United States Science Advisory Board EPA-SAB-EPEC-97-002
Environmental 1400 January 1997
Protection Agency Washington, DC
&EPA AN SAB REPORT: REVIEW
OF THE AGENCY'S DRAFT
ECOLOGICAL RISK
ASSESSMENT GUIDELINES
PREPARED BY THE ECOLOGICAL
PROCESSES AND EFFECTS
COMMITTEE (EPEC)
-------
NOTICE
This report has been written as a part of the activities of the Science Advisory Board, a
public advisory group providing extramural scientific information and advice to the Administrator
and other officials of the Environmental Protection Agency. The Board is structured to provide a
balanced expert assessment of scientific matters related to problems facing the Agency. This
report has not been reviewed for approval by the Agency; and hence, the contents of this report
do not necessarily represent the views and policies of the Environmental Protection Agency or
other agencies in Federal government. Mention of trade names or commercial products does not
constitute a recommendation for use.
-------
ABSTRACT
The Ecological Processes and Effects Committee (EPEC) of the Science Advisory Board
(SAB) met in Washington DC on September 19-20, 1996 to review the Agency's draft Ecological
Risk Assessment Guidelines (Eco RA GLs). The Committee commended the Agency on its
multi-year effort and its frequent contact with the outside scientific community during the
development of the Eco RA GLs. The Committee responded to a charge covering ten specific
areas: a) Risk Assessor/Risk Manager Interactions; b) Assessment Endpoints; c) Tiered
Assessments; d) Field Data; e) Background Variability; f) Landscape Ecology; g) Multiple
Stressors; h) Risk Estimation Approaches; i) Terminology; and j) Future Products.
While making suggestions for further improvements and for specific extensions of the
work, the Committee endorsed the process and product of Agency's efforts on this project.
Keywords: ecological, risk assessment, guidelines, endpoints, landscape
-------
US ENVIRONMENTAL PROTECTION AGENCY
SCIENCE ADVISORY BOARD
ECOLOGICAL PROCESSES AND EFFECTS COMMITTEE
CHAIR
DR. MARK A. HARWELL1, Rosenstiel School Marine and Atmospheric Science,
University of Miami, Miami, FL
VICE CHAIR
DR. ALAN W. MAKI2, Exxon Company-USA, Houston, TX
MEMBERS
DR. WILLIAM J. ADAMS, Kennecott Utah Copper Corp., Magna, UT
DR. STEVEN M. BARTELL1, SENES Oak Ridge, Inc., Oak Ridge, TN
DR. KENNETH W. CUMMINS, South Florida Water Management District, West Palm
Beach, FL
DR. VIRGINIA DALE, Oak Ridge National Laboratory, Oak Ridge, TN
DR. CAROL A. JOHNSTON, Natural Resources Research Institute, Duluth, MN
DR. ANNE MCELROY, New York Sea Grant, SUNY at Stony Brook, Stony Brook,
NY
DR. FREDERIC K. PFAENDER, Carolina Federation for Environmental Programs,
University of North Carolina, Chapel Hill, NC
DR. JERRY SCHUBEL3, New England Aquarium, Boston, MA
DR. WILLIAM H. SMITH1, School of Forestry and Environmental Studies, Yale
University, New Haven, CT
DR. TERRY F. YOUNG, Environmental Defense Fund, Oakland, CA
1 Recused, due to participation in development of Agency's document.
2 Acting Chair for the review of Ecological Risk Assessment Guidelines
3 Absent for the review
111
-------
CONSULTANTS
DR. THOMAS W. LA POINT, Clemson University, Clemson, SC
DR. CHARLES A. PITTINGER, Procter & Gamble, Cincinnati, OH
SCIENCE ADVISORY BOARD STAFF
MS. STEPHANIE SANZONE, Designated Federal Official, Ecological Processes and
Effects Committee, Science Advisory Board (1400), US EPA, Washington, DC
MS. CONNIE VALENTINE, Staff Secretary, Science Advisory Board (1400), US EPA,
Washington, DC
DR. DONALD BARNES, SAB Staff Director, Science Advisory Board (1400), US
EPA, Washington, DC
IV
-------
TABLE OF CONTENTS
1. EXECUTIVE SUMMARY 1
2. INTRODUCTION 4
2.1 Statement of the Charge 4
2.2 Process 5
3. COMMITTEE FINDINGS AND RESPONSE TO THE CHARGE 7
3.1 Overall Findings
7
3.2 Responses to the Charge Questions 7
3.2.1 Risk Assessor/Risk Manager Interactions 7
3.2.2 AssessmentEndpoints 8
3.2.3 Tiered Assessments 10
3.2.4 Field Data 10
3.2.5 Background Variability 12
3.2.6 Landscape Ecology 12
3.2.7 Multiple Stressors 13
3.2.8 Risk Estimation 14
3.2.9 Terminology 15
3.2.10 Future Products 15
4. ADDITIONAL POINTS 19
4.1 Use of accurate language 19
4.2 Case studies 19
4.3 Hypothesis testing 20
4.4 Adaptive Management 20
4.5 Risk Estimation Techniques 22
4.6 Risk Characterization 22
4.7 Different levels of sophistication for different needs 23
4.8 Lines of evidence 23
5. CONCLUSION 25
REFERENCES R-l
-------
1. EXECUTIVE SUMMARY
The Ecological Processes and Effects Committee (EPEC) of the Science Advisory Board
(SAB) met in Washington, DC on September 19-20, 1996 to review the Agency's draft Ecological
Risk Assessment Guidelines (Eco RA GLs) (5). The Committee commended the Agency on its
multi-year effort on the project and its frequent contact with the outside scientific community
during the development of the Eco RA GLs.
The Committee responded to a charge covering ten specific areas:
a) Risk Assessor/Risk Manager Interactions - "Overall how compatible are these
Proposed Guidelines with the National Research Council concept of the risk
assessment process and the interactions between risk assessor, risk managers, and
other interested parties?"
Answer: The Committee found the Eco RA GLs to be generally compatible with
the recent NRC reports. They recommended further elaboration in some parts of
the GLs.
b) Assessment Endpoints - "Assessment endpoints are an 'explicit expression of the
environmental value that is to be protected'...Some reviewers have recommended
that assessment endpoints also include a decision criterion that is defined early in
the risk assessment process (e.g., no more than 20% reduction in reproduction, no
more than a 10% loss of wetlands.)...[T]he Proposed Guidelines suggest that such
decisions are more appropriately made during discussions between risk assessors
and managers in risk characterization at the end of the process. What are the
relative merits of each approach?"
Answer: The Committee agrees with the approach proposed in the Eco RA GLs,
at the same time noting the importance of retaining an appropriate balance between
allowing risk managers/interested parties to inform the risk assessment process, on
the one hand, and while maintaining the independence of the risk assessor and the
risk assessment process, on the other. Given the intended broad application of the
Eco RA GLs to the full range of ecological issues, the Committee does not believe
that a single set of quantitative decision criteria are appropriate from a technical
point of view. Further, the Committee recognizes the inherently policy aspect of
reaching final quantitative decisions. Scientific information can-and should—
inform the selection of quantitative numbers in any decision, but there are factors
beyond those in the strictly technical realm that the risk manager must also
consider.
c) Tiered Assessments - "While the Proposed Guidelines acknowledge the
importance of tiered assessment, the wide range of applications of tiered
-------
assessment makes more specific guidance difficult. Given the broad scope of the
Proposed Guidelines, what additional principles for conducting tiered assessment
can be discussed?"
Answer: The Committee endorses the concept of tiered testing as an efficient use
of resources in ecological risk assessment. The Committee also encourages the
development of guidance criteria for determining when higher tiered assessment is
necessary.
d) Field Data - "[Beyond the material already in the Proposed Guidelines], what, if
any additional material should be added on these topics and, if so, what principles
should be highlighted?"
Answer: The Committee report provides a discussion of the comparative
advantages of field data and lab data, including a set of "principles" to guide in the
evaluation and use of field data.
e) Background Variability - "[Beyond the material already in the Proposed
Guidelines], are there additional considerations concerning this issue that should be
included in the Proposed Guidelines."
Answer :The Committee recommends specific approaches to addressing this
problem: 1) use of information on the mechanisms involved, particularly for
bioaccumulative substances; 2) use of caged organisms; and 3) use of improved
statistical techniques for dealing with highly variable data sets.
f) Landscape Ecology - "...given the Agency's present interest in evaluating risks at
larger spatial scales, how could the principles of landscape ecology be more fully
incorporated into the Proposed Guidelines?"
Answer: The Committee agrees that the Agency should pay more attention to
landscape effects and its report offers several specific suggestions for expanding
the Eco RA GLs so that they more effectively include consideration of landscape
issues. Among the suggestions are improvements in the Figures, presentation of a
"hierarchy of scales", and use of landscape-based indicators.
g) Multiple Stressors - "[Beyond the material already in the Proposed Guidelines],
what additional principles can be added?"
Answer: The Committee acknowledges this difficult problem and proposes specific
statistical methods of analysis that would illuminate the situation. Specifically,
path analysis, multiple-regression analysis, canonical analysis, or partial correlation
analysis are often useful in discerning the effects of multiple stressors on an effect
variable.
-------
h) Risk Estimation Approaches - "The Proposed Guidelines use simple exposure and
stressor-response curves to illustrate some aspects of risk estimation. What other
examples could be used to represent risk estimation for nonchemical stressors or
more complex situations?"
Answer: The Committee provides several examples of empirically derived non-
chemical stressor-response phenomena that illustrate the generality of such curves
for decision making. In addition, they note the utility of process models for
estimating the shape of stressor-response curves, even in the absence of field data.
Finally, the Committee notes that many phenomena appear to exhibit practical
thresholds that may be useful benchmarks in the risk management decision-making
process.
i) Terminology - The Agency asked for SAB input on the use of a few specific terms
and phrases.
Answer: The Committee felt that the terminology issues were simply a question of
semantic choices and decided not devote time to reaching a consensus position on
them.
j) Future Products - "[Beyond the material already in the Proposed Guidelines and]
considering the state of the science of ecological risk assessment, what topics most
require additional guidance?
Answer: The Committee generated a list of items which they recommend that the
Agency study in order to generate additional information and/or guidance.
In addition to responding to the specific elements of the charge, the Committee went
further and provided additional comments on issues that arose during the course of the review.
These additional comments cover the following areas:
a) The use of accurate language
b) Case studies
c) Hypothesis testing
d) Adaptive management
e) Risk estimation techniques
f) Risk characterization
g) Different levels of sophistication for different needs
h) Lines of evidence
-------
2. INTRODUCTION
The U.S. Environmental Protection Agency (Agency) has a considerable history in
assessing risks associated with environmental stressors and adverse effects to human health. Over
the years, in order to conduct such assessments in a systematic, consistent manner, the Agency
has developed a range of risk assessment guidelines (RA GLs) that articulate the general
procedures to be followed when the Agency assesses these risks (1). Given the significance and
wide applicability of these RA GLs, the Agency has asked the Science Advisory Board (SAB) to
conduct an independent, public review of the scientific basis for these documents (2).
In the mid-1980s the Agency became increasingly concerned about ecological risks
associated with environmental problems, in addition to any risks that those problems might pose
to human health. In its 1990 report on comparing risks, Reducing Risk, the SAB explicitly
recommended that the Agency "attach as much importance to reducing ecological risk as it does
to reducing human health risks" (3). This report increased awareness of the need for systematic,
consistent procedures for assessing ecological risks that would match those developed for human
health risks. Consequently, the Agency embarked on a multi-year effort to develop ecological risk
assessment guidelines (Eco-RA GLs) in a manner that would involve participation and review by
a wide range of scientists across the country, including the SAB.
The current report—the seventh Eco RA GLs-related product since 1990 (4)~documents
the SAB's review of the Agency's draft Eco-RA GLs (5), which builds, in particular, upon the
Agency's Ecological Risk Assessment Framework document (6). This review is an objective
assessment of the document, informed by—but independent of—the earlier efforts. This
independence results from the fact that the composition of the Ecological Processes and Effects
Committee (EPEC) has changed significantly, in the intervening years. In addition, as noted in the
roster at the beginning of this report, three EPEC members recused themselves from this review
due to their participation in previous aspects of the development of the Eco-RA GLs. They have
been replaced for this exercise by two consultants who have not been involved in prior SAB
reviews.
2.1 Statement of the Charge
The charge to the SAB from the Agency (7) covered ten questions:
a) Risk Assessor/Risk Manager Interactions - "Overall how compatible are these
Proposed Guidelines with the National Research Council concept of the risk
assessment process and the interactions between risk assessor, risk managers, and
other interested parties?"
b) Assessment Endpoints - "Assessment endpoints are an 'explicit expression of the
environmental value that is to be protected'. Some reviewers have recommended
that assessment endpoints also include a decision criterion that is defined early in
-------
the risk assessment process (e.g., no more than 20% reduction in reproduction, no
more than a 10% loss of wetlands.)...[T]he Proposed Guidelines suggest that such
decisions are more appropriately made during discussions between risk assessors
and managers in risk characterization at the end of the process. What are the
relative merits of each approach?"
c) Tiered Assessments - "While the Proposed Guidelines acknowledge the
importance of tiered assessment, the wide range of applications of tiered
assessment makes more specific guidance difficult. Given the broad scope of the
Proposed Guidelines, what additional principles for conducting tiered assessment
can be discussed?"
d) Field Data - "[Beyond the material already in the Proposed Guidelines], what, if
any additional material should be added on these topics and, if so, what principles
should be highlighted?"
e) Background Variability - "[Beyond the material already in the Proposed
Guidelines], are there additional considerations concerning this issue that should be
included in the Proposed Guidelines."
f) Landscape Ecology - "...given the Agency's present interest in evaluating risks at
larger spatial scales, how could the principles of landscape ecology be more fully
incorporated into the Proposed Guidelines?"
g) Multiple Stressors - "[Beyond the material already in the Proposed Guidelines],
what additional principles can be added?"
h) Risk Estimation Approaches - "The Proposed Guidelines use simple exposure and
stressor-response curves to illustrate some aspects of risk estimation. What other
examples could be used to represent risk estimation for nonchemical stressors or
more complex situations?"
i) Terminology - The Agency asked for SAB input on the use of a few specific terms
and phrases.
j) Future Products - "[Beyond the material already in the Proposed Guidelines and
]considering the state of the science of ecological risk assessment, what topics
most require additional guidance?
2.2 Process
As announced in the Federal Register (8), the Ecological Processes and Effects Committee
(EPEC) convened in public session on Sept 19-20, 1996 in Conference Room 17 of the
Washington Information Center at Waterside Mall near EPA's Washington Headquarters Offices
-------
to carry out the review of the draft Eco RA GLs (5). The Committee was also briefed by Agency
scientists who were most directly involved in preparation of the Eco RA GLs over a period of
years. As noted above, three EPEC members, including the Chair, Dr. Mark Harwell, recused
themselves and did not attend the meeting. The group was augmented by two consultants for this
meeting. After leading the Committee through the review, the Acting Committee Chair (EPEC
Vice-Chair), Dr. Alan Maki, speaking in public session, summarized the overall findings and
recommendations to Agency representatives and expressed the Committee's appreciation for the
Agency's cooperation throughout the process.
-------
3. COMMITTEE FINDINGS AND RESPONSE TO THE CHARGE
3.1 Overall Findings
The Agency's Ecological Risk Assessment Guidelines are the product of an almost 10-year
effort within EPA to produce a model document that can guide the implementation of ecological
risk assessments. The conceptual outline for these ecorisk principles was first developed at the
Pellston workshop in 1977 (10). At that time, the mutual importance and co-dependency of
environmental chemistry and fate testing with ecotoxicology were first identified. The current
guidelines advance ecological risk assessment well beyond that early conceptual stage to what is
now a truly functional practice.
The proposed guidelines outline the entire process and provide details on how each of the
steps of problem formulation, analysis, risk characterization, and risk management can be
successfully completed. Sufficient balance between specific guidance and flexibility is maintained
so that the user can adapt the process to highly variable problems. The SAB review comments
below are generally supportive of the guidelines while offering numerous suggestions for
improvement. Comments address a wide range of issues, including risk assessor/risk manager
interaction, assessment endpoints, use of tiered assessment, field data, and multiple stressors.
The Eco RA GLs document provides a template for dealing with many environmental
programs and problems; e.g., new chemicals, water and air quality issues, waste site assessment,
and spills. As such, these guidelines, like the Agency's ecological risk assessment framework from
which they sprang, will likely be widely accepted and used by regional, state, and local
authorities, as well as by many in the private sector. Indeed, the Eco RA GLs appear destined to
be as seminal and trend-setting in their field, as the Cancer RA GLs have been in theirs. This
achievement reflects well on science at EPA and all those who were involved in the guidelines
development process.
3.2 Responses to the Charge Questions
3.2.1 Risk Assessor/Risk Manager Interactions
Among the refinements that the Agency has made to the Framework for Ecological Risk
Assessment is the addition of an expanded discussion of the type of interaction that should occur
among risk assessors, risk managers, and other interested and affected parties at the outset of the
Eco RA process.2 The importance of this dialogue has been further emphasized by adding an
explicit "planning" phase, with associated outputs, to the process. The Committee fully supports
these additions and the manner in which they have been incorporated in the Guidelines.
" Interested and affected parties" should include local residents, who, especially in rural areas, may have important ecological knowledge that is
important to include in the analysis.
-------
At the same time, the Committee recognizes the importance of conducting the scientific
risk assessment analysis in a manner that keeps it insulated from inappropriate influence by other
components of the decision making process. This relationship has recently been described in a
National Research Council report as an iterative process of "analysis", on the one hand, and
"deliberation", on the other. (9b) The Committee feels that this delineation of the scientific
component of the process-separate from social, political, and economic considerations—is
important and, therefore, supports the Agency's use of the "bold black line box" in Figure 1-2 of
the Eco RA GLs (See Figure 1-2 from p. 34 of Eco RA GLs - overleaf) to highlight the boundary
that should separate the consensus process from the scientific process. We further recommend
that the following statement, that now appears in the text on page 36, be added to the Executive
Summary:
"... it is imperative to remember that the planning process is distinct from the
scientific conduct of an ecological risk assessment. This distinction helps ensure
that political and social issues, while helping to define the objectives for the risk
assessment, do not bias the scientific evaluation of risk."
In order to further clarify the draft Eco RA GLs' discussion of the respective roles of risk
assessors, risk managers and other interested parties, we suggest that "interested parties" be
added to the planning box in the framework diagram that appears first as Figure 1-2. In addition,
a slightly expanded discussion of the benefits of tapping the scientific expertise of the risk
assessment team could be added (perhaps at the top of page 37). For example, while the current
text makes it clear that the risk manager can help to make the risk assessment more relevant to
decision making by clearly articulating societal goals, the document should also note the benefit
associated with information flowing in the opposite direction; e.g., the risk assessor can help to
define management options that are more likely to achieve the stated goals because they are
ecologically grounded. (See also Section 4.4 on "Adaptive Management".)
The Committee felt strongly that Eco RA GLs should have a fuller discussion of the
continuing communication that should occur between risk assessors and managers throughout the
entire process. In concert with the recent reports of the National Research Council on risk
assessment (9), the Committee views the decision-making process is an iterative one, involving
continual interaction between the parties. The "dotted line" at the bottom of Fig 1-2 that links
"Risk Management" and the "As Necessary: Acquire Data, Iterate Process, Monitor Results"
should be replaced with a solid line to emphasize the need for and utility of these continuing
connections as the scientific assessment and the risk management needs evolve.
3.2.2 Assessment Endpoints
The Committee agrees with the definition for assessment endpoints, as an entity and an
attribute, and with the approach described in the draft Eco RA GLs to generate them. The
Guidelines should not, in general, contain specific decision criteria for these endpoints, although
they should acknowledge that some decision criteria may be mandated by statute or established by
regulatory policy.
-------
-------
The balanced involvement of interested parties; e.g., industrial/environmental and
public/private communities, is necessary in order to identify endpoints that are valued by society.
At the same time, the involvement of technical experts is necessary to ensure identification of
endpoints that are ecologically sound and relevant to decision making.
Development of assessment endpoints for landscape-scale risk assessment may benefit
from explicit consideration of a hierarchy of ecological scales. For example., assessment
endpoints might be derived at the population, community, ecosystem, and landscape scales. At
each of these scales, consideration could be given to structural, compositional, and functional
attributes. Using this type of hierarchical "checklist" would help to assure the ecological
relevance of the assessment endpoints, as describe in Section 3.3.1.1 of the draft Eco RA GLs.
3.2.3 Tiered Assessments
The Committee strongly agrees with the need for and value of tiered testing in ecological
risk assessment. As a general rule, the scope of a particular assessment should be commensurate
with the scope and importance of the risk management decision. The Eco RA could better
describe some of the considerations and criteria for determining when higher tiered testing is
necessary, particularly in those cases in which adequate data are available to conduct such a
higher tier assessment. Some of those criteria would be science-based, with a firm technical logic
connecting the results of lower tiered testing to the need for additional testing in higher tiers. At
the same time, the Committee recognizes that there are non-technical, risk management
considerations that can legitimately come into play (e.g., time and resources available,
"importance" of the decision, etc.) when the Agency is reaching a decision on how much
additional information to gather and assessment to conduct. In any case, risk managers need to be
sensitive to the need for a firm scientific foundation for the decisions they make.
3.2.4 Field Data
The Committee believes the Eco RA GLs should provide more discussion on topics
related to the use of field observational data in ecological risk assessments. Historically, the use
of field data to make ecological decisions has preceded and subsequently paralleled the use of
laboratory data. The first Pellston Workshop (10) endorsed the idea of collecting data in
sequential tiers, the last of which would be the collection of field data in a "confirmatory" tier.
The concept embedded here is that field data can be an important component of a risk assessment
and, when obtained from properly designed studies, can be used in lieu of, or in support of,
decisions based on laboratory data or theoretical constructs.
The Committee, while not responding to all aspects of the EPA "field data" question,
proposes the following ecological principles relative to the use of field data:
a) The overall intent of ecological risk assessments is to evaluate risk to organisms
(populations, communities) in the field. Therefore, data from well-defined field
studies, both exposure and effects, is usually preferred over other forms of data.
10
-------
b) Field studies should be designed with sufficient statistical rigor to define one or
more of the following:
1) Stress-response regime for organisms and endpoints of interest.
2) Differences in measurement endpoints between reference sites and or controls
from study areas or treatments, respectively.
3) Demonstrate a lack of observed field effects in a study with an appropriate
statistical design capable of detecting reasonable changes (detrimental) in
measurement endpoints.
c) Field data may be used to help characterize the stressor-response regime, as done
in the laboratory, but are often used to document the "health" of the ecosystem
(population, community) relative to the same site historically, a reference site, or a
conceptual reference site. Biocriteria are often used for these comparisons.
d) Field data are not needed for all risk assessments. For example, screening level
assessments and prospective risk assessments often do not incorporate field data.
e) Field data are more important for some risk assessments than for others. This
typically occurs when assessments are performed on large areas, such as whole
watersheds, where multiple stressors are present, or where site-specific factors
significantly control the stressor-response regime.
Additional factors favoring the use of field data that the Agency might include in their Eco
RA GLs discussions include the following:
a) Species of interest and assessment endpoints of interest can often be
observed/measured directly in the field.
b) Uncertainty associated with extrapolating from laboratory data to field
relationships, from acute to chronic ratios, and from one species to species
extrapolation can often be avoided by utilizing field data.
c) Exposure regimes in the field are often vastly different from those in laboratory
experiments. The use of field data can reduce uncertainty associated with the
exposure regime, including those factors associated with stressor duration or fate
and transport, as in the case of chemicals. At the same time, the Eco RA GLs
should note that most site or watershed risk assessments will be site-specific and
that data collected there will not necessarily reflect other situations.
11
-------
3.2.5 Background Variability
There are several approaches that could be described in the Eco RA GLs that can be used
to investigate potentially important effects that are difficult to distinguish from "background"
variability in natural systems.
One approach would emphasize the importance of understanding the nature of the
chemical or physical insult to the population or community at risk. Usually low-level stressors
would be expected to have little toxicity or—by definition—have a relatively low or diffuse impact
on the population, community or ecosystem under assessment. However, in some cases, such
stressors may be potentially dangerous because they bioaccummulate to levels sufficient to induce
developmental, teratogenic, or tumorigenic effects; e.g., PCBs in the Great Lakes. Therefore, the
risk assessment of such low-level stressors needs to take into account the potential for
bioaccumulation.
A second approach to assessing the significance of low-level stressors is to use previously
unexposed organisms caged or maintained for periods of time on-site in order to enhance the
potential for measuring uptake and/or effects. It is possible to obtain direct measures of body
burden, biomarkers, and behavior characteristics during these extended exposures. An example of
such research is the work of Dickerson et al. (11), who are exposing previously unexposed
rodents on the Rocky Mountain Arsenal, CO and trapping them after season-long exposures.
Biochemical and physiological biomarkers are being linked to population-level effects; e.g., the
relationship between body burden data and reproductive success. This approach will explain
some, but not all, of the variability seen in the population of the native animals that have had
varying degrees of exposures in the field. Therefore, there will be a greater possibility of
detecting effects due to the stressor.
A third approach is to employ appropriate statistical techniques to ferret out the variance
components due to natural variation from those due to exposure to the stressors of interest. A
range of statistical analytic methods could be employed to address this issue, including factor
analysis methods (e.g., principal components) and canonical or partial correlation techniques.
Non-linear distributions can be treated by analysis of frequency approaches, appropriate
transformations, or polynomial regression techniques.
3.2.6 Landscape Ecology
Ecological risk assessment is typically applied to toxins that are stressors at a particular
site. The Eco RA GLs attempt to move the ecological risk assessment concept to larger spatial
scales which means that the process must consider multiple stressors, multiple ecosystem types,
and multiple levels of biological organization, as well as spatial heterogeneity. Some of these
issues have been addressed more directly and more completely than others. Overall, the
Committee has some concerns that the attempt to extend the ecological risk concepts to broad
spatial scales is not fully balanced across the document. The Committee recommends that future
12
-------
versions of the Eco RA GLs incorporate guidance at the landscape scale. Some particular cases
where landscape concepts should be considered are given below:
a) The document should keep the temporal and spatial scale in mind when data
evaluation is discussed (cf. p.75 and following).
b) The Figures should include concepts and examples that are not toxins and that are
spatial in nature.
c) The concept of thresholds needs to be added as a criterion. That is, while
theoretical linear stressor-response curves do not have ecologically based break
points (e.g,, "thresholds"), non-linearities or chaotic responses are particularly
possible "in the real world" where broad scales issues, such as multiple factors on
levels or organization, come into play. (Also see Section 3.2.8 below)
d) As a general matter, the Agency should consider presenting a variety of concepts
in the document through a "hierarchy of scales" that considers both spatial
resolution and extent. Among the concepts that seem well-suited for such a
presentation are setting management goals (p.38), integrating information (p.46),
defining "key" information (p.46) and ecological relevance (p.50), and selecting
assessment endpoints (pp 56-57).
e) A discussion of landscape-based indicators needs to be added. For example, on p.
52 of the document there is a discussion of fragmentation, but it does not include
species gap crossing ability as a critical metric (12).
f) Although the discussion of exposure includes mention of multiple stressors, the
concept of landscape influences on these stressors is not addressed; cf. pp 52-53.
For example, topography can affect the temporal and spatial distribution of
exposure regimes, such as temperature and air-borne pollutants.
g) A regional scale analysis would not only include impacts on single organisms as a
metric of stress, but it might also include physical indicators on a complex of biotic
responses as indicators.
3.2.7 Multiple Stressors
The multiple stressor problem is very likely the "rule" in ecosystems, as discussed briefly
in 5.2.2, "Determining Ecological Adversity." The difficulty lies in discerning which of the
stressors have a predominant influence on the endpoint of interest, in knowing how the multiple
stressors interact, and in understanding how they contribute to the added risk for the population
or community of interest in the risk assessment.
Multiple stressors can be analyzed using many of the same statistical techniques described
in Section 3.2.5 above on "low-level effects." Path analysis, multiple-regression analysis,
13
-------
canonical analysis, or partial correlation analysis are often useful in discerning the effects of
multiple stressors on an effect variable.
3.2.8 Risk Estimation
Stressor/response curves are commonly used to depict the effects of chemical stressors on
organisms. Such curves are much less commonly used to describe the effects of non-chemical
stressors on organisms and the effects of non-chemical stressors on non-biological responses,
particularly at ecosystem and landscape scales.
However, just as with chemical stressors, empirical relationships can be developed
between non-chemical stressors and non-biological responses. For example, there is a
demonstrable relationship between wetland extent and downstream flood peak (13) and between
land use activities and downstream nutrient concentrations (14). In addition, the U. S.
Geological Survey and USDA Natural Resources Conservation Service have developed many
empirical relationships pertaining to water quantity and quality (15).
Stressor-response curves can also be generated through the development and manipulation
of process models. That is, the magnitude of a stressor can be varied incrementally so as to
generate response curves. For example, a non-point source pollution model could be used to
predict the response of stream sediment transport to increases in cropland area.
Risk managers are increasingly acknowledging the connection between watershed
stressors and downstream responses. Examples include the effect of upstream land use activities
on the water quality of Chesapeake Bay and New York City' s water supply. It is often difficult to
prove cause and effect of stressors/responses at such large scales because a) the location of
stressor generation is often different from the location of the response, b) there are often multiple
sources of a particular stressor; e.g., phosphorous from septic system leakage, accelerated
erosion, and run-off from impervious surfaces, and c) base line values are not sufficiently well-
known. In such cases, the use of process models is usually the only viable method of risk
estimation. If a process model is spatially explicit, specific locations for implementing risk
management practices may be suggested. "Spatial decision support systems", in which a process
model is linked to GIS input and displayed, may be useful in such complex cases.
If a stress/response relationship is non-linear, inflection points in those relationships may
suggest natural thresholds that could be useful for risk management purposes. For example, the
Everglades Forever Act explicitly directs a coalition of Agency scientists and their consultants to
determine the "threshold" of phosphorous concentrations (somewhere between 0 and 50 ppm)
that leads to changes in Everglades Wetlands plant community composition; e.g.. replacement of
sawgrass by cattail. Also, work by Oberts (16) has shown that in watersheds in which wetlands
constitute less than 10% of the land area, the total suspended solids loadings to streams draining
those watersheds are significantly greater than in streams draining watersheds that have more than
10% of their area in wetlands.. Similarly, there may be natural thresholds of stress below which
14
-------
no response occurs because an organism or ecosystem can tolerate or assimilate that stress
without an adverse response. For example, some amount of phosphorous input to a wetland
could be assimilated by soil and vegetation uptake processes, while excessive amounts would be
detrimental to the wetland.
3.2.9 Terminology
The Committee felt that the terminology issues were simply a question of semantic choices
and decided not to devote time to reaching a consensus position on them.
3.2.10 Future Products
The Committee commends the Agency on the significant contributions it has made to the
field of ecological risk assessment through the development and dissemination of the Framework
(6) and these draft Eco RA GLs (5). At the same time, it is clear that more work needs to be
done as a part of an overall effort to continually improve ecological risk assessment. In this
subsection, the Committee briefly lists a number of the more important outstanding issues the
Agency should address in a timely manner through exposition and guidance.
a) Reference sites
The selection of appropriate reference sites for use in data collection or field
studies is a difficult issue. However, recent EPA documents (17) have presented
imaginative approaches to selection of reference sites, even when there are little or no field
data available. The Agency has discussed categories of reference site information,
including a hierarchy of actual reference conditions; i.e., from "best system" (which can
mean least disturbed or most pristine), to "best site" or population in an ecosystem, to
present defined conditions, to historical reference data.
A useful new approach is to define conceptual reference conditions which
represent a best professional judgement of where a particular site fits along a spectrum of
sites exhibiting the full range of possible values of all (or as many as possible) relevant
parameters, both those sites for which data exist and those sites for which they do not.
This approach is free from the compromises that are necessary in selection of specific
reference sites and flexible in that the position of the specific site along the spectrum can
change as more data are generated or site conditions change. An added advantage of the
approach is its capability to compare different ecosystems or sites; i.e., they are simply at
different points along the gradients. Further, it is also a useful way to compare—albeit
conceptually—different management scenarios or alternatives based on how they would
move the ecosystem along the various gradients.
In the face of this significant progress in addressing selection criterion for reference
sites, using of reference data, and exploring new approaches to reference concepts, the
Committee recommends that the Agency consider pulling this information together into a
15
-------
single guidance document that could be used in many places, both inside and outside the
Agency.
b) Landscape issues
The question of how risk assessment approaches can be applied to a complex
ecosystem or at landscape scales has not been addressed in either a research or an applied
mode. Since there is increased recognition that there are important environmental
protection issues at the landscape scale, it is important there be explicit guidance on how
to conduct ecological risk assessment on this scale. Issues to be resolved include how the
multiple stressors should come into play, how spatial heterogeneity affects the stressor
impact, how human actions may enhance or deter any effects, and how risk analysis should
be viewed over the long-term. The Committee encourages the Agency to provide needed
leadership by pressing forward to investigate these important issues in the coming years.
c) Risk Characterization: How to...
The Committee recommends that in the future the Agency prepare a supporting
guidance document summarizing techniques for synthesizing exposure and effects data
into a risk characterization. The entire process of risk assessment is focused on providing
a sound technical basis for a decision on potential risk posed by the stressors of interest.
Hence, the process of deriving risk estimates and formalizing the risk characterization is
very important.
In particular, techniques for preparing risk characterizations have been emerging
and expanding over the past 3-5 years. For example, there are now methods available for
estimating risk that go beyond the traditional hazard quotient-point estimate approach.
Other recent advances include the use of probability and cumulative distribution functions,
as referenced in the Eco RA GLs. In addition, the current technical literature contains
numerous approaches for using these results in making decisions about ecological risk.
Therefore, the Committee recommends that the Agency collect this information
and synthesize it into a guidance document for use both inside and outside the Agency.
d) Risk Communication
The question of how information about risk is communicated to the public is
crucial to the success of any overall risk management approach. The characteristics and
perceptions of the various stakeholders need to be included and evaluated throughout the
process (9). How best to communicate with interested parties in a given situation has yet
to be determined. Options include everything from direct one-on-one exchanges to
16
-------
worldwide internet linkups. In addition, the appropriate language for communication of
risk information needs to be carefully considered, since it is clear that scientists and the
public do not always share a common understanding of such terms as risk, biodiversity,
landscape, etc.
In short, there is an important role for the Agency to play in exploring better ways
to communicate risk information—including the results of Eco RAs—to the public. While
this Committee is not in a position to describe the means by which this might be
accomplished, it has identified the need.
e) Indicators: biological and others
The role of indicators in risk assessment needs to be fully explored. Indicators
may be physical or biological metrics and may be as simple as one species or as complex
as a diverse ecosystem that only occurs in a particular physical environment. An example
of an important physical indicator is a change in temperature during a critical phase of the
life cycle of the organism. An example of a physical indicator relevant at a larger scale
might be the length of shaded riparian habitat along a river segment (a measure of edge
habitat quality and extent) or the frequency of flood events large enough to transport
gravel and sand-sized sediment. The questions to be addressed here include the
following:
1) How do these indicators change with different levels of a stressor?
2) How does one measure an indicator or a surrogate for it?
3) How does one identify situations under which physical indicators may be
sufficient to represent the risk of the species at hand?
4) How might these indicators change with geographic region and/or changes in
spatial or temporal scales?
5) Are particular life stages of an organism or success!onal stages of an ecosystem
more susceptible to a stressor?
The Agency could improve the quality and future utility of ecological risk
assessment methods by developing guidance on the appropriate selection and use of
indicators.
f) QA/QC
In conducting ecological risk assessments, the assessor is often faced with a dearth
of site-specific data, compounded by a lack of time and/or resources available to collect
the needed information. In such instances, the assessor most often seeks existing data
17
-------
from other sites, organisms, or stressors to bridge the gaps. The proposed Eco RA GLs,
however, provide little guidance to help the assessor select, evaluate, and use such
"foreign" data.
Clearly, there are significant differences in the quality and utility of these foreign
data that depend, in part, on their intended use in the assessment. For example, there is a
gradient between well-conducted field and laboratory data collected specifically for a
particular risk assessment, on the one hand, and data from the literature on a different
species in a different environment, on the other.
Judgments can and should be made to clarify the situation. For example, published
values of LD50 for the same species, partition coefficients for a specific chemical of
concern, etc. can probably be used with little impact on the uncertainty of the overall
assessment. Toxicity data from related species (e.g. two different species offish) should
be used with caution for specific assessment endpoints, and data from totally different
species (e.g. birds) probably should not be used at all. The mechanisms of the effect
should be considered in making such decisions. For example, some changes to a system
may have different effects within a taxa, but similar effects for unrelated organisms with
similar behavioral patterns.
In short, there is a need for a guidance document on the selection and use of
"foreign" data in risk assessments. This guidance should address both issues of
appropriateness and of quality and how they are related to the overall uncertainty of the
assessment.
18
-------
4. ADDITIONAL POINTS
During the course of the Committee's deliberations, a number of items were discussed that
went beyond the questions in the charge per se. In this Section, the Committee presents its
observations on these additional points in the belief that attention to these issues will further
enhance the Eco RA GLs.
4.1 Use of accurate language
In a number of areas, the text makes generalizations that may not apply to all
applications of ecological risk assessments. Some statements specifically relate to
a) Prospective vs retrospective assessments
b) Assessments used for 1) chemical evaluation, 2) site evaluation, or 3) natural
resource evaluation.
c) Assessments used for 1) chemical stressors, 2) biological stressors, or 3) physical
stressors.
The document would be improved by more precisely qualifying statements according to
the contexts listed above.
4.2 Case studies
The Committee applauds the Agency for taking the time to carry out some model
ecological risk assessments in the watershed case studies. The utility of these studies to potential
practitioners of ecological risk assessment is so great that their prompt completion should be an
Agency priority.
Specifically, the Committee urges the Agency to consider carefully its watershed case
studies and the SAB's pending report on their review of that information. There are relevant
insights and advice in those documents that merit incorporation into the Eco RA GLs For
example, the discussion of the outputs of the planning phase—goals and objectives—and their
relationship to the assessment endpoints developed in the problem formulation phase could be
expanded and clarified through reference to some of the watershed case studies. Specifically, the
Eco RA GLs text currently provides a very brief definition of management goal and only mentions
the development of more specific objectives by reference to text note 2-6. Because the
development of goals and the more specific, ecologically-based objectives is so critical to the
success of the planning phase, the Committee suggests that an expanded explanation of
management goals and associated objectives be included in the Eco RA GLs, and as well as the
Waquoit Bay management goal and its ten associated objectives.
19
-------
In the Waquoit Bay example, the management goal defines the ecological value that the
community wants to protect. The objectives translate this goal into more detailed subunits that
incorporate the properties of the ecosystem that are relevant to achieving the goal. Assessment
endpoints can then be derived that not only meet the tests for ecological relevance and scientific
validity but can also be related directly back to the objectives. Using this progression, the
scientific rationale for translating the goal through the objectives into the assessment endpoints
can be illustrated quite well.. The process of translating the general goal into the objectives also
provides a focal point for productive early interaction among the risk manager, interested parties,
and the risk assessor and holds the promise of generating results that are socially valued, as well
as ecologically meaningful.
Also, once the watershed case studies are completed and the SAB has reviewed them, the
Agency should prepare a synthesis of "lessons learned" from the experience. Approaches that
proved to be effective should be highlighted, along with identification of approaches that proved
to be ineffective. The compilation of the Eco RA GLs, the case studies, and the lessons learned
will be an extremely useful compendium for those faced with conducting ecological risk
assessment.
In addition, the SAB suggests that, over time, the Agency compile a list of completed case
studies that span a wide variety of stressors, environments, and levels of complexity and scale.
This list should include both prospective and retrospective analyses of single and multiple
stressors, different sites, and various resource evaluations. Since risk assessments usually are not
published in their entirely in the open literature, compiling a list of representative studies would
provide a wide range of risk assessors and managers with access to information they might
otherwise miss.
4.3 Hypothesis testing
The Committee believes that the Eco RA GLs should have a greater emphasis on the use
of the scientific method, specifically a sharper focus on hypothesis testing and guidance for its use.
The rigor of the risk assessment process would be significantly enhanced if the problem
formulation discussion provided more and better guidance on hypothesis development and if the
analysis phase contained more and better guidance on hypothesis testing. For example, problem
formulation should articulate what questions are being asked, how the answers to those questions
will be sought, and the possible interpretations of the answers that might emerge. In many
applications, by following the hypothesis testing model in risk assessment methodology, the entire
ecological risk assessment process would be more useful, better understood, and more widely
accepted.
4.4 Adaptive Management
The concept of "adaptive management" has proven beneficial in a number of areas.
The Agency should seek to maximize the benefits derived from "an adaptive management mind
set" as it implements the new Eco RA GLs.
20
-------
First, the Eco RA GLs should make the point even more clearly that the consideration of
assessment endpoints need not be constrained by data from monitoring programs nor even to the
availability of data per se. In some cases, an ecologically relevant assessment endpoint may be
one for which data are currently unavailable, but for which data could begin to be collected. The
assessment endpoint would then be a useful tool for adaptive management, would be able to
inform future iterations of the risk assessment, and would contribute to data generation that might
be useful in other arenas.
Second, Figure 1-2 should be modified in order to emphasize the need for "technology
transfer" that will inform risk managers of research findings on a continuing basis. A general
model for the process of technology transfer is found in the health arena in which pharmaceutical
companies and other suppliers of health care related products move continuously and aggressively
to transfer findings from the research lab to clinical application. The important point is that
neither the pharmaceutical researcher (cf, risk assessor, including both conceptual modeling and
analytical work, in the present context) nor the doctor/surgeon clinician (cf. risk manager in the
risk context) performs the tech transfer function, but rather a third for-profit group. However, no
such separate group is established in the ecological risk area. For example, in Figure 1-2, an
arrow should be used to indicate the adaptive management link between analysis (i.e., ecological
response analysis) and the conceptual modeling activity. In this model, the tech transfer that
undoubtedly occurs is performed largely by the risk assessor. Yet, as the diagram clearly shows,
this is not a primary function of the risk assessor activity.
In the current draft, technology transfer is included in problem formulation; cf, Fig 1-2
depicting the integration of available data which involves linking the assessment of impacts and
conceptual models. Although not specifically characterized as such, the elements included in the
"side-bar" in the Figure; i.e., "Acquire Data, Iterate Process, Monitor Results", are also
technology transfer activities.
It is not clear that each of the technology transfers would be the sole responsibility of the
risk assessor per se; cf, the functions described on pp.29, 69, 153, and 196, for example. The
Eco RA GLs should acknowledge the need for coordinated efforts of several experts in order to
effect the appropriate data collection, data analysis, and, most important, the translation of this
new information into formats (such as conceptual models) that are useful for risk managers.
This problem could be addressed by discussing the fundamental importance of explicit
adaptive management feedback loops throughout the process. As an example that was noted
earlier, the dashed double arrow line at the bottom of the Figure should be changed to solid and
bold. The other "doubled-headed" arrows depicting feedback loops could be made more explicit.
Also, the Agency should consider modifying the text to clarify the technology transfer process
involved and to describe explicitly the agents responsible for such transfers, whether they are to
be performed by risk assessors or other parties.
21
-------
4.5 Risk Estimation Techniques
The Committee applauds the Agency's inclusion of information and references on risk
estimation techniques that go beyond the traditional "comparison of point estimates." Since risk
estimation is arguably the most critical part of the risk assessment and new techniques are being
introduced regularly, users of the Eco RA GLs would benefit from a fuller discussion of available
techniques, together with additional examples of how data have been or could be utilized to
provide risk estimates.
The discussion on mixtures in section 5.1.2 references Broderius (1991) and indicates that
overall "Caution should be used when predicting that chemicals in a mixture will act
independently of one another." While caution should be used in utilizing approaches which
attempt to accommodate component toxicity from multiple chemicals, the general conclusion to
be drawn from the literature is that on the basis of acute toxicity to aquatic organisms, chemicals
do act independently and the responses are generally additive.
Similar comments apply to the discussion section 5.1.3 also. The document should have
examples illustrating the applicability of the approach to multiple species, in addition to a stressor-
response curve (Figure 5-4 on p. 138) for a single species. While it may seem that this is an
obvious extension of what the Eco RA GLs already present, the additional detail will demonstrate
the greater depth to the risk estimation approach.
4.6 Risk Characterization
It should be recognized that the ecological risk assessment process includes a number of
science policy positions that are mandated by Congress or have been formulated by the Agency.
Examples of such science policy positions include the following:
a) The specification of test organisms and test conditions for certain tests.
b) The manner in which multiple effects measures are aggregated and summarized.
c) The selection of exposure concentrations used in quotient methods; e.g. 50th vs
90th percentile and mean-flow vs low-flow conditions.
d) The use of 10-fold "uncertainty" factors.
For this reason, it is imperative that all science policy assumptions be clearly distinguished
in the risk characterization portion of the report. In fact, these admonitions are contained in the
Agency's Risk Characterization Policy (18), which should be highlighted in the Eco RA GLs.
22
-------
4.7 Different levels of sophistication for different needs
The Eco RA GLs describe an in-depth model for an ecological risk assessment. That is,
they provide guidance for the broadest and most comprehensive of ecological risk assessment that
the Agency (or those outside the Agency) is likely to perform. In point of fact, given the
logistical constraints existing inside and outside the Agency, explicit performance of all steps and
elements of the idealized model may not be necessary, appropriate, or feasible. For example,
screening level assessments performed by the Office Pollution Prevention and Toxics in its Pre-
Manufacturing Notification reviews may not require an explicit and unique conceptual model or
analysis plan for each review.
In summary, ecological risk assessments conducted with little data, few resources, or brief
time spans may be adequate for certain applications.
4.8 Lines of evidence
In discussing the contents of Section 5.2.1 on "Lines of Evidence", the Committee noted
that confidence in risk assessment conclusions can be increased using several lines of evidence
(weight of evidence) which may be derived from different sources. Recognizing the overall
importance of the quality of the information and its utility for reaching decisions, the Committee
offers the following general guidance on the use of data from different sources (lines of evidence)
since not all data are equal in "weight" or value in supporting risk estimates:
a) The strength and value of data for risk assessments is usually derived from the
rigor of the experimental design, the careful manner in which the data are
collected, and the quality assurance and quality control built into the study.
b) Data obtained via qualitative approaches may be of less weight than similar data
obtained via quantitative approaches.
c) Data from screening level studies may be of less weight than data from more
definitive studies related to the same endpoint.
d) Data obtained by species-to-species extrapolation may be of less weight than data
for the species of interest from a similar site outside of the study area.
e) Data generated by quantitative structure-activity relationships are of less weight
than similar data obtained in well-designed laboratory or field studies.
f) Data generated by computer model simulations are valuable, but may be of less
weight than quantitative laboratory or field data, depending on the extent of the
model validation and the strength of the laboratory or field data.
23
-------
g) Data generated in laboratory studies may be of less weight than comparable data
generated from well-designed field studies.3
In general, lab studies provide for control of variables at the expense of reality, while field studies maximize reality but may offer little insight into
cause and effect, since few, if any, parameters can be controlled under field conditions. Thus, the weighting of lab vs. field data from equally well-
designed studies must carefully consider the ecological question being asked. For example, if the principal concern is the effect of a single stressor,
then a lab study might well be weighted more heavily than any field study. On the other hand, if there is clear evidence of multiple stressor
involvement or of overriding control by a complex of environmental parameters, the field may be the only appropriate place to pursue the cause and
effect possibility, since there are simply too many variables to control. The compromise, i.e., the field experiment where some subset of environmental
factors are more or less controlled, may potentially merit the heaviest weight, but only if the lines of evidence are clear and some sort of "control site"
or "control treatment" is included in the study.
24
-------
5. CONCLUSION
The Committee is generally pleased with the Agency's draft Eco RA GLs. While this SAB
report identifies a number of ways in which the document could improved, the Agency is to be
commended for carrying out this work. This seminal work will evolve over time—and
appropriately so—in response to comments and suggestions such as those in this report.
However, those further improvements and enhancements will be building upon the solid
foundation that is established in the Agency's draft.
25
-------
REFERENCES
1. For example,
a) USEPA. October 4, 1995. Proposed Guidelines for Neurotoxicity Risk Assessment,
Federal Register 60(192):52032-52056.
b) USEPA. October, 1996. Guidelines for Reproductive Toxicity Risk Assessment,
Federal Register 61(212):56274-56322.
2. Complementary examples,
a) Science Advisory Board, Review of Guidelines for Reproductive Toxicity Risk
Assessment, EPA-SAB-EHC-95-014, 1995.
b) Science Advisory Board, Review of Guidelines for Neurotoxicity Risk, in draft, 1996.
3. Science Advisory Board, Reducing Risk: Setting Priorities and Strategies for Environmental
Protection. EPA-SAB-EC-90-021, USEPA, September, 1990.
4. a) Science Advisory Board, Ecological Risk Assessment Consultation, EPA-SAB-EPEC-90-
LTR-005, 1990.
b) Science Advisory Board, Review of EPA's Ecorisk Assessment Research Program, EPA-
SAB-EPEC-92-006, 1992
c) Science Advisory Board, Review of Process and Rationale for Developing Ecological Risk
Assessment Guidelines, EPA-SAB-EPEC-92-023, 1992
d) Science Advisory Board, Commentary on Ecological Risk Assessment for the Proposed
RIA for RCRA Correction Action Rule, EPA-SAB-EPEC-COM-94-001, 1994
e) Science Advisory Board, Consultation on the Selection and Use of Case Studies in the
Ecological Risk Assessment Guidelines, EPA-SAB-EPEC-CON-95-003, 1995
f) Science Advisory Board, Review of Watershed Cases Studies to Illustrate Use of
Ecological Risk Assessment Guidelines, report underdevelopment.
5. USEPA, "Proposed Ecological Risk Assessment Guidelines", EPA/630/R-95/002B, August, 1996.
6. USEPA, "Framework for Ecological Risk Assessment Framework", EPA-/630/R-92-001, Risk
Assessment Forum, 1992.
7. Wood, Bill. August 23, 1996. Memorandum from the Executive Director of the Risk Assessment
Forum to Stephanie Sanzone, Designated Federal Official of the SAB's Ecological Processes
and Effects Committee, transmitting the charge for the SAB review.
8. Federal Register. August 23, 1996. Vol. 61, #165, p. 43545.
9. a) National Research Council, Science and Judgment in Risk Assessment. Committee on Risk
Assessment of Hazardous Air Pollutants, Board on Environmental Studies and
Toxicology, Washington, DC, National Academy Press, 1994.
R-l
-------
b) National Research Council, Understanding Risk: Informing Decisions in a Democratic
Society. Committee on Risk Characterization, Commission on Behavioral and Social
Sciences and Education, Washington, DC, National Academy Press, 1996
10. Cairns, J., Jr., Maki, A. W., and Dickson, K. L., "Estimating the hazard of chemical substances
to aquatic life", ASTM Special Technical Publication 657, American Society for Testing and
Materials, 100 Barr Harbor Drive, West Conshohocken, PA, pp. 153-187, 1978.
11. Dickerson, R.L., M.J. Hooper, N.W. Gard, G.P. Cobb and RJ. Kendall. 1994. Toxicological
Foundations of Ecological Risk Assessment: Biomarker Development and Interpretation
Based on Laboratory and Wildlife Species. Environ. Health Perspectives 102:65-69.
12. a) Harris, L.D. and P.B. Gallagher. 1989. New initiatives for wildlife conservation: the need
for movement corridors. Pages 11-34 in G. MacKintosh, editor. Preserving
communities and corridors. Defenders of Wildlife, Washington, D. C.
b) Harrison, R.L., 1992. Toward a theory of inter-refuge corridor design. Conservation
Biology 6:293-295.
c) Dale, V.H., Offerman, H., Pearson, S. and O'Neill, R.V. 1994. Effects of forest
fragmentation on neotropical fauna. Conservation Biology, 8:1027-1036.
13 a) Novitzki, R.P. 1979. Hydrologic characteristics of Wisconsin's wetlands and their
influence on floods, stream flow, and sediment. In P.E. Greeson, J.R. Clark, and I.E.
Clark (eds.) Wetland Functions and Values: The Sate of Our Understanding.
American Water Resources Association, Minneapolis, Minnesota, pp 377-388.
b) Johnston, C.A., N.E. Detenbeck, and GJ. Niemi. 1990. The cumulative effect of
Wetlands on Stream Water Quality and Quantity: a landscape approach.
Biogeochemistry 10:105-141.
14. Omernik, J.M., 1987. Ecoregions of the conterminous United States. Annals of the Association
of American Geographers. 771:118-125.
15. a) Conger, D.H 1971. Estimating Magnitude and Frequency of Floods in Wisconsin. U.S.
Geological Survey, Madison, WI. Open file Report.
b) Jacques, J.E. and D.L. Lorenz. 1988. Techniques for estimating the magnitude and
frequency of floods in Minnesota. U.S. Geological Survey, St. Paul, MN. Water
Resources Investigations Report 87-4170.
16. Oberts, G. L. 1981. Impact of wetlands on watershed water quality. In Richardson, B. (ed.)
Selected Proceedings of the Midwest Conference on Wetland Values and Management.
Freshwater Society, Navarre, MN. p. 213-26.
17. a) USEPA. May, 1995. Lakes and Reservoirs Bioassessment and Biocriteria: Technical
Guidance Document, Office of Water.
R-2
-------
b) USEPA. May, 1996. Biological Criteria: Technical Guidance for Streams and Small
Rivers, Revised Ed., Office of Water, EPA-822-B-96-001 .
18. USEPA, Memo to EPA managers from Administrator Carol Browner, "EPA Risk
Characterization Policy", March, 1995.
R-3
-------
DISTRIBUTION LIST
The Administrator
Deputy Administrator
AA Office of Research and Development
Director, National Center for Environmental Risk Assessment
Director, Risk Assessment Forum
Director, Office of Science Policy
Assistant Administrators
Regional Administrators
EPA Headquarters Library
EPA Regional Libraries
Chair, FIFRA Scientific Advisory Panel
Chair, ORD Board of Scientific Counselors
Library of Congress
National Technical Information Service
-------
January 21, 1997
EPA-SAB-EPEC-97-002
Honorable Carol M. Browner
Administrator
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
RE: Science Advisory Board (SAB) Review of the Agency's Ecological Risk Assessment
Guidelines
Dear Ms. Browner:
In its 1990 report, Reducing Risk, the Science Advisory Board (SAB) recommended that
"EPA should attach as much importance to reducing ecological risk as it does to reducing human
health risks". During much of the 1990s the Agency has responded by paying increased attention
to ecological risks.
Now, we are pleased to transmit to you the results of the SAB's review of the Ecological
Risk Assessment Guidelines (Eco RA GLs) which mark the culmination of a nearly decade-long
development process that has involved scientists both inside and outside the Agency. The goal
has been to generate a guidance document that will provide structure and discipline to the
assessment of ecological risks.
We believe that the Agency has successfully achieved that goal.
The conceptual outline for these ecorisk principles was first developed at the Pellston
workshop in 1977. At that time, the mutual importance and co-dependency of environmental
chemistry and fate testing with ecotoxicology were first identified. The current guidelines
advance ecological risk assessment well beyond that early conceptual stage to what is now a truly
functional practice.
The proposed Eco RA GLs outline the entire process and provide details on how each of the
steps of problem formulation, analysis, risk characterization, and risk management can be
successfully completed. Sufficient balance between specific guidance and flexibility is maintained
-------
so that the user can adapt the process to highly variable problems. The SAB review comments
are generally supportive of the guidelines, while offering numerous suggestions for improvement
on a wide range of issues, including risk assessor/risk manager interaction, assessment endpoints,
use of tiered assessment, field data, and multiple stressors.
The Eco RA GLs document provides a template for dealing with many environmental
programs and problems; e.g., new chemicals, water and air quality issues, waste site assessment,
and spills. As such, these guidelines, like the Agency's ecological risk assessment framework from
which they sprang, will likely be widely accepted and used by regional, state, and local
authorities, as well as by many in the private sector. Indeed, the Eco RA GLs appear destined to
be as seminal and trend-setting in their field, as the Cancer RA GLs have been in theirs. This
achievement reflects well on science at EPA and all those who were involved in the guidelines
development process.
We look forward to your response to these comments.
Sincerely,
Dr. Genevieve Matanoski, Chair
SAB Executive Committee
Dr. Alan Maki, Acting Chair
Ecological Processes and Effects Committee
------- |