological risk assessor should reevaluate the feasibility or adequacy of the sampling design.
Any proposed changes to the WP or SAP must be agreed upon by both the risk assessor and
the risk manager.
The analysis phase of the ecological nsk assessment consists of the technical
evaluation of data on existing and potential exposures and ecological effects and is based on
the information collected during Steps 1 through 6. Analysis of exposure and effects is
performed interactively, and follows the data interpretation and analysis methods specified in
the WP and SAP Site-specific data obtained during Step 6 replaces many of the assumptions
that were made for the screening-level analysis in Steps 1 and 2. Evidence of an exposure-
response relationship between contamination and ecological responses at a site helps to
establish causality. The results of Step 6 are used to characterize ecological risks in Step 7.
STEP 6, Page 7
-------
August 21. 1996 DRAFT
STEP 7: RISK CHARACTERIZATION
OVERVIEW
In risk characterization, data on exposure and effects are integrated into a
statement about risk to the assessment endpoints established during problem
formulation. A weight-of-evidence approach is used to interpret the implications
of different studies or tests for the assessment endpoints. In a well-designed
study, data analysis should be straightforward, because the procedures were
established in the WP and SAP. The risk characterization section of the baseline
ecological risk assessment should include a qualitative and quantitative
presentation of the risk results and associated uncertainties.
7.1 INTRODUCTION
Risk characterization is the final phase of the risk assessment process and includes two
major components: risk estimation and risk description (U.S. EPA, 1992a; Exhibit 7-1). Risk
estimation consists of integrating the exposure profiles with the exposure-effects information
(Section 7.2) and summarizing the associated uncertainties (Section 7.3). The risk description
provides information important for interpreting the risk results and, in the Superfund Program,
identifies a threshold for adverse effects on the assessment
endpoints (Section 7.4).
It is U.S. EPA policy that risk characterization should be consistent with the values of
"transparency, clarity, consistency, and reasonableness" (U.S. EPA, 1995f). "Well-balanced
risk characterizations present risk conclusions and information regarding the strengths and
limitations of the assessment for other risk assessors, EPA decision-makers, and the public"
(U.S. EPA, 19950- Thus, the documentation of risks should be easy to follow and to
understand, with all assumptions, defaults, uncertainties, professional judgments, and any
other inputs to the risk estimate clearly identified and easy to find.
7.2 RISK ESTIMATION
Documentation of the risk estimates should describe how inferences are made from the
measurement endpoints to the assessment endpoints established in problem formulation. As
stated earlier, it is not the purpose of this document to provide a detailed guidance on the
STEP 7, Page 1
-------
Aueusi I!. !?96
DRAFT
EXHIBIT 7-1
Risk Characterization
Risk Estimation
Irrttgrsttoo
RlskOtsc
rirxtofl '
Urtc*rUUnty
r
Risk
Summary
'
'
ln»«rpr»utlon
Ecologies
Slgnlflcsnc*
DlsctASion 3«tor*«n th«
RJsk ASMSMX and Rlak
(R««ulU)
B
n
a
c
O
a
2
o
3
O
Risk
STEP 7, Pagt 2
-------
August:!. 1996 ^ DRAFT
selection and utilization of risk models. The risk assessor should have developed and the risk
manager should have agreed upon the model used to characterize risk, its assumptions,
uncertainties, and interpretation in Steps 3 through 5. This agreement is specified in the site
WP and SAP and is the purpose of the SMDPs in Steps 3 through 5.
Unless the site investigation during Step 6 discovers new information, the risk
assessment should move smoothly through the risk characterization phase, because the
analysis procedures were specified in the WP and SAP. While it might be informative to
investigate a data set for trends, outliers, or other statistical indicators, these investigations
should be secondary to the data interpretations specified in the SAP Unless a data
interpretation process is specified in the SAP and followed during nsk characterization,
biased, seriously conflicting, or superfluous conclusions might be obtained. Those outcomes
can divert or confound the nsk characterization process.
For ecological risk assessments that entail more than one type of study (or line of
evidence), a strength-of-evidence approach is used to integrate different types of data to
support a conclusion. The data might include toxiciry test results, assessments of existing
impacts at a site, or risk calculations comparing exposures estimated for the site with toxicity
values from the literature. Balancing and interpreting the different types of data can be a
major task and require professional judgment. As indicated above, the strength of evidence
provided by different types of tests and the precedence that one type of study might have over
another should already have been established during Step 4. Taking this approach will ensure
that data interpretation is objective and not biased to support a preconceived answer.
Additional strength-of-evidence considerations at this stage include the degree to which DQOs
were met and whether confounding factors became evident in the site investigation and
analysis phase.
For some biological tests (e.g., toxicity tests, benthic macroinvertebrate studies), all or
some of the data interpretation process is outlined in existing documents, such as in toxicity
testing manuals. However, in most cases, it will be necessary for the SAP to provide details
on how the data are to be interpreted for a site. The data interpretation methods also should
be presented in the risk characterization documentation. For example, if the triad approach
was used to evaluate contaminated sediments, the risk estimation section should describe how
the three types of studies (i.e., toxiciry test, benthic invertebrate survey, and sediment
chemistry) are integrated to draw conclusions about nsk.
Where exposure-response functions are not available or developed, the quotient
method of comparing an estimated exposure concentration to a threshold for response can be
used, as in Step 2. Whenever possible, however, presentation of full exposure-response
functions provides the risk manager with more information on which to base site decisions.
This guidance has recommended the use of on-site contamination gradients to demonstrate on-
site exposure-response functions. Where such data have been collected, they should be
STEP 7, Page 3
-------
August I! 1996 _ DRAFT
presented along with the nsk estimates. Hazard indices, the results of in situ toxicity testing,
or community survey data can be mapped along with analytic chemistry data to provide a
clear picture of the relationship between areas of contamination and effects.
7.3 RISK DESCRIPTION
A key to nsk description for Superfund sites is documentation of environmentaJ
contamination levels that bound the threshold for adverse effects on the assessment endpoints
(Section 7.3.1). The nsk descnption also provides information to help the nsk manager judge
the ecological significance of the estimated nsks (Section 7.3.2).
7.3.1 Threshold for Effects on Assessment Endpoints
Key outputs of the nsk characterization step are contaminant concentrations in each
environmentaJ medium that bound the threshold for estimated adverse ecological effects based
on the uncertainty inherent in the data and models used. The lower bound of the threshold
would be based on consistent conservative assumptions and NOAEL toxicity values. The
upper bound would be based on observed impacts or predictions that ecological impacts could
be occurring. This upper bound would be developed using consistent assumptions, site-
specific data, LOAEL toxicity values, or an impact evaluation.
The approach to estimating environmental contaminant concentrations that represent
thresholds for adverse .ecological effects should have been specified in the study design.
When higher-trophic-level organisms are associated with assessment endpoints, the study
design should have descnbed how monitoring data and contaminant transfer models would be
used to back^alculate an environmental concentration representing a threshold for effect. If
the site investigation demonstrated a gradient of ecological effects along a contamination
gradient, the nsk assessor can identify and document the levels of contamination below which
no further improvements in the measurement (or assessment) endpoints are discernable. If
departures from the onginal analysis plan are necessary based on information obtained during
the sue investigation or data analysis phase, the reasons for change should be documented.
When assessment endpoints include populations of animals that can travel moderate
distances, different ways of presenting a threshold for adverse effects are possible. Various
combinations of level of contamination and areal extent of contamination relative to the
foraging range of the animals can result in similar contaminant intake levels by the animals.
In this case, a point of departure for identifying a threshold for effect would be to identify
that level ot contamination, which if uniformly distributed both at the site and beyond, would
not pose a threat. The assumption of uniform contamination has been used to back-calculate
water quality cntena to protect piscivorous wildlife in the Great Lakes (U.S. EPA, 1995a).
Again, use of this approach should have been specified in the study design.
STEP 7. Page 4
-------
August 21. 1996 DRAFT
7.3.2 Additional Risk Information
In addition to developing numerical estimates of existing impacts, risks, and thresholds
for effect, the nsk assessor should put the estimates in context with a description of their
extent, magnitude, and potential ecological significance. Additional ecological nsk
descriptors are listed below:
The location and areal extent of existing contamination above a threshold for
adverse effects;
The degree to which the threshold for contamination is exceeded or is likely to
be exceeded in the future, particularly if exposure-response functions are
available; and
The expected "half-life" (qualitative or quantitative) of contaminants in the
environment (e.g., sediments, food chain) once the sources of contamination are
removed.
To interpret the information- in light of remedial options, the risk manager might need to
solicit input from specific experts.
7.4 UNCERTAINTY ANALYSIS
There are several sources of uncertainties associated with Superfund risk estimates.
One is the initial selection of substances of concern based on the sampling data and available
toxicity information. Other sources of uncertainty include estimates of toxicity to ecological
receptors at the site based on limited data from the laboratory (usually on other species), data
from other ecosystems, or data from the site over a limited period of time. Additional
uncertainties result from the exposure assessment, as a consequence in the uncertainty in
chemical monitoring data and models used to estimate exposure concentrations or doses.
Finally, further uncertainties are included in nsk estimates when simultaneous exposures to
multiple substances occurs.
Uncertainty should be distinguished from variability, which arises from true
heterogeneity or variation in characteristics of the environment and receptors. Uncertainty, on
the other hand, represents lack of knowledge about certain factors which can be reduced by
additional study.
This section briefly notes several categories of uncertainty (Section 7.4.1) and
techniques for tracking uncertainty through a risk assessment (Section 7.4.2). Additional
STEP 7, Page 5
-------
August :i. 1996 DRAFT
guidance on discussing uncertainty and variability in risk characterization is provided in U.S.
EPA's (1992D Guidance on Risk Characten-anon for Risk Managers and Risk Assessors.
7.4.1 Categories of Uncertainty
There are three basic categories of uncertainties that apply to Supcrfund site risk
assessments: (I) conceptual model uncertainties; (2) natural variation and parameter error; and
(3) model error. Each of these is described below.
There will be uncertainties associated with the conceptual model used as the basis to
investigate the site. The initial characterization of the ecological problems at a Superfund
site, likely exposure pathways, chemicals of concern, and exposed ecological components,
requires professional judgments and assumptions. To the extent possible, the nsk assessor
should describe what judgments and assumptions were included in the conceptual model that
formed the basis of the WP and SAP
Parameter values (e.g., water concentrations, tissue residue levels, food ingestion rates)
usually can be characterized as a distribution of values, described by central tendencies,
ranges, and percentiles, among other descriptors. When evaluating uncertainty in parameter
values, it is important to distinguish uncertainty from variability. Ecosystems include highly
variable abiotic (e.g., weather, soils) and biotic (e.g., population density) components. If all
instances of a parameter (e.g., the weight of all members of a population) could be sampled,
the "true" parameter value distribution could be described. For realistic sampling efforts,
however, only a fraction of the instances (e.g., a few of the members of the population) can
be sampled, leaving uncertainty concerning the true parameter value distribution. The risk
assessor should provide either quantitative or qualitative descriptions of uncertainties in
parameter value distributions.
Finally, there is uncertainty associated with how well a model (e.g., fate and transport
model) approximates true relationships between site-specific environmental conditions.
Models available at present tend to be fairly simple and at best, only partially validated with
field tests. As a consequence, it is important to identify key model assumptions and their
potential impacts on the nsk estimates.
7.4.2 Tracking Uncertainties
In general, there are two approaches to tracking uncertainties through a nsk
assessment: d) using various point estimates of exposure and response to develop one or
more point estimates of nsk (e.g., central tendency and high end); and (2) conducting a
Monte Carlo simulation to predict a full distribution of risks based on a distribution of
exposures and exposure-response information.
STEP 7, Page 6
-------
August 21. 1996 DRAFT
Although Monte Carlo approaches to traciung uncertainty have become widely used in
risk assessment in recent years, the potential pitfalls of such an approach must be appreciated.
It is critical that parameters that covary are not modeled as though they are independent of
one another. A variety of techniques can be used to establish a covanance in a Monte Carlo
simulation, depending on the software being used. However, if data to describe the
distribution of input parameters are limited, the level of confidence in the output distribution
is similarly low. Finally, combining both uncertainty (e.g., no site-specific information on
what an animaJ eats, many non-detects in the chemical analyses) and natural variability (e.g.,
in body weight, measured contaminant levels) to run a single Monte Carlo is not particularly
helpful. It is more appropriate to evaluate natural variability within a single simulation and to
evaluate uncertainty using a sensitivity analysis.
7,5 SUMMARY
Risk characterization integrates the results of the exposure profile and exposure-
response analyses, and is the final phase of the risk assessment process. It consists of risk
estimation and risk description, which together provide information to help judge the
ecological significance of risk estimates in the absence of remedial activities. The risk
description also identified a threshold for effects on the assessment endpoint as a range
between contamination levels identified as posing no ecological risk and the lowest
contamination levels identified as likely to produce adverse ecological effects. To ensure that
the risk characterization is transparent, clear, and reasonable, information regarding the
strengths and limitations of the assessment must be identified and described.
STEP 7, Page 7
-------
August 21. 1996 DRAFT
STEPS: RISK MANAGEMENT
OVERVIEW
Risk management at a Superfund site is the responsibility of the site risk
manager, who must balance risk reductions associated with cleanup of
contaminants with potential impacts of the remedial actions themselves. In Step
7. the nsk assessor identified a threshold for effects on the assessment endpoint as
a range between contamination levels identified as posing no ecological nsk and
the lowest contamination levels identified as likely to produce adverse ecological
effects. In Step 8, the risk manager evaluates several factors in deciding whether
or not to clean up to within that range.
8.1 INTRODUCTION
Risk management is a distinctly different process from nsk assessment (NRC, 1983,
1994; U.S. EPA, 1984b, 1995f). The nsk assessment establishes that a risk is present and
defines a range or magnitude of the risk. In risk management, the results of the risk
assessment are integrated with other considerations to make and justify risk management
decisions. Additional risk management considerations can include the implications of existing
background levels of contamination, available technologies, tradeoffs between human and
ecological concerns, and costs of alternative actions, to decide what actions to take.
8.2 ECOLOGICAL RISK MANAGEMENT IN SUPERFUND
According to section 300.40 of the NCP, the purpose of the remedy selection process
is to eliminate, reduce, or control risks to human health and the environment. The NCP
indicates further that the results of the baseline nsk assessment will help to establish
acceptable exposure levels for use in developing remedial alternatives in the feasibility study
(FS). Based on the criteria for selecting the preferred remedy and, using information from the
human health and ecological risk assessments and the evaluation of remedial options in the
feasibility study (FS), the risk manager then selects a preferred remedy.
The nsk manager must consider several types of information in addition to the
baseline ecological nsk assessment when evaluating remedial options (Section 8.2.1). Of
particular concern for ecological risk management at Superfund sites is the potential for
remedial actions themselves to cause adverse ecological impacts (Section 8.2.2). There also
STEP 8, Page 1
-------
August :i 1996 DRAFT
exists the opportunity to monitor ecological components of the site to gauge the effectiveness
(or impactsi of the selected remedy (Section 8.2.3).
8.2.1 Other Risk Management Considerations
The baseline ecological nsk assessment is not the only set of information that the risk
manager must consider when evaluating remedial options during the FS phase of the
Superfund process. The NCP specifies that each remedial alternative should be evaluated
according to the following cntena:
(1)
Overall protection of human health and the environment;
(2) Compliance with applicable or relevant and appropriate requirements (ARARs)
(unless waiver applicable);
(3) Long-term effectiveness and permanence;
(4) Reduction of toxicity, mobility, or volume of hazardous wastes through the use
of treatment;
(5) Short-term effectiveness;
(6) Implementability;
(7) Cost;
(8) State acceptance; and
'9> Community acceptance.
Additional factors that the site ask manager takes into consideration include existing
background levels (see U.S. EPA. 1994g); current and likely future land uses (see U.S. EPA,
I995c); current and Likely future resource uses in the area; and local, regional, and national
ecological significance of the site.
Consideration of the ecological impacts of remedial options and residual asks
associated with leaving contaminants in place are very important considerations. These
considerations are described in the next secuon (8.2.2).
STEP 8, Page 2
-------
August 21. 1996 DRAFT
8.2.2 Ecological Impacts of Remedial Options
Management of ecological risks must take into account the potential for impacts to the
ecological assessment endpomts from implementation of various remedial options. The risk
manager must balance: (1) residual risks posed by site contaminants before and/or after
implementation of the selected remedy with (2) the potential impacts of the selected remedy
on the environment independent of contaminant effects. The selection of a remedial
alternative could require tradeoffs between long-term and short-term risk.
The ecological risks posed by the "no action" alternative are the risks estimated by the
baseline ecological risk assessment. For all other remedial options, there might be some
ecological impact associated with the remedy. This impact could be anything from a short-
term loss to complete and permanent loss of the present habitat and ecological communities.
In instances where substantial ecological impacts will result from the remedy (e.g., dredging a
wetland), the risk manager will need to consider ways to mitigate the impacts of the remedy
and compare the mitigated impacts to the threats posed by the site contamination.
During the FS, the boundaries of potential risk under the no-action alternative (i.e.,
baseline conditions) can be compared with the evaluation of potential impacts of the remedial
options to help justify the preferred remedy. As indicated above, the preferred remedy should
minimize the risk of long-term impacts that could result from the remedy and any residual
contamination. When the selected remedial option leaves some site contaminants presumed to
pose an ecological risk in place, the justification for the selected remedy must be clearly
documented.
In short, consideration of the environmental effects of the remedy itself might result in
a decision to allow contaminants to remain on site at levels higher than the threshold for
effects on the assessment endpoint. Thus, selection of the most appropriate ecologically-
based remedy can result in residual contaminant levels and ecological impacts for which the
PRPs remain responsible.
8.2.3 Monitoring
Ecological risk assessment is a relatively new field with limited data available to
validate or calibrate its predictions. At sites where remedial actions are taken to reduce
ecological impacts and risks, the results of the remediation efforts should be compared with
the predictions made during the ecological risk assessment.
While it often is difficult to demonstrate the effectiveness of remedial actions in
reducing human health nsks, it often is possible to demonstrate the effectiveness of
remediations to reduce ecological risks, particularly if a several-year monitoring program is
established. The site conceptual model provides the conceptual basis for monitoring options,
STEP 8, Page 3
-------
August:!. :?96 DRAFT
and the sue investigation should have indicated which options might be most practical for the
site Monitoring aJso is important to assess the effectiveness of a no-action alternative. For
example, monitoring sediment contamination and benthic communities at intervals following
removal of a contaminant source allows one to test predictions of the potential for the
ecosystem to recover naturally over time.
8.3 SCIENTIFIC/MANAGEMENT DECISION POINT (SMDP)
The risk management decision is finalized in the Record of Decision (ROD). The
decision should minimize the nsk of long-term impacts that could result from the remedy and
any residual contamination. When the selected remedy leaves residual contamination at levels
higher than the upper-bound estimate of the threshold for adverse effects on the assessment
endpoint, the risk manager should justify the decision.
8.4 SUMMARY
Risk management decisions are the responsibility of the sue manager (the nsk
manager), not the nsk assessor. The nsk manager should have been involved in the nsk
assessment from the beginning; knowing the options available for reducing risks, the nsk
manager can help to frame questions during the problem formulation phase of the nsk
assessment.
The nsk manager must understand the nsk assessment, including its uncertainties,
assumptions, and level of resolution of the assessment. With an understanding of potential
adverse effects posed by residual levels of sue contaminants and posed by the remedial
actions themselves, the nsk manager can balance the ecological costs and benefits of the
available remedial options. Understanding the uncertainties associated with the risk
assessment also is cntical to evaluating the overall protectiveness of any remedy.
STEP 8. Page 4
-------
August 21. 1996 ^ DRAFT
GLOSSARY
This glossary includes definitions from several sources. A superscript number next to
a word identifies the reference from which the definition was adapted (listed at the end of the
Glossary).
Abiotic. Characterized by absence of life; abiotic materials include non-living environmental
media (e.g., water, soils, sediments); abiotic characteristics include such factors as light,
temperature, pH, humidity, and other physical and chemical influences.
Absorption Efficiency. A measure of the proportion of a substance that a living organism
absorbs across exchange boundaries (e.g., gastrointestinal tract).
Absorbed Dose. The amount of a substance penetrating the exchange boundaries of an
organism after contact. Absorbed dose for the inhalation and ingestion routes of exposure is
calculated from the intake and the absorption efficiency. Absorbed dose for dermal contact
depends on the surface area exposed and absorption efficiency.
Accuracy.4 The degree to which a measurement reflects the true value of a variable.
Acute.5 Having a sudden onset or lasting a short time. An acute stimulus is severe enough
to induce a response rapidly. The word acute can be used to define either the exposure or the
response to an exposure (effect). The duration of an acute aquatic toxicity test is generally 4
days or less and mortality is the response usually measured.
Acute Response. The response of (effect on) an organisms which has a rapid onset. A
commonly measured rapid-onset response in toxicity tests is mortaliry
Acute Tests. A toxicity test of short duration, typically 4 days or less (i.e., of short duration
relative to the Ufespan of the test organism).
Administered Dose.2 The mass of a substance given to an organism and in contact with an
exchange boundary (i.e., gastrointestinal tract) per unit body weight (BW) per unit time (e.g.,
mg/kgBW/day).
Adsorption.14 Surface retention of molecules, atoms, or ions by a solid or liquid, as opposed
to absorption, which is penetration of substances into the bulk of a solid or liquid.
Area Use Factor. The ratio of an organism's home range, breeding range, or
feeding/foraging range to the area of contamination of the site under investigation.
GLOSSARY, Page 1
-------
Augusi I!. 1996 DRAFT
.Assessment Endpoint.6 An explicit expression of the environmental value that is to be
protected.
Benthic Community. The community of organisms dwelling at ihe bottom of a pond, river,
lake, or ocean.
Bioaccumulation 5 General term describing a process by which chemicals are taken up by an
organism either directly from exposure to a contaminated medium or by consumption of food
containing the chemical.
Bioccumulation Factor 3 The ratio of the concentration of a contaminant in an organism to
the concentration in the ambient environment at steady state.
Bioassay.5 Test used to evaluate the relative potency of a chemical by comparing its effect
on living organisms with the effect of a standard preparation on the same type of organism.
Bioassay and toxicity tests are not the samesee toxicity test.
Bioassessment. A general term referring to environmental evaluations involving living
organisms; can include bioassays, community analyses, etc.
Bioavailability 4 The degree to which a material in environmental media can be assimilated
by an organism.
Bioconcentration 5 A process by which there is a net accumulation of a chemical directly
from an exposure medium into an organism.
Biodegrade. Decompose into more elementary compounds by the action of living
organisms, usually referring to microorganisms such as bacteria.
Biomagnification." Result of the process of bioaccumulation and biotransfer by which tissue
concentrations of chemicals in organisms at one trophic level exceed tissue concentrations in
organisms at the next lower trophic level in a food chain.
Bicmarker.' Biochemical, physiological, and rustological changes in organisms that can be
used to estimate either exposure to chemicals or the effects of exposure to chemicals.
Biomonitoring. Use of Living organisms as "sensors" in environmental quality surveillance
to detect changes in environmental conditions that might threaten living organisms in the
environment.
Body Burden. The concentration or total amount of a substance in a living organism;
implies accumulation of a substance above background levels in exposed organisms.
GLOSSARY, Page 2
-------
August 21. 1996 DRAFT
Breeding Range. The area utilized by an organism during the reproductive phase of its life
cycle and dunng the time that young are reared.
Bulk Sediment. Field collected sediments used to conduct toxicity tests; can contain
multiple contaminants and/or unknown concentrations of contaminants.
Characterization of Ecological Effects.6 A portion of the analysis phase of ecological risk
assessment that evaluates the ability of a stressor to cause adverse effects under a particular
set of circumstances.
Characterization of Exposure.6 A portion of the analysis phase of ecological nsk
assessment that evaluates the interaction of the stressor with one or more ecological
components. Exposure can be expressed as co-occurrence, or contact depending on the
stressor and ecological component involved.
Chemicals of Potential Concern.2 Chemicals that are potentially site-related and whose data
are of sufficient quality for use in a quantitative risk assessment.
Chronic.5 Involving a stimulus that is lingering or continues for a long time; often signifies
periods from several weeks to years, depending on the reproductive life cycle of the species.
Can be used to define either the exposure or the response to an exposure (effect). Chronic
exposures typically induce a biological response of relatively slow progress and long duration.
Chronic Response. The response of (or effect on) an organism to a chemical that is not
immediately or directly lethal to the organism.
Chronic Tests.9 A toxicity test used to study the effects of continuous, long-term exposure
of a chemical or other potentially toxic material on an organism.
Community.6 An assemblage of populations of different species within a specified location
and time.
Complexation.14 Formation of a group of compounds in which a pan of the molecular
bonding between compounds is of the coordinate type.
Concentration. The relative amount of a substance in an environmental medium, expressed
by relative mass (e.g., mg/kg), volume (ml/L), or number of units (e.g., parts per million).
Concentration-Response Curve.5 A curve describing the relationship between exposure
concentration and percent of the test population responding.
GLOSSARY, Page 3
-------
August 211906 DRAFT
Conceptual Model.6 Describes a series of working hypotheses of how the stressor might
affect ecological components. Describes ecosystem or ecosystem components potentially at
risk, and the relationships between measurement and assessment endpomts and exposure
scenarios.
Contaminant of (Ecological) Concern. A substance detected at a hazardous waste site that
has the potentiaJ to affect ecological receptors adversely due to its concentration, distribution,
and mode of toxiciry.
Control.5 A treatment in a toxicity test that duplicates all the conditions of the exposure
treatments but contains no test material. The control is used to determine the response rate
expected in the test organisms in the absence of the test material.
Correlation.10 An estimate of the degree to which two sets of variables vary together, with
no distinction between dependent and independent variables.
Critical Exposure Pathway. An exposure pathway which either provides the highest
exposure levels or is the primary pathway of exposure to an identified receptor of concern.
Degradation.'4 Conversion of an organic compound to one containing a smaller number of
carbon atoms.
Deposition.14 The lying, placing, or throwing down of any material.
Depuration.5 A process that results in elimination of toxic substances from an organism.
Depuration Rate. The rate at which a substance is depurated from an organism.
Dietary Accumulation. The net accumulation of a substance by an organism as a result of
ingestion in the diet.
Direct Effect (toxin). An effect where the suessor itself acts directly on the ecological
component of interest, not through other components of the ecosystem.
Dose. A measure of exposure. Examples include (1) the amount of a chemicaJ ingested,
(2) the amount of a chemical absorbed, and (3) the product of ambient exposure concentration
and the duration of exposure.
Dose-Response Curve." Similar to concentrauon-response curve except that the dose (i.e. the
quantity) of the chemicaJ administered to the organism is known. The curve is plotted as
Dose versus Response.
GLOSSARY, Page 4
-------
A"gust:i. 1996 DRAFT
Duplicate. A sample taken from and representative of the same population as another
sample. Both samples are carried through the steps of sampling, storage, and analysis in an
identical manner.
Ecological Component. Any part of an ecosystem, including individuals, populations,
communities, and the ecosystem itself.
Ecological Risk Assessment.6 The process that evaluates the likelihood that adverse
ecological effects may occur or are occurring as a result of exposure to one or more stressors.
Ecosystem. The biotic community and abiotic environment within a specified location and
time.
Ecotoxicity. The study of toxic effects on nonhuman organisms, populations, or
communities.
Estimated or Expected Environmental Concentration.5 The concentration of a material
estimated as being likely to occur in environmental media to which organisms are exposed.
Exposure.6 Co-occurrence of or contact between a stressor and an ecological component.
The contact reaction between a chemical and a biological system, or organism.
Exposure Assessment.2 The determination or estimation (qualitative or quantitative) of the
magnitude, frequency, duration, and route of exposure.
Exposure Pathway.2 The course a chemical or physical agent takes from a source to an
exposed organism. Each exposure pathway incudes a source or release from a source, an
exposure point, and an exposure route. If the exposure point differs from the source,
transport/exposure media (i.e., air, water) also are included.
Exposure Pathway Model. A model in which potential pathways of exposure are identified
for the selected receptor species.
Exposure Point.2 A location of potential contact between an organism and a chemical or
physical agent.
Exposure Point Concentration. The concentration of a contaminant occurring at an
exposure point.
Exposure Profile.6 The product of characterizing exposure in the analysis phase of
ecological nsk assessment. The exposure profile summarizes the magnitude and spatial and
temporal patterns of exposure for the scenarios described in the conceptual model.
GLOSSARY, Page 5
-------
August 21. 1996 DRAFT
Exposure Route.: The way a chemical or physical agent comes in contact with an organism
(i.e., by mgestion. inhalation, or dermal contact).
Exposure Scenario.6 A set of assumptions concerning how an exposure takes place,
including assumptions about the exposure setting, stressor characteristics, and activities of an
organism that can lead to exposure.
False Negative. The conclusion that an event (e.g., response to a chemical) is negative when
it is in fact positive.
False Positive. The conclusion that an event is positive when it is in fact negative.
Fate.5 Disposition of a material in various environmental compartments (e.g. soil or
sediment, .water, air, biota) as a result of transport, transformation, and degradation.
Food-Chain Transfer. A process by which substances in the tissues of lower-trophic-level
organisms are transferred to the higher-trophic-level organisms that feed on them.
Forage (feeding) Area. The area utilized by an organism for hunting or gathering food.
Habitat.1 Place where a plant or animal lives, often characterized by a dominant plant form
and physical characteristics.
Hazard. The likelihood that a substance will cause an injury or adverse effect under
specified conditions.
Hazard Identification. The process of determining whether exposure to a stressor can
cause an increase in the incidence of a particular adverse effect, and whether an adverse
effect is likely to occur.
Hazard Quotient.' The ratio of an exposure level to a substance to a toxicity value selected
for the nsk assessment for that substance (e.g.. LOAEL or NOAJEL).
Home Range. The area to which an animal confines its activities.
Hydrophilic." Denoting the property of attracting or associating with water molecules;
characteristic of polar or charged molecules.
Hydrophobic.1- With regard to a molecule or side group, tending to dissolve readily in
organic solvents, but not in water, resisting wetting, not containing polar groups or sub-
groups.
GLOSSARY, Page 6
-------
August 21. 1996 DRAFT
Hypothesis. - A proposition set forth as an explanation for a specified phenomenon or group
of phenomena.
Indirect Effect.6 An effect where the stressor acts on supporting components of the
ecosystem, which in turn have an effect on the ecological component of interest.
Ingestion Rate. The rate at which an organism consumes food, water, or other materials
(e.g., soil, sediment). Ingestion rate usually is expressed in terms of unit of mass or volume
per unit of time (e.g., kg/day, L/day).
lonization. The process by which a neutral atom loses or gains electrons, thereby acquiring
a net charge and becoming an ion.
Lethal.5 Causing death by direct action.
Lipid.1 One of a variety of organic substances that are insoluble in polar solvents, such as
water, but that dissolve readily in non-polar organic solvents. Includes fats, oils, waxes,
steroids, phospholipids, and carotenes.
Lowest-Observable-Adverse-Effect Level (LOAEL). The lowest level of a stressor
evaluated in a toxicity test or biological field survey that has a statistically significant adverse
effect on the exposed organisms compared with unexposed organisms in a control or
reference site.
Matrix.14 The substance in which an analyte is embedded or contained; the properties of a
matrix depend on its constituents and form.
Measurement Endpoint.6 A measurable ecological characteristic that is related to the valued
characteristic chosen as the assessment endpomt. Measurement endpomts often are expressed
as the statistical or arithmetic summar-es of the observations that make up the measurement.
As used in this guidance document, measurement endpoints can include measures of effect
and measures of exposure.
Media.15 Specific environmental compartmentsair, water, soilwhich are the subject of
regulatory concern and activities.
Median Effective Concentration (EC50).5 The concentration of a substance to which test
organisms are exposed that is estimated to be effective in producing some sublethal response
in 50 percent of the test population. The EC50 usually is expressed as a time-dependent
value (e.g., 24-hour EC50). The sublethal response elicited from the test organisms as a
result of exposure must be clearly defined.
GLOSSARY. Page 7
-------
August ::. .^96 DRAFT
Median Lethal Concentration (LC50).5 A statistically or graphically estimated
concentration that is expected to be lethaJ to 50 percent of a group of organisms under
specified conditions.
Metric.16 Relating to measurement; a type of measurementfor example a measurement of
one of various components of community structure (e.g., species richness, % similarity).
Mortality. Death rate or proportion of deaths in a population.
No-Observed-Adverse-Effect L«vel (NOAEL).5 The highest level of a stressor evaluated in
a toxicity test or biological field survey thai causes no statistically significant difference in
effect compared with the controls or a reference site.
Nonparametric.17 Statistical methods that make no assumptions regarding the distribution of
the data.
Parameter.18 Constants applied to a model that are obtained by theoretical calculation or
measurements taken at another time and/or place, and are assumed to be appropriate for the
place and time being studied.
Parametric u Statistical methods used when the distribution of the data is known.
Population.6 An aggregate of individuals of a species within a specified location in space
and time.
Power.10 The power of a statistical test indicates the probability of rejecting the null
hypothesis when it should be rejected (i.e.. the null hypothesis is false). Can be considered
the sensitivity of a statistical test. (See also Appendix D.)
Precipitation. In analytic chemistry, the process of producing a separable solid phase
within a liquid medium.
Precision.' -\ measure of the closeness of agreement among :ndividual measurements.
Reference Site. A relatively unconiaminated site used for comparison to contaminated sites
in environmental monitonng studies, often incorrectly referred to as a control.
Regression Analysis.10 .Analysis of the functional relationship between two variables; the
independent van able is described on the X axis and the dependent variable is described on the
Y axis (i.e. the change in Y is a function of a change in X).
GLOSSARY. Page 8
-------
August 21. 1996 _ DRAFT
Replicate. Duplicate analysis of an individual sample. Replicate analyses are used for
quaJitv control.
Representative Samples 18 Serving as a typical or characteristic sample: should provide
analytical results that correspond with actual environmental quality or the condition
experienced by the contaminant receptor.
Risk. The expected frequency or probability of undesirable effects resulting from exposure
to known or expected stressors.
Risk Characterization. A phase of ecological risk assessment that integrates the results of
the exposure and ecological effects analyses to evaluate the likelihood of adverse ecological
effects associated with exposure to the stressor. The ecological significance of the adverse
effects is discussed, including consideration of the types and magnitudes of the effects, their
spatial and temporal patterns, and the likelihood of recovery.
Sample. Fraction of a material tested or analyzed; a selection or collection from a larger
collection.
Scientific/Management Decision Point (SMDP). A point during the risk assessment process
when the risk assessor communicates results of the assessment at that stage to a risk manager.
At this point the risk manager determines whether the information is sufficient to arrive at a
decision regarding risk management strategies and/or the need for additional information to
characterize risk.
Sediment.20 Paniculate material lying below water.
Sensitive Life Stage. The life stage (i.e., juvenile, adult, etc.) that exhibits the highest degree
of sensitivity (i.e., effects are evident at a lower exposure concentration) to a contaminant in
toxiciry tests.
Species.13 A group of organisms that actually or potentially interbreed and are reproducuvely
isolated from all other such groups; a taxonomic grouping of morphologically similar
individuals; the category below genus.
Statistic.10 A computed or estimated statistical quantity such as the mean, the standard
deviation, or the correlation coefficient.
Stressor.6 Any physical, chemical, or biological entity that can induce an adverse response.
GLOSSARY, Page 9
-------
August 21. 1996 DRAFT
SublethaJ5 Below the concentration that directly causes death. Exposure to sublethaJ
concentrations of a substance can produce less obvious effects on behavior. biocherrucaJ
andyor ph>siologicaJ functions, and the structure of cells and tissues in organisms.
Threshold Concentration.5 A concentration above whjch some effect (or response) wiiJ be
produced and below which it will not.
Toxicity Assessment. Review of literature, results in toxicity tests, and data from field
surveys regarding the toxicity of any given material to an appropriate receptor.
Toxicity Test.5 The means by which the toxicity of a chemical or other test material is
determined. A toxicity test is used to measure the degree of response produced by exposure
to a specific level of stimulus (or concentration of chemical).
Toxicity Value." A numerical expression of a substance's exposure-response relationship that
is used in nsk assessments.
Toxin. A poisonous substance.
Trophic Level. A functional classification of taxa within a community that is based on
feeding relationships (e.g., aquatic and terrestrial plants make up the first trophic level, and
herbivores make up the second).
Type I Error. Rejection of a true null hypothesis (see also Appendix D).
Type II Error.10 Acceptance of a false null hypothesis (see also Appendix D).
Uptake.- A process by which materials are transferred into or onto an organism.
Uncertainty.1' Imperfect knowlcuge concerning the present or future state of the system
under consideration, a component of risk resulting from imperfect knowledge of the degree of
hazard or of its spatial and temporal distribution.
Volatilization.14 The conversion of a chemical substance from a liquid or solid state to a
gaseous vapor state.
Xenobiotic.6 A chemical or other stressor that does not occur naturally in the environment
Xenobioucs occur as a result of anthropogenic activities such as the application of pesticides
and the discharge of industnal chemicals to air, land, or water.
GLOSSARY, Page 10
-------
August 21. 1996 ^ DRAFT
ENDNOTES
1 Krebs 1978, 2 U.S. EPA 1989, 3 CaJow 1993. 4 Freedman 1989, 5 Rand and Petrocelli
1985, U.S. EPA 1992a, 7 Pucklefs 1990, 8 U.S. EPA 1992b, 9 ASTM 1993a, 10 Sokal and
Rohlf 1981, n Suter 1993, 12 Wallace et aJ. 1981, 13 Curtis 1983. 14 Parker 1994, 15 Sullivan
1993, 16 U.S. EPA 1990, 17 Zar 1984, 18 Keith 1988, 19 Gilbert 1987, 20 ASTM 1993b, 21
Huggett et al. 1992, 22 Stedman 1995.
REFERENCES
American Society for Testing and Materials (ASTM). 1993a. ASTM Standard E 943.
Standard terminology relating to biological effects and environmental fate.
American Society for Testing and Materials (ASTM). 1993b. ASTM Standard E 1525.
Standard guide for designing biological tests with sediments.
Calow, P (ed.). 1993. Handbook of Ecotoxicology. Volume 1. Boston, MA: Blackwell
Publishing.
Curtis, H. 1983. Biology, Fourth Edition. New York, NY: Worth.
Freedman. B. 1989. Environmental Ecology. The Impacts of Pollution and Other Stresses
on Ecosystem Structure and Function. New York, NY: Academic Press.
Gilbert, R.O. 1987. Statistical Methods for Environmental Pollution Monitoring. New York,
NY: Reinhold.
Keith, L.H. (ed.). 1988. Principles of Environmental Sampling. American Chemical Society.
Krebs, C.J. 1978. Ecology: The experimental analysis of distribution and abundance.
Second edition. New York, NY: Harper & Row.
Huggett, R.J.; Kimerle, R.A.; Nehrle, P.M. Jr.; Bergman, H.L. (cds.). 1992. Biomarkers:
Biochemical, Physiological, and Histoloqical Markers of Anthropogenic Stress. A
Special Publication of SETAC. Chelsea, MI: Lewis Publishers.
Parker, S.P. (ed.). 1994. Dictionary of Scientific and Technical Terms. Fifth Edition. New
York, NY: McGraw-Hill.
"Rand, G.M.; Petrocelli, S.R, 1985. Fundamentals of Aquatic Toxicology. Methods and
Applications. New York, NY: McGraw Hill.
GLOSSARY, Page 11
-------
August II. !996 DRAFT
Ricklefs. R.E. 1990. Ecology. Second Edition. New York. NY: W.H. Freeman.
Sokal. R.R.. RohJf. F J 1981 Biometr.. Second Edmon. New York. NY: W.H. Freeman.
Stedman. T L. 1995. Stedman's Medical Dictionary: 26th Edition. Baltimore, MD:
Williams and Wilkins.
Sullivan, T F P 1993. Environmental Regulatory Glossary. Government Institutes. Inc.
Surer. G.W EL 1993. Ecological Risk Assessment. Ann Arbor, MI. Lewis.
U. S. Environmental Protection Agency (U.S. EPA). 1989. Risk Assessment Guidance for
Superfund: Volume 1 Human Health. Washington, DC: Office of Emergency and
Remedial Response; EPA/540/1-89/002.
U. S. Environmental Protection Agency (U.S. EPA). 1990. Macroinvertebratc Field and
Laboratory Methods for Evaluating the Biological Integrity of Surface Waters.
Washington. DC: Office of Water; EPA/600/4-90/030.
U. S. Environmental Protection Agency (U.S. EPA). 1992a. Framework for Ecological Risk
Assessment. Washington, DC: Risk Assessment Forum; EP.V630/R-02/011.
U.S. Environmental Protection Agency (U.S. EPA). 1992b. Sediment Classification Methods
Compendium. Washington, DC: Office of Water; EPA7823/R-092/006.
Wallace, R.A.. King. J.L.; Sanders, G.P 199x. Biology. The Science of Life. Second
Edition IL. Scon, Foresman & Co.
Zar, J.H. 1984 Biostanstical Anal\sis. Princeton, NJ: Prentice-Hail.
GLOSSARY, Page 12
-------
August II. 1996 DRAFT
BIBLIOGRAPHY
This combined reference list and bibliography is intended to provide a broad, but not
all inclusive, list of other materials that may provide useful information for ecological risk
assessments at Superfund sites. These documents include other Superfund Program guidance
documents, standard guides for toxicity testing, other EPA program office references with
potential applications at Superfund sites, and other ecological nsk assessment reference
materials. References cited in the text are marked with an asterisk (*).
American Public Health Association (APHA). 1989. Standard Methods for Examination
of Water and Wastewater, 17th edition. Washington, DC: APHA.
American Society for Testing and Materials (ASTM). 1994a. Annual Book of ASTM
Standards. Philadelphia, PA: ASTM.
American Society for Testing and Materials (ASTM). 1994b. Standard guide for conducting
sediment toxicity tests with freshwater invertebrates: ASTM Standard E 1383-94.
American Society for Testing and Materials (ASTM). 1993a. Standard terminology relating
to biological effects and environmental fate: ASTM Standard E 943-93.
American Society for Testing and Materials (ASTM). I993b. Standard guide for designing
biological tests with sediments: ASTM Standard E 1525-93.
American Society for Testing and Materials (ASTM). 1993. ASTM Standards of
Aquatic Toxicology and Hazard Evaluation. Philadelphia, PA: ASTM.
American Society for Testing and Materials (ASTM). 1992. Standard guide for
conducting sediment toxicity tests with freshwater invertebrates: ASTM
Standard E 1383-92.
American Society for Testing and Materials (ASTM). 1992. Standard guide for
conducting 10-day static sediment toxicity tests with manne and estuanne
amphipods: ASTM Standard E 1367-92.
American Society for Testing and Materials (ASTM). 1990. Standard guide for
collection, storage, characterization, and manipulation of sediments for
tcxicological testing: ASTM Standard E 1391-90.
American Society for Testing and Materials (ASTM). 1988. Standard guide for
conducting early life-stage toxicity tests with fishes: ASTM Standard E
1241-88.
BIBLIOGRAPHY, Page 1
-------
August 21. 1996 DRAFT
American Society for Testing and Materials (ASTM). 1984. Standard Practice for
conducting bioconcentration tests with fishes and saltwater bivalve mollusks:
ASTM Standard E 1022-84.
American Society for Testing and Materials (ASTM). 1980. Practice for conducting
acute toxiciry tests with fishes, macroinvertebrates, and amphibians: ASTNf
Standard E 729-80.
American Society for Testing and Materials (ASTM). 1980. Practice for conducting
static acute toxiciry tests with larvae of four species of bivalve moiiusks:
ASTM Standard E 724-80.
Anderson and Hickey. 1972. [reserved]
Ankley, G.T.; Thomas, N.A.; Di Toro, D.M.; et al. 1994. Assessing potential bioavailability
of metals in sediments: a proposed approach. Environ. Manage. 18: 331-337.
Aulerich, R.J.; Ringer, R.K. 1977. Current status of PCB toxicity to mink and effect on
their young. Arch. Environ. Contam. Toxicol. 6: 279-292.
Bamhouse, L.W.; Suter, G.W.; Barteil, S.M.; et al. 1986. User's Manual for Ecological Risk
Assessment. Oak Ridge, TN: Oak Rjdge National Laboratory, Environmental
Sciences Division Publication No. 2679.
Barteil. S.M.: Gardner, R.H.; O'Neill, R.V 1992. Ecological Risk Estimation. New
York, NY: Lewis Publishers.
Baudo, R.; Giesy, J.P.; Muntau, H. 1990. Sediments: Chemistry ind Toxiciry of In-place
Pollutants. Ann Arbor, MI: Lewis Publishers.
Beyer, W.N.; Heinz, G.H.; Redmon-Norwood, A.R. (eds.). 1996. Environmental
Contaminants in Wildlife: Interpreting Tissue Concentrations. A Special Publication
of the Society of Environmental Toxicology and Chemise-/ ;SETAC), La Point, T.W.
(senes edj. Boca Raton, FL: CRC Press, Inc., Lewis Publishers.
Burton, G.A., Jr. ted.). 1992. Sediment Toxiciry Assessment. Ann Arbor, MI: Lewis
Publishers.
Butler. 19" 1. [reserved]
Cairns, J. Jr . Niederlehner, B.R. 1995. Ecological Toxiciry Testing: Scale, Complexity, and
Relevance. Boca Raton, FL: CRC Press, Inc., Lewis Publishers.
BIBLIOGRAPHY, Page 2
-------
August 21. 1996 DRAFT
Calabrese, E.J.; Baldwin, L.A. 1993. Performing Ecological Risk Assessments.
New York, NY: Lewis Publishers.
Carter. M.R. (ed.). 1993. Soil Sampling and Methods of Analysis. Boca Raton, FL: CRC
Press, Inc.. Lewis Publishers.
"Chapman et al. 1992. [reserved]
*Colborn, T.I. 1991. Epidemiology of Great Lakes bald eagles. J. Environ. Health Toxicol.
4: 395^53.
Calow, P. (ed.). 1993. Handbook of Ecotoxicology, Volume 1. Boston, MA: Blackwell
Publishers.
V
Cochran, W.G. 1977. Sampling Techniques. 3rd ed. New York, NY: John Wiley and
Sons, Inc.
Cochran, W.G.; Cox, G.M. 1957. Experimental Design. New York, NY: Wiley.
Cockerham. L.G.; Shane, B.S. (eds.). 1994. Basic Environmental Toxicology. Boca Raton,
FL: CRC Press, Inc., Lewis Publishers.
*Cooke, A.S. 1971. Selective predation by newts on frog tadpoles treated with DDT.
Nature 229: 275-276.
Cowardin, L.M.; Carter, V.; Golet, F.C.; LaRoe, E.T. 1979. Classification of Wetlands
and Deepwater Habitats of the United States. Washington. DC: U.S. Fish and
Wildlife Service; FWS/OBS-79/31.
Crawley, M.J. 1993. GUM for Lcologists. Oxford, UK: Blackwell Scientific Publications.
Curtis, H. 1983. Biology; Fourth Edition. New York, NY: Worth.
Daniel, W.W. 1990. Applied Nonparamemc Statistics. Boston, MA: PWS-KENT
Publishing Company.
Davis, W.S.; Simon, T.P. 1995. Biological Assessment and Criteria: Tools for Water
Resource Planning and Decision Making. Boca Raton, FL: CRC Press, Inc.,
Lewis Publishers.
Diggie, P.J. 1990. Time Series: A Biosatistical Introduction. Oxford Statistical Science
Senes No. 5. Oxford, UK: Clarendon Press.
BIBLIOGRAPHY. Page 3
-------
August ::. 1996 DRAFT
Dilworth et aJ. 1972. [reserved]
*Dourson. M.L., Stara, J.F 1983. Regulatory hjstory and experimental support of
uncertajnry (safety) factors. Reg. Toxicol. Pharmacol. 3. 224-238.
Finney, DJ 1964. Statistical Method in Biological Assay. London. UK: Charles Griffin
and Company.
Finney. D.J. 1970. Probit Analysis: A Statistical Treatment of the Sigmoid Response Curve.
Cambridge, UK: Cambridge University Press.
*Foley, R.E.; Jackling, S.J.; Sloan, R.J. et al. 1988. Organochionne and mercury residues in
wild rruak and otter: comparison with fish. Environ. Toxicol. Chem. 7: 363-374.
Freedman, B. 1989. Environmental Ecology. The Impacts of Pollution and Other Stresses
on Ecosystem Structure and Function. New York. NY: Academic Press.
*Fnberg et al. 1986. [reserved]
"Fuller and Hobson. 1986. [reserved]
Geller, M.D. 1979. Dynamics of three populations of Microtus pennsylvanicus in the
Northwestern United States. PhD Thesis. Binghamton, NY: State University of New
York.
Gilbert, R.O. 1987. Statistical Methods for Environmental Pollution Monitoring. New York,
NY: Reinhold.
Green, R.H. 1979. Sampling Design and Statistical Methods for Environmental Biologists.
New York, NY: Wiley.
Hamelink. J L.; Landrum, P.F.; Bergman, H.L.; Benson, W.H. fcdsi 1994. Unavailability:
Physical, Chemical, and Biological Interactions. Boca Ratcn. FL: CRC Press, Inc.,
Lewis Publishers.
Hill. I.R.; Matihiessen, P.; Heimbach, F. (eds.j. 1993. Guidance Document on Sediment
Toxiciry Tests and Bioassays for Freshwater and Marine Environments. From the
Workshop on Sediment Toxiciry Assessment, Renesse, The Netherlands, November 8-
10. 1993. Society of Environmental Toxicology and Chemistry - Europe.
"Hill, A.B. 1965. The environment and disease: Association or causation? Proceed. Royal
Soc. Med. 58: 285-300.
BIBLIOGRAPHY. Page 4
-------
August:!. 1996 DRAFT
Hoffman, D.J.; Rattner, B.A.; Burton. G.A. Jr.; Cairns, J., Jr. (eds.). 1995. Handbook of
Ecotoxicology. Ann .Arbor, MI: CRC Press, Inc., Lewis Publishers.
Howard, P.H.; Jarvis, W.F.; Meyland, W.M.; Michalenko, E.M. 1991. Handbook of
Environmental Degradation Rates. Boca Raton, FL: CRC Press, Inc., Lewis
Publishers.
Hugget, R.J.; Kimerle, R.A.; Mehrle, P.M., Jr.; Bergman, H.L. 1992. Biomarkers:
Biochemical, Physiological, and Histological Markers of Anthropogenic Stress. A
Special Publication of the Society of Environmental Toxicology and Chemistry
(SETAC), Ward, C.H.; Walton, B.T; La Point, T.W. (senes eds.). Boca Raton, FL:
CRC Press, Inc., Lewis Publishers.
Kabata-Pendias, A.; Pendias H. 1984. Trace Elements in Soils and Plants. Boca Raton, FL:
CRC Press, Inc.
Keith, L.H. (ed.). 1996. EPA s Sampling and Analysis Methods Database; Version 2.0.
Boca Raton, FL: CRC Press, Inc., Lewis Publishers.
Keith, L.H. (ed.). 1988. Principles of Environmental Sampling. American Chemical
Society.
Kendall, R.J.; Lacher, I.E. (eds.). 1994. Wildlife Toxicology and Population Modeling:
Integrated Studies ofAgrotcosystems. A Special Publication of the Society of
Environmental Toxicology and Chemistry (SETAC), La Point, T.W. (series ed.). Boca
Raton, FL: CRC Press, Inc., Lewis Publishers.
*Klemm. D.J.; Lewis. P.A.; Fulk, F.; Lazorchak, J.M. 1990. Macro invertebrate Field and
Laboratory Methods for Evaluating the Biological Integrity of Surface Waters.
Washington, DC: U.S. Environmental Protection Agency. EPA/600/4-90/030.
Kraus, M.L. 1989. Bioaccumulation of heavy metals in pre-fledglmg tree swallows,
Tachycimeta bicolor. Environ. Contam. Toxicol. 43: 407-414.
Krebs, C.J. 1978. Ecology: The Experimental Analysis of Distribution and Abundance;
Second Edition. New York, NY: Harper & Row.
*Krebs, C.J.; Valiela, I.; Harvey, G.R.; Teal. J.M. 1974. Reduction of field populations
of fiddler crabs by uptake of chlorinated hydrocarbons. Mar. Pollut. Bull. 5:
140-142.
BIBLIOGRAPHY, Page 5
-------
August 21 1^(3 DRAFT
Landis, WG.; Yu, M. 1995. Introduction to Environmental Toxicology: Impacts of
Chemicals upon Ecological Systems. Boca Raton. FL: CRC Press, Inc., Lewis
Publishers.
Landis. W G.; Hughes, J.S.; Lewis, M.A. (eds.). 1993. Environmental Toxicity and Risk
Assessment. Philadelphia. PA: American Society for Testing and Materials.
Long, E.R.; Chapman, P.M. 1985. A sediment quality tnad: measures of sediment
contamination, toxicity, and infaunaJ community composition in Puget Sound.
Mar. Pollut. Bull. 16: 405-415.
Lyon, J.G. 1993. Practical Handbook for Wetland Identification and Delineation. Boca
Raton, FL: CRC Press, Inc., Lewis Publishers.
Manahan, S. 1994. Environmental Chemistry; Sixth Edition. Boca Raton, FL: CRC Press,
Inc., Lewis Publishers.
Maughan, J.T. 1993. Ecological Assessment of Hazardous Waste Sites. New York,
NY: Van Nostrand Remhold.
"McNamara, B.P 1976. Concepts in health evaluation of commercial and industrial
chemicals. In: Mehiman, M.A.; et al. (eds.). Advances in Modern Toxicology, Volume
/, Part 1: New Concepts in Safety Evaluation. New York, NY: John Wiley & Sons.
*McNamara. 1971. [reserved]
Mead, R. 1988. The Design of Experiments. Cambridge, UK: Cambridge University Press.
'Melacon and Lech. 1983. [reserved]
Moltmann. J F . Rombke, J. 1996. Applied Ecotoxicology. Boca Raton, FL: CRC Press,
Inc.. Lewis Publishers.
Morgan, B J. 1993. Analysis of Quantal Response Data. London, UK: Chapman and Hall.
Mudroch. A., Azcue, J.M. 1995. Manual of Aquatic Sediment Sampling. Boca Raton, FL:
CRC Press, Inc., Lewis Publishers.
Mudroch. A; Mac Knight, S.D. (eds.). 1994. Handbook of Techniques for Aquatic Sediment
Sampling; Second Edjtaon. Boca Raton, FL: CRC Press, Inc., Lewis Publishers.
BIBLIOGRAPHY, Page 6
-------
August 21. 1996 DRAFT
Murdoch, A.; MacKnight, S.D. (eds.). 1991. CRC Handbook of Techniques for Aquatic
Sediments Sampling. Boca Raton, FL: CRC Press
National Oceanic and Atmospheric Administration (NOAA). 1987, Guidelines and
Recommendations for Using Bioassessment in the Superfund Remedial Process.
Seattle, WA: Ocean Assessments Division. Prepared by Chnstopherson, S.,
and Field, L.J., National Oceanic and Atmospheric Administration, and Dexter,
R.N., E.V S. Consultants.
National Research Council (NRC). 1994. Science and Judgment in Risk Assessment.
Washington, DC: National Academy Press.
'National Research Council (NRC). 1993. Issues in Risk Assessment. Washington, DC:
National Academy Press.
'National Research Council (NRC). 1983. Risk Assessment in the Federal Government:
Managing the Process. Washington, DC: National Academy Press.
*Nebeker, A.V.; Cairns, M.A.; Wise, C.M. 1984. Relative sensitivity of Chironomus tetans
life stages to copper. Environ. Toxicology and Chemistry 3: 151-158.
Neilson, A.H. 1994. Organic Chemicals in the Aquatic Environment: Distribution,
Persistence, and Toxicity. Boca Raton, FL: CRC Press, Inc., Lewis Publishers.
Newman, M.C. 1995. Quantitative Methods in Aquatic Ecotoxicology. Boca Raton, FL:
CRC Press, Inc., Lewis Publishers.
Newman, M.C.; Mclntosh, A.W. (eds.). 1991. Metal Ecotoxicoiogy: Concepts and
Applications. Boca Raton, FL: CRC Press, Inc., Lewis Publishers.
Oak Rkige National Laboratory (ORNL). 1994. Manual for PC-Data Base Screening
Benchmarks for Ecological Risk Assessment. Environmental Sciences Division, Health
Sciences Research Division. Prepared by Martin Marietta Energy Systems, Inc., for
the U.S. Department of Energy. ORNL/TM-12898.
Oak Ridge National Laboratory (ORNL). 1994. Toxicological Benchmarks for Screening
Contaminants of Potential Concern for Effects on Sediment-Associated Biota: 1994
Revision. Prepared by Hull, R.N.; Suter, G.W., El; Energy Systems Environmental
Restoration Program, ORNL Environmental Restoration Program (managed by Martin
Marietta Energy Systems, Inc., for the U.S. Department of Energy). ES/ER/TM-
95/R1.
BIBLIOGRAPHY, Page 7
-------
August 21. 1996 DRAFT
Oak Rjdge National Laboratory (ORNL). 1994. Toxicological Benchmarks for Screening
Potential Contaminants of Concern for Effects on Aquatic Biota: 1994 Revision.
Prepared by Suter, G.W . II. Mabrey, J.B.; ORNL EnvironmentaJ Sciences Division,
for the U.S. Department of Energy. ES/ER/TM-96/R1.
Oak Ridge National Laboratory (ORNL). 1994. Toxicological Benchmarks for Wildlife:
1994 Revision. Prepared by Opresko, D. M., Sample, B., E., and Suter, G. W. n,
ORNL EnvironmentaJ Sciences Division, for the U.S. Department of Energy.
ES/ER/TM-86/R1.
Oak Ridge National Laboratory (ORNL). 1994. Toxicological Benchmarks for Screening
Potential Contaminants of Concern for Effects on Terrestrial Plants: 1994 Revision.
Prepared by Will, M. E., and Suter, G. W. n, ORNL Environmental Sciences Division,
for the U.S. Department of Energy. ES/ER/TM-85/R1.
Ostrander, G. (ed.). 1996. Handbook of Aquatic Toxicology Methods. Boca Raton, FL:
CRC Press, Inc., Lewis Publishers.
Ott, W.R. 1995. Environmental Statistics and Data Analysis. Boca Raton, FL: CRC Press,
Inc., Lewis Publishers.
Pain, D.J. 1995. Lead in the environment. In: Handbook of Ecotoxicology. pp 356-391.
Parker, S.P. (ed.). 1994. Dictionary of Scientific and Technical Terms; Fifth Edition. New
York. NY: McGraw-Hill.
*Pascoe, D.; Williams, K.A.; Green, D.W.J. 1989. Chronic toxicity of cadmium to
Chironomus ripanous Meigen effects upon larval develecment and adult emergence.
Hydrobiologia 175: 109-115.
Peakall. 1975. [reserved]
'Phillips. D H. 1978. The use of biological indicator organisms fo quantitate
organochlorine pollutants in aquatic environments - a review Environ. Pollut.
16: 167-227.
'Phillips, D H. 1977. The use of biological indicator organisms to monitor trace metal
pollution in marine and estuarine environments - a review Environ. Pollut.
13. 281-317
BIBLIOGRAPHY, Page 8
-------
August 21. 1996 DRAFT
Ramamoorthy, S.; Baddaloo, E.G. 1995. Handbook of Chemical Toxicity Profiles of
Biological Species; Volume 1: Aquatic Species. Boca Raton, FL: CRC Press, Inc.,
Lewis Publishers.
Rand, G.M.; Petrocelli. S.R. 1985. Fundamentals of Aquatic Toxicology. Methods and
Applications. New York, NY: McGraw Hill.
Renzoni, A.; Fossi, M.C.; Lan, L; Mattel, N. (eds.). 1994. Contaminants in the
Environment. A Multidisciplinary Assessment of Risks to Man and Other Organisms.
Boca Raton, FL: CRC Press, Inc., Lewis Publishers.
Ricklefs, R.E. 1990. Ecology; Second Edition. New York, NY: W.H. Freeman.
Siegel, S. 1956. Non-parametric Statistics. New York, NY: McGraw-Hill.
Sokal, R.R.; Rohlf, F.J. 1981. Biometry. Second Edition. New York, NY: W.H.
Freeman.
Sullivan, T.F. 1993. Environmental Regulatory Glossary. Government Institutes, Inc.
Suter, G.W., EL 1993. Ecological Risk Assessment. Ann Arbor, MI: Lewis Publishers.
*Tanabe. 1988. [reserved]
Talmage, S.S.; Walton, B.T. 1991. Small mammals as monitors of environmental
contaminants. Reviews of Environmental Contamination and Toxicology 119: 95.
Trapp, S.: McFarlane, J.C. (eds.). 1995. Plant Contamination: Modeling and Simulation of
Organic Chemical Processes. Boca Raton, FL: CRC Press. Inc., Lewis Publishers.
U.S. Department of the Interior (U.S. DOI). 1991. Plant toxicity testing with sediment and
marsh soils. Technical Report NPS/NRWRD/NRTR-91/03.
U.S. Department of the Interior (U.S. DOD. 1987. Guidance on Use of Habitat Evaluation
Procedures and Suitability Index Models for CERCLA Application. Washington, DC:
U.S. Fish and Wildlife Service, National Ecology Center, PB86-100151.
*U.S. Environmental Protection Agency (U S. EPA). 1996a. Ecotox Thresholds. ECO
Update, Intermittent Bulletin, Volume 3, Number 2. Washington, DC: Office of
Emergency and Remedial Response, Hazardous Site Evaluation Division; Publication
9345.0-12FSI; EPA/540/F-95/038; NTTS PB95-963324.
BIBLIOGRAPHY, Page 9
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1996b. Proposed Guidelines for
Carcinogen Risk Assessment. Washington, DC: Office of Research and Development,
April. EPA/600/P-92/003C.
*U.S. Environmental Protection Agency (U.S. EPA). 1996c. Representative Sampling
Guidance Document, Volume 3; Ecological. Washington, DC: Office of Emergency
and Remedial Response.
U.S. Environmental Protection Agency (U.S. EPA). 1995a. Great Lakes Water Quality
Initiative Criteria Documents for the Protection of Wildlife. Washington, DC: Office
of Water. EPA/820/B-95/008.
*U.S. Environmental Protection Agency (U.S. EPA). 1995b. Ecological Significance and
Selection of Candidate Assessment Endpoints. ECO Update, Intermittent Bulletin,
Volume 3, Number 1. Washington, DC: Office of Emergency and Remedial
Response, Hazardous Site Evaluation Division; Publication 9345.0-11FSI; EPA/540/F-
95/037; NTIS PB95-963323.
*U.S. Environmental Protection Agency (U.S. EPA). 1995c. Land Use in the CERCLA
Remedy Selection Process. May 25 Memorandum from Elliot P. Laws, Assistant
Administrator, to EPA Regional staff. OSWER Directive No. 9355:7-04.
*U.S. Environmental Protection Agency (U.S. EPA). 1995d. Great Lakes Water Quality
Initiative Technical Support Document for the Procedure to Determine
Bioaccumulation Factors. Washington, DC: Office of Water; EPA/820/B-95/005.
U.S. Environmental Protection Agency (U.S. EPA). 1995e. Great Lakes Water Quality
Initiative Criteria Documents for the Protection of Aquatic Life in Ambient Water.
Washington. DC: Office of Water; EPA/820/B-95/004.
*U.S. Environmental Protection Agency (U.S. EPA). 1995f. EPA Risk Characterization
Policy. March 21 Memorandum from Carol Browner, Administrator, to EPA staff.
Washington. DC: Office of the Administrator.
U.S. Environmental Protection Agency (U.S. EPA). 1995g. Technical Support Document
for the Hazardous Waste Identification, Rule: Risk Assessment for Human and
Ecological Receptors, Volume I. Washington, DC: Prepared for the Office of Sold
Waste under Contract No. 68-D2-0065, 68-W3-0028; August.
BIBLIOGRAPHY, Page 10
-------
August 21. 1996 DRAFT
*U.S. Environmental Protection Agency (U.S. EPA). 1995h. Technical Support Document
for the Hazardous Waste Identification Rule: Risk Assessment for Human and
Ecological Receptors, Volume II. Washington, DC: Prepared for the Office of Sold
Waste under Contract No. 68-D2-0065, 68-W3-0028; August.
U.S. Environmental Protection Agency (U.S. EPA). 1995. Ecological Risk: A Primer for
Risk Managers. Washington, DC: EPA/734/R-95/001.
U.S. Environmental Protection Agency (U.S. EPA). 1995. Draft Science Policy Council
Statement on EPA Policy: Cumulative Risk Framework, With a Focus on Improved
Characterization of Risks for Multiple Endpoints, Pathways, Sources, and Stressors.
Washington, DC: Science Policy Council.
*U.S. Environmental Protection Agency (U.S. EPA). 1994a. Memorandum from Carol
Browner, Administrator, to Assistant Administrators concerning "Toward a Place-
Driven Approach: The Edgewater Concensus on an EPA Stategy for Ecosystem
Protection. May 24.
*U.S. Environmental Protection Agency (U.S. EPA). 1994b. Using Toxicity Tests in
Ecological Risk Assessment. ECO Update, Intermittent Bulletin, Volume 2, Number I.
Washington, DC: Office of Emergency and Remedial Response, Hazardous Site
Evaluation Division; Publication 9345.0-051; EPA/540/F-94/012; NTIS PB94-963303.
*U.S. Environmental Protection Agency (U.S. EPA). 1994c. Catalogue of Standard Toxicity
Tests for Ecological Risk Assessment. ECO Update, Intermittent Bulletin, Volume 2,
Number 2. Washington, DC: Office of Emergency and Remedial Response,
Hazardous Site Evaluation Division; Publication 8345.0-051; EPA/540/F-94/013; NTIS
PB94-963304.
*U.S. Environmental Protection Agency (U.S. EPA). 1994d. Field Studies for Ecological
Risk Assessment. ECO Update, Intermittent Bulletin, Volume 2, Number 3.
Washington, DC: Office of Emergency and Remedial Response, Hazardous Site
Evaluation Division; Publ. 9345.0-051; EPA/540/F-94/014, NTTS PB94-963305.
*U.S. Environmental Protection Agency (U.S. EPA). 1994e. Selecting and Using Reference
Information in Superfund Ecological Risk Assessments. ECO Update, Intermittent
Bulletin, Volume 2, Number 4. Washington, DC: Office of Emergency and Remedial
Response, Hazardous Site Evaluation Division; Publication 9345.101; EPA/540/F-
94/050; NTTS PB94-963319.
BIBLIOGRAPHY, Page 11
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1994f. Guidance for the Data Quality
Objectives Process; EPA QA/G-4. Washington, DC: Quality Assurance Management
Staff: Final. September.
*U.S. Environmental Protection Agency (U.S. EPA). 1994g. Establishing Background
U\els. Quick Reference Fact Sheet. Washington, EXT: Office of Solid Waste and
Emergency Response. OSWER Directive 9285.7-F9FS. Publication PB94-963313;
EPA/540/F-94/030.
U.S. Environmental Protection Agency (U.S. EPA). 1994h. Peer Review Workshop Report
on Ecological Risk Assessment Issue Papers. Washington, DC: Office of Research
and Development, Risk Assessment Forum; EPA/630/R-94/008.
U.S. Environmental Protection Agency (U.S. EPA). 19941. A Review of Ecological
Assessment Case Studies from a Risk Assessment Perspective, Volume II. Washington,
DC: Office of Research and Development, Risk Assessment Forum; EPA/630/R-
94/003.
U.S. Environmental Protection Agency (U.S. EPA). 1994J. Ecological Risk Assessment Issue
Papers. Washington, DC: Office of Research and Development, Risk Assessment
Forum; EPA/630/R-94/009.
U.S. Environmental Protection Agency (U.S. EPA). 1994. Methods for Measuring the
Toxiciry and Bioaccumulation of Sediment-associated Contaminants with Freshwater
Invertebrates. Washington, DC: Office of Research and Development. EPA
600/R-94/024.
U.S. Environmental Protection Agency (U.S EPA). 1994. Methods for Measuring the
Toxiciry of Sediment Toxiciry of Sediment-associated Contaminants with Estuarme and
\ianne Amphipods. Washington, DC. Office of Research and Development. EPA
600/R-94/025.
U.S. Environmental Protection Agency (U.S. EPA). 1993a. Wildlife Exposure Factors
Handbook Volume I. Washington, DC: Office of R-search and Development;
EPA/60Q/R-93/187a.
U.S. Environmental Protection Agency (U.S. EPA). 1993b. Wildlife Exposure Factors
Handbook Volume II: Appendix. Washington, DC: Office of Research and
Development; EPA/600/R-93/187b.
BIBLIOGRAPHY, Page 12
-------
August 21. 1996 DRAFT
*U.S. Environmental Protection Agency (U.S. EPA). 1993c. Data Quality Objectives
Process for Superfund. Washington, DC: Office of Emergency and Remedial
Response; Interim Final Guidance; EPA/540/G-93/071.
*U.S. Environmental Protection Agency (U.S. EPA). 1993d. Data Quality Objectives
Process for Superfund. Workbook. Washington, DC; Office of Emergency
and Remedial Response; EPA/540/R-93/078.
*U.S. Environmental Protection Agency (U.S. EPA). 1993e. Wildlife Criteria Portions of the
Proposed Water Quality Guidance for the Great Lakes System. Washington, DC:
Office of Water, EPA/822/R-93/006.
*U.S. Environmental Protection Agency (U.S. EPA). 1993f. Guidance for Planning for
Data Collection in Support of Environmental Decision Making Using the Data
Quality Objectives Process. Interim Final. Quality Assurance Management
Staff; EPA QA/G^.
U.S. Environmental Protection Agency (U.S. EPA). 1993. A Review of Ecological
Assessment Case Studies from a Risk Assessment Perspective. Washington,
DC: Risk Assessment Forum; EPA/630/R-92/005.
U.S. Environmental Protection Agency (U.S. EPA). 1993. Technical Basis for Deriving
Sediment Quality Criteria for Nonionic Organic Contaminants for the
Protection of Benthic Organisms by Using Equilibrium Partitioning.
Washington, DC: Office of Water; EPA/822/R-93/011.
U.S. Environmental Protection Agency (U.S. EPA). 1993. Guidelines for Deriving Site-
Specific Sediment Quality Criteria for the Protection of Benthic Organisms.
Washington. DC: Office of Water; EPA/822/R-93/017.
*U.S. Environmental Protection Agency (U.S. EPA). 1992a. Framework for Ecological Risk
Assessment. Washington, DC: Risk Assessment Forum; EPA/630/R-92/001.
U.S. Environmental Protection Agency (U.S. EPA). 1992b. Developing a Work Scope for
Ecological Assessments. ECO Update, Intermittent Bulletin, Volume I, Number 4.
Washington, DC: Office of Emergency and Remedial Response, Hazardous Site
Evaluation Division; Publ. 9345.0-051.
*U.S. Environmental Protection Agency (U.S. EPA). 1992c. The Role of Natural Resource
Trustees in the Superfund Process. ECO Update, Intermittent Bulletin, Volume 1.
Number 3. Washington, DC: Office of Emergency and Remedial Response,
Hazardous Site Evaluation Division; Publ. 9345.0-051.
BIBLIOGRAPHY, Page 13
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1992d. Briefing the BTAG: Intitial
Description of Semng, History, and Ecology of a Site. ECO Update, Intermittent
Bullenn, Volume 1. Number 5. Washington, EXT: Office of Emergency and Remedial
Response, Hazardous Site Evaluation Division; Publ. 9345.0-051.
'U.S. Environmental Protection Agency (U.S. EPA). 1992e. Draft Report: A Cross-species
Scaling Factor for Carcinogen Risk Assessment Based on Equivalence of mg/kg*3/4
per day; Notice. Federal Register. 57(109): 24152-24173 (June 5).
U.S. Environmental Protection Agency (U.S. EPA). 1992f. Guidance on Risk
Characterization for Risk Managers and Risk Assessors. February 26 Memorandum
from F Henry Habicht II, Deputy Administrator, to EPA Assistant Administrators and
Regional Administrators. Washington, DC: Office of the Deputy Administrator.
U.S. Environmental Protection Agency (U.S. EPA). 1992. Sediment Classification Methods
Compendium. Washington, DC: Office of Water; EPA/823/R-092/006.
U.S. Environmental Protection Agency (U.S. EPA). 1992. Superfund Ecological
Assessment Process Case Studies. Washington, DC: Office of Emergency and
Remedial Response, Hazardous Site Evaluation Division, Toxics Integration
Branch. Prepared by The Cadmus Group, Incorporated.
U.S. Environmental Protection Agency (U.S. EPA). 1992. Peer Review Workshop
Report on a Framework for Ecological Risk Assessment. Washington, DC:
Rask Assessment Forum; EPA/625/3-91/022.
U.S. Environmental Protection Agency (U.S. EPA). 1992. Report on the Ecological Risk
Assessment Guidelines Strategic Planning Workshop. Washington, DC: Risk
Assessment Forum; EPA7630/R-92/002.
U.S. Environmental Protection Agency (U.S. EPA). 1992. Guidelines for Exposure
.Assessment. Federal Register. 57: 22888-22938 (May 29).
U.S. Environmental Protection Agency (U.S. EPA). 1992. Dermal Exposure - Principles
and Applications; Final; Washington, DC: Office of Health and Environmental
Assessment; EPA/600/8-91/01 IB.
U.S. Environmental Protection Agency (U.S. EPA). 1992. Science Advisory Board's
Review- of the Draft Final Exposure Assessment Guidelines (SAB Final Review
Draft, August 1991). Washington, DC: Science Advisory Board;
EPA/SAB/lAQC-92/015.
BIBLIOGRAPHY, Page 14
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1992. Interim Guidance on the
Interpretation and Implementation of Aquatic Life Criteria for Metals.
Washington, DC: Office of Water, Office of Science and Technology, Health
and Ecological Criteria Division.
"U.S. Environmental Protection Agency (U.S. EPA). 1991a. .Risk Assessment Guidance
for Superfund: Volume I Human Health Evaluation Manual (Part B,
Development of Risk-Based Preliminary Remediation Goals), Interim.
Washington, DC; Office of Emergency and Remedial Response; 9285.7-0IB.
*U.S. Environmental Protection Agency (U.S. EPA). 199lb. Risk Assessment Guidance for
Superfund: Volume I Human Health Evaluation Manual (Part C, Risk Evaluation of
Remedial Alternatives), Interim. Washington, DC: Office of Emergency and
Remedial Response; 9285.7-01C.
*U.S. Environmental Protection Agency (U.S. EPA). 1991c. Ecological Assessment of
Superfund Sites: An Overview. ECO Update, Intermittent Bulletin, Volume I, Number
2. Washington, DC: Office of Emergency and Remedial Response, Hazardous Site
Evaluation Division; Publ. 9345.0-051.
*U.S. Environmental Protection Agency (U.S. EPA). 199Id. The Role of BTAGs in
Ecological Assessment. ECO Update, Intermittent Bulletin, Volume I, Number I.
Washington, DC: Office of Emergency and Remedial Response, Hazardous Site
Evaluation Division; Publ. 9345.0-051.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Guidance on Oversight of
Potentially Responsible Party Remedial Investigations and Feasibility Studies, Volume
1. Office of Solid Waste and Emergency Response, Washington, DC. OSWER
Directive 9835.l(c). EPA/540/G-91/OlOa.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Guidance on Oversight of
Potentially Responsible Party Remedial Investigations and Feasibility Studies, Volume
2, Appendices. Office of Solid Waste and Emergency Response, Washington, DC.
OSWER Directive 9835.l(c). EPA/540/G-91/010b.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Methods for Measuring :he
Acute Toxicity of Effluents and Receiving Waters to Freshwater and Marine
Organisms; Fourth Edition; Washington, DC: Office of Research and
Development; EPA/600/4-90/027.
BIBLIOGRAPHY, Page 15
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1991. Technical Support Document for
Water Quality-based Toxics Control. Washington. DC: Office of Water; EPA/505/2-
90/00 1.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Shon-term Methods for
Esnmaung the Chronic Toxicity of Effluents and Receiving Waters to
Freshwater Organisms; Third Edition. Washington, DC: Office of Research
and Development; EPA/600/4-91/002.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Short-term Methods for
Estimating the Chronic Toxicity of Effluents and Receiving Waters to Marine
and Estuanne Organisms; Second Edition. Washington, DC: Office of
Research and Development; EPA/600/4-91/003.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Methods for Aquatic Toxicity
Identification Evaluations: Phase I, Toxicity Characterization Procedures.
Duluth, MN: Office of Research and Development: Environmental Research
laboratory; EPA/600/6-91/003.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Supplemental methods and
status reports for short-term saltwater toxicity tests. G. Mormon and G.
Chapman. ERL contribution No. 1199. Narragansett, RI: Environmental
Research Laboratory.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Assessment and Control of
Bioconcentratable Contaminants in Surface Waters. June 1989 Draft prepared
by EPA's National Effluent Toxicity Assessment Center, Environmental
Research Laboratory - Duluth. MN. Washington, DC: Off.cc of Water
Regulations and Standards; and Cincinnati, OH: Office of Health Effects
Assessment.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Rote of the Baseline Risk
Assessment in Superfund Remedy Selection Decisions. Don R. Clay, Assistant
Administrator, Office of Solid Waste and Emergency Response. Washington,
DC. Office of Solid Waste and Emergency Response Directive 9355.0-30.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Summary Repon on Issues in
Ecological Risk Assessment. Washington, DC: Risk Assessment Forum;
EPA, 625/3-91/018.
BIBLIOGRAPHY, Page 16
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1991. Ecological Exposure and
Effects of Airborne Toxic Chemicals: An Overview. CorvaJlis, OR: Office of
Research and Development, Environmental Research Laboratory; EPA/600/3-
91/001.
U.S. Environmental Protection Agency (U.S. EPA). 1991. Technical Support Document
for Water Quality-based Toxics Control. Washington, DC. Office of Water
Regulations and Standards; EPA/440/4-85/032.
*U.S. Environmental Protection Agency (U.S. EPA). 1990a. Macromvenebrate Field and
Laboratory Methods for Evaluating the Biological Integrity of Surface Waters.
Washington, DC: Office of Water, EPA/600/4-90/030.
U.S. Environmental Protection Agency (U.S. EPA). 1990b. Hazard Ranking System.
Federal Register. 55: 51625-51662 (Dec 14).
U.S. Environmental Protection Agency (U.S. EPA). 1990. Guidance for Data Useability in
Risk Assessment. Washington, DC: Office of Solid Waste and Emergency Response;
EPA/540/G-90/008.
U.S. Environmental Protection Agency (U.S. EPA). 1990. EPA Oversight of Remedial
Designs and Remedial Actions Performed by PRPs. Quick Reference Fact Sheet.
Washington, DC: Office of Emergency and Remedial Response, Hazardous Site
Control Division; Publication 9355.5-01/FS.
U.S. Environmental Protection Agency (U.S. EPA). 1990. Guidance Manual for Evaluation
of Laboratories Performing Aquatic Toxicity Tests. Washington, DC: Office of
Research and Development; EP.A/600/4-90/031.
U.S. Environmental Protection Agency (U.S. EPA). 1990. Biological Criteria, National
Program Guidance for Surface Waters. Washington, DC: Office of Water
Regulations and Standards; EPA/440/5-90/004.
U.S. Environmental Protection Agency (U.S. EPA). 1990. Managing Contaminated
Sediments: EPA Decision-Making Processes. Washington, DC: Sediment
Oversight Technical Committee; EPA/506/6-90/002.
U.S. Environmental Protection Agency (U.S. EPA). 1990. National guidance: wetlands
and nonpoint source control programs. Memorandum from Martha G. Prothro,
Director. Office of Water Regulations and Standards; Washington, DC: Office
of Water (June 18).
BIBLIOGRAPHY, Page 17
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1990. Water Quality Standards for
Wetlands-National Guidance. Washington, DC: Office of Water Regulations
and Standards; EPA/440/S-90/011.
U.S. Environmental Protection Agency (U.S. EPA). 1989a. Risk Assessment Guidance
for Superfund: Volume I Human Health Evaluation Manual, Interim Final.
Washington, DC: Office of Solid Waste, Office of Emergency and Remedial
Response: EPA/540/1-89/002.
U.S. Environmental Protection Agency (U.S. EPA). 1989b. Risk Assessment Guidance
for Superfund: Volume 2 - Environmental Evaluation Manual, Interim Final.
Washington. DC: Office of Solid Waste and Emergency Response;
EPA/540/1-89/001 A.
U.S. Environmental Protection Agency (U.S. EPA). 1989c. Rapid Bioassessment
Protocols for Use in Streams and Rivers: Benthic Macromvertebraies and
Fish. Washington, DC: Office of Water; EPA/444/4-89/001 (Authors:
Plafkin, J.L.; Barbour, M.T.; Porter, K.D.; Gross, S.K.; Hughes, R.M.).
U.S. Environmental Protection Agency (U.S. EPA). 1989. Briefing Report to the EPA
Science Advisory Board on the Equilibrium Partitioning Approach to
Generating Sediment Quality Criteria. Washington, DC: Office of Water
Regulations and Standards; EPA/440/5-89/002.
U.S. Environmental Protection Agency (U.S. EPA). 1989. Survey of State Water Quality
Standards for Wetlands. Washington, DC: Office of Wetlands Protection.
U.S. Environmental Protection Agency (U.S EPA). 1989. Report 10 the Sediment
Criteria Subcommittee: Evaluation of the Apparent Effects Threshold (AET)
Approach for .Assessing Sediment Quality. Washington, DC. Office of the
Administrator. Science Advisory Board; SAB-EETFC-89-027
U.S. Environmental Protection Agency (U S. EPA). 1989. Sediment Classification
Methods Compendium; Final Draft. Washington, DC: Office of Water,
Watershed Protection Division (Juno.
U.S. Environmental Protection Agency (U.S. EPA). 1989. Water Quality Criteria to
Protect Wildlife Resources. Corvallis. OR: Office of Research and
Development, Environmental Research Laboratory; EPA/600/3-89/067
BIBLIOGRAPHY, Pag« 18
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1989. Ecological Assessment of
Hazardous Waste Sites: A Field and Laboratory Reference. Corvallis, OR:
Office of Research and Development, Environmental Research Laboratory;
EPA/600/3-89/013.
U.S. Environmental Protection Agency (U.S. EPA). 1989. Superfund Exposure
Assessment Manual - Technical Appendix: Exposure Analysis of Ecological
Receptors. Athens, GA: Office of Research and Development, Environmental
Research Laboratory (December).
U.S. Environmental Protection Agency (U.S. EPA). 1989. Protocols for Short-Term Toxicity
Screening of Hazardous Waste Sites. Office of Research and Development,
Environmental Research Laboratory; EPA/600/3-88/029.
U.S.' Environmental Protection Agency (U.S. EPA). 1989. Scoping Study of the Effects
of Soil Contamination on Terrestrial Biota. Washington, DC: Office of Toxic
Substances.
*U.S. Environmental Protection Agency (U.S. EPA). 1988a. Guidance for Conducting
Remedial Investigations and Feasibility Studies Under CERCLA. Washington,
DC: Office of Emergency and Remedial Response; OSWER Directive No.
9355.3-01.
*U.S. Environmental Protection Agency (U.S. EPA). I988b. (original human exposure
factors handbook).
U.S. Environmental Protection Agency (U.S. EPA). 1988. Estimating Toxicity of Industrial
Chemicals to Aquatic Organisms Using Structure Activity Relationships. Office of
Toxic Substances, Washington, DC: EPA/560/6-88/001.
U.S. Environmental Protection Agcucy (U.S. EPA). 1988. CERCLA Compliance with Other
Laws Manual, Part 1. Washington, DC: Office of Emergency and Remedial
Response; OSWER Directive 9234.1-01.
U.S. Environmental Protection Agency (U.S. EPA). 1988. Short-term Methods for
Estimating the Chronic Toxicity of Effluents in Receiving Waters to Marine and
Estuanne Organisms. Cincinnati, OH: Office of Research and Development, Office
of Environmental Monitoring and Support Laboratory; EPA/600/4-87/0928.
BIBLIOGRAPHY, Page 19
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1988. Estimating Toxiciry of
Industrial Chemicals to Aquatic Organisms Using Structure Activity
Relationships. Washington, DC: Office of Toxic Substances; EPA/560/6-
88/001.
U.S. Environmental Protection Agency (U.S. EPA). 1988. Short-term Methods for
Estimating the Chronic Toxiciry of Effluents in Receiving Waters to Marine and
Estuanne Organisms. Cincinnati, OH: Office of Research and Development,
Environmental Monitoring and Support Laboratory; EPA/600/4-87/0928.
U.S. Environmental Protection Agency (U.S. EPA). 1988. Methods for Aquatic Toxiciry
Identification Evaluations: Phase II, Toxicity identification Procedures.
Duluth, MN: Environmental Research Laboratory; EPA/600/3-88/035.
U.S. Environmental Protection Agency (U.S. EPA). 1988. Methods for Aquatic Toxicity
Identification Evaluations: Phase III, Toxicity Confirmation Procedures.
Duluth, MN: Office of Research and Development, Environmental Research
Laboratory; EPA/600/3-88/036.
U.S. Environmental Protection Agency (U.S. EPA). 1988. Superfund Exposure
Assessment Manual. Washington, DC: Office of Solid Waste and Emergency
Response Directive 9285.5-1; EPA/540/1-88/001.
U.S. Environmental Protection Agency (U.S. EPA). 1987. Data Quality Objectives for
Remedial Response Activities: Development Process. Washington, DC: Office of
Solid Waste and Emergency Response, Office of Emergency and Remedial Response
and Office of Waste Programs Enforcement, OSWER Directive 9355.0-7B;
EPA/540/G-87/003.
U.S. Environmental Protection Agency (U.S EPA). 1987a. Data Quality Objectives for
Remedial Response Activities: Example Scenario: RJ/FS Activities at a Site with
Contaminated Soils and Ground Water. Washington, DC Office of Solid Waste and
Emergency Response, Office of Emergency and Remedial Response and Office of
Waste Programs Enforcement, OSWTR Directive 9355.0-~B. EPA/540/G-87/004.
U.S. Environmental Protection Agency (U.S. EPA). 1987. Permit Writer's Guide to Water
Quality-Based Permitting for Toxic Pollutants. Washington, DC: Office of
Water Regulations and Standards; EPA/440/4-87/005.
U.S. Environmental Protection Agency (L'S EPA). 1987. Guidelines for Deriving
Ambient Aquatic Life Advisory Concentrations. Washington, DC: Office of
Water Regulations and Standards (unpublished).
BIBLIOGRAPHY, Page 20
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1987. A Compendium of Superfund
Field Operations Methods. Washington DC: Office of Solid Waste and
Emergency Response, Office of Environmental and Remedial Response;
EPA/540/P-87/001.
U.S. Environmental Protection Agency (U.S. EPA). 1987. Role of Acute Toxicity
Bioassays in the Remedial Action Process at Hazardous Waste Sites. Corvallis,
OR: Office of Research and Development, Environmental Research
Laboratory; EPA/600/8-87/044.
U.S. Environmental Protection Agency (U.S. EPA). 1987. Ecological Risk Assessment
in the Office of Toxic Substances: Problems and Progress 1984-1987.
Washington, DC: Office of Toxic Substances, Health and Environmental
Review Division (Author. Rodier, D.)
*U.S. Environmental Protection Agency (U.S. EPA). 1986a. Guidelines for the Health Risk
Assessment of Chemical Mixtures. Washington, DC: Office of Health and
Environmental Assessment; EPA/600/8-87/045.
U.S. Environmental Protection Agency (U.S. EPA). 1986. Engineering Support Branch,
Standard Operating Procedures and Quality Assurance Manual. Region IV,
Environmental Services Division.
U.S. Environmental Protection Agency (U.S. EPA). 1986. Guidelines for Deriving
Numerical Criteria for the Protection of Aquatic Organisms and Their Uses.
Washington, DC: Office of Water Regulations and Standards.
U.S. Environmental Protection Agency (U.S. EPA). 1986. Qualm Criteria for Water 1986.
Washington, DC: Office of Water Regulations and Standards; EPA/440/5-86/001.
*U.S. Environmental Protection Agency (U.S. EPA). 1985a. Ambient Water Quality Criteria
for Copper-1984. Washington, DC: Office of Water, Regulations and Standards,
Catena and Standards Division. EPA/440/5-84-031. PB85-227023.
U.S. Environmental Protection Agency (U.S. EPA). 1985. Development of Statistical
Distributions of Ranges of Standard Factors Used in Exposure Assessments.
Washington, DC: Office of Health and Environmental Assessment, OHEA-E-
161: EPA/600/8-85/010.
U.S. Environmental Protection Agency (U.S. EPA). 1985. Guide for Identifying Cleanup
Alternatives at Hazardous Waste Sites and Spills. Washington, DC: Office of
Solid Waste and Emergency Response; EPA/600/3-83/063, NTIS PB86-144664.
BIBLIOGRAPHY. Page 21
-------
August 21. 1996 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1985. Methods for Measuring the
Acute Toxiciry of Effluents to Freshwater and Marine Organisms. Cincinnati,
OH' Office of Research and Development, EnvironmentaJ Monitoring and
Support Laboratory; EPA/600/4-85/013.
U.S. Environmental Protection Agency (U.S. EPA). 1985. Shon-term Methods for
Estimating the Chronic Toxicity of Effluents in Receiving Waters to Freshwater
Organisms. Cincinnati, OH: Office of Research and Development,
Environmental Monitoring and Support Laboratory; EPA/600/4-85/014.
U.S. Environmental Protection Agency (U.S. EPA). 1984a, Risk Assessment and
Management: Framework for Decision Making. Washington, DC: Office of
Policy, Planning, and Evaluation; EPA/600/9-85/002.
U.S. Environmental Protection Agency (U.S. EPA). 1984. Estimating "Concern Levels"
for Concentrations of Chemical Substances in the Environment. Washington,
DC: Office of Toxic Substances, EnvironmentaJ Effects Branch.
U.S. EnvironmentaJ Protection Agency (U.S. EPA). 1984. Technical Support Manual:
Waterbody Surveys and Assessments for Conducting Use Attainability Analyses:
Volume II: Estuarine Systems. Washington, DC: Office of Water Regulations
and Standards.
U.S. Environmental Protection Agency (U.S. EPA). 1984. Technical Support Manual:
Waterbody Surveys and Assessments for Conducting Use Attainability Analyses:
Volume III: Lake Systems. Washington, DC: Office of Water Regulations and
Standards.
U.S. Environmental Protection Agency (U.S EPA). 1983. Technical Support Manual:
Waterbody Surveys and Assessments for Conducting Use Attainability Analyses.
Washington, DC: Office of Water Regulations and Standards (November).
U.S. Environmental Protection Agency (U.S. EPA). 1983. Environmental Effects of
Regulates Concern Under TSCA \ Pisinon Paper. '-Visrungton, DC:
Office of Toxic Substances, Health and Environmental Review Division
(Author Qements, R.G.)
U.S. Environmental Protection Agency (U S EPA) and Department of the Army, U.S. Army
Corps of Engineers (US ACE). 1994b. Evaluation of Dredged Material Proposed for
Discharge in Waters of the U.S. - Testing Manual (Draft I: Inland Testing Manual.
Washington. DC: EPA Office of Water. EPA/823/B-94/002.
BIBLIOGRAPHY, Page 22
-------
August 21. 1996 . DRAFT
Watras. C.J.; Huckabee, J.W. (eds.). 1995. Mercury Pollution: Integration and Synthesis.
Boca Raton. FL: CRC Press, Inc., Lewis Publishers.
*Weil, C.S.; McCollister, D.D. 1963. Relationship between short- and long-term
feeding studies in designing an effective toxicity test. Agr. Food Chem. 11:
486-491.
Wentsel, R.S.; LaPoint, T.W.; Simini, M.; Checkai. R.T.; Ludwig. D.; Brewer, L. 1994.
Procedural Guidelines for Ecological Risk Assessments at U.S. Army Sites, Volume I.
Aberdeen Proving Ground, MD: Edgewood Research, Development, and Engineering
Center, U.S. Army Chemical and Biological Defense Command. Rept. No. ERDEC-
TR-221.
BIBLIOGRAPHY, Page 23
-------
APPENDIX A
EXAMPLE ECOLOGICAL RISK ASSESSMENTS
FOR HYPOTHETICAL SITES
-------
INTRODUCTION
Appendix A provides examples of Steps 1 through 5 of the ecological risk assessment
process for three hypothetical sites:
(1) A former municipal landfill from which copper is leaching into a large pond
down-gradient of the site (the copper site);
(2) A former chemical production facility that spilled DDT, which has been
transported into a nearby stream by surface water runoff (the DDT site); and
(3) A former waste-oil recycling facility that disposed of PCBs in a lagoon from
which extensive soil contamination has resulted (the PCS site).
These examples are intended to illustrate key points in Steps 1 through 5 of the ecological
risk assessment process. No actual site is the basis for the examples.
The examples stop with Step 5 because the remaining steps (6 through 8) of the
ecological risk assessment process and the risk management decisions depend on site-specific
data collected during a site investigation. We have not attempted to develop hypothetical data
for analysis or the full range of information that a site risk manager would consider when
evaluating remedial options.
-------
August 21. 1996 DRAFT
EXAMPLE 1: COPPER SITE
STEP 1: SCREENING-LEVEL PROBLEM FORMULATION AND ECOLOGICAL
EFFECTS EVALUATION
Site history. This is a former municipal landfill located in an upland area of the
mid-Atlantic plain. Residential, commercial, and industrial refuse was disposed of at this site
in the 1960s and 1970s. Large amounts of copper wire also were disposed at this site over
several years. Currently, minimal cover has been placed over the fill and planted with
grasses. Terrestrial ecosystems in the vicinity of the landfill include upland forest and
successional fields. Nearby land uses include agriculture and residential and commercial uses.
The landfill cover has deteriorated in several locations. Leachate seeps have been noted on
the slope of the landfill, and several seeps discharge to a five-acre pond down-gradient of the
site.
Site visit. A preliminary site visit was conducted and the ecological checklist was
completed. The checklist indicated that the pond has an organic substrate; emergent
vegetation, including cattail and Phragmites, occurs along the shore near the leachate seeps;
and the pond reaches a depth of five feet toward the middle. Fathead minnows, carp, and
several species of sunfish were observed, and the benthic macromvertebrate community
appeared to be diverse. The pond water was clear, indicating an absence of phytoplankton.
The pond appears to function as a valuable habitat for fish and other wildlife using this area.
Preliminary sampling indicated elevated copper levels in the seep as well as elevated base
cations, total organic carbon (TOC), and depressed pH levels (pH 5.7).
Problem formulation. Copper is leaching from the landfill into the pond from a
seep area. EPA's ambient water quality criteria document for copper (U.S. EPA, 1985)
indicates that it can cause toxic effects in aquatic plants, aquatic invertebrates, and young fish
at relatively low water concentrations. Thus, the seep might threaten the ability of the pond
to support macromvertebrate and fish communities and the wildlife that feed on them.
Terrestrial ecosystems do not need to be evaluated because the overland flow of the seeps is
limited to short gullies, a few inches wide. Thus, the area of concern has been identified as
the five-acre pond and the associated leachate seeps. Copper in surface water and sediments
of the pond might be of ecological concern.
Ecological effects evaluation. Copper is toxic to both aquatic plants and aquatic
animals. Therefore, aquatic toxicity-based data will be used to screen for ecological risk in
the preliminary nsk calculation. The screening ecotoxicity value selected for water-column
exposure is the U.S. EPA chronic ambient water quality criterion (12 ug/L at a water hardness
of 100 mg/L as CaCO3). The screening ecotoxicity value for copper in sediments was
identified as 34 mg/kg (U.S. EPA, 1996).
APPENDIX A, Page 1
-------
August 21. 1996 DRAFT
STEP 2: SCREENING-LEVEL EXPOSURE ESTIMATE AND RISK CALCULATION
Exposure estimate. Preliminary sampling data indicate that the leachate contains
53 ug/L copper as well as elevated base cauons, total organic carbon (TOO, and depressed
pH (pH 5.7V Sediment concentrations range from 300 mg/kg to below detection (2 mg/kg),
decreasing with distance from the leachate seeps.
Risk calculation. The copper concentration in the seep water (53 ug/L) exceeds the
chronic water quality criterion for copper (12 ug/L). The maximum sediment copper
concentration of 300 mg/kg exceeds the screening ecotoxicity value for copper in sediments
(34 mg/kg). Therefore, the screening-level hazard quotients for both sediment and water
exceed one. The decision at the Scientific/Management Decision Point (SMDP) is to continue
the ecological risk assessment.
Similar screening for the levels of base cations generated hazard quotients below 1 in
the seep water. Although TOC and pH are not regulated under CERCLA, the possibility that
these parameters might affect the biota of the pond should be kept in mind if surveys of the
pond biota are conducted. Sediment concentrations of chemicals other than copper generated
hazard quotients (HQs) of less than 1 at the maximum concentration found.
STEP 3: BASEUNE RISK ASSESSMENT PROBLEM FORMULATION
Based on the screening-level risk assessment, copper is known to be the only
contaminant of ecological concern at the site.
Ecotoxicity literature review. A review of the literature on the ecotoxicity of
copper to aquauc biota was conducted and revealed several types of information. Young
aquatic organisms are more sensitive to copper than adults (Demayo et al., 1982; Kaplan and
Yon, 1961; Hubschman, 1965). Fish larvae usually are more sensitive than embryos (McKim
et al., 1978; Weis and Weis, 1991), and fish become less sensitive to copper as body weight
increases (Demayo et al., 1982). Although the exact mechanism of toxicity to fish is
unknown, a loss of osmotic control has been noted in some studies (Demayo et al. 1982;
Cheng and Sullivan, 1977).
Flowthrough toxicity studies in which copper concentrations were measured revealed
LC50 values ranging from 75 to 790 ug/L for fathead minnows and 63 to 800 ug/L for
common carp (U.S. EPA, 1985). Coldwater fish species, such as rainbow trout, can be more
sensitive, and species like pumpkinsecds (a sunfish) and blue gills are less sensitive (U.S.
EPA, 1985). Although fish fry usually are the most sensitive life stage, this is not always the
case; Pickering et al. (1977) determined an LC50 of 460 ug/L to 6-month-old juveniles and
an LC50 of 490 ug/L to 6-week-old fry for fathead minnows. A copper concentration in
APPENDIX A, Page 2
-------
August :i. 1996 _____ DRAFT
water of 37 ug/1 has been shown to cause a significant reduction in fish egg production
(Pickering et al.. 1977).
Elevated levels of copper in sediments have been associated with changes in benthic
community structure, notably reduced numbers of species (Winner et a]., 1975; Kraft and
Sypmewski, 1981). Studies also have been conducted with adult Hyalella azteca (an
amphipod) exposed to copper in sediments. One of these studies indicated an LC50 of 1,078
mg/kg in the sediment (Cairns et al., 1984); however, a no-observed-adverse-effect level
(NOAEL) for copper in sediments was not identified for an early life stage of a benthic
invertebrate.
A literature review of the ecotoxicity of copper to aquatic plants, both algae and
vascular plants, did not reveal information on the toxic mechanism by which copper affects
plants. The review did indicate that exposure of plants to high copper levels inhibits
photosynthesis and growth (U.S. EPA, 1985), cell separation after cell division (Hatch, 1978),
and iron uptake (reference). Several studies conducted using Selenastnun capricornutwn
indicated that concentrations at 300 ug/L kill algae after 7 days, and a value of 90 ug/1 causes
complete growth inhibition after 7 days (Harriett et al., 1974).
The literature indicates that copper does not biomagnify in food chains and does not
bioaccumuiate in most animals because it is a biologically regulated essential element.
Accumulation in phytoplankton and filter-feeding mollusks, however, does occur. The
toxicity of copper in water is influenced by water hardness, alkalinity, and pH.
Assessment endpoints and conceptual model. Based on the screening-level
risk assessment and on the ecotoxicity literature review, development of a conceptual model
for the sue is initiated. Copper can be acutely or chronically toxic to organisms in an aquatic
community through direct exposure of the organisms to copper in the water and sediments.
Threats of copper to higher trophic level organisms are unlikely to exceed threats to
organisms at the base of the food chain, because copper is an essential nutrient which is
effectively regulated by most organisms if the exposure is below toxic levels. Aquatic plants
(particularly phytoplankton) and filter-feeding mollusks (e.g., clams), however, are poor at
regulating copper and are likely to be sensitive receptors. In addition, fish fry can be very
sensitive to copper in water.
Based on these receptors and the potential for both acute and chronic toxjcity, an
appropriate general assessment endpomt for the ecosystem would be the maintenance of the
community composition of the pond. A more operational definition of the assessment
endpomt would be the maintenance of pond community structure typical for the locality and
for the physical attributes of the pond, with no loss of species or community alteration due to
copper toxicity.
APPENDIX A, Page 3
-------
August :i. 1996 DRAFT
A flow diagram was added to the conceptual model to depict the environmental
pathways that could result in impacts of copper to the pono s biota (see Exhibit A-l). Direct
exposure to copper in the pond water and sediments could cause acute or chronic toxicity in
early life stages of fish and/or benthic invertebrates, and in aquatic plants. Risks to filter-
feeding mollusks and phytoplankton as well as animals that feed on them are not considered
because the mollusks and phytoplankton are unlikely to occur in significant quantities in the
pond. The exposure pathways that will be evaluated, therefore, are direct contact with
contaminated sediments and water.
Hypothesis formulation. The testable hypothesis is that concentrauons of copper
present in the sediments and water over at least part of the pond are toxic to aquauc plants or
animals. A further question is at what copper concentration in sediments do adverse effects
become detectable?
STEP 4: MEASUREMENT ENDPOINTS AND STUDY DESIGN
To answer the hypothesis identified in Step 3, three lines of evidence were considered
when selecting measurement endpoints: (1) whether the ambient copper levels are higher
than levels known to be directly toxic to aquauc organisms likely or known to be present in
the pond; (2) whether water and sediments taken from the pond are more toxic to aquatic
organisms than water and sediments from a reference pond; and (3) whether the aquatic
community structure in the site pond is simplified relative to a reference pond.
Measurement endpoints. Since the identified assessment endpoint is maintaining
a typical pond community structure, the possibility of directly measuring the condition of the
plant, fish, and macroinvertebrate communities in the pond was considered. Consultation with
experts on benthic macroinvertebrates suggested that standard measures of the pond benthic
invertebrate community probably would be insensitive measures of existing effects at this
particular site because of the high spatial variation in benthic communities within and among
ponds of this size. Measuring the fish community also would be unsuitable, due to the
limited size of the pond and low diversity of fish species anticipated. Since copper is not
expected to bioaccumulate or biomagmfy in this pond, direct toxiciry testing was selected as
an appropriate measurement endpoint. Because early life stages ;end to be mere sensitive to
the toxic effects of copper than older Life stages, chronic toxiciry would be measured on early
Life stages. For animals, toxicity is defined as a statistically significant decrease in survival or
juvenile growth rates of a population exposed to water or sediments from the site compared
with a population exposed to water or sediments from a reference site. For plants, toxicity is
defined as a statistically significant decrease in growth rate \vith the same comparison.
APPENDIX A, Page 4
-------
August 21. 1996
DRAI-T
EXHIBIT A-1
Conceptual Model for the Copper Site
MEASUREMENT I NDIVMNI
(Sediment loxkily to Hyaltllu aileca)
PRIMARY SOURCE
(Landfill)
SECONDARY SOURCE
((iroundwaler seer.)
TERTIARY SOURCE
(Sediment, exposure point
for aquatic receptors)
TERTIARY SOURCE
(Surface water, exposure
point for aquatic receptors)
AQUATIC RECEPTOR
H
AQUATIC RECEPTOR
MEASUREMENI ENDPOIN I
(Surface water toxkily to Stlenairum
capricornalum and Pimtphales promelas)
APPENDIX A. Page 5
-------
August 21. 1996 DRAFT
One toxicity test selected is a 10-day (i.e., chronic) solid-phase sediment toxicity test
using an early life stage of Hyalella azieca. The measures of effects for the test are mortality
rates and growth rates (measured as length and weight increases). Two water-column toxicity
tests will be used: (1) a 7-day test using the alga Selenastrum capncornutum (growth test)
and (2) a 7-day larval fish test using Pimephales promelas (mortality and growth endpomts).
The H. azieca and P promelas toxicity tests will be used to determine the effects of copper
on early life stages of invertebrates and fish in sediment and the water column, respectively.
The test on 5. capncornutum will be used to determine the phytotoxiciry of copper in the
water column.
Study design. To answer the questions stated in the problem formulation step, the
water column tests will be run on 100 percent seep water, 100 percent pond water near the
seep, 100 percent reference-site water, and the laboratory control. U.S. EPA test protocols
will be followed. Five sediment samples will be collected from the pond bottom at intervals
along the observed concentration gradient, from a copper concentration of 300 mg/kg at the
leachate seeps down to approximately 5 mg/kg near the other end of the pond. The sediment
sampling locations will transect the pond at equidistant locations and include the point of
maximum pond depth. All sediment samples will be split so that copper concentrations can
be measured in sediments from each sampling location. A reference sediment will be
collected and a laboratory control will be run. Test organisms will not be fed during the test;
sediments will be sieved to remove native organisms and debris. Laboratory procedures will
follow established protocols and will be documented and reviewed prior to initiation of the
test. For the water-column test, statistical comparisons will be made between responses to
each of the two pond samples and the reference site, as well as the laboratory control.
Statistical comparisons also will be made of responses to sediments taken from each sampling
location and responses to the reference sediment sample.
Because leachate seeps can be intermittent (depending on rainfall), the srudy design
specifies that a prc-sampling visit is required to confirm that the seep is flowing and can be
sampled. The srudy design also specifies that both sediments and water will be sampled at
the same time at each sampling location.
As the work plan (WP) and sampling and analysis plan (SAP) were finished, the
ecological nsk assessor and the nsk manager agreed on the sue concepmal model, assessment
endpomts. and srudy design (SMDP).
STEP 5: FIELD VERIFICATION OF STUDY DESIGN
A site assessment was conducted two days prior to the scheduled initiation of the site
investigation to confirm that the seep was active. It was determined that the seep was active
and thai the site investigation could be initiated.
APPENDIX A, Page 6
-------
August 21. 1996 DRAFT
REFERENCES
Banlett. L., Rabe, F W.. Funk, W.H. 1974. Effects of copper, zinc, and cadmium on
Seienastrum capricornutum. Water Res. 8: 179-185.
Cairns, M.A.; Nebeker, A.V.; Gakstatter, J.H.; Griffis, W.L. 1984. Toxicity of copper-spiked
sediments to freshwater invertebrates. Environ. Toxicol. Chem. 3: 345-445.
Cheng, T.C.; Sullivan, J.T. 1977. Alterations in the osmoregulation of the pulmonate
gastropod Biomphalaria glabrata due to copper. J. Invert. Path. 28: 101.
Hatch, R.C. 1978. Poisons causing respiratory insufficiency. In: L.M. Jones, N.H. Booth
and L.E. McDonald (eds.), Veterinary Pharmacology and Therapeutics. Iowa State
University, IA: Ames Press.
Hubschman, J.H. 1965. Effects of copper on the crayfish Orconectes rusticus (Girard). I.
Acute toxicity. Crustaceana 12: 33-42.
Kaplan, H.M.; Yoh, L. 1961. Toxicity of copper to frogs. Herpetologia 17: 131-135.
Kraft, K.J.; Sypmewski, R.H. 1981. Effect of sediment copper on the distribution of benthic
macroinvertebrates in the Keweenaw Waterway. J. Great Lakes Res. 7: 258-263.
McKim, J.M.: Eaton, J.G.; Holcombe, G.W. 1978. Metal toxicity to embryos and larvae of
eight species of freshwater fish. II. Copper. Bull. Environ. Contain. Toxicol. 19:
608-616.
Pickering, Q.; Brungs, W.; Gast, M. 1977 Effect of exposure time and copper concentration
of fathead minnows, Pimephales promelas (Rafinesque). Aquatic Toxicol. 12: 107.
U.S. Environmental Protection Agency CU.S. EPA). 1996. Ecotox Thresholds. ECO Update,
Intermittent Bulletin, Volume 3, Number 2. Washington, DC: Office of Emergency
and Remedial Response, Hazardous Site Evaluation Division; Publication 9345.0-
12FSI; EPA/540/F-95/D38; NTIS PB95-963324.
U.S. Environmental Protection Agency (U.S. EPA). 1985. Ambient Water Quality Criteria
for Copper. Washington, DC: Office of Water; EPA/440/5-84/031.
Weis, P.; Weis, J.S. 1991. The developmental toxicity of metals and metalloids in fish. In:
Newman, M.C.; Mclntosh, A.W. (eds.), Metal Ecotoxicology: Concepts and
Applications. Boca Raton, FL: CRC Press, Inc., Lewis Publishers.
APPENDIX A, Page 7
-------
August!]. 1990 PR APT
Winner. R.W , KelJmg, T.; Yeager. R.; et al. 1975. Response of a macroinvertebrate fauna
to a copper gradient in an expenmentaJly-polluted stream. Verb. Int. Ver. Limnol. 19:
2121-2127.
APPENDIX A. Page 8
-------
August :i. 1996 DRAFT
EXAMPLE 2: DDT SITE
STEP 1: SCREENING-LEVEL PROBLEM FORMULATION AND ECOLOGICAL
EFFECTS EVALUATION
Site history. This is the site of a former chemical production facility located
adjacent to a stream.. The facility manufactured and packaged dichlorodiphenyltrichloroethane
(DDT). Due to poor storage practices, several DDT spills have occurred.
Site visit A preliminary site visit was conducted and the ecological checklist was
completed. Information gathered indicates that surface water drainage from the site flows
through several drainage swales toward an unnamed creek. This creek is a second-order
stream containing riffle-run areas and small pools. The stream substrate is composed of sand
and gravel in the pools with some depositional areas in the backwaters, and primarily cobble
in the riffles.
Problem formulation. Previous sampling efforts indicated the presence of DDT
and its metabolites in the stream's sediments over several miles at a concentrations up to 230
mg/kg. A variety of wildlife, especially piscivorous birds, use this area for feeding. Many
species of minnow have been noted in this stream. DDT is well known for its tendency to
bioaccumulate and biomagnify in food chains, and available evidence indicates that it can
cause reproductive failure in birds due to eggshell thinning.
The risk assessor and risk manager agreed that the assessment endpoint is adverse
effects on reproduction of high-trophic-level wildlife, particularly piscivorous birds.
Ecological effects evaluation. Because DDT is well studied, a dietary
concentration above which eggshell thinning might occur was identified in existing U.S. EPA
documents on the ecotoxicity of DDT. Moreover, a no-observed-adverse-effect-level
(NOAEL) for the ingestion route for birds also was identified.
STEP 2: SCREENING-LEVEL EXPOSURE ESTIMATE AND RISK CALCULATION
Exposure estimate. For the screening-level exposure estimate, maximum
concentrations of DDT identified in the sediments were used. To estimate the concentration
of DDT m forage fish, the maximum concentration in sediments was multiplied by the
highest DDT bioaccumulation factor relating forage fish tissue concentrations to sediment
concentrations reported in the literature. Moreover, it was assumed that the piscivorous birds
obtain 100 percent of their diet from the contaminated area.
APPENDIX A. Page 9
-------
August 21. !996 DRAFT
Risk calculation. The predicted concentrations of DDT in forage fish were
compared with the dietary NOAEL for DDT in birds. This nsk screen indicated that DDT
concentrations measured at this site might be high enough fo cause adverse reproductive
effects in birds. Thus, transfer of DDT from the sediments to the stream and biota are of
concern at thus sue.
STEP 3: BASELINE RISK ASSESSMENT PROBLEM FORMULATION
Based on the screening-level risk assessment, potential bioaccumulation of DDT in
aquatic food chains and effects of DDT on reproduction in piscivorous birds are known
concerns. Dunng refinement of the problem, the potential for additional ecological effects of
DDT was examined.
Ecotoxicrty literature review. In freshwater systems,M2DT can have direct effects
on animals, particularly insects. A literature review of the aquatic toxicity of DDT was
conducted, and a NOAEL and LOAEL identified for the toxicity of DDT to aquatic insects.
Aquatic plants are not affected by DDT. Additional quantitative information on effects of
DDT on birds was reviewed, particularly to identify what level of eggshell thinning is likely
to reduce reproductive success. A number of studies have correlated DDT residues measured
in eggs of birds to increased eggshell thinning and egg loss due to breakage. Eggshell
thinning of more than 20 percent appears to result in decreased hatching success due to
eggshell breakage (Anderson and Hickey, 1972; Dilworth et al., 1972). Information was not
available for any piscivorous species of bird. Lincer (1975) conducted a laboratory feeding
study using American kestrels. Females fed a diet of 6 mg/kg DDE1 (1.1 mg/kgBW-day)
produced eggs with shells which were 25.5 percent thinner than archived eggshells collected
prior to widespread use of DDT. Based on this information, a LOAEL of 1.1 mg/kgBW-day
was selected to evaluate the effects of DDT on piscivorous birds.
Assessment endpoints and conceptual model. Based on knowledge of the
fate and transport of DDT in aquatic systems and the ecotoxiciry of DDT to aquatic
organisms and birds, a conceptual model was initiated. DDT buned in the sediments can be
released to the water column during resuspension and redistribution of the sediments. Some
diffusion of DDT to the water column from the sediment surface aJso will occur. The benthic
community would be an initial receptor for the DDT in sediments, which could result in
reduced benthic species abundance and DDT accumulation in species that remain. Fish that
feed on benthic organisms might be exposed to DDT both in the water column and in their
food. Piscivorous birds would be exposed to the DDT that has accumulated in the fish, and
could be exposed at levels sufficiently high to cause more than 20 percent eggshell thinning.
DDE is a Jegradauon product of DDT. typically, field measures of DDT arc reported as the sum of the
conccntrauons ?f DDT, DDE. and DDD lanoiher degradation product).
APPENDIX A. Page 10
-------
August 21. 1996 DRAFT
EXAMPLE 3: PCB SITE
STEP 1: SCREENING-LEVEL PROBLEM FORMULATION AND ECOLOGICAL
EFFECTS EVALUATION
Site history. This is a former waste-oil recycling facility located in a remote area.
Oils contaminated with polychlorinated biphenyl compounds (PCBs) were disposed of in a
lagoon. The lagoon was not lined, and the soil was composed mostly of sand. Oils
contaminated with PCBs migrated through the soil and contaminated a wide area adjacent to
the site.
Site visit. During the preliminary site visit, the ecological checklist was completed.
Most of the habitat is upland forest, old field, and successional terrestrial areas. Biological
surveys at this site have noted a variety of small mammal signs. In addition, red-tailed hawks
were observed.
Problem formulation. At least 10 acres surrounding the site are known to be
contaminated with PCBs. Some PCBs are reproductive toxins in mammals that, when
ingested, induce (i.e., increase concentrations and activity of) enzyme systems in the liver
(Melancon and Lech, 1983). The enzymes are not specific for PCBs and also will enhance
the degradation of steroid hormones (Peakall, 1975). The impaired reproduction that has been
observed in mammals of several species exposed to PCBs might be caused by PCB-induced
reduction in circulating steroid hormones (Tanabe, 1988). Other effects, such as liver
pathology, also are evident at high exposure levels (Fuller and Hobson, 1986). Given this
information, the screening-level ecological risk assessment should include potential exposure
pathways by which mammals could be exposed to PCBs.
Several possible exposure pathways were evaluated for mammals. PCBs are not
highly volatile, so inhalation of PCBs by animals would not be an important exposure
pathway. PCBs in soils generally are not taken up by most plants, but are accumulated by
soil macroinvertebrates. Thus, herbivores, such as voles and rabbits, would not be exposed to
PCBs in most of their diets; whereas insectivores, such as shrews, or omnivores, such as deer
mic-, could be exposed to accumulated PCBs in their diets. PCBs also arc known to
biomagnify in terrestrial food chains; therefore, the ingestion exposure route needs evaluation,
and shrews and/or deer mice would be appropriate mammalian receptors to evaluate in this
exposure pathway.
Potential reproductive effects on predators that feed on shrews or mice also would be
important to evaluate. The literature indicated that exposure to PCBs through the food chain
could cause reproductive impairment in predatory birds through a similar mechanism as in
APPENDIX A. Page 16
-------
August I!. 1396 DRAFT
U.S. Environmental Protection Agency (U.S. EPA). 1989. Rapid Bioassessment Protocols
for Use in Streams and Rivers: Benthic Macroinvenebrates and Fish. Washington,
DC. Office of Water (Plafkin. J.L.. Barbour, M.T., Poner. K.D., Gross, S.K.. and
Huehes, R.M.. authors); EPA/440/4-89/001.
APPENDIX A, Page 15
-------
August 21. 1996 DRAFT
success in kingfishers due to site contamination is likely, and an HQ of less than one implies
impacts due to site contaminants are unlikely (see text for a description of the limitations of
HQs;.
STEP 5: FIELD VERIFICATION OF STUDY DESIGN
A field assessment was conducted and several small fish collection techniques were
used to determine which technique was the most effective for captunng creek chub at the site.
Collected chub were examined to determine the size range available and to determine if
individuals could be sexed.
Seine netting the areas targeted indicated that the creek chub might not be present in
sufficient numbers to provide the necessary biomass for chemical analyses. Based on these
findings, a contingency plan was agreed to (SMDP), which stated that both the creek chub
and the longnoscd dace (Rhinichthys cataractae) would be collected. If the creek chub were
collected at all locations in sufficient numbers, then these samples would be analyzed and the
dace would be released. If sufficient creek chub could not be collected but sufficient
longnosed dace could, the longnosed dace would be analyzed and the creek chub released. If
neither species could be collected at all locations in sufficient numbers, then a mix of the two
species would be used; however, for any given site only one species would be analyzed. In
addition, at one location, preferably one with high DDT levels in the sediment, sufficient
numbers of approximately 20 gram individuals of both species would be collected to allow
comparison (and calibration) of the accumulation between the two species. If necessary to
meet the analytic chemistry needs, similarly-sized individuals of both sexes of creek chub
would be pooled. Pooling two or more individuals will be necessary for the smaller dace.
Samples will be collected by electro-shocking. Field notes for all samples will include the
number of fish per sample pool, sex, weight, length, presence of parasites or deformities, and
other measures.
REFERENCES
Anderson, D.W.; Rickey, JJ. 1972. Eggshell changes in certain North American birds. In:
Voos, K.H. (ed.), Proceedings: XV International Ornithological Congress. The
Hague, Netherlands; pp. 514-540.
Dilworth, T.G., Keith, J.A.; Pearce, P.A.; Reynolds, L.M. 1972. DDE and eggshell thickness
in New Brunswick woodcock. J. Wild! Manage. 36: 1186-1193.
Lincer, J.L. 1975. DDE-induced eggshell thinning in the American kestrel; a comparison of
the field situation and laboratory results. J. App. Ecol. 12: 781-793.
APPENDIX A, Page 14
-------
August 21. 1996 DRAFT
sediments. Adult creek chub average 10 inches and about 20 grams, allowing for analysis of
individual fish. Creek chub also have small home ranges dunng the spring and summer, and
thus it should be possible to relate DDT levels in the chub to DDT levels in the sediments.
For the assessment endpomt of maintaining stream community stnicrure, the selected
measurement endpomts were several metrics describing the abundance and trophic structure of
the stream benthic macromvertebrate community.
Study design. Creek chub will be collected at several locations with known DDT
concentrations in sediments. The fish will be analyzed for body burdens of DDT, and the
relationship between DDT levels in the sediments and in the creek chub will be established.
The fish DDT concentrations can be used to evaluate the DDT threat to piscivorous birds
feeding on the fish at each location. Using the DDT concentrations measured in fish that
correspond to a LOA£L and NOAEL for adverse effects in birds, the corresponding sediment
contamination levels can be determined. These sediment DDT levels then can be used to
derive a cleanup level that would reduce threats of eggshell thinning to piscivorous birds.
The study design for measuring DDT residue levels in creek chub specifies that 10
creek chub of the same size and sex will be collected at each location. Each creek chub
should be at least 20 grams, so that individuals can be analyzed. In addition, at one location,
QA/QC requirements dictate that an additional 10 fish are collected. In this example, it is
necessary to verify in the field that sufficient numbers of creek chub of the specified size are
present to meet the tissue sampling requirements. In addition, stream conditions must be
evaluated to determine what fish sampling techniques will work best at the targeted locations.
The study design and methods for benthic macroinvertebrate collection followed the
Rapid Bioassessmem Protocol (RBP) manual for level three evaluation (U.S. EPA. 1989).
Benthic macromvertebrate samples were co-located with sampling for fish tissue residue
levels so that one set of co-located water and sediment samples for analytic chemistry could
serve for comparison with both tissue analyses.
The study design also specified that the hazard quotient (HQ) method would be used
to evaluate the effects of DDT on the kingfisher dunng nsk characterization. To determine
the HQ. the estimated daily dose of DDT consumed by the kingfishers is divided by a
LOAEL of 1 1 mgAgBW-day for kestrels. To estimate the DDT dose to the kingfisher, the
DDT concentrations in the chub will be multiplied by the fish ingestion rate for kingfishers
and divided by the body weight of kingfishers. This dose will be adjusted by the area use
factor. The area use factor corresponds to the proportion of the diet of a kingfisher that
would consist of fish from the contaminated area. The area use factor is a function of the
home range size of kingfishers relative to the area of contamination. The adjusted dose will
be compared to the LOAEL. A HQ of greater than one implies that impaired reproductive
APPENDIX A, Page 13
-------
August 21. 1996
DKAIT
EXHIBIT A-2
Conceptual Model for the Stream DDT Site
MEASUREMENT ENOI'OIN I
(DDT concentration in fish tissue.
CHpoiurc point for kingfisher)
SECONDARY RECEPTOR
(Fish)
II KIIAKY Kid I'lOK
(Kingfisher)
PRJMARY SOURCE
(Plant site)
SECONDARY SOURCE
(Surface drainage)
TERTIARY SOURCE
(Stream sediment*, exposure point
for fish and macroinvcrtebralcs)
A *
I'KIMAHY Kl( I I'lOK
(llentliic macroinvenchrates.
poinl. fish)
MEASUKUMENT I NDPOIN I
(Denthic macroinvcrtebrale
community struclurc)
APPENDIX A, Page 12
-------
August 21. 1996 DRAFT
Based on this information, two assessment endpomts were identified: (1) maintaining stream
commumt> structure typical for the scream order and location, and i2) protecting piscivorous
birds lYom eagshell thinning that could result in reduced reproductive success.
A flow diagram of the exposure pathways for DDT was added to the conceptual model
(Exhibit A-I). The diagram identifies the primary, secondary, and tertiary sources of DDT at
the site, as well as the primary, secondary, and ternary types of receptors that could be
exposed.
Hypothesis formulation. Two hypotheses were developed: (1) the stream
community has been affected by the DDT, and (2) food-chain accumulation and transfer of
DDT occurs to the extent that 20 percent or more eggshell thjnmng occurs in piscivorous
birds that use the area.
STEP 4: MEASUREMENT ENDPOINTS AND STUDY DESIGN
Measurement endpoints. For the assessment endpoint of protecting piscivorous
birds from eggshell thinning, the conceptual model indicated that DDT in sediments could
reach piscivorous birds through forage fish. Belted kingfishers are known to feed in the
stream. They also have the smallest home range of the piscivorous birds in the area, which
means that more kingfishers can forage entirely from the contaminated stream area than can
other species of piscivorous birds. Thus, one can conclude that, if the risk assessment shows
no threat of eggshell thinning to the kingfisher, there should be minimal or no threat to other
piscivorous birds that might utilize the site. Eggshell thinning in the belted kingfisher
therefore was selected as the measure of effect.
Data from the literature suggest that DDT can have a bioaccumulation factor in
surface water systems as high as six orders of magnitude (10 1; however, in most aquatic
ecosystems, the actual bioaccumulation of DDT from the environment is lower, often
substantially lower, than 10 Many factors influence the actual accumulation of DDT in the
environment. There is considerable debate over the parameters of any proposed theoretical
bioaccumulation model; therefore, it was decided to measure tissue residue levels in the
forage fish at the site instead of estimating the tissue residue levels in forage fish using a
bioaccumulation factor.
Existing information on the distribution of DDT in the stream indicates that a general
gradient of DDT concentrations exists in the sediments, and five locations could be identified
that corresponded to a range of DDT concentrations in sediments. Based on information
available on fish communities m streams similar to the one in the sue area, creek chub
(Semotilus airomaculatus] were selected to measure exposure levels for kingfishers. Creek
chub feed on bcnthic invertebrates, which are in direct contact with the contaminated
APPENDIX A, Page 11
-------
*"8usi:i. 1996 DRAFT
mammals. The prey of red-tail hawks include voles, deer mice, and various insects. Thus,
this raptor could be at nsk of adverse reproductive outcomes.
Ecological effects evaluation. No-observed-adverse-effect levels (NOAELs) for
the effects of PCBs and other contaminants at the site on mammals, birds, and other biota
were identified in the literature.
STEP 2: SCREENING-LEVEL EXPOSURE ESTIMATE AND RISK CALCULATION
Exposure estimate. For the screening-level risk calculation, the highest PCB and
other contaminant levels measured onsite were used to estimate exposures.
Risk calculation. The potential contaminants of concern were screened based on
NOAELs for exposure routes appropriate to each contaminant. Based on this screen, PCBs
were confirmed to be -the only contaminants of concern to small mammals, and possibly to
birds, based on the levels measured at this site. Thus, at the SMDP, the risk manager and
risk assessor decided to continue to Step 3 of the ecological risk assessment process.
STEP 3: BASELINE RISK ASSESSMENT PROBLEM FORMULATION
The screening-level ecological risk assessment confirmed that PCBs are of concern to
small mammals based on the levels measured at the site and suggested that predatory birds
might be at risk from PCBs that accumulate in some of their mammalian prey.
Ecotoxicity literature review. A literature review was conducted to evaluate
potential reproductive effects in birds. PCBs have been implicated as a cause of reduced
reproductive success of piscivorous birds (e.g., cormorants, terns) in the Great Lakes
(Colborn, 1991). Limited information was available on the effects of PCBs to red-tailed
hawks. A study on American kestrel indicated that consumption of 33 mg/kgBW-day PCbs
resulted in a significant decrease in sperm concentration in male kestrels (Bird et al., 1983).
Implications of this decrease for mating success in kestrels was not evaluated in the study, but
studies on other bird species indicate that it could increase the incidence of infertile eggs and
therefore reduce the number of young fledged per pair. The Great Lakes International Joint
Commission (LJC) recommends 0.1 mg/kg total PCBs as a prey tissue level that will protect
predatory birds and mammals (LJC, 1988). (This number is used as an illustration and not to
suggest that this particular level is appropriate for a given site.)
Assessment endpolnts and conceptual model. Based on the screening-level
risk assessment for small mammals and the results of the ecotoxicity literature search for
birds, a conceptual model was initiated for the site, which included consideration of predatory
APPENDIX A, Page 17
-------
August 21. 1996 DRAFT
birds (e.g., red-tailed hawlcs) and their prey. The ecological nsk assessor and the risk
manager agreed (SMDP) that the general assessment endpomt for the site would be the
protection of predatory birds from reproductive impairment caused by PCBs that had
accumulated in their prey. They identified an operational assessment endpoint as reproductive
impairment in red-tailed hawlcs.
An exposure pathway diagram was developed for the conceptual model to identify the
exposure pathways by red-tailed hawks could be exposed to PCBs originating in the soil at
the site (see Exhibit A-3). While voles may be prevalent at the site, they are not part of the
exposure pathway for predators because they are herbivorous and PCBs do not accumulate in
plants. Deer mice (Peromyscus maniculatus), on the other hand, also are abundant at the site
and, being omnivorous, arc likely to be exposed to PCBs that have accumulated in the insect
component of their diet.
Hypothesis formulation. Based on the assessment endpoint and conceptual model,
the testable hypothesis is that concentrations of PCBs in the prey of the red-tailed hawk
exceed levels known to impair reproduction in predatory birds.
STEP 4: MEASUREMENT ENDPOINTS AND STUDY DESIGN
Measurement endpoints. To test the hypothesis that PCB levels in prey of the
red-tailed hawk exceed levels that might impair their reproduction, PCB levels will be
measured in deer mice taken from the site (of all of the species in the diet of the red-tailed
hawk, deer mice are assumed to accumulate the highest levels of PCBs). The measures of
exposure will be compared with measures of effect reported in the literature.
Study design. The available measures of PCB concentrations in soil at the site
indicated a gradient of decreasing PCB concentration with increasing distance from the
unlmed lagoon. Three locations along this gradient were selected to measure PCB
concentrations in prey. The study design specified that eight deer mice of the same size and
sex will be collected at each location. Each mouse should be approximately 20 grams so that
contaminant levels can be measured in individual mice. With concentrations measured in
eight individual mice, it is possible to estimate a mean concentration and an upper confidence
limit of the mean concentration in deer mice for the location. In addition, QA/QC
requirements dictate that an additional eight deer mice should be collected at one location.
For this site, it is necessary to verify that sufficient numbers of deer mice of the
specified size are present to meet the sampling requirements. In addition, habitat conditions
must be evaluated to determine what trapping techniques will work at the targeted locations.
APPENDIX A, Page 18
-------
August 21. 1996
DRAIT
EXHIBIT A-3
Conceptual Model for the Terrestrial PCB Site
MI-ASURI:MI:NI i NDI-OINI
(PCIIs in iiwmse (issue,
cxsposufc point for red UilcJ hawk]
PRIMARY SOIWCi:
(Waste lagoons)
SI-CONDARY SOURCE
(Sue soils)
PRIMARY RUCI'PTOR
(Deer IIHIUSC)
SECONDARY Rl CM'IOR
(Red tailed hawk)
APPENDIX A. Page 19
-------
August 21. 1996 DRAFT
The study design specifies further that the hazard quotient (HQ) method will be used to
estimate the nsk of reproductive impairment m the red-tailed hawk from exposure to PCBs in their
prey To determine the HQ, the measured DDT concentrations ;n deer mice will be divided by the
LOAEL of 33 mg/kgBW-day for a decrease in sperm concentration in kestrels. To estimate the dose
to the red-tailed hawk, the PCB concentrations in deer mice will be multiplied by the quantity of deer
mice that could be ingested by a red-tailed hawk each day and divided by the body weight of the
hawk. This dose will be adjusted by a factor that corresponds to the proportion of the diet of a red-
tailed hawk that would come from the contaminated area. This area use factor is a function of the
home range size of the hawks relative to the area of contamination. A HQ of greater than one implies
that impacts due to site contamination are likely, and an HQ of less than one implies impacts due to
site contaminants are unlikely.
STEP 5: FIELD VERIFICATION OF STUDY DESIGN
. A field assessment using several trapping techniques was conducted to determine (1)
which technique was most effective for capturing deer mice at the site and (2) whether the
technique would yield sufficient numbers of mice over 20 grams to meet the specified
sampling design. On the first evening of the field assessment, two survey lines of 10 live
traps were ser for deer mice in typical old-field habitat in the area believed to contain the
desired DDT concentration gradient for the study design. At the beginning of the second day,
the traps were retrieved. Two deer mice over 20 grams were captured in each of the survey
lines. These results indicate that collection of deer mice over a period of a week or less with
this number and spacing of live traps should be adequate to meet study objectives.
REFERENCES
Bird, D.M.; Tucker, P.H.; Fox, G.A.; Lague, PC. 1983. Synergistic effects of Aroclor 1254
and rrurex on the semen characteristics of American kestrels. Arch. Environ. Contam.
Toxjcol. 12: 633-640.
Colbom, T.I. 1991. Epidemiology of Great Lakes bald eagles. J. Environ. Health Toxicol.
4: 395-453.
International Joint Commission (LJQ of United States and Canada. 1988. Great Lakes Water
Quality Agreement. Amended by protocol. Signed 18 November 1987. Ottawa,
Canada.
Fuller and Hobson. 1986. [reserved]
Melacon and Lech. 1983. [reserved]
APPENDIX A, Page 20
-------
A"g"st:i. '996 DRAFT
Tanabe. 1988. [reserved]
APPENDIX A, Page 21
-------
APPENDIX C
SUPPLEMENTAL GUIDANCE ON LITERATURE SEARCH
-------
August I!. 1996 DRAFT
APPENDIX C
SUPPLEMENTAL GUIDANCE ON LITERATURE SEARCH
A literature search is conducted to obtain information on contaminants of concern,
their potential ecological effects, and species of concern. This appendix is separated into two
sections; Section C-l describes the information necessary for the literature review portion of
an ecological nsk assessment. Topics include information for exposure profiles,
bioavailability or bioconcentration factors for various compounds, life-history information for
the species of concern or the surrogate species, and an ecological effects profile. Section C-2
lists information sources and techniques for a literature search and review. Topics include a
discussion of how to select key words on which to base a search and various sources of
information (i.e., databases, scientific abstracts, literature reviews, journal articles, and
govemmeht documents). Threatened and endangered species are discussed separately due to
the unique databases and information sources available for these species.
Prior to conducting a literature search, it is important to determine what information is
needed for the ecological risk assessment. The questions raised in Section D-l must be
thoroughly reviewed, the information necessary to complete the assessment must be
determined, and the purpose of the assessment must be clearly defined. Once these activities
are completed, the actual literature search can begin. These activities will assist in focusing
and streamlining the search.
C-1 LITERATURE REVIEW FOR AN ECOLOGICAL RISK ASSESSMENT
Specific information. During problem formulation, the risk assessor must
determine what information is needed for the nsk assessment. For example, if the nsk
assessment will estimate the effects of lead contamination of soils on terrestrial vertebrates,
then literature information on the effects of dissolved lead to fish would not be relevant. The
type and form of the contaminant and the biological species of concern often can focus the
literature search. For example, the toxicity of organometallic compounds is quite different
from the comparable inorganic forms. Different isomers of organic compounds also can have
different toxic effects.
Reports of toxicity tests should be reviewed critically to ensure that the study was
scientifically sound. For example, a report should specify the exposure routes, measures of
effect and exposure, and the full study design. Moreover, whether the investigator used
accepted scientific techniques should be determined.
The exposure route used in the study should also be comparable to the exposure route
in the nsk assessment. Data reported for studies where exposure is by injection or gavage are
APPENDIX C, Page 1
-------
August Z!. ! 996 DRAFT
not directly comparable to dietary exposure studies. Therefore, an uncertainty factor might
need to be included in the nsk assessment study design, or the toxicity report should not be
used in the nsk assessment.
To use some data reported in the literature, dose conversions are necessary to estimate
toxiciry levels for species other than those tested. Doses for many laboratory studies are
reported m terms of mg contaminant/kg diet. That expression should be converted to mg
contaminant/kg bodyweight/day, so that estimates of an equivalent dose in another species can
be scaled appropriately. Average ingesuon rate and body weight for a species often are
reported in the original toxiciry study. If not, estimates of those data can be obtained from
other literature sources to make the dose conversion:
Dose = (mg contaminant/kg diet) x ingestion rate (kg/day) x
(I/body weight (kg)).
Exposure profile. Once contaminants of concern are selected for the ecological risk
assessment, a general overview of the contaminants' physical and chemical properties is
needed. The fate and transport of contaminants in the environment determines how biota are
likely to be exposed. Many contaminants undergo degradation (e.g., hydrolysis, photolysis,
rrucrobial) after release into the environment. Degradation can affect toxiciry, persistence,
and fate and transport of compounds. Developing an exposure profile for a contaminant
requires information regarding inherent properties of the contaminant that can affect fate and
transport or bioavailability.
Bioavailability. Of particular importance in an ecological risk assessment is the
bioavailabihty of site contaminants in the environment. Bioavailability influences exposure
levels for the biota. Some factors that affect bioavailability of contaminants in soil and
sediment include the proportion of the medium composed of organic matter, grain size of the
medium, and its pH. The aerobic state of sediments is important because it often affects the
chemical form of contaminants. Those physical properties of the media can change the
chemical form of a contaminant to a form that is more or less toxic than the original
contaminant. Many contaminants adsorb to organic matter, which can make them less
bio?vailable.
Environmental factors that influence the bioavailability of a contaminant in water are
important to aquatic risk assessments. Factors including pH, hardness, or aerobic status can
determine both the chemical form and uptake of contaminants by biota. Other environmental
factors can influence how organisms process contaminants. For example, as water
temperatures nsc, metabolism of fish and aquatic invertebrates increases, and the rate of
uptake of a contaminant from water can increase.
APPENDIX C, Page 2
-------
August 21. 1996 DRAFT
If the literature search on the contaminants of concern reveals information on the
bioavailability of a contaminant, then appropriate bioaccumulauon or bioconcentration factors
(BAFs or BCFs) for the contaminants should be determined. If not readily available in the
literature, BAF or BCF values can be estimated from studies that report contaminant
concentrations in both the environmental exposure medium (e.g., sediments) and in the
exposed biota (e.g., benthic macroinvertebrates). Caution is necessary, however, when
extrapolating BAF or BCF values estimated for one ecosystem to another ecosystem.
Life history. Because it is impossible and unnecessary to model an entire ecosystem,
the selection of assessment endpoints and associated species of concern, and measurement
endpoints (including those for a surrogate species if necessary) are fundamental to a
successful risk assessment. This process is described in Steps 3 and 4. Once assessment and
measurement endpoints are agreed to by the risk assessor and risk manager, life history
information for the species of concern or the surrogate species should be collected. Patterns
of activity and feeding habits of a species affect their potential for exposure to a contaminant
(e.g., grooming activities of small mammals, egestion of bone and hide by owls). Other
important exposure factors include food and water ingestion rates, composition of the diet,
average body weight, home range size, and seasonal activities such as migration.
Ecological effects profile. Once contaminants and species of concern are selected
during problem formulation, a general overview of toxicity and toxic mechanisms is needed.
The distinction between the species of concern representing an assessment endpoint and a
surrogate species representing a measurement endpoint is important. The species of concern
is the species that might be threatened by contaminants at the site. A surrogate species is
used when it is not appropriate or possible to measure attributes of the species of concern. A
surrogate for a species of concern should be sufficiently similar biologically to allow
inferences on likely effects in the species of concern.
The ecological effects profile should include toxicity information from the literature
for each possible exposure route. A lowest-observed-adverse-effect level (LOAEL) and the
no-observed-adverse-effect level (NOAEL) for the species of concern or its surrogate should
be obtained. Unfortunately, these toxicity values are available for few wildlife species and
contaminants. If used with caution, toxicity data from a closely related species can be used to
estimate a LOAEL and a NOAEL for a receptor species.
C-2 INFORMATION SOURCES
This section describes information sources that can be examined to find the
information described in Section 3-1. A logical and focused literature search will reduce the
time spent searching for pertinent information.
APPENDIX C, Page 3
-------
August 21. 1996 DRAFT
A first step in a literature search is to develop a search strategy, including a list of key
words. The next step is to review computerized databases, either on-line or CD-ROM-based
information systems. These systems can be searched based on a number of parameters.
Scientific abstracts that contain up-to-date listings of current, published information
also are useful information sources. Most abstracts are indexed by author or subject.
Toxicity studies and information on wildlife life-histories often are summarized in literature
reviews published in books or peer-reviewed journals. Original reports of toxicity studies can
be identified in the literature section of published documents. The original article in which
data are reported must be reviewed before the data are cited in a nsk assessment.
Moore (1980) provides further insights on conducting a literature search, including
techniques to limit a search, selection of key words, and the location of dissertations.
Moore's examples relate to information on wildlife species, but apply to all components of an
ecological risk assessment.
Key words. Once the risk assessor has prepared a list of the specific information
needed for the nsk assessment, a list of key words can be developed. Card catalogs,
abstracts, on-line databases, and other reference materials usually are indexed on a limited set
of key words. Therefore, the key words used to search for information must be considered
carefully.
Useful key words include the contaminant of concern, the biological species of
concern, the type of toxicity information wanted, or other associated words. In addition,
related subjects can be used as key words. However, it usually is necessary to limit
peripheral aspects of the subject in order to narrow the search. For example, if the risk
assessor needs information on the toxicity of lead in soils to moles, then requiring that both
"lead" and "mole" are among the key words can focus the literature search. If the risk
assessor needs information on a given plant or animal species (or group of species), key
words should include both the scientific name (e.g., genus and species names or order 01
family names) and an accepted common name(s). The projected use of the data in the risk
assessment helps determine which key words are most appropriate.
If someone outside of the risk assessment team will conduct the literature search, it is
important that they understand both the key words and the study objectives for the data.
Databa»«*. Databases are usually on-line or CD-ROM-based information systems.
These systems can be searched using a number of parameters. Pnor to searching databases,
the risk assessor should determine which database(s) is most likely to provide the information
needed for the nsk assessment. For example, U.S. Environmental Protection Agency's
(EPA's) AQU1RE database (AQUatic Information REtrieval database) provides information
speciScally on the toxicity of chemicals to aquatic plants and animals. U.S. EPA's IRIS
APPENDIX C, Page 4
-------
August :i. 1996 DR\FT
(Integrated Risk Information System) provides information on human health risks (e.g..
references to original toxicity studies) and regulatory information (e.g., reference doses and
cancer potency factors) for a variety of chemicals. Other useful databases include the
National Library of Medicine's HSDB (Hazardous Subs'inces Data Bank) and the National
Center for Environmental Assessment's HEAST Tables (Health Effects Assessment Summary
Tables). Commercially available databases include BIOSIS (Biosciences Information
Services) and ENVTROLINE.
Several states have Fish and Wildlife History Databases or Academy of Science
databases, which often provide useful information on the life-histories of plants and animals
in the state. State databases are particularly useful for obtaining information on endemic
organisms or geographically distinct habitats.
Databases searches can yield a large amount of information in a short period of time.
Thus, if the key words do not accurately describe the information needed, database searches
can provide a large amount of irrelevant information. Access fees and on-line fees can apply;
therefore, the selection of relevant key words and an organized approach to the search will
reduce the time and expense of on-line literature searches.
Abstracts. Published abstract compilations (e.g.. Biological Abstracts, Chemical
Abstracts, Applied Ecology Abstracts) contain up-to-
-------
August 21. 1996 DRAFT
and Wildlife Services (U.S. FWS) has published several contaminant-specific documents that
list toxicological data on terrestrial, aquatic, and avian studies (e.g., Eisler, 1988). The U.S.
EPA publishes ambient water quality cntena documents (e.g.. U.S. EPA, 1985) that list all
the data used to calculate those values. Some literature reviews critically evaluate the original
studies (e.g., toxicity data reviewed by Long and Morgan, 1990). The Wildlife Exposure
Factors Handbook (U.S. EPA, 1993a,b) provides pertinent information on exposure factors
(e.g., body weights, food mgesuon rates, dietary composition, home range size) for 34
selected wildlife species.
Literature reviews can provide an extensive amount of information. However, the risk
assessor must obtain a copy of the original of any studies identified in a literature review that
will be used in the nsk assessment. The original study must be reviewed and evaluated
before it can be used in the nsk assessment. Otherwise, the results of the nsk assessment
could be based on incorrect and incompleie* information about a study.
References cited in previous studies. Pertinent studies can be identified in the
literature cited section of published documents that are relevant to the risk assessment, and
one often can identify several investigators who work on related studies. Searching for
references in the literature cited section of published documents, however, takes time and
might not be very effective. However, this is probably the most common approach to
identifying relevant literature. If this approach is selected, the best place to start is a review
article. Many journals do not list the title of a citation for an article, however, limiting the
usefulness of this technique. Also, it can be difficult to retrieve literature cited in obscure or
foreign journals or in unpublished masters' theses or doctoral dissertations. Although this
approach tends to be more time consuming than the other literature search approaches
described above, it probably is the most common approach used to locate information for a
risk assessment.
Journal articles, books, government documents. There arc a variety of
journals, books, and government documents that contain mformauon useful to nsk
assessments. The same requirement for retrieving the original reports for any informs''0^
used in the nsk assessment described for other information sources applies to these sources.
Threatened and endangered species. Threatened and endangered species are of
concern to both federal and state governments. When conducting an ecological nsk
assessment, it often is necessary to determine or estimate the effects of site contaminants to
federal threatened or endangered species. In addition, other special-status species (e.g.,
species listed by a state as endangered or threatened within the state) also can be the focus of
the assessment. During the problem formulation step, the U.S. FWS or state Natural Heritage
programs should be contacted to determine if these species are present or might be present on
or near a Superfund site.
APPENDIX C, Page 6
-------
August 21. 1996 DRAFT
Once the presence of a special-status species is confirmed or considered likely,
information on (his species, as well as on surrogate species, should be included in the
literature search. There are specific federal and state programs that deal with issues related to
special-status species, and often there is more information available for these than for non-
special-status species used as surrogates for an ecological nsk assessment. Nonetheless, the
use of surrogate species usually is necessary when an assessment endpoim is a special-status
species.
REFERENCES
Eisler, R. 1988. Lead Hazards to Fish, Wildlife, and Invertebrates: A Synoptic Review.
U.S. Fish and Wildlife Service Patuxent Wildlife Research Center, Laurel MD: U.S.
Department of the Interior, Biological Report 85(1.14), Contaminant Hazard Reviews
Rep. No. 14.
Long and Morgan. 1990. [reserved]
Moore. 1980. [reserved]
U.S. Environmental Protection Agency (U.S. EPA). 1993a. Wildlife Exposure Factors
Handbook Volume I. Washington, DC: Office of Research and Development;
EPA/600/R-93/I87a.
U.S. Environmental Protection Agency (U.S. EPA). 1993b. Wildlife Exposure Factors
Handbook Volume II: Appendix. Washington, DC: Office of Research and
Development; EPA/600/R-93/187b.
U.S. Environmental Protection Agency (U.S. EPA). 1985. Ambient Water Quality Criteria
for Copper-1984. Washington, DC: Office of Water. Regulations and Standards,
Catena and Standards Division. EPA/440/5-84-031. PB85-227023.
APPENDIX C, Page 7
-------
APPENDIX D
STATISTICAL CONSIDERATIONS
-------
1996 _ _ DRAFT
APPENDIX D
STATISTICAL CONSIDERATIONS
In the biological sciences, statistical tests often are needed to support decisions based
on alternative hypotheses because of the natural variability in the systems under investigation.
A statistical test examines a set of sample data, and, based on an expected distribution of the
data, leads to a decision on whether to accept the hypothesis underlying the expected
distribution or whether to reject that hypothesis and accept an alternative one, The null
hypothesis is a hypothesis of no differences. It usually is formulated for the express purpose
of being rejected. The alternative or test hypothesis is an operational statement of the
investigator s research hypothesis. An example of a null hypothesis for toxicity testing would
be that mortality of water fleas exposed to water from a contaminated area is no different
than mortality of water fleas exposed to water from an otherwise similar, but uncontaminated
area. An example of the test hypothesis is that mortality of water fleas exposed to water
from the contaminated area is higher than mortality of water fleas exposed to uncontaminated
water.
D-1 TYPE I AND TYPE II ERROR
There are two types of correct decisions for hypothesis testing: (1) accepting a true
null hypothesis, and (2) rejecting a false null hypothesis. There also are two types of
incorrect decisions: rejecting a true null hypothesis, called Type I error; and accepting a false
null hypothesis, called Type n error.
When designing a test of a HIGHLIGHT BOX D-1
hypothesis, one should decide what Rule of Thumb for Sample Size
magnitude of Type I error (rejection of a
true null hypothesis) is acceptable. Even An empirically determined rule of
when sampling from a population of known thumb for field sampling for tissue residue
parameters, there are always some sample levels or b.oaccumulauon studies is that
, . , , , , , ., Tf a = O.I0 is an acceptable level of statistical
sets which, by chance, differ markedly. If . v
f i i j . significance.
one allows 5 percent of samples to lead to a
Type I error, then one would on average '-^^^^^i
reject a true null hypothesis for 5 out of
every 100 samples taken. In other words, we would be confident that, 95 times out of 100,
one would not reject the null hypothesis of no difference "by mistake" (because chance alone
produced such deviant results). When the probability of Type I error (commonly symbolized
by a) is set at 0.05, this is called a significance level of 5 percent Setting a significance
level of 5 percent is a widely accepted convention in most experimental sciences, but it is just
that, a convention. One can demand more confidence (e.g., a = 0.01) or less confidence
(e.g., a = 0.10) that the hypothesis of no difference is not rejected by mistake.
APPENDIX D, Page 1
-------
August 21. 1996 DRAFT
If one requires more confidence for a given sample size that the null hypothesis is not
rejected by mistake (e.g., a = 0.01), the chances of Type H error increase. In other words,
the chance increases that one will mistakenly accept a faJse null hypothesis (e.g., mistakenly
believe that the contaminated water from the site has no effect on mortality of water fleas).
The probability of Type n error is commonly denoted by (3. Thus:
p (.Type I error) = a
p (Type II error) = P
However, if one tnes to evaluate the probability of Type II error (accepting a false hypothesis
of no difference), there is a problem. If the null hypothesis is false,.then some other
hypothesis must be true, but unless one can specify a second hypothesis, one can't determine
the probability of Type El error. This leads to another important statistical consideration,
which is ;he power of a study design and the statistical test used to evaluate the results.
D-2 STATISTICAL POWER
The power of a statistical test is equal to (1 P) and is equal to the probability of
rejecting the null hypothesis (no difference) when it should be rejected (i.e., it is false) and
the specified alternative hypothesis is true. Obviously, for any given test (e.g., a toxicity test
at a Superfund site), one would like the quantity (1 - P) to be as large as possible (and P to
be as small as possible). Because one generally cannot specify a given alternative hypothesis
(e.g.. mortality should be 40 percent in the exposed population), the power of a test is
generally evaluated on the basis of a continuum of possible alternative hypotheses.
Ideally, one would specify both a and P before an experiment or test of the hypothesis
is conducted. In practice, it is usual to specify a (e.g., 0.05) and the sample size because the
exact alternative hypothesis cannot be specified.1 Given the inverse relationship between the
likelihood of making Type I and Type n errors, a decrease in a will increase p for any given
sample size.
To improve the statistical power of a test (i.e., reduce P), while keeping a constant,
one can either increase the sample size (N) or change the nature of the statistical test. Some
statistical tests are more powerful than others, but it is important that the assumptions-
required by the test (e.g., normality of the underlying distribution) are met for the test results
to be valid. In general, the more powerful tests rely on more assumptions about the data (see
Section D-3).
Alternative study designs sometimes can improve statistical power (e.g., stratified
random sampling compared with random sampling if something is known about the history
'With a specified aJtemauve hypothesis, once a and (he sample size (N) are set, P is determined.
APPENDIX D. Page 2
-------
August 21. 1996 DRAFT
and location of contaminant release). A discussion of different statistical sampling designs is
beyond the scope of this guidance, however. Several references provide guidance on
statistical sampling design, sampling techniques, and statistical analyses appropriate for
hazardous waste sites (e.g., see Cochran, 1977; Green, 1979; Gilbert, 1987; Ott, 1995).
One also can improve the power of a statistical test if the test hypothesis is more
specific than "two populations are different," and, instead, predicts the direction of a
difference (e.g.. mortality in the exposed group is higher than mortality in the control group).
When one can predict the direction of a difference between groups, one uses a one-tailed
statistical test; otherwise, one must use the less powerful two-tailed version of the test.
Highlight Box 0-2
Key Points About Statistical Significance, Power, and Sample Size
(1) The significance level for a statistical test, a, is the probability that a statistical test will
yield a value under which the null hypothesis will be rejected when it is in fact true.
In other words, a defines the probability of committing Type I error (e.g., concluding
that the site medium is toxic when it is in fact not toxic to the test organisms).
(2) The value of P is the probability that a statistical test will yield a value under which the
null hypothesis is accepted when it is in fact false. Thus, |3 defines the probability of
committing Type £1 error (e.g., concluding that the site medium is not toxic when it is
in fact toxic to the test organisms).
(3) The power of a statistical test (i.e., 1 P) indicates the probability of rejecting the null
hypotheses when it is false (and therefore should be rejected). Thus, one wants the
power of a statistical test to be as high as possible.
(4) Power is related to the nature of the statistical test chosen. A one-tailed test is more
powerful than a two-tailed test. If the alternative to the null hypothesis can state the
expected direction of a difference between a test and control group, one can use the more
powerful one-tailed test
(5) The power of any statistical test increases with increasing sample size.
D-3 STATISTICAL MODEL
Associated with every statistical test is a model and a measurement requirement. Each
statistical test is valid only under certain conditions. Sometimes, it is possible to test whether
the conditions of a particular statistical model are met, but more often, one has to assume that
APPENDIX D, Page 3
-------
Augusi 21. 1996 DRAFT
they are or are not met based on an understanding of the underlying population and sampling
design. The conditions that must be met for a statistical test to be valid often are referred to
as the assumptions of the test.
The most powerful statistical tests (see previous section) are those with the most
extensive assumptions. In general, parametric statistical tests [e.g., t test, F test) are the most
powerful tests, but also have the most exacting assumptions to be met:
(1) The "observations" must be independent;
(2) The 'observations" must be drawn from a population that is normally
distributed;
(3) The populations must have the same variance (or in special cases, a known
ratio of variances); and
(4) The variables must have been measured at least on an interval scale so that it is
possible to use arithmetic operations (e.g., addition, multiplication) on the
measured values (Siegel, 1956).
The second and third assumptions are the ones most often violated by the types of data
associated with biological hypothesis testing. Often, distributions are positively skewed (i.e.,
longer upper than lower tail of the distribution). Sometimes, it is possible to transform data
from positively skewed distributions to normal distributions using a mathematical function.
For example, many biological parameters turn out to be log-normally distributed (i.e., if one
takes the log of all measures, the resulting values are normally distributed). Sometimes,
however, the underlying shape of the distribution cannot be normalized (e.g., it is bimodal).
When the assumptions required for parametric tests are not met, one must use
nonparametric statistics (e.g., median test, chi-squared test). Nonparametric tests are ia
general less powerful than parametric tests because less is known or assumed about the shape
of the underlying distributions. However, the loss in power can be compensated for uy an
increase in sample size, which is the concept behind measures of power-efficiency.
Power-efficiency reflects the increase in sample size necessary to make test B (e.g., a
nonparametnc test) as efficient or powerful as test A (e.g., a parametric test). A power-
efficiency of 80 percent means that in order for test B to be as powerful as test A, one must
make 10 observations for test B for every 8 observations for test A.
For further information on statistical tests, consult references on the topic (e.g., sec
references below).
.APPENDIX D. Page 4
-------
AugUS[:i-1996 DRAFT
REFERENCES
Cochran, W G. 1977 Sampling Techniques. 3rd ed. New York, NY' John Wilev and
Sons, Inc. }
Gilbert, R.O. 1987. Statistical Methods for Environmental Pollution Monitoring. New York
NY: Reinhold.
Green, R. H. 1979. Sampling Design and Statistical Methods for Environmental Biologists
New York, NY: Wiley. ' '
Ott, W.R. 1995. Environmental Statistics and Data Analysis. Boca Raton, FL: CRC Press,
Inc., Lewis Publishers.
Siegel, S. 1956. Non-parametric Statistics. New York, NY: McGraw-Hill.
APPENDIX D, Page 5
-------