UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
                             WASHINGTON, D.C. 20460
                                                               OFFICE OF THE ADMINISTRATOR
 EPA-SAB-EPEC-ADV-94-001                                    SCIENCE ADVISORY BOARD

 September 26, 1994

 Honorable Carol M. Browner
 Administrator
 U.S. Environmental Protection Agency
 401 M Street, S.W.
 Washington, D.C. 20460

      Subject:    Advisory on the Development of a National Wildlife Criteria
                  Program

 Dear Ms. Browner:

      On April 27-28, 1994, the Wildlife Criteria Subcommittee of the  Ecological
 Processes and Effects Committee (EPEC) of the Science Advisory Board (SAB) met
 in Washington, D.C. to hear briefings and engage in a consultation with Agency
 representatives on plans for the development of a national methodology for
 developing wildlife criteria.  Although SAB consultations typically last only a few
 hours and do not result in a formal report, at the request of Agency staff from the
 Office of Water, we spent nearly seven hours in briefings and  discussion with
Agency staff on this topic.  Our sense from Agency staff was that a proposed
 national methodology was unlikely to come to the SAB for another year or so.
Therefore, based  on  these discussions and our review of the  background materials
provided to  the Subcommittee, we would like to provide you with our assessment
of the overall program and  the approaches being considered  for developing wildlife
criteria.

      We congratulate the Agency managers and researchers for assembling
thoughtful and concise background materials to facilitate our discussions.  The
presenters were enthusiastic about the topic of wildlife criteria, prepared insightful
questions to guide our consultation,  and assembled a well-organized set of briefing
materials.  Clearly, a main strength of the wildlife criteria development effort is
ths multidisciplinary nature of the Agency staff and contractors involved.  In large
part, the future success of the program will relate to how well the group
maintains communication as the chemical, physical and biological elements
necessary to develop relevant criteria are assembled.
                                                                     Recycled/Recyclable
                                                                     PrtnMd on p«p«r that contain*
                                                                     at toatt 75% ncydad Oxr

-------
       Prior to the consultation in April, the Subcommittee was provided with. ~.
 set of 34 specific questions (enclosed) regarding the development of wildlife
 criteria.  Although we have not attempted to answer each of these,  our comme:..
 are grouped in seven major categories which summarize the questions posed.

 1.  ECOLOGICAL RISK ASSESSMENT AND THE DEVELOPMENT OF
 WILDLIFE CRITERIA

       Recommendation:  The wildlife criteria program should be guided by the
       Agency's Ecorisk Assessment Framework, which provides a paradigm for
       considering integrated risks to wildlife populations from both chemical and
       non-chemical stressors and multiple routes of exposure.  In particular,
       questions arising from the problem formulation phase of ecorisk assessment
       should be addressed.

       In 1992, several committees of the SAB  reviewed technical documents
 supporting the Great Lakes Water Quality Initiative (GLWQI), including those
 portions which related to wildlife  criteria for the Great Lakes.  A number of the
 recommendations in  the final report (EPA-SAB-EPEC/DWC-93-005)  are relevant to
 the national wildlife  criteria program (e.g., the relationship of wildlife criteria to
 ecological risk assessment, protection of populations vs. individuals,  and the need
 to consider additional wildlife species) and are reiterated by the current
 Subcommittee in this letter.

      The SAB report on the GLWQI recommended that wildlife criteria based on
 aquatic routes of exposure should be considered in the broader context of
 ecological risk assessment.  In other words, the GLWQI, which focuses on point
 sources of pollution,  should be related to an ecosystem  management approach
 which incorporates multiple pathways of exposure and additional endpoints  of
 effects for wildlife.

    - We recognize that the Agency staff associated with the wildlife criteria
program are operationally familiar with  the Agency's landmark document,
Framework for Ecological Risk Assessment (EPA/630/R-92/001), commonly referrei
to as  the Ecorisk Framework. Although Agency staff made reference to evolving
linkages between the Ecorisk Framework and the national wildlife criteria effort,
we urge that  the Framework be used more directly.

      The Ecorisk Framework provides  the essential and most relevant outline for
development of science-based wildlife criteria.   The current phase of the national
wildlife criteria program is  directly analogous to the initial problem  formulation
phase of ecorisk assessment.  Careful initial scoping of relevant issues and

-------
 questions, through a thorough review of available toxicological field data for
 wildlife populations, will focus subsequent assessments of exposure and effects "• y
 clarifying issues of uncertainty, prioritizing  long-term research needs, and
 providing the framework for integration of laboratory and field data.  Such an
 exercise will also address external criticism  of the need for wildlife criteria.

       The Agency's guidance  for developing wildlife criteria should discuss the
 relative risk of chemical and non-chemical stresses on wildlife populations.  In
 keeping with the tenets of the problem  formulation phase  of the Ecorisk
 Framework, the Agency needs to clearly indicate where and when wildlife
 populations remain at risk when existing criteria for aquatic life are met.  A
 critical review  of real-world  case studies should be conducted in the future to
 identify examples where wildlife populations are at incremental risk.  Considering
 wildlife criteria in the context of ecorisk assessment automatically provides a
 framework for considering that chemical stressors are but  one  stress imposed on
 wildlife populations.  Additional  stressors may include habitat loss/fragmentation,
 introduction of exotic species,  and various biological and physical stressors. In
 cases where these non-chemical stressors are more significant than chemical
 stressors, wildlife criteria values based on chemical stressors may be insufficient ;o
 protect wildlife  populations.

       Preliminary knowledge of the properties of the perturbation or chemical of
 concern can be  used to decide which endpoints are likely to be most relevant (e.g.,
 hydrophobic, tightly-sorbed,  non-persistent chemicals are not likely to be
 significant for exposure via aquatic prey species and non-bioaccumulative
 compounds are  unlikely to impact higher trophic levels).

       In addition, chemical stressors for wildlife may have multi-media routes of
 exposure (e.g., soil and atmospheric routes, as well as water).  For example,
 researchers  have found PCB/DDT/DDE loadings  in the forest soil at Hubbard
 Brook Experimental Forest in New Hampshire to be higher than sediment and
 water-column loadings to Long Island Sound . In areas with important non-
 water  exposure  of wildlife to toxic chemicals, exclusive examination of water-based
 wildlife screening could seriously underestimate wildlife  population risks.
             W.H.,  R.C.  Hale,  J.  Greaves,  and  R.J.  Huggett.   1993.   Trace
Organochlorine Contamination of the Forest Floor of the White Mountain National
Forest, New Hampshire.  Environmental Science and Technology, vol. 27, pp. 2244-
2246.

-------
 2.  ALTERNATE APPROACHES AND METHODS FOR DERIVING WILDLCF I1,
 CRITERIA

       Recommendation:  Both approaches proposed by the Agency for deriving
       wildlife criteria, 1) an adaptation of the human health non-cancer approach
       proposed in the GLWQI and 2) a modification of the aquatic life criteria
       approach, are in the early stages of development, but each holds promise
       and should be pursued further.

       The primary approach presented to the Subcommittee is based largely on
 the proposed GLWQI and is an adaptation of the human health non-cancer
 approach with an uncertainty factor to account for interspecies variability.  The
 majority of our comments relate to that approach, which, while clearly in the early
 stages of development, promises to be an innovative and valuable new  method for
 understanding the fate and effects of contaminants in the  environment.  An
 alternate method based on modification of the aquatic life criteria approach  was
 also presented to the Subcommittee and appeared worthy of further research.  At
 this early stage in methodology development, we encourage the Agency to continue
 pursuing both options.

       While the research staff involved in the wildlife criteria program are aware
 of the regulatory framework into which this approach must ultimately  be
 coordinated, we think it is important to explicitly state how this developmental
 effort will evolve into the determination of criteria and enforceable standards.  For
 example, the Agency should  decide up-front on the level of uncertainty (i.e.,
 variability in the data sets and level of statistical confidence) which is acceptable
 for specific uses of the methodology. This information in turn will help define
 data requirements  and research needs.

      Another option discussed at  the meeting involved the use of egg-injection
 studies to assess the effects of chemicals on the avian embryo.  In most cases,
 embryonic development represents  the most sensitive stage in reproduction.
 However, extreme care must be taken in interpreting data on embryo toxicity
 derived from such studies.  Egg-injection studies ignore many important
physiological and behavioral factors which affect exposure  (e.g., fertility and
nesting behavior).  In addition, the female may incorporate the chemical of
concern into the egg in a different manner from that used in the  egg-injection
study.  For example, fat-soluble chemicals are deposited into the yolk portion of
the egg which has  a metabolic regime different from that of the albumen. In
short, experimental results may not accurately reflect reponses that would occur in
nature. Other sources of experimental error include excess mortality in the
embryos  due to poor injection technique and adverse  effects caused by the vehicle

-------
employed as a carrier.  In summary, we do not recommend the use of these
studies since more valid data on effects can usually be obtained by using natural
transfer of the chemical of concern from the female to the embryo and using
residue analysis to quantify exposure.

3.  NATIONAL VS. REGIONAL WILDLIFE CRITERIA

       Recommendation:  The objective of the Agency's wildlife criteria program
       should be to develop a national methodology which can hi turn be used to
       derive regional or site-specific wildlife criteria.

       Development of national criteria per se would not be defensible  given the
many factors affecting regional and local wildlife populations.  For example,
feeding ecology and behavior of a particular species can vary markedly in  different
geographic locations and may exceed interspecies variability  in some  instances.
The natural history work included in the Agency's proposed long-term research
strategy will be important to define  appropriate geographic areas for given wildlife
criteria.  This effort should be coordinated with the activities of the National
Biological Survey.  In addition, although one tends to  think  of regional or
geographically-based criteria, protection of some species requires consideration of
migratory patterns and home ranges.  This may result in a need for  criteria
specific to  species groups  in some circumstances.

4.  MODEL VALIDATION

       Recommendation: The Agency should consider  the use of hind-cast testing
       or case studies using existing  data for well-studied  chemicals of interest to
      validate  the models and methodologies and to focus future research needs.

      Many of the proposed  models  and algorithms for developing wildlife criteria,
and indeed the entire approach, can be fully  evaluated using existing toxicology
data. _An initial focus on  data-rich scenarios (e.g., the well-studied effects  of
DDT/DDE on eagle reproduction and the effects of selenium on waterfowl
reproduction) would allow the theoretical basis of the  model to be evaluated, as
well as its applicability  to real-world events.  Also,  this use of case studies  should
help focus  future research needs  for the development of wildlife criteria.  As noted
by Agency researchers,  however,  data cannot always be used directly from peer-
reviewed literature, but instead must be reviewed for quality assurance and quality
control.  Many published toxicological studies do not fully confirm  or measure dose
to the  organism, relying instead on unconfirmed estimates of exposure.  We
recommend direct consultation with  authors of published  studies when possible to
double-check accuracy of data prior to its use by the Agency.

-------
 5.  CHARACTERIZATION OF TOXIC EFFECTS

       Recommendation:  The focus of the wildlife criteria program should remain
       on the protection of wildlife populations (as opposed to individuals) from the
       direct effects of stressors, although consideration of indirect effects should
       be added hi the future.  Methods to refine both the benchmark dose and
       calculated NOAEL should proceed.  Wildlife criteria approaches should
       consider both chemical and biological transformations of chemicals when
       characterizing toxic effects.

       The Subcommittee strongly endorses the focus on population rather than
 individual organism effects, except in special cases such as endangered species.
 This emphasis  was also recommended in the previous SAB review of the GLWQI
 (EPA-SAB-EPEC/DWC-93-005).  We recognize, however, that due to the difficulty
 of measuring wildlife population effects, it is sometimes necessary to  measure
 effects on individuals and extrapolate to population-level effects.

       One method of protecting wildlife populations within the human non-cancer
 risk assessment paradigm is to use toxicological endpoints which can be related to
 population effects. Of the potential endpoints under  consideration by the Agency,
 reproductive endpoints (e.g., number of viable young per female) are clearly the
 most relevant and most readily related to population  effects and are therefore the
 clear first choice when data are available or are being generated.  Mortality,  while
 also a  useful endpoint, is not always related to the response of a population to a
 stress.  Growth and development endpoints are even  harder to relate to population
 effects, and published data are often not comparable due to differences in
 experimental design.  Growth and development endpoints, therefore, are not
 recommended for use when other alternatives exist.

      Given the early stage of wildlife criteria development, we agree with the
Agency's initial emphasis on direct effects of chemicals on wildlife versus indirect
 effects.  Due to the better data sets available on direct vs. indirect effects, this
 focus should improve  the chances for success in initial model development.
 However, indirect effects of chemicals on wildlife (e.g., via impacts on habitat) are
 extremely important  and should be considered in the  context of the ecorisk
 assessment paradigm  as model development becomes more sophisticated.

      With  regard to the use of no observed adverse  effect level (NOAEL) or
 lowest observed adverse effect level (LOAEL) values  to establish criteria, the SAB
previously recommended that the Agency develop guidance for the selection of
NOAELs appropriate  for protection  of wildlife populations as distinct from the
protection of individuals  and cautioned that uncertainty factors relating LOAEL to

-------
 NOAEL are highly dependent on the dosage-spacing used in the chronic
 studies (EPA-SAB-EPEC/DWC-93-005).  In other words, NOAEL values are
 sometimes extrapolated from effects data, rather than being based on direct
 observations.  LOAEL or benchmark dose values, if based on  actual measured
 effects, represent an adverse effect level, but provide a value with minimal
 uncertainty. Since no clear preference for a benchmark dose  vs. a calculated
 NOAEL currently exists among human and ecological  toxicologists, methods  to
 refine both approaches should proceed.

       The methodologies presented by the Agency (the human health non-cancer
 algorithm and the aquatic life criteria-derived method) use effects terms whica
 characterize the animal's response  to a chemical insult.  Missing from both
 methods is  an explicit factor acknowledging chemical and biological
 transformations of the chemical that may occur between the time the chemical
 enters the environment and the time the organism  is exposed. Such
 transformations may substantially alter the severity and nature of the effect  and/or
 the bioavailability of the compound. Selenium and  mercury are cases in point:  in
 aquatic environments, both are naturally transformed from inorganic to organic
 forms which are far more toxic. We recommend that these  potential pathways be
 recognized explicitly within the effects and exposure terms for both methodologies;
 when the approach is broadened to include non-bioaccumulating chemicals, the
 bioaccumulation factor (BAF) term will  not serve this purpose.  Other
 transformations which should be included in the physiologically-based toxicokinetv;
 (PB/TK) models, are those occurring within the animal that produce metabolites
 whose effects may differ  from those of the parent compound.  In the 1992 review
 of the GLWQI, the SAB  referred to the need to consider chemical speciation,
 bioavailability, and persistence in the environment when establishing criteria;
 these considerations remain valid in the  development of a national wildlife criteri .
 methodology.

      The uncertainty factors included in the exposure term represent at least two
 qualitatively different types of uncertainty.  One type,  exemplified by the
 interspecies uncertainty factor, will be relevant to any  methodology and therefore
 deserves considerable  attention.  The current attempts to establish a range of
 uncertainty and guidance for deriving an uncertainty factor  for specific  chemicals
 should be considerably refined and  augmented.  Another type, exemplified by the
LOAEL to NOAEL uncertainty factor, is necessary only where there is  a paucity
of relevant data.  In this case, a large uncertainty factor probably renders the
resulting criterion indefensible,  and suggests that the Agency is better served
improving the data set rather than expending substantial resources bounding the
uncertainty factor.

-------
        Allometric relationships can provide useful ways to analyze data but should
  not be the sole basis for selection of an interspecies uncertainty factor. Many
  examples can be proffered that render such an approach invalid.  For  example,
  animals  with similar body weight can exhibit very different sensitivities to a
  chemical, as seen in the case of hamsters, which are very resistant to  dioxin,  and
  guinea pigs, which are very  sensitive to dioxin. We recommend that the Agency
  collect toxicological data on  a greater number  of wildlife species to improve the
  ecological relevance  of an interspecies uncertainty factor.

  6.  CHARACTERIZATION OF WILDLIFE EXPOSURE

        Recom mendation: While we agree with the initial focus on persistent,
        bioaccumulative, organic contaminants, we recommend that the wildlife
        criteria methodology be developed and tested for other chemical groups, in
        other  ecosystems and  regions than the Great Lakes, and for additional
        wildlife species.

        During the initial stages of wildlife  exposure characterization, it is
  imperative that expert opinion and professional judgement be used to determine,
  to the extent practical, what exposure will dominate, which species are likely to  be
  maximally exposed, and which endpoints will be most ecologically meaningful  at
  the population level.  As mentioned previously, chemical, physical, and biological
  transformations which may alter either the fate of the compound or its biological
  effects must  also be  considered. We note that  the Agency is focusing its initial
'  wildlife criteria efforts on persistent organic contaminants that are bioaccumulated
  by wildlife primarily through trophic transfer through aquatic systems. While this
  is an appropriate place to begin, the wildlife criteria methodology should be
  developed and tested on other chemical groups as well, since exposure  pathways,
  chemical forms, and primary impacted species may differ with other categories of
  contaminants.  The methodology should also be tested in ecosystems and regions
  other than the Great Lakes.  Specifically, trace element  exposure  in a western
  ecosystem might be used to test the robustness of the models on a qualitatively
  different system.

        The wildlife criteria program is currently focusing exclusively on avian and
 mammalian species due to  the lack of toxicological data  for reptiles and
 amphibians.  We  agree that development of wildlife criteria at this time should
 focus on species which have  substantial databases and upon which considerable
 toxicological  understanding already is available. However, as stated in the SAB
 report on the GLWQI, the Agency should consider incorporating reptiles and
 amphibians into the  program in the future so that wildlife criteria derived with
 the national  methodology will adequately protect these species.  The importance  of

-------
 developing empirical databases for these species should be recognized in the
 Agency's long-term research agenda.

 7.  LONG-TERM RESEARCH NEEDS

       Recommendation:  In addition to continuing Agency research in the
       identified high-priority areas, the Agency should take greater advantage of
       wildlife toxicology expertise in professional societies and at other federal
       agencies.

       The Subcommittee strongly endorses the Agency's four proposed research
 areas to support development  of science-based water quality criteria to protect
 wildlife:

       a)     Integrated Wildlife Population and Toxicology Research
       b)     Species and Exposure Extrapolation Research
       c)     Watershed Characterization and Diagnosis Research
       d)     Natural History Surveys

 These research areas are consistent with recommendations made at a recent
 conference and book sponsored by the Society of Environmental Toxicology and
 Chemistry (SETAC) on the need to integrate wildlife population and  toxicology
 research .  As noted previously, we urge  that model development and wildlife
 criteria promulgation be consistent with the Agency's ecorisk assessment paradigm.

       In order to  define the need for additional research on toxicological endpoirv. j
 and on the relevance of test species to target wildlife species, we urge the Agency
 to establish criteria for judging a priori the level of variability in the  data sets
 which is acceptable for various policy and regulatory applications of the wildlife
 criteria methodology.  As mentioned, we also recommend research on additional
 types of chemicals and ecosystems other than Great Lakes, as well as research to
 develop toxicological data on amphibians  and reptiles.

      Finally, since wildlife biology/toxicology expertise within the Agency is
 limited, we encourage  the Agency to take greater advantage of the professional
 expertise that is available outside the Agency.  In particular, professional societies
 can work in partnership with the Agency to define critical scientific issues, develop
 research needs and strategies, plan and prioritize research goals,  and provide
    "Kendall, R.J. and T.E. Lacher, Jr., Editors.   1994.   Wildlife Toxicology and
Population Modeling:   Integrated Studies of Agroecosystems.   CRC Press/Lewis
Publishing, Boca Raton, Florida.  579 pp.

-------
 external peer review.  These services are often available at nominal or no cost to
 the Agency.  Relevant professional societies include SETAC  (ecotoxicology),  Society
 for Toxicology (toxicology), American Society for Microbiology (biogeochemistry,
 bio transformation), and the American Chemical Society (analysis, chemical
 reactions).

      Additionally, we encourage EPA scientists and managers to coordinate their
 programs with scientists in other federal agencies;  science relevant to the
 development of wildlife criteria is also taking place  at the U.S. Fish and Wildlife
 Service, NOAA, Department of Defense, Department of Energy, U.S. Forest
 Service, National Institute of Environmental Health Science, and U.S. Department
 of Agriculture.

      We appreciate the opportunity to hear  about  the Agency's progress in
 developing a national methodology for deriving wildlife criteria, and we hope our
 comments will assist the Agency in that effort.  We look forward to your response
 on the issues raised and to further review of the Agency's national methodology
 for wildlife criteria as  it develops.

                               Sincerely,
                               Dr. Genevieve M. Matanoski, Chair
                               Executive Committee
Dr. Kenneth L. Dickson, Chair        Dr. Alan W. Maki, Chair
Ecological Processes and              Wildlife Criteria Subcommittee
  Effects Committee
Enclosure
                                      10

-------
              U.S. ENVIRONMENTAL PROTECTION AGENCY
                       SCIENCE ADVISORY BOARD
          ECOLOGICAL PROCESSES AND EFFECTS COMMITTEE
                  WILDLIFE CRITERIA SUBCOMMITTEE

                            April 27-28, 1994

                                ROSTER
CHAIR

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

MEMBERS/CONSULTANTS

      Dr. Ronald J.  Kendall, Institute of Wildlife and Environmental Toxicology,
      Clemson University, Pendleton, SC

      Dr. Anne McElroy, State University of New York at Stony Brook, Stony
      Brook, New York

      Dr. Frederic K. Pfaender, Carolina Federation for Environmental Studies,
      University of North Carolina, Chapel Hill, North Carolina

      Dr. Robert Ringer, Institute for Environmental Toxicology, Michigan State
      University, East Lansing, MI

      Dr. William H. Smith, School of Forestry and Environmental Studies, Yale
      University, New Haven, Connecticut

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

SCIENCE ADVISORY BOARD STAFF

      Ms. Stephanie Sanzone, Designated Federal Officer, US EPA, Science
      Advisory Board (HOOF), 401 M Street, SW, Washington, DC  20460.

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

-------
                     U.S. Environmental Protection Agency

                                   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.

-------
                            DISTRIBUTION LIST

Administrator
Deputy Administrator
Assistant Administrators
Deputy Assistant Administrator for Research and Development
Deputy Assistant Administrator for Water
EPA Regional Administrators
EPA Laboratory Directors
EPA Regional Libraries
EPA Laboratory Libraries

-------
                          Questions to Focus SAB Discussion at
                                  SAB Commentary on
                               National Wildlife Criteria
                                     3/31/94 DRAFT
 Agenda Item: Characterization of Toxic Effects

 Population vs. Individual Level Assessments
 Wildlife criteria are intended to support risk assessments for populations of wildlife exposed
 to chemical pollutants.  These criteria are intended to address the direct effects of chemicau
 on wildlife populations.  They are not intended to  protect individual organisms (except in
 special cases where that is desired) or to provide protection for wildlife habitat and associate;"!
 biological communities.  Used in conjunction with other criteria and diagnostic approaches,
 wildlife criteria provide an essential component for carrying out risk assessments at the
 watershed scale.

1)     Given  the objective of protecting wildlife populations, do you think toxicological
       endpoints  such as mortality, growth, reproduction, and development,  are appropriate
       for deriving criteria because they can be related, at least qualitatively, to density-
       independent factors which influence populations?

2)     Should other endpoints such as biochemical or histopathological changes be used :o
       support criteria development as well?  Should use of these types of endpoints be
       restricted to cases where they can be causally-linked to mortality rates, fecundity
       rates, etc.?

3)     Currently  toxicological effects are extrapolated from NOAELs because of lack  of
       certainty concerning the amount of increased mortality or decreased fecundity that cr.n
       be absorbed in different populations and because dosing  regimes in most bioassays
       result in very steep (i.e., "all or nothing") dose-response curves. Are you aware of
       any credible population dynamic models that can be  used to guide  selection of effect
       thresholds from toxicological studies?
Effects Analyses
Given the initial goal of deriving water-based wildlife criteria, a paradigm adapted from the
human health non-cancer approach was generally endorsed in two separate national
workshops and proposed in the Great Lakes Water Quality Initiative (GLWQI). Although
based on the human health paradigm, there are significant differences between the human
health and wildlife approach.  Because the wildlife approach is designed to prptect
populations and not individuals, the wildlife paradigm does not include  an intraspecies
uncertainty factor (although exceptions can be made in special cases where protection of
individuals is desired) and wildlife assessments are typically restricted to a smaller set of
more gross endpoints.  This last point is illustrated in the GLWQI assessment for DDT
where the NOAELs used to calculate the mammalian wildlife value and the human health

-------
 Questions for SAB Commentary on Wildlife Criteria
criteria were from the same rodent bioassay.  However, in the wildlife assessment the
endpoint was the number of viable young per female while in the human health assessment
the more sensitive endpoint of impacts on liver pathology was used.

4)     Because of the limited number of toxicity tests,  what  is a reasonable minimum
       database? Should acute and/or chronic data be required?

5)     Efforts  are in progress  to tabulate and analyze existing toxicity databases to further
       evaluate interspecies differences and the type and nature of chronic endpoints tha. ^
       available for criteria derivation.  These analyses are designed to support the
       development of guidance on the selection  of interspecies, subchronic to chronic,
       LOAEL to NOAEL, and intraspecies uncertainty factors. Do you have any
       suggestions regarding this effort? Are you aware of additional data sources or
       analyses that could be used?
Allometric Scaling
To date, decisions made concerning the selection of interspecies uncertainty factors have
been based on professional judgement concerning underlying toxicokinetic (i.e., absorption,
distribution, metabolism and excretion rates of chemicals) and toxicodynamic (i.e.,
modes/mechanisms of toxic action and associated physiological and biochemical processes)
principles and available empirical data.  Allometric scaling (e.g.; based on surface area,  m2/3;
or,physiological time, m3M) can be used to describe general toxicokinetic trends and
contribute to "weight-of-evidence" deliberations.

6)     Should allometric relationships provide the basis for the selection of an  interspecies
       uncertainty factor or should  toxicodynamic issues also be incorporated into
       determining an interspecies uncertainty factor?  If so,  how?

7)     Are you aware of any other mechanistically-based toxicokinetic and toxicodynamic
       models that are currently available that can be employed to better predict interspecies
       variability?

-------
 Questions for SAB Commentary on Wildlife Criteria
 Agenda Item:  Characterization of Wildlife Exposure

 Representative Species in the GLWQI Approach
 One effort being conducted is to identify the most vulnerable (i.e., exposed and/or sensitive)
 species to contaminants in the Great Lakes to compare with the initial five species (i.e., bald
 eagle,  osprey, kingfisher, mink, and otter) used to propose the wildlife criteria.  The
 approach is to identify those wildlife species for which adverse impacts due to toxic
 contaminants have been documented in the Great Lakes region at one time or another over
 the past three decades.  In  addition  to the bald eagle, osprey, mink, and otter, which have
 already been included in the analysis, species for which impacts  have been documented
 include, but may not be limited to,  the Caspian tern, Forster's tern, common tern, ring-billed
 gull, herring gull,  double-crested cormorant, and black-crowned  night heron.  We will use
 dietary information and allometric equations for food ingestion rates to determine if any of
 these other species are likely to be more exposed to contaminants in aquatic food chains than
 those already included in the criteria development.

 S)     Is this approach adequate for defending our selection of wildlife species to examine
        for proposing criteria for the Great Lakes region?

 9)     Would this approach be appropriate for other regions of the country?

10)     For areas for which adverse  impacts have not yet occurred,  what might be the most
        appropriate alternative approach?
 Exposure Assessments for National Wildlife Criteria
 Efforts are underway to estimate the trophic level of wildlife species that may be at risk in
 the Great Lakes region based on dietary studies of selected wildlife species in the Great
 Lakes region and elsewhere.  Information on both the species and size of aquatic prey taken
 in specific locations is being compiled.  To estimate the trophic level of the prey species,
 data on the feeding habits of the aquatic prey species (and the prey species' prey or forage,
 etc.) are~being compiled from the literature.  These data will be used to estimate likely
 trophic levels for the prey species and then for the wildlife that feed on them and to illustrate
 how wildlife trophic  level is likely to vary by location, season,  and type of aquatic
 ecosystem.
          1
 As part of this effort, trends in  food web structure of aquatic ecosystems also are being
 investigated.  Food web structure and potential food chain length depend on many attributes
 of the aquatic ecosystem and also on the history of introductions of new species to the
 system.  For example, lakes in  which Mysis have been introduced to "improve" the forage
 base for fish often have a  food chain that is in effect one step longer than food chains in
 similar lakes in which Mysis has not been introduced.   Also, glaciated lakes tend  to support
 longer food chains than non-glaciated lakes in North America.

-------
 Questions for SAB Commentary on Wildlife Criteria
 These analyses will be used for three purposes:
        a)     To identify the potential range and central tendencies of trophic levels for ea-.'1.
               selected species (e.g., bald eagle) in the Great Lakes Region and  across the
               US, as well as sources of uncertainty in these estimates (which include both
               the wildlife feeding studies and knowledge of the aquatic food web).
        b)     To identify habitat attributes that are likely to influence trophic level for
               wildlife species, which may help define appropriate geographical  regions for
               establishing wildlife criteria.
        c)     To iu^ntify key data gaps that contribu^ to the uncertainty in estimates of
               wildlife trophic levels on a site- or region-specific basis to help identify which
               studies (e.g.,  tissue residue monitoring for trout) might help most in improving
               estimates of water quality criteria to protect wildlife.

11)     Given the highly variable nature of aquatic food chains, is analysis of site-specific
        wildlife diets and aquatic food webs suitable for the three purposes outlined above?

12)     What other approaches should be examined to identify trophic levels for  the model
        used to estimate water quality criteria to protect piscivorous wildlife? On a regional
        basis?  On a more local basis?
 Agenda Item:  Adaptation of 1985 Aquatic Life Criteria Approach

13)    We have concluded from our compilation and analysis of laboratory toxiciry data that
       variability is so large as to made a purely statistical treatment highly uncertain and
       probably overconservative.  Do you agree with this conclusion?

-------
Questions for SAB Commentary on Wildlife Criteria
Agenda Item:  U.S. Fish and Wildlife Service/National Biological Survey Approach   :
Dioxin-like Compounds

Use of Field Data
14)    The GLWQI allows for the use of field-derived data to take precedence over
       laboratory data to develop a water quality criterion for a chemical. When should
       field-derived data on species be used to develop a criterion?  What type of specific
       guidelines should be developed for the evaluation of field data for use in deriving
       wildlife criteria?

15)    Can the SAB suggest a mechanism whereby information from the field and specific
       sites can be  incorporated into the development of a wildlife criterion?
Egg-injection Studies
Egg-injection studies are useful in assessing potential reproductive/development effects, in
diagnosing responses in the field, and in deriving NOAELs for avian effects assessments.  In
addition, many wildlife species can not be maintained in the laboratory making it impossib-2
to conduct reproductive studies yet direct injection of the chemical into the egg provides
important and relevant toxicity information. However, egg-injection studies  do not
incorporate potential effects of metabolism  or effects on the female/male reproductive
physiology or behavior.  Therefore these studies could underestimate potential reproductive
hazards for some modes of toxic action.

16)    Are egg-injection studies appropriate for deriving NOAELs for avian effects
       assessments?

17)    Should the use of egg-injection studies be limited to chemicals where evidence
       indicates the chemicals direct effects are on the  developing embryo?

18)    Results from egg-injection studies are being considered for use in deriving NOAELs
       m avian effects assessments.  Although effects data from these  studies can be useful
       from a number of perspectives,  do you agree that these data are currently problematic
       in deriving water or sediment-based wildlife criteria because of limited
       bioaccumulation and toxicokinetic data that relate chemical concentrations in bird eggs
       to concentrations in the female bird  and/or fish, sediments,  or water associated with
       the  habitat?
Mode of Action Considerations
The best example of a class of compounds which are known to have the same mode of tcxic
action are the dioxin-like compounds. This class includes PCBs, PCDDs, and PCDFs. l"he
concept of assessing these using toxic equivalency factors and an additive model of toxicity
has been the topic of much discussion. The conclusions have generally been that this is a

-------
Questions for SAB Commentary on Wildlife Criteria	Pa--- -

good concept, that there are examples where it has worked very well, but there are problems
related to a universally applicable set of toxic equivalence factors.  These is, however, a
fairly large data set on avian related toxicity equivalency factors—particularly related to
embryo mortality. In  most instances,  we know that a multiplicity of compounds are acting in
concert and trying to assess them as individuals does not make sense.

       Is the available data adequate to assess classes  of compounds  that act through the
       same toxic mode of action?

-------
 Questions for SAB Commentary on Wildlife Criteria
 Agenda Item: General Discussion of Three Approaches

 General Questions
20)    Because of the paucity of data for amphibians and reptiles, wildlife criteria efforts
       thus far have focused on avian and mammalian wildlife species.  Given current
       information concerning exposures and effects of bioaccumulative compounds on some
       species of mammalian and avian wildlife, is it appropriate to proceed with credible
       criteria for these taxonomic  classes or should criteria/guidance be withheld until
       empirical databases and predictive techniques can be created  for amphibians and
       reptiles?

21)    Does the SAB have any suggestions on how best to incorporate effects on reptiles and
       amphibians?

22)    Are approaches other than these we have presented available to relate toxicity and
       exposure data to derive criteria? If so, given the constraints on the availability of
       wildlife toxicity data, is it feasible to apply any alternative approaches to deriving
       wildlife criteria?

23)    What is your assessment  of the strengths and weaknesses  of the three alternative
       approaches presented?
Benchmark Dose Approach
Human health non-cancer assessments are considering using the benchmark dose approach
for developing RiDs for developmental endpoints in the near future.  We recognize that the
benchmark dose approach may provide certain advantages to the current hypothesis-testing
approach (NOAEL), specifically, a LOAEL to a NOAEL uncertainty factor is no longer
applicable and intraspecies variability is incorporated in the dose-response curve.

24)    Should the benchmark dose approach be pursued for future wildlife criteria derivation
      'efforts?

25)    Should EPA continue with the hypothesis testing methodology or wait until the
       benchmark dose method is further developed?
Short-term Research Needs
In establishing a wildlife criteria approach, we have established efforts to address several
near-term research needs. This includes our efforts at quantifying wildlife exposures from
the aquatic food chain and compiling and analyzing wildlife toxicity data.  These efforts are
essential for future criteria derivation, uncertainty factor analyses, and identification and
verification  of critical data gaps for species-types, endpoints and modes of toxic action.

-------
, Questions for SAB Commentary on_Wild'ife Criteria	2SS.!,?.

 26)    Are there additional short-term research efforts which we should consider?

 27)    Do you have any suggestions for modifications to our research priorities?

 28)    Do you think there is sufficient scientific consensus concerning actual or potential
        adverse effects of chemical stressors on wildlife to establish a formal wildlife criteria
        approach for inclusion in ecological risk assessments.  Why or why not?
                                                                    •
29)     Based on current information, it is scientifically credible  to assume that aquatic life
        and human health criteria for persistent bioaccumulative chemicals are necessarily
        protective for wildlife?  Why or why not?
 Currently, development of criteria is not based on all possible mammalian and avian
 populations, but rather on groups of avian and mammalian species which are likely to be
 most at risk based on species' feeding habits (exposure analyses) and the properties of the
 chemical, (e.g.; for highly bioaccumulative compounds which are not metabolized in
 aquatic ecosystems, the most exposed species would be piscivorous species which feed at
 higher trophic levels).  If initial criteria development is based on species with the greatest
 exposures, subsequent analyses can be undertaken to determine if less-exposed species would
 be at risk based on greater sensitivity.

30)     Should criteria be derived  in this way, to protect populations identified as most at
        risk,  or,  should they be derived to protect a percentage of all possible species,  such as
        is done with aquatic life criteria?  (It should be noted that a percentage approach  can
        be applied in  the GLWQI approach by using a distribution of exposures for wildlife
        species.  The toxicity database appeals  too limited to enable developing distributions
        of wildlife impacts considerate of both toxic effects  and exposure assessments using
        the GLWQI approach.

31)     Should approaches for deriving wildlife criteria be premised on the bioaccumulative
        potential and  mode of action of the chemical of concern?

32)     Should national criteria be developed based on exposure analyses of wildlife
        populations in the entire U.S., or should analyses be derived for smaller geographical
        settings  in which the mammalian and avian populations most at risk can be identified
        (e.g., ecoregions, classes  of similar  watersheds, individual watersheds)?
 Geographical Considerations
 An alternative to providing national wildlife criteria is to provide national guidance to
 support criteria development in the context of specific ecoregions or smaller geographical
 sites (e.g., watersheds).  This strategy would require a national partnership to insure that
 consistent approaches are used in developing criteria for similar watersheds, with similar

-------