August 1, 1997

EPA-SAB-EPEC-ADV-97-001

Honorable Carol M. Browner
Administrator
U.S. Environmental Protection Agency
401 M. Street, SW
Washington, DC  20460

      Subject:     Advisory on the Problem Formulation Phase of EPA's Watershed
                  Ecological Risk Assessment Case Studies

Dear Ms. Browner:

      The Watershed Ecorisk Subcommittee of the Environmental Processes and
Effects Committee (EPEC) of the Science Advisory Board (SAB) met on July 18-19,
1996 to review five Watershed Ecological Risk Assessment Case Studies being
developed under the auspices of the Agency's Risk Assessment Forum.  The case
studies were undertaken to demonstrate the utility of ecological risk assessment for
setting management priorities for watersheds. This initial review, which focuses on the
planning and problem formulation steps of the ecorisk process, is considered an SAB
advisory since it provides peer review of an Agency work-in-progress.  The goal of an
SAB advisory is to provide suggestions to the Agency for mid-course corrections that
will refine the ultimate product. In this case, the intent of an early SAB review of the
planning and problem formulation step is to facilitate the completion of high quality
case study examples. The Subcommittee expects to review the completed case
studies in late 1997,  at which time a significant number of new participants will be
added to the reviewing panel - by changes in EPEC membership and/or inclusion of
additional consultants - to ensure independent assessment of the Agency's work.

      However, following the review, Agency Staff have indicated to EPEC that efforts
to complete the cases studies all the way through the characterization phase may be
suspended. This change of plans is of concern because much of the value of the case
studies will lie in the Agency's ability to demonstrate to communities the utility of the
ecorisk assessment process in supporting decisions and in capturing lessons leaned
from these five examples.  Therefore, we urge the Agency to complete the case study
exercises and fully utilize the substantive and procedural lessons that they are sure to
yield.

-------
      Although the Agency's Framework for Ecological Risk Assessment (U.S. EPA,
1992) (ecorisk framework) and Proposed Guidelines for Ecological Risk Assessment
(U.S. EPA, 1996) describe a process for assessing ecological risks from chemical,
biological, and physical stressors, the Agency felt that these approaches should be
tested in a watershed context relevant to communities.  Place-based ecological risk
assessment highlights the need to deal with multiple stressors that affect the watershed
or place, as well as the importance of interactions between risk assessors and
community decision-makers.  The five watershed case studies are intended to provide
an illustration of the application of the ecorisk assessment process in watersheds and,
when completed,  to serve as the basis for guidance to the Agency and communities
who wish to conduct ecorisk assessment to guide community-level management
decisions.

      The materials provided to the SAB for review consisted of draft guidance on
problem formulation in a watershed context and draft case study examples for five
watersheds: Big Darby Creek,  OH; Clinch River Valley, VA; Middle Platte River
Floodplain, NE; Middle Snake  River, ID; and Waquoit Bay,  MA.  The charge to the
Subcommittee (Attachment A) contained 14 specific questions in the following six
categories:

      a)    establishing management goals (questions 1-3);
      b)    selecting assessment endpoints (questions 4-7);
      c)     selecting measurement endpoints (questions 8-9);
      d)    developing  conceptual models and the analysis plan (questions 10-12);
      e)    the overall process for watershed risk assessment (question 13); and
      f)     case study  presentation (question 14).

Although the Subcommittee reviewed all five preliminary case studies, the focus of our
comments, as requested by the Agency,  is on the overall problem formulation process
that was used by the case study teams,  especially those aspects  that depart from the
problem formulation process described in the ecorisk framework.  Differences  include
the  approach  for developing management goals and defining and selecting assessment
endpoints, and the introduction of analysis plans.

1. General Comments

      The Subcommittee agrees that case study applications of the ecological risk
assessment framework are essential to test and improve its applicability to watershed-
based analyses, and we  commend the Agency for their willingness to undertake this
effort. The case studies were well selected and represented diverse environmental  and

-------
historical situations.  Further, the composition and leadership of the case study teams
were especially good.  Consistent application of the framework across the case studies
was recognized, and the comparisons among the case studies were instructive and
useful. The efforts to secure public involvement were good, and the utility of "place-
based" and "value-initiated" risk assessments was demonstrated. We are pleased that
the Agency views the ecorisk framework as an adaptable strategy, and we agree that
the proposed changes are an improvement.

2. Establishing Management Goals (Charge Questions 1-3)

      The Subcommittee applauds the Agency for its efforts to obtain stakeholder
input to establish place-specific values.  To be most effective, however, the Agency
should develop a consistent strategy to  ensure that a full complement of stakeholders
and technical experts is brought together from the early stages of planning and problem
formulation. Although local values should clearly drive the establishment of
management goals, development of these goals  should be fully informed by the
scientific community.  For this reason, the process should include open discussion of
the ecological considerations associated with specific goals and  other ecological
attributes of the system that may be overlooked if local management goals are too
narrowly focused.

      The Subcommittee does not think that a strict separation between the planning
and problem formulation steps is desirable.  In fact, this separation may be
counterproductive to establishing appropriate goals. The interaction between  risk
assessors and risk managers, however, should in no way compromise the scientific
rigor of the risk assessment to be conducted. When the watershed risk assessment
teams include diverse viewpoints and expertise, the participants  should recognize that
different elements  of the team will take the lead during different parts of the process,
although the iterative consultation should proceed throughout.

      The effectiveness of the approaches used to define meaningful goals and
objectives varied considerably among the five case studies, demonstrating that the
success of the process depends  on the  composition of the team assembled. The
process appeared  to work best where the teams  assembled for problem formulation
involved a broad representative group of the public and individuals with technical
expertise, as in the Waquoit Bay study.  It did not appear to work as well in the Middle
Snake River study, for example, where the public was not effectively engaged.

       In some cases, the team's choice of ecologically relevant assessment endpoints
did not connect well with the management objectives. Nonetheless, the endpoints were

-------
appropriate because the selected endpoints address the integrity and sustainability of
component ecosystems independent of societal values.  One goal of any risk
assessment should be public education. Thus, for these case studies, the
environmental situation in the watersheds being evaluated should be put in context
relative to the ecological landscape and other ecosystems regardless of the selected
management goals, and this in turn may allow reassessment of these goals.

      An important first step in the planning process, and a good team-building
exercise, is for the risk assessor/risk manager team to gain a historical perspective on
the social, political, geological, and ecological history of the region, drawing upon
either the knowledge of team members or outside experts. Indeed, a sufficiently broad
team should include members with knowledge in each of these areas

      In the problem formulation stage of the risk assessment, it is important to
recognize that watersheds have many boundaries  in addition to  the topographic
boundaries of water flow.  These include their relationship to ecological, geographic,
and political boundaries. In the formulation of assessment goals, all of these
boundaries need to be recognized and considered in order to select appropriate
assessment and measurement endpoints, as well as to translate the risk assessment
into risk management.  For example,  a surface watershed is also part of an airshed and
a groundwatershed, which almost certainly will not have the same boundaries and
could have a significant impact on the watershed.  The team needs to be aware of
these other units, particularly in identifying stressors and formulating management
strategies.  In addition, a number of the case study areas represented only portions of a
watershed. For these reasons, it is critical to consider factors external to the
assessment area that may be important influences on watershed processes.

      The five case study areas, while highly variable in terms of their spatial scales,
were justified as conveniently bounded systems for assessing ecological risks.
However, the concerns focused on ecological structure and function within the
watershed (or portion of the watershed), not on ecological structure and function  at the
scale of the watershed. For example, habitat manipulation within a watershed suggests
different kinds of risks (and assessment methods) than atmospheric deposition at
scales encompassing entire  watersheds.  Thus, the Agency should clarify what is
meant by the term "watershed-level ecological risk assessment" and should
acknowledge that once the hydrologically determined boundaries are set, the scale of
the assessments is often effectively limited to the "people shed", not the watershed.
Since place-based ecorisk assessments frequently do not address a whole watershed,
but only a segment of it, watershed teams should identify and explicitly state what
portion of the watershed their assessment concerns; describe the extent to which the

-------
assessment may be applicable to other portions of the watershed; and state clearly
those issues or portions of the watershed for which the assessment is not applicable.

      Finally, it was not clear from the oral presentations or the draft case study
materials whether watershed management is to be conducted within the framework for
ecological risk assessment, or whether ecological risk assessment should support
watershed management. A distinction should be made because the two are quite
different; ecorisk  assessment provides an estimate of the adverse effects of
environmental stressors on ecological resources, whereas ecosystem or watershed
management strives to achieve specific ecological and societal goals for a regional
environment.

3. Selecting Assessment Endpoints (Charge Questions 4-7)

      The draft guidance on problem formulation for watersheds identifies three key
components for selecting assessment endpoints: a) linkage to management goals; b)
ecological relevance; and c) susceptibility to stressor.  We  believe these are relevant
selection criteria, although they do not appear to have been consistently applied by the
case study teams.  For example,  as noted in the previous section, the derivation of
assessment endpoints from management goals was evident in the Waquoit Bay case
study, but not in the Middle Snake River case study; in the  latter case, the assessment
endpoints are related to biota-fish, benthic fauna, macrophytes, and algae-while the
management goals are related primarily to water quality and quantity.

      In addition, the Subcommittee suggests a slight modification of the third criterion.
Sensitivity of response to a stressor is desirable,  especially in terms of implementing
and measuring success toward a management goal, but  it should not always be used
as an a priori criterion for excluding a given assessment endpoint.  In some instances,
the most ecologically relevant endpoint may not be the most sensitive, but should still
be included in the risk assessment. For example, if an important management goal is
the protection of bird populations, then selection of a population endpoint for
amphipods, although it might  be the most sensitive, would not be the most ecologically
relevant assessment endpoint. Thus, assessment endpoints should be selected
primarily for their ecological relevance and their ability to be measured.

      The Subcommittee supports the efforts evident in  the case studies to clarify
terminology in the ecorisk framework dealing with  assessment and measurement
endpoints. The use of the word "measures" as a replacement for the term
"measurement endpoints" appears clearer and more inclusive since it can include
stressors as well  as ecological effects. With respect to the definition of "assessment

-------
endpoints" as "entities and their attributes," the Subcommittee did not feel that this
assisted in clarification or in providing separation of assessment endpoints from
management goals. We suggest that a better definition of an assessment endpoint
should be an ecological characteristic that the stakeholders identify as being at-risk,
based on their identification of values to be protected.

      Ecological risk assessments can be performed either using known stressors for
purposes of selecting assessment  endpoints or using assessment endpoints
(associated with known or perceived effects and management goals) to build
conceptual  models and identify stressors. The latter approach was appropriately used
in the case  studies and is an  effective technique for focusing the ecological risk
assessment.  It can also be used to provide technical guidance on  achieving
management goals.

4.  Selecting Measurement Endpoints (Charge Questions 8-9)

      Data availability is certainly  one aspect to consider when selecting what to
measure in support of assessing ecological risk.  However, ecological measures should
be selected primarily based on their relevance to what has  been identified as being
potentially at risk; e.g., risk to the ecological structure and function  of the watershed as
a whole.  If the corresponding data necessary to support the risk assessment do not
exist, then they should be collected, or relevant data from similar systems that address
the issue might be substituted in a  preliminary risk assessment.  General guidance
about the selection of measurement endpoints needs to be expanded; additional points
might include selecting different and independent measures of the same assessment
endpoint, and selecting measures  at different levels of ecological organization that
address the endpoint.  For example, a population-level risk (e.g., unacceptable decline
in the production of a commercially valuable species) might be measured at the level of
the population (e.g., birth rate, death rate) or at a physiological process-level (e.g.,
consumption, respiration).

      When selecting measurement endpoints, assessors  must decide whether
segregation of the relative contributions of different stressors is an  important purpose of
the watershed-level risk assessment. Teasing apart the relative contributions of
different stressors to a single ecological effect remains an issue at  the forefront of
ecological research. Perhaps the only reasonable guidance that might be given is to
break down further the effect of interest into greater detail,  in the hope that the different
modes of action of the different stressors might become apparent.  For example, if
unacceptable increases in noxious algae are of concern, and the set  of stressors
includes nutrient enrichment,  metals, and organic contaminants,  then primary

-------
production might be examined at the level of the physiological processes that
determine it, in this case the effects of nutrients on increased photosynthesis and
growth.  This in turn might be contrasted with the relative toxic effects of the metals and
the organics. Such a perspective may allow nutrient effects to be separated from toxic
effects and might also permit examination of the indirect effects, for example, the
reduced grazing pressure on algae resulting from toxic impacts on consumer
populations in the system.

5. Developing Conceptual Models and the Analysis Plan (Charge
      Questions 10-12)

      The use of conceptual models, as envisioned in the draft guidance, and the
hierarchical nested approach leading from the general to the specific, in at least three
tiers, is quite good. However, a broader, watershed-level conceptual model would be
useful as a common starting point for all watershed case studies, present and future
(i.e.,  as part of the "lessons learned" guidelines).  Such a model might include, for
example, stream order,  long-term seasonal runoff and thermal regime patterns, and
seasonal schedule of riparian zone inputs in order to define the boundary conditions for
organism life cycles in watersheds.  Also, some of the paradigms developed to explain
the structure and function of stream river ecosystems could be  used as general models
to set the stage for watershed risk assessment.  These paradigms include the River
Continuum Concept (RCC),  Flood Pulse Concept (FPC), Riparian Influence Concept
(RIC), Nutrient Spiraling Concept, and Patch Dynamics Concept (Gushing et al., 1995;
Cummins, in press). The RCC, FPC, and RIC are conceptual models that address
generalized spatial and  temporal relationships in watersheds.  The RCC and FPC
propose a model describing  how running water ecosystems change in predictable
fashion longitudinally (RCC) and laterally with their floodplains  in the lower reaches.
The RIC conceptualizes interactions between the terrestrial fringe and running waters
along the riparian ecotone.  There are also some conceptual models that could help set
the stage for formulating risk hypotheses; e.g., the Serial Discontinuity Concept (SDC)
would have been useful in the Snake River Case Study. The SDC is applicable to an
impounded river, such as the Snake, because it addresses structural and functional
attributes that are displaced  by damming.

      Conceptual models allow for the identification of multiple stressors and facilitate
the consideration of cumulative, interactive effects.  To this end, continued
development and refinement of multi-layered conceptual models are, and  likely will
continue to be, the most effective method to ensure the development of the most
appropriate analysis plans and the identification of complex, cumulative effects.
Incorporation of an analysis  plan in problem formulation has been correctly identified

                                       7

-------
as an important step.  It is also important to examine multiple exposure pathways
simultaneously in the analysis.

6.  How the Process Can Be Improved (Charge Question 13)

      Many of the opportunities for improvements in the case study process are
related to research needs that will be identified during the ecological risk assessments.
For example, there is not a great deal of "watershed level" data available because past
research has been either narrowly focused on a specific stream or stream segment, or
broadly focused on an ecosystem. In order to improve the watershed case study
process, the Subcommittee recommends that consideration be given to the following:

      a)     watershed management requires  continued improvement in terms of
            understanding the science of watershed-scale processes (e.g., energy
            flow, productivity, and nutrient cycling);

      b)     watershed management requires more data actually collected over full
            watershed systems;

      c)     identification of multiple watershed boundaries is critical and must
            consider the distinction of ecological, geo-physical,  and political
            elements;

      d)     the influence of societal values and forces external to the watershed may
            be important in watershed processes and should be recognized;

      e)     the current case studies focus on surface-water issues - full watershed
            assessment must also consider terrestrial, atmospheric, and groundwater
            components and the important linkages among these elements;

      f)     river segments are not "watersheds" and thus their assessment must
            include upstream and downstream perspectives; and

      g)     assessments should strive to provide only the minimum  information
            necessary to inform decisions.

7.  Case Study Presentation (Charge Question 14)

      Capture of "lessons learned" will be a very important output of the case study
exercise, and while it is useful to include this information in summary documents aimed

                                      8

-------
at community-level risk assessors and managers, it is also important that these lessons
be incorporated into the Agency's Proposed Guidelines for Ecological Risk
Assessment.  Therefore, it is important that these case studies be carried through to
completion and that a special effort be made to extract the generic lessons learned and
to use these to develop a brief narrative and a template that other groups can use to
apply the place-based, ecological risk assessment approach to new situations.

      The planning and problem formulation sections of the draft Waquoit Bay case
study contain an appropriate level of detail and meet the general project objectives.
While the case study drafts represent a good start,  the Subcommittee believes much
more work is required to meet the objectives of the  individual case studies and the
overall project. For example, we offer the following specific suggestions for improving
the present documents: a) each case study report should have a parallel structure,
including an executive summary-not an abstract-up front and summary and
conclusions sections at the end of each case study report; b) lessons learned should
be highlighted and accompanied by a brief description of how they would change the
process if the study were re-started, and how they would affect future applications; and
c) all case study  reports should be written concisely, clearly, and with a minimum of
jargon.

      In addition to the case study reports designed for a science audience generally
familiar with the ecological risk assessment process, less technical documents with
minimal (or no) jargon would be necessary for other audiences (e.g., state and local
governments, community groups, and other institutions).  We recommend, therefore,
that the Agency  develop a concise, consumer-friendly, version for use by communities
that uses "plain English" to explain the concepts and principles of ecological risk
assessment and  something of the process that will be conducted. Illustrations and
graphic material would also be helpful in communicating the ecorisk approach to a
broader audience.  This non-technical version should include a brief section on the
advantages and  limitations of ecological risk assessments, points that should also be
discussed as part of the team education process at the beginning of any watershed
ecological risk assessment.

      For both the technical and non-technical writeups of the watershed case studies,
the power of the  place-based ecological risk assessment approach will be far more
evident and the reasons for using it more compelling if the basic underlying principles
and processes and the benefits are clearly understood by the reader. Relating these
concepts to other decision-making models would further strengthen the argument for
using the ecorisk paradigm, rather than weaken it.

-------
      In conclusion, the Subcommittee found the planning and problem formulation
stages of the watershed case studies generally to be well done, the assessment
endpoints selected by the watershed teams to be ecologically relevant, and the
conceptual models developed to be good summaries of the science available. We
emphasize the importance of completing these cases studies and look forward to
reviewing the completed analysis and risk characterization stages for each of the case
studies in the near future.

                              Sincerely,
                                    /signed/
                              Dr. Genevieve M. Matanoski, Chair
                              Executive Committee
      /signed/                                   /signed/
Dr. Mark A. Harwell, Chair                   Dr. William H. Smith, Chair
Ecological Processes and                   Watershed Ecorisk Subcommittee
 Effects Committee
Attachments
                                      10

-------
                           LITERATURE CITED
Cummins, K.W.  In Press. The River Continuum Concept: A Worldwide Model for
      Running Water Ecosystems. Prospettive di Ricerca in Ecologia delle Acque,
      Institute di Ricerca sulle Acque del C.N.R., Italy.

Gushing, C.E., K.W. Cummins and G.W. Minshall (eds.). 1995. Ecosystems of the
      World, 22. River and Stream Ecosystems. Elsevier, Amsterdam.  882p.

U.S. Environmental Protection Agency. 1992.  Framework for Ecological Risk
      Assessment. EPA/630/R-92/001, February 1992.

U.S. Environmental Protection Agency. 1996.  Proposed Guidelines for Ecological Risk
      Assessment. EPA/630/R-95/002B, August 1996.
                                    R-1

-------
              ATTACHMENT A: CHARGE TO THE SUBCOMMITTEE

               WATERSHED ECOLOGICAL RISK ASSESSMENT

                       Science Advisory Board Review
                               18-19 July 1996

Subject:     Review of watershed level ecological risk assessment case study
            planning and problem formulation

Requesting Organizations:    Risk Assessment Forum

Contact:     Suzanne Marcy, Technical  Panel Chair
            National Center for Environmental Assessment, ORD (8601)
            (202) 260-0689

      The Risk Assessment Forum is requesting the review of documents describing
and illustrating the process of watershed ecological risk assessment in the form of five
case study examples and a summary of what was learned from their development. This
is the first part of a two stage review process. The initial review (termed an SAB
Advisory), will focus on the planning and problem formulation sections to obtain early
feedback on the process used to: (1) obtain watershed level management goals;  (2)
interpret the goals for the risk assessment; and (3) define the problem at the landscape
scale for multiple stressors,  including a plan for analyzing data and characterizing risk.
SAB review of the process used  for planning and problem formulation will facilitate the
completion of quality case study  examples.  A second review will be requested in  FY97
to focus on the analysis and risk characterization component of the case study
examples.

BACKGROUND

      As the Agency shifts emphasis from command and control toward voluntary
compliance and community-based environmental protection, it becomes critical that
EPA provide the scientific basis for community-level management decisions.  States
and local organizations need a process and tools they are able and willing to use for
determining what ecological resources are at risk and how best to protect those
resources through management  action.

      In 1992, the Agency published the Framework for Ecological Risk Assessment
                                    A-1

-------
(Framework). The Framework provided a good foundation for conducting ecological
risk assessments for individual chemical or physical stressors.  However, it does not
address risk from multiple chemical, physical and biological stressors that combine
within watershed ecosystems, the kinds of problems communities will need to manage.
The case studies were initiated to address this need. They were intended to provide a
process for learning how to apply the principles of ecological risk assessment to a
watershed scale problem, and when completed, provide case study examples of the
process.  Lessons learned from case study development were intended to be
incorporated into future Agency Guidelines for Ecological Risk Assessment.  Because
of interest by the Office of Water in watershed protection, and the Risk Assessment
Forum in testing and expanding the ecological risk assessment Framework, the case
studies were jointly sponsored by these offices to illustrate the application of the
Framework to watersheds impacted by multiple stressors.

      The project officially began in September 1993 with the initiation of five
watershed ecological risk assessment case studies:  Big Darby Creek, OH; Clinch
River, VA; Middle  Platte River Wetlands, NE; Snake River, ID; and Waquoit Bay
Estuary, MA.  Watersheds were nominated for inclusion in the project by organizations
concerned about the watershed. Specific selection criteria included data availability,
willingness of local professionals to participate, diversity of stressors, and the presence
of significant and unique ecological values. The case studies are being conducted by
interdisciplinary and  interagency teams of risk assessors and risk managers.  The
projects are organized to approximate the kinds of expertise, resources, and data likely
to exist in communities that would use this guidance.

CHARGE

      We request that the Committee focus primarily on the process used to conduct
problem formulation at the watershed level, including those aspects of the process that
depart from that described in the Framework. These include changes in: the process
based on why the  risk assessment is initiated, how management goals were  developed
and defined, selection and definition of assessment endpoints, development of
conceptual models for watersheds  and multiple stressors, and the introduction of
analysis plans.  Suggestions for improving individual case studies, while valuable, are
considered tangential to the primary intent of the review. The Waquoit Bay case study
will be used to illustrate the process of planning and problem formulation, with
additional examples being drawn from the other case studies where appropriate.
Committee members should read the entire Waquoit Bay Watershed case study and at
least one other case study in full, and the executive summaries for each case study.
                                      A-2

-------
      We ask the Committee to evaluate the following specific issues:

Establishing Management Goals

1)    The watershed risk assessment teams found it essential to work with local, state
      and federal managers and scientists to generate management goals prior to
      conducting the problem formulation stage. To emphasize the importance of this
      step, we established a full  "planning" section in the case study which describes
      the goal and how it was derived.  How well does this process make the link
      between risk management and risk assessment? How responsive is the  problem
      formulation to the identified management goals?

2)    In the Framework, interactions between managers and risk assessors (planning)
      was considered to be outside of risk assessment and was not emphasized. We
      have maintained this separation between planning and problem formulation to
      help ensure the scientific foundation for the risk assessment while being
      responsive to management concerns. Currently within the Agency, there is
      interest in  adopting the expression "problem formulation" and redefining it as a
      combination of planning and problem formulation.  How important is the
      separation of risk management from the scientific evaluation of ecological risk?
      How effective was the approach used in the case studies in maintaining
      separation but ensuring management input?

3)     Goals  for  a watershed or other landscape ecological unit, when generated by a
      diverse constituency, are often very general and broad. Teams confronted with
      these general goals defined their meaning by generating sub-goals or objectives
      that made  implicit assumptions in the goal explicit. How effective was this
      process in defining the goal in ecologically relevant ways? What other
      approaches may do this more effectively?

Selecting Assessment Endpoints

4)    Assessment endpoints were defined based on the management goals and
      objectives.  Each assessment endpoint was evaluated  by comparing its
      relationship to specific management objectives, and  by determining its
      ecological  relevance and susceptibility to stressors within the watershed. How
      effective is this approach for selecting a quality set of assessment endpoints?
      How well do the assessment endpoints reflect their respective sub-goals?
                                     A-3

-------
5)    Because of the breadth of the goals and potential assessment endpoints, teams
      had difficulty selecting which set of assessment endpoints to use. Some are
      specific, some may be better as measures, some remain fairly broad.  What
      guidance can reviewers provide on the level of specificity appropriate for
      assessment endpoints and the best complement of endpoints to use for this type
      of risk assessment?

6)    Assessment endpoints are defined as entities and their attributes (e.g., bird
      reproduction, wetland areal extent). This definition is more specific and differs
      from other definitions (e.g., USEPA, 1992; Suter,  1993). It was used to clarify
      and focus the problem formulation on definable characteristics that were not
      goals or adverse effects. The definition is used to clarify the focus of the
      assessment, and to separate assessment endpoints from management goals.
      How effective is this definition in providing separation from management goals?
      How well do assessment endpoints defined in this way establish a framework for
      generating appropriate conceptual  models?

7)    Assessment endpoints were selected prior to targeting stressors in the
      watershed.  This was a significant,  although seemingly simple, change in the
      process of ecological risk assessment where  generally stressors and observed
      affects initiate the process instead of ecological values.  This provided a basis
      for developing conceptual models containing  multiple stressors.  How effective
      was this approach for focusing the assessment on management goals, and
      multiple stressors?

Selecting Measurement Endpoints

8)    The Framework describes the selection of measurement endpoints, intended to
      be measures of the effects on assessment endpoints from exposure to stressors,
      as occurring in direct conjunction with the selection of assessment endpoints.
      However, as the case studies were developed, greater success was achieved
      where teams identified an array of measures  of effect, exposure, and ecosystem
      and life history characteristics at a later stage, during conceptual model
      development and analysis planning. This approach helped to prevent the
      selection of measures based primarily on data availability, and targeted other
      measures important to the process.  Do these measures encompass the array
      needed for these risk assessments and is there a more systematic approach for
      their selection that the reviewers can identify?
                                      A-4

-------
9)    One challenge of value-initiated risk assessments is to ensure that there are
      good conceptual connections between the assessment endpoints and the
      measures chosen to evaluate their potential response to stressors. This is
      particularly problematic when there are multiple stressors that may combine to
      cause the same type of response in an assessment endpoint.  There  has been
      mixed success in the case studies in achieving distinctions among stressors
      through the measures chosen. What guidance can reviewers provide to
      overcome this problem?

Developing Conceptual Models and the Analysis Plan

10)   The teams approached the development of watershed conceptual models at
      multiple levels. Teams developed a conceptual model for the watershed as a
      whole, featuring the entire array of identified sources, stressors, ecological
      effects and assessment endpoints. These models were augmented by a second
      level of conceptual models that featured a single assessment endpoint, and all
      the stressors relevant to that assessment endpoint.  More detail on processes
      and relationships was included at this second level. A third level is also
      appropriate, though not well  represented in the case studies,  where the
      pathways for one stressor are shown for one assessment endpoint. What mix of
      conceptual models should be developed to best  represent the watershed
      ecosystem,  communicate important relationships, and highlight important risk
      hypotheses?

11)   One of the principal objectives of these case studies is to feature the
      assessment of multiple stressors.  The potential for interaction among multiple
      stressors is primarily represented in the conceptual models. Much work remains
      to be done to develop analysis plans that help us evaluate the combined and
      cumulative effects of multiple diverse chemical,  physical and biological
      stressors. What guidance can reviewers provide on the process, conceptual
      models and analysis plans to help evaluate combined and cumulative effects?

12)   One significant addition to problem formulation was the incorporation of an
      analysis plan. Although implicit in the Framework,  the need for an explicit formal
      plan seems essential to problem formulation for watershed risk assessments. As
      currently  envisioned, the plan would lay out specifics on data, analyses and their
      expected results (including presentation format for management decisions),
      uncertainties and research needs. In the Committee's view, what are the critical
      elements of such an analysis plan?
                                     A-5

-------
Overall Process

13)   A multi-step winnowing process was developed to help focus the risk
      assessment from the broad management goals to a plan for data analysis.  The
      process included the development of management sub-goals and objectives,
      definition of assessment endpoints, development of conceptual models and
      prioritization of assessment endpoints during analysis planning.  How effective
      was this process? How could it be improved?  What other factors or steps could
      assist the teams in their selection process?

Case Study Presentation

14)   The case study write-ups are intended to communicate to a diverse audience,
      including the scientific community, risk assessors and managers, and members
      of the local watershed communities.  The write-up is intended to clearly describe
      the process of risk assessment, including how and why decisions were made
      and results obtained. The document should also educate decision-makers about
      potential risks to the watershed and provide the basis for management action.
      Finally, the document must be  scientifically valid.  How well do the planning and
      problem formulation sections of the draft Waquoit Bay case study fulfill these
      objectives? Does the draft contain the appropriate level of detail? What other
      format would fulfill these objectives more effectively?
                                     A-6

-------
                                   NOTICE
      This report has been written as 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 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, nor of other agencies in the
Executive  Branch of the Federal government, nor does mention of trade names or
commercial products constitute  a recommendation for use.

-------
                      EPA SCIENCE ADVISORY BOARD
            ECOLOGICAL PROCESSES AND EFFECTS COMMITTEE
                  Watershed Ecological Risk Subcommittee

CHAIR
Dr. William H. Smith, Professor of Forest Biology, Yale University, New haven, CT

MEMBERS
Dr. William Adams, Kennecott Utah Copper Corp, Magna, UT

Dr. Steven M. Bartell, SENES Oak Ridge, Inc., Oak Ridge, TN

Dr. Kenneth W. Cummins, South Florida Water Management District, W.  Palm Beach,
FL

Dr. Virginia Dale, Oak Ridge National Laboratory, Oak Ridge, TN

Dr. Mark A. Harwell, Rosenstiel School of Marine and Atmospheric Science, University
of Miami, Miami, FL

Dr. Carol Johnston, Natural Resources Research Institute, University of Minnesota,
Duluth, MN

Dr. Alan W. Maki, Exxon Company, USA, Houston, TX

Dr. Anne McElroy, SUNY at Stony Brook, Stony Brook, NY

Dr. Frederick K. Pfaender, University of North Carolina, Chapel Hill, NC

Dr. Jerry Schubel, New England Aquarium, Boston, MA

Dr. Terry F. Young, Environmental Defense Fund, Oakland, CA

SCIENCE ADVISORY BOARD STAFF
Ms. Stephanie Sanzone, Designated Federal Official, US EPA, Science Advisory Board
(1400), 401 M Street, SW, Washington, DC 20460

Ms. Connie Valentine, Staff Secretary, US EPA, Science Advisory Board  (1400), 401  M
Street, SW, Washington,  DC 20460

-------
                      DISTRIBUTION LIST

Administrator
Deputy Administrator
Assistant Administrators
Deputy Assistant Administrator for Science, ORD
Director, Office of Science Policy, ORD
EPA Regional Administrators
EPA Laboratory Directors
EPA Headquarters Library
EPA Regional Libraries
EPA Laboratory Libraries
Library of Congress
National Technical Information Service
Congressional Research Service

-------