-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE -
  1       the landscape scale (e.g., watershed,
  2       region), involve multiple environmental
  3       issues, may not have overarching enabling
  4   '    legislation or regulations, and may have a
  5       principal goal of developing a conceptual
  6       model that prioritizes risks.  There may or
  7       may not be direct evidence of damage to  a
  8       specific ecological resource, a specific
  9       stressor of concern, or a regulatory
10       violation. Rather, the value itself is
11 -     considered sufficiently important to require
12       better management because of increasing
13       human demands.  As a result, value-
14       initiated assessments tend to be highly
15       complex, involving multiple assessment
16   ".  endpoints and multiple chemical, physical,
17       and biological stressors'. The geographic
18       area of concern may be administered and
19       managed under local, state, and federal
20       laws  and regulations and be subject to
21       multiple jurisdictions. These types of
22       assessments present special challenges to
23       the risk assessor.                                                        .         '
24          The primary difference in the flow of value-initiated assessments is the early selection of assessment
25       endpoints. Those selected should best represent the ecological values expressed in the management goals,
,26       while meeting other key characteristics of effective assessment endpoints (see section 3.4 for further
27       discussion).  After the ecological values  are translated into assessment endpoints, the risk assessor can then
28       consider which stressors and potential routes of exposure are likely to present a risk to the selected endpoints.
29        Problem formulation then proceeds using the same elements as those used in stressor-initiated and effects-
30       initiated assessments. For example, once assessment endpoints are selected for a particular ecological
31       resource, stressor characterization would be completed for identifiable sources of stressors within a
 Text Box 3-3.  Example of Effects-Initiated
 Assessment: Special Review of the Granular
 Formulations of Carbofuran Based on Adverse Effects
 on Birds
 An effects-driven assessment is illustrated with a case:
. study prepared by the Office of Pesticide Programs (OPP),
 The following brief descriptions are derived from
 Houseknecht, 1993 (Appendix A, Case A-2),
 *      Management goal: Predetermined under the
        Federal Insecticide, Fungicide, and Rodenticide
        Act, which authorizes the cancellation of
        registration for a pesticide that poses an
        unreasonable risk to humans or the environment.
 •      Observed ecological effect: OPP initiated a
        special review of the granular formulations of the
        pesticide carbofuran because of multiple suspected
        carbofuran-related bird kills.
 •      Exposure characterization: Based oh the
        observed effects, exposure was found to be
        associated with ingestiori of carbofuran granules or
        contaminated prey by birds foraging for food at
        sites where the pesticide was applied, for  ,
        agricultural: reasons.
 •      Assessment endpoint: The ecological resource
        affected by the pesticide: survival of birds
 •      Stressor characterization:  Carbofuran is an
        acefylcholinesterase inhibitor. A great deal of
        information was available due to prior use of the
        pesticide.
                                                         29
                                                10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1       geographic area that are likely to have an impact on the assessment endpoint.  A combination of stressors
  2       such as habitat alteration from increasing urbanization, hazardous waste site contamination, and sewage
  3       treatment plant discharges might be characterized relative to their potential effects on the reproduction of
  4       commercially valuable fish.  Ecological effects would also be characterized by evaluating observed changes in
  5       the assessment endpoints and in ecological resources upon which the assessment endpoints may depend (e.g.,
  6       declines in the important fishery and a decline in water quality).
  7           In value-initiated assessments, stressors and ecological effects are evaluated in terms of their relationship
  8       to the selected assessment endpoints. Thus, it is important to select assessment endpoints early in value-
  9       initiated risk assessments. Without them, the risk assessor can easily become overwhelmed by the
 10       complexity of the problem. Once assessment endpoints are selected, the elements of the risk assessment
 11       process are the same as for stressor-initiated and effects-initiated assessments.
 12           Management goals, while essential  to the development of assessment endpoints in all risk assessments,
 13       take on special significance in value-initiated assessments.  The diversity of the issues and the breadth of the
 14       assessment often requires the public to become involved in management goal development.  Management
 15       goals for these assessments may be derived by consensus, by combining federal, state and local regulations,
 16       or other complex processes.  Management goals and risk assessments such as these are likely to be required
 17       for place-based or community-based environmental protection.
 18           Two value-initiated assessment examples are provided in text box 3-4. The first is a published case
 19       study that contains one primary stressor and one principal ecological effect. The second is based on an
                                                                                                          /
20       ongoing watershed risk assessment that focuses on several stressors and assessment endpoints.  Discussion
21       about broadening the problem formulation to incorporate multiple stressors and endpoints is provided in the
22       assessment endpoint and conceptual model development sections.
23           Once assessment endpoints are established, the elements of the risk assessment process for value-
24       initiated assessments are the same as for stressor-initiated and effects-initiated assessments.
25                             '        ^       -           '   '     _    ^  /""'\
26       3.3. ASSESSMENT OF AVAILABLE INFORMATION
27           The initial step in problem formulation is evaluating available information on the sources of stressors and
28       stressor characteristics, the ecosystem(s) potentially at risk,  and ecological effects (see figure 3-1).  The order
29       of consideration and emphasis placed in each of these areas  will depend on the type of assessment (section
30       3.2) and the extent of the available data. Detailed discussion of data evaluation is deferred to the analysis
                                                         30
10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
        Text Box 3-4. Ecological Value Initiated Assessments                       '  *                    •

        Modeling Future Losses of Bottomland Forest Wetlands and Changes in Wildlife Habitat Within a
        Louisiana Basin: An ecological value-initiated assessment is illustrated below with a case study derived from
        Brody et al, 1993 (see Appendix A).

        •       Management goal: Not directed by statute.  The case was based in part on the National Environmental
                 Policy Act. It can also:be based on the Clean Water Act's objectives of restoring and maintaining the
                 physieali chemical, and biological integrity of the Nation's waters including wetlands.            .
        *       Assessment endpoihts: Forest community structure and habitat Rvalue to wildlife species and species:
                 composition of the wildlife community
        •       Ecological;; effect: Potential flooding of cypress; and tupelo wetland forests that could alter forest
                 community species and dynamics over timei                  -       •
        •       Stressor characterization;  Focused on the primary concern:. the rate and magnitude of water level
                 changes over time (i.e., hydroperiod);
        ,» .      Exposure characterization: Exposure was related'to change in hydroperiod from land subsidence
                 related: to the construction of artificial levees and other factors.

        Watershed Level Ecological Risk Assessment: Waquoit Bay: An assessment under development by EPA's
        Office of Water and Risk Assessment Forum.                   .    •                           •

         *       Management goal: Directed by a consortium of concerned citizens in a variety of local and regional
                 organizations and supported by the State of Massachusetts and federal government under the Clean
                 Water Act and NOAA research program. The goal is to restore and maintain water quality and habitat
                 conditions to support self-sustaining aquatic life and reverse degradation (see text box 2-4).
         •       Assessment Endpoint: Multiple assessment endpoints were generated. Two examples are: area!
                 extent of eel grass beds as primary habitat for shellfish and fish in the bay, and trophic status of
                 freshwater ponds and rivers.                                                      •
         •       Ecological effect:  Multiple ecological effects with primary emphasis on loss of eel grass and scallops,  ,
              .   and increased macroalgal growth, in the bay and in similar Cape Cod bays, eutrophication in freshwater
                 ponds and rivers, and potential reproductive and growth effects of groundwater contaminants.
         •       Stressor characterization: Multiple stressors characterized:including land development, nutrient
                 loading from groundwater and air, potential pesticide contamination from cranberry bogs and potential
                 toxicity of contaminated groundwater plumes from a Superfund site.                     .
         •  •'     Exposure characterization: Exposure to nutrients from septic system loadings from "build-out" of
                 residential areas, fate and transport models of nutrients and fate and transport of contaminated Superfund
                 plumes.   .             .                          .            •  .  ..  '
1

2

3

4

5
phase (section 4).  The brief summary provided in this section follows the Framework Report and draws
from the Conceptual Model Development issue paper (Barnthouse and Brown, 1994).


3.3.1.  Source and Stressor Characteristics
    The source of a stressor can be natural (e.g., tar pits) or anthropogenic (e.g^, oil spills), geographically

well defined (e.g., point source discharge) or geographically diffuse (e.g., air pollutants from automobile
                                                           31
                                                                                                10/13/95

-------
                                  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1      emissions; nonpoint source sedimentation). For many stressors there may be multiple sources, perhaps
  2      multiple point sources, a combination of point and nonpoint sources, or multiple nonpoint contributions., A
  3      well designed risk assessment will allow risk managers and risk assessors to distinguish among diverse
  4      sources and stressors.                                                           '                    '
                                                        ::,    ' •:  „,','',  I*"1,'i,»"   'ill1,,!! v :•„ I',!,..1 , "', '!'  • ;, i1., ' ' :,,•"' , , .»r, |!i,'  ,  |, . .   »,,',, Jill  " H,;*' ;•' " ."i|||i||i; ,,,1,1 ,'•
  5     "     Preliminary assessment of source and stressor characteristics may be based on actual, inferred, or
  6      estimated data. These data are used to determine which sources and stxgs.s.prg are of potential concern, identify
  7      ecological components that may be at risk, identify risk hypotheses and develop the conceptual model.
  8      Identifying the source of a stressor provides the basis' for determining the scale, duration and frequency of
  9      stressor occurrence and its potential for co-occurrence with particular ecosystems and assessment endpoints.
 10          Stressor- and source-initiated assessments are based on a known source or stressor so the problem is
 11      formulated to predict ecological effects from exposure to those stressors. In effects-driven assessments, the
 12      stressor or its source may be unknown. The magnitude, scale and type of ecological effects as well as known
 13      human activities in the location of observed effects are used to identity the stressor and its source. For value-
 14      initiated risk assessments, the assessment endpoints provide the basis for selecting the most important
 15      sources likely to serve as stressors to the assessment endpoints. Information on human activities, any known
                      .  :                ;           ' ••..'.  •••••:•/;•,•./•  ',;'':"(;. • I}!1!;'?1!:!,1';1;",,;/«! / - '•}•'•••'"' $>>'!"", "I""! I1' •' •""'v * >'•'•*£,''.WS31*	•'!,	 • "Klin i1 "1,11L; "I-"I, ""•	"'  ' • I'll'"1",1!! '• • , ' ' i,,1 ,,.,i'  «i  ."."i .''n' t !|''	T1'1 ii iKLi'H ""I'
 17      ecological processes upon which the assessment endpoints depend, are all considered when determining
 18      \vhichstressorsandsourcesneedtobecharacterized.
 19          The characteristics of a stressor influence how an assessment endpoint will become exposed and respond
20      to the stressor.  Evaluation of these characteristics is key to developing the conceptual model.  Key
21      characteristics are summarized in text box 3-5.
22                     ,                          :     '  .   '',   '  	 ;,••	'	 . 	• •   ,    •'    	• .•;..;	'
23      3.3.2.  Considerations of the Ecosystem Potentially at Risk                        .
                                                                       /            ,                         i
24          Once  exposure to a stressor occurs, the type and severity of affects depends on how the stressor acts upon
25      the ecosystem, what other stressors may also be acting upon the system, and the characteristics of the
26      ecosystem itself, as discussed above.
27          To assemble available information on an ecosystem potentially at risk, the first issue is defining its              I
                     1 ,,'!'      ( „    ,     '     '•„',"      „   	, i   . „  .,,„•; ,,,"| ,. i« • ||||||||L,,||,| "-pi:, li!.,. •  ,.n	 • in »	II '	,'ii'  t  	 '  «.	;	'• 	 |
28      geographic boundaries. These boundaries can vary widely (e.g., a single pond, a  watershed and a region can
29      all represent ecosystem boundaries).  Management goals may establish where the ecosystem is located and its
                      11 n1"!:1 .      ,       , "'. . '    '   "ij1 ",   '"  ,'     '      I ' j II  III  III i ^ I                "    I              » f
30      general size and complexity. Ecological factors determine how to translate these goals into ecologically
31       relevant boundaries.
                                                          32
10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1          This issue is particularly relevant in value-initiated assessments done on a landscape scale.  The risk
 2      assessor should consider biotic and abiotic factors                              ,
. 3      that serve as "forcing functions" for the system to
 4      help establish the spatial and temporal boundaries
 5.      for the assessment (Leibowitz, etal, 1992).
 6      Wetlands, for example, depend on hydrologic
 7      function over a wider area than the wetland itself
 8      (e.'g., the Florida Everglades depend on the
 9      drainage area of South Florida; the spatial scale of
10      a risk .assessment for the Everglades would have to
11     , consider the hydrologic contributions of this larger
12      area). Precipitation patterns may directly influence
13 .     what temporal scale  is relevant. In large systems
14      like the Everglades, or watersheds like the Waquoit
15      Bay estuary (Appendix A, Case A-4), a landscape
16      perspective is important because managing small
17      portions of these ecosystems would make it
18-    difficult to achieve management goals for the larger
19      system. Understanding the function of important
20      components (e.g., a habitat mosaic of wet
21      meadows, sandbars and aquatic backwater habitat)
22      is a prerequisite for-successful management.
23          Once the boundaries are defined, the
24      characteristics of the ecosystem can be described,
25   "   including diversity of habitats, physical structure,
26      biological characteristics, hydrology and other key
27      factors. These characteristics can directly influence
28      how different stressors are likely to alter basic
29      ecological functions  and impact resident organisms.
30      They are critical to understanding how the
Text Box 3-5. Key Stressor Characteristics
•       Intensity; Examples include concentration
        or dose (chemical stressors), magnitude or
        extent: (physical stressors), or
        density/population size (biological
        stressors)
*       Frequencies: A stressor event can be
        isolated, episodic or continuous. Events
        can be described by their periodicity (e.g.,
       : daily, lunar, seasonal, annual) or the
        absence of such influences (i.e., stochastic
        or chaotic),
•       Duration.  Stressor characteristics can
        influence how long the stressor persists in
        the environment. Stressor duration is
        directly relevant to issues  of ecological
        recovery and bioaccumulation.
•       Timing. Stressors are frequently episodic,
        occurring in greater and lesser degree at
        different times of day, different seasons or
        annual cycles. The timing of exposure to a
        stressor can be critical relative to organism
        life cycle or ecosystem events (e.g.,
        reproduction, lake overturn).
•       Scales. The spatial extent or influence of
        a stressor can range from local to global
        and:from habitat specific to  all habitats .
        within an exposed ecosystem.
•       Mode of Action.  Information on how a
        stressor acts upon organisms or ecosystem
        functions provides valuable  insights on the
        kinds of ecological components likely to be
        impacted by the stressor.
•       Type. Chemical, physical and biological
•-•      Distribution: How the stressor moves
        through the environment:  chemical fate
        and transport, physical transport, and life
        history dispersal characteristics.
                                                         33
                                         10/13/95

-------
                                  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1       ecosystem may be susceptible to an array of potential stressors. Text: box 3-6 provides questions to consider
  2       when assembling available information on the ecosystem potentially at risk.
  3                '     '   _   '      '_/     '  .'   _  ' -  '  ^  ' "   .   '";"';
  4       3.3.3.  Ecological Effects Considerations            ,
  5           Stressors can cause a wide range of ecological
  6       effects once exposure occurs in the environment:
  7       The effects could be general, impacting a diverse
  8       array of organisms and ecological processes, or •
  9       specific to one identifiable organism.  The extent
10       of effects could be broad, covering the continent, or
11       local, depending  on exposure as well as mode of
12       action. The type and severity of effect the stressor
13       has once exposure occurs depends on how the
14       stressor acts upon the components of the
15       ecosystem, what  other stressors may also be acting
16       upon the system, and the characteristics of the
17       ecosystem itself as discussed above. The effect of
18       one stressor may be altered significantly by the
19       presence or absence of other stressors. These are
20       issues that require careful evaluation when
21       assembling available information on observed
22       ecological effects. The information will help risk
23       assessors identify those stressors most likely to be '
24       responsible for observed affects.           .   •
25           The type of risk assessment influences how
26       information is assembled and used. In stressor-
27       initiated assessments, key stressor characteristics
28       (see text box 3-5) help identify which ecological
29       components in the target ecosystem are likely to be
30       susceptible to the stressor and show an effect. Information on their susceptibility to the particular stressor or
31       similar stressors under similar exposures may significantly aid the risk assessor in predicting potential
  Text Box 3-6. Questions Concerning
  Ecosystems Potentially at Risk and Ecological
  Effects
  Ecosystems Potentially at Risk
  What abiotic factors are most important in
  structuring the ecosystem (e.g., climatic factors,
  geology, hydrology, soil, water quality)?
  What habitat types are present?
  Where and how are functional characteristics
  driving the ecosystem, (e.g., energy source and
  processing, nutrient cycling)?
  What are the structural characteristics of the
  ecosystem (e.g., species  number and abundance,
  trophic relationships)?
  Ecological Effects
  What is the nature and extent of available
  ecological effects information (e.g., field tests,
  laboratory tests, or structure-activity
  relationships)?
  Given the nature of the stressor (if known), which
  organisms , habitats or processes is the stressor
  most likely to affect and how specific are those
  effects?
  What are the resource needs of organisms
  potentially influenced by the stressor? How are
  those resources affected by the stressor?
34

-------
         .          -  .       .   -  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE                    .    :
  1      effects, or identifying effects that may be occurring already if exposure has occurred. These ecological
 2      components may become good candidates for assessment endpdints. Effects-initiated assessments are based
 3      on observed ecological effects. These effects would require systematic evaluation as described in section
 4      3.2.2.  In value-initiated assessments, assembling available information can be a difficult task, but the
 5      principles used for stressor,characterization and ecological effects are combined to attempt to distinguish
 6      among the array of effects and determine which are likely to be related to human activities. Text box 3-6
 7      provides some questions to consider when assembling available information on ecological effects.
 8'     -       '                   '          " '•      .                            .
 9      3.4. SELECTING ASSESSMENT ENDPOINTS
10          Assessment endpoints are "explicit expressions of the actual  environmental value that is to be protected"
11    •  (U.S. EPA, 1992a). Assessment endpoints are critical to problem formulation because they define the focus
12      "of conceptual model development. Their relevance to assessment of risk is determined by how well available ,
13      information was used to select ecologically appropriate endpoints.  Their ability to provide the basis for a risk
14      assessment is determined by whether they are measurable characteristics of the ecosystem that adequately
15      represent the management goals. This section describes criteria for selecting and defining assessment
16      endpoints.-                            ,               .
17
18      3.4.1.  Selecting What to Protect
19          The ecological resources selected to represent the management goals for ecological resource protection
20      become the assessment endpoints that.drive ecological risk assessments. Suter( 1993a) defined effective
21      assessment endpoints as those that identify "... the valued attributes of the environment that are considered
22    •  to be at risk arid defining these attributes in operational terms." Much confusion about assessment endpoints
23      has come from different interpretations of what "environmental value" really means. Despite ongoing   '   .
24      discussions about the appropriate meaning of "value," it is clear that the focus of a risk assessment should be
25      on ecological resources that are valuable because they are protected by law, provide critical resources,
26      ecosystem function would be significantly impaired if the resource were altered, or human perceptions of
27      value would be lost.
28          The Framework Report identifies three criteria to consider when selecting assessment endpoints:
29      •.  policy goals and societal values,                                •
30      •   ecological relevance, and
31      •   susceptibility to the stressor.
                                                        35
10/13/95

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
                        DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                                           1      ',,'',' 	 :  „,,' Hit, I1"	!'.',! ,. •  ,.;,
Assessment endpoints that meet all three criteria'
provide the best foundation for an effective risk
assessment (e.g., see text box 3-7).
3.4.1.1. Policy Goals and Societal Values
    Ultimately, the value of a risk assessment
depends on whether it is used to make quality
management decisions.  Risk managers are more
willing to use a risk assessment for making
decisions when the assessment is based on values
and organisms that people care about. Thus
assessment endpoints that reflect policy goals and
societal values add to the potential use of the
assessment for decision-making.   .
    Management goals, as discussed in section 2,
                 r"
are based on policy goals and societal values for
ecological resources potentially at risk.  Assessment
endpoints are derived from management goals to
effectively translate the goals into a form that can
be directly or indirectly measured for a risk
assessment. Candidates for assessment endpoints
might include entities such as endangered species
and commercially or recreationally important species, or functional attributes that support food sources or
flood control (wetlands, for example), and aesthetic values, such as clear air in national parks or the existence
of charismatic species like eagles or whales.
    Many resources that may be potential endpoints because of their importance to an ecosystem are often
                                  ,:       '•  '" ' ' ' "              I        I      I                  I   III
not considered valuable because humans are indifferent to them or find them annoying.  Midges, for example,
•    ,    "   :••*•      '     ,   „,-,  . •   : •'•  • :•  • .'•>•.:;• - •:;•            |i         I      i         	I
may be considered pests but can represent the base of a complex food web that supports a popular sports
fishery. In this case, it would be better to choose the fishery as the basis for a risk assessment and select
midges as a critical ecological component to measure. In cases where the appropriate assessment endpoint is
Text Box 3-7. Salmon and Hydropower: Why
Salmon Would Contribute to a Good
Assessment Endpoint
A hydroelectric.dam is to be built on a river in-the
Pacific Northwest where anadromous fish such as
salmon spawn. To evaluate the risk of the dam,
assessment endpoints must be selected. Of the
anadromous fish, species of salmon that spawn in
the river would be appropriate choices because they
meet the criteria for good assessment endpoints.
Salmon fry and adults are important food sources
for a multitude of aquatic and terrestrial species
and are major predators on aquatic invertebrates
(ecological relevance). Salmon are sensitive to
changes in sedimentation and substrate pebble size,
require quality cold water habitats, and have
difficulty climbing fish ladders. Hydroelectric
dams represent significant and normally fatal
habitat alteration and physical obstacles to
successful salmon breeding and fry survival
(susceptibility).  Finally, salmon support a large
commercial fishery, some species are endangered,
and they have ceremonial importance and are key
food sources for Native Americans (societal
value).  Salmon reproduction and population
maintenance is a good assessment endpoint for the
risk assessment, and if salmon populations are
protected, other anadromous fish-populations are
likely to be protected as well.
                                                        36
                                                                                          10/13/95

-------
  1
  2
  3
  4
  5
  6
.  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30  .
31:
                         DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
unpopular with the public, the risk assessor may choose to select a more desirable organism or resource that
is directly dependent on the appropriate endpoint, or present a persuasive case in its favor.
    A complicating factor in the selection of assessment endpoints can be people's changing perceptions of
ecological relevance (Suter, 1993 a).  For example, wetlands were formerly regarded as wasted acreage that
could be reclaimed by draining and filling for development, or used as garbage dumps. Now there is much
greater awareness of the values wetlands provide such as wildlife habitat and flood mitigation. Assessment
endpoints for wetland risk assessments.conducted 30 years ago would differ markedly from those completed
today.
    Public meetings during the initial stages of problem formulation can be very useful in getting the public
involved,, elucidating local concerns, selecting effective assessment endpoints, and gaining support for the
risk assessment process.                            -                                      "

3.4.1.2.  Ecological Relevance
    Ecologically relevant endpoints "reflect important characteristics ,of the system and are functionally
related to other endpoints" (U.S. EPA 1992a). These are endpoints that help sustain the natural structure and
function of an ecosystem.  Ecological relevance becomes most important when risk assessors are identifying
the potential cascade ofadverse effects that could result from the loss or reduction'of one.or more species.
Species are considered ecologically relevant when they provide a significant food base, maintain community
structure, provide shelter for other species, promote regeneration of critical resources, or serve another
important function in an ecosystem.  Certain major categories of organisms (e.g., principal primary
producers, forage species, keystone predators) and ecosystem processes (e.g., primary production, nutrient
cycling) are generally considered ecologically relevant. Changes in these species or processes can result in
unpredictable and widespread effects. They  are appropriate entities to select as assessment endpoints in an
ecological risk assessment.
    Determining ecological relevance in specific cases requires expert judgment based on site-specific
information, preliminary site surveys or other available information. If assessment endpoints in a risk
assessment are not ecologically relevant, the results of the risk assessment will fail to predict risk and could
 •'                                                       -                         /
lead to misguided management and significant environmental risk.
                                                         37
                                                                                           10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      3.4.1.3.  Susceptibility to the Stressor
 2          Ecological resources are only considered susceptible to a human-induced stressor when they are sensitive
 3      to a stressor to which they are exposed.
 4          Sensitivity refers to the likelihood that one individual or species may be more or less affected by a
 5      particular stressor than another.  Measures of sensitivity include mortality or adverse reproductive effects
                    , , „    "           , •;        •    :("  .    • ',  ,  i	  i  nnij	I|M| J	 |. |	|    	I    I       i I    I J| IJ II ill II
 6      from exposure to toxics, as well as behavioral abnormalities, avoidance of significant food sources or nesting
 7      sites, or loss of offspring to predation because of the proximity of stressors  such as noise, habitat alteration or
 8      loss, community structural changes, or other factors. Toxicity testing is normally used to determine
 9      sensitivity to chemical stressors. Sensitivity to other kinds of stressors requires other types of information.
10          General life history characteristics are normally evaluated to determine  potential sensitivities. For
11      example, populations of species with long life cycles and low reproductive rates will be more vulnerable to
12      extinction from increases in mortality than species with short life cycles and high reproductive rates
13      (Barnthouse, i993). Species with large home ranges may be more sensitive to habitat fragmentation than
14      species with small home ranges where the entire home range is within a fragment.  Sensitivity is also related
15      to the life stage of an organism when exposed to a stressor.  Frequently the  young  of an animal  species are
16      more sensitive to stressors than adults. For example, Pacific salmon eggs and fry  are very sensitive to
17      sedimentation from forest logging practices and road building because they can be smothered.            •
18          Exposure is the other key determinant in susceptibility. To evaluate susceptibility, it is important to
19      evaluate the proximity of an ecological resource to  the stressor, the timing of exposure (both in terms of
20      frequency and duration), and the intensity of exposure along with the life stage of organisms during.
21          Exposure can mean co-occurrence, contact, or the absence of contact, depending On the stressor and
22      assessment endpoint. Exposure to a chemical stressor normally occurs when direct contact is made with the
23      organism. Issues of concentration, duration, and type and location of exposure are considered when
24      evaluating contact. Co-occurrence becomes exposure when the existence of a stressor results in an adverse
25      effect. For example, a highway built through a wetland may not only directly expose wetland species to
26      adverse habitat alteration at the site of the highway, it may drive off roosting birds because the  birds require
2*7      unobstructed views where they roost.  Direct contact with the highway is  not required; co-occurrence is
28      sufficient to cause significant adverse effects (e.g., loss of critical roosting habitat for whooping cranes, an
29      endangered species). The presence of degraded habitats can be translated into exposure to unsuitable feeding,
30      resting, or breeding habitat.  The spatial extent of these conditions is key to understanding the potential risk
31      of habitat changes to particular assessment endpoints.
                                                          38    ,                    '                 10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12  '
13
14
15
16
17
18
19  .
20
21
22
23
24
25
26
27
28
29
30
31
    If a species is unlikely to be exposed to the stressor of concern, it is inappropriate as an assessment
endpoint.  For example, deer and turkey may be appropriate for evaluating potential sensitivity of ungulates •
and birds to a chemical contaminant. However, it would be erroneous to conclude that since deer and turkey
did not live at the contaminated site, the chemical did not pose a risk. Deer and turkey may only serve as
good surrogates for determining the sensitivity of similar animals living in the area. Appropriate assessment
endpoints in this case would include birds and ungulates that live in the area and are likely to be exposed.
    The timing of exposure is often linked to sensitivity.  Adverse effects of a particular stressor may be
important during one part of an organism's life cycle, such as early development or reproduction.  Often, fish
toxicity tests are conducted during their developmental stages because adverse reactions tend to be higher
during these life stages. Sometimes sensitivity refers to the absence of exposure to a necessary resource
during a critical life stage.  For example, if fish are unable to find suitable nesting sites during their
reproductive phase, risk is significant even though water quality is high and food sources are abundant.  The
interplay between life stage and stressors can be very complex (e.g., see text box 3-8).
    Problem formulations based on assessment endpoints that are insensitive and unlikely to be exposed,to
the stressor will not be relevant to management concerns and can lead to erroneous decisions.
3.4.2.  Defining Assessment Endpoints
    Assessment endpoints interpret management
goals and public values into operationally defined
e'cological endpoints that can be measured directly
or through indirect measures. Assessment
endpoints can translate vague management goals
such as "ecosystem integrity" into relevant  .
endpoints for the system under evaluation.
    To operationally define an assessment  '
endpoint, two elements are required. The first is
the valued ecological entity. This can be a species
(e.g., eel  grass, piping plover), a functional group of
species (e.g., raptors) an ecosystem function or
characteristic (e.g., nutrient cycling) or specific
valued habitat (e.g., wet meadows). The second
Text Box 3-8. Sensitivity and Secondary
Effects:  The Mussel-Fish Connection
Native freshwater mussels are endangered in many
streams.  Management efforts have focused on
maintaining suitable habitat for mussels because
habitat loss has been considered the greatest threat
to this group. However, larval unibnid mussels
must attach _to the. gills of a fish host for one mpnth
during development.  Each species of mussel must
attach to a particular host species of fish. In
situations, where the fish community has been.
changed, perhaps due to stressors to which'mussels.
are insensitive, the host fish may no longer be  .
available. Mussels will die before reaching     :
maturity as a result. Regardless of how well.
managers restore mussel habitat, mussels will be
.lost from this system unless the fish community is
restored. In this case, the absence of exposure to a
critical resource is the source of risk.
                                                         39
                                                                                           10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1       necessary element is the characteristic about the entity of concern that is important to protect and potentially
  2       at risk. For example, it is necessary to define what is important for piping plovers (e.g., nesting and feeding
  3       success) eel grass (e.g., areal extent and patch size), and wetlands (e.g., endemic wet meadow community
  4       structure and function). For an assessment endpoint to provide a clear interpretation of the management
  5       goals and provide the basis for measurement in the risk assessment, both an entity and attribute are required.
  6       Assessment endpoints are not management goals. They do not contain words like "protect," "maintain," or
  7       "restore." Nor do assessment endpoints indicate a direction for change such as "loss" or "increase" or
  8       represent adverse responses like "mortality."  They are descriptions of the entity of value and the
  9       characteristics or attributes of the entity that are to be protected (see text box 3-9).
10          Defining assessment endpoints can be difficult. They may be too broad, vague, or narrow or
11       inappropriate for the ecosystem requiring protection.  "Ecological integrity" is a frequently cited, but'vague"'
                    , ,,;,    ' -       ' . ' -. ,  •   /'-:'.    ' "•:.,'  '.'            | I I           I           |       |' \  l|| II |
12       goal and an even more vague assessment endpoint.  "Integrity" can only be used effectively when its meaning
13       is explicitly characterized for a particular ecosystem, habitat or entity. This may be done by selecting key
14       entities and processes of an ecosystem and describing characteristics that best represent integrity for that
15       system.  For example, general integrity goals for Waquoit Bay were translated into several assessment
16       endpoints including "areal extent and patch size of eel grass beds" (see box 2-4).
17          Expert judgement and an understanding of the function of an ecosystem are important to translating
18       general goals into usable assessment endpoints.  Endpoints that are too narrowly defined, however, may not
19       support effective risk management. For example, if an assessment is focused on protecting the habitat of an
20       endangered species, the risk assessment may fail to include other critical variables (see text box 3-8).
21          Assessment endpoints can also be inappropriate for the ecosystem of concern. Selecting a game fish that
22       grows well in reservoirs to meet a "fishable" management goal would be inappropriate for evaluating risk
23       from a new hydroelectric dam if the ecosystem of concern is a stream in which salmon spawn (see text box 3-
24       7). Although the game fish will satisfy the fishable goal and may be highly desirable by local fishermen, a
25       reservoir species does not represent the ecosystem at risk. A vague "viable fish populations" assessment
26       endpoint could result in a completely inappropriate risk assessment.
27          Well-defined assessment endpoints reduce uncertainty in a risk assessment.  They provide clear direction
28       and boundaries for the risk assessment and minimize miscommunication. They also influence the analysis
29       and interpretation of data in the analysis phase. For example, clearly defining the boundaries of an
30       assessment to mean "genetic exchange in regional populations" of an organism rather than "genetic exchange
                                                  III	
40
                                                                                                   10/13/95

-------
                       DRAFT-DO NOT QUOTE, CITE, 'OR DISTRIBUTE
Text Box 3-9. Examples of Management Goals and Assessment Endpoints (See Appendix A)
  Case  -
Regulatory/Management Goal
Assessment
End point
 New Chemical.
General: Protect "the environment" from "an
unreasonable risk of injury" (TSCA §2[b][l] and
[2]);. protect the aquatic environment
Specific; Exceed a. concentration of concern by 20
days or less a year
Survival, growth, and
reproduction of fish,
aquatic invertebrates,
and algae
  Carbofurah: •"
Gerier'alf- prevent... "unreasonable;adverse effects on
the environment"1 (FIFRA §§3[cir5J:and 3[c][6j);
: using ebsfcfaeriefit cohsiderafionstno .regularly
repeated bird kills
Individual bird survival
 Bottomland
 Hardwood
General: National Environmental Policy Act may
apply to environmental impact of new levee
construction; also Clean: Water Act: §404::
(1) Forest community
structure and habitat
value to wildlife species
(2) Species composition
of wildlife community
  Chilean Log Importation
General: This assessment was done to help provide
a basis for any necessary regulation of the
importation of timber and timber products into the
United States.                •
Survival and growth of
tree species in the
western United States
 Baird and McGuire
 Superfund Site
 (terrestrial component)
General: Protection of the environment
(CERCLA/SARA)
(1) Survival of soil
invertebrates
(2) Survival and
reproduction of song
birds   '         ~
  Waquoit Bay Estuary
General: Clean Water Act - wetlands protection;
water quality criteria - pesticides; endangered
species. .National Estuarine Research Reserve,
Mass. Area of Critical Environmental Concern.  Re-
establish and maintain water quality and habitat
conditions to support diverse self-sustaining
commercial; recreational; and native fish, water
dependent wildlife, and shellfish,- and reverse
ongoing degradation
(1) Estuarine eel grass
habitat abundance and
distribution
(2) Estuarine fish
species diversity and
abundance
(3) Freshwater pond
benthic invertebrate
species diversity arid
abundance      .
                                                41
                                                                   10/13/95

-------
   Text Box 3-10.  Common Problems in Selecting
   Assessment Endpoints
           Endpoint is too vague (e.g., ecosystem
           integrity)
           Ecological resource is better as ,a
           measurement endpoint (e.g., midges
           example)
           Ecological resource is not exposed to the
           stressor (e.g., turkey and deer example)
           Ecological resources are irrelevant to the
           assessment (e.g., lake fish in salmon
           stream)
           The values of a species or attributes of an
           ecosystem are not fully considered (e.g.,
           mussel-fish connection, text box 3-8)
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1       in local populations" directly influences how
 2       heterogeneity within those boundaries will be
 3       described. Common problems encountered in
 4       selecting assessment endpoints are summarized in
 5       text box 3-10.
 6           The presence of multiple stressors should
 7       influence the selection of assessment endpoints.
 8       When it is possible to select one assessment
 9       endpoint that is sensitive to many of the identified
10       stressors, yet responds in different ways to different
11       stressors, it is possible to consider the combined
12       affect of multiple stressors while still discriminating
13       among effects.  For example, if recruitment of a fish           .
14       population is the assessment endpoint, it is important to recognize that, recruitment may be adversely affected
15       at several life stages, in different habitats, through different ways, by different stressors. The measures of
16       effect, exposure and ecosystem and receptor characteristics chosen  to evaluate recruitment provides a basis
17       for discriminating among the different stressors and their effects and evaluating their combined effect.
18       Although the data used to evaluate ecological effects may be diverse, the assessment endpoint can provide a
19       basis for comparison if carefully selected.  For example, the National Crop Loss Assessment Network  (Heck,
20       1993), selected crop yields as the assessment endpoint to evaluate the cumulative effects of multiple
21       stressors. Although the primary stressor was ozone, the crop yield  endpoint allowed them to consider the
22       effects of sulfur dioxide and soil moisture. Carefully defined assessment endpoints are essential for
23       addressing multiple stressors.  As noted by Suter (1993a)
24           "'The assessment of multiple stresses demands that well-defined endpoints be used that are applicable to
25           ways in which all of the stresses act on the target biota. It is not possible to combine toxic effects
                       '"         ,   ' •!;'   ' ' ; '"''I           ',".'.          I    II I      '                          I III  Illl
26           expressed as multiples of an MATC, fishing effects expressed  as tons harvested, and habitat degradation
                                     :,"   '  : ••   «,  ••  -   ! •...: • . !.•          i  i i       i                      11 i     in i
27           expressed as hectares of salt marsh filled. Instead, an endpoint such as recruit abundance (the abundance
28           of one-year-olds) must be used so that all effects can be expressed in the same units (Barnthouse et al., 1990)."
42
                                                                                                    10/13/95

-------
                                 DRAFT-DO NOT QUOTE; CITE, OR DISTRIBUTE
  1       For example, the National Crop Loss Assessment
  2       Network (Heck, 1993) selected crop yields as the
  3       assessment endpoint in assessment primarily
  4       concerned with the effects of ozone.  Use of the
  5       crop yield endpoint also facilitated evaluation of
  6       the effects of sulfur dioxide and soil moisture.
  7          Assessment endpoints are effective only when .
  8       they are accessible to prediction and measurement,
  9       Assessment endpoints must provide the basis for
 10       generating and evaluating hypotheses about the
 11       relationships among the assessment endpoints and
 12,     stressors  to which they are exposed. If the
 13   ,  •  response  of an assessment endpoint cannot be
 14      directly measured, or be predicted from measures
 15       of responses by surrogate or similar entities, it
 16      cannot be assessed.  In many applications, "the best
 17      assessment endpoints are those for which there are
. 18      well-developed test methods, field measurement
 19      techniques, and predictive models" (Suter, 1993a).
 20 ,     Measures that will -be used in the risk assessment
 21      are often identified during conceptual model  .
 22      development and specified in the analysis plan.
 23          Once assessment endpoints are  selected to best
 24      represent the management goals for the particular
 25      ecological value, the risk assessor should discuss
 26      the endpoints with the risk manager, providing the
 27"     rationale for their selection.  Problem formulation .
 28      should only proceed when both risk assessor and
 29      risk manager  agree that the assessment endpoints
 30      adequately reflect the management goals, and
Text Box 3-11. How Do Water Quality Criteria
Relate To Assessment Endpoints?
Water quality criteria'(U.S. EPA, 1986a) have been
developed for the protection of aquatic life from
chemical stressors. This text box shows how the
elements of a water quality criterion correspond to
management goals; assessment endpoints, and
measures.                  •           ,
Regulatory Context:
•-•"     Clean Water Act,  §101: Protection of the
        chemical, physical, and biological integrity
        of the nation's waters
Management Goal:
»•      Protect 99% of individuals in .95% of the
        species in aquatic communities from acute
        and chronic effects resulting from
        exposure to a chemical stressor
Assessment Endpoints:
*      Survival of fish, aquatic invertebrate, and
        algal species under acute exposure
•      Survival, growth, and reproduction of fish,
        aquatic invertebrate.; and algal species
        under chronic exposure
Measures'of Effect:
•      Laboratory LC50s for at least eight species
        meeting certain requirements
•      Chronic NOAELs for at least three species
        meeting certain requirements
Measure of Ecosystem; and Receptor
Characteristics:
»      Water hardness  (for some metals)
*      pH     -';••     '•-.••'.'.•'.•
The water quality criterion is a benchmark level
derived from a distributional analysis of single
species toxicity data. It is assumed that the species
tested adequately represent the composition and
sensitivities of species in a natural community.  .
                                                         43
                                        10/13/95

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
31
              	           DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
define the goals in such a way that they can be evaluated in the risk assessment with scientific rigor.
3.5.  CONCEPTUAL MODELS              	^	;	,	
    A conceptual model in problem formulation is a verbal description and visual representation of predicted
responses by ecological components to stressor's to which they are exposed, and includes ecosystem processes
that influence these relationships.  The conceptual model consists of a series of integrated risk hypotheses and
predictions about these relationships.
    Risk hypotheses are assumptions made in order to evaluate logical or empirical consequences. They are
formulated using initial integration and evaluation of information  on the ecosystem at risk, potential sources
of stressors, stressor characteristics, and observed or predicted ecological effects on selected or potential
assessment endpoints. These hypotheses may predict the effects of a stressor event (e.g., a chemical release)
on an ecological component before,it happens, or they may postulate why observed ecological effects
occurred, and whether these events are caused by the stressor(s) of concern. Depending on the. scope of the  '
risk assessment, conceptual models may be very simple, predicting the potential effect of one stressor on one
receptor, They may be extremely complex, as is typical in value-initiated risk assessments that often include
predictive and retrospective hypotheses about the relationship of multiple complexes of stressors on diverse
ecological receptors. The relatively broad scope of the conceptual model becomes focused in the analysis
             111 i   •          i    ,'      •   '•  •.  ;•••,:•.. ",   ••.	;.y,'.•,'»!	i;"S;K"",i :«!,	:,!""!'I;")1 :: riM "V M'' I":1 : •"!",;	.,'i.i 1,'' :,:,';:: :•  • ';  nv:1*:::1!"
plan when key risk hypotheses are selected as the subject of the risk assessment.  It is then that justifications
             "	        „             :/     „  , ;	'' "" i	•'	Mil" "' 'Ilil/iiNlL	Ml1,,,!"':!!" .", "„,,":' tr ;„!„:: '''„} '"	 ,»(>,:.	:  "	MilK'ii ..III1;1!	li'L  iiiijllh ,!!,, rv,
for selecting and not selecting hypotheses are documented
    Conceptual models include many relationships. Exposure scenarios may qualitatively link land-use
activities to sources and their stressors; describe primary, secondary, and tertiary exposure pathways; and
describe co-occurrence between exposure pathways, ecological effects, and ecological receptors. Selection of
critical relationships to include in the conceptual model and pursue in the risk  assessment is based on several
criteria, including:            "
                                                                                                  !	!'"!f!	'fliS'll	"Hv:'!:!!	
•   data availability,
•   strength of data establishing relationships between stressors and effects,
•   relative importance of endpoints,
•   relative importance or influence of stressor, and
•   importance of effects to ecosystem function.
    Conceptual models require three basic elements to establish relationships that lead to risk: .stressor,
              	i,    ••   '   .''.•'''.'   '•'.,•"' '.  ':.'.'••   	   fi 	i   •             i          J    i
exposure, and ecological entity. Depending on what initiated the assessment, different elements are known
                                                                                                     ft" Ml!*''1:*. -"!'i: 1
                                                           44
                                                                                              10/13/95
                                                                                              '" '/."IS,I' I1",:'"i|""	,:l.'!,,,i TIHil,,,!H!!
                                                                     .It/-»ift-i	SW:	i: ,.':l;"' i1 '••>	M, «	I'1.	i:.1,::':, 4,:A1:',»«?r	:i;l-:!W	i»,ilijt:Nl:!

-------
                                DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      and unknown.  Conceptual models help to show these relationships, and identify where hypotheses must be
 2      generated. In stressor- or source-initiated assessments, the stressor or source is known. The risk assessor
 3      must determine potential routes of exposure for the stressor, and .identify entities that may be exposed and
 4      would be appropriate assessment endpoints. In effects-initiated assessments, the entity and effect are known.
 5      The affected ecological entity often becomes the assessment endpoint. The risk assessor must determine
 6      what stressors  may be causing the effect and how the assessment endpoint became exposed. In value-'
 7      initiated assessments, assessment endpoints are normally selected based on management goals.  These goals
 8      are derived from recognized values of ecological resources, and often because of undesirable changes
 9      observed over  time in these resources. Assessment endpoints then become the focus for defining a variety of
10'     stressors, exposure pathways, and potential effects that are incorporated into the conceptual model.
11          The complexity of the conceptual model depends on the complexity of the problem, the number of
12      stressors, number of assessment endpoints, and the characteristics of the ecosystem. For single stressors and
13      single assessment endpoints, conceptual models can be relatively simple relationships.  For value-initiated
14      risk assessments, where conceptual models describe the pathways of individual stressors and assessment
15      endpoints and  the interaction of multiple and diverse stressors and assessment endpoints, several submodels
16      will normally be required to describe individual pathways.  Other models may then be used to hypothesize
17   *  how these individual pathways interact.
18          Conceptual models may account for one of the most important sources of uncertainty in a risk
19      assessment.  If important relationships are missed or misspec'ified, risks could be seriously under- or over-
20      estimated in the analysis phase.  Uncertainty can arise from lack of knowledge of how the ecosystem
21      functions, in identifying and interrelating temporal and spatial parameters, and in describing a stressor or
22      suite of stressors (Smith and Shugart, 1994). In some cases, little may be known about how a stressor moves
23      through the environment or causes adverse effects, In most cases, multiple stressors are the norm and a  _
24      source of confounding variables, particularly for conceptual models that focus on a single stressor.  Opinions
25      of experts on the appropriate conceptual model configuration may differ.  While simplification and lack of
26      knowledge may be unavoidable, Smith and Shugart (1994) discuss the need to document what is known,
27      justify the model, and rank model components in terms of uncertainty.
28          Uncertainty associated with conceptual models can be reduced by developing alternative conceptual
29      models for a particular assessment to explore relationships and generate additional hypotheses.  Part of the
30      purpose for, conceptual model development is to select the most important relationships to pursue in analysis.
31      As a result, treatment of unassessed risk hypotheses.becomes an important issue. Risk assessors use
                                                        45
10/13/95'

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
                         DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  ''       "    , I?  :          'I  ,      „    	     i ': I     'i    	',"• „'':,,:• M 'l>: :P'5'MII!|	'|:|»I'\  ' I/  "   , 	n:;" ,  'i'1                I
 professional judgement to select among hypotheses and provide narrative rationales to justify their inclusion
r or exclusion. In cases where more than one conceptual model is plausible, the risk assessor must decide
 whether it is feasible to follow separate models-through in the analysis phase or whether the models can be
 combined into a better conceptual model.  It is important to revisit, and if necessary revise, conceptual models
 during risk assessments to incorporate new information and re-check rationale.
    The principal products of conceptual model development include:
 •  A set of risk hypotheses that describe predicted relationships between stressor, exposure and assessment
    endpoint response, along with the rationale for their selection, and
 •  A flow diagram that graphically depicts the relationships presented in the risk hypotheses.

 3.5.1. Risk Hypotheses
    Risk hypotheses are assumptions about relationships among assessment endpoints and their predicted
 responses to stressors when exposed. While hypotheses should be developed  even when infprmation is
 incomplete, the amount and quality of data will affect the specificity and level of uncertainty associated with
 risk hypotheses and the conceptual models they form.  These hypotheses provide the basis for specific
 predictions about links among stressors, exposure, endpoints and responses.  The predictions can then be
 evaluated systematically either through new data collection, or by using available data, during the analysis
 phase. Hypotheses and predictions set a framework for using data to evaluate functional relationships (e.g.,
            . •   '       ,      -        •.''••           '      i
 stressor-response curves).
    The plausibility of specific risk hypotheses helps risk assessors sort through potentially large numbers of
 Strcssor-effect relationships, and the ecosystem processes that influence them, to identify those appropriate
 for the conceptual model and analysis phase.  As noted in the Framework Report, "only  those hypotheses that
 are considered most likely to contribute to risk are selected for further evaluation in the analysis phase." As
 discussed previously, it is important to provide the rationale for selecting, and not selecting, risk hypotheses
 and that data gaps and uncertainties are acknowledged. Examples of risk hypotheses are provided in text box
3-12.            '            "'"'   '..   '.     .'_  '"'

3.5.2.  Flow Diagrams
    Flow diagrams are visual representations of conceptual models. They may be based on theory and logic,
empirical data,  mathematical models or probability models. There is no ideal configuration for a flow
                                                         46
                                                                                      „	 10/13/95	

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE'
 1      diagram; it can take many forms. However flow  -
. 2 "     diagrams that show relationships clearly and simply
 3      provide the best learning and communication tools.
 4          Three flow diagrams are provided here as
,5      examples. As described in Barnthouse and Brown

 6      (1994):

 7          "Figure 3-2 is a.more conventional flow chart.

 8         ' based on the physical movement of a toxic

 9          contaminant from a source, through
10.       '  environmental media, to direct and indirect
11          effects on a fish population (i.e., food-chain

12          effects). This type of flow chart has the
13       .   advantage of corresponding more directly to

14-         the quantitative environmental fate models that

15        .  are often used in risk assessment It is only

16          applicable, however, for chemical stressors.
17          Also, it does not lead as directly to
18          consideration of alternatives."
19          "The flowchart shown in-Figure 3-3

20          depicts the influence of hydrology on the
21          structure and function of a bottomland
22          forest ecosystem (Brody, et al., 1993; see
23          'Appendix A, Case A-1),,drawn in'energy

24          circuit language'(Odum, 1971). In this
25          chart, the various symbols represent energy

26   -       sources and transformation processes,,
27          environmental influences, and internal
28          regulatory mechanisms. Energy circuit
29          notation is quite general, allowing
30          representation of chemical, physical, and
31    ,      biological processes. A person familiar
Text Box 3-12. Examples of Risk Hypotheses,

Hypotheses include known information that sets,
the problem in perspective, and the proposed
relationship that needs evaluation.        .   "

Stressbr-initiated: Chemicals with a high Kow
tend to bioaccumulate.  PMN chemical A has a
KO^, of 5.5 and similar molecular structure as
known chemical stressor B. Hypotheses:  Based
on the Kow of chemical A, the mode of action of
chemical B, and the food web of the target.
ecosystem, when the PMN chemical is released at a
specified rate it will bioaccumulate sufficiently in.
five years to cause developmental problems in
selected wildlife and fish.

Effects-initiated: Bird kills were repeatedly
observed in golf courses following the application
of the pesticide, carbofuran, which is highly toxic.
Hypotheses:  Birds die when they consume
recently applied granulated carbofuran; as the level
of application increases, the number of dead birds,
increases. Exposure also occurs when dead and
dying birds are consumed by other animals. Birds
of prey and scavenger species will die from eating
contaminated birds.

Ecological value^initiated: Waquoit Bay, MA
supports recreational boating, commercial and ,
recreational shell fishing-and is a significant
nursery for fish. Large mats of macroalgae clog
the estuary, most of the eel grass has died and
scallops are gone. Hypotheses: Nutrient loading
from septic systems, air pollution and lawn
fertilizers cause eel grass loss by shading from  .
algal growth, sedimentation and direct nitrate
toxicity.  Fish and shellfish populations are
decreasing because of loss of eel grass habitat, and
periodic hypoxia.
                                                         47'
                                        10/13/95

-------
                                             I   Pliiiiiin
                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                               Piscivorous
                                birds and
                                mammals
                                             Direct exposure pathways
                                                Food chain exposure pathways
Figure 3-2. Diagram of contaminant transport processes in an aquatic ecosystem. (Based on Davis
and Bascietto [1993]; from Bamthouse and Brown, 1994).
                                            48
10/13/95

-------
                    DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                    ORGAMCUMEJUL
                    CON TENT OF SOX.
Figure 3-3. Dynamics contained in FORFLO. (Pearlstine et al., 1985; from Bamthouse and Brown,
                                         49
10/13/95

-------
                     :	; „';,   .   ' •  DRAFT-DO NOT QUOTE, CITE, Oft' DISTRIBUTE '.;..',   ,"	; „'.,.'	'... '"71™;."
  1           with the notation can extract from the diagram the essential features of a quantitative model of energy
  2           sources, transformations, and sinks for the system being represented.  Such diagrams are often quite
  3           complex, however, and are not readily comprehensible to nonexperts."
  4           "For the granular carbofuran study a highly simplified flow diagram was used (figure 3-4a).
  5           Although the diagram leaves out most of the complex ecological processes that occur in agricultural
  6           ecosystems, it provides a reasonable representation of the conceptual model actually used in EPA's
                      '_    '     -«  ••   :'.   •   :;   • ,"   ! :,' > '•  "?„:•                .11             '            i'I
  7           special review of this pesticide.  The only assessment endpoints identified are ground-foraging birds
  8           and the raptors that prey on them. The exposure pathways of concern are (1) ingestion of granules
             , •.  ..  ::   *.•).  '       .,.;.;. •  ••  .';.•  ..  • •.. :„".:•'v  .'.V'         I  i lim	      I                      nil"
  9           by ground-foraging birds,  (2) ingestion of contaminated invertebrates, and (3) secondary poisoning
 10           of raptors feeding on poisoned prey. The only relevant environmental fate data are pesticide
 11           application rates. The  only biological data employed are bird kill data (i e., incident reports and
 12           experimental field trials), laboratory toxicity studies, and results from raptor autopsies.  Figure 3-4b
 13           shows a slightly more complex flow diagram that includes ecological processes that could have been
 14           investigated but that were  deemed irrelevant to the special review.  This diagram, which includes a  .
 15           more detailed representation of the exposure process and considers raptor population dynamics
 16           would be appropriate if the assessment had required quantitative estimates of the impacts of granular
 17           carbofuran on raptor populations." (See Appendix A, Case A-2)
 18           When developing flow diagrams to represent the conceptual model there are a number of factors to
 19       consider: the number of relationships depicted, the comprehensiveness of the information, the certainty
20       surrounding the pathway, and the relationship to methods for measurement.
21           The number of relationships that can be depicted in one flow diagram depends up how comprehensive
22-     each relationship is. The more comprehensive, the fewer relationships that can be shown with clarity. The  -
23       diagram for carbofuran (figure 3-4), shows only a few key relationships. This diagram is easy to understand
24       and follow. Hqwever, what must be weighed against communication value is the loss of information that
25       results from simplification. In stressor-initiated assessments where one stressor and one effect are being
26       diagramed, fairly complete  information can be represented in a single diagram. In complex risk assessments
27       more typical of value-initiated  assessments, multiple stressors, assessment endpoints and responses occur and
28       interact. Models representing these relationships quickly become overwhelming and confusing. This
29       problem can be addressed by producing a series of flow diagrams, that represent different parts and processes
30       of an ecosystem and stressor-response relationships.  Flow diagrams that are complete but cannot be
31       comprehended provide little value.
                                                         50
10/13/95
                                                                                                     . Ill1 ' in"i flip1	H,.'	

-------
                        DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
       a
Bird kill Incident      Raptor mortality
Pesticide
application
to field


Ingestion of
particles by
birds


1
Raptor
predation and
scavenging
                                      Bird kill incident
             application rate
              physical form
             degradation rate
              foraging rate
Pesticide
application
to field
""


^.
Ingestion of
particles by
birds
                           mode of action
                           subtethal effects
                            threshold
                          J.D50
                                                                                      raptor
                                                                                     population
                                                                                     dynamics
                                                                                       7
                                             Raptor population
                                                  reduction
raptor
predation and
scavenging
i
L


secondary
poisoning
i
l
/^ ^v
                               consumption
                                  rate
                            \prey preferences/
 LD50
residues
Figure 3-4. (a) Flow diagram and conceptual model for the granular carbofuran case study;
           (b) Expanded flow diagram for the carbofuran example. (From Bamthouse and
               Brown, 1994)
                                               51
                                                             10/13/95

-------
                                                                      	Ill
                                                                                                           /nil Id «l	
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1          When showing multiple relationships in one flow diagram, it is important to distinguish among
 2      relationships, show how different relationships interact,- and show the degree of confidence the risk assessor
 3      has in each relationship. Flow diagrams that highlight the most important relationships, and relationships
 4      where data are abundant or scarce, can provide insights on how the analyses should be approached.  Such
 5      flow diagrams can also help communicate why certain pathways were pursued and others were not.          *
 6          Flow diagrams that correspond to particular quantification methods provide clear direction for analyses
 7      and can provide a solid basis for the risk assessment. However, using this approach can restrict the
 8      consideration of alternative relationships or factors that may not fit the quantification well.  Some
 9      relationships may be missed or misrepresented, so it is important to think of alternative approaches outside of
10      the accepted quantification method before relying exclusively on a particular approach or set of approaches.
11          Flow diagrams provide a working and dynamic representation of relationships. They should be used to'
12      explore different ways of looking at a problem before selecting one or several to guide analysis. Once the risk
13      hypotheses are selected and flow diagrams  drawn, they set the framework for final planning for the analysis
14      phase.
is                       '                             '	'	,""	'.	"'..'"'   "",.'""  '.'.'"  :	'	,	"	':
16      3.6.   ANALYSIS PLAN
17          The analysis plan is the final stage of problem formulation and represents a specific step, implied but not
18      identified in the Framework Report. Here, risk hypotheses presented in the conceptual model are evaluated to
19      determine how these hypotheses will be assessed using available and new data.  The design of the assessment,
20      data needsj measures, and methods for conducting the analysis phase of the risk assessment are delineated.
21          In the analysis planning stage, decisions must be made about which risk hypotheses can be pursued in the
22      risk assessment and which hypotheses are not feasible to include.  This decision will be based on factors such
23      as management goals, data availability, available  methods and analytical tools, and financial resources. It is
24      critical during this stage of problem formulation that the risk assessor clearly articulates justification for
25      decisions about what was done and, in particular, what was not done.
26          The analysis plan includes the most important pathways and relationships identified during problem
27       formulation that will be pursued in the analysis phase.  Data availability will determine how well these
28      pathways can be pursued  Where data are not available, recommendations for new data collection should be
29      part of the problem formulation. To  improve the quality of the assessment where data are few, data can be
30      used from other locations, on other organisms, where similar problems exist and data are available. Models
31      may.be generated from these data to predict relationships for the planned risk assessment.  When using data
                                                         52
10/13/95

-------
  Text Box 3-13.  Examples of Assessment
  Endpoints and Measures
  Assessment Endpoint:  Coho salmon breeding
  success and fry survival.
  Measures of Effects (formerly measurement
  endpoints)
  •      egg and fry response to low dissolved
         oxygen
  •';';.'.-    adult behavior in response to obstacles
  *;     spawning; behavior and egg survival in:
         response td sedimentation
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      that require extrapolation, justification for using
 2      the data, and associated uncertainty, must be.
 3  -    clearly stated in the analysis plan.  The analysis
 4      plan also includes a clear description of
 5      assumptions made during the development of
 6      hypotheses and models.
 7         -It is in the analysis planning stage that
 8      measures are identified to evaluate the risk
 9      hypotheses. There are three categories of
10     _ measures.  Measures of effect are measures used
11      to evaluate the response of the assessment endpoint
12      when exposed to a stressor (formerly, measurement
13      endpoints).  Measures of exposure are measures
14      of how exposure may be occurring, including how a
15      stressor moves through the environment and how it
16      may co-occur with the assessment endpoint.
17 .     Measures of ecosystem and receptor
18      characteristics include ecosystem characteristics
19      that influence the behavior and location of
20      assessment endpoints and the distribution of a
21      stressor, and life history characteristics of the •
22      assessment endpoint  that may affect exposure or
23      response to the stressor (see text boxes 3-11 and 3-
24      13). These diverse measures increase in important
25      .as the complexity of the assessment increases and
26      are particularly important for value-initiated risk assessments.
27          The analysis plan provides a synopsis of measures that will be used to evaluate risk hypotheses, the
28      extrapolations and models and their formats for presenting the relationships among stressors and assessment
29      endpoints, and the type of data (including quality) and analyses (with specific tests for different types of data)
30      to be completed. This presentation should include a description of how the results will be presented upon
31      completion.
  Measures of Ecosystem and Receptor
  Characteristics
  •       water temperature, velocity and physical
          obstructions
  •       abundance and distribution of suitable
          breeding substrate
  •      .-abundance and distribution of suitable
          food sources for fry
  •       feeding, breeding, resting, reproductive
          cycles
  •       natural population structure .(proportion of
          different size and age classes)
  . ••      laboratory evaluation1 of reproduction,
          growth, mortality
  Measures.of Exposure
  •       number and height of hydroelectric dams-
  *      toxic chemical concentrations in watery
          sedimeht,iish tissue
  •       nutrient and dissolved oxygen levels in
          ambient waters                :
53
                                                                                                  10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE                 :   "
 1          The plan includes explanations about how data analysis will distinguish among hypotheses, provides an
 2      explicit expression of the approach to be used, and justifications for the elimination of hypotheses and
 3      selection of others. The measures to be used to test hypotheses and distinguish among them are clearly
 4      articulated, not just listed. A quality plan contains explicit statemenls for how measures were selected, what
 5      they are intended to evaluate, and which analyses they support. The plan also includes the types of models to
 6      be developed and which stress-response relationships will be generated.  During analysis planning,
 7      uncertainties associated with selected measures and analyses are; articulated and, where possible, plans  for
 8      addressing them are made. Key questions for the assessor to ask include:
 9      •  How will the heterogeneity in the environment or receptor characteristics be described?
10      •  How will gaps between available data and needed information be handled for the ecosystem, receptors of
11          interest and exposure?
12      •  What plans are needed for quality assurance and quality control?
                    ,,; i r  .:;  •     •'•..;••. -,ik  "   •,  ,.'. ,:...•! '•.'•,..' •/ .•»,;. > is"- «| i i.' •, Bi f«: is. ,- ••"! 4,1i,,	•'	•?! i' •  A ' ::i|S':,-'Jl i.",," iF,:;; .•.tw.igwa
13      •  How have all important decision points been documented?
14          The analysis plan is a risk manager-risk assessor checkpoint,  the analysis plan should be presented to
15      the risk managers to ensure that plans for analyses will provide the type and extent of information that  the
16      manager can use for decision-making. The plan presented should include the measures selected, analytical
17      methods planned, and the nature of the risk characterization options and considerations  that will be generated
18      (e.g., quotients, narrative discussion, stressor-response curve with probabilities). Here the risk manager and
19      risk assessor should agree on what can and cannot be done based on the preliminary evaluation of problem
20      formulation, including which relationships to portray for the risk management decision.  A reiteration of the
21      planning discussion is important to ensure that the appropriate balance among the requirements for the
22      decision, data availability, and resource constraints .is established for the risk assessment,
23                        '                 -      -                  '           ..'"''•.-
24      3.7.  UNCERTAINTY IN PROBLEM FORMULATION
25           Throughout the process of problem formulation, ambiguities, errors, and disagreements will occur, all of
26      which contribute to uncertainty'in conclusions about risk. There are six main areas where uncertainty is likely
27      to occur (see table 3-1). Wherever possible, uncertainty should be efiminatedthrough'.better planning.. Wlien
28       uncertainty cannot be eliminated, then a clear description of the nature of the uncertainties should be
                    n "::..   '      ' 	     I.      ' t   , "   , '	: . ,.."..'' » i, »:,      II  "l 111 ' "       '    '     I l" '  I   '      II  l|l I 111
29       summarized at the close of problem formulation, table 3-1 shows the kinds of uncertainties frequently
30       encountered in problem formulation and some of the strategies for reducing them.
 "'           '         „!,».'  '"'	       , ' '  , 	  I' ''' •"  • I •   " i. „  , '  ' ' 	'"''.:'    II     II '  I                           I I   III I  I  III

                                                           54                     '      '              10/13/95

                 •   "•'  '    '•    ';  ::  :.'"'':v.::	''^>' A;:>;C;1&111^
                                            •  '.  ,   '„' •''•'•.' '"'f' ••  »• 1|1!;11'1 ' '•; i*'1 'I;1'"' '|;	"'-"I""'.'.!"	'-'i.r ,;:':• 'i '„*'! i  .*... >''  !:.'*i. .i,' .In''!' ^ ,'i'" | , , 	'>	i	i,1!	J«l'j',:ii,
                                         ": i., , , • , "' . • : '!  ,	' . !, ..V ,i ,„:„;, ''.'f •,!«:,;: iii;!1 '" iiPtt        •,,••,;! ",;,"!	. •" i'• ri1',,l	fur"!'1 i !|lk ,'• L (i1'" ""f'T',:,','
fl      ,      '   '"<     l             '      '': ;.i ,.,;;: V "ii"1, •;'is.' r ;:",<,< "1;:^ :;i:,':i!; i:	^ l;''!ii;;1    '      ^                •           'sSRSm
 	I   J  ., „] ,   " ,,i'; •!. 'i.,,:  	_^	      	''• "illlT l|l|i|,i .,! . P llil'li'l .HlBliH	HI	if ,r Hit !'l!	IP'Jt I'V I	IllfcilllSI ifill ll'T.Villl'i'llliillllllllllllii'llllliillllllilllllllliliiillljIIIIIIIIIIIIIIIIIIJIi 	'  	

-------
DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
1
2-;
' 3
4
5
6
.7 .
8
9
10
11
12
13
14
15
16
17....
18
19
20
•21, . •
Table 3-1. Uncertainty Evaluation in Problem Formulation
So'urce of Uncertainty1
Unclear
Communication
Variability '
Lack of Knowledge:
Model Structure
Uncertainty
Lack of Knowledge:
Extrapolation
-Uncertainty
Lack of Knowledge:
Measurement Error
Simplification and
Approximation
Human Error
Example
Healthy populations vs.
Populations with
individuals that can
survive, reproduce, and
grow.
Differences in species
' sensitivity within the
aquatic community;
variations in weather
patterns.
Choosing the critical
scenarios of exposure and
effects in conceptual
model development:
Difference between
responses of laboratory
rats and field mice
Uncertainty in the
chemical concentration of
a soil sample
Use of long term average
exposures to compare
with chronic effects data.
Mistyped computer code
Problem Formulation Phase Strategies '
Ensure that the assessment endpoint includes both an entity and
attribute .that can be measured directly or indirectly. .Be as clear as
possible in defining assessment endpoints. '.- .
Clearly define the group boundaries and characteristics of interest.
Discuss the strengths and limitations of the conceptual model. '
Identify key assumptions; describe approaches used and their
rationales.
Identify key sources of measurement, error. -
Discuss key assumptions and model simplifications.
Document important decisions; evaluate data sources in terms of"
QA/QC ' ' .
The use of these terms is discussed in section 1.5. • ,-
                  55
10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE-
                                    M    h     ,  , 	     •*
  1                                          4. THE ANALYSIS PHASE
                                            jj"       „',
 2            •   ..   * •       ' •.'•   •   ''' I. • ,:"•"   '; •;'.'.'"'
 3       4.1.  INTRODUCTION
 4          The analysis phase (figure 4-1) consists of the
 5       technical evaluation of data on the potential effects
 6       and exposure of the stressor(s) identified during
 7       problem formulation. Inputs to the analysis phase
 8       include data on exposure (which may include
 9       source information, measurements of stressor
                .' :•   '•:!	      '  ' . i   i    "	k . - '.;•. u...	i
10       levels, or direct measurements of exposure), the
11       ecosystem, and biological effects.  This
12       information is used to conduct exposure and
13       ecological response analyses,  the primary
14 •      activities in this phase of risk  assessment. The
15       outputs of the analysis phase are summarized in
16       exposure and stressor-response profiles, which are
17       integrated in the next phase of the assessment, risk
18       characterization.
                    W  • , ••  '   '  '  :•"   ,  •'.;'•• •',  •••. i'1  \.
19          In some complex assessments, the
20       characterizations of exposure  and effects are so intertwined that they can be difficult to distinguish. In
21       addition, the distinction between the analysis phase and risk estimation can become blurred.  These activities
22       are conducted in close iteration when  secondary stressors are formed through biological processes or when
                                     '    V ..'.". 	, .• .:                          r
23       secondary (indirect) effects are important and risks to different ecological components are estimated
24       iteratively. For example, estimating risks of fish population decline as a result of nutrient addition to an
25       aquatic ecosystem may require estimating increases in productivity and decomposition rates of the plant
26       community, then the decline in dissolved oxygen levels in the water column, and finally mortality of fish
27       encountering areas of low dissolved oxygen. Parts of this complex scenario may be evaluated using
28       combined exposure and effects models, making it difficult to tease apart the different aspects of the analysis
29       and integration. These guidelines retain the distinctions of exposure and effects characterization and risk
30       estimation so that assessors clearly recognize the types and sources of information that are needed.
Text Box 4-1. What is Different in the Analysis
Phase Diagram?
   V    The left-hand side of figure 4-1 shows the
general process of characterization of exposure,
and the? right-hand side shows the characterization
of ecological effects. These two aspects of analysis
must closely interact to produce compatible output
that can be integrated in risk characterization. The
dotted line and hexagon that includes both the
exposure and ecological response analyses
(changed slightly from the Framework Report's
figure 3) emphasize this interaction.
        In addition, the left-hand  box under
Characterization of Exposure now reads Relevant
Exposure Da fa to reflect the wide variety of data
that serve as input to exposure analyses. Such data
may include source-term information,
measurements of stressor levels in the environment,
or direct evidence of exposure (e.g., body burdens
of chemicals)..
                                                         56
                                         10/13/95

-------
                            DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                                                    PROBLEM FORMULATION

                                                    ANALYSIS

                                                    RISK CHARACTERIZATION
                                  PROBLEM FORMULATION
          Characterization of Exposure
Characterization of Ecological Effects
Relevant
Exposure Data


Ecosystem
Characteristics;
Blotlc
Abiotic


Relevant
Effects Data
I
                              I
                             Exposure
                             Analysis
          I
      Ecological
      Response
       Analysis
                                                         Stressor-
                                                         Response
                                                          Profile
                                  RISK CHARACTERIZATION
                                                                                         O
                                                                                         to
                                                                                         e.
o
3
                                                                                         fig
                                                                                         a

                                                                                         .o
                                        (Q
1    . Figure 4-1. Analysis phase
                                                 57.
                                       10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, C)R DISTRIBUTE
 1          Many of the issues in the analysis phase are associated with the evaluation and analysis of data. This
 2      section discusses common sources of information and analysis approaches used for risk assessment, and their
           i    " ,    mijliii  |.i,  i' ' , '"' • "' . '	 i, !  C  ' ' "''•!• " 'i'li'ii •'!'„''":"'!	"[ I      I     II  II I II  11                !r   "              II
 3      associated strengths and limitations. Cited references can provide the reader with additional information on a
 4      particular topic.  While a detailed treatment of data acquisition and model development is beyond the scope of
 5      these guidelines, the risk assessor should consider the following points.
 6      •   Data used in the assessment should meet applicable data quality objectives (DQOs).  Additional
 7          information on sampling strategies and data considerations may be found in EPA's exposure assessment
 8          guidelines (U.S. EPA, 1992d), the EPA report Guidance for the Data Quality Objectives Process (U.S.
 9          EPA, 1994d), and the issue papers (Barnthouse and Brown, 1994; Sheehan and Loucks,  1994).
10      •   Mathematical models used in an assessment should be appropriate to the available data and address
11          pertinent scientific and decision-making questions. Modeling protocols are important whatever the type
12          of model selected.  In addition to scale, resolution, and boundary conditions, the goals of the modeling
13          should be described in a way that facilitates comparisons between goals and modeling results.
14          Procedures for model calibration and uncertainty analysis also should be described.  Model results will be
15          most useful if the following information is provided:
                 ,    ,	  ;"",   i  .  •.  ;••    :                   _       i   i i    i        i               i    i
16           »•    Tables of all parameter values used for analysis
                    ,:;!']:,  '  „     !\ i   ;,, •'. •'..    ^                     I   I II  (I  I  I          I
17           "    Parameter estimation techniques and associated uncertainties
                     1 i  ,   •         ,.• •                                 in  i                           iii
IS           *•    Tables or graphs of results
19           >•    Accuracy of results
20           »•    Parameter sensitivity analysis to evaluate model responses to changes in input parameters
21           Additional  information on principles for model building, aggregation, and uncertainty is summarized in
22           the uncertainty issue paper (Smith and Shugart, 1994). Further guidance on modeling for hydrogeologic
23           systems has been developed (U.S. EPA, 1994f).
24           The approach to the analysis phase depends on the outcome of problem formulation. The following
25      questions describe some of the information from problem formulation that will influence the assessor's
26      approach to the analysis phase.  Depending on the particular assessment, certain questions may be
27      emphasized over others, and answers to all may not be available prior to beginning the analysis phase.
28           Is the primary stressor(s) of concern a chemical, physical, or biological entity^or a combination o/        I
29      multiple stressors?  By the close of problem formulation, the assessor should have a clear idea of the
30      principal stressors of concern. Currently, the most common approach to risk assessment is to evaluate
31      stressors individually, and this section of the guidelines is organized accordingly. Multiple stressors are
        mi
        iii 11
                                                         58
10/13/95

-------
                          •      "DRAFT-DO NOT QUOTE, CITE,-OR DISTRIBUTE     •..'.--
 1      evaluated by aggregating risks attributable to individual stressors or by working.from an aggregate measure
 2      of effect to identify the principal stressors responsible.
 3        -Is the stressor already present in the ecosystem under evaluation?  Some stressors occur naturally in
 4      ecosystems. For example, many soils contain heavy metals.  Similarly, some ecosystems have evolved under
 5      the influence of disturbances such as floods or fire.  In these cases, an evaluation of baseline or background
 6      levels of the stressor is often included when characterizing exposure, and adaptive or potentiating
 7      mechanisms are explored when characterizing effects."
 8           What level of biological organization is being evaluated? What adverse effects are likely?  The
 9      assessment endpoint identifies the level of biological organization and type of effect that are the subject of the
10      assessment.                                        ,                •
11         •  Are there ecosystem characteristics or intermittent events that will influence fate and transport or'
12      mitigate physical or biological stressors? Ecosystems can be characterized in an almost unlimited number
13      of vyays, so it is important that the assessment focus on those characteristics that will influence the behavior
14      and effects of stressors. Both abiotic (e.g., storm events) and biotic (e.g., trophic status) characteristics of
15   .   ecosystems can influence risk. In addition, processes acting at the landscape level (e.g., patterns of
16      disturbance, population sources and sinks) can be important factors and will undoubtedly play an increasing
17      role in ecological risk assessment as this field continues to develop.
18     .      Will secondary> stressors be, formed,  and are  they of 'concern?  Will secondary effects be evaluated?
19      Interactions between stressors and the ecosystem can produce additional (secondary) stressors that may be of
20      concern. In addition, a biological effect may generate additional effects in the ecosystem through interspecies
21      interactions (Figure 4-2). The analysis of secondary stressors and effects can greatly expand the effort.
22      required to perform an ecological risk assessment, so it is important that they be identified as  early as
23      possible.  However, the outcome of these interactions can be difficult to predict; later identification may
24      require an  additional iteration through the analysis phase.
25           This section is organized by stressor type.  Sections 4.2,4.3, and 4.4 address chemical, physical, and
26      biological  stressors, respectively, and section 4.5 discusses the substantially more complex issue of multiple
27      stressors.  The remainder of this section discusses the use of illustrative examples and the evaluation of
28      uncertainty.
                                                          59
10/13/95

-------
                                                                           ll III It  III III 11
                                                                           11IIII III III  111 I Mill
                   DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
       Primary Stressor
       (e.g., building logging roads)
           i
    Secondary
    Stressor
    (e.g., increased
    siltation of stream)
             ,;;,;;;;,;   	 - N.
       Interaction with   N.
         ecosystem       \
    (e.g., slope, soil type)^/
   Exposure
   ofreceptor
                              Exposure
                              cf iF
                              (e.g., no primary exposure
                              pathway for logging
                              road example)
Primary Effect
(e.g., smothering of
benthic insects)
                                               >X^ Interspecies   N.
                                            /     interaction      x.
                                            N.  (e.g., food, habitat,  /
                                              ^V , competition)   jS
                                               « ^              /
                                           Secondlary (Indirect) Effect
                                           (e.g., decreased abundance
                                           of insectivorous fish)
                                      , ir ;; '';,' i';,!'. „; ,it	I1 \,,,,: ",„«,! pi':': iiiiii1 ii] i* !|j:: S|t i!!11'1 j*1 ™' i \| ^s11 w^^^^          	!"'';' 'i1 ""!!l
                                          1 "':,.i:	•«.; I'si'v«' 'JtHWiM-'iVft.*)!;1 ''"'i1 t'f'Xi;»,:,fTii;i..i-i'.ii .-'tr1 ";:'! •
Figure 4-2. Relationships between primary and secondary stressors and effects
                             '.v1'1' :	'• 60
                                         10/13/95
                  ".	;i;:''.i « Si IS	*!" N • 3,"1'!:." !/ •.! ''i	;; I. I';,': ",•, •-, ,»•": i>', i3	Bi:' ifS!! .iDS^	jlii I
             ''•llllliL
J:4lL'U&^

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1                  •                              •            '•'''.     '•-•'-           •           -
  2       4.1.1.  Illustrative Examples
  3            While figure 4-1 depicts an overarching view of the analysis phase, the process can be implemented in a
  4       wide variety of ways. Sample flow diagrams capture some of this variety throughout this section. These
  5       samples do not represent all possible approaches; instead, they illustrate how the components of the analysis
  6       phase were brought together in actual cases.                                 "
  7       ;        '      .       '    '•                   .       .•''..•         ""''.
  8       4.1.2.  Characterizing Uncertainty in the Analysis Phase
  9            The objective of characterizing uncertainty in the analysis phase is to describe and, where possible,
 10       quantify what is known and not known about exposure and effects in the system of interest. Table 4-1
 11       summarizes sources of uncertainty that are commonly encountered in the analysis phase;  Effective treatment
 12       of uncertainty in problem formulation makes the assessor's task in the analysis phase easier.
 13            The uncertainty evaluation in the analysis phase has both quantitative and qualitative aspects/  For
 14,      -example, while the strengths and limitations  of the conceptual model should have been discussed in problem
 15       formulation, the analysis phase should include a discussion and evaluation of key assumptions and
                      >                                      .                              '
 16       simplifications. Discussion of mathematical models used in the analysis phase should include similar points.
 17       Clear communication can become an important issue when evaluating literature sources of information. The
 18       boundaries and characteristics of the system  under study must be critically evaluated to determine their
 19       relevance to the assessment at hand. Uncertainty may arise where these attributes are insufficiently
 20       described.  Human error can be controlled through adherence to principles of quality assurance and quality
 21       control and by ensuring that calculations or computer codes can be and are cross-checked.
 22     "       The quantitative aspects of uncertainty  analysis address variability, extrapolation uncertainty and
 23       measurement error.  As used here,  the term variability refers to the true heterogeneity in^a characteristic.
.24       Examples in ecological risk assessment include the variability in soil organic carbon, seasonal changes in,
 25       stream flow, and seasonal differences in the diet of animals.  As discussed in section 1.5, the description of
 26       this heterogeneity is usually conducted as part of uncertainty analysis, although heterogeneity may not reflect
 27       a lack of knowledge, and usually cannot be reduced by further measurement.  Variability can be quantified by
 28       presenting a distribution, or by  presenting one or more points of a distribution (e.g., mean, 95th percentile).
 29       When presenting information on variability,  the assessor should identify the source of information, the group
 30      that it is intended to represent, and discuss its relevance to the assessment. Particular care should be taken
 31       \vhencombininginformationfromseveraldifferentsources.
                                                          61
10/13/95

-------
i1, ;"' \ 	 'i:1;1!!1
1
2
3
4
5
6
4
7
8
9
10
11
12
13
14
15
16
17
18
19
	 ' '!',!. '„ - i ,,','i "I."1, ' i 	 n!" :••" liifiii!11;;™;!! n i ii HI in i MII n in n gj in
PI 1 ill 1 i 1 	 1 1 	 1 	 liilllillilliliill ill III 	 Illllliilllilllllliillilillill In! liilliilllillililil 	 liili(ilililililllllllllllil 	 lllililil illllililil 	 (lull nil 	 Ill 	 lilillliillllM
i nil n i i in inn inn n i n i i in i i n n i i n nun mi n i nil niiiini inn I iiiiini innnnnnnnninn
> If1!
	 i 	
DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
Cable 4-1. Uncertainty Evaluation in the Analysis Phase
Source of Uncertainty1
Unclear
Communication
Variability
ILack of Knowledge:
Model Structure
Uncertainty
Lack of Knowledge:
Extrapolation
Uncertainty
Lack of Knowledge:
Measurement Error
Simplification
Human Error
Example
Reporting of average
concentrations without
specifying whether it is
an arithmetic or
geometric mean.
Differences in species
sensitivity within the
aquatic community;
variations in weather
patterns.
The use of a linear model
to estimate uptake of
metals in plants
Difference between
responses of laboratory
rats and field mice
Uncertainty in the
chemical concentration of
a soil sample
Use of long term average
exposures to compare
with chronic effects data.
Mistyped computer code
Analysis Phase Strategies
Critically review objectives, design and study group boundaries of
literature studies.
Describe heterogeneity using point estimates (e.g., central tendency and
high end) or by constructing probability or frequency distributions.
Evaluate power of designed experiments to detect differences.
Differentiate from measurement uncertainty and systematic error.
Distinguish between data that can be reduced with further data and that
which cannot.
Discuss key assumptions underlying model choices.
Evaluate whether alternative models should be'combined formally or
treated separately.
Identify key assumptions; describe approaches used and their
rationales.
Use standard statistical methods to construct probability distributions or
point estimates (e.g., confidence limits).
Discuss key aggregations and model simplifications.
Conduct QA/QC; ensure that data sources followed good laboratory
practices.
|U| " .'-!]„, ' i' ni; " . ' 	 '" i. ' ',:'.' .'"" i" '"'• ! /'"I!1"'1 	 " 	 > 'I II 1 1 II 1 1 III III!
The use of these terms is discussed in section 1,5.
i ii in
62
10/17/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1           In contrast to variability, both measurement error and extrapolation uncertainty can potentially be
• 2      reduced by taking further measurements  Measurement error is the difference between the true value and the
 3.      measured value  It arises from, random variation in the characteristic of interest When that characteristic is
 4      biological response, measurement error can greatly influence the ability of the study to detect effects
 5      Properly designed studies will specify sample sizes sufficiently large to detect important signals
 6      Unfortunately, many studies have sample sizes such that only gross changes in conditions can be detected
.7      (Smith and Shugart,  1994; Peterman, 1990)  The analysis phase"uncertainty discussion should highlight
 8 -    situations where the power to detect differences is low
 9           Even if data from well-designed experiments are available, the attribute being investigated is rarely the
10      attribute of interest in the risk assessment (extrapolation uncertainty)  Examples of extrapolation uncertainty
11   ,   include measurements of responses in laboratory animals when the response in the wild population is of
12      interest, or measurement of bioaccumulation in one field situation that is different from the system of interest
13      Common approaches to characterizing extrapolation uncertainty .for ecological effects are discussed in section
14      4232  The assessor should discuss key extrapolation assumptions, and describe the approach used and its
15      rationale  In cases where the data were collected for purposes other than risk assessment, the assessor should
16      address the relevance of the data for the assessment  Mismatches between hypotheses increase uncertainty
17      and can lead to potentially misleading conclusions (Smith and Shugart, 1994)
18           One of the more important objectives of characterizing uncertainty in the analysis phase is to distinguish
19      variability from uncertainties arising from lack of knowledge (e g ; extrapolation uncertainty and
20      measurement error) (U S EPA, 1995d)  This distinction facilitates the interpretation and communication of
21    '  results  For example, in their food web models of herons and mink, Macintosh et  al (1994) separated
22      variability expected among individual animals from the uncertainty in the mean size and concentration in prey
23      species  In this way, the assessors could place error bounds on the distribution of exposure among the
24      animals using the site and estimate the .proportion of animals that might exceed a toxicity threshold
25           Methods to analyze and communication uncertainty remain an area of active research  Sensitivity
26      analysis can be used to evaluate the influence of different parameters on the risk assessment, and it is often
27    .  suggested as one of the first steps of any quantitative analysis  The calculation of one or more point
28      estimates is one of the most common approaches to presenting analysis results; point estimates that reflect
29     . different aspects of uncertainty can have great value if appropriately developed and communicated  The
30      development of easy-to-use software for Monte Carlo analysis has greatly increased the application of this
31      particular method  for uncertainty analysis; readers are encouraged to follow the best practices that are   ..
                                                  •      63       .                                  10/17/95

-------
                                  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 I
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
emerging for this method (e.g., Burmaster and Anderson [1994] and guidance prepared by U.S. EPA Region
III [U.S. EPA, I994g]). Other methods (e.g., fuzzy mathematics, Bayesian methodologies) are available, but
have not been extensively applied to ecological risk assessment problems (Smith and Shugart, 1994).  These
guidelines do not endorse the-use of any one method and note that the poor execution of any method can
             'rill'   	 . • . .  "i , :'  l"   "'i"  ! j  :' .'	,i.  c,	Hi'"'	•••	      	Pi;	»'Wt*M'l11»W*ni**!liflIVW.»ilftil7*';
obscure rather than clarify the impact of uncertainty on an assessment's results. No matter what technique is
used, the sources of uncertainty discussed above should be addressed.
4.2. ANALYSIS OF CHEMICAL STRESSORS
4.2.1.  Introduction          _	
     Chemical stressors are currently the focus of much of the risk assessment activity in EPA. Included in
this category are a wide range of substances, including pesticides, hazardous substances, nutrients, and
radionuclides. Types of ecological risk assessments of chemical stressors also vary widely within the
Agency, from purely predictive assessments of single chemicals—for example, the evaluation of new chemical
releases—to retrospective assessments of site-specific mixtures as are often encountered at hazardous waste
sites. The assessments may focus primarily on direct effects from exposure to contaminated media such as
food, water, sediment, and soil or may follow more complex pathways; of secondary effects, as is often
necessary when evaluating the risks associated with adding nutrients. The principles discussed in this section
are intended to be flexible enough to be applicable to this wide range of analyses.
     The organization of section 4.2 follows the two principal activities in the analysis phase,
characterization of exposure (section 4.2.2) and characterization of ecological effects (section 4.2.3). These
two activities are easily identifiable for many chemical assessments, although they are often iterative in
assessments that encompass secondary effects. Figure 4-3 illustrates how these concepts were implemented
in an actual case, the assessment of new chemical releases under the joxic Substances Control Act (TSCA)
(Lynchetal.  1994).   '"   """"''   " ' 	'''\ '"'"''	''"''' ''''"^.T'™?
                                                        '!' ',1 'I1''!	,!''i r'lil'n, jf'i	i ]'' illllliHW^^	R '"• •,',' v j" ii,,"' ': ,1."!,"'' I 'II''»; :i * Jf I'1!', ''I III'!' , I	 '" 	'llll'li" W'I'1' •' I'', ','i	''' ."i'll'i I:»' illi i "'P'1' ill
4.2.2.  Characterization of Exposure to Chemicals
         ,",'   	iU'i'i! '-I1 i, ,.:,".'"",", ";" I'1- -   i •' i!	/.i,,' -.i "".in," «  •!•
     The characterizatiSn of exposure to chemicals encompasses three objectives: to characterize (1) releases       I
  -        ;, •; I'll" i" ;;,. • i.   ;:, • '":',5 ;•' •:;. ?•  '<<"¥:}" -{f .;"•!:""";;,;'?jI'isrfei^a^'JW'^iwiRiiF^'iiiM	'is	lisa                          I
into the environment (section 4.2.2.1), (2) the subsequent spatial and temporal distribution within the
     ',!   I,   '  ii''!i!'!i!;!|  ','''" '  '  ' 	:' j1 IM 'i	•"' '  in,' i ' ',ii",";';; ii"/ \" n. i" "™: IP' '"ii fi''»'"i1.", "ii,1"!:1!';,' ^''TfL'''^^?'1^          '!.!""i",i 'i1'"';!!!,1' ..'"'iii''! i1'"!'"„''''!!'i"',,1,' i,'''1'. '.'iJ!! ",',"! »""''!„'!, i""1'!" ",1'i'i;1" ,1:,''"'';„!;!; '!'„ !il';";!;,:'!;, ,r/!!'!!;''i,;iii;!i/.!!_^
environment (section 4.2.2.2), and (3) contact with the  ecological component of concern (section 4.2.2.3).
     '•' '  • ''   i'    ;:i" •  " '"•'' "•:.:."•  ;•'•.. • •'-'' .''''^ •        '-'.'	T'' *:-!f:':
                                                           64
                                                                                                 1-,!!!"*1" '°:i!:;	lili'l;!''i	^I^HM'Sii™^       I
                                                                                                 ''  ll::i%<^\iiiii!,;iiu^        I

-------
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
            Release:
            from POTW or point
            source discharge
  Source
  assumptions
               Chemical:
               New Chemical
  Transport
  and .
  transformation
  model
         Spatial and
         Temporal Distribution
         in Aquatic Environment
        jS  Exposure:    N.
     v
     ^v    to PMN in water    /
        >v    column      x^
 Ecosystem:
 Pelagic waters
                      Receptor:
                      Pelagic community
                     Life history
                     information

                     Selection of
                     indicator
                     organisms

Natural history characteristics
                                   influencing exposure
                               Stressor-response
                               Relationship:
                               PMN concentration
                               vs. mortality, growth
                               reproduction of
                              , indicator organisms
                          Causal Evidence:
                          Analogy,
                          Mechanism of
                          action
           Exposure
            Profile
                                                                      i
            Stressor-Response
                  Profile
Figure 4-3. Analysis example:  new chemical
                                          65
                                            10/13/95

-------
 1
 2
 3
 4
 5
 6
 7
 8  .
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
                         DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE

The results of exposure characterization are
summarized in the exposure profile, discussed in
section 4.2.2 A. While the pursuit of these goals is
presented in a stepwise fashion in this guidance, it
is important to recognize that the exposure
characterization process  might be entered at any
one of these steps, depending on the information at
         " ""   hiii'iil   I1!  ...   " in  ,,, " ' Ii' i ,  ,	'  :  !!:1	,,,	'r .;„ ',,:ii	 ,!
hand and the scope and purpose of the assessment.
         1 " i   ii'ii|iilll,   * ,„' M . 	„'  ,. ii	ii,ii: ''/ ,.' ', "', ;,ip	: , ;' ,,^	!: „  :'!' ii
In addition, although most exposure
  	  -,-    '-	fi   !•',  - ii,  • ;  .la	!";:;"•. •;•,•,,,  	•;.'• ..;  •':>.;  "";,"'
characterizations address all three objectives, the
amount of emphasis on each  will  depend  on the scope and purpose of the assessment. This section of the
    	   •:.  '•   II                                 I         Mill 111 HI HIM 1   II   I  I III HI  IP       I   Hill I  I III  II  11(111 I I,1
guidelines draws extensively from concepts found in the Characterization of Exposure issue paper (Suter et
                                                   Text Box 4-2., Example: The Assessment of
                                                   New Chemical Releases Under TSCA--
                                                   The Exposure Characterization Process
                                                          The exposure characterization process for
                                                   the new chemicals case study is shown on the left-
                                                   hand side of figure 4-3.  As can be seen, the case
                                                   study combined information on releases, the
                                                   ecosystem, and receptors to estimate the spatial and
                                                   temporal distribution of the chemical in the aquatic
                                                   systems and Its contact with aquatic organisms.
al., 1994), but the issue paper materials have been modified as necessary to meet Agency needs.
                                                   Text Box 4-3. New Chemical Example:
                                                   Analysis of Sources and Releases
                                                          In the new chemical case, investigators
                                                   used an entirely predictive approach. Effluents
                                                   containing the substance would first be treated in
                                                   publicly owned treatment works (POTW).  To
                                                   estimate the release of the substance by POTWs to
                                                   pelagic water, assessors used data from laboratory-
                                                   scale wastewater treatment experiments and the
                                                   output from mathematical wastewater treatment
                                                   simulations.
4.2.2.1.  Characterizing Sources and Releases ~
     The first objective of the analysis phase of
many chemical assessments is to define the source
term, that is, the type, magnitude, and pattern of
chemical(s) released. Suter et al. (1994) discuss
the wide range of sources: mobile or stationary
(e.g., cars vs. sewage treatment plants), point or
nonpoint (smokestacks vs. agricultural runoff), and
deliberate (pesticide applications), adventitious
(brake fluid leaks), or accidental (spills).  Releases may be the result of ongoing human activity or result from
past activity.
     A complete source characterization includes the specific content, timing, duration, location, and intensity
of any releases. In addition, the source characterization should consider whether other constituents emitted by
             ,	i'|i •. i,• •,  •'!• • •' •;;: "i:.•• a • • '.:;>  :'	:-\•";~*.•  \:;:''-:;>• •;!'»tia11;snare	81, iftiiBWH*1.'^*M-tiMXHW:** f^isiyr^ss^ssv
the source will influence transport, transformation, or bioavailability.  For example, the presence of chloride
in the fee4 stock of a cpal-fired power plant will influence whether mercury will be emitted in divalent (e.g.,
         .i •   iti!1!*! i<<:  .  •	                          i   in i i in  ii iiiiiiii il i     i  i  i  i i  i       i         inn i
as mercuric chloride) or elemental form (Meij, 1991). Finally, the assessor should consider whether there are
                     .•ll
                                                          ';^:i!::V
                                                             ....... .................. .............. ..................................... ....................... [[[ ....................
                                                         66

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3
  4
  5-
  6
  7 .
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
.18
 19
 20
 21'
 22
 23
 24
 25
 26
 27 '
 28
 29
 30
 other sources of chemicals in the ecosystem.  The issue of defining background concentrations is of,
 particular concern when the chemicals occur naturally (e.g., most metals),'are generally widespread from
 other sources (e.g., PAHs in urban ecosystems), or have significant sources outside the .boundaries of the
 current assessment (e.g., atmospheric nitrogen deposited on the Chesapeake Bay). In these cases, an
 evaluation of background concentrations (along with an explanation of how background^ being defined)
 may be an important component of the assessment.

 4.2.2.2. Characterizing the Spatial and Temporal Distribution of Chemicals in the Environment
     The second objective of exposure characterization is to estimate the spatial and temporal distribution of
 the chemical(s) in the environment, and may include
- the air, soil, sediment^ water or biota. Fate-and-
 transport modeling and measuring concentrations in
 environmental media are two common approaches
 used to accomplish this task. Site:specific
 measurements can be used to directly estimate the
 spatial and temporal distribution of chemicals or to
 confirm the results of a model.  They can greatly
 increase the confidence in the assessment when
Text Box 4-4. New Chemical Example:
Distribution of Chemicals in the Environment
       To estimate concentrations in receiving
streams, assessors divided the mass of chemical
released per day by estimated stream flow.
Assessors used mean and low flow in U.S. streams,
and used the 1 Oth and 50th percentiles of these
distributions.
 collected in accordance with an appropriate design. Guidance on taking chemical measurements and selecting
 and using fate and transport models is outside the scope of these guidelines.  While many issues are similar to
 those encountered in human exposure assessments for chemicals discussed in Exposure Assessment
 Guidelines (U.S. EPA, 1992d), ecological risk assessments may also consider the modeling of chemical
 movement .through food webs (e.g., bioaccumulation, biomagnification) when stressors include persistent^
 lipophilic chemicals or certain metals.
     Chemical distribution in the environment is influenced by characteristics of both the chemical and the
 ecosystem into which it is released. Some important considerations are listed below.
   •  What specific chemical or chemical form is of toxicological-concern?  The characterization of
 exposure and ecological effects steps must be coordinated so that the lexicologically important chemical or
 chemical forms are addressed. This is a particularly important issue for metals and complex mixtures (e.g.,
 PCBs or dioxins), where individual chemicals and chemical forms have very different properties influencing
                                                        67
                                                                                          10/13/95

-------
                    '•ii           DRAFTS-DO NOT QUOTE, CITE, OR DISTRIBUTE
     1 .    	,     '  ••..(! •"•    ,                                         i  i                                MM
 1      fate and transport and toxicity. In addition the nature of chemicals can change depending on environmental
 2      conditions.  In aquatic systems, factors such as pH, hardness, and the presence of mitigating substances such
        !    ''     "   ''Slfi   ' , ;•                                  I I    I   II II I 111 II I      I     III   I  I I II II    II II 111 Illllll 111  I 1111
 3      as dissolved prganic matter can^change the form or bioavailability of a chemical.
 4           Will transformation processes generate additional chemicals that should be assessed?  In addition to
 5      being transported, chemicals may be transformed through biotic or abiotic processes, Additional chemicals
 6      (metabolites or degradation products) may be formed through these processes and may also be of concern.
 7      For example, many azo dyes are not toxic because of their large molecular size But, in an anaerobic
 8      environment, the polymer is hydrolyzed into more toxic water-soluble units.  Microbial action'increases the
 9      bioaccumulation of mercury by transforming it the inorganic form to organic forms,  A related issue is the
10      formation of secondary stressors through ecosystem processes. For example, nutrient inputs into an estuary
11      can result in decreased dissolved oxygen concentrations by increasing rates of production and subsequent
12      decomposition. Evaluating the formation of secondary stressors often falls within the purview of the
13      exposure characterization. Coordination with the ecological effects characterization is important to ensure
14      tha't all important chemicals and secondary stressors are evaluated.
15           How will ecosystem characteristics that influence transport be characterized (e.g., physical and
16      chemical attributes, seasonal and intermittent events) ?  In many predictive assessments, the assessment is
17      not tied to a specific location. In these cases, a range of potential values characterizing the ecosystem is often
18      used. For example, the new chemical case study used stream flows intended to represent mean and low flow
19      in U.S. streams.  Another approach is to define a "canonical" or "reference" environment based on current or
20      historical data (e.g., a warm, "blackwater" southeastern U.S. stream; a western mountain lake; a southwestern
21      irrigation canal) and estimate exposure within it. If this approach is used for location-specific assessments,
22      the reference system is selected.to match the site as closely as possible in all aspects except the presence of
23      thestressor.  More often, site-specific measurements are taken to characterize environmental attributes that
24      may influence transport.
25             '      '"  i            ^    '' ^   ."  ' '   \	'    \-Zr:'.,3""".''.'."','',.'•	'".'•'	",'	::'"•'".
26      4.2.2.3.  Estimating Chemical Exposure
27           The third objective of exposure characterization is to estimate chemical exposure.  Conceptually,
28      exposure can be thought of as the intersection of a chemical with a receptor in time and space (figure 4-4).
29           Most approaches to estimating exposure have been targeted at the individual organism, EPA has
30      defined chemical exposure of humans in terms of contact of a chemical with the outer boundary of an
                                                     ;::::• T68.'
10/13/95

-------
                                 DRAFT-DO NOT'QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3
  4
  5
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
-2.1
 22
 23
 24
 25
 26
 27
 28
 29
                                       Exposure
Figure 4-4. Exposure is the intersection of the
chemical with the receptor in space and time.
organism (US. EPA, 1992d), Exposure is
commonly quantified as the amount of a
chemical ingested, inhaled, or in material
applied to the skin (potential dose), or as the
amount of chemical that has been absorbed
and is available for interaction with
biologically significant receptors (internal
dose); These concepts work well for
exposure of nonhuman individuals as well,
     Exposure assessment for a population
can be accomplished by incorporating the
variability in exposure among individuals
within the population,  Exposure estimates can be presented as a distribution of exposure in the population or
as point estimates (e.g., the number or proportion of individuals incurring exposures above a particular
threshold value).  Exposure at higher levels of organization (e.g., communities, ecosystems);        -
     .., can be accomplished by estimating the exposure of the component parts or by establishing an
     operational boundary around the entire unit, of interest (e,g,, the perimeter of a lake'including 30 cm of .
     sediment). For the component-part approach, exposure routes can be evaluated as described above [for
     individuals or populations] but attention must be paid to the full distribution of exposure to each
     component, For the operational-boundary approach, exposure routes can be evaluated as fluxes across
  ,   the boundary (e.g., atmospheric deposition,  inflow, burial,in deep sediment, outflow) (Suter et.al,,
     1994).      .•-                      '

Estimation Approaches                                      .                                   ;
     Two common techniques used to estimate ecological exposure (Suter et al,  1994; U.S, EPA, 1992d) are
(1) estimation of effective concentrations or doses in media and (2) measurement of residues or biom'arkers in
the receptors.
                                                        .69
                                               10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1           In selectifig the most appropriate approach to estimating exposure, the assessor should consider the
       , '            "Sii     " •:••.  .' i,	 !!!',;:•  ••' • -: (M*..', "        I     |	|,,l II'I              '                    M
 2      following:
 3           In what form are the effects data (e.g., ppm diet, tissue concentrations, 'dose)? For the results of an
 4      exposure assessment to be useful, they must be expressed in a form that can be compared with the stressor-
 5      response profile generated in the effects assessment.  Because the effects assessment is often based on data
 6      that have been previously collected (e.g., a laboratory toxicity test), often the units and dimensionality of the
 7      effects data'must be matched during the exposure assessment. It is also important during the exposure
 8      assessment to identify situations where the conditions under which the effects data were collected may either
 9      be inappropriate or substantially vary from actual exposure conditions.  Currently, most stressor-response
10      relationships express the amount of stressor in terms of media concentration or potential dose.3  Fewer
11      express it as tissue concentration, and fewer still in terms of a biomarker or bioassay.  For mis reason, tissue
12      concentrations and biomarkers are less frequently used to characterize exposure.
13           What is known about the bioavailability of the chemical? Bioavailability refers to the fraction of the
14      total chemical in the surrounding environment that is available for uptake by organisms (Rand and Petrocelli,
15      1985),  Bioavailabifity is a function of the chemical (e.g., form or valence state), the medium (e.g., sorptive
16      properties or presence of solvents), the biological membrane (e.g., sorptive properties), and the organism
17   .   (e.g., sickness, active uptake) (Suter et al., 1994). Because of interactions among these four factors,
18      bioavailability factors will vary on a site-specific basis.  In some cases, factors that influence bioavailability
19      arc well known and easily measured (e.g., organic carbon content and nonpolar organic compounds in soil,
20      complexed versus free cyanide in water).  In many other cases, these factors are either not known or
21      insufficiently characterized.   If bioavailability is expected to be a significant issue, and the factors
22      influencing bioavailability are not known, tissue measurements, biomarkers, or other bioassessment methods
23      may be more useful in estimating or confirming exposure,
24           Is enough known to appropriately interpret tissue concentrations or biomarkers?  While tissue
25      concentrations provide direct information on internal exposure, they can be difficult to interpret without
26      information on the chemical's distribution and metabolism within the organism and knowledge of the
27      organism's behavior prior to capture.  Biomarkers (i.e., biochemical or physical changes in an exposed
                                                                - .'•f.'Wi'M! •ll*;,'-:;)'1!1:.
                                                                               '• f.-;-,;,(! -i >i nil1 Sii","..:;.'	.'!,! I..*1 •>! !•'•<,Sv':'i.-KKXffltllr'i',*maic!'af:U, I
           3As discussed above, potential dose is defined in U.S. EPA (1992d) as the amount of chemical ingested,
         inhaled, or in material applied to the skin.
                                                          70
10/13/95
                                                                                                      	I"!	|«lh I'l'li",	I!

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      organism) also measure responses that are at or near the action site.  In addition, they'may persist longer in
 2      the organism than the chemical (e.g., acetylcholinesterase activity in brain tissue) and may respond to suites  •
 3      of chemicals that have similar modes of action and occur as mixtures in the environment (e.g., the H4IIE
 4      bioassay used to estimate exposure to dioxins). However, biomarkers can be difficult to interpret because
 5      they may not be diagnostic of specific contaminants, and they can be modified by extraneous factors, such as
 .6      temperature and season.  For these reasons, care must be taken in using tissue residue and biomarker  .
 7      information as a primary source of exposure estimates. These measurements, however, can provide valuable
 8      confirmatory information that exposure has occurred.
 9           Irrespective of the approach used to estimate exposure, the characteristics of the ecosystem and exposed
10'     organisms need to be considered to make appropriate conclusions about exposure:                 !'
11           Are there abiotic factors that will influence the degree of contact? Ecosystem attributes may increase
12      or decrease the amount of chemical contacted by receptors: For example, the presence of anoxic areas above
13      contaminated sediments in an estuary may reduce the amount of time that bottom-feeding fish spend in
.14      contact with the contaminated sediments, and-thereby reduce exposure.
15           Are there biotic factors that will influence the degree of contact?  Community-level interactions can
16      also influence exposure. For example, if several organisms compete for the same contaminated resource, it
17      may reduce exposure of a particular individual. Alternatively, competition for high quality resources may
                '      >        •      '               '  •  ' • '         -
18      force some organisms to utilize contaminated areas.
19           Will the behavior of receptors influence initial or subsequent exposures? The interaction between
20   .   exposure and receptor behavior can influence both the initial and subsequent exposures. .For example, some
21      chemicals reduce the prey's ability to escape predators, and thereby may increase predator exposure to the
22      chemical.  Alternatively, organisms may avoid areas, food, or water with contamination they can detect.
23   "  While avoidance can reduce exposure, it may have additional ramifications by altering habitat usage or other
24      behavior.                                 ,
25                           :   •                                              -      •     '  .
26      Exposure Dimensions                                             '  '. .     "              -
27           Three dimensions must be considered when estimating chemical exposure: intensity, time (duration,
28      frequency, and timing), and space. Intensity,  the most familiar dimension, is a function of the concentration
29      of a chemical in an environmental medium and a contact rate. In its simplest form, intensity is quantified as a
30      medium concentration (e.g., concentrations in water, soil, or food), with the assumptions that the
                                                         71
10/13/95

-------
-,:: iii. •   IIBSI ! ,>,,	, •';,;	-; i"1 .••;.;, us <\ »<[ - ,v is: 'i i; j	;i;r 'iW'f«!»t;!iiH&	i;n	; iris	:»	f i,	IB
                   - • I"' ;,>•'. "  • v - , , - -	i-  •,,,, ;.'i»	l-i ," i'	«'li:"li;:i;is .;t !'* • '<«. IUIIM mt, ,< '.• ,<• '"f""«' -si ViJKiiK	r "iiuS, .>,	:, i	' IIFilt s "i1.;	n	illliw^^


                   DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                                          UPTAKE VIA SILLS
                                                         LOSS VIA GILLS      LOSS BY METABOLISM
                                            UPTAKE FROM FOOD
                                                                  LOSS BY EOESTION
                                                                                 GROWTH DILUTION
 1      contaminants are well mixed and that the organism contacts a representative concentration.  This approach is
 2      commonly used for respired media (e.g., water for aquatic organisms!, air for terrestrial organisms).  For
 3      ingested media (e.g., food, soil), another common approach combines modeled or measured concentrations of
 4      a contaminant with assumptions or parameters describing the contact rate (U.S. EPA, 1993c). If the
 5      estimation of^an internal dose is desired, toxicokjnetic .approaches may be useful (figure 4-5).
 6           The temporal dimension of exposure has
 7      aspects of duration, frequency, and timing.
 8      Duration can be expressed as the time over
 9      which exposure occurs, exceeds some threshold
10      intensity, or over which intensity is integrated.
11      If exposure occurs as repeated, discrete events
12      without significant variation in duration (e.g.,
13      discrete chemical spills), the frequency of
14      recurrence is the important temporal dimension
15      of exposure.  If the repeated events have                      	
i/:      „•„„'«   t   A   -MA   *•     n,    f, ft.       Figure 4-5. Mechanisms of chemical uptake and loss
16      significant and variable durations, then  both       rr-u/j   * * e     /•*  i  ^.i  *nnn\
          e                       ,                     for fish (adapted from Clark etal., 1990)
17      types of temporal dimensions must be
18      considered. In cases where the timing of an exposure influences the extent and magnitude of effects (e.g.,
19      influx of hydrogen ions and aluminum during snow melt), this factor should be described in addition to the
20      duration and frequency.                                   •
21           The duration oyer which intensity  is
22      integrated is determined by considering both the
23      ecological effects of concern and the likely pattern
24      of exposure.  In general, the statistic of interest is
25      the total (sum or integral) intensity of exposure
26      over the event or time period of lexicological
27      significance (O.S. EPA, 1992dj.  For example,
28      shorter times are  used for acute and developmental effects; longer times for chronic effects.
    1 •           ' ,|n    '„',' 'j!!!I1   . '    '"    . ,.!"	 • •
29           Several simplifications are commonly employed to make exposure calculations easier. When the contact
30      rate for each exposure event does not vary substantially within the time period of interest (say, the daily
                                           Text Box 4-S. New Chemical Example:
                                           Estimation of Exposure
                                                   Exposure was estimated for the new
                                           chemical by first estimating the average chemical
                                           concentration over one day of the chemical in
                                           streams. Aquatic organisms were assumed to
                                           contact this concentration.
                                          72
                                                                                                    10/13/95

-------
                                DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      ingestion of a prey item over six months), total exposure Can be calculated by multiplying the temporally
 2      averaged contact rate by the temporally averaged medium concentration by the total number of exposure
 3      events. However, as concentrations or contact rates become more episodic or variable, this simplification
 4      becomes more problematic.  In extreme cases, averaging may not be appropriate at all, and assessors may
 5      need to use a toxicodynamic model.
 6           Spatial extent is addressed in ecological risk assessments by defining the area contaminated, the area
 7      above a particular threshold, or the average concentration in a biologically-relevant area (e.g., foraging
 8  "    range). At larger spatial scales, however, the shape or arrangement of contaminated areas may be an
 9      important issue and area alone may not be the appropriate descriptor of spatial extent for risk assessment.  A
10      general solution to the problem of incorporating pattern into ecological assessments has yet to be developed;
11      this issue is normally addressed on a case-by-case basis.  .
12                                                                               *
13      4.2.2A.  Exposure Profile
14   ,        The exposure profile is the output of the characterization of exposure process (figure 4-1).  The purpose
15      of the exposure profile is to summarize the exposure analyses so that they can be best used in risk
16      characterization. If an empirical approach is used, the exposure profile may be expressed as a point estimate
17      or distribution.  If a mechanistic model is being developed, then the output of exposure characterization may
18      be an exposure module of a larger model that integrates exposure and effects. -Itrany case, by the close of the
19      exposure characterization process, the assessor should be able to present the following information;
20   ,        Describe the group represented by the exposure profile- A succinct definition" of the group described
21      by the exposure profile includes its level of biological organization and spatial and temporal extent. For
22      example, the exposure profile may represent the local population of piscivores feeding on a specific lake
23      during the summer months.                                                           .
24           Summarize the most important exposure pathways,                      •                .
2*5           Summarize the three dimensions of exposure: As discussed above, for the results of an exposure to be
26      useful, they must be commensurate with the stressor-response relationship generated by the effects
27      characterization. The assessor should state how each of the three general dimensions of exposure (intensity,
28      time, and space) was treated and why that treatment is necessary or appropriate. Continuing with the
29      piscivore example, intensity might be-expressed as a distribution of potential doses in piscivores feeding on
30      the lake averaged over the summer months.
                                                        73
10/13/95

-------
                                  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  I            Describe the uncertainty associated with the exposure estimates: In the exposure profile important
  2       uncertainties are summarized (see section 4.1.2 for a discussion of the different sources of uncertainty). In
  3       particular, the assessor should:
  4       •    identify key assumptions and describe how they were handled;
  5       •    distinguish between variability and measurement and systematic uncertainty;
  6       •    identify the most sensitive variables influencing exposure;
  7       •    identify which uncertainties can be reduced through the collection of more data.
  8            Summarize the methods used to perform the analyses: Include any statistical and modeling techniques.
  9      "               '            '            	      •	,	•'•"'  ;  •;   "' •;.	;	'	'"^
10       4.2.3.  Ecological Effects Characterization
11            There are two activities associated with characterizing ecological effects  (see figure 4-1).  The first
12       activity is ecological response analysis, and the second is the preparation of a stressor-response profile.
13       Ecological response analysis is essentially a "number crunching" activity. The stressor-response profile
               • . "     i,i!:ii	•    •• .     ,r  >"' • ;  	.:  ,  u .• ',';•„.  '	i.''",>, .:•.	i,:1!.11..111;;"!,1"!	i'!'1;1!;:'M-iiiiH	nisii:: KM - jtiirjMT'i-i"**1	-•> :\»r ir,	ii-	fit.-i-Tis: "i'M	]<	IKVIM .>• nupr	  I
14       summarizes and discusses the results of the ecological response analyses and serves as input to risk
15       characterization.
          '.   " "   •     iSi'f;  i ,•. . ,    ,  i": 	"    ' '"  . ;| •       I          '      n I I I I      II           II     'Mill  111 I,
16            During ecological response analysis, the assessor quantifies, to the extent possible, the relationship
            v   *      , *	!,:,:•'    	  „   T   ' "„  :	  N,,                         I |||         1                     ^
17       between the types and magnitudes of effects elicited by a chemical and exposure to the chemical.  A critical
18       part of ecological response analysis is relating measures  of effect to the assessment endpoint(s) identified
                	   :i"'i!H!.,i  :      ' ,   .|«.'  I1'., '".  ' 	• "  ' i M1 '„ ' , •; f Vl .|«  ' i.!i         I    I II  I
19       during problem formulation. An analysis of cause and effect relationships is also conducted when necessary.
20       The latter activity is particularly important when the risk assessment is being driven by the observation of
                      "v""i .  '     "  .  l4  "• ,:   •  ••  "i '!•" i  ', 	'	 •': .,' •„'';„;,:::!, cil	ii!	Litiih	till	ic'iili-'. •	«itMi.r<	i 	'ji	ill!	i jwivMitt'1	f-:>	ii'.s!	I'*	lii^ill.tillliiiilitiii'Ill!	    I
21       adverse ecological effects.  Examples include bird or fish kills and a decline or shift in the flora or fauna of a
22       given area (e.g., Florida Everglades, hazardous waste sites).  Input for the ecological response analysis
23  •     includes relevant effects data gathered during problem formulation and ancillary activities, i.e., data
24       acquisition,  verification, and monitoring.
25            Upon completion of the ecological response analysis, the  assessor prepares a stressor-response profile.
26       The profile summarizes the ecological response analysis and discusses associated methods, assumptions, and
27       uncertainties. Uncertainties can include such factors as the accuracy and relevance of the data for the purpose
28       at hand, the  absence of information deemed important to the analysis, and how well the methodologies
29       employed address the effects of a given stressor. It is important that the stressor-response profile summarize
30       the data in a form that is compatible with the risk characterization method chosen in problem formulation.
                                                           74
10/13/95

-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE     .

 1      For risk assessments that are effects-driven, it is important for the assessor to clearly and objectively discuss
 2     . the uncertainties associated with the evidence or data for cause.-and-effect relationships.
 3           The discussion of ecological effects characterization is subdivided into five sections.  Section 4.2.3.1-
 4      discusses the sources and types of data analyses that may be used to address primary effects, also known as
 5      direct toxic effects, to individuals and populations.  It does not cover toxicological principles or how to
 6      conduct tests. Section 4.2.3.2 addresses the use of extrapolations and other methods (such as models) to
 .7      relate measures of effect with the assessment endpoint(s).  In some risk assessments, the assessment endpoint
 8      and the measures of effect are the same, but more often they are not and some type of extrapolation or other
 9      approach is necessary to relate the two.  Section 4-.2.3.3 discusses (indirect) secondary effects.  For the
10 .     purposes of discussing chemicals, secondary effects can be thought of as effects to one or more assessment
11      endpoints induced by toxic effects to other trophic levels or organisms-. These other trophic levels or
12      organisms serve as food, habitat, or even regulate the assessment endpoint itself (e.g., predator/prey
13      relationships).  The mussel-fish connection presented in text box 3-8 is an example of a secondary-effect. To
14      complicate matters, many chemicals can cause both direct and secondary effects.  Section 4.2.3.4 discusses
15      the various approaches in dealing with causality and is particularly important for assessors confronted with
16      effects-driven risk assessments.  The final section, 4.2.3.5, offers general guidance preparing a stressor-
17      response profile,    .        .  •       •
18           To supplement or illustrate a particular principle or application, case illustrations of assessments (U.S.
19      EPA, 1993a, 1994a) are included as text boxes or discussed in the text itself.  These cases are intended only
20      as adjunct material  and are not examples to be followed in a rote manner. One example, the new chemical
21      case study (Appendix A, Case A-5), has been selected to illustrate how risk assessments are conducted on a
22      predictive basis for chemicals. Figure 4-3 is a flow chart of the overall assessment and, where appropriate, it
23      will be referred to in the aforementioned sections.             .    ,   "
24    :              :   ,           '   •     .--     •  .               ^       ^      "   •      .''••-..--.
25\    4.2.3.L  Estimating.Primary Effects                            .                .      •-
26            Chemicals can cause a wide variety of biochemical or physiological disruptions that can range from
27      enzyme inhibition'and inactivation to effects on membranes and other cell components.  The results of these
28      effects can vary from mutagenicity to growth and reproductive impairment and death. For the most part, the
29      characterization of ecological effects for chemicals has concentrated on evaluating effects that are readily
30      observed. These effects include mortality, growth, and reproduction and are measured via selected test
                                                          75
10/13/95

-------
                  •. I , v ' I'li'iil ' < !':• ........ .i ..... P., . "'5' :
                                   :!:!
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE

 1      as to how the magnitude of the toxic effect varies with incremental changes of exposure to the chemical,
 2      Stressor-response information is important not only for conducting a more complete and credible risk
 3   -   characterization, but also for providing information on the effectiveness of potential risk mitigation options,
 4      For example, it may be unclear as to what effect decreasing an exposure concentration by some amount will
 5      have on the assessment endpoint if only a single point estimate is available. A Stressor-response Curve
 6      enables' the assessor to quantitatively evaluate the consequences of incremental reductions in exposure,
 7           Four sources of information or data that may be utilized by an assessor are presented below. The choice
 8      of one or more of these sources will largely depend upon the scope of the assessment. In addition to being
 9      used for characterizing ecological effects, the methods may also prove useful in problem'formulation. For
10      example, structure-activity relationships can be used as a preliminary screen to prioritize or rank chemicals
11      for additional analyses as well as for developing a Stressor-response profile,                      '
12   „        Structure-Activity Relationships, When little or'no toxicity data are available for a chemical,
13      structure-activity relationships (SAR) aid in estimating ecological effects (Auef et al., 1990; Clements et al.»
14      1988; Clements and Nabholz, 1994; U.S. EPA,  1995e; Nabholz et al.,  1993a,b)  The simplest application of
15      SAR is to identify a suitable analog for which data are available and use the data to estimate the toxicity of
16   .   the compound for which data are lacking; More advanced applications involve the use Of quantitative
17      structure-activity relationships (QSAR). QSAR is a quantitative relationship between chemical strUCtufe and
18      a specific biological effect and is derived using information on a series  of related chemicals (Auer et al. 1990;
19      Aueretal., 1993).
20           There are important limitations associated with the use of SAR/QSAR, Clearly, the selection of
                                                                                              __        <
21      appropriate SAR/QSAR for a given chemical is critical. For example,  some SAR/QSAR approaches used to
22      predict toxicity to aquatic organisms are based solely on chemical classes. Others attempt to classify
23      compounds based on common modes of toxic action and can either combine or subdivide classical classes
24 -    (Bradbury, 1994).  Whichever approach is used, a detailed explanation of the SAR/QSAR selection process
25      should be provided to  impart an understanding of the model structure uncertainty associated with the
26      prediction. Typically, QSARs reported in the literature provide statistical information concerning the
27      standard error of estimate and variability associated with slopes and intercepts.  These additional data are
28      important for evaluating reported regressions. Because QSARs yield single point estimates (e.g., an.LCso)
29      they do not provide information regarding the slope of the Stressor-response Curve,
                                                        77
10/13/95
-------
                                   DRAFT--DO NOT QUOTE,, CITE, OR EJISTRIBUTE
  I           SAR/QSAR can be useful in problem formulation as well for characterizing ecological effects and can
  2      be used to ranJc or prioritize chemicals for further assessment. Also  SAR/QSAR can help estimate what
  3      effects a chemical may elicit, thereby providing insight into the kinds of tests that may be needed, if testing is
       "          •     ", "HI  '   I   .',;/              I  I    II _  I    ll l| II IN II Hi II 111 III
  4      required to complete an assessment.  In the new chemical case example, QSAR. was used initially to ascertain
  5     ' what effects were likely. Based upon the chemical's high octanol-water partition coefficient, assessors
  6      concluded that short-term effects (mortality) were not likely because the chemical would be taken up in
  7      insufficient amounts to elicit effects over a short term period ( 96 hrs.).
  8           Single-species assays. Toxicity testing with single species is the most common method for evaluating
  9      the toxic effects of chemicals to terrestrial and aquatic animals and plants. These tests, which use standard or
  11 ,„               • i   fi.|iiii|i:| 	  	 , "r "  i»,,  ,11- ,»  '., " „	   	   "• *•                •*•
 10      surrogate species, tend to be cost-effective because they are typically used in a tiered fashion.  That is, short-
 11      term tests are conducted first.  These tests are designed to evaluate effects such as lethality and immobility. If
      !            „,'''' '  " 'ijii ii  ''>  '...,.;        |                   ^      i  i i|i j | |i |i|  ill nil i ii 11 i  MI   ii  n  in ii in i i inn MI   ii i ill in ii • IIP • inn i iiinin
 12      the chemical exhibits high toxicity or a preliminary risk characterization indicates a risk, then more
 13      expensive, longer term tests can be conducted. Longer term tests tend to measure sublethal effects such as
 14   '   effects on growth and reproduction. The tests can include part or all of an organism's life cycle. For instance,
 15      a fish early life stage test measures toxic effects to the fertilized egg and the fry.  A full life cycle test follows
 16      fertilized eggs through the fry, juvenile, and adult stages, including egg laying by the latter. The tests can also
 17      proceed from single-species tests to microcosms (q.v.) and field studies (q.v.).
                  i  . 	i'!1 in  „, ""   "                                                         «              1    n
 18           Typically, short-term test results use statistical estimation techniques to estimate a median effect level
111     "                 , •"	Hi i      •      %
 19      such as LCjo or LD50 if lethality is the endpoint of concern. Short-term effects other than lethality (e.g.,
 20      effects on growth) are expressed as an EC50 or ED50.  Stephan (1977) discusses several statistical methods to
      1 ' '   '  "'"      '  'frtllH i' !'' . •   ' • if, ' :   ! T '' '• hi  "'i i ""1 * ' 'I'                           II
 21      estimate LCsoS. Median effect concentrations or doses are commonly used because they provide a consistent
 22      approach for ranking and comparing the toxicity of a  wide array of chemicals.  For characterizing both
 23      ecological effects and risk, concentrations eliciting other than a median effect can be valuable.  For example,
 24      the use of a concentration which elicits effects to 5 percent of the exposed population has been recommended
 25      as a benchmark for evaluating the lethality of pesticides to aquatic organisms under short periods of exposure
 26      (SETAC, 1994a).
 27           The most commonly used approach for analyzing test data from longer-term or chronic tests is
 28      hypothesis testing (not to be confused with risk hypotheses, which are discussed in section 3.5.1).
 29      Hypothesis testing involves  setting up a null hypothesis that usually assumes there are no differences in the
 30      responses in the dosed and undosed test animals. The test results are generally described in terms of a no-
                                                           78
10/13/95
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1       observed-adverse-effect-level (NQAEL) and a lowest-observed-adverse-effect level (LOAEL). The NOAEL
  2  -.    is the highest dose or concentration for which no statistically significant effects,~compared to control
  3       organisms, were observed.  The LOAEL is the lowest concentration or dose for which a statistical difference
  4       was noted. The range between the NOAEL and.the LOAEL is often referred to as the maximum acceptable
  5       toxicant concentration (MATC). The MATC also can be expressed as the geometric mean of the, LOAEL
  6       and the NOAEL. The geometric MATC is also known as the chronic value because it is obtained from longer
  7       term (chronic) tests (Stephanetal, 1985).          ..-.'.-
  8    _  -    Hypothesis testing has been criticized for several reasons.  For example, statistically significant effects
  9       do not necessarily correspond to biologically significant changes, and poor testing procedures can increase
 10       test variability, thereby reducing test sensitivity to toxic effects (Stephan and Rogers, 1985). Some
 11       investigators (Stephan and Rogers, 1985; Suter, 1993a) have proposed using regression analysis as an
 12     ™ alternative approach to hypothesis testing. Regression analyses enable, an assessor to estimate effects over a
 13       wide range of exposures. In situations where it is desirable to estimate or protect at a certain level of
 14       exposure where effects are minimal (e.g., 5%) regression analyses would likely be -the preferred method.
 15       However, some toxicity test data that are amenable to hypothesis testing may not support regression analysis.
 16           hi the new chemical example (figure 4-3), single species assays were used to both confirm the QSAR
 17       predictions and to evaluate the long-term effects of the chemical on the survival, growth, and reproduction of
 18       surrogate aquatic species.                                 .              •"      . '
 19           Multispecies assays. In contrast to tests conducted with a single species or axenic culture, multispecies
 20       tests involve two or more species in the same test vessel. Most often the species represent different trophic
 21       levels and as such are intended to evaluate community level effects.  In addition, multispecies tests have been
 22       proposed for many purposes such as  the evaluation of bioengineered organisms and secondary effects.
 23       Multispecies assays can range from small-scale (battery jar) laboratory aquatic or terrestrial tests to larger
 24       indoor or outdoor tests such as mesocosms. The use of microcosms and mesocosms is discussed in section
 25      -4.2.3.3.  '--.''.                    "      "-••..
 26           Field experiments. Field experiments and data provide the assessor with a degree of realism that
 27       usually cannot be obtained under controlled laboratory conditions.  This category of data includes field
 28       experiments, observations studies, and laboratory testing with media collected from the field. Field •
,29       experiments can be used in various ways. For example, they can be used to confirm (or refute) effects
 30       observed under laboratory conditions. In hazardous waste sites, field studies are often an integral part of
                                                         79
10/13/95
-------
                                                            	I" 11 PI I III II ill "I!
                                                                          111(11(11111 (ill1
                                                                                      iiiiiiiiiip mi	i illinium i mi 11 in in iiiiiiiiiiiiiiiiiiiiiii
                                                                                                 i
                                  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1       ecological risk assessments. They are frequently used at these sites because of lack of data on bioavailability,
  2       bioaccumulation, subtle ecological effects, and toxicity due to multiple pathways of exposure (soil, water,
  3       air), Field data are most useful when the assessment endpoint is;a.t the community or higher level or
  4       organization, when multiple stressors are present (section 4,5), or when factors influencing bioavailabilty are
  5       uncertain,, The. design of these experiments or observations, including the selection of appropriate reference
  6       sites, is extremely important. See reviews by Suter (1993a) regarding; the problems associated with
  7       pseudoreplication and Peterman (1990) regarding the importance of statistical power in accepting or rejecting
  8       null hypothcsesT Applicationof field studies to sedmient rontarm^ation and biological surveys in streams are
  9       described in Cirpne and Pastprak (i993) and U.S. EPA (1989a), respectively,
 10     '"             '"  ^   '.''  ^	""	'"'	'r.'".i'Z"!Tr""i	
 11       4,2,3,2.  Extrapolations
 12            Often, actual data on an assessment endpoint are lacking or incomplete. For example, the assessment
 13       endpoint might be survival and reproduction of estuarine fish species but the only data available are for
 14       freshwater species. Pragmatic constraints dictate that not every spectes can be tested,  An obvious fact is the
 15       very large number of potential test species present in any system.  In addition, many species are too large or
              .  » r t'   f«;|wl  <   '  ' ;; J1,!1 |  Ifj'iV/	'" "	Jl >!l ^iffl11111'*';1]*1'11, /|l!"l|;"i"l|i':il'|l"|;1'111: [r^Cf!'!^^                       I III I 111   111 III 1N1  I III llllllll Illllll I Illlllllllllill  111
 16       difficult to maintain under laboratory conditions,  OtheVspecies'"are protecte36y state and federal statutes and
 17       require permits for their collection and  possession. Because of the constraints, ecotoxicologists have selected
 18       species which can be reared or readily cultured, are relatively sensitive to toxicant, and are representative of
 19       naturally occurring species, genera, and! families"  Thus, me assessor may•'be confronted witn test clata on an
	 .    -'  '     '    '  	"imiiH  , I :i,' '' r "..v1"1 l;l11 i'i •" I'i'ij'l : 'Sji':'. ;,.j''.M>i&J'j' ^     •:ii!i:><>llil                                   •
 20       organism other than the species of concern. Furthermore, the test data may be incomplete,  In the previous
         .         •'; '  . /mil : ".'I „  ; .:,-, ,	,;-, i", (i;!;'':'	.; 'it,,;>;'' I.!"':''::';!'.)!'!!!1!;;;':!1!;1'  i'V Ib^iJ'ffi'iitll1^                                             ,
 21       example, the assessor may need to evaluate lethal and sublethal effects of a toxicant to the estuarine fish but
 22       only has toxicity data on lethal effects.
 23            In cases where estimates have to be made regarding effects on an assessment endpoint and the
 24       appropriate data are not available, extrapolations may be necessary, Depending cm the particular taxoh
 25       involved, the data bases may be extensive or scant.  For instance, while extensive multiple species data bases
 26       are available for effects, of toxicants on fish, comparable data bases for mammals, amphibians, or reptiles are
 27       virtually nonexistent, Extrapolations require  both common sense and professional judgment, Obviously, one
 28       would not extrapolate toxic effects on paramecia to effects on whales. However, there are many instances
 29       where extrapolations can be made in a credible manner, Methods ancl approaches are discussed by
 30       Bamthouse et a^
                                                           80
10/13/95
                                         i;	;• iii". jii'iiif ll|liii|!:iiii;n;Hi||:i;'iiliiiinit	fta: Ji
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1
 2
 3
 ,4
. 5
 6
 7
 8
 9
10
11  '
12
13
14
15
16
17
18
19
20
21
22
23
24-
25
26
27
28
29
30
     Extrapolations Between ^Responses.  Risk assessors often face situations where there are data only on-
the short-term effects of a chemical (e.g., lethality) and questions arise regarding longer-term or sub-lethal
effects of a certain stressor. Clearly, the best approach is testing, but this is not always possible.  The
alternative to testing is to estimate sublethaleffects from lethal effects via extrapolation. For aquatic
organisms, two of the oldest methods for extrapolating from lethal to sublethal effects are the application
factor (Mount, 1977) and the acute-to-chronic ratio described by Kenaga (1982). The application factor is
derived by dividing the maximum acceptable toxicant concentration by an acute value, such as an LC50, for
the same chemical. This factor is then used to calculate the MATC for another chemical.  The acute-to-
chronic ratio does the reverse, the acute value is divided by the chronic value, and the ratio is used to calculate
an MATC, Although developed for aquatic organisms, the approach could be used for terrestrial organisms
as well;
     Mayer et al. (1994) noted the above methods have limitations because the ratios actually represent.
different responses.  Typically, acute tests measure lethality, whereas the chronic tests used in the MATC
calculation measure sublethal effects. Mayer proposed the use of regression analyses to estimate chronic no-
effect levels for lethality rather than application factors or acute-to-chronic ratios.  This method requires
stressor-response data for 24-, 48-, 72-, and 9.6-hour test durations.
     Empirically derived uncertainty factors have been used to extrapolate from lethal to sublethal effects,
species to species sensitivity, and laboratory to field effects. These factors range from 10 to 100. Use of
these factors is contingent on the assessor's knowledge about the chemical and the class to which it belongs, '
     In addition to extrapolating from one toxic
                                                Text Box 4-7. Methods for Extrapolating Effects
                                                from Individuals to Populations
                                                •      Population Models (e.g., Leslie Matrix }
                                                •      Individual-based population models
                                                •  •    Life Tables        '          .
effect to another, extrapolations can be made
from individuals to populations (see text box 4-
7).  Population models have been used
extensively both in ecology and fisheries
management. Excellent reviews are presented
by Barnthouse et, al. (1986), Barnthouse (1993), and Wiegert and Bartell (1994).  Population models have
also been used to assess the impacts of power plants and toxicants on specific fish populations (Barnthouse et
al,, 1987; Barnthouse et al., 1990).  Population models are useful in answering questions related to the
probability that a certain population will fall below a, specified number and/or the probabilities of various
percent reductions in a population.  Proper use of the models requires a thorough understanding of the natural
                                                         81
                                                                                            10/13/95
-------
 Ti ;   |"!,"f'! i1 • ii1""11 ,  • ""' wi' 'I. "'I!' •«	i, i T ' ' , '' fllllll1!1' ' i; ,*'"' r, Hi,:''!!'.', •* W"' -nil!:,' '• 'fi. P w: i";; Illlil1' "iM! • 1' • "*!;'!: 'li!1'"1 T »,:• I* I!	• »!l *i	PS! ^ "fVT», iiilP! Li* Wi' ilHn^illlUilllillllilllllllllllilllllillllllll
     ; ,.,.  „     ; •  -  •1;i., •  i:j:|j .; *• . ji • • v v'  :*j  .^v,j;; ,;'i ;:;|;; X™ ;* fi^v' '-|j^Jli'flfyl^                                     i.	,,,,,„....,,.,.
                      „ I:™; '!i||i j  • ,/'  , i',' '„';,' , ! ',  j. ,•' ;i	., ' '!»i n ;!, •»'';''; •!:'!]!:["!»; jj1,!; '!,;  i; "i,,,;'•. '!• ]• '•",,;, .,•:.,"''"; .ijii if 'i' :iji,'ji ]!i: ill1!'!'" jj!1!1 i|! i] jjjif I1 j1 |;,i|! iiiiniiiiiiw!! •!' ii!!!' ii'n1' wiiiiii i j, ii' i!. li JI Ji ,il» 'iiii'ii1 nwiil1 liiiiH i1'! .ii"lu 'iiw',.' 'ilil'!', niii! ''w f I „' siii:" iii'iin1 iiiiiiiiiiiiiiiiiiiiiiiiii'iiiiijiiaiiiiiniiii ii'iiii'liiiiiiiiiiiijiiiuliiiiiiii,

                                                         •	': ;,„„ *.!	ii: -'? 'r^i'.';;!	; M > \i • Pttwali^	l^'i i"" f I!1' :i;;': i;i|i!:;I,]'.' '••>*! ifl3',', jv ^5»I:':]1!1" i;|l'::^!!!'K'^         I
                      .,;.- ,: i  • .;,;," DRAFT-DONOT QUOTE, CITE, 6RlSlSTRIBUTE "_	/_"	" ';_";'	'""""	™"	~~|

         history of the species under consideration.  Parameters include longevity, age to sexual maturity, fecundity,         I
    "''.'  ',  • •  , •    ,.'   "53I , !•.'. '.:• .•>:• ""'>'  [f:-I:': "Vi'i!:"'"' •*;«,*:i;1"-!••'.'Kif'H;i'1'!-Svs»is^
 2       and percent survival among the various age classes. In addition, the fype of density-dependent function is           I
 •        ! <    ' '  "!!j'   iiyi'i'! •"> '•''•' >'	" '	;:< "i"1"'1''1'":'' •	'-'	I!:-' •'' :'ir" .^''li'iiiiiii'"111';'i'!1:!)!'1!               .                    '            	iiitM            I
 3       also important (Person et al.,  1989).  Although a model cannot distinguish between good and bad data, the          I
 4       assessor must be able to supply the correct data or the results can be very misleading or inaccurate, ^tameid'       I
 5       and Bleloch (091) describe the  formulation and use of models for m anaging wildlife ancf proviSe real-life          I
 6       examples obtained from wildlife parks in Africa. Population models are a useful tool, and their use in              I
 7       quantifying risks to assessment endpoints is encouraged.                                                         I
 8            Individual-based models are relatively new approaches to assessing risks to populations.  This approach       I
 9,      simulates the behavior of individuals to a particular stressor, then calculates the net effect on the overall            I
10       group. EPA has not used individual-based models in  a regulatory context.  Hallametal. (1990) developed         I
11       an individual-based model for Daphnia magna and studied the effects of neutral organic compounds  on   "         I
12       survival  of the copepods, but  the model requires a sophisticated mainframe computer and does not have a           I
13       user-friendly interface.          '      ,                                                                         I
14            The use of life tables to calculate an intrinsic rate of population increase can also provide meaningful          I
15       data regarding the effects of a toxicant to an assessment endpoint. Gentile et al. (1982) studied the effects of       I
16       nickel and cadmium on mysid shrimp (Mysidopsis bdhid) survival, sexual maturity, time to first Brood             I
17       release, brood duration, and total juveniles produced.  The data was then used to calculate an intrinsic rate of        I
18   i    ' growth.   '   .'"'   '           """'   "'  ' "  	:   	   '     "	"	'	"	"	'	''"""	,"	'	'	"	          "      I
19            Extrapolations Between Taxa. Many times, the assessor may be confronted with a situation where the ,     I
20       risk assessment is focused on a certain geographic area, but the effects assessment is based on data Tor             I
21       surrogates that do not occur or are not representative of the species iii that area. For instance, a risk                I
22       assessment may be addressing risk to estuarine fish species, but the only effects data available are for              I
23       freshwater fish. Or the risk assessment may assess several geographic areas or species, but the effects data         I
24       are limited to one  or two species. In lieu of actual testing, the only practical approach is to attempt to
25       extrapolate from one taxon to another. The taxa can range from species, genera, families, and perhaps orders.
26            Suteretal. (1983), Suter (1993a), and Barnthouseet al. (1987,	1990J	hayeJeyeippedmethpHs to
27       extrapolate toxicity among  freshwater and marine fish and arthropods. As noted by Suter (1993a), the
     , '      ')  '     '   i:;/	11  .  ,	::<: :::       .         '         	iiii>i><:>:>M,i:u^^^^^^^^   	ta:
28       uncertainties associated with extrapolating between orders, classes, and phyla tend to be very high. However,
29       extrapolations  can be made with fair certainty between aquatic species within genera and genera within
                    '  »"•;,!   '   •  .    '"  !•• ".   » ''''•,-           Ml  I    II   III Ihli III
30       families.
            1  '    '    >"!v;:j  .       •'   ,;  : ,: ' ••   •:                '    .          11 ii 11 hi        	|       	1 hi ill I l|l (Win II ii 111
                                                             82                                        ,  10/13/95
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE _
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
     Dose-scaling has also been used to extrapolate the effects of a toxicant to several species and is used for
human health risk assessment but has not been applied extensively to ecological effects (Suter, 1993a). The
most common unit used with human health is weight, expressed as milligrams toxicant per kilogram of body
weight. This has been used with avian species (Kenaga, 1973). Allometric approaches have also been used to
extrapolate between species. Allometry has been used in two main ways: to relate the growth and size of one
body part to the growth and size of the whole organism (U.S. EPA, 1995e) and to relate body size with other
biological parameters. These regressions can then be: used for species for which data are not available (Suter,
1993a; U.S. EPA, 1993). Allometric  scaling according to body mass is based on-broad differences in
physiology between species that can mostly be related to general toxicokinetic differences that occur between
species, owing to the many complexities of toxicokinetic and toxicodynamic processes thaHnfluence an
organism's response to a toxicant (U.S. EPA, 1995e).
     Allometric regression has been used to a limited extent for estimating effects to marine organisms based
on their length.  It is important that the regression equations be confined to a specific taxonomic group.
Thus, allometric equations developed  for avian species may not be applicable to amphibians.  As with any   •
other regression analysis, the assessor must carefully consider whether the variables being regressed (e.g.,
length or metabolic rate) represent biological reality.  Could other factors such as morphology or
physiological activity be more important than the parameter being regressed?  Therefore, allometry
adjustments should be used only as one part in the overall process of estimating interspecies difference's.
     Extrapolating from Laboratory to Field Effects. A frequently asked question is, how well do effects
observed in the laboratory predict effects under natural conditions? As a general rule, most toxicity tests are
designed to prevent any masking or mitigation of the toxic effects of a compound. Thus, exposure tends to be
maximized.  It can be reasonably argued that if a similar exposure profile exists under field conditions, then
one can expect the same effects with regard to type and magnitude.. Often, however, there are mitigating
conditions existing under field conditions that reduce toxicity. The duration of exposure might be different,
the organisms may be able to move to different areas, and.biotic and abiotic factors may contribute to
lowering the duration of exposure.  Absent data to the contrary, however, a reasonable (and experimentally
proven)-assumption is that laboratory  effects do represent field effects when the exposure profiles are similar.
     Extrapolating From One Geographic Area to Another:  Frequently, the assessor may be confronted
with useful  data that are germane to the assessment endpoint of interest, but the data are obtained from a
different geographical area.  For instance, as assessment may be centered around the effects of an aquatic
                                                         83
                                                                                          10/13/95
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE

 1      herbicide in a northeast watershed but the data were obtained from a midwestern site.  In extrapolating such
 2      effepts, a default assumption could be that the effects should be similar qualitatively but not necessarily
 3      quantitatively Principal factors to be considered when extrapolating from one area to another are differences
 4      in environmental considerations or forcing functions (e.g., light and temperature) and spatial conditions
 5      (Bedford and Preston, 1988). Obviously, extrapolations from one geographic to another require coordination
 6      between the exposure and effects analyses.
 7                        .       -             •                                          ,   , ;   ,    '.    , •;
 8      4.2.3.3.  Secondary Effects
 9           Often, secondary (indirect) effects to assessment endpoints are more subtle and difficult to detect than
                 	         ,    „       „   	   	     i                       : 	;	
10      direct effects.  They also are sometimes the most difficult to quantify and relate to a risk manager. Methods
11      that quantify secondary effects to assessment endpoints are more desirable than narrative approaches, which
12      are often difficult to follow (Rodier and Mauriello, 1993).
13           Artificial ecosystems can be used to measure or quantify the response to disturbances or stressors.
14      Wiegert and Bartell (1994) recognize two main types: " 'cosms" (microcosms, mesocosms, etc.) and field
15      experiments.  Cosms are the most common form of physical model. Cosms can vary in size  and complexity
16      from small flask or battery jar types (Taub and Reed, 1982) large scale mesocosms. A variation to cosms is
17      the littoral enclosure approach (Siefert et al., 1989) and microcosms that are contained in the laboratory but
'"    .if  ', •'        ,    .''iii'i, ',  i  , ;	, i«i ' i1' i, • fir „'•'„,:.,,              i    i    111  i i  ii inn in nn n n   n  11  < nn i n in i li i  in n in  i i n i in n n in I in 11 inn i in
18      governed by natural ambient conditions (Perez et al., undated). Cosms and large-scale field experiments have
19      been used to characterize effects and exposure and have also  been used to evaluate risks predicted by
20      empirical approaches and process models (SETAC, 1994b).  Physical models offer a higher scale of realism
21      than do single species tests.  As the cosm increases in scale and complexity, however, variability among the
22      control and treated replicates can mask certain effects. Cost and time is certainly a factor when contemplating
23      the use of these systems.
24           Ecosystem models (Bartell etal., 1992) and microcosm models (Swartzman and Rose, 1984;
25      Swartzman and Kalunzy, 1987) offer a useful  method for evaluating secondary effects. The assessor can use
     ,       '\	" '  n  in    . •'                    ii       i   i in i n n n   n
26      them for both risk characterization as well as developing a stressor-response profile. .Unfortunately, models
27      that assess secondary effects to terrestrial organisms are not as extensive as aquatic ecosystem models.
28       Emlen (1989) has reviewed  models that can be used for terrestrial risk assessment.
29           Many models have been developed to mimic natural ecosystems such as forests and lakes. Practical
30       applications have included evaluating nutrient stimulation in lakes (eutrophication) and climate change.
     Ill I 111 111 II III I  I 111
                                                         84
10/13/95
-------
                                , DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2  .
  3
  4
  5  •
  6
  7
  8
  9  .
10
11
12
13
14
15
16
17  -
18
19,
20
21
22
23
24
25  •
26
27
28
29
O'Neill et al. (1982) were among the first to apply ecosystem models to evaluate the risks of xenobiotics,
specifically synfuels. The method modified an existing lake model known as CLEAN to a more simple model
known as the Standard Water Column Model (SWACOM). Within the EPA,-the approach known as
Ecosystem Uncertainty Analysis has been used to evaluate the risks of chlbroparaffms (Bartell et al., 1992;
Rodier and Mauriello, 1993).  Problems encountered with this approach have been the use of assumptions in
assigning toxicity values to the various species simulated in the model and unvarying concentrations over the
one year simulation period.                                .
     Ecosystem models can be very useful in assessing secondary effects to assessment endpoints.  Expert
judgment is heeded to interface with the model and requires personnel familiar with the underlying
assumptions and components contained in the model.  Swartzman and Kalunzy (1987) and Swartzman and
Rose (1984) provide a- useful source for understanding how ecosystem models are created and evaluated.

4.2.3.4.  Causality
     The need for a vigorous analysis of causality will vary with the type of risk assessment and the need to   •
address confounding variables. Predictive risk.assessments that utilize controlled laboratory test data
typically rely on hypothesis testing, regression analysis, or other statistical techniques to infer but not
absolutely prove causality. Thus, statistically speaking, effects deemed  significant at the 0.05  level will occur
5 percent of the time  simply due to randomness; the other 95 percent of the time they will be associated with
the stressor.  Hypothesis tests reject or support a null hypothesis, and regression analysis quantifies how well
two or more variables are correlated with each other. For most predictive risk assessments these methods
suffice for establishing a credible relationship between exposure to the stressor and the types and magnitude
of the effects elicited, because confounding variables or factors are minimized. Additional studies
(mesocosms, field studies) can provide useful corroborating evidence.
     For effects-driven assessments, the assessor may be confronted with a wide array of data, some of which
may be anecdotal or incomplete. The data may originate from reference and affected sites, historical records,
and perhaps some experimental studies,  The assessor first needs to ascertain the types of stressor(s) that may
be responsible and then present evidence that they could have caused the observed effects in a plausible
manner.  There are several approaches to evaluating and presenting cause-and-effect approaches.
                                                        85
                                                                                          10/13/95
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE

 1      Ecotoxicologists have modified Koch's postulates for pathogens to provide evidence of effects caused by
 2      pollutants (Adams, 1963; Woodman and Cowling,  1987):
 3      •   The injury, dysfunction, or other putative effect of the toxicant must be regularly associated with
 4           exposure to the toxicant and any contributory  causal factors.
 5      •   Indicators of exposure to the toxicant must be found in the affected organisms.
 6      •   The toxic effects must be seen when normal organisms or communities are exposed to the toxicant
 7           under controlled conditions, and any contributory, factors should be manifested in the same way during
 8           controlled exposures.
 9      •   The same indicators of exposure and effects must be identified in the controlled exposures  as in the
10 "          field."   '	' '   ""      "   "'  '	'   	'		,  	'	'  	"	
11      In addition to the modifications to Koch's postulates, Hill's criteria (table 4-2) can also be used to present
12      evidence of causality. As noted in the Framework Report, proof of causality is not a requirement for risk
13      assessment, but for many effects-driven assessments it is often an important component.
14                                           ;  '     ^ ,   ,  ,    (   ;  •'   '	  ^   (   ^    '.'     .  •  ••„   ; ;"'	;	;
15      4.2.3.5.  Stressor-Response Profile
\ 6           The stressor-response profile is a succinct summary of the ecological response analysis and serves as
17      input both to the risk characterization as well as part of the documentation for the overall risk assessment.  A
18      useful approach in preparing the stressor-response  profile is to imagine that it will be used by someone else to
19      perform the risk characterization.  In fact, many times the risk characterization is performed by other
20      specialists.  Using this approach, the assessor may  be better able to extract the information most important to
21      the risk characterization phase. The assessor may want to review the analysis plan to be certain that all facets
22      identified earlier have been addressed to the extent possible and that the results are presented in a format that
23      is compatible with the chosen risk characterization method.  Because the types and scopes of assessments
24      vary, it is not possible to provide a comprehensive  checklist of what should be contained a particular stressor-
25      response profile. At a minimum, the information listed below should be included.
26       *    Taxa represented by the stressor-response profile. The taxa could be either specific species or more
27            general groupings (e.g., Birds). The life stages (juveniles, adults) upon which the stressor-response
28            profile was based should be reported.

                     :"5 ,    , •   , 	   .   ,,.;,  ,. ,.;:, »,.:••• ], ;• i:„-,.,,-, i,l,'i	Civ.JI •!M;'•¥ PW briCKM	/'"'jltMltajJiri.' inliMryiki^&ljt'rdiiMHtPVIK'fiinH&H |
                     ;;	j! .,.     : '•"..' !•'!'*'  ''I'TJ 'r • ]\ ''"ivt  !'.- .I'l''1'.." .''••: Li'j'iiS('!tf^i-Jii^\i^i|^i^5|i°-'!iii'''j!i'tfif^if:i? iii^'j^Ji'S^'flSi'^S^K^BES^BKS
                                                          86                                 .   .    10/13/95
                                                                                                         	 fill (	
-------
                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
Table 4-2.  Hill's (1965) Factors for Evaluating the Likelihood of Causal Association in
           Epidemiological Assessments
Strength
Consistency
Specificity
Temporality
Biological Gradient
Plausibility ,
Coherence
Experiment"
Analogy
A stronger response to a hypothesized cause is more likely to indicate true
causati'dn. This means either a severe effect or a large proportion of organisms
responding in the exposed areas native to the reference areas, and a large
increase in response per unit increase in exposure. In other words, a steep
exposure-response curve situated.low on the exposure scale. ,
A more consistent association of an effect with a hypothesized cause is more
likely to indicate true causation. Hill's discussion implies that the case for
causation is stronger if the number of instances of consistency is greater, if the
systems in which consistency is observed are diverse, and if the methods of
measurements are diverse. ._ ' .
The more specific the effect, the more likely it is to have a consistent cause.
This is equivalent to our suggestion that regular association is more readily
established if a characteristic effect is identified. Also, the more specific the
cause, the easier it is to associate it with an effect. For example, it is easier to
demonstrate that localized pollution caused an effect than that a regional
pollutant caused an effect.
A cause must always precede its effects.
The effect should increase with increasing exposure. This is the classic
requirement of toxicology that effects must be shown to increase with dose.
Given what is known about the biology, physics, and chemistry ofthe
hypothesized cause, the receiving environment, and the affected organisms, is it
plausible that the effect resulted from the cause?
Is the hypothesized relationship between the cause arid'effect consistent with the
available evidence? ~ • ' •
Changes in effects following changes in the hypothesized cause are strong
evidence of causation. Because Hill was concerned with effects on humans, he
emphasized "natural experiments" jather than the controlled exposures required
. by Koch's postulates. An example would be observations of recovery of a
receiving community following abatement of an effluent.
Is the hypothesized relationship between cause and effect similar to any well-
established cases?
Source: Suter, 1993 a.
                                           87
10/13/95
-------
                      ill
                                    I III I III
                                                        	Ill I ill II (I i|"l I'M	
                                     DRAFT-DO, NOT QUOTE, CITE, OR DISTRIBUTE
                                                                                                                       111 in i
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
•    Types of effects and a stressor-response profile.  To the extent possible and depending upon the nature
     of the chemical, both short- and long-term ecotoxicological effects on the assessment endpoint should be
     reported, including stressor-response information.
•    Causality: when appropriate discuss causal evidence.
•    Uncertainty:  examples include extrapolations, data deficiencies, and the assumptions used. The
     assessor may want to  identify areas in the analysis where uncertainty could be lowered substantially
     through the acquisition of additional information or data (see section 4.1.2).
•    Methods:  A summary and documentation of the methods used in the ecological response analysis (these
     methods may  include  statistical analyses, extrapolation methods, and models).
                                     n           	    	,	•	•  	;	/	
4.3.  ANALYSIS  OF PHYSICAL STRESSORS

4.3.1.  Introduction
   1 ' ,„ "	 ,,.i|;!i"':,!'||  « '' ,  < '  •: .  •: '	i'  i  'I'll, ,i'i! „ ' ' ••  "ii I :n'' ' i1"'..' »" V" "l" vi I1!!':,"*;1!  i:."!'1,,,"'!!!1")!	At 'l^ff^X! iliBU'Wi'livi'li" Illlliliiilliifipiill	ir^iiiWi.jiill!'1, :l|ll^lJ^l;|,l>ifijjiij  ,•!'..,  •  ;ii ,,|, „  „"i,'" |j,j,i » , • „,' ij;.  ,' '»i;',' j,,,;''ii•• I,,,;:!"!1 i'i,i,,''  i ,*ii'|h''i	, j r"'i i"1j,, ly"iji;i1 jirfii;1 vjirilji i|<4||i!', fl*"I j, N«jiFj jjjjJIJjjjiiljjjjjjijjijijiijjjjjjijijij'jji iip|ji .jjiiijiiii "iji|' M"'y"||ilsi'ii:1"jii;'Vj* •:|||i|ji!Jjj\f\ 'f*>"*?'\fig ij1',jjif•' »j,"iiJ',«'J!'jjjb
heat. Also included are human.activities thatcause; direct disturbances such as t% d^
Finally, physical stressors include the exploitation'and harves"tmg"pf'resources. Table 4-3 illustrates the wi3e
            1 l'i 111   i,;"' , ,   • ' , :  fl,  ii 	i	 *  : "' '< , mill I ";,<< II11,, ,« ,«r':	i	L1", • I1 :<;, 	if '>, Tt .'''PJIIIi/llllli'lllllllL.i'illillP.r llllllh'l."!!!!!!!!!!," IIIIPIHIiNiPIIIMIIiliiili,!	rnlii'llilili',":;!!:,"!, 'i| ..iii'i'ln.!''!!1]!,!!!	ll'SijI'llllHE'iTii,,"1, 'ifPI1,!!'!1!1" I" 'Ifr PPHII..!!!'!' 'illliiJIDpi'1;, 'I'llilJlll'riiillD.hi.lillllllPlllliljIPIJ: ^IIIIIIIPIHI
range of physical stressors  that may be of interest.  Although EPA has Had less experience evaluating
physical stressors than chemical stressors, the SAB ranked Habitat alteration in the highest category in its
         ;   ,;,,;I'KI|,  	h   ,,	ii: i,' I:^•,: „,'^ ';;!:,,,•, ;„'",, :' ;iii,'j	;?:; ";' ri,":;ri;1 -,! i™! W';l -nt!'!"!"ilfi1 |l|i:wiiiis^                          	ijs,,'iii,e,sii;';iiiii^
relative risk linking exercise (tlS EPA, 1990). EPA expectsi'contmued"interest in evaluating these issues in
the Agency, particularly in'assessments done in partnership with offier agencies.
     Many of the concepts in this section are drawn from the disturbance literature and also from the
characterization of exposure issue paper (Suter et al., 1994).  While these guidelines use disturbance
terminology where  it  is the least awkward, the term disturb has been used to describe both trie initial event
(e.g., the backhoe disturbs  a wetland) and as well as the resulting status of the  environment (e.g., the filled
            j!':,;'i!,;|ii!  < :,„   ,',, ,  ; •.. >i;i' 	|,i, ;;  " ;„ i,. .MI,;;,,, v'A	KT.i." „., I,*'"* I," ,MM^"Wrt)HS	**';H(iiliL."j',KBWPU'JtitriHI	1;j"it-lV*fl1tlp'	.'''"'ST/i'iliii'Si'Si	\"VfH:	iTIlfBCai
wetland is disturbed habitat to tfie wndlife that may have used it).
     The analysis process shown in figure 4-1 is generally applicable for physical stressors.  For physical
        i1:,    ii'lillll'" : "   - ;, •:',	  "	!' i: „! „:' S  : t'1":,,!','!,,' l	i :>: ill;,1 ^il :;,:,;l,'i n iii-; y,»\ f<;i,:; iisi flKHKnuVi tit!,, IIK^^     	Viii, i	K>>- 'SMaiXS'-lKM	!•	siif ZZiBIIH^^^^^^^^
stressors that Constitute additions to the environment, the two principal activities of exposure and effects
characterization are readily identifiable. For example, trie amount of silt added to a stream is estimated in
       i ";   '.ij"!!   ,;	:, •",,   > '•  it1!",:,,' ' •"";: :i,"::•:>,•	Kf",,1,: C,	HVf'ii se ,'T	KiiS	ini'h'ni'Hr'-iiiaHKitftKT^ilV	VAIftWOSK,	iBRtllINB'9ljBU'.«iej-
charactenzatipn pf exposure, and the effect of siltatipn on the benthic cprnrnunity is eyaluated in
characterization of effects.  Hpweyer, for many physical stressprs, secondary effects such as changes in
        ',    •;„,SI    ,„,     ' ' "  '•,,"'  if,"1 . : •:,  :,,,|!':,i:"1	;";"£!,  i"; f jl;,,;";;*'; "i !'"|V'^'fl*l	fiif >f,9Kf8fc$foV'$"(1.jVi'^
-------
                    DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
Table 4-3.  Examples of Physical.Disturbances
Source or
Management
' Practice
Construction of
levees
Dredging of river
channel
Logging
Fill material
Bridge
Diversion of water
Primary
Stressor
Increased downstream
floods
Suspended sediments
Logging
Fill material
Bridge
Increased frequency
of low flow
Primary Effect
Altered riparian
community
Altered benthic
community
Altered forest
community
Eliminated wetland
Loss of sandbar habitat
with unobstructed views
Loss of riffle habitat
Secondary (Indirect) Effect
Altered wildlife community
Altered fish community
Altered wildlife community,
extirpation of species
Increased downstream
flooding
Decreased downstream water
quality
Decrease in whooping crane
abundance
Extirpation of endangered
snails
                                        89
10/13/95
-------
                                                       Ill 111 III I 111 III II 111 I 111 IIII 111 III 111
                                                                             lllllllllllllllllllll
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 I
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
wildlife communities are often of as much concern as the initial effects, such as the loss of forest community
from logging.  Because of the errtphasis on secondary effects, .this section is organized into characterizing
primary exposure and effects (section 4.3.2) and characterizing secondary exposure and effects (section
4.3.3). The outputs of the analysis process  are profiles of exposure and effects.  Figure 4-6 illustrates how
these concepts were implemented in an actual case, the bottomland hardwoods case study (Brody et al.,
1993).                     '
4.3.2.  Characterizing Primary Exposures and Effects
     The first objective of many assessments of physical stressors is to characterize the magnitude, extent,
and pattern of primary effects, that is, the disturbed environment. This characterization relies on two
principal activities, exposure and effects characterization.
     As with chemical stressors, this analysis often begins with a characterization of the source, which may
be an identifiable outfall, but more often is a management practice or action. The assessor should consider the
     > ', :r     i iii             i              i       i     * ii i  i   10  	»	                i              f
characteristics of the source that will influence the
nature, magnitude, and spatial and temporal
patterns of subsequent disturbances. For example,
whether a dam discharges water from the top or
bottom of the reservoir will influence the water   ,
quality downstream and the entrainment offish in
the turbines.
     Disturbance is part of every ecosystem, and
many anthropogenic disturbances have natural
counterparts. Human activities may change the magnitude or frequency of natural disturbances. For
example, development may decrease the frequency but increase the severity of fires, or increase the frequency
and severity of flooding in a watershed. When evaluating these types of disturbances, a characterization of
natural disturbance cycles is an important cdmponent of the analysis, phase.
Text Box 4-8. Bottomland Hardwood Example:
Characterizing Sources/Releases
           f-
        Changes in the hydrologic regime were
expected from the construction of levees were.
based on a hydrological model. These.changes
were superimposed over net subsidence (decrease
in sedimentation from earlier levees and global sea
level rise) and the decreased gradient of the river
from deposition of sediments at the mouth.
     Approaches to estimating primary exposure and effects will depend on the type of physical stressor
being evaluated. Physical stressors that constitute additions to the environment (clean sediment, heat, soil
moisture) are often assessed in a similar manner to chemicals as described above. The amount of the addition
    • '  "  : "   111              i      i             i   i    i  mi nil i ii i in n 111 mi mi 11  ii 1  n i i 11 i n i i  i i in  i in n  n i  n  i n i n 11 n inn in in iiiiiiiiiiii i iiiii
is measured or modeled, and these estimates are combined with estimates of the biological response to the
                                                         90
                                                                                           10/13/95
-------
                      DRAFT-DO NOT'QUOTE, CITE, OR DISTRIBUTE  '
     jS Source:
   
-------
                                  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE

  1       addition. Information on biological responses
  2       often relies on field data, and considerations are
  3       similar to those discussed in section 4.2.3.1.
 4       Assessments of physical stressors often focus on
 5       cornmunity-level effects. Because extrapolation of
 6       responses from one community to another is an
 7       area of great uncertainty, these assessments often
  8       rely on empirical data collected in the community
 9       of interest. An alternative is the use of mechanistic
Id       cpmniunity models (e.g., see text box 4-9). These
11       models often require a substantial resource
12       commitment and detailed information on the
13       ecosystem of interest, such as soil type, climate,
14       and vegetation composition and cover.
15           For disturbances that eliminate parts of or
T6       entire ecosystems, such as logging activity or the
    -•'.'..      »i              1           	Mil 11 li'n I   IIIIIII illllll Ilil I 'I (I  Hill	(Pi  ill |||||P||| ill 11   II i Id l|l|i|i|li|||illl|l  11	1
17       construction of. dams or parking lots, the characterization of primary effects is a relatively simple process.
, -.  . \ ;•   '•   ''"         	p[  "   "•        MI         i HI   11    i, in 11 (I  |i||i|i |i   i|i|||||i|i||ii| mi in  in mill i mi  mi iiii i in 11  i inn | i  '  n|| iiiiiii ii|iiii|i|i i iiii|iii|
18       Unlike exposure to chemical, biological, and physical agents, assessments of these types of disturbances
19       require no modeling of exposure pathways or measurement of contaminant concentrations, biomarkers, or
      'i    ',        ii in            i         i    i         t   i i   n i n  in 11   i iiiiii  n i    i     iiiiii   i  i  n  i i   i n • i iiiiii  iiiiiiiii
20 ,      radiation levels—the wetland is filled, the fish are harvested, the valley is flooded.  For these direct
     !     ,      "i'1"1' 'v •ill'IIIII"  '/,?''", - , !'''?;;."! "'If' fit'.'' 'I, i'i«it'.'^il!:W!'' ::!  II        I II II  I ' I  II Illllll I II    I II    II  II II   111 IIIIII   I 111 III   I I II I II IIII Illllll I  lllllIN
21       disturbances of the environment, the assessment burden is often on evaluating secondary effects (see section
22       4.3,3).  The increased availability of geographic information systems (GIS) has greatly expanded the options
     i, ,   ,'. ,-,„""  '""':: , all1  -i'; ";,.	 •'"••' • ••'!"'":' !A,-' '/.!*:: vJi'::.;'i''i'l*?;w",".ij i  II I  in   I il II II H I  I |ll 11 111 Illllll Illllll t
23       for analyzing and presenting the spatial dimension of disturbances.  These analyses Often take the form of
24       map overlays of potential disturbances with ecological resources (e.g., areas proposed for logging overlaid on  _
25       old growth forests).  In working with GIS, it is still important to recognize the difference between exposure
26       and effects and to address issues of causality when map overlays are used to show correlations between the
27       locations of stressors, ecological resources, and effects.
28           The dimensions of intensity, time, and spatial extent must also be considered when describing primary
29       effects.  Intensity can be expressed as the amount of physical stressor added or the severity of disturbance.
  Text Box 4-9.  Bottomland Hardwood Example:
  Characterization of Primary Exposure and
  Effects
          Mfbttnation on the hydrologic regime, the
  ecosystem, and biological response of the plant
  community v/as combined using the FORFLO
  modeL  The hydrologic regime was measured for
  current conditions using gauge readings and
  estimating water taSle depths.  Future hydrologic
  conditions were estimated based on a combination
  of modeling and professional judgment based on
  subsidence rates.  The ecosystem was characterized
  with data on soil type, climate, and measurements
  of plant community characteristics.  Plant response
  to soil moisture in terms of seedling germination,
  survival, and growth was characterized by grouping
  plant species according to their tolerance of  •
  waterlogging. The FORFLO output yielded curves
  of plant community composition in terms of basal
  area by species over time.
92
                                                                                                     10/13/95
-------
                                DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      Temporal aspects include.the duration and frequency of the disturbance and seasonal timing, when
 2      appropriate. Spatial dimensions can include area, pattern, and the landscape context,
 3        -         -                    '           '  '       .. .     .      ..•'•'•'
 4      4.3.3.  Characterizing Secondary Exposure and Effects
 5           As discussed in the introduction to this section,
 6      assessment endpolnts are often more closely linked
 7      with secondary effects than with the initial
 8      disturbances. One of the most challenging aspects of
 9      this part of the analysis involves identifying the
10      specific consequences of the disturbance that will
1-1      affect the assessment endpoint. For example, the
12   .   removal of riparian vegetation can generate  multiple
13 '     secondary disturbances; yet it is the resulting increase
14      in stream temperature that appears to be the primary
15      cause of adult salmon mortality (Suter et al.s 1994).
16           Characterizing secondary effects for
17      disturbances has much in common with secondary
18      effects assessments for chemicals (discussed in section 4.2,3.3). Both types of analyses rely heavily on life
19      history-characteristics of receptors.  Life history models (e.g., based on Leslie-style matrices) and semi-
20      quantitative methods such as Habitat Suitability Indices developed by the U.S. Fish and Wildlife Service can
21      be particularly useful for evaluating the impacts of disturbances on specific wildlife species,
22  .'     ..  ~                                   .  •                      '-...'
23      4.3.4.  Exposure and Stressor-Response Profiles
24     ..     the output of the analysis phase is a profile or profiles summarizing the results of the process. In cases
25      where a combined exposure and effects model was used, or when secondary effects predominate, this output
26      may be presented in a combined form,  In any case, by the close of the analysis phase, the assessor should be
27      able to present the following information:                                            '
28   ,.       Describe the boundaries of the analysis: The boundaries of the analysis should describe the level of  .
29      biological organization and the spatial and temporal boundaries of both the primary and secondary effects.
30           Summarize the most important pathways of exposure and effects.
text Box 4-10; Bottomland Hardwood
Example: Characterization of Secondary
Exposure and Effects
        Habitat Suitability Index (HSI) models
were used to identify specific attributes of the
bottomland hardwood ecosystem important to the
indicator species: gray squirrel, sWamp rabbit,
mink, downy woodpecker, and wood duck.
Important ecosystem attributes included canopy
closure, mast production, annual flood duration,
diameter breast height (DBH) of overstory  trees,
basal area, number of snags, and amount of winter
cover. The HSI models were also used to, quantify
changes in the habitat value for the indicator
species.
                                                        93
                                     10/13/95
-------
 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
•' . i '" ii-"",,,' „  ;• -:1,, •:'•••'•.  •: 'ii: -r •-,- •, :,•:•:> ;	it	'iiuvwwrfl* VAtf	iiiti	\pj-tjt-t'••:	naif	"iii,1" :.»[«,"[-»iii	ii.i:»! ii	b'etffiHNWHnr'
  1            Summarize the methods used to perform the analyses: including any statistical and modeling
  2       techniques. In addition, stressor-response and exposure information (or assumptions) used in the analysis
  3    ...  should be sumrnarized,	  t	,	,„.	                             ;
  4            Describe the three dimensions of the primary and secondary effects: The dimensions of intensity,
  5       time, and space should be addressed for both the primary and secondary effects.
  6            Describe the uncertainty associated with the estimates: An important component of the output from
  7       the analysis phase is a summarization of the important uncertainties (see section 4.1.2 for a discussion of the
  8       different sources of uncertainty). In particular, the assessor should:
 i   ' :<   ' '  ' . '• - '•',    I '111              '    '    I   '   I I   'I  ."  '  '        .
 . ,,   ;   -    ' ,',	# '  ' '    IFI                         ,    I   I  I    'I ml Sill t'llllllii I x I'llliiaH'iii11	1'	Ill	I ihliliiiiln	!!±: i lltrii'hriit'r	i	T i; 'i l.i
  9       •    identify key assumptions and describe how they were handled;
 10       •    distinguish between variability and measurement and systematic; uncertainty;
 11       •    identify the most sensitive variables influencing exposure and effects;
     • " '  ii     '    '   ""ih ilf <"T !,  " 'r mi1 i"'", i, "iii... ,j,! '„ ft1,:,M' i  'n!', i' '|;!' a',;,|;'::; ill•'ilj«f! „!,;•'" ii; ,'!«••• "ii' j ii-;;vii: • /iif,,• |li d'All Wliliii1!!!!!1"!;Iffi'!;!,K"ij!l In l|llllllllll''llltfliliriBHIi!!:!IIV. nfliiitlu        ,>: 4T,	^Wf:\:,L:;f ''till1']';:'''iil|.,l
 12       •    identify which uncertainties can be reduced through the colJTectipn of more data.
 13   "'      '    '    'i"1"'   ••"•'•1L-l!;"^il •"-"•  ' ••-•--^^^••^	-^'	-1'	-i'''^1	•  '" '         -  -'	^-	-'*        	•	•	r"
 14       4.4;  ANALYSIS OF BIOLOGICAL IP4TRODUCTIONS
 15             This section of these guidelines draws extensively from concepts found in the Biological Stressors
 16       issue paper (Simberloff and Alexander, 1994), but the issue paper materials have been modified as necessary
 17       to meet Agency needs.
 18    "  '	             _""'  '  |   /  '";  \	 n	^'/'"  ;
 19       4.4.1.  Introduction
20 ,           The evaluation of the ecological risks of biological stressors poses a distinct challenge to EPA.  Many of
21       the stressors are the result of state-of-the-art advances in biotechnology, such as genetically engineered
22       Rhizobium spp. (McClung and Sayre, 1994). Som'e are native parasites or pathogens that are cultured en
      in i  '    I :'< "  ,,   iii 'Ml:'!1 ', '!',:„ ii 	,,'i    II                  I  I  I    III  I III I I I  ill l||l l| I III  MM   III I I II   I  III I  III I I I I II  Illllll IIII III 11II III
23       masse and then introduced to control specific agricultural and horticultural pests (U.S. EPA, 1989b).  Others
24       are non-native species introduced for a specific purpose.  For example, the introduction of the grass carp in
25       the southern United States was intended to be a more environmentally friendly alternative to herbicides to
26       control noxious aquatic weeds.  However, the carp did not distinguish between noxious and beneficial plants
27       and unanticipated adverse effects to aquatic communities occurred. Watershed evaluations may be
28       confronted with an analogous use of natural organisms to ameliorate a particular problem.  As will be
             " »' 	    Iii	Ill : I-	 ii, '"'i :                                    1                ,                   ,
29 •      discussed later in this section, many exotic species that were deliberately or inadvertently introduced into the
30       United States have resulted in serious consequences to the environment.  The lessons learned from these past
                     ;"||•';:,;" .  "; t '",                                   | I III  ||| i            wtsra^
                  	'...            '                      .94  •
                                                                     10/13/95
-------
                                  DRAFT-DO NOT QUOTE, CITE, OR'DISTRIBUTE
  1
  1.  .
 3
  4
.  5
  6
  7
.8
  9
10
11
12
13
14'-
15.
16
17
18
19
20
.21
22
23
24
25
26
27
28
29
incidences have resulted in approaches that emphasize prevention and containment.  Risk assessment is a
useful approach in dealing with biological stressors, but the main emphasis is prevention rather than
mitigation.    .
     While some have questioned whether the ecological risk framework is suitable for assessing biological
stressors, the risk concepts in the framework offer a flexible and logical approach that is generally applicable.
However, there are important differences between        .   .             .  .   '
biological and physical and chemical stressors (text
box 4-11). Associated with these differences are
the potential consequences when a biological
stressor becomes established in a new environment.
                                                    Text Box 4-11. Unique Features of Biological
                                                    Stressors.  (Siftiberloff and Alexander, 1994)
                                                    Biological stressors can:-
                                                           Reproduce and multiply.
                                                           Disperse in a number of ways.
                                                           Interact with other organisms in ways that
                                                           are often difficult to predict.
                                                           Evolve over time.
 Table 4-4 lists five examples of biological stressors
 and the primary and secondary effects associated
 with each.  The consequences range from extremely
 severe (chestnut blight) to effects that may not (and
 perhaps never will) result in widespread ecological disturbance (introduction of the caiman). Opinions about
 whether a particular biological stressor is beneficial or deleterious can also differ.  For example, the
 proliferation of Hydrilla in the Potomac River is the bane of pleasure boaters, but Hydrilla has also, improved
 the water clarity and provided food and habitat for fish and wildlife. The sport fishery has improved to such
 an extent that the Potomac River is now one of the premier largemouth bass (Microptenis salmoides) fishing
 areas in the country.
      Because of the uniqueness of biological stressors, .the separation of exposure and effects analysis is not
 always easy. For instance, the movement of .bacteria or fungal spores in soil can be modeled in an analogous
 fashion to nonliving colloidal or larger particles.  However, depending on the species, considerations such as
 the formation of resting spores and ability to exist as a saprophyte when suitable living hosts are not present
 require that the analysis carefully explore the life cycle of the microorganism, including (but not limited to)
.effects such as pathogenicity, free-living potential, and the requirement for multiple hosts to complete the life'
 cycle (e.g., cedar-apple rust  and wheat rust). Thus it is important that the exposure and effects analyses are
. coordinated,and that individuals with appropriate expertise (microbiology, entomology, etc.) are involved as
 necessary in the assessment.
                                                          95
                                                                                            10/13/95
-------
                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
,     ' '  ;.     ii                   i                  	  IP	II	' i   i i i I   	In I  " ill 
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3-
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
•14
 15
 16
 1.7
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
     It is not possible to: address all biological stressors in these guidelines.  Therefore, a case study that
evaluates the risks of insects and fungal pathogens becoming established as the result of the importation of
Chilean logs into the northwest United States will be used to illustrate important principles that may be
applied to other biological stressors. A brief summary of the case is presented in text box 4-12. An overall
schematic of the analysis phase process is shown in figure 4-7,
     Because highly quantitative exposure and stressor-response profiles are not generally attainable at this
time for biological stressors, this section focuses on general principles and the use of expert judgement.  The
exposure aspects of biological stressors are discussed, including their ability to enter, survive, proliferate, and
disperse in a new environment, followed by consideration of the possible effects  of biological stressors once
they are released or become established in a new environment.
     To date, ecological risk assessments for biological stressors, such as those performed by the U.S,
Department of Agriculture (US DA), frequently  incorporate a delphic approach into the risk assessment (Orr
et al., 1993). We have relied on USDA's methods in these guidelines in view of USDA's considerable
experience in this area.  Wiegert and Bartell (1994)                •        ,
provide' additional approaches for evaluating events
that, although they have a low probability of
occurrence, have major consequences if they occur.
The use of fault trees can also be useful in scoping
out pathways that can lead to the introduction and
establishment of biological stressors. The method
was in fact developed for identifying causal
pathways for events with catastrophic
consequences (Barnthouse et al., 1986).
 Bamthouse and Brown (1994) discuss this
 approach in developing conceptual models for
 chemicals, but the same principles can apply to
 biological stressors, and the fault tree method could
 be a useful adjunct for problem formulation.
Text Box 4-12. Chilean Log Case Study at a
Glance
        The Animal and Plant Healthinspection
Service (APHIS), USDA, regulates the importation
of foreign plants and animals. The timber industry
:wants to import Chilean logs, predominantly
Monterey pine (Pinus radiata}, for processing into
lumber. A team of six APHIS experts evaluated
the probability and consequences of foreign insects:
and diseases becoming established in the western
forests; of the United States.  Although 14 pests
were evaluated for the Monterey Pine, only one, a
bark beetle (Hylurgus ligniperda) will .be used in
this case study summary.  This bark beetle-is
important not only because it can infest the trunk of
conifers^ but it also can serve as a vector for black
stain root disease (Leptographiumwagenerf).
                                                          97
                                                                                           10/13/95
-------
                             DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
             ••/ Source: \
            /   Proposed   N.
            \ importation of/
              \Chilean \ogy
<                Colonization
                 Potential:
               jf pests at entry
                   points
                #'''''"11
                  Spread
                 Potential:
                f pests beyond
                 entry point
                Ill
                               Volume
                  Entry:   \^
                 ofinfested    \
                logs into US  .S
               Likelihood of
              Establishment:,
                of pest and
               Exposure: of
             resources of concern
            Exposure Profile
rocessing Ecosystem/Receptor:
oint of entry Western U.S. forests,
estination for logs potential host species
Climate
>Ho.st suitability
^ Life history of pests
Dispersal mechanisms
Host availability
Climate
Geographic barriers


Similar pests on
U.S. hosts.
Pest effects on
U.S. species in
Chile,
Pest effects in
other countries
Characterization
of Effects/
Consequences of
Establishment:
by analogy to similar
species and mechanism of
action
\
r
ISffects Profile
,      Figure 4-7. Analysis example:  importation of Chilean logs
                '.•	I
                                                                            ^ai^
                                                                            iii ;,:;:;:„;!!	ft 'tfte''i
                         iiKS
-------
                                  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1           It is impossible to propose a universal checklist of questions that, if addressed, would result in a "fail-
  2      safe" assessment of biological stressors.  The following considerations, however, may aid in developing an
  3  . '   assessment.
 ,4      What natural history information is available?
  5           Knowing the natural history is a very important consideration, although the lessons learned from
  6      Chestnut Blight (Endothia parasitica) tell us that we can never be sure. In retrospect, an evaluation of the
  7      naturalhistory of this organism in Asia would have shown that it was basically a saprophytic organism with
  8      weak pathogenic capabilities. In evaluating microorganisms, however, the potential, pathogenicity (to plants
  9      or animals) is certainly an important consideration.
 10           Keeping current about the taxoriomie status of an organism may assist the assessor in making more
 11      informed, and credible decisions.  For instance, many fungi have only been classified on the basis of their
 12      asexual stage, and an artificial method of classification is used. When the perfect or sexual stage is known,
 13      the fungus is then classified in the normal manner (i.e., by .the type of sexual spore and fruiting body formed).
 14 ..    The perfect stage provides the assessor with much more information about the natural history of the fungus
•15      than the imperfect stage.  ,                     .......
-1.6      Wliat is known about similar organisms or past incidents?
 17           While they will not provide complete answers, lessons learned from past experiences with analogous or
 18      closely related organisms often are critical in trying to predict whether a stressor will survive, reproduce, and
 19      disperse. The consequences associated with analogous organisms may allow the assessor to  at least make a
 20      conservative or. worst-case estimate.        .                                           '
 21      What is the most appropriate source of data/information?
 22           In the evaluation of genetically engineered organisms, a combination of laboratory  and field experiments
 23      are conducted as well as the^inclusion of expert judgment (McClung and Sayre, 1994).  Nonetheless, most of
 24      these studies are designed to evaluate the efficacy of a given organism, not the potential adverse effects that
 25    ,- itmay elicit.  For instance, laboratory and field trials of genetically engineered Rhizobia are designed to,
 26      evaluate how well they enhance nitrogen fixation over native Rhizobia. The enhancement of nitrogen fixation
 27      is measured through the yield of the control and inoculated legumes.  Likewise, laboratory and, to some
 28      extent, field trials are used to evaluate how well an organism can control some type of pest. Actual
                        *            '      •          >               .             •
 29      experience in evaluating adverse effects caused by the organisms under the Agency's jurisdiction is quite
 30      limited.      .-   .-                  .                                 .
                                                         99
10/13/95
-------
                                                     if i 'Isrt1 nil 1 ..... jJi'i ' ; :it: »; «
                                                   i! if!:l : • , : ' ' :rv.'^ 1 1 «'ii' '^ »'' '[ «H'. i i- i 'I
                                                   ~>'i ...... r''y:i;iS:ij>«Hi« jj
                                                                                          '     '          "
                                    DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
11
18
19
20
21
22
23
24
25
26
27
28
29
30
 4.4.2.  Exposure Considerations
               i                          ,                    ••   \                 '         '   >        '
      This section outlines exposure considerations for biological stressors, including the likelihood of entry,
         1    ,1             '  I"    I   	I   |l  "I "ll'l'l'lll'll	I JFill Hi.  Ml    I    .   I      1       ,     I     I
 survival, and dispersal.  Simberloff and Alexander (1994) provide further discussion of these topics.  As
 noted in table 4-4, many biological introductions have resulted in dire consequences to the environment. As a
 result, the overall regulatory philosophy regarding unwanted exotic biological stressors is to prevent their
 occurrence rather than embark on extensive and costly eradication  efforts, which often do not work (e.g.,
 eradication of the imported fire ant). In cases where potential biological stressors are introduced, careful
 consideration is given to containment and, if needed, control procedures, if the unanticipated occurs.

 4.4.2.1.  Likelihood of Entry
      In the case of foreign biological stressors not present in the country, an initial consideration is the
 likelihood that one or more pests may enter the United States as a result of international trade (importation of
 agriculture produce, nursery stock) or activity (presence  of pests in airplanes or ships, dumping of bilge
          '•   Mai .. i"'. .   c. ',' "•.'. ',* . i'ii'",! i ..'	'• i)"1!"1;: iif'f'iiSS1!"!'':: ' 	'  " ni" i;:!;;1 :ij.'!,,/:' FJ ';>!\,. )!'!:.,, >;, 'i;*	i^iiijiijiii;; iv < '^''fl'i'ii'^ijtiljii iMiiS^            !!!I!!!!!B^^^^^^^
 field design to prevent genetically engineered Miizobia from migrating out of the plots.  There  are several Icey
     !;:  •;',-'''•, vial ;l!";:". :'i"•:"";:"• I.':1',.I3:i!',i:'.!!;:^,-,:Sil&''jJMfStl^S	
 considerations regarding entry potential.
                                                                                                   «i i'iw^           I
   Will the pest be present on a transportable item?
.  ,'   ' 1'  ijl|| ;. ;' .:  ;','i; 	, ' is;;;lii-l,A'; :'• •:,••.,;;);«•,L;'l-jgS;::''J !:§,-.i!"t,B,,i;i;ii	i!H:;ilft"M»	• WliWBS
   Will the pest be present on the item at the time of export?
     1 ''1   !iik!1i|i  , 	 '• '.'M| :  ",,:   Ii!,'miij !! , !;:ill'i^^>j;l,\!:i>NiA!ri:!i:i!i!lii:;:'l>>;' ,„  i'!;	/'jrl,; :Uiit!,.'"lyjlrs;iliWlilll'SBl. ^	/til;!	:1:!«! f	,	v	
 •    Is the new,environment similar to,OT^tferent,From ffie^e	tHe,biological stressor originated from?            |
      	•  ,'   fiisiti,'.."•" "'	";"•'.: "•'!r' "ie; * -1-:'i: \ i^-i wmm '^ winSiw £::S^^    	iliii                               	
      •	, -;   	i'si; ]•',.,••' '• t'''.;j;' ;,• • iiii'1;1	:!	,,,8',/ f •,".~aia:>;• •";	.;;>.ax	ai^taiMii^iiat^'11
    "'   : >,';, ij;*ti:"• • i';.  : -	">; • ul^.: I	'i , ?"!;?;: W«lr;	$'£&	K	SSM	li	lli	f	OHM	,	•	•,	i;	•	i,	is?	i	,	r	xn	
          	;•	:•:•••: "w- •  ••::'.-••.	••	••••:":i-;L'r-:lOO     '  '                             .        10/13/95
                                                                     'II' "Hi"!! '"'i Itl/ ,1"',' 'PIPPllP'flPliPKlPfP, T!1:1!":!;!,1"!'! il!" iPiil1!"!! I'ViP"1)!':1
                       ,;;{!
                              ;," i  " '" ,• '  	 ,,,'[,;	,   11!' i, ,.'•;" :;,' "•;||;| f •, i!h|' •; .it f1],!, a, ";i •	''" av i ii i' ;iii» • iiitiLiiiiiii 'i; KIHH p';:," IN • i. i,if i,," f a, •:;	j,4 jiPKi1' in1 iii,,,' nil!. • fl "i ";"1::" ?i",i an u • „, J'sa;; •- la j1 "HI: lae iiHiijfiii: ijtHinn
                           •|i;j:;!" ^; = ::;::"::::!Ii^-:;;!;r':i:::^^                                             	B^^^^^^^^^I
-------
                         DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
   1
  2
  3
 .4
  5
 .6
  7
  8
  9
 10  .
 11
 ,12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23  .
 24
 25
 26
 27
 28
 29
•30   '
     What are the habitat and host needs of the  .
     biological stressor?
     What survival mechanisms do the stressor
     possess (e.g., sclerotia, chlamydospores,
     ability to aestivate during drought conditions)?
     Is there an opportunity for repeated
     introductions into the new environment?
     How many generations does the stressor
     produce in one year?
     What are the fecundity and survival rates?
    Text Box 4-13. Chilean Log Case Study::
    Survival/Proliferation of Hylurgus ligniperda
    •      The climates of Chile and the northwest
           United States are similar.
    .»   _ , Similar species of pine occur in the United
           States and Chile.
    •      The genus Hylurgus is found in Europe,
           Great Britain, and western Siberia and has
           been introduced into Japan, Australia, New
           Zealand, and South Africa, It is very
           adaptable!                      .
•    Is there a presence or absence of natural predators, parasites, or diseases?                           •
  N '  The survival of microorganisms can be measured to some extent under laboratory conditions. However,
if the microorganism has certain survival mechanisms (e.g., spores), it may be impossible to determine how
long it could'survive under adverse conditions.  Many resting stages can remain viable for years. More
complex-organisms such as insects, fish, reptiles, and mammals become increasingly difficult to evaluate
under laboratory conditions.  Therefore, a knowledge of the stressor's life history, including temperature
tolerance, food preferen'ce, potential (or absence of) predators, is important.

4.4.2.3. Likelihood of Dispersal          .
     After considering the likelihood of survival of a given biological stressor, the assessor needs to consider
if the biological stressor can spread beyond the colonization area.  Two major factors must be considered:  the
possibility of dispersal and the rate of dispersal.
Text box 4-14 shows several ways biological
stressors can be dispersed. Text box 4-15
summarizes disposal considerations used in the
Chilean log case study.
     Methods of dispersal center around the  life
history of the particular stressor. Often one or
more mechanisms are important. Many plant
                              V
pathogens become established via airborne
Text Box 4-14. Mechanisms of Dispersal
       Air currents
       Rivers, lakes, streams    -
       Over and/or through the soil .surface
       Through ground water
       Splashing or raindrops
       Human activity (boats, campers)
       Passive transmittal by other organisms
       Biological vectors
                                                101
                                           10/13/95
-------
DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                          Text Box 4-16.  Chilean Log Case Study:
                          Potential Effects Caused by Hylurgus
                          ligniperda
                                  The.black stain root disease is vectored by
                          Hylurgus. APHIS cites a.moderate potential for
                          economic damage as a result of the disease and a
                          moderate potential for environmental damage
                          caused by the death of trees (specifics not
                          provided).
 1      recommended by Simberloff and Alexander (1994) to help estimate what types of effects a biological stressor
• 2     .is likely to cause. Species .traits involve the compilation of relevant characteristics such as the ability to grow
 3      quickly, survive under harsh conditions, or reproduce rapidly. An important consideration discussed by
 4      Simberloff and Alexander (1994) is that a stressor does not have to reproduce to cause adverse ecological  -
 5      effects (e.g., introduction of sterile grass carp in aquatic systems). Text Box 4-16 summarizes the effects that
 6      might result from the 'establishment of the bark beetle, Hylurgus ligniperda.
 1           As noted earlier, one of the unique                                      *
 8      characteristics of biological stressors is their ability
 9      to evolve. One aspect of this consideration in the,
10   •..  potential for hybridization, as illustrated by the
11      genetic change induced in the Siberian Weasel
12      (Mustela sibirica itatsi) through hybridization with
13      the introduced (to Japan) of the Korean Weasel
14     . (M.s. coreana).  In the United States, weed pests
15"   ,  such as Johnson Grass (Sorghum halpense) have
16      become more  aggressive  invaders as a result of hybridization with cultivated Sorghum (Hordeum vulgarum}.
17           In summary, a risk assessor has a difficult task in predicting the possible ecological effects posed by a
18      particular stressor.  Clearly, no one assessor can be an expert in all of the fields required to evaluate a
19.     biological stressor.  Therefore it is very important that a team approach be utilized to ensure that the risk
20    '  assessment is  scientifically credible.
21  ....                      :  ,       ,              '                            .        -.   .
22      4.4.4. Exposure and Stressor-Response Profiles
23           Because  most evaluations of biological stressors involve a team approach with professional judgement
24      as the main input, stressor and exposure profiles for biological stressors are likely to differ from those
25      prepared for chemical and physical stressors. While the profiles will not usually be quantitative, they can still
26      be quite useful.  For many organisms such as plant pathogens, models can estimate exposure considerations
27     < such as dispersal. For others, qualitative estimates are the only means for assessing both exposure and
28      effects.            .     .                          "                                             .   .
29           An important component of both the exposure and stressor response profile -is the uncertainty in the
30    .  analyses.  A biological stressor's ability to reproduce, disperse, and evolve imparts a great deal of uncertainty
                        103
                                                                                                    10/13/95
-------
                                                             Ill Illlllllllllllllllllllllllll
                                                                       (I Illllll Hill
                                   DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
"'2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 13
 19
20
21
22
23
24
25
26
27
28
29
30
with regard to predictions as to whether a given stressor will colonize one or more areas and what the effects
will be, As noted by Simberloff and Alexander (1994), however, decisions will continue to be made in the
face of uncertainty.
     Given the uncertainty, what is the best method to estimate risk? Simberloff and Alexander appear to
support a delphic approach, using a team of experts to evaluate a given biological stressor.  This approach
has been used not only by the USDA but also the State of California to assess risks posed by the
Mediterranean fruit fly and by the National Institutes of Health Recombinant DNA Advisory Committee.
The majority of the assessments, however, have been conducted on agricultural commodities where there is
some knowledge about the major pests that occur overseas. Even with pes'ts that have been studied
•   |; •   •';  m , r , •- ••'": „'.! •  /" £  i  .i',1"  'i,:1,	    i1',:,,! ii,,)"",!,!,: iii if::; MI	:•;•    i       i iiiipii in  n n  ii in ill n MI   i,  |   | pin   pi  inn  •   i  n ,11111 in |inii|ii in
    " •  : ';" i  - 'Sil'HI!  If  ,,:,'"	1	  I1	Hlllli,:;",! :i!|,i,i,,",i:ii	,-!;,!',,1,"HI"1:;::",':,", I          I III III II  II I      '      ,1   I      'I     I      I  l I I I
house sparrow in the United States.  The former did not spread beyond St. Louis for more than a century,
 '  .' " ' '•],' ' ''SI, I"1., „" "";!;; 'i''1-/:„::"" I,?! "US '• f''.|::^Ji|jSiS:ij;,tiS;!f'	'	I	ri	"i"!	1 "'Ftp	fiiiliiTiiii	iiiii	"I	 lllji iiillilil mi llPli  I.n I • Ii in (i|i|lliillllllil
whereas the Jjguse, sparrow spread throughout North America.  Both, species are similar in habits and
  1 '!  "'  ,;| ;"  '"ijinill i ,l'	  i	 ,  iV" ,'',|| I?' I;:':*;,, ":"'i:'/Mvi!::^               | |  |    || |   g|  ||||gg| 111 1111 l l ill || l 1 l  iiiii   l  111  IIII Illllll  11 in 1 I || || 11| 11|||
mqrphology. Why one colonized extensively and one did not remains a mystery (Simberloff and Alexander,
           ffliiiiii  i "   i ',',' 'i'"1  M i;,i: :,:,, 'Kii,'""",,,"',,,"!,!;:,1"!!:1"1,,!!":	iiM/1:,-"!:         HIM MI   linn nil in i IN in IN     inn mi  i  HIM   n ill i in in 11 in iiiiiiiini
1994).   '   ™  "...  	'"';'".!T."'',"",!'™'"Z'",
      ' '  'Kill1 '.' "• i" '''"'ii';,  !'.'•• h, •'.)' ii,,  ,!;'':: :,,,i! "Ml,:1"1!:!;!;;":!!,,!/- "i'-Sili;!,  ""*    i i n in    i i  iiiiiiiini i nun n n i i i i in   in  n  11 i  i  in iiiiininiiiiiiiii
     Thus a combination of a delphic approach coupled with a knowledge of past case histories is probably
                                                                                                          I	l	f
                                                                                                         iiiiiiii iiiiiiii |
 the, best approach. It would be naive to expect that
    „!  '" ,  ,   ",:,,!iill, ,' ,",„:: i|M, 	:,,»',",'; /ill	';,,,i '.n,1'	."iiii!1 :' '!l';i'::;iiJil'1'!llil1::1 'i, ,1,,, ''ill, I"-ill!,,,
 a complete uncertainty analysis could be performed
n.    '       - Miiiii ,•  ,' • ,  '.''„  'i  , ii ',• i,i>i lip f  ;.,;>, M ,« "ny	ii'ijiiiiiiiriXJ  	TI ii'iu :iH'<:<'
 for many organisms, particularly when the natural-
 history of a given stressor is only partially
 understood. |n the Chilean log case study, APHIS .
 experts  admitted that for some pests there were
 little or  no information. However, they also
 stressed that no information is not equivalent with
 Ipw risk.
              II          1     I                     *        f                          I   I II  II      L I III
      The Chilean log case study indicated that no mitigation considerations were made during the risk
 assessment. If the risks were deemed sufficiently high, however, appropriate mitigation steps would be •
 considered.
                                                      Text Box 4-17.  Chilean Log Case Study:
                                                      Uncertainties
                                                      •       Can Hylurgus ligniperda vector fungal
                                                              palhogens other than Leptographium
                                                              species?-
                                                      •       What are the risks of pests  for which there
                                                              is incomplete information regarding their
                                                              life history? •
                                                            104
                                                                                                10/13/95
                                                              III III Hill 111 nllllli linn IIIIIH   III I I In 11 III 11 III Illllll IIIIIIIIIIIIIIIII lill 1 Illllll 1111II Illlll Illill Illlllliliil
                                                                                                        	II	(I
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  f
 ,2
 3
 4
 5
 9
10
11
12
1-3
14
15
16
17
18
19
20,
21
22
23
24
25
26
27
28
29'
30
4.5. ANALYSIS OF MULTIPLE STRESSORS

4.5.1.  Introduction
     The result of exposure to multiple stressors is generally termed cumulative risk or cumulative impact
(see section 1.6.2). Within EPA, cumulative risks are receiving increasing attention as more emphasis is
placed on community- or ecosystem-based risk assessments (Irwin and Rodes, 1992;  U.S. EPA, 1994e).
Historically, the cumulative impacts of Federal actions have been considered in environmental impact
statements prepared under the National Environmental Policy Act. Consideration of the cumulative impacts
of discharges of dredge or fill material is also required under Section 404 of the Clean Water Act (Leibowitz
et al., 1992).
     At any'point in time, multiple stressors are present in any  give ecosystem. Some are natural; others are
anthropogenic.  Although several approaches for evaluating risks associated with chemical mixtures are
available, our ability to conduct risk assessments'involving multiple chemical, physical, and biological
stressors, especially at larger-spatial scales, is limited by our understanding of the ecological processes
operating at these scales. When effects are observed (e.g., a declining fishery resource), there may be
insufficient data to accurately weigh the individual contributions of multiple stressors or even recognize all
the  stressors that may be present. Nevertheless, the risk assessment process offers a valuable systematic
approach to organizing and evaluating available.information in a way that can be useful to a decision-maker.
     This section suggests some options and considerations relevant to evaluating the effects of combinations
of chemical and nonchemical stressors.  General analysis phase approaches for predicting and measuring the
effects associated with multiple stressors are reviewed, and the  problem of establishing causal linkages
between stressors and observed effects is discussed. Some of the concepts discussed in this section are drawn
from'the Effects Characterization issue paper (Sheehan and Loucks, 1994), but the issue paper materials
have been modified as necessary to meet Agency needs.                                       •

4.5.2.  Predicting Effects Of Multiple Stressors
     This section evaluates our ability to predict the effects of multiple stressors given a knowledge of the
individual stressors. The effects of.chemieal mixtures may be tested directly when the mixtures are available
(e.g.,'Contaminated soil,  wastewater effluents), synthetic mixtures may be used, or toxicity of a mixture may
be predicted from toxicity information on the individual constituents. This section focuses on predictive
                                                         105
                                                                                           10/13/95
-------
                                   DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                                                             i ii	l( ill .1 i| l	
                                                            1 I   '  in I,	 ill i  '  i     11
                        in i in i nil	n it (1111
                          1  II III! I I
   1      approaches and discusses the suitability of an assumption of additivity of toxicity among mixture
   2    '  components.  	                	" "
   3           The work of Plackett and Hewlett (1952), which describes four models for the quantal response (e.g.,
   4      mortality) of organisms to chemicals in a mixture, provided a starting point for many subsequent approaches
   5      to predicting the toxicity of chemical mixtures.  To make the problem tractable, most researchers have
 	  6      assumed a lack of interaction between the indiyidual chemicals in the mixture,M leaving two models: simple         I
      i. ,  ' '    , .   ' ,: '  .iiiiiiiii,  i1    '::  T "'»  >'•, in vi i,1;,  ,ii ,11, •::	' ,.„:	i. ''.'iivjiir'-'ii", < i ^1,111; !,!• j, ira ^niii,1:1;,,,! ''I'lnviuiiriHiiii	lOK'''!!1 i    , '   ,  ,          ,     , , , •     •
   7      similar action (concentration addition) and independent action (response addition).  The widely used toxic
   8      unit concept (Marking and Dawson, 1975) is based on the concentration addition model, where the fractions
   9      of an LCs"	IL-;	IJ/.TI	        .                     	fi&^i^'^'lli&niiffilflKMKIifB I
  10   '   individual concentration'in the mixture aridTts"t C50.	An LC"5o	for th"e™rnix"ture"is" p re3icted	to	bccuFwFen the
      •I'., '' .'Wl ' "<  "    <  ' <" '  jlill'l' .1 '; . p i1 V ',;!." '""'« ""'"• ;;' i'i I';!" !llllll < .IVV '' '"I1', ".i,1!1!! S .lib''1'11'"! ;'l'' i' l^^r li1.''1™'1 I"I:|H<:"III!	Ill'tiJIIIi''^!!!'!^!^;!^!!''!'!!*!!!	Hi	1!|N!l|||M                  l:illlllllllil|l!iMill|l!llilli*Jil':i'i,IVIIirT4l,!!:liilliHIIIIIIIIIIP!lll»^          I
      : •'  "'	"....;,  irii ;' 1 •.  •.''' '< •;' r • • it .'I  ! i;i: L(l.;;;.ki*t*1) •! i:;!:E«i ;:N i, i •/"	'*J&	Siililliia	Bli	,	islli	1	3	J	i	«f	I	f >'	te	-	i»n	'	f	»"	•	•	I	'!	1	11
  11      sum of the fractions equals one.  Concentration addition assumes that the sites and mode ot action ot the
  12      chemicals are similar in the mixture; response addition assumes that the sites and mode of action are
      *:	   . •  • •  ...;"	   fin"'.!.,; ! •••	- >.-:•',;p	••;;zw t\;$.$&w•«• ww$'mwMM^mmsiMmsm^mm'mm'mM
  13   ^   diflerent,	t	  	f ^	r ^	'	b	_	,	|>;	^	^	i	|	B	rir	t^	h|	^	^
  14           When the modes o'f action of chemicals  in a mixture are known to be similar, additivity (concentration
      i. •" "••  •   " • •  . ' i .  l|:"li ':	;;,•''. •:'	'.i;!!':"';'.! i;;,;,: III:: •;	:";:i17;<$!SSi* *!I,-:K^	i^IS^ISpfiSt£1^^.-.-"."!"~zrj;,-T,::;;"1,';i:: .:^',Z^p^±
,  15      Addition) may be appropriate.  Tests with aquatic organisms and hydrophobic organic chemicals acting by an
  16      apparent generalized narcotic mode of action seeni to  show additivity for mortafity [e.g., Konemanh,  !9SI;  •
•  17  '    Broderius andKahl, 1985). "indeed, additivity'for "lethality	as	an'eTilipo^rwas'lfoundl	in'rriixtiiires'fiaviiig	up	to	
  18      50 chemicals, with each chemical at a  concentration of only 0.02 of its LC50 (Konemann, 1981; Hermens et
  19      aL, 1984a).  Similar studies using reproduction in Daphnia magna as the endpoint gave results that were
  20      close to but somewhat less than additive (e.g., Hermens et al., 1984a,b). Nevertheless; there is insufficient
  21      justification for assuming  additivity in mixtures where the.mode of action of constituent chemicals in
  22      unknown or partially known.  Thus, no single approach can be recommended for predicting the effects of
  23.   ',   chemical mixtures (U.S. EPA, 1986b).  	'^	^^	^ r,_;	^	^	^	^
  24           Some factors the risk assessor should consider when attempting to .predicti the effects of a chemical
   •                     i                      i      ,    t 	•'.     11111  iiiiiiiii  mi mi i mil nil 'iHrfJwnnwpMtf'f'  (     in   *    in           I
  25      mixture from individual constituent data are listed below.                              .
  26      •   Do the exposure measurements for mixture components accurately reflect component bioavailability to
  27        .   the organism?
  28      •   Are interactions between mixture constituents likely to affect exposure?
  29      •   How will the composition of the mixture change with time?
                                                           106
                    Miks
                                                            i in	* if UNI	tiii fill	II
.                 	liiiiu'l'i' il|«^   	      I
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3
  4
  5 '
  6
  7
  8,
  9
 10
 11.
 12
_13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
'25
 26
 27
 28
 29
 30,
 •    Are toxicity data for individual chemicals'comparable? For example, in an aquatic toxicity test
      involving multiple metals, were the individual metals tested under similar water quality conditions?
 • ..  Do the chemicals in the mixture have the same toxicological mode of action?
      While it is difficult to generalize concerning toxicity predictions for chemical mixtures, it is even more
 difficult to do so for stressor combinations that-include physical and biological stressors. One possible
 approach would be to evaluate the potential effects of each stressor separately, then combine them in the risk
. characterization phase.  It may be possible to draw supporting data from other situations in which the same or
 similar stressors have been observed in the past.  There is little basis for assuming additivity. The. most
 accurate/predictions will result from a sound understanding of system structure and function.
 4.5.3.  Measuring Effects of Multiple Stressors
      Many of the same techniques may be used for
 measuring the effects of single or multiple
 stressors, but the key difficulty in many "cases is to
 link an observed effect to any one particular
 stressor.  For determining the potential effects of a
 chemical mixture, it is preferable to directly test the
 mixture of concern (U.S.  EPA, 1986b). This
 approach is commonly used to evaluate the toxicity
 of wastewater effluents as well as contaminated
 sediments and soils.  Some issues that the risk
 assessor should consider  when using such an
 approach are listed below.                 ,                                '
      Is the sampling scheme for the mixture adequate to provide representative samples for testing?.   .
      Are issues of sample stability and storage before testing adequately defined? .•
      How do the selected measures of effect relate to the assessment endppint(s)?
 •   How will test results be interpreted in light of spatial and temporal variability in the mixture?
      Another option for the risk assessor is to use a synthetic mixture. Such testing allows for manipulation
 .of the mixture and investigation of how varying the components present or their ratios may affect mixture
 toxicity.  However, this approach requires additional assumptions about the relationship between effects of
Text Box 4-18. The Apparent Effects
Threshold Approach
The Apparent Effects Threshold (AET) is an
example of a field method that has been used for
evaluating the effects of individual chemicals,
occurring in mixtures in sediments. The AET for a-
chemical is its concentration in the sediment above
which an adverse effect to the sediment biota has
always been observed (Cirone and Pastorak, 1993).
The. adverse effect may be toxicity observed .in a
laboratory bioassay or alteration in benthic
community structure compared to a reference site.
Cause and effect are inferred from the presence of
contaminants in the sediments and the
corresponding condition or presence of sediment
organisms.
                                                         107
                                                                                             10/13/95
-------
iii'i?;!'1 'I;,, „;"	IIMI'i r!1:!!*"1"!1"!!
   i
"   2
   3
   4
"  5
   6
   7
   8
   9
  10
  il
  12
  13
  14
  15
  16
  17
  18
  19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
       ; r'. ST'iiK1'.''' in11! i' iiiiiii:	„''' i| iiiii1"" il! f, j: ii	IIIBII	! iKm, n TIIIIIIV f ill1'111!! w: 'aa	Mirinii	mil:	,	.	.	.	,	
.,; , .'   i "i. '•.. ;;; , {:ii||,[ , ['.,"";,;:,.•;" .,	i;, J.J;!!	iiii.:.; •;.' .i1,,.;..                                                    _
  1  ,•   .'  : „;  ' liSI; i:.''. "... ' • : /i1 i •;: -...i;1! ii,"'ffwi^ iii,.:|^
      |i    'II,   	 ., ill,     '! ,, , 11 	 '< i: i' < ,  ,,'  'ilr1 ! [""M < , , lif',,,1, lip,1;1!!' ln,i<;|<: ''.nili „ !!„,"![ I W ri, Vlil' /'iiiiliiifPpLilPiiiliiJIIIIiriii iliiiililiilPliliil/JPiiiJilil IlllllliliilPIIII''!^'!!!!!!!!!!! INPilJi!	111,1' Ji1 nil' 'ill'llr Ilili'i

     • ^  '     ;';;'   • ^ • • •  ',. DRAFT-DO NOT ^UOT	^	^	 _	;	\	

           	pJii,!, i    , ' "•   "  , • • PJI,, ''	 < , ill1	i  "I,!" 1, i ii mi; :, :>,„, t ', 'I'1",, |, I'm, i. >ii''	ii,;1,:,	lilJiSlPliiipftlsll'ilih'-l'iirilRIISt11 iili!i|lilinilni!lllll!l!!;lllliiiiill1||i|ll|i;ill||:|!!i|ili:|i.i!|!l.iii:iililiiii||i! II1!	SIW	WllP'i'ii'ii'ii1'	!i!l::iip;iii:;lill,i,i!lJ!|»i!ili|1i,,li!lli!l!Hi!!,l",JP'!'"!!Pl I'fiJlllliiiiilillllllillllliillllliilillllllliilllllll'lli;
 the synthetic mixture and those of the environmental mixture^	Results	ofjes|s gilji^mxtiirgg	having varied  >
 component ratios can be fit to models such as those described in section 4.5'.2 (Suter, I993a). This  approach
 ....   . .i,.	>; i. ,,(!j:;i|| I I".;',:" ;,.;,(,"v i, .i'-ii_ pi!1 ^:, •   '              • ^                         	!(H    	     	     r1^	,	,	•	>
 is more frequently used to answer research questions than to directly support risk assessments.
      For physical stressors, empirical stressor-response relationships have been used to separate the relative
      ,! '   ' , ,'  ii'iliMl, • !""; '	  ,,' I,-1!1!,: '• : "ti. !*:•!!, •," • »fi ii,!1, n, aKJ'i1l»lll	IBS	SUM	KJISSfl-^'i:
 (contaminant concentrations in' sediments, sediment toxicity, and bottom dissolved oxygen) and habitat
 indicators (e.g., water depth, salinity, temperature, and sediment characteristics).  Investigators compared the
 areal extent of "degraded" benthos with trends in the exposure and response indicators.  This technique is
most useful as part of a screening/problem formulation approach, since the data collected may not be
sufficient to establish cause-effect relationships or to distinguish between natural processes and
anthropogenic stressors (Gentile et al., 1994).
1',,    i"1    n1,,,  ' • !ii,:i!i,!i!!!i' • '!„ ''"''i " 	'•                  ,                        i  '                                    "
      Another example is the synoptic index, which was developed to rank the relative risk of cumulative
wetland impacts between landscape subunits (e.g., counties, watersheds, or ecoregions) (Leibowitz et al.,
 1992). Specific synoptic indices are selected for an assessment based on the goals of the assessment, the
        1  •   i »i .                        'i     	in	ini	i   i                  'M
nature of the ecosystems and stressors involved, and an understanding of wetlands and their relationship to
landscape level processes.  Each index is composed of landscape indicators chosen based on data quality and
availability.  Index values are computed  for each landscape subunit and displayed on a map. Leibowitz et al.
note the importance of documenting the rationale for index selection,  assumptions, and data quality
                                                             •108
                                                                                                  10/13/95
                                                                                                                  in n iiiiiiii 11 mi 11
                                                                                                                  n iii ii i li 11
-------
                        DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1

  2

  3

  4

  5

  6

  7

  8

"  9

 10

 11

 12

 13

 14

 15

 16

 17

 18"

 19

 20

 21

 22

 23

 24

 25

•26

 27

 28

 29

 30
considerations and critically evaluating the accuracy of the assessment.. The synoptic index approach relies

heavily on professional judgment and, as with the EMAP example, is best considered a screening/problem

formulation level approach.
     Indices are used as part of rapid

bioassessment protocols developed by EPA for use

at field sites (U.S. EPA, 1989a).  These protocols

involve the collection of basic information on the

presence or absence of individuals from classes of

organisms which are expected to occur in certain

types of ecosystems.  Some elementary physical

and chemical data are also obtained to characterize

the habitat and exposure. The effect of stressors is
inferred by the condition of the environment with

respect to some known or expected reference

condition.

     Potential problems with the application of

indices can arise, however, when heterogeneous

measurements are combined into a single index.

Some of these limitations are summarized in text

box 4-19.  To reduce these limitations, the risk
assessor could focus on a real property of an
ecosystem or component rather than using an index

to measure vague concepts such  as "ecosystem

health" or "ecological integrity." "Appropriate

theoretical considerations can also be used in
selecting and combining variables. Indices that
include both exposure and effect measures should

be avoided unless there is-a clear, logical basis for
such combinations.  Multivariate statistical

techniques might provide alternative approaches.
Text Box 4-19. Potential Problems With
Indices (Suter, 1993b; Ott, 1978)

•    .  Ambiguity. A low index value could
       mean slight changes in several variables or
       a large change in one variable:
*:•;•    : Eclipsing. Low values of one component
       variable may be hidden by high values of
       ^another          ,
• •••     Arbitrary combining functions and
       variance.  Combining variables in
       different ways (additive, multiplicative,
       etc.) may greatly influence the index value
       and variance, but there may be little basis
       (biological or otherwise) for choosing one
       approach over another.
•      Unreality. Indices.may not reflect "real-
       world" properties. For example, an index
       value of 0.8 or 4 does not provide any real
       measure that would be directly useful in
       decision-making.
•      Unitary response scale:  Using a single
   -   "" index for multiple heterogeneous variables
       implies only one type of response by  -
       ecosystems to disturbance and one mode
       of action for the stressors involved.
•      No diagnostic  results. Combining several
       responses into a single index may hide
       variations in individual responses that
       could be useful in determining which type
       of stressor is responsible for the observed
       effects.
•      Disconnected from testing and
       modeling. Causal inferences derived from
       indices cannot be verifie'd using controlled
       laboratory tests or through available
       models.
                                                109
                                        10/13/95
-------
                         DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                                      I	I i     '  i i>| 	  (ill	'IPi  I	 »  '
4,5.4.  Evaluating Causal Evidence for Linking Observed Effects .to Stressors
•              '          "                             '
                                                                                                     K	           I
                                                      i in ..... i |i
  1
   «•  r:   '•   .                                                                      _

  2           Although causality is covered elsewhere in these guidelines (section 4.2.3.4), a brief discussion is

      , ,. . ' 'f  . "I"  , • '"i „!!!'  , ' Hllllllil'.,.  ',,„'                                        1                                      1
  3      included here to address separating the contribution of individual stressors to an observed combined effect.


....... 4      In the laboratory, when tests of a chemical mixture suggest toxic effects, there is still a question about which


  5      components of the mixture are causing the effect.  EPA has developed a toxicity identification evaluation


  6      (TIE) to address this problem. By using fractionation and other methods, the TIE approach can help identify
      ,111   n,     • , n,  ' » ,, ' iJIHIO  >, 'i , ," ' < "I!,!   1                    I              I  I I    I Illl III         il             I  I   I 1 1   111  I I

  7      chemicals responsible for toxicity and show the relative contributions to toxicity of different chemicals in


  8      aqueous effluents (U.S. EPA, 1988a, 1989c,d) and sediments (e.g., Ankley et al., 1990).


  9           When effects are observed in field situations, assigning causality to individual stressors can be much


 10      more difficult. Sometimes, a stressor may have a distinctive mode of action that suggests its role. Yoder and


 11  *   Rankin (1994) found that patterns of change observed in benthic fish communities could serve as indicators


 12      for different types of anthropogenic impact (e.g., nutrient enrichment vs. toxicity). Chemical stressors may


 13      provide biomarkers of exposure that can be used to provide a link to observed effects.  Often, the risk
     .............                                                                                                q  1

 1 4      assessor may have to rely on a weight of evidence analysis in risk characterization to evaluate the causal


 15      factors involved.
                                                                                                          I
                                                                                                   iiiiiiiiiiiiiiiiiiiiii
                                                               I	
                                                no
                                                                                                       in inn nil i n

                                                                                                     10/13/95
                                                          ,1 in i	iiiiiii iii ii 'in
                                                                                       (   HI  I             I
-------
  1 ,
•  2

  3
  4
  5
.  6
  7
  8

 .9   .
 10
 11
 12

 13
 14

 15  -
 16
 17 .
 18
 19
 20
 21
 22
 23
 24
L25 .
-26
 27
 28
,29

 30  -
 31
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE

                            5.  RISK CHARACTERIZATION PHASE
5.1.  INTRODUCTION

     Risk characterization (figure 5-1) is the final

phase of risk assessment.  There are two stages to~

this final phase of risk assessment: risk estimation

and risk description.   ,

     At the risk estimation stage the risk assessor

integrates the exposure and stressor-response

profiles from the analysis phase and estimates the

likelihood of adverse effects to the assessment

endpoint of concern. The qualitative and

quantitative elements of uncertainty are also

included with the risk estimate.

     The description of the risk assessment is the

second stage of risk characterization. The risk

description synthesizes an overall  conclusion

regarding risk estimates that is complete: and

informative; addresses the uncertainty,

assumptions, and limitations; and  is useful for the

decision-makers (NRC," 1994).  In addition to

describing the risk assessment, .the risk assessor

should acknowledge the iterative nature of risk

assessment and point out the need for additional data or analysis (US. -EPA, 1992a).

     The goal of the risk characterization is to fully disclose the strengths, weaknesses and assumptions in

the risk assessment. The lines of evidence supporting or refuting the risk assessment should be carefully

described in this stage of the risk characterization,'

      This section of these guidelines draws from concepts found in the Risk Integration Methods,

Ecological Recovery, and Ecological Significance issue papers (Wiegert and Bartell, 1994; Fisher and

Woodmansee, 1994; and Harwell  et al., 1994, respectively), but the issue paper materials have been modified

as necessary to meet Agency needs.
Text Box 5-1. What is Different in the Risk
Characterization Phase Diagram? •

       Experience with the application of risk
characterization as outlined in the Framework
Report suggests the need for several modifications
in this process.
       Risk estimation entails the mechanistic
integration of exposure and effects estimates along
with an analysis of uncertainties. The process of
risk estimation outlined in the Framework Report
explicitly separates integration and uncertainty.
The original purpose for this separation was to
emphasize the importance of estimating
uncertainty. This separation is no longer needed
since uncertainty analysis is now explicitly
addressed in most risk integration methods.
       The description of risk is expanded beyond
simply a risk summary and interpretation of
ecological significance, as presented in the
Framework Report. The additional topics that
should bejncluded in the risk description are: ,
•      the weight of evidence supporting
       causality, including linkages between
       measurement and assessment endpoints;
       and
•      a summary of the assumptions,
       weaknesses, and strengths in the
       assessment.
                                               Ill
                                        10/13/95
-------
DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                           PROBLEM FORMULATION


                           ANALYSIS


                           RISK CHARACTERIZATION
              ANALYSIS
            Risk Estimation
            Risk Description
          Risk estimate summary

          Weight of evidence

          Ecological significance

          Major assumptions and
          uncertainties
                  i
        Discussion Between the
    Risk Assessor and Risk Manager
               (Results)
                 I
           Risk Management
                                                                 a
                                                                 Q>
                                                                 sr
                                                                o
                                                                3
                                                                                      03

                                                                                      a


                                                                                      o
                                                                3
                                                                i:<:;-'' •   iw:!!<'>::!:::: ;
Figure 5-1.  Risk characterization phase

   ....................   t::'^
                                                           ....... EK^^^^^^               ^
                     ^i	^lf?!!K^;;;lw^^^^^^^^^^^^^^^^^^^^^                                         	
                     •;:*. i-;" • K ^K' -!• i i.    '             '   VliiiiilJ iBWiliK i'il-jiliiji;!:1	l<::' 'iiiiii;,! fliii-1	          -     I

-------
                              DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3  .
  4
  '5''
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26 .
 27
'28-
 29
 30   -
 31
 5.2.  RISK ESTIMATE
      Risk estimation is the first stage of risk characterization and is the process for integrating the exposure
 and effects analyses.  There is a wide array of risk estimation methods available to the risk assessor. Any
 approach that considers both effects and exposure can be considered to Be a risk estimation method.
 5.2.1.  Qualitative, and Quantitative
 Assessments
      While risk estimations can vary from highly
 qualitative to highly quantitative, most estimations
 have elements of both. Quantitative assessments
 are mathematical approximations of the
 relationship of stressor-response and exposure.
 Qualitative assessments rely solely on knowledge
 of ecological interactions to predict the likelihood
 of exposure and effects. The selection of which
 approach is most appropriate will usually- be
 determined in the analysis plan during problem
 formulation.  Situations that require a great deal of
- professional judgment will have to rely on a more
 qualitative approach, whereas those that utilize
 measured laboratory or field data will be more
 quantitative, even though the interpretation may be
 qualitative. Likewise, highly quantitative models may contain parameters such as the presence or absence of
 species for which data may be lacking. Thus, professional judgment may be needed, imparting a qualitative
 element to a quantitative model. In a qualitative  assessment, potential direct effects associated with a stressor
 could be ranked as high, medium, or low, or as yes/no alternatives (Wiegert and Bartell, 1994).
      Qualitative estimates of secondary effects may indicate the likelihood (high, medium, or low) of
 ecosystem destabilization or the reduction or elimination of species. Food web interactions may be Used to
 interpret the movement of contaminants through  ecosystems and to describe the effects of exposure to variety
 of organisms.                                •
Text JBox 5-2, Qualitative and Quantitative
Risk Estimation
*r     The Pest Risfc Assessment of the
 ;       Importation of Logs from Chile (Appendix
        A, Case A-3) relied on expert judgment to
        characterize the potential for colonization
        and spread of the bark beetle, Hylurgus
        ligniperda. The experts expressed their '
        opinion as a high, medium, or low
        expectation of .damage.
•       In Modeling Losses of Bottomland Forest
        Wetlands (Appendix A, Case A-l), a  .
       .simulation approach (FORFLO model)
        was: used. The hydrologic changes were
        simulated from numerical estimates of
        hydrologic conditions. The output of the
        FORFLO model was coupled with habitat
        suitability indices to estimate biological
        changes. The results of these simulations
        were discussed in terms of qualitative
        theories of ecological processes as; well as
        quantitative relationships;
                                                        113
                                                                                          10/13/95
-------
                                                             '
                                   • i'i	! / a1;	Jn .mi.
  1
  2
  3 '
  4
  5
 '6
  7
  8
;  *
 10
 11
 12
 13
 14
 15
 16
 I?
 is
 1$
 20
 21
 22
 23
,,24,
 25
 26
 27
 28
 29
 30
 31
    ,; ;  ,« „; i -< •, / ";P!RAFT--DO|| NOT C^UOTE CITE,	OR DlgTRIjBUTE
  Regardless of the estimation method"used, the risk as'sessof shbuIS "adhere" tolfieprinciples fistecl below.
  Accuracy of the data is as important as the integration method used.  A simple approach that uses
  measure;d exposure and effects information is more credible than a complex assessment that uses
  estimated or extrapolated exposure and effects information.
  Whenever possible, the risk estimation should provide a range of risks rather than a single-point
  ""•  :• •  lIrT:i!H'  > "•.''. i;::.fif s,:"]"'„ ;uwtf,&.v.;S''lW	                .                  .  	     .      .         ~  '
  estimate. A distribution of exposure and effects offers a vast improvement over single-point estimates.
 ;" .„"'<  „' -: :  MB ' In'•"' :•' ",:'.;„;;,>:•1:),;	i tBvj! 'Y l:ill	i : ,,„',! jif:| 'IB1 '.:;!, i;j( "ii II1": | ]^t,A 'VSK^ .}(•, t,'«fV,, \M« [Miil^^^             Ilii                        !»ifiiii|li:Fi;i:iiiii|iiii:iaiiiiiuiiiiiiiiii,i!iiiinijipiiiiiiiiiiiiiiif i;;liii
  later in this section, some methods only provide a single numerical figure.  Although useful for
 ';:  ii;"f:!; 131 :f?::f\;:)::#ft y$	>!<	fi	,   	i^mm	              ,      ,  ,     	,
        .•   r t  _  /•       i*-ii   • i   	i_  • _ j_i	j_ i	A1	't;	fi._i'	  t£*il	
     i1  , ' -I, > :, ">,H|ii| 	|";, .. '• , 'i:	' ,••,:»,: 1 l>ir>>>, .visl' 'jilr (>'!',•• i            	            .     .   .     T
      separating  large from negligible risks, such methods have their limitations. If these methods are
      supplemented with approaches such as population or ecosystem modeling, the.risk assessor can prepare
      more thorough analyses about the impact of the risk to the assessment endpoint of interest.
 •    All methods should be carefully documented, particularly if a new method or approach is being used..
    ,
-------
                           .   DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      such as an LC50 or LD50), adjusted by uncertainty or modifying factors (U,S. EPA, 1984; Nabholz, 1991;
 2      Urban and Cook, 1986).  The magnitude of the uncertainty factor is usually inversely related to the amount
 3      and quality of the data available to the assessor.  The higher the quotient the higher the risk.
 4           The interpretation of any quotient depends on the exposure and effects estimates on which it is based.
 5      .For example, a quotient that uses an LC50 for the effects point estimate will be interpreted differently from
 6      one that uses an LCOI. Similarly, a quotient based on a typical exposure estimate will be interpreted
1 7      differently from one based on worst-case estimates. For this reason, the assessor should discuss the origin of
 8      the estimates, the scenario or group they are intended to represent, and .their associated uncertainties (see
 9      section 5.2.4 below).                    ,                                „
10          .'-•'•.-     .    '-'        -
11      5.2.2.1. Single Value Quotient Method.
12           The quotient method has several advantages, but also has limitations.  The principal advantage is that it
13,    is simple, quick to use, and risk .assessors and managers are familiar with it, particularly for screening-level
14      risk assessments where both effects and exposure data are limited. The quotient method provides a quick,
15      inexpensive means of screening out situations of high or low risk, making further, more resource intensive
16      evaluations unnecessary.  Under .these circumstances, the quotient method provides a consistent approach for
17      decisions that.have to be made on a wide array of chemical stressors.
18           The.limitations of the quotient method have been discussed at length by many authors (see reviews by
19      Smith and Cairns, 1993; Suter,  1993 a).  Some of these limitations are summarized below.
20      •   It has little predictive capability. The unit ratio"is used to establish a finite limit on the tolerable
21      _     concentration of a chemical. If the ambient concentration exceeds the tolerance limit, it is assumed that
22           there is a  100 percent chance of adverse effects occurring.
23      •   Because the method utilizes the results of standard single species toxicity assays, the measured endpoint
24           (e.g., LC50) may not be appropriate for the assessment. LC50s are commonly'used because of the degree
25           of certainty in estimating 50% mortality.  However, the LC50 may not be protective of the ecological.
26           component at risk. The magnitude of response (LC^-LCjo) may be as important as the endpoint (e.g.,
27           reproductive effects).
28  '•   In some cases, the toxicity measurement may have been modified to account for variability arid lack of
29           knowledge about the effect. It is important to ascertain the full description of the toxicity endpoint, and
30           to carry this description into risk characterization.                                 .   "
                                                        115
10/13/95
-------
                                     " '.!'• -*iViiiirii <>'l"""          '                 ^  (• in n| p11 ii 1 in  11 ii i n n| (
                                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                                                                                                  	lull	i	ii
      :' 1
       2
       3
       4
       5
       6
       »!ii
       7
       8
       9
      10
      U
      12
      13
      14
      15
      16

      17
      18
      19
      20
      21
      22
      23
      24
      25
      26
      27
      28
      29
      30
       Secondary effects are not readily evaluated by the method. For example, a particular risk assessment

       may have fish populations as the assessment endpoint of concern for a particular chemical.  The
       stressor-response profile reveals that the chemical is more toxic to aquatic invertebrates than it is to fish,

       and the quotient method confirms that there is a risk to aquatic invertebrates. Without the use of other

       methods, such as ecosystem models, the risk manager may not be able to relate the effects on aquatic

       invertebrates to effects on the fish populations.
       :•'•', ;' .;	! ail1;.. i i', i.:;,;, • •!";» •.!»', -IS,:	I ?,I'!';1 *, I "li'i	fi	• &»•, •;	j«,«	ii.«,;~»	• •	j;•«	•	i	,;	st-	,; i
          	    hi mill  I ",  ;  '„';, ii! < "",'1'!', "i	' '! HI"!::,''I i ill , "':<	"I! Ir '.'il" '	 Illmllillllii1 .li,,!.',1" li'll!1,,,,!'",!'!" illl'ill* I	i	ill, "I	UlllUlillilMPiiIPlilli	I iWl!%il!  :: ;:; !:';•'?  if,	li!'''"i:!'"(	C*
       this method.
, i •    1     "   lifiiiwll'l ''' i 'Li i ' iii iifnn ' 'i1'1 irk	 .I1'! i"'''' ii i"	ii' • «iil •''!'!'
  5.2.2.2.  Distributions for Exposure and Effects,

       Instead of using discrete values for the

  exposure concentration arid the toxicity data,
,, Isl jltlil'lllli1 11 M"1!1']!! .....
  distributions of exposure may be compared to

  distributions of toxicity data.  Wiegeft and Barteil

  (1994) discuss the use .of joint distributions for
  11  ..... ' i i ..... ',:,'»  ' Will , i ,| , • • ' '; ,''„; ,„.;', "., •„,,:"':'• i, » ||, *• I!' .<" kuLll'-iil;; ''IM11'!.!' ...... IIKIIi'illlli'HI1!!, Iliull11' ml; ij|l«i I'll ill uL
  estimating the risks of metals to aquatic life. Also,
     ''  ...... ' 'Ii,   ISICIIlli |.,l ..... "', „, ."'"i;11 i11!,!, ,:1i:\: pF,,1'!! <„, Vl|l,!, ......... 'lljfl;1 ...... 1'll'lllill.':!!!:;.!.!,!!,, ...... ,1,:,, Isl jltlil'lllli1 11
  the Dutch ^Government  has used joint distributions

  to evaluate risks to aquatic life '(SETAC, ''l994aX

  Their system attempts to identify the most sensitive

  5% of the aquatic populations through the use of

  chronic data and NOAELs.  Similar approaches

  have been recommended for evaluating the risks of
      i ,       i  in              i     ii  i °
  pesticides to aquatic life in the United States

  (SETAC, 1994a). The use of effect distributions^

  generally requires testing of several species for a '

  given chemical or access to data bases for chemicals that have already been tested.  These methods are useful

  when the assessment endpoint is the protection of a wide array of aquatic life, as opposed to a particular

  species.
Text Box 5:-3. Going Beyond the Quotient
Method

        The: EPA Office of Toxic Substances uses
a Probabilistic Dilution Model (PDM3).  This
method estimates how often, over a one year
period, a pairticular effect concentration is exceeded
(Nabholz et al., 1993b;  U.S. EPA, 1988b).  Thus,
if a concern concentration was based on a 96-hour
or 4-day test, a risk would be identified if that
particular concern concentration was exceeded four
days or more (not necessarily consecutive).  The
four day limit may not be appropriate for all
endpoints. For example, mortality may occur in the
first 24 hours, and subtle short term effects may be
missed if the, resylts are only tabulated on the
fourth day.
        The same approach would apply to chronic
tests, which typically run from 28 days or longer.
The PDM3 model is conservative in that it assumes
instantaneous mixing in the water column and no
losses due to physical, chemical, or biodegradation
effects.
                                                                 116
                                                                                                 10/13/95
'.',.'•, *'  .•"!)'•. tail...	t-:i:	'!	':	iMtii^.^il^LiiH
                                                     iillii||||||||||iilllll«llllliilM	             -	«	                 i	i	
-------
                                 DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
    1   ,         Parkhurst et al. (1981) combined several exposure models to determine the percent exceedence of acute
    2       and chronic toxicity values to catfish for the pesticide toxaphene in the Yazoo River. This kind of integration
    3       method is intermediate between single-value comparisons (i.e., the quotient method) and the comparison of
    4       joint distributions.  Another empirical approach is to use regression analysis to establish a relationship'
  .5       between cause and effect. For example, nutrient levels can be used to predict  productivity in aquatic
    6       ecosystems.       '.
    7   .         Regardless of the method that is used for estimating a quotient, it is important for the risk assessor to
    8       consider the questions listed below.       ,
    9       •    What is the relevance of the point estimates used in the quotient?
  10       •    How does the effect concentration relate to  the assessment endpoint?
  11"   •    What extrapolations are involved?
  12       •    How does the point estimate of exposure relate to potential spatial and temporal variability in exposure?
  13       •    What are the critical assumptions or facts used in calculating .the point estimates of exposure and effects
  14            and the quotient?                 .                                                    >
  15_       •    Would it be appropriate to provide a range of quotients reflecting different assumptions (e.g., average
  16            Chemical concentration, high-end concentration, etc.)?
  17    '   "    '                       '  '          .               ""              ' •'.             '.•"•-.
  18       5.2.2.3.  Physical Models and Field Surveys.
  19            Physical models provide physical analogues of a real system (cosms) and/or representations of a real
  20       system (field-scale experiments). Although  physical models are described in the analysis phase of these
  21       guidelines [sections 4.2.3.1 and 4.2.3.3], they may also be used for risk estimation since they integrate
  22       exposure and effects information.           .                                   .
  23       "     Field surveys involve the collection of actual exposure and effects data at locations of concern.
  24       Frequently, field survey data are used at hazardous waste sites because of lack of predictive data ,on
  25       bioavailability, bioaccumulation, subtle biological effects, or toxicity due to exposure to soils, sediments, and
.  26       air emissions.  Results from field surveys may be expressed as: 1) descriptive results (observed defects); 2)
  27     s  statistical correlations of effects with elevated concentrations of stressors, 3) food chain effects extrapolated
  28       from tissue residues; or 4) biological indices of community structure or function. Interpretation of field
  29       survey data should consider the statistical power of the experimental design, .the possible influence of factors
  30       unrelated to the stressor(s) of concern (e.g., natural variability), and whether there is sufficient  causal
  31       evidence to link observed effects  with particular stressors.
                                                            117
10/13/95
-------
  1
  2  .;.
 ,3  "
  '4  ;
  5  ',
  6  ;.
  7
  8
  9
 10  ;"
 11
 l?
 13  ''"
 i4  !1
 is  ;;
 16
 17
 18   '
                        DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
              111 in i             i  i     i      ^                  i   11  p	nil	i	i	r i ri	i	n	ii	i	nun	i	i	i	i	T11 T   i
      Physical models can provide a critical "reality check" on the predictions of theoretical models and are
             II IIllll  I   '  i        II      I I   III II '     111    II 11  I 111 111 111 1111 llll III                           	,	....	;	    I
 more representative of type kinds of associations and interactions found in natural systems than single species
                                                                                                                   iiniiTii|i	inn
 toxieity "tests™	Nevertheless"^" the" risk 'assessor' sh'ould keep in min3 potential deficiencies such as expense
 (compared with simulation models), lack of representativeness or the system at risk, and, especially for
 cosms, difficulties relating the spatial and ecological scale of the test
•'• '  • i ;•	;"I::; VII, 'l>!'=:':,(;':?;	:;; H!:f:;fl/'. ^WWii'1^:-!*'.*	m	MmXfflKffi
 and Bartell, 1994).
:	 . '•  ,'•' . •' '• '  	lliilll  :  '	 ,  viV

                                                                    iii w n»«»^^                    	•"iiiiiii ijiiiliiiiiin

                                          r!, •,:;	vi	" .irriK	ii	liiiiiiBiiiiiiit iiiiiiib^^^^^^^^^^ 111	itiiiiH^^   )«,„,< i:::i	i:«r	fiiiiu	.SIIK;«
 5.2.3. Simulations
      This category includes methods, typically computer-based applications, that are designed to assess risk
                                      ..!'!i. 'Him,' :!hniii" ,.""1,1,! •<''«i	,:„» i,1,,. "
 to populations, communities, ecosystems, and landscapes.  Wiegeit and Bartell (1994") describe these
 simulations of processes or mechanisms as:
             liiliiill! ;';,i,,!,;'::i', ;" j J , ,;""';i1	:' ;i ,, »: m\ ill' '," ^ jlii:!" „;!", [if,," ii!""	iir'lliiiu
       "... an attempt to mathematically represent the physical, chemical, and biological processes that
     „":. >" I/!   .         	   .          	  .                           	ill		i	,	,r	  	in,	
VL.i,: if1! \:j|liliii.,•<";!>:;,	i:;:'!':	s!;i»::Ifl	,K	':!	              ,	H	•	 ,	,	IT	„	,	,	i	Ii	
  determine the dynamics of ecological systems and to formulate the lexicological processes that translate

  stress into response."                                                    ...
 In contrast to empirical approaches, simulations include statements of causation.  Simulations are also
 amenable to predictions.  Their predictions are less restricted than empirical models with respect to the range
 of stressor magnitude, frequency, and duration that may be simulated.
      An overview of the various methods within this category is presented by Wiegert and Bartell (1994),
 ">   ">:::Y:::/; !IIV::rt'^:::'!;y(:°^^^^^^^^^^^^^             	      .             .       .      .                      ,
 Suter (1993a), and Bartell et al. (1992).  (Additional discussion of simulation models is presented in section
   ,', "! ' ' ,:, ,,:! ) " .,!	llylL,, 1	';, 1 :,. 	, >>l":i	.;!' ;,: t •	:("'. iii: ,j •_ ,i,,i»i! 'pj;:i.!!'JIBlK .'» fill-,I I'tsi ffilraHHwlltKIMi'HBHinHnK'L^	^^KSStSSSSSTSSSSIi^StKSSSSSSSSi
                                                                             :h the model itself but rather
                                                                             i'ii-a^^^^^^^^^               	iiEi
22
  .
2/
24
25
26
27
28
29
30
•3 1
         ick of knowledge cegarding basic natural history"data for many species of concern and incomplete
         •;"!' '• ' fiiiil '«"';'".->:" V •:'' f":: !•>'.'! ii' -i; ;!!*. ,::;i i i1 "•!i>i-
-------
                       .        DRAFT-DO JSTOT-QUOTE, GITE, OR DISTRIBUTE-
  1      5.2.4.  Uncertainty Analysis for the Risk Estimate                      .    :
  2           An important counterpart to the risk estimate is a, description of the associated uncertainties.  Readers
  3      are referred to the discussion of uncertainty in problem formulation (section 3.7) and the analysis phase
  4      (section 4.1.2); effective treatment of uncertainty in these phases will make description easier during risk
  5      characterization (see summary in table 5-1). The objective during risk characterization is to review and
  6      summarize the predominant sources of uncertainty in the assessment and the approaches that were used to
  7      address them,  hi particular, the assessor should consider the points listed below.
  8      •    Describe how variability was characterized. For example, if the quotient method was used, describe
  9           what group the quotient is intended to represent.
 10   ,   •    Describe how measurement error was characterized.  If the quotient method was used, describe whether
 11           the point estimates are means or upper-bound estimates.
 12      •    Describe how extrapolation uncertainty was addressed. Identify extrapolations that were addressed
 13           using assumptions (e.g., field response is assumed to be equivalent to laboratory  response).
 14      •    Identify which parameters of the assessment have the greatest impact on risk.
 15      •    Identify which uncertainties have the greatest potential for reduction.  '                            .
 16                                            . -'        •                 -                      .
 17      5.3. RISK DESCRIPTION
 18           The risk description should include: (1) a summary of all the risk estimate(s); (2) a discussion of the
 19      evidence supporting the risk estimate(s) (weight of evidence); and (3) an interpretation of ecological
 20      significance of the estimate(s). The risk summary should provide a brief review of the conceptual model   .
 21      including a rationale for selecting risk hypotheses that were or were not studied and a brief synopsis of the
 22      critical assumptions, limitations, information gaps, confidence, and variability.  Boundary conditions may
 23      have been refined, focused, or mpdified-during the analysis phase; any  changes should be clearly articulated
 24      in the risk summary.  The description does not have to be quantitative.   Professional judgment can be used to
 25      combine inferences based on accepted ecological theory and practice with supporting evidence to prepare
 26      risk descriptions.                          .-'.-.
 27           In addition to describing the risk estimate, the risk summary should include:
.28      •    a brief explanation of the stressors, levels of biological organization, or  geographic areas that were
.29           specifically excluded from  the review;
 30      •    the types and quantity of data sources, reviews, and databases that were utilized;
 31      •    a brief discussion of the key issues from the reports or data sources used to make the risk assessment;
                                                         119
10/13/95
-------
'""I!!!1'!,;.'"1:  [I PI1'    '!	U   "i
                                                                                                                                                                                      Ill    II       II   II   III III III  11   II  III   III  I  III 11 111 II   Illllllllllllllllllllllllllllllllllllllllllllllllllllllllllll
1
" 2
' ': 3 '. '..."
,• • 4 ,
5
. 6 " 	 ;
!'" "'
' ' ''»."'' il
, 7 ;, 	 ,
8
9
10
"'II
'^.,'J2 !',
13
';/ '; ;"•*
:: • , 14 :. : ,
	 .' •. ;
i'j'15 ^
i i , '!i,,
:-;;l6 „;"",„
17
18
19
20
21
1 •' ',! I''* ;
'-:1 1 . !" •! 	 I.!':'
llil II ill 1 U 1 1i i i ii 1 1 i N ii 1 iii in i Inn 1 1 i Hi ii UK UN inn i in i i •
III 1 1 1 1 1 1 III 1 1 KM III m 1
.{i i ' i , 	 ,' i ii i 	 '"ir i''!n in 	 (ii1! 	 11 i 'HI 	 ill in1 iii 	 a." itiiiiiii 	 HI 	
DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
« i i n i i i i n in i i i i
Table 5-1. Uncertainty Evaluation in Risk Characterization
, ' ,; ' ^ 1 ' II 1 II 1 1 1 Hi nln ! Ii L 1 i 11 1 i 1 II • hull II » h III!
Source of Uncertainty1
Unclear
Communication
Variability
Lack of Knowledge:
Model Structure
Uncertainty
Lack of Knowledge:
Extrapolation
Uncertainty
Lack of Knowledge:
Measurement Error
Simplification
Human Error
Example
Healthy populations vs.
Populations with
individuals that can
survive, reproduce, and
grow.
Differences in species
sensitivity within the
aquatic community;
variations in weather
patterns.
Choosing the critical
scenarios of exposure and
effects in conceptual
model development.
Difference between
responses of laboratory
rats and field mice
Uncertainty in the
chemical concentration of
a soil sample .
Use of long term average
exposures to compare
with chronic effects data.
Mistyped computer code
Risk Characterization Phase Strategies
Describe risks in terms of the assessment endpoint..
Describe how variability is reflected in the final risk estimates.
Distinguish between uncertainty that can be reduced with further
data and that which cannot.
Discuss the strengthsrand limitations of the models used. If
alternative models are plausible, discuss implications of their use.
Identify key assumptions; describe approaches used and their
rationales. , • •
Describe how measurement error is reflected in the final risk
estimates (e.g., by using uncertainty bounds).
Discuss key aggregations and model simplifications.
Describe steps taken to minimize human error.
m i mini i " • • i i n in " i n i i " i i nun i i in i n 1
1 The use of these terms is discussed in section 1.5.
: 120 , 10/13/95
i 'i II1! Ill II ill i ' " ' " .
•
-------
                              DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1  -     •   a discussion of the incomplete knowledge and absence of consensus concerning scientific phenomena
 2           that were evaluated in the risk assessment;
 3 -      •   identification of the major data gaps and, where appropriate, an indication whether gathering particular
 4           data would add significantly to the overall certainty of the risk; and
 5 •      •   an indication of where scientific judgments or default assumptions were used to bridge information
 6           gaps, and an explanation of the basis for these judgments/assumptions.
 7            -       -                        .      ._•"''.
 8       5.3.1. Weight of Evidence
 9           At this point in the assessment, the assessor may have several lines of evidence; for example, the results
10       of a quotient risk estimate, a field experiment, or an observational study.  A powerful approach to increasing
11       confidence in the overa.ll conclusions of  a risk assessment is to combine these multiple lines of evidence.
12       Each line of evidence should be evaluated to ascertain the reliability of the information.  Rather than simply
13       listing all the factors which support or refute the risk, the risk assessor should carefully examine each factor
14       and justify its inclusion in the risk summary.
15            the evaluation of each line of evidence is commonly determined using professional judgment based on
16       the following, consideration's.                          .                        -    '         .
17       •    The relevance of the evidence to the assessment endpoint.  Often, lines of evidence that are directly
18            linked to the assessment endpoints  have greater importance.
19       •    The relevance of the evidence to the conceptual model.  When evaluating ancillary evidence, it is
20            particularly important to evaluate the purpose for which the information was collected.   Some-lines of,
21            evidence may be particularly useful in verifying parts of the conceptual model. For example, biomarker
22           results may confirm that exposure has occurred, or field observations may demonstrate consistency with
•23      ,      model predictions.
24      •    The relevance of the evidence to the exposure scenario of interest.  Lines of evidence that are most
25            relevant to the exposure scenario of interest are given greater weight (all other factors being equal).
26      •    The confidence in the risk estimate or other information. Confidence in each risk estimate is a function
27           of the reliability of the information entering into the analysis and the competence of the integration and .
28.          translation of this information into estimates of risk.  Uncertainty has been discussed throughout these
29       ,    guidelines in sections 1.4, 3.7,4.1.2, and 5.2,4. These same considerations also apply to evidence other
30    .       than risk estimates (e.g., field incident reports). In particular, when evaluating lines of evidence, it is
                                                         121
10/13/95
-------
   1
  2
 -3
  4
  5
  6
  7
  8
  9
 10.
 11
 12
 13
 14
 15
 16
'.17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
                    ;  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
     Interpreting ecological significance can provide the risk manager with some context,for comparing the
risk estimate with other cases', incidents, or regulatory .actions.  Evaluation of ecological significance takes
place during problem formulation arid risk characterization. At the problem formulation stage, the risk
assessor relies on ecological significance in selecting assessment endpoints and when defining the geographic
and ecological boundaries of the assessment.  These boundaries may be modified during the analysis phase
of the risk assessment.  At the risk characterization stage, the risk assessor describes the significance of the
risk estimate with respect to three primary aspects an effect:  its intensity, scale, and where appropriate, the
potential for recovery.  Ecological significance should also be considered in light of natural system variability:
While the considerations used to interpret ecological significance may be quantified where possible, in many
cases they will involve professional judgment.
                           •   .  _                  "      .  '       i      .
5.3.2.1.  Nature and Intensity.
     Intensity refers to  the severity of the anticipated risk.  High intensity could involve an extremely severe
effect such as acute mortality or widespread loss of wetland habitat. Intensity of risk depends both on the
stressor and the ecosystem upon which the stressor acts. For example, a certain change in temperature that
has little or no consequence for a temperate estuary might have severe consequences in a coral reef, where
organisms are less well adapted to temperature fluctuations.     '

5.3.2.2.  Scale.
     Interactions on a spatial and temporal continuum need to be considered in assessing significance.  There
are at least four spatial  factors that need to be considered in regard to ecological significance: absolute  area
(km2), the percentage of the landscape, the extent of landscape fragmentation, and the role or use of the area
at risk within the landscape. All of these spatial factors are important when considering the territorial range
and refugia of populations. Linkages between one or more landscapes are important because they provide
refugia for affected populations. It is also important to consider whether there are adequate corridors for
successful migration.
     Habitat designations for certain portions of the landscape can be a critical element of determining the
significance of the risk estimate.  For example, in river systems both the riffle and pool habitats provide,
important microhabitats that maintain the structure and function of the total river ecosystem.  Stressors  acting
on part or all of these microhabitats may present a significant risk to the entire system. The percentage  of the
landscape that is at risk and how it relates to the territorial range and refugia of the populations and
                                                          123
                                                                                           10/13/95
-------
JI1:.!	•"''	• !« •'' PII'IIT111:	r:" ">' !II = ,:;M Wi	^illli	IHTK'S'l	nC^fWII	1«!!HBBI	il W.'WH
 '!';"  :• ,	' „ » >•'  „ i:"  '" (i i:"	  • • ':•; i'  '" iiii ! :!, ..i	i'1[' !!!ij m•"!, 'i;	!f«'' r&lftfW. 1- «;	gj                            •

          1111         '     , :  1	ill  I—    "	 ,  '"'i  L/M'II,; : 	,/l',i	", I'lill'iS,'I1 "I'l' ;!':]	I!'! il1,ii,|,^l|i|l^1iH^^
           .,..   '   ' ,  ,    '''f^,;;';\:,'ft!^^	
       1      communities are of concern! For example, in the forest or river, is there an adequate stock available for
 ;     2      recruitment^ or are there satisfactory linkages fietween one or more landscapes that provide refugia for,
      3      affected populations?  Are there adequate corridors for successful migration? Often overlooked in aquatic
      4      systems is'the secondary'risk' to migratoryspecies (e.g., the secondary risk to migratory species (e.g., anadromous
      S      acceptable uses of certain portions of water bodies for specific times of the year.
      6            Ecological significance changes with the area affected.  A  larger affected area may be: (1) subject to a
      7      greater number of other stressors, increasing the complications from stressor interactions; (2) more likely to
      8      contain sensitive species or habitats; or (3) more susceptible to landscape-level changes because many
      9      ecosystems may be altered by the stressors. Nevertheless, it should not be assumed that "large" means
 •, ' 	-. •,:  	:!;> i •' 1 •;  :',•";. :i  :' •/" i"»  iff. j • & •,:"; f, ;„ - f:' IV:":; E	,•:" „" lift' V,	:;ii	:! iffiF- "'<      . i;K^                   .              .          •
 •     10      greater risk.  Destructionibf small but"unique area's",'such''as"pothole wetlands in a desert environment, may
          1 j, i  . -I.'   "' i ,   ,    i .iiii ', j;';,.	-1.,'', •	i,"	' ii i., nr1;	' ii -"in!	c;it,!	'•	ac'i,1':,, !-H"r.»,i uuri!	•/ Mi"	if	iiii	MTWv!	IE s=ii"^f v^f xsr^z.-^zszs.'zzzxxs.	:=::=^^^^
      11      have devastating effects on local wildlife populations.                       •            •
      12            The temporal scale for ecosystems can vary from minutes (photosynthesis, prokaryotic reproduction) to
      13      days (dissolved oxygen declines).  Changes within a forest ecosystem can occur gradually over decades or
      14      centuries and may be affected by slowly changing external factors such as climate. \^he"n interpreting
      15      ecological significance, the risk assessor should consider that the time scale of stressor-induced changes
      16      operates within the context of these  multiple natural time scales.  In addition, temporal responses for
 ;-    17     	ecosystems may involve intrinsic time lags such" tfiaf'a^verse'resp'orises	from	a'ltres"s'or""may "Be	gefay"e'3_ 'fn'{s
      1,8      is important when distinguishihg the fong-ferm'implicTs'of'a stress  from the immediately visiBle effects. Thus
      19      caution must be used in ecological risk assessments to ensure that important but time-lagged adverse effects
            i  •>'',	'	i,	•" ft i ," ilipiil	:i,	'I'V;",/ ••i',v. irtji'jS	I1:1: /'•! •, 'i; ji' fsi/i'i	{Kirran^^                                  •                 •
 1   i1'1 ,i,   ,    I,  , n  '"'  1,1,1 I1   'I	H !!; I,,, ,  IJlllWll J ;, , ,,„ \ iii,, „ ,:	M Ililliiv1 j!,;,,' ; "iii1 „ l,i,!!1" Jill' i';,i»i|i-jslllliflillij^                                                  	i?i:il|il|ii||i|N::iH9^^^        	
  1   20      are considered For example, visible changes resulting "from eutrophi cation of aquatic systems (turbidity,
 :  ,; '  " ' Lv: .I  '"]*:i^.:>^y 'lil'i'fr'1''1'!'':1'-1-	P	]l	m^KM,	"ill	!:!%•!":«	           	                                     . . .
     21      excessive macrophyte gro\vth, species decline) may not become evident for- many years after initial increases
          ,      '  , '"!''   i'" ii!,1' ' '  f'liff'Uli	':,,	",!•!,• :	" i,,  in1 „ ri» •'"»[	<',':	, lli» u\ \ 'i::t\'"m\,\K,::«v «'Mi '''I1 lih'iS'itih1:1: l,">!i;il,li|i|i!,i	ifJiiiiHiiiJijiiLMiiHiniii1!!,:1:	ijii'illinilllllllililliiiillillllllllil'iillliililiniiilillllliliiliin^    	IlliliiliiiiliilllillillU1; n,r'iiiiiCIIIII'iiliJlliiliSiiSlllilHllllllllliiriy     "I1"!!!;:!!!!!!!:!"1!!:!!!1!!!!:::!::!!!!!!!:1!!!!:!1!"!!!!!!!
     22	  in nutrient ieyejs.    ,..  ,,	,	 	.'.	:,	, „,	'^.	.,.	'' 	  '„,'	',"	'	'	"	'	".	'„"'	'	'	'„"'	'	'	"	'„"	",,',i"
 ',''..,; 23  ;.'  ",   ',   	    ,;;',;  ,;;,'"	"	'"	;	;\  ;;;	;;;;'"..	"!'.;•	                    iz."i.;z„ .'„ 1".	7~_~,.!:^iiii'i
     24      5.3,2.3. Recovery.     ' ' 	'	  "   """"""'	"                              .       ,       '••.>.  '•
     25            The risk assessor should recognize the dynamic nature of ecosystems (Landis et al., 1993) and realize
  '" 	I,' ,'    ' ,;/, i';  '  ,  - "' ' 	'  lilliiiiiiii  liiM,,:"'-  'f.V'iii'iiT'llT  ill!''JS,"" -iiiii Vl,iij1"»ir*ISlilji;rn'*'fiu:i«iiwii  I 1111111)1  IIII III II 111 111 ill III III 111 III 11 III III11  II I 111 I I I I I 111  1111 11 III II 111 11111    .       '
     26      that ecosystems constantly change as systems undergo succession or respond to long-term changes in the
 •i   , ,i   „'; ,ii:,,ii, '  '   i ,  	ii111  Siiii'i	 i1!*" 'in ,i ^i^i'V^WhiJ^                           I 1 nil in in i in in in ii in                                   T;	i"i	
     27      physical environment (weather, natural catastrophes, etc.). Given this natural variability, sustaining an
     28      ecosystem at some static condition is both ecologically inappropriate and scientifically insupportable.
     29      However, it is still possible  for risk  assessments to provide valid estimates of potential recovery from
     30      projected perturbations.

                            	 "	124.        '   '•'.       '      "  ',     '  '  '     10/13/95
-------
                              •DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3
  4
  5
  6
  7
.  8
  9
10
11
12
13
14
15
16
17.
18
19
20
21
22
23
24
25
26
27
28
29
30
      Recovery may be defined as the partial or full return of a population or community to a condition that
existed before the introduction of a stressor. The evaluation of recovery is made more difficult because
ecosystems are dynamic and will not return exactly to a preexisting state. Therefore, the attributes of a
"recovered" system need to be defined.  The physical, chemical,  and biological nature of the stressor is
relevant to recovery or reversibility of effects.  For example, the  potential recovery of an ecosystem from
exposure to low levels of a labile chemical stressor (ammonia) is likely to occur more rapidly than recovery
from the physical alteration of habitat.                  .        -
     The spatial and temporal characteristics of the components of an ecosystem will influence both the
degree and rate of recovery. For example, the time needed for a fisheries stock to recover might be a decade
of more; the recovery of a benthic infaunal community could require three to five years; a planktonic  >
community can completely recover within weeks to months; while reforestation may take several decades.
The common ecological factors in these examples are the temporal scales of organisms' life histories and the
interspecific and trophic dynamics of the populations comprising the ecosystem or landscape that is
potentially at risk.  An additional factor is the availability of adequate stock within the landscape for
recruitment.
     An ecosystem that has been subjected to repeated disturbances may be more vulnerable to extinction or
irreparable loss of habitat.  Continuous logging of .old-growth forests will eventually eliminate the forest
ecosystem. Aquatic organisms that experience repeated acute water depletion due to dam operations are
likely to experience loss of individuals and ultimately may be unable to recolonize their natural habitat.
Physical alterations such as deforestation in the coastal hills of Venezuela have changed soil structure, seed
sources, and the local physical environment such that forests cannot easily grow again, even after the areas
are abandoned by humans. This phenomenon also was seen in the irreversible loss of the great forests in
Britain during the Neolithic period, when humans cleared land for agricultural production and energy
resources (Fisher and Woodmansee, 1994).
     There are a number of issues to consider when attempting to characterize the ability of an ecosystem to
recover.  One consideration is how to define recovery given that a system cannot return to an undisturbed
state. The following examples from Fisher and Woodmansee (1994) illustrate a range of options that could
be used to characterize a recovered system:
•    When the pools of nutrients in a eutrophic system return to  their prestressed state;
•    When a specific species has reestablished its population at a particular density; or
                                                         125
                                                                                           10/13/95
-------

                        DRAFT-DO NQT QUOTE, CITE, OR DISTRIBUTE
   1
.,'2,,


"'  4

••.  s,"
   6
   7
:,  8,,,
 "  9'
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 2.2
 23
 24

,25,

•''27
 28
 29
      When the residues of a toxic chemical in sediments or in biological tissues have decreased below some
      Uureshold.
     ';••••  .-» ',  iiliiill
                                               WRlte	
      \Vhen considering the reversibility of a stressor-induceH effect, th'e method of remediating the stressor

 should be compared with" the original "effects of tfie stressor.  For example, at hazardous waste" sites tHat have
 received a variety of tqxic chemicals, it may be better to rely on natural processes to detoxify these chemicals

 rather than attempting to remove them and thereby destroying habitat. In some cases, it may be preferable to
: I	   ' '• *  *, ,i,', ' n> • liJilli1' .,(!«'-!! ••, '  :,' »:F 'V .;;: t >-,•;, mi;*'! it vis.vir1:1' iQ/iiHij»i"ir«.*twn #Hh.MKMEHM	iiiiHK'ii1	iiiiiiM                  :  •	EiifiVJBS
 Jeave o^l in gravel beds in high energy aquatic environments rather than attempting to remove the oil by
                                                                                          inlw^^             	"»	iiiiniiiiiiiiiiiiii	in
                                                                                                      )ved
                                                                                                                 I
 but also on how any mitigation efforts impact the recovery process.
,»  ,  :•  "  '"AK ji	win ' i ;..:, ' ,	:'';,*';!, ,,HJ;,T : jjiiii ^L, ,:• ~t 'il,,!' 'li11' ''io^piiiiiih1;'. iniuiiiiw   wnlw^         iiiiiii™^	!iiiii>i      	iniir1	hiiniiiinv
 pressure washing. How well a given area will recover depends not only on how quickly a stressor is removec
   '  >>r! ;  ' ' -a •   lltillt :	i ••,:', i' -i'" .;;'•• "'', It!'!	• •!!" !'. 'iilHi •:;:!:: '3'	hi;*!*! •	!>!:! ftai, !«,!« B!"if SltiiHi	JIH	£i!B^^^^^^^   	iiilii IIIIIIM^^^^
 5.3.2.4.  Natural Variability and Disturbances.

      Determinations of the ecological significance

 pf risks should be considered along with the

 existing condition of the environment.  Natural

 disturbances such as droughts, wind and rain
 storms, or geological upheavals may affect a

 system such that a risk may be more significant

 than.i); would Jigye beenm the,absence of such ip  	;	
•;'   :, ,:'   " ' -   ."IS  'I :"'";'::: „''' i''?'> :.i'!i'K:'''• •'! W{"': WW^i'I1 'i1,!;*'W'ni'ilii
 natural stressors.

      Given  the occurrence of natural disturbances"

 and the inherent variability of ecosystems, it is

 important for the risk assessor to ask whether

 changes in assessment endpoints are
                                                      Text Box 5-4.  Importance of Understanding
                                                      Natural Disturbances

                                                             . The risk assessment for the middle Snake
                                                      River identified multiple physical, chemical and
                                                      biological stressors as the cause of the decline of
                                                      the native mollusks and fish species.  In addition to
                                                      these anthropogenic stressors, the area has recently
                                                      experienced 6 years of drought.  Sampling and
                                                      analysis of water quantity and quality over 30 years
                                                      has provided some evidence of the dynamics of this
                                                      ecosystem  Analysis of these data indicate that the
                                                      initial stress to the ecosystem from dam
                                                      construction, and irrigation withdrawals has been
                                                      exacerbated by the meteorologic conditions
                                                      (Bowler etal., 1992).          .
 distinguishable from the natural variability 'of'the response being measured.  For example, natural
                                   	
                              " I,' If ,i i!1	 '-I:!11,' 11. WP"! :'i<"! i," I'll« I JllliHI il|l|,!,, Ml !',""111':: iliLIP,' >ll ilu ilil'i'iil!! <	IHIillV Hi' .Jll!,;! '"
  1 .  .:   i ' "Ji. , '"  liilEO" , i i  , 'I., 	US'": i,, JM i ijllli' nil; , „ ::< 'lillll! [„!', «¥" < ilu1 ""> 'iPr IB1"'!!,!'! i "i ,i i! . 1, > rl JiiNII'liUli'lllllP''!!'!.!!'!!!, wililliilKHiillliillillll li|i|lllllllllllil(l:HIIIIIII!illllllfflll:!il™         	iii:fiiiii^i:iii«iiiiiuiipi	iiiiiiiiiiiiiiiiiiiiiii:i,i<<<	iiiiii.iiiimiiiiiiiiiS'ifiiiiiiiiiiiiiiir :.iiiiii;:„„",>'	,',,!!':	i;	""M1":,",,!!!:1!!;,'!,,;/" r:,,,,!!,'!1,",!1,	IlifFliH,"	IMI	lliil'iEaiiiJIJia^^^^^^                    	llllKIEB^^^                 	ill	iliuniplllillll
 of historical (time series) data or by inadequate information on life history characteristics and controlling
 r ,   :;C   ."ft	  11!,'i *isr'W'•'•''•	sJ	!1'!!:]H:
-------
                              DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1      6.  RELATING ECOLOGICAL INFORMATION TO RISK MANAGEMENT DECISIONS
  2'          ,.         .-"...-         "                    .             ',..-.'
  3      "     Risk characterization must accurately and concisely eonvey the ecological risks and, equally as
  4      important, address how reliable the risk estimates are in relation to the current scientific knowledge.
  5      Discussion with the risk manager should form a coherent picture at a level of detail appropriate for the
  6      decision being supported.  Accordingly, greater emphasis is placed on ensuring clarity, consistency, and
  7      reasonableness of the risk picture and transparency of the risk assessment process :as an input to the decision-
  8      making process than on reformatting or otherwise reiterating conclusions of risk assessment components that
  9      precede risk characterization.                                ' . _-      <
10           The risk manager should be able to read the risk characterization and know what the major risks (or
11      potential risks) are at some level of biological organization (organism, population, community, ecosystem)
12      and have an idea of whether the conclusion is supported by a large body of data or if there are significant data
13      gaps. It is recognized, however, that sometimes there isi insufficient information to characterize risk at an
14      appropriate level of detail due to a lack of resources, a lack of a consensus on how to interpret information, or
15      other causes, hi these situations,  the issues, obstacles, and correctable deficiencies should be clearly
16      articulated for the,risk managers' consideration.               .                                  ,
17           InrMarch 1995, the EPA issued a policy statement requiring that risk characterizations be prepared "in a
18      manner that is clear, transparent^'reasonable, and consistent with other risk characterizations of similar
19      scope prepared across programs in the Agency" (U.S. EPA 1995d).
20           Clarity may be achieved with' the procedureslisted below.                             '   \
21                •   Brevity is achieved and jargon is avoided.            ,                                ;
22       .         •   The language and organization of the risk characterization are understandable to EPA risk
23           .   - -     managers and the informed lay person.
24                •   Unusual issues specific to a particular risk assessment are fully discussed and explained.
25           Risk characterization should also be transparent.
26                ••  The conclusions drawn from the science are identified separately from policy judgments.
27                •    Major differing viewpoints.surrounding the scientific judgments are clearly articulated.
28                •   The purpose of the risk assessment is defined and explained (e.g., regulatory purpose, policy
29,                   analysis, priority setting).                         '                              j
30                •   Assumptions and biases (scientific and policy) are fully explained.
31     .            :     •  '              •.        '             .  '   .      '••'•.'•-'••
                                                        127
10/13/95
-------
                             DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3
.  4
  5
  6
  7
7. REFERENCES

Adams, D.F. '(1963). Recognition of the effects of fluorides on vegetation.-J. Air Pollution Control Assoc.  13:
     360-362.  Cited in: Suter, G.W. II. (1993a) Ecological risk assessment. Chelsea, MI: Lewis Publishers.

Alder,,H.L.; Roessler, E.B. (1972) Introduction to probability and statistics. San Francisco, CA: W.H.  ,   .
     Freeman and Co.
  9
 10
"11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
-24
 25
 26
 27
 28
Andrewartha, H.G.; Birch, L.C. (1984) The ecological web. Chicago, IL: University of Chicago Press.

Ankley, G.T.; Katko, A.; Arthur, J.W. (1990) Identification of ammonia as an important sediment-associated
     toxicant in the Lower Fox River and Green Bay, Wisconsin. Environ. Toxicol. Chem. 9:313-322.

Auer, C.M.; Nabholz, J.V.;. Baetcke, K.P. (1990) Mode of action and the assessment of chemical hazards in
 .  •  the presence of limiting data: use of structure-activity relationships (SAR) under TSCA, Section 5.
   ,  Environ. Health Perspect. 87:183-197.             .           ...'.-',

Auer, C.M.; Zeeman, M.; Nabholz, J.V.; Clements, R.G. (1993) SAR - the U.S. regulatory perspective.  In:
     SAR and QSAR in environmental research, volume 2. Gordon and Breach Science Publishers S.A.; pp.
     29-38.         ,  '  .                •-                 .   .          .    "  ...
                                        '
Barnthouse, L.W. (1993) Population level effects. In: Suter, G.W. II, ed. Ecological risk assessment. Chelsea,
     MI: Lewis Publishers.

Barnthouse, L.W.; Brown, J. (1994) Issue paper on conceptual model development. In: Ecological risk
     assessment issue papers.  Washington, DC: Risk Assessment Forum, U.S. Environmental Protection
     Agency; pp. 3-1 to 3-70.  EPA/630/R-94/009/                   •
                                                      129
                                                                                       10/13/95
-------
                        DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 Barnthouse, L.W.; O'Neill, R.V.; Bartell, S.M.; Suter, G.W. II. (1986) Population and ecosystem theory in
      ecological risk assessment. In: Poston, T.M.; Purdy, R., eds. Aquatic ecology and hazard assessment,
r...                                                                 PI         _              i          j
      9th symposium. Philadelphia, PA: American Society for Testing and Materials; pp. 82-96.
 Barnthouse, L.W.; Suter, G.W. II; Rosen, A.E.; Beauchamp, J.J. (1987) Estimating responses offish
      populations to toxic contaminants. Environ. Toxicol. Chem. 6.811-824.
 :  • •  „  ',1
 •  ' -  :  2
      ''• 3
".'.''.  ','  4
 •••:  •:'  5
        6

•:'-"'  .  >.ii'   ;:;'i  '''>:'•   ,     i    11,              I  I   I   	i    I	,                          id1 iiiiiiiiiiiii I iiiii mini iii lyiiui in
        8       Barnthouse, L.W.; Suter, G.W. II; Rosen, A.E. (1990) Risks of toxic contaminants to exploited fish
        9            populations: influence of life history, data uncertainty, and exploitation intensity. Environ. Toxicol.
 ;''   '  10    ,i _'ii  '   ii:Chem.'9:297-311._" (  " ["'^^'""_i''_'"['"'ii"""J™. ^'"'"'"	^™™	i	i	'	^	'	i	"i	i	i	i	

        12       Bartell, S?M.; Gardner, R.H.; O'Neill, R.V. (1992) Ecological risk estim
  	;     13    ;;i>     .  ,,  i  _    ^ _'    "	;;;;	'~'~  	;_;;;i;("_^ti^;;^''r'~"t~'^,	'~^.":"	I	,;	~	'~'^~	"^~	"^"""i","!"",.^","	:	i	j	7t	~	']	~,~	
  / ,   14   "',  Bedford,, B iC'-'YrestonVK                     	the	scwntific^asis:^r	a'sse'ssing^'un^latrve1 effects of
  	"   „      i .'i  ,,:'!'    ; .,',1  .. 1!  , i|,  '  'ijijjfijjigjl ! 'in,.'i. ,' '	,"! .< ,',,,iij,;' HII,,, ,i,:||' ...ijVllllii1! ''	Ill, ^l11:!!!!,!11'"'^     K.llKKiiHIOE'lHHI^^^^^^^^^^^^^                        	lilnlW'illlliNLiM^^^^^^^^     	U	H	n	I1"	•	•	I	HI1
        15            wetland loss and degradation tin landscape functions: status, perspectives, and prospects. Environ.
,   i1    ;    l>"i  =!  :.  '"/'v,  i>;'  •'•4i%i;\:!!:''ll!-'i:^     	'?i':i»;iiii	'!7:!.^i}s:;.i!iHi   	fiiii	i»^^^^^                 	                                  ,
        16            Manag. 12:751-772.
   :   •  17	,  „	'	,	,	''	'	;	i	,,i	!	'	!	*	:	ii	'I	;	,	,,i	-K
  '•-',  ;'  18   	Bowler,  P.A.-"iWatspn^C:. ^.ilVearsleyjI'^'rCirpnCP^
  ':, ••/  19   .'.   ,n  	' impact on respiirce'ajlpcatio^
  . i'  • '!i  " •    ii;.;1.   '• i'!'. ", '.*;i';,,'  ||S »''ii,''!*::",i'Mi V:.• lia:'""iii."iii1" >vSlaii'ia.1 T"'i'lS*'1"^IM>Jmwin	14?i^nSffinSaiil:S^SS^^^r^r:^^^;^'^;;!:^!^!:!^^^^^^^
        20            Council Volume XXIV. 42-51.                                .       .
        2i      i,:',  >?'.  '^li';;  illl;:l;;'!:::;.i::::!l^	^J'i/lili^:!!!:1'-'^                   	!;:i^^^^^^^^^^^^^^^^^^^^                         	H^^^^^^^^^^^^^^^^^^^
  1 I , i, ',   <^l<>''l!illl||J||llI||i||||plll||!|i|||•^                     ,iilirl«l^l\|illlliiilliltlllirHllllilllilllI" 'I'llll'llilll'l1 il'iriBIIIIII'iillllilllllBBIIIHIHIIH
        22       Bradbury, S.P. (1994) Predicting modes of toxic action from chemical structure: an overview.  In: SAR and
        23            QSAR in Environmental Research Vpl 2: 89-104. Gordon and Breach Science Publishers S.A.
    .  '  24  . _   .  .  '.  "";    I;	"'  ", J 	;/	^	']	"	 ^  ^^'^'^	™  ^'^	^7Ji'~f    ,          ,                            '"_^	;	I
        25       Broderius, S.; Kahl, M. (1985) Acute toxicity of organic chemical mixtures to the fathead minnow. Aquatic
    •  •      i, ", i  .  .'!	:'• •	i,,1'-:  '•' IJII  i1111!:;:,;'""'":,!*;:'11,,'	ft,,!>['„I!-,	1	>r!!"•')'l	-'',,Ht	tSll;(WPB!mSAfl'^KlttBnilllJ	I111'                                    '	'	
        26  	Toxicoi:6:307-322'.	:	                 _
        27
        28       Brody, M.S.; Troyer, M.E.; Valette, Y. (1993) Ecological risk assessment case study: Modeling future losses
        29           of bottomland forest wetlands and changes in wildlife habitat within a Louisiana basin. In: A review of
                                                                	II
                                                                                                10/13/95
                                                                  IIH   ilk
                                                                                                                 in i mi iii nil I iiiiliiiiiiii i ill)
-------
  1
  2
  3
 '4'
  5
  6
;  .7
  8
*  9
•10
 11
 12
 13
 14
 15
16
 17
 18
 19
20
21
22
23
24
25
26
27
28
29
30
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
     ecological assessment case studies from a risk assessment perspective. Washington, DC-Risk
     Assessment Forum, U.S. Environmental Protection Agency; pp. 12-lto 12-39. EPA/630/R-92/005.

Bin-master,D.E.; Anderson, P.O. (1994) Principles of good practice for the use of Monte Carlo techniques
     in human health and ecological risk assessments. Risk Anal. 14: 477-481.

Cirone, P.A.; Pastorak, R.A. (1993) Ecological risk assessment case study: Commencement Bay tidelands
    , assessment. In; A review of ecological.assessment case studies from a risk assessment perspective.   .
     Washington, DC: Risk Assessment Forum, U.S. Environmental Protection Agency; pp. 5-1 to 5-32.
     EPA/630/R-92/005.                       '                 ,                      ...'..•'

Clark, K.E., Gobas, F.A.P.C.; Mackay, D.  (1990) Model of organic chemical uptake and clearance by fish
     from food and water. Environ. Sci., Technol. 24:1203-13

Clements, R.G.; Johnson, D.W.; Lipnick, R.L.; Nabholz, J.V.; Newsome, L.D. (1988) Estimating toxicity of
     industrial chemicals to aquatic organisms using structure-activity relationships. Washington, DC: Office
     of Toxic Substances, U.S. Environmental Protection Agency. EPA-5 60-6-8 8-00.

Clements, R.G.; Nabholz, J.V. (1994) ECOSAR: A computer program for estimating the ecotoxicity of •
     industrial chemicals based on structure activity relationships, user's guide. Washington DC:
     Environmental Effects Branch, Health and Environmental Review Division (7402), U.S. Environmental
     Protection Agency, Washington, DC.

Cohrssen, J.; Covello, V.T. (1989) Risk analysis: a guide to principles and methods for analyzing health and
     ecological risks. Washington,. DC: Council on Environmental Quality.

Council on Environmental Quality (CEQ). (1986) Regulations for implementing the procedural provisions of
     the National Environmental Policy Act.  40 CFR Parts 1500-1508. Washington, DC: U.S. Government
     Printing Office.,Cited in: Suter, G.W. II. (1993a) Ecological risk assessment. Chelsea, MI: Lewis '
     Publishers.                                            :
                                              131
                                                                                                10/13/95
-------
      ,F""!,i 	 ,;!1!!1!'1, I!'1, ;„	
                                           	Vi^r^.-^^^
  .«( •'.'  .,* ., i,
                                                                                  IfljUi.^^^^^.^!!!
••  10
  11
...... 12
  13
 • 14
   18
   19
   20
   21
   22
   23
    !! ,
   24
   25
   26'
   27
           •'' : '••''. ':''(•   A^:\:''^^M^^^lK^^^
                                           r--M'Not;Qpo^^
           :    	i, „  ,I   S!	ill , •••",  . ':: .:!•:',:•>• >:I; f1,'« ,'F! •;•':' !1 •*•>,:	»i ;wrapWrK5M»rt!W                •	i	,	,	™,	i	i	 	,	i	,	i«	:	-	
           Davis, W.S.; Bascietto,1(1993) Ecological evaluation of a freshwater stream and wetlands near an inactive
                ! colce plant. In: A review of ecological 'assessment casestudies from a risk assessment perspective.
          !:'"'   '.V.	"S1"; "ffl.W'.'i^i'W1'*	It:	-"""'-ii"!	^SiiK	Wffi&m	HWBKItMm	S	'	I	™	!	•	:l	%	?=	5	5
    "'Washington^DC:" Risk'XssessmentTorurnrtl.1^	^nv^onmental	pTc^ction'Agency;	pjT^-'lTo^'-'CS.
     i • ",', iiiii< 'J ' I'l'ijiii "i,,,1':!'1,,!,, i j1 "v,;, '"'i'";1'""'ii1';1 i''"!:' ' iiiii;!!:iiM in. •!!. !i. v i»,,: 'o':,ii:%io'i:!niii:Tsi i'''^;.^!!/'^''^.]/!.';;!*
     EPA/630/R-92/005.
   5
   ""6  ,  •
   7
   8  '':
Detenbeck, N. (1994) Ecological risk assessment case study: effects of physical disturbance on water quality
     status and water quality improvement function of urban wetlands. In: U.S. Environmental Protection  ,
     Agency! A review of ecolbgicaf assessmeni case studies from a risk assessment perspective, volume.II.
               " Washington, DC: •|^sj^lggles'™nFF^^                     	i^tecltr6n'A"geificy'rpp;	4-1 to 4-58.
               ;',; •! T || ,<;; •';  ii||}|| •;,;!};: • „!'<,"" ;j>t '!, •_	it;, : :,l,;1v sji!1;"*! i11-,;!!;; i;1;!!,:,;1 ji;*;if^f*j|!"kiii'i!;!) 'fif ^LffitiJM	'	
' Emlen^" J.M. (|1!i"989) terrestrial popu'falion'm^^                                 	a	state-of-the-art review.
                  , ,in |H,: <,  jjlLUSit ' ,,"" ', 	'" iii, , 	 ' 	"!,'i,",I'',;, 'n,,  , «III	,'i	*,", „, > "jullk I,ii:,|||L;ilh'	"I?J'!'11" i', :'lA'lfcjlil'ljd ll|||liij|h;iIjUlllP!!i'llllllllli'iillJllll'llili'lllilllIllllllll!I,lll'll'I'lllllLllil11!!"!!!!	I11:"!!,	LiiliilLil,!*!	Illll!:!!!	Li:!,1!'I,'i!!!1!!!!!!!!:,,iJ'll',1	1!::'!!, lull!!:illilSf i"'I™I'llll'i'll1," :' ''Illlrt1!1 li!!!1!!!!!:!!!!!!!;!!!!!!!:!'!!!
           Person, S  ; Girizburg,L; Silvers, A. "(1989)Extreme "eventriskanalysis for age-structured populations. Ecol
                  * ':'«•'": <  ;T iiiiiiii:""'". i',,' 11,	«''! ' ii1' '.in"''	"" i ii" 111;, ,i ,i i" K '" ,;:i,' • , •',»	< ,1,'': !i	'i; ,,ii; n, t:  ' m'; ii"' r; >s	?":,">: *.¥•$&' '••. -
 Fisher, S.G.; Woodmansee, R. (1994) Issue paper on ecological recovery.
                                                                                                 .           „	
                                                                                             Environmental Protection
             .            „
. In: U.S. "           '
      Agency. Ecological risk assessment issue papers, Washington, DC: Risk Assessment Forum, U.S.
                               i- illII :!»!!, ' li'-i'yrSllLSi:	'Jliii.liI'LiSillHIiW: 'hl''.!!!''/'!...!'''!!.'!^^!*'!'^!!^.' W'P"'
   , i.i ' j	'•/.- Rill i fn •!, :f,i,::" 'iai-i-i,.-,
      Environmental Protection Agency; pp. 7-1
                                        t'IJ |n,| |>'	M III, i|.: , 'IW"! if ,i • it:.':.!: \m#	i	"i-it	:-m	«[:/:!!• ini	IIK«	i	•<	«	i	5	•	»*	:	=•=	: -,	>	t	a-	« a	,- _ ft , *y—
      evaluating the chronic toxicity of pollutants toMysidopsis bakia. Hydrobiologica 93: 179-187.
            GentileJ.R-J>cot^K" 1; Pau'ClF'; LmthursC'EA'	(K^I'lcblogicarassessment case study: the role of . ^
                 monitoring in ecological risk assessment: an EMAP example. In: U.S. Environmental Protettfori
                                                                                      BililMilrtn     	bil	iiiiilliii	iiiiliiiiilil^^^^^^
   30            Agency! A review of ecological assessmeht'case studies from a nsk assessment perspective, volume II
                                                                132
                                                                                                   10/13/95
-------
  1
  2.
  3.
  4
  5
  6
  7
  8
 .9
 10
 11
-12
 13
 14
 15
 16
 17
 18
 19 '.
20
21
22
23
24
25 '
26
27
28
29
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
   •  Washington, DC: Risk Assessment Forum, U.S. Environmental Protection Agency; pp. 5-1 to 5-40.
     EPA/630/R-94/003.       .           ,       '                     ' .'         '  . :  '  '

Hallam, T.G.; Lassiter, R.R.; Li, J.; McKinney, W. (1990) Toxicant-induced mortality in models of Daphnia
     populations. Environ. Toxicol. Chem. 9:597-621.

Harwell, M.J.; Norton, B.; Cooper, W.; Gentile, J. (1994) Issue paper on ecological significance. In:
     Ecological risk assessment issue papers. Washington, DC: Risk Assessment Forum, U.S. Environmental
     Protection Agency; pp. 2-1 to 2-49. EPA/630/R-94/009. ^      .

Heck, W.W. (1993) Ecological assessment case study: the National Crop Loss Assessment Network. In: U.S.
     Environmental Protection Agency. A review of ecological assessment case studies from a risk
     assessment perspective. Washington, DC: Risk Assessment Forum, U.S. Environmental Protection
  ., -Agency; pp. 6-1 to 6-32. EPA/630/R-92/005.

Hermens, J.; Canton^H.; Janssen, P.; De Jong, R. (1984a) Quantitative structure-activity relationships and
     toxicity studies of mixtures of chemicals with anaesthetic potency: acute lethal and sublethal toxicity to
     Daphnia magna. Aquatic Toxicol. 5:143-154.                                               ,  .

Hermens, J.; Canton, H.; Steyger, N.; Wegman, R. (1984b) Joint effects  of a mixture of 14 chemicals on
     mortality and inhibition of reproduction of Daphnia magna. Aquatic Toxicol. 5:315-322.

Hill,  A.B. (1965) The environment and disease: association or causation? Proc. Royal Soc. Med. 58:295-300.

Houseknecht, C.R. (1993) Ecological risk assessment case study: special review of the granular formulations
     of carbofuran based on adverse effects on birds.  In: U.S. Environmental Protection Agency. A review of
     ecological assessment case studies from a risk assessment perspective. Washington, DC: Risk
   .  Assessment Forum, U.S. Environmental Protection Agency; pp. 3-1 to 6-25. EPA/63 O/R-92/005.   '
                                              133
                                                                                               10/13/95
-------
                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
            , , .ill                        1 1 'I ' i '  ' i ' ........ 1 1 " ii ..... Ifl I III1 illlili'lf'ii  •   ....... " ' " ' ' " ' ' ' " ' ' ...... " |l111"11111" ''' "l"1 '" ' ' """ "' ''' " ' "' "IW " ' ' ll1"1'1111 IK| M^ " ll111 l||l|illini11 I
Huntsberger, D.V.; Leaverton, P E. (1970) Statistical inference in the biomedical sciences. Boston, MA:
     Allyn and Bacon, Inc.

Irwin, F.; Rodes, B. (1992) Making decisions on cumulative environmental impacts: a conceptual framework.
     Baltimore, MD: World Wildlife Publications.
              ^             Jl
                  Kenaga, E,E. (1973) Factors to be considered in the evaluation of the toxicity of birds to the environment.
                       Environ. QuaLSaf. 2:16^6-181.   '
                                                                     	'iii i mil	i	i	1111
                                                                                                         	in
        i
        2
..'.: '    3
        4
:  ' ,    5
 '" 	,    6
        7
4  ' '    8
:1""     9
       10       Kenaga, E.E. (1982) Predictability of chronic toxicity from acute toxicity of chemicals in fish and aquatic         I
       11            invertebrates. Environ. Toxicol. Chem. 1:347-358.
       12  	
       13       KSnemann, H. (1981) Fish toxicity studies with mixtures of more than two chemicals: a proposal for a   .
       14            quantitative approach and experimental results. Aquatic Toxicol. 19-229-238
 '•   ;	15  i(      	u   ivrii]	i  _	(|	_	;r	,	•        t   •
    ,   16       Landis  W Gl; Matthews, RIA.;" Warkiewicz^AT; Matthews, G.B.  (1993) Multivariate analysis of the
 ii;•  ••':    i.  '"•';:   . •:; . '*> ^ '" •••• Ilig •'         .     it'- 4Pi &[TOi?iB8HfeHKi   . .  ill ill (l >     .	  ii.   I ill il i' "i 111	in (i n	 f niiiiiiiiiiiiiiiiniiii i iiiiiiiiiiiiiiiiiii'iiii'iiiiiiiii
       17            impacts of the turbine fuel JP-4 in a microcosm toxicity test with implications for the evaluation of
       18-          ecosystem dynamics and[risk assessment. Ecotoxicol. 2:271-300.

 ,  . 	20       Leibowitz, S.G.'Abbruzzese, B/Adamus, P.R.; Hughes, L.E.; Irish, J.T. (1992) A synoptic approach to
       21            cumulative impact assessment. Corvallis, OR:  Environmental Research Laboratory, U S. Environmental
       22            Protection Agency. EPA/600/R-92/167.
  ,,    23,"; 	^     |	^   i	    ,	,;r,ii;;,i	„,,„„	...^	.,	,..4.	h.	(	,jf>i	
" "'•;; •,;   24  :	„'  'Upton",'!/.; Galbraith,'!!.;" Burger, jr.J;'''vVartenb'erg,'	6"	(i"5§3")	^'paraHt^Tor	ecbTogicaf	nslc'assessment.	
       25            Environ.Manag. 17:1-5
       26                                   •         '     '             '     "   '•''••'	   '•   '' '   '   •'•
       27   ,   ,    ,     ,'.    '	I	„ . .'^ i  .'.
       28                      ' :             '  _                          •..       '  ^
      ' 29'      „    ,';    „:;,  '',""    	;"";	;"';'';„,' „.;'';; •;;..;;.,;„:;, •;.: :„::,:;i,:,,;:	:::::,!::::::;„:;„:	:,:	:::,,:	::,;',;:	;	;::::
           ,,-  :       <  ,    jjiiiij, ,	;  ; . , '	 ""r'rr,;1;',^ 'IE.,  '; i: .1", vi'f.ji;	::'	ni:,")11!::.	,,	^Li;,!-!,;.::!!':!:;^	Hiiiiii	>, iHiiiii  f t	•.,«V "v •,,	, i	••,;:."!;,:, :•,; ai-i!' if 'i >~:isK3 'Imsfm^SSSm iiiH^^^^^	


                           ,<'1!:Hi!l ' .«' i'1'' i'1 !,i< i'n ul ,,',,!,,'i!"i tV'\	iiiliii'^ni'lliaiilllNi'i'llllPlliplllllliil,!1	
-------
  1
  2
  3
  4
  5
  6
  ?'
  8
  9
.10
 11
 12
 13
 14.
 15
 16
 17
 18
 19
 20
 21
 22
 23.
 24
 25
 26
 27
 28
 29
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
Lynch, D.G.; Macek, G.J.; Nabholz, J.V.; Sherlock, S.M.; Wright, R.. (1994) Ecological risk assessment
     case study: assessing the ecological risks of a new chemical under the Toxic Substances Control Act.
     In: U.S. Environmental Protection Agency. A review of ecological assessment case studies from a risk
     assessment perspective, volume II. Washington, DC: Risk Assessment Forum, U.S. Environmental
     Protection Agency; pp. 1-1 to 1-3-5, EPA/630/R-94/003.                        ;

Macintosh, D.L.; Suter II, G.W.; Hoffman, F.O. 1994. Uses of probabilistic exposure models in ecological
     risk assessments of contaminated sites. Risk Analy. 14(4) 405-419            -  '  .    "

Marking, L.L.; Dawson, V.K. (1975) Method for assessment of toxicity or efficacy of mixtures of chemicals.
     U.S. Fish and Wildlife Service Investigation of Fish Control 67:1-8.    .,

Mayer, F.L.; Krause, G.F.; Buckler, D.R.; Ellersieck, M.R.; Lee, G. (1994) Hazard assessment, predicting
     chronic lethality of chemicals to fishes from acute toxicity test data: concepts and linear regression
     analysis. Environ. Toxicol. Chem. 13:  671-678.        .     •      .

McClung G.;  Sayre P.G. (1994) Ecological risk assessment case study: Risk assessment for the release of
     recombinant rhizobia at a small-scale agricultural field site. In: A review of ecological assessment case
     studies from a risk assessment perspective, volume II.  Washington, DC: Risk Assessment Forum, U.S.
     Environmental Protection Agency;. pp. 2-1 to 2-35. EPA/63O/R-94/003.

Meij, R. (1991) The fate of mercury in coal-fired power plants and the influence of wet flue-gas
     desulphurization. Water Air Soil Pollut 56:21-33.               -

Merriam-Webster Dictionary (1974) New York: G. & C.  Merriam Co., Simon & Schuster, Inc.

Mount; D.I. (1977) An assessment of application factors  in aquatic toxicology. In: Recent advances in fish
     toxicology: a symposium. U.S. Environmental Protection Agency; pp 183-190. EPA-600/3-77-085.
                                                       135
                                                                                        10/13/95
-------
    pv,^
                                                                               1 nijiilji11'lU:1: : Jiiiii iiiihijiiiiiliiji»h ,ii< iii'i'iiEiifiiiibiViiiii;,:!,,!!, ,i:i:!(!iiili!iii;!'",;iiii;!;ii iia
                                                                               I lilliiiliiil'LiililiriiX Tiiil'ill'ill'iil ill T Iti	I'U! iiMilllllilliiiiSlllllillil'iHif !U': ,;;::!!"ii' till lilliiiiil'il.HmillTa1 'fll
                                                                               I niiill Ml V1 KOfi JE KKWHMWi 1111!!!1!. ftA.        liiiillli
                                DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 I
 2
.3'.
       abhplz^ J.V31991) Environmental hazard 'anj'nslc a'ssessinent	unclef	tEe"Unifed"Sfates"Toxic Substances	
'•	'   	•  •'• Control	Act Science'Total	Environ.	IbWiTb":649"-665~
	, '''I1 .  , 'i, '" 	"	;'  . will11'!	a!	.-; ;•'	-f ,!,:"';, ^ft'lilM ::	E.; c; i': - ••!!;::';	,,  tiiin:	:,."<	 _	 iini,)	       .     	      ,              	         _ 	    ^t	,	,	P	•	••	,	
              C,G.; LaPqint, T-W., eds. Environmental toxicology and risk assessment, 2nd volume. Philadelphia, PA:
    I11'''' '''•  '    .      :ill'¥^^^            	             ,     -               ,   c,        s
              American Society for Testing and Materials; pp. 571-590. AS1M STP 1216.
.10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
     Nabholz, J.V";Millerj P^; 'Zeeman,"M.'(r^^Sb^Hnvir'onmehtal'nsKassessment "of new chemicals under the
          Toxic Substances Control Act (TSCA) section five: In: Landis, W.G.; Hughes, S.G.; Lewis, M, eds.
          Environmental toxicology and risk assessment. Philadelphia,. PA: American Society for Testing and
          Materials; pp. 40-55. ASTM STP 1179.

     National Research Council. (1983) Risk assessment in the federal government:  managing the process.
          Washington, DC: National Research Council, National Academy Press.
                                                                  in mi in iiiiiiiiiiiiiiiii i 11 MI
                                                                                                i in  in inn i iiiiiiiiliii|ii|iii iiiiiii in i iiiiiii
     National Research Council. (1993) A paradigm for ecological risk assessment. In: Issues in risk assessment.
{;'.'  , •,. >  f ' <;;;:':"'"  |    | HI   I I   ''I   (  !   •    ' || |	I  |  I, I, |l|l "  < 11 |, j |  | ^ HI | I Illllll I llllil 111 |l|l I |||||l||l(|||||||l|l 111 Illllllllll illll	1111 1II 11  II 111 l||| IIII 111) |	Hi )|l| 1111 	||l | I 11||| I Illllll 111IIIIII11 I Illllllllllllllllll |
          Washington, DC: National Research Council, National Academy Press.
    '::: "'•;:•.        nun             ill   I   i  i        ill nil  I  i  ill 11 ill I (ill ill illllll iiiiiii (ill ill i ill HI ill ill ill
     National Research Council (1994) Science and judgment in risk assessment.  Washington, DC:  National
          Academy Press.
                            , "ii:',', I ii :;!'i|'" W • j'lllll	i!:r::7 i!,i"!''!'!i iMIll1* I1T i»	'i ,	!L i1!11,1 !,M" i'
     O'Neill, R.y,; Gardner, R.H.; IBamffiouse, L.W.; Suter, ^W:^
          Ecosystem risk analysis: a new methodology. Environ. Toxicol. Chem. 1:167-177.

     Orr. R.L,: Cohen. S.D.; Griffin, R.L. (1993) Generic non-indigenous pest risk assessment process.
>•"  	: >  "-I;;  ,,';i:" •ilIlil-:ii
-------
                             DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3
  4
   /
  5
  6
  7
  8
-  9
10
11
12
13
14
15
16
I.7
18
19
20
21
22
23
'24
25
26
27
28
29
30
Parkhurst, M.A.; Onishi, Y.; Olsen, A.R. (1981) A risk assessment of toxicants to aquatic life using
     environmental exposure estimates and laboratory data. In: Branson, D.A.; Dickson, K.L., eds. Aquatic
     toxicology and hazard assessment. Philadelphia, PA: American Society for Testing and Materials; pp.
     59-71. ASTM STP 737. ,                   •                        .       •

Pearlstine, L.; McKellar, H'.; Kitchens, W. (1985) Modelling the impacts of a river diversion on bottomland
   ,  forest communities in the Santee River Floodplain, South Carolina. Ecol. Model. 29:281 -302.

Perez, K.T.; Davey, E.W.; Morrison, G.E.; Soper, A.W.; Lackie, N.F. Winslow, D.W.; Blasco,-R; Johnson,
     R; Marino, S. (Undated) Environmental assessment of a direct substituted benzidine based azo dye,
     direct blue-15, in experimental marine microcosms: Narragansett, RI: U.S. Environmental Protection
     Agency. ERNL Contribution No. 882..
                                 •-                •                           •  *
Peterman, R.M. (1990) The importance of reporting statistical power: the forest decline and acidic
     deposition example. Ecology 71:2024-2027                     -         -,

Plackett, R.L.; Hewlett, P.S. (1952) Quantal responses to mixtures of poisons. J. Royal Stat. Soc. B14:141-  .
     163.                    "           • '          :  ,          .

Rand, G.M.; Petrocelli, S.R. (1985) Fundamentals of aquatic toxicology, methods and applications. New
     York, NY: Hemisphere Publishing Corporation, pp. 221-263.

Reckhow, K.W. (1979) Empirical lake models for phosphorus: development, applications limitations and
     uncertainty. In: Scavia, D. Robertson, A. eds.  Perspective in lake ecosytem modeling. Ann Arbor, MI:
     Ann Arbor Science; pp 193-221.

Rodier, D.J.; Mauriello, D.A. (1993) The quotient method of ecological risk assessment and modeling under
     TSCA: a review. In: Environmental toxicology and risk assessment. Philadelphia, PA: American Society
     for Testing and Materials; pp. 80-91. ASTM STP 1179. ASTM Publication Code Number 04-011790-
     16.
                                                      137
                                                                                       10/13/95
-------
                                                                                          I,!:*:1:,:*!!!'	
  8
  9
 10
 11
 12
 13
 14.
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
•?$
   ',
 27
 28
 29
         •' .  • J"; ", .	' j" •„ i'Slll,' -J: •;,:''., ii! >.' ill: i:" "> >	!ii:i!;:•"!>':	/ li::"
-------
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2
  3
  4
  5  •
  6   .
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18   .
 19  .
 20
 21
 22
 23   -
 24
 25
.26
 27  .
 28
Society of Environmental Toxicology and Chemistry (SETAC). (1994b) Aquatic mesocosms in ecological
     risk assessment. In: Graney, R.L.; Kennedy, J.H.; Rpdgers, J. eds. .SETAC Special Publication Series.
    .Boca Raton, FL: Lewis Publishers.  .

Star-field, A.M.; Bleloch, A.L. (1991) Building models for conservation and wildlife management. Edina,
     MN: Burgess International Group, Inc.

Stephan, C.E. (1977) Methods for calculating an LC50. In: ASTM Special Technical Publication 634.
     Philadelphia, PA: American Society for Testing and Materials; pp. 65-88.

Stephan, C.E.; Rogers, J.R. (1985) Advantages of using regression analysis to calculate results of chronic
     toxicity tests. In: Banner, R.C.; Hanse, D.J. eds. Aquatic toxicology and hazard assessment. Eighth  ,
     symposium. Philadelphia, PA: American Society for Testing and Materials, pp. 328-39.

Stephan, C.E.; Mount, D.I.; Hansen, D.J.; Gentile, J.H.; Chapman, G.A.; Brungs, W.A. (1985) Guidelines for
     deriving numerical national water quality criteria for the protection of aquatic organisms arid their uses.
     Duluth, MN: Office of Research.andDevelopment, U.S.  Environmental Protection Agency.

Suter, G.W. II (1990) Endpoints for regional ecological risk assessments.  Environ. Manag. 14:19-23.

Suter, G.W. II (I993a) Ecological risk assessment. Chelsea, MI: Lewis Publishers.

Suter, G.W. II (1993b) A critique ofecosystem health concepts and'indexes. Environ. Toxicol. Chem.
     12:1533-1539.',                           ,                -           '•   ^     -  . .   -

Suter; G;W. II; Vaughan, D.S.; Gardner, R.H. (1983) Risk assessment by analysis of extrapolation error. A
     demonstration for effects of pollutants on fish. Environ! Toxicol. Chem." 1:369-377.     .
                                               139
                                                                                                 10/13/95
-------
   1

  3
  4
  5'
  6
  7
  8
  9
 IP
 11
.12
I13"
•'14'
:• 1,5'
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25;
 26
 27';
 28
 29
                         II
                           DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                                                        II i hi HI'ill if (^
    Suter, G.W. It Gillett, J.W.; Norton, S.B. (1994) Issue paper on characterization of exposure. In: U.S.
1	IlilliW	       I
                  (fin
                                                                        I in Hi
         Environmental Protection Agency. Ecological"risk assessment issue papers. Washington, DC: Risk
         Assessment Forum, U.S. Environmental Protection Agency; pp. 4-1 to 4-64. EPA/630/R-94/009.
   	                                                                              u                        r
    Svvartzman, G.; Rose, K.A. (1984) Simulating the biological effects of toxicants in aquatic microsystems.
   	Ecol. Model 22:123-134,	_	:	

    Swartzman, G.L.; Kalunzy, S.P. (1987) Ecological simulation primer. New York, NY: Macmillan Publishing,
    Taub, F.B, j Read, P.L. (1982^ Model	ecosystems;" standardized aquatic microcosm protocol. Vol. II, Final  -
         Report.  Washington, D.C.: U.S. Food^and Dnig Adrnmistrati^^                                      •  '
    U.S. Environmental Protection Agency (1979) Toxic Substances Control Act. Discussion of premanufacture
         testing policies and technical issues: request for comment. Federal Register 44:16240-15292.
i." ,.  . '  :" :, '	;••; .. liiiiii, i	i'	•	-; • j< /"a; i:tf J-TIJI • > • :<; • ?w .a	; K	IP	«,;,	iriri	isi; ;i	'filsi	fm*isif,	nfii;;!,.: j;,,::,;	;:	::	:	;;	,	;;;;:.:!:_:;,	;:	;„;	;:	,;	i1,:	;;;,;;;

    U.S. Environmental Protection Agency (i984) Estimating concern levels for concentrations of chemical
:;	>!• *, •'.. '• •• • 'i;,1 /' •;,:"; >, am w ;>; 5 .Hi"1 xi box!"1,"	ii •,	ati'ii x	HRUM*?.*1* uiiiiiK^^^^^^^^^^^^^^^  	IIIH^^^^^^^^^^^                            	ifIH^

         substances in the environment. Washington, DC: U.S. Environmental Protection Agency, Health and
         Environmental Review Division, Environmental Effects Branch.

11 'iti 'Hi1     i; „                                          i nn n  n   in i  nn i nn  nun nn inn n
111:1,1;,!", v i ,  ,:  M	<   t                                  H    ii n  ii ii i  i inn i in in nun inn inn nnnnninninn nn innnnnnnninnninnniini n n in iiinii n nini n  in iniinin  in in in n nun i innii i nn i  i iiininnnnn niiinnnini
    U.S. Environmental Protection Agency (1986a) Quality Criteria for V/ater.  Washington, DC: Office  of
         Water, US. Environmental Protection Agency.  EPA-440/5-86-001.
    U.S. Environmental Protection Agency (1986b) Guidelines for the health risk assessment of chemical
        ''	"	 .
                                  "'''i1!'", V. :, '„ •'" :,!ii::> i 1" ITv \ >»I*»! 311:;,»'rill! in ,t 'i I" 'i! i, >!». iii" M; 'ill JiiHi'lH:!:, i i, l:,i,»',!,"": !!•!
         :."',,:'. i', Jilllll ' «i," :•>:, t, li; i: > i1! "1i -, -i ;!• -J	 ;#i, I: i"ii>;::	ilK'' iJllVilifi:;:::;;:!^^^^^^^^^    i 1; i!i«^  	illlllH                                         	41H^^^^^^^^^^^
         Enyirpnnjental Prgtection Agency (1988a) Methods for aquatic toxicity identification evaluations: phase
         '''i'  .   . . lilt -^-;;-3	"'	;:'":		i	i!M^^^^^^        	I	    	i!	
         I  to.xidt^^haracterization^rocedures.	Duluthj	MN;	Environmental	Research Laboratory, U.S.
         Environmental' Protection" Agency.' EPA-6QW3-&8-Q34.
                                                             'ii< jiiiiiii"	"»,' "jfM'1"" nij	mi, ;i!»i»»nvi:: iiiniiu'i iMiinHi nnnnnititniniiininnnnji am rum "' '•"'»•' .i," •	"',':l ,! '1|''' Hi i i":»::'!' »•':" 4i:." • r '•. • In * t^l • i!!' 11,!,"1:'' liiSlfi' ^ • iii:;i* *i ?!:: Jff i "k tyft',' rfW' i^lijiiliilliiiliiiliiiiiiiin,,' liiipiiiiiiiii'l 1,511
                                                                                      i iiiiiiuiiiiin'iiniiiiiiiniii	iiiiiniiv mxt m\; Kjiijiiiiii;? ini;,iyi:i r iiiiiH'1*;'"! ^ ;iiiiii,ii;|:ijiiiiiiiiijiiiiiif iiiiiijiiiiiiinF nil
                                                                 ,	;	iFiH'jiiiiiiiii'aiiiii '"I'liiiiiiifiiiiiiiiLSii ,:jid!'"!i inii||liii|iii|ii|i'iiiiii<'iiiiiiiiii!iii in jii11'1!;,1!":!!!
                                                                                 iri'iiiiiiiiiLiini "iiiiiiiiii iinniiilii:!'jiMJ	iiiiiiin,il:	wiii'iiiiii
                              	_.	,  . ,,	_..,,, .,,	,.;	_,,,.	:I,4Q,;,,,;,

                              •?i'ti'iI-1 /  '••fyf;'"-'1!!:	^fei'^'^^'i^^-f'SS
                                                                                                           10/13/95
-------
                             DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1
 2
 3
 4
 5
 6
 7
 8
 9
10,
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27'
28
29
U.S. Environmental Protection Agency (1988b) User's guide to PDM3, final report; EPA Contract No. 68-
    02-4254, task no. 117. Exposure Assessment Branch, Exposure Evaluation Division (TS-798), Office
   , of Toxic Substances. Washington, DC: U.S. Environmental Protection Agency.                      ;

U.S. Environmental Protection Agency (1989a) Rapid bioassessment protocols for use in streams and rivers:
    benthic macroinvertebrates and fish. Washington, DC: Office of Water, U.S. Environmental Protection
 '   Agency.  EPA/440/4-89/001.'

U.S. Environmental Protection Agency (1989b) Subdivision M of the pesticide testing guidelines, microbial
    pest control agents.  Office of Pesticides and Toxic Substances, U.S. Environmental Protection Agency,
    Washington, DC.                            -.-..-''.

U.S. Environmental Protection Agency (1989c) Methods for aquatic toxicity identification evaluations: phase
    -II toxicity identification procedures. Duluth, MM: Environmental Research Laboratory, U.S.
    Environmental Protection Agency. EPA-600/3-88-035.

U.S. Environmental Protection Agency (1989d) Methods for aquatic toxicity identification evaluations: phase
    III toxicity confirmation procedures. Duluth, MN: Environmental Research Laboratory, U.S.
    Environmental Protection Agency. EPA-600/3-88-035.

U.S. Environmental Protection Agency (1990) Reducing risk: setting'priorities and strategies for
    environmental protection. Washington, DC: Science Advisory Board. SAB-EC-90-021.

U.S. Environmental Protection Agency (1991) Summary report on issues in ecological risk assessment.
    Washington, DC: Risk Assessment Forum, U.S. Environmental. Protection Agency. EPA/625/3-91/018.

U.S. Environmental Protection Agency (1992a) Framework for ecological risk.assessment. Washington, DC:
    Risk Assessment Forum, US. Environmental Protection Agency. EPA/630/R-92/001.              ,  •
                                                      -141
                                                                                       10/13/95'
-------
,|l, 'I	   i' .  I'1  Hi1,;'"!1  , '   ,i,'|,i'! , '   f i "l! J lii ||||"i|||l| , '''i'!" I ml'1' < ', ,!' *',• ''i  ''I1!1!1 fl; 11!,,, |>. ,t| ,|M ,' h !,|< ,,, . IV i,,' • i!., | |i|ii' I'llf'S'«'||	I'llnf1 . i||P' Illl^ii ^ll|l||||||luillll|!:;:il||i ll|illill|l|i|iill|l!, ! i I;" „«' !: !-' '"I. >:ii;'"i;i 11'1'; ,i in:"' v:1 : '<•:<, :!li|in< i,	"	ii'i/ii'iitiw;!1!"! n«i	iiv'Tiii'iiiiii/if jibia	il:Hi^^^^^
      2
      3
    .  4
      5
      6"
    .'.7
    ,  8
     >
     10
     II
     12
     13
     14
     15
     18
     23
     26
     27
     28
     29
    -30
                                   <, i ,,i < ii.iiiiii1 "tW3 '"'ii	'ii"'!||i'|J|1 . iiilP1 "i'.ii1; niiiB.iii;:;!!;,,!! iiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiuiiiiiiiiiiiiiiiiiiniiiiiiiini'i.iii
                                   	'"I, Iiill ;;i'iiiii|U|,' i| 'i: 1'il'U.,'H'I<'! llll.|!|;n,i|i7ll!!lil|i|i|i|,,"f ±,'tiii'lit'viillllli^lllllliililllll'illillhBllliilllllllllillhlllil
                                    •             	Un^ii	IS u	|!1:H^^^^
                                     D^RAFT-DO	NQTqUOTE
U,S- Environmental Protection Agency (1992b) Peer review \yorkshop report on a framework for ecological
     risk assessment. Washington, DC: Risk Assessment'Forum, U.S. Environmental Protection Agency.
     EPA/625/3-91/022.
               I III              I      I      I     I   III II    II  I  III II II I I Jill  ||
   , ,n                                                              L     ,

U.S. Environmental Protection Agency (1992c) Ecological risk assessment guidelines strategic planning
.  .  ••                                                          ii
     workshop. Washington, DC: Risk Assessment Forum, U.S. Environmental Protection Agency.
     EPA/63Q/R-92/002.
                                                                                            iiihi (I'i ii'iii 1 ........ viiiiii ..... mn
                   *                                                j                          «
U.S. Environmental Protection Agency (1992d) Guidelines for exposure assessment: notice. Federal Register
     57:22888-22938.
U.S. Environmental Protection Agency (1993a) A review of ecological assessment case studies from a risk
     assessment perspective. Washington, DC: Risk Assessment Forum, U.S. Environmental Protection
     Agency. EPA/630/R-92/005.	'	:	:	'.	'	:	
               	            '          ,..,,,,     	        M            ,    ,            ,
U.S. Environmental Protection Agency (1993b) A guidebook to comparing risks and setting environmental
",.'  "..'                                                              ,                    "       i>
     priorities. Washington, DC: Office of Policy, Planning, and Evaluation, U.S. Environmental Protection
      ,
     Agency. EPA/230-B-93-003.
                                                     , ,:!,ili;i' •''	' ill "ii 'i'1!!;:	sm vmiK:	   'iJlJiiillN •" t	,".."  , ^;:!!:;;, l • h." :. H?j-; :. ^ i1 •*"",; t f;' 7] iff	ft-: f4i jf»,"!,!" 5, M	,;, kJtW	K
                                                            w, »;!,. ii , "i w<\; i1;, 'Jiii,:™
                                                                                  •IIIIIIIIIIIII1III Bill'WI1 lianilllfli!! 1: !ii;,!' '•' "I,', II Clllnillllllll lll< flm InPtUtlnill" iWiililli"! Illllilin1':'' IIIRliU1;!?1!!1!!,!,

U.S. EtiYirO'nmental protection Agency (1.994a) A review of ecological assessment case studies from a risk
     	i  ' l;"' '1 ii' j 'Killilii •,	,'' -: ~-:ll f; 4Sf:^ T .•III*"1!,! •':" - i£ ;!
-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
.16
17
18
19.
20
21.
22
23
24
25
26
27
28
29
30
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 U.S. Environmental Protection Agency (1994c) Peer review workshop report on ecological risk assessment
     issue papers. Washington, DC: Risk Assessment Forum, U.S. Environmental Protection Agency.
     EPA/630/R-94/008.

 U.S. Environmental Protection Agency (1994d) Guidance for the data quality objectives'process,
     Washington, DC: Quality Assurance Management Staff. EPA QA/G-4.                            .

 U.S. Environmental Protection Agency (1994e) Ecosystem protection. Memorandum from Robert Perciaspe,
     David Gardiner, and Johnathan Gannon to Carol Browner (March).
        /                    .                         -

 U.S. Environmental Protection Agency (1994f) Environmental Services Division guidelines.  Hydrogeologic
     modeling. Seattle, WA: Region X, U.S. Environmental Protection Agency.

 U.S. Environmental Protection Agency (1994g) Use of Monte Carlo simulation in risk assessments. Region
     III Technical Guidance Manual, Risk Assessment. Philadelphia, PA: Region III, U.S. Environmental
     Protection Agency. EPA 903-F-94-001.       "                                             •

,U.S. Environmental Protection Agency (1995a) An SAB report: ecosystem management - imperative for a
     dynamic world. Washington, DC: Science Advisory Board. EPA-SAB-EPEC-95-003.

• U.S. Environmental Protection Agency (1995b):Draft Science Policy Council statement on EPA policy:
    "cumulative risk framework, with a focus on improved characterization of risks for multiple endpoints,
     pathways, sources, and stressors. Washington, DC: Science Policy Council, U.S. Environmental
     Protection Agency.                      •                                           ,

 U.S.Environmental Protection Agency (1995c) Ecological risk: a primer for risk managers. Washington, DC:
',    U.S. Enviromental Protection Agency. EPA/734-R-95-001,

 U.S. Environmental Protection Agency (1995d) Memo to EPA managers from  Administrator Carol Browner,
     "EPA Risk Characterization Program" (March 1995).
                                                     143
                                                                                       10/13/95
-------
                              DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1                                   APPENDIX A - CASE ILLUSTRATIONS
  2                           —                                                 '.          '
  3                                                  Introduction
  4                    '  •              ""•''•.•         '  ..
                          '                       "^
  5           These six case, illustrations are intended to provide an overview of how the ecological risk assessment
  6       process might apply in widely varying situations.  Criteria used to select the cases include:
  7     •  •   representative of a broad range of potential assessments, based on the categories listed below. ,
  8           >•    spatial scale (local to national)
  9           »•   . stressor type (chemical, physical, or biological)                   '           '    -
 10           >   • ecosystem type (aquatic, terrestrial, wetland)
 11           »•    level of biological organization (individual/population, community, ecosystem, landscape)
 12           >•    data rich or data poor
 13       •   real rather than hypothetical examples.                                           '
 14       •   priority given to peer-reviewed cases previously sponsored by the EPA Risk Assessment Forum (U.S.
 15           EPA, 1993 and 1994)..                                                       '          '
 16       •   include cases relative to EPA's legislative mandates. (This was intended to be inclusive of some cases,
 17           not exclusive of others).
 18           These cases were adapted from summaries prepared by Dr. Charles Menzie (Menzie-Cura and  .
 19       Associates) under subcontract to Eastern Research Group, Inc., an EPA contractor. Each case contains a
 20       short verbal description of how the approach used corresponds to the various elements of the ecological risk
.21       assessment process. Please consider the following points when reviewing the cases.
 22       •   The cases are general illustrations of how the ecological risk approach might be used indifferent
 23           circumstances. The cases are not standards to^be followed.
 24       •   Although the cases have been structured as described in the Framework Report and these guidelines, not
 25           all were originally planned and conducted as risk assessments. To some extent, all have been retrofitted
 26           to the framework process and are not totally consistent with the procedures recommended in these
 27           guidelines.                       .    .  -
 28       •   Details on the cases are intentionally limited, and recommendations are not made regarding the utility of
 29           specific methodologies.' Given the typically long intervals between EPA guidelines revisions, any
 30           recommended methods could be outdated before new guidelines could be issued.
                                                        A-l
10/13/95
-------
                                i;-'^
                                          <:fis ,,,:::	I-'< • •": >, •'•.t	;«	'  ' '•  if;,1:'	,	;:!	-s	Kt.-.nsfA	mKiSSM	
              1         •     Further discussions of the strengths and limitations of each assessment may be found in the references

             2    ••   ,         listed for each case.  	   '  	/	     '.                ,
                               -  '",  "" „ i"!  sioiiu   i;;. \,,i »:}	', i<	J:ii 6'1 ',i  -• '•/, •' r i<  fi	:IT	^ '"fr'tTi	ifr'' 'riiV^j'fi'U^

                              V:   :t:'ii'.J|KHV	,-	l^^i	•',••.<	I"?	Iff	W3	Ifiill	

                                                                                                ";;: ii, 11!" r ii i	ijiviiifl^^^
                                  i"!    '  	 I I   ,'   • i 	!!,,' ", '!,'!,	 !	<  "»    i  '•" >  I  '• ,.L ,i fl  j 	'Ii ,': i,,',|,"!'l,:	   	    	
                                                                                                                        jij^^
                                   ,  ,  '"  :i  'i r  ', i',i	,  M
                                                                 i"  '	V'i'Hi i,;:.11 iiHhii'l 'ilic: ia.
                                      it •• •  aiiB , -  t,	:  "  " • '' '• i'	!,:" • ti", 1	''  I! ii1! "I* "'i,'!' i K\. ' it	I'; ifll. t1	•• W	m	MUf lllEBir!«	
                                         ,       r f,,',:.:'• ..i. :uiv • fif j .,''';!;'.•4;*i!; ^tffii: ii*' 3,i'iil]li:i,.,,.

                                "'•: , •;:  ''' iifii  	!;:	,'* ;:J:":",* •  •;,;=!>!!'•' II;-'' *" ;lMW        IlilH^^^^^^^^^^^^                    	;iiH^      fii


                                     ": ,„,"   liiiiiiiiiiiii'  , i'	'i'1"', ,  " '!!":'i',,:' ,ii':, i!i "'iii" iiil i'iiK'1 i'n, iPli	i'l' 1:" 'i ii'xilikiiu!',' i	% iiiiii'ipi'L.hi";"! ''Wiiiii'':!!'	jniiiiiiiiv                      twii Jiiii'iKiii'iliiifln^              'iiiiiiiiiiiiiii^^
                                    '4     i',,: i ,1'i'l' 'I i|	"'      :, ii  , 'iii  11,,:  1:.';	i' / ii  i i, '" 'in "nu. If':, I'linHi ,;|i n: r ih!!:,  ;;",,.	in,/ m,!! Mil; n,, 1 '.u IIIHB" u	liiiillijlll	illillii' i''l I J!ilii!:iji'i:i':i:iii' i'< iH^^^^^
                                     	"  .) 'fiiililll " it	".I,'1'' "'i'S !i""'::'i-n"
   	«  "  ' •'  •   ,;,;   "I." .;•!,•      ; j •'"	'  ''	  :'  'III •, •! i'' „"	•:.;!	: 'iv, „ l;i«. t \  •.	S- • l •d^':iM... ">-''!!>« "iEBi iii	•]	iililj;
	ii. V   *   ':": 'i'.i'  'i:	"•'•''  *  •  '•	ll'':'ifi':V'l;:^^^,1!:^':^;:"!i<              	^H$34^	sl^m
;•;>	r    ';:':  !i;!.;a!   -11-':-^	it  ft; *	>	:i
-------
  1
  2
  3
  4
  5
  6
'  7
  8
  9
10
11
12
13
14
15
16
17
18
19-,
20  .
21
22
23
24
25
26
27
28
29
30-
31
                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                        A-l. Modeling Losses of Bottomland Forest Wetlands
           i     .                       -
     This case illustration is presented in detail in Brody et al. (1989,1993), and Connor and Brody (1989).
The case focuses on estimating the ecological consequences (risks) of long-term changes in hydrologic
conditions (water level elevations) for three habitat types in the Lake Verret Basin of Louisiana.

Problem Formulation      '                                •
     Relation to Environmental Management Decisions: The project was intended to provide a habitat-
based approach for assessing the environmental impacts of federal water projects under the National
Environmental Policy Act and Section 404 of the Clean Water Act.  Output from the models was intended to
provide risk managers with information on how changes in water elevation might result in ecological
alterations.                                                      •      •                     ,
     Source and Stressor Characteristics: The primary stressor was changes jn hydrologic regime, including
the degree, duration, and frequency of water level changes.  Possible sources for these changes include sea
level rise and land subsidence.  Land subsidence results from a variety of natural and anthropogenic
processes. The primary anthropogenic source addressed in this assessment was artificial levee construction
for flood control, which contributes to land subsidence by reducing sediment deposition in the floodplain.  A
decreased gradient of the river due to sediment deposition at its mouth also contributed to increased water
levels.
     Ecological Receptors Potentially at Risk: Ecological receptors include three habitat types (drier
bottomland hardwood forest, wetter hardwood forest, and cypress-tupelo swamps) and associated wildlife in
the Lake Verret Basin of Louisiana.
     Ecological Effects:  The analysis considers direct effects of water level changes on plant community
composition and habitat characteristics.  Secondary effects on wildlife associated with changes in the habitat
provided by the plant community also are considered.                                                 ,
     Endpoint Selection: Assessment endpoints include forest community structure and habitat value to
wildlife species and the species composition of the wildlife community. Measures of ecosystem and receptor
characteristics included the vegetative, hydrologic, and life history input data required for the forest
community (FORFLO) and wildlife habitat suitability index (HSI) models used in the analysis phase.
Examples of these measures include tree species presence and abundance, canopy closure, and individual tree
size. Changes in wildlife populations were assessed indirectly through the HSI models for five selected
        •  . ••    •                              A-3                                       10/13/95
-------
I",!!!!" :,i'l SiV'il1', ,1! f
111! II I iJW'Tli'1 i'
"i:;i'':"i"i'i»ii!";i|!",,;1
                            ; :j llllilDlli ill}! "IB aiili; i;VfiliPIt »ll"!!i! Ml II	I1 W '*! IIIIRi "III!'., l|lll!:!'1!! '•!!, iPIIII'fll! I lln
                            • ;l!;;|ili|llllli;„ •„	"C'i ,•<	\„j|,;,.'!,'Vf* • t Mi,i"! l:illli«:i'	II'. 'US,tfKM I,ii,ftl'lXHI 111:OOfV'ittfAflUKftJSiiiiiJiillliilliillllllllllllllllllllllllliillllillillllllllllJIllllllllllllJllllilllE^
                             •I1!1;,!!	i'" • • 't' •'( i;";! in":"-' {;;'": fi;;;!	I.".L iKiiii1!1!!!!	i:	it	       .            qqmiw
                             tliSlii < ",:'i'	: <	•: .< •;. -I!	:: »!' 111!1, •!, :,.	1 ii • to ;>,;,''1«i!Rt ';,*.	!! .>!!!:;:! n-S	KIK^	I!1H^
        1"
        2
        '3,
        4"
Kfrw.
! I1'   ' ' I   "  rl' JUii ,
;  ,'  ,;	•>'"  "	:.
        10
        ii.
        12
        13
        14
        15
        16
        17
        18
        19
        20
        21
        22
       ,,23
       ,24.
        25;
        26
        2?"
        ,28
        29
        30
        31
   , • ,  • • ; , •'. I," • .• -i • -V, ;';. - ; P.RAFT--DQ._NOT .QUOTE, ....... CITE, ....... OR DISTRIBUTE [[[
                                                        11    i,  '  ,' ' i '   ;  ' ,       H       '  '
    wildlife species (gray squirrel, swamp rabbit, mink, wintering wood duck, and downy woodpecker). The
    species were selected fbr'evaiuation'blise^^                               (0 sensitivity to hydrologic
    ...................   [[[ " ................................................ ' ........ : .................. ' [[[ : ......................... ! ................ : ..............                        '
                                   . '||!| ''• I. !iii| ,l',ll,i!l HI1 I IKS .....
                                                »'!' 'mi' '.i"!'."!*1! SI "'III i:,1:,!1 ''imiliTiB
                                                                                    iiFiiiiP'i ii ii ii 'i ill PIP "in
    changes, (2) representation in a large segment of the wildlife community, (3) representation in various niches
    (trophic status'use of habitat),and (4) cbmniercIaCfecfeatibnai, or social importance.  Measures of exposure

                included estimates of increasing water levels" ffiafresulted'fFdni ayariety of processes. Decreased sediment
                deposition in tne' flood plain due to 3ie construction oF artificial levees for flood protection was an important
    germination,
    ";, '   .....  V -i! :•< : iiHIIII ..... 'I ......
         Conceptual Model: " Alterations'ui
                                       ' • ...... :•'  ! l-vili1 Mllilf'l 'i, V
                                                       Hi)	In!	WltMWiftMW	               ,        	i!4M            	'!
                                                         '	'	"	"Jl1	'	'	'"•'	""'	'""*"'	'	"	'•'	4	with" ^hinges	in	hycfrblbgic	
    conditions are linked to changes in wildlife values for species that are dependent on particular types of plant
                                                            1  '        -             -                    i    , i,
    communities for habitat. The risk hypotheses are implicit in the predictive models selected for use in this
    i  '•  .• »!"• i £	: •'• liiiiilili"r,i:"> ;jj i V• }••('
    case (FORFLO and HSI).

    Analysis (see ifigure 4-6)
          Ecological Receptor Characterization: For the plant community, baseline data on key habitat
    characteristics were obtained from the literature or by making observations at field sites. For wildlife, the
    'casts relied primarily on"published"iiterature	concerning tSie'Habirtats	an"3 t'He wildlife they' support.
                                                                                                •in'IIIIU.IIIHIIlinilllli'lll.llnlnilHI.III'SII
          Characterizationof Exposure: fhe exposureregime for changes in hydrologic conditions was
    .""!'.; .''F "<  i'.';"' yijl"t."i""'!,;„', "i;.!'!,">.:'-.,;.fi:;.;|	ji;:-^	n? .  n;)mf,wii-i	iHt	,     .       ^                 •
    simulated by the FORFLO model based on" estimates'bf n"et	suBsISence'	rate"	Since	forests can" respond
    slowly to these hydrologic changes, model simulation was conducted over periods ranging from 50 and 100
    years up to 280 years.
    i   . - u' ,                                                      i   I      I         i        j            j   I  I
          Characterization of Ecological Effects:  Changes in plant communities and the habitat they provide
    weres simulated using the FORFLO model based on laboratory studies of plant response to moisture (seed
                                                                                                   model .predictions
     germination, survival).  The model tracks the species type, diameter, and age of each tree on simulated plots
     from the time "the'tree 'enters
     '  i. • ;l .  li"i'!''•""IS!!! • 'ii1;!n! Jk'iv''''":i:it
     were used as input to the H
      	';'   ""'     -'
from the time "the "treei'enters "ffie plot'ay a's^Hrig or'sprbut'until it dies. The :
 ': i.  • 5.   '.Vi:'• -"IS!!! • f<;!'i'v Jk'iv1'''i:|:ititi iir i:=i" -•>'!ii3il!M                                                  	IMC*	•*
were used as input to the  HSI model.  HSI input vanables for the sel<;cted wildlife species included:
percentage of"caribpy"closure'bf'1iani''ni1astitrees	(squirrel), numb'er'of trees that produce hard mast (squirrel),

     percentage of tree canopy closure (squirrel and rabbit), average diameter oFov'erstoty tree's" (squirrel)',
     percentage of shrub cro\vncbve'ir (squmfei)",' annual flood duration (rab"b"it and mih^), percentage of all
     vegetation canopy closure (JnjnJ^ tree ba'sararea1 Woodpecker), nun-iber of snags (woodpecker), and
     percentage ofwafer surface covered by winter cover (wood duck).
                                                                   A-4
-------
 1-
 2
 3
 ,4
 5
 6
 7
.8
 9
10
11
12
13
14
15
16
                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
Risk Characterization
     /?/sfcs to Vegetation: Output from the FORFLO model was presented as a series of time plots
illustrating changes in botanical characteristics for predominant plantspecies: These plots further illustrated
the composition and relative standing stock or production of plant species for each habitat over time.  These
characteristics were examined in regard to different rates of change in hydrologic conditions, with the results
discussed in terms of the rate at which one habitat would replace another.
     Risks to Wildlife: Risks to wildlife were illustrated by comparing tabulated values for current habitat
suitability (as measured by the HSI models) to future habitat suitability as simulated with the FORFLO
model. The'future HSI values showed a general trend toward loss of wildlife values; however, this occurred
at different rates for the various species and habitats. Indeed, in several cases habitat value increased.  The
exercise provided a basis for examining the underlying factors contributing to changes in habitat values.
     Uncertainties: While the models capture major features of the response of vegetation to flooding
(FORFLO) and factors affecting the habitat of selected wildlife (HSI), they do not account for wildlife
disease, predation, competition, or colonization or for the effects of other stressors.  The analysis is
applicable only for the selected species, and effects on other wildlife must be evaluated in a qualitative way
based on professional judgment.
                                                        A-5
                                                                                            10/13/95
-------
  :: 2
   3
 .. ,4
 .} 5
  "... 6
  - 7
 . ,,   ,1
VM2 ......
  Jll .....
   15
   16
   17
   18
   19
   20 .....
   21
   22
   23
  .24
  ,1 i!11: •
   25
   26
   27
  .28.
   29
   30
                         aii "'i,,),.
                                  ' DRAFT--DO NoTQUptE^^
                                  .'ir4                                                                 -                •
               2. Special Review of Granular Formulations of Carbofuran Based on Adverse Effects on Birds
         This easels presented in detail in U.S. EPA (1989) and Housekriecht (19S3).  The issue leading to
   EPA's special review of carbofuran concerned risks to birds.
   Problem Formulation
         Relation to Environmental Management Decisions:  If EPA determines that the risks appear to
   outweigh the benpfits of a pesticide, the Agency can initiate action under the Federal Insecticide, Fungicide,
   and Ko^enticiclei Act "(FIFRA) to> cancel", suspend^and/or 'require "modification of the terms and conditions of
   '4|. •  , '	 ••' ,:,	;	' jium' • 'HI.' ',". „: (,' ai, ':" t- f. • lit,*, ii«,»;r jyuVMHfajM ~*»i> ::ij. i	i	; #,	imitrnm	i	ui.'Ma	I	!«	i	J'"1	•	i	i	B	i	1	:	j>	i	•	;	''
   registration. EPA initiated a special review of granular carbofuran formulations because of multiple instances
   '•         '''
   z. "ev-ivy .;	!,TO>«i'WM^fi	lit ;:;t.IK1:	
   of carbofuran-related bird kills.
   ': i "v', „,'" •'!	,,z: j. 'X'Sii1 i' it/,", i j ;,!,:> •;>>« i''	i	"' if i;»:!, ,ii;ii>iii, wsiisiiM * .'login* nftWMZHWHtfUKW •      	               •	,	i	
   	•   " ^ .Lit'!..: ill:..: .ir:""11''..,:!!!;; '"'"ii; ju.1; .11;;:! lff"!i":i; fS, I'tjii^'iLlii^^^^^^^^^^^^^^^^^^^^^ ;;|i«^^  	SIIXH^^^^^^^^^^^^^^^       IIV^^^^^^^^^^^^^^^^^^^    	I	      '
                                                      I fill.!1
           applied primarily in granular form on 27 crops as well as forests and pmeseed orchards.
                 Ecological Receptors Potentially at Risk: The assessment focused on Girds that may inicTaentally
        I         inn 11 in i             ill        wi'i'.	ii *i,i*ii" * , ' 11 vSi'i1's	i'^ii1'1 .ii I. 'i"' ijiii'»iii'hi,iii 'B1 i*iii'"!»: Diiniiiiiiii iiiiiiiiiiii1" ipiniliiiiiiiiiin, iiiiiiiiiiiiiii' • iiSiiiiiiiiiiiiiiiiiiiiiiii!iiiiiiiii;i',,iiii»iiiiiiiiiiiiii ii-iii liisl	ii liiiii' i llili;:;;ilil.li;illilllllll»jiii,iii	iiiiii'miiiim / 'iiii ii™ wiuni iiiiii< iiiiiiiiiiiwiiiiiiqi	isiiiiiii=:iiiiiiiii:iiiiiiiiiiiiiii!ii|iii'is;;:::
    ingest granules as they forage or that may eat other animals that contain granules or  residues. Watertowl and
    songbirds are considered most at risk.    	'"'	
         Ecological Effect's: Carbofuran "is "an.'"acute "toxicarit'Siat1mHibrtscEpIines'Ferase. The primary effect is
    death of the birds.  Secondary poisoning of animals feeding on contaminated birds has alsSbeen observed.
  .,  "'   ! "' ';.	"  ,.i; ii:'.:.  ii,i:.ii "lj ' .ii;i.;!'l:l MI i.	!'	i1.? ; .i.:1:*:1.'...!. IJi ..ftii .;	'....'..li..;.,!.,'!!':1.!!.,.	i;i!;;!"!.if IllliillJCiW'.....!..';''!!!!!! Ij:111..;'	I	Ill	i:	/"I!1!.. tiUIHIM^^^^^^      	I	:.iin	i	I	I	
         Endpoint Selection: The assessment endpoint is survival of birds that forage in agricultural areas where
    carbofuran is applied. Measures of effects include lethal toxicity data from laboratory studies and  •
    observations on, bird kills in the field following carbofuran applications.  Measures of exposure include the
    number of exposed carbofuran  granules per square foot of soil.
                                         I" ruin;" fii'.i'i'iJlirl'fl, > I'lWl "'I "*, II.I...'.I/11 ",
    Analysis
         Ecological Receptor Characterization: Anlnventory"of~blfci siJecles that may be"exposed following
''f!;',;::V:!:,  ,^'V'^ ' M!if:i*!ii.:.i:;'*1':fl','," '"''V:1':11,;"!' '*Jv'i,:i.1:'*iirHt;i'§f;'••;?;'tfi^l
                                     - ;:,,iiiii:.,,,hl'.,i"'i" IEl""ni"i.' ':'
-------
                              DRAFT--DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      furrow applications. The amount of granules exposed at the surface of the soil was estimated from these
 2  .    percentages for application rates of 1 and 3 pounds of active ingredient per acre.  For field observations
 3      where bird deaths had occurred, levels of exposure were estimated from information on application rates and
 4      application methods (band or in-furrow)-.
 5           Characterization of Ecological Effects: Field studies were used to document the occurrence of bird  '
 6      deaths following applications. Data were converted into mortality rates per acre and compared to chemical
 7   .   levels in the field, estimated using information on application rates and methods.  Single-dose toxicity studies
 8      were conducted for several species of birds.  These data were used to construct a toxicity statistic expressed
 9      as granules per LD50 (the single dose that kills 50 percent of the test birds), assuming 0.6 mg of active
10      ingredient (AI) per granule and average body weights for the birds tested.            ,
11   '                            '    • .   "            '                  .       '           '    "    '
12      Risk Characterization
13           Risks to Birds: Risks were evaluated using a weight-of-evidence approach that considered laboratory
14      toxicity data, estimated exposure data, field studies, and incident reports. The approach included a form of
15      the quotient method wherein estimated exposure levels of granules in surface soils (number/ft2) were divided
16      by the granules/LD50 statistic. The calculation yields values with units of LD50s/ft2.  The higher the value, the
17      more likely a bird is exposed to levels of granular carbofuran; at the soil surface that can result in death.
18      Minimum and maximum values for LD50s/ft2 were estimated for songbirds, upland game birds, and waterfowl
19      that may forage within or near 10 different agricultural crops.   '
20    '       The potential magnitude of bird mortalities from direct poisoning was estimated from the number of
21      acres of agricultural land treated each year and the mortality in the field studies conducted on com. Assuming
22      similar bird mortality occurs in all crops, an estimated several million birds could be killed each year from the
                                                                                !
23      use of granular carbofuran; because mortality in the field studies is likely to be underestimated, the risk
24      estimate (number of birds killed annually) may be  low.                          ,
25           Uncertainties: Several  areas of uncertainty were identified in this case study. First, while a large
26      number of bird species could  be exposed to granular carbofuran, data on the effects of carbofuran are
27      available for only a limited number. It is unlikely that the most sensitive species was tested or identified from
28      available studies. Despite this uncertainty in the analysis, EPA's Office of Pesticide Programs (OPP)
                                     •v                                         '
29      concluded that carbofuran posed a risk to birds such that the continued use of granular carbofuran outweighs
30      possible benefits; therefore, OPP concluded,-registratiqn of granular formulations should be canceled.
                                                         A-7
10/13/95
-------
                  . . • 1 •  . " ' • ' "if "  I'mitRI • • , ,!' j n <•
                                        ' 1 *fl ,]„" *| "B»i| i f ' ....... 1 1 ..... '! t ...... \ \ . »:
                                                                                    ll!,ni IIIIKinii'lPllii' iifllllli I .P II: BiBIBB BIIIIIBBBIBBIIi illllBIIBIIBilllBBIBflBflllB	'BIBBBIB
,i'!!U  ,,i'i T'   'HI  li.l'
                        ,;•  ;f y^	'«	••>"•¥«•	ii»!mM$&«	Bmwiwn	mm	i	ifi	•
                                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
       1
     "  2
       3
    ,  4
     •  5
       <$
    ;;it•',•('".; iii..:-:;' »ii:>>!> ;, -i"is/"ii,Jtjiift'S;	~\"t.":ii xsirmi^,iiii&
 exposed granules.
••'," 'M'!"";  w- '	81;MS'. M':,'-S"i»;ypSi;-v,:!r'y^:ifr&it	^^t^K^K»3KKri.«KytveaKstt	rfKEiafflsaacaEHS	mas	•.    , .
      Third, while it is  assumed that the higher the LDsos/ft2 values the higher the nsk, the actual relationship
          .mil.1 , i1, lliiiillllliil: ',„ "ii11:,	;„  i'ti i<	Il1	 Sri I"
 is not Jknown-'VlTherefore,. the availability of field data from more than 40 actual incidents of bird mortality is
 an important component of the overall weight-of-evidence approach in reducing uncertainty associated with
 the assessment of risk.
                                                                                                              -
                                                                                                  •'lIF'I'IIIIIBBlMl'll'Fl1:!'!!!!!:!!!!!'1!!!!!
              • , "  ' , '!'
          M1,, ,  'j:   | ,, hi!,'

          I."1""'  •"'  !m' !
                                        '»', if' i ii,; i „ i;;1, " , i"  , •;: "i' ,i, ir :•',; n11 •;•«:	ii!1 ••' s •>' .!•., ii", :& • 1,1 .!• ^ivw • u;»"iH^iiiiiw
                           i;hihffli'-' „::''• "IIP

                                                                                              HI, iiiniR', IIP 1'iip, i1, riiii'i'.iiinii, in iiiimiii'ii.iui'iiiiiiiii! u»: iiiiiiiiliiiii	",	i	i	ii"ii"p"<	i	<
                                                                                    I|IGI ll|il|l!l!tiii!i|lli!|l|iliPi|i illilllll'iiiijillL'P iflPlNlii'.!!!.,:.!	Pllll B1|:ll S1'!'!! 11|illil!illiPlii:>Hi 4pi'i:; j'f III flU''!^!!!^'^';*!!!.!^^'1!!!!! 'W.iVllllllliniiniiiiillillii'illlllllim^^^^                     	lilll!lliiliNi|ilii;W^
          ' ;!  ,...,;  '' ,h-  ' „ A  S;""> i....•;':,;••'•:'|"';•,tin-1 '-,\"}:JJ	i".lii^il'i^it;:::![^
                                                                 A-8
                                                                                                 10/13/55
-------
  1
  2
  3
  4
'  5
  6-
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
.20
 21
 22
 23'
 24
"25
 26
 27
 28
 29
 30
 31
                       DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                     3. Pest Risk Assessment of the Importation of Logs from Chile

      This case is presented in detail in USDA Forest Service (1993). The case considers potential risks to
 U.S. forests due to the incidental introduction of insects, fungi, and other pests inhabiting logs harvested in
. Chile and transported to U.S. ports.

, Problem Formulation
      Relation to Environmental Management Decisions:  This risk assessment was used to determine
 whether actions to restrict or regulate the importation of Chilean logs were needed to protect U.S. forests.
 Based in part on this assessment'and others like it, a regulation covering the importation of timber and timber
 products into the United States was prepared and published (7 CRF Parts 300 and 319).
      Source andStressor Characteristics: Stressors include insects, forest pathogens (e.g., fungi), and
 other pests.  Preliminary evaluation focused the analysis on 14 individual pest species of Monterey pine. The
 source was the proposed importation of potentially infested Chileanlogs.  '                     '.   -
      Ecological Receptors Potentially at Risk: Receptors were the managed and native conifer forests near
 areas where Chilean logs are imported. The analysis focused on the forests of the western United States
 because these resources were considered most vulnerable. This region has a climate similar to that in Chile,
 the forest resources are of great economic and ecological value, and most log shipments are destined for
 Pacific coast ports.
      Ecological Effects: Depending on pest species, damage can occur to leaves (needles), roots, phloem,
, and bark, resulting in the decreased growth or death of the tree.
      Endpoint Selection: The assessment endpoint is the survival and growth of tree species in the western
 United States.  Damage that would affect the commercial value,of the trees as lumber was clearly of interest.
 Measures of effect included data on the effects of the 14 pest species (or other, similar .species) on the
 survival and growth of U.S. trees or related species.  Measures of ecosystem and receptor characteristics
 included data on pest climatic requirements, life history, and host specificity.

 Analysis (figure 4-7 provides an interpretation of this process)  '
      The analysis was carried out by a six-member scientific team that assembled and evaluated available
 information on forests in Chile and the United States as  well as on the pests that could be transported.  Risk
 was summarized in terms of Pest Risk Potential (PRP), a formal determination of the Individual Pest Risk
                                                        A-9
                                                                                            10/13/95
-------
                              II  III,I 111
                                                               I	I1I>	lllllllli'illllll	IIIIIH    	Il'lllilill	il	I	f ....... flaimwiii'i&'EBWiffi ...... .  .....   .              .           „ ...... ,,.,,. ...... «» ......   ,   c
    western slopes of the Cascade Mountains were identified as supporting some of the highest quality stands of
 j -• ' j r  ,i j:i ( i«> iiiifii ''*•> ',:•: ...... ' i: ,;-'! ........ , •i:,i:!:>:i:ii!iill. ! r a .    . :"i>Ai<'i;i ..... iiiiiiTi'iiifC'MsniH ..... iw ...... im ...... .   ....... i ...... iiiiiiiiiii .....          . ..... n ........ i ...... • ........ . ....................................... . .................. > ......... ._. ........ I ........... i ..... • ........... • ..... ..... » ...... i ........ ! ............ ..... i ....... i ....... in ............
    origin (the pest's capability to "hitchhike" on log shipments); (2) entry potential (the probability that the pest
:"'i ...... ' y'ii  •.  .'i^J..^^*.^^!'^-!-^^^'^:;'^ ....... ldi*3(«ffl ....... i! ..... \ ...... i«^^^^^^^^^^^^^^^^^^^^^^^^^^^         ....... JBHIMM ..... il«^^^^^^^^       .....         ........ i ...... • ...... • [[[          I
    will survive while in transit from the country of origin to the United States); (3) colonization potential (the         |
    probability that the pest, once introduced, will colonize the local area — factors include number of pests,
    reproductive requirements, and host specificity); and (4) spread potential (the probability that the pest will
    spread beyond the colonization site — factors include natural and assisted dispersal, genetic plasticity, and
    distribution of potential hosts). Each of these  four elements is" carefully considered and assigned a judgment-
    based vaiue of high, medium, or low. These values are supported by detaire'H'EIoIogical sta'ternerifs' that take
    into account conditions in the Chilean forests,  knowledge of the U.S. forests, historical observations of the
    pest species or related species, and the biology of potential pest and host species.
         The analysis identified insects that inhabit the inner bark and wood of imported Monterey pine logs as
    having a higher probability of being introduced than other insect pests.  This group included bark beetles   In
                                                     A-10
                                                                                                         111 fill 11IIII III III I III I I IIIIIIIIIII
                                                                                                          10/13/95
                                                                                                            iiini up i
-------
                             DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
 1      addition, several pathogens (principally fungi) may inhabit the .heartwood and sapwood of pine and be
 2    '  introduced,
 3           Characterization of Ecological Effects: The IPRA methodology for evaluating the "consequences of
 4      establishment" was utilized for evaluating potential effects. Like the "probability of establishment"
 5      methodology, this is a judgment-based procedure with the following components: (1) economic damage
 6      potential (the likelihood of economic impacts with regard to the economic importance of hosts, crop loss, and
 7      effects on subsidiary industries); (2) environmental damage potential (the likelihood of ecosystem
 8      destabilization, reduction in biodiversity, loss of keystone species, reduction or elimination of
 9      endangered/threatened species, and effects of control measures); and (3) perceived damage potential (the
10      likely impacts from social and/or political influences). The scientific team assembled and analyzed the
11      information to reach judgments concerning each of these components.
12     ,      Meristematic insects and defoliators are considered unlikely to be potential pests of quarantine
13      importance on logs. The European bark beetle is of concern because it could serve as a vector of black stain
14      root disease, caused by the fungus Leptographium wageneri. Saprophytic fungi causing blue  stains on
F5      freshly cut wood were also identified as potentially important.
16                              '       _••'/".
17  ,    Risk Characterization .....'•            ..                                        •               '
18           Risk Analysis: The seven judgment-based "risk values" of high, medium, or low derived for the
19      "probability of establishment" and "consequences of establishment" are combined into a final PR? for each
20      pest of major concern following the IPRA methodology.  Final PRP categories can include low,
21      moderate/low, moderate, and high. The assessments were completed by the six-member team  and distributed
                                                                                   .       *-
'22      for peer review. The final assessment includes reviewers comments and responses to these comments.
23           The analysis resulted in a "high" rating for the European bark beetle because of its abundance in Chile
24      and its importance as a vector for black stain root disease; other insects were rated low or moderate. Fungi
25      that can result in stains were given a rating of moderate/high.
26           Uncertainties:  Natural history information is limited for many of the insects and other organisms
27      inhabiting Chilean trees.  Although these species could be important, they are not considered in the
28      assessment. The analysis is based on the judgments of a six-member scientific team concerning each of the
29      seven elements of the IPRA process.                                      ,
                                                       A-11
10/13/95
-------
     1
    2
 :• . 3
 ;!  "4
 :.  -
 ; ..... ^   6
    7
    ?'
 •;'  ,9
   10
   it
   12
   13
 i' ......   I,,,,
   14
   15
   16
 H7.
   18
   19
   22
 I '
'>  23
   26
   27
   28
   29
   30
   31
                       ,  DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE    '
                                                                     i i              i i i                 i  |i
                                                                                                         I,
 M
 :   .            i                          4. Waquoit Bay Estuary

                                                               II                     !                  Ill
       The case outlines a risk assessment approach for a watershed that comprises a broad range of stressors
  and receptors at various spatial and temporal scales. As an example of an ecological value-initiated
  assessment (section 3.2), this case is organized somewhat differently than the other cases in Appendix-A.
  This case is presently under development at EPA and is not yet complete.
            r                     '                               II

  Planning
       Management goal:  The goal for the risk assessment in Waquoit Bay was derived from common
  elements of established goals by a consortium of'local, state, regional and federal agencies and private and
  public organizations. Among the organizations were the Citizen Action Committee,  Association for the
 I''1             niiinii            i i n n  n   i    in i i 1 1 i "        i n  i n i i in n i i in mi n n  11111 i|iiiiiiiiiiiii|iiiiiiiiiiii in in nil in i n 1 1 mi mi in inn n ilin 1
  Preservation of Cape Cod, Cage Cod Commission, Atlantic States  Marine Fisheries

  Massachusetts Coastal Zone ManagemenJ, Nationd Marine Fisheries Service, U.S. EPA, U.S. Fish and
  Wildlife Servjee, National Oceanic and Atmospheric Administration, and the National Estuarine Research
11 1 '     ',;  : ;. '• . ,i -  mm - • ft .-•• : , ....... ' .r:, . ..... ' i. a > • [iv ..... v ;< ,  -    3, taSK f::»i wa n tiffi Zffjii&iil ...... iii»                              n in ........ M ..... m ...... ii ...... ( .......... iliiii ......... iii |
  Reserve System: Waquoit Bay).  The risk assessment team developed a risk assessment management goal
                                                                                     i       '  , •  '         |
  and presented the goal to representatives of these groups at a public meeting. After review, the following
  goal was established for the risk assessment:
       Re-establish and maintain water quality and habitat conditions in Waquoit Bay and associated wetlands,
II ,   : ...... I . ' '  ; ';"   iliilllll ' "' I 'V1''!'1 'I i  'I;,!';:! "!•!•, 'j- til?! fit,S'\ ..... *}Bi&li>>ffiUIWWt !!' KINI lull iilBiliVifK^                    ..... II lilillH^^^^      ..... liun^^^^^  ..... it ...... lillllllilM^^^^^
       freshwatgr rivers and ponds to (1) support diverse self-sustaining commercial, recreational, and native
      ? fish, wa|ej-dependent wildlife, and shellfish populations, and (2) reverse ongoing degradation of
 1 .   • : ' :   ' ..... i,  _ 'S ' f, : '  ' ": " ::,SK "k j 'It w ..... "' n, Sri I , if':"'; SriwKn i:i! «"«» "*'a ':. ..... 1 ,. ..... .wil /j
           ,          ,   -   , .      ,.  .               , ......                                             [[[
  For use ir» risk assessment, the risk assessment team interpreted the goal into 10 sub-goals or objectives: The
 ":      .....               ...... ^ ......                     •                                                        .
    _   _     :
  objectives listed below identify conditions in Waquoit Bay supportive of achieving of the management goal.
 -'  :-','\i^ ............ !;a^iKlirlii^,^;;t::'i-;ii ....... ';.!!:^             ...... mm ..... ms, ..... ;M^^^^^^                                ....... i ..... i ...... JM^^^^^^^^^^^^^^^
                  ..,. i. .  ,II      ,,   « ,, , « ,    ni<  ,, • <  .n,,,!   ,«    ,  ,ii'.
                 educe or eliirimate hypoxic or anoxic events
                    * '!:	'XSiif ' ''r! ' ; '."!ii  • iilii'1'1	F Jn ' ifli."ii	, f  iji1",!'!'1 !l:.l^^!i.'vllill!lllillj1li!;lJr|:l'^l iijfiNi.'^ii!V'^iS	'iiiiii'v'*!!!	v^;,iiniii?aiii »	uw
       prevent toxic levels of contamination in water, sediments and biota
      :.'..: ,v. •• .iii';!::,: t,, i iMf :.\® - mi i	::>*• a vwsm 'mm	in	iB^^^^^^^^^^^^^^^^^^^^^^^      am	
       re-establish viable eel grass beds and associated aquatic communities in the bay
       re-establish a self-sustaining scallop population in the bay that can support a viable sport fishery
       protect shellfish beds from bacterial contamination that results in fishing closures
       reduce or eliminate nuisance macroalgal growth
       restore and maintain self-sustaining native fish populations and their habitat
       prevent eutrophication of rivers and ponds
                        ,:,» in i ,r i,
                        •i, II •'','.
                         Wii
' ' ' -l'!1"', ! ' '• ' ..... l ifil! SIS1 HHi '  i" '! •'T ' ! I.:!-' ........ IC'I'fnK ...... fll:!:!!! ..... ill' il!!!!!"' 4 ........ IK .liif IM     til ...... IE"!! K> ..... IE^^
'•;'' • "'"' • ' ':; "::" ?f:"Bi!! ' '*i;¥:if, A- 12'         '      '
" '  „ '' I '! ' D } ! iilli "iii 'i \!i i 'V M ill!! T ' n'i i! 1 1' ' !i ¥11 III il ' lilll?1 IL i V
                                                                                                 ilin^^^^^^^^^^            |
                                                                                                   10/1 3/95
                                                                                              l; I, ' ilH'1;!1 i m1 ill WhUHI, •illllMlllliliilW^^^^^
                                                                                  niJIi:,,* Mi* JiJplin, nil; M1!,, 'filuJliillili'l'illl 'liliii1' lilMlliiik "llili'WK i I'll	iliil; I/

                                                            	Tl*,^)M«r'*i*	i i!:	:	ri.iiijiitjiiiiiijiiiiiiiiii.iifjriiiii	i:,	I;:,.,;!!1.!'!	Ufa	MJ	iiiijii^iijjf^iiijti^iiciiiu  	iiii'.!!:'!)!:,;!,	' •   '     	      I
                                                    ,n	ni ni t; sj ::iii:;i.;iiiniiii
-------
  1-
  2
  3
  4
  5
  6
  7  -
  8
  9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
•    maintain diversity of native biotic freshwater communities
•    maintain diversity of water-dependent wildlife

                       1 .  '         '                   /•
Problem Formulation   •
     Assessment Endpoints:  Multiple assessment endpoints were selected based on the management goal
and an assessment of available information. Endpoints include: eel grass, habitat abundance and distribution;
the species diversity and abundance of estuarine benthic invertebrates, migratory fish, water-dependent
                             '- :                .                                    *
wildlife, estuarine fish, and freshwater fish and invertebrates; and freshwater pond trophic status.
Source andStressor Characteristics: Major sources of stressors in the system were identified to include
residential development, industrial uses, agricultural activities, marine activities and activities occurring
outside'of the watershed that influence the watershed ecosystem (e.g., armoring of coasts, air pollution).
Seven physical, chemical and biological stressors were identified in Waquoit Bay watershed.  These stressors
include: excess nutrient loadings, suspended and resuspended sediments, physicalalteration of estuarine
habitat, toxic chemicals, eel grass disease, fish harvest pressure and altered river flow.  The stressor
considered most dominant is excess nutrient loadings from septic contamination of groundwater and air
pollution.  Tbxicity from contaminated ground water plumes flowing froma Superfund site represent a
potential current risk to freshwater ponds and a future risk to the estuary.
     Ecological Effects: The waters of Waquoit Bay and associated freshwater ponds are exhibiting signs of
water quality degradation.  Ecological effects include loss of eel grass habitat and associated species
(especially scallops), alterations in species composition in estuarine and freshwater communities; and
declining abundance of commercially important fish and shellfish.  Shellfish closures from bacterial
contamination are an increasing problem.  Significant growth of macroalgal matts are covering the bay and
are believed to be responsible for eel grass loss and fish kills. Preliminary data on fish in ponds indicate
potential toxic effects from contaminated groundwater plumes.  Pond eutrophication is a concern.
     Ecological Receptors Potentially at Risk: In value-initiated assessments, the management goals define
valued resources.  These values were translated into assessment endpoints that represent both valued and
susceptible ecological receptors. Conceptual models were developed based on assessment endpoints,
ecological effects and stressors presented above.
     Conceptual Model: A watershed conceptual model was developed that included multiple sources  of
stressors, the.primary stressors, and the primary and secondary effects they have on the assessment endpoints.
                                               A-13
                                                                                                    10/13/95
-------
 ••'t, 	i; i-r:;;,', ,', "vii ^4-'
""'i;  .-7')iti?!y !'". :  '; '-v;'
    2

    3

    4

    5
                                     DRAFT-DO NOT QUOTE, CITE, OR .DISTRIBUTE".
An effects matrix for the Waquoit Bay watershed, derived from the conceptual model was generated using a

consensus  "fuzzy set" approach (Harris, et al. 1994) to prioritize risk hypotheses.
!l	' •! ImJIi If*, i|, i'", * iJ» i;:, "H • ifi1! i','! |||,, nR „. !„!"«1 ,# ji-i1' M, 1 il l ,n'i ,J:,,, t'1'1 iiPf iliiii i'f i HflB!1!1 !lll!ll|lf|P. JihiillP1 Iliiill11'1,«'!'! a™!!!!,!!
Analysis

      Primary risk hypotheses generated for the first round of analyses focused on nutrient loading effects on

                                                                                                    ;	at	
            eel grass and macroalgal growth. The analysis plan is being developed to evaluate nutrient load

            different locations in Waquoit Bay and other similar estuaries in the region and compare loading to losses in
            ii",!,''  f >  '  '  V •  ' "IJI'liBlf % i'' ''"'," •'"" • !.,, ":'	,! ''..III1'1! •,.  II'! 'I'!"i, i11,, H1 S "W,	!!|Mi••I'MIMi!1 W!1*!!, "1,1 .''"'iili1. iiii'lK1'1!!'	k.i	'IMIiWIf'iT'liKillliBI SIP'".'II	iliiinm,"!!'	I	•	n	,	f .1,1,,,	•
            eel grass over time.  Other risk hypotheses will be selected for further development.
                                                                       /'..fii'iili'if'Jiiiiwii:1 ji'ii! wiiii'ii, '"ii
                                                                                                      I11111'!111!!1'!::!!!!!!1)!!'1!','1!!'1!:!!1!'!!!!;!,!,1111'1!1:1'!!:!1'!1!1!1 '!!!!"'
                                                                            -                     •
                                                      ',, i; ,ii,:: •»: ,ft,*, / ,."!•; ,F , f, * i:"ii, * ai.iii.iMii'jthii' t'1*'^^         	iT!*"!!1111!!!!'!!;!™!!!:!™!1!!'1, 'T!'!'"'.'!	I!1!"'"1 "!,'Il!Tir!Mr"™^^
                                                      ',":'V       '         	                       '
                                        jjr^^iift	^i:!:p^p3|3ii|i
                                                            .f ^                                             	
                                                            '»;	lilMi* ,1! P. ,•:', I'1'	,,,"'ii. '1.' ! :„ !lf, ii i	'! 1 III!1	',»., I fli linikll	lill' IIIIIIHDI Illllllinil UK:	; :i^ IIIIM i 'I lill'n r All JdHi M
                                    :: |i,i " «; ." '' ;	n: ,' ;i'' I  1 ...... '''iii                        '"  'lvVli'pK     " "wi1 J^'d'Hi'W
                                                                                   'I— — .!!!!*i™v:,::,^,,1:!!!1! ',:'!:!„„", ; :,!"'! :!,:::,,::,'":::, I!!,:,1:::1!:!!;,;1:,:, ; 1!T:::i.;::\i:;:;1;,ii;r:,1 i1;,:,:,,,, "..T'lv:::;!:.1;:::?::;:^^^

-------
  1
  2
  3
  4
 o '
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
20
21
22
23
24
25.
26
27
28
29
30
31
                      DRAFT-DO NOT:QUOTE, CITE, OR DISTRIBUTE
            5. Assessing Risks of a New Chemical Under the Toxic Substances Control Act

     This case is presented in detail in Lynch et al. (1994).  The case examines the iterative approach used by
the Office of Pollution Prevention and Toxics (OPPT) to evaluate ecological risks associated with a new
chemical regulated under the Toxic Substances Control Act (TSCA).    '-.'"'               "

Problem Formulation                                                ,                         •
     Relation to Environmental Management Decisions: The assessment is used to determine if, how,
and/or where the chemical identified in the premanufacturing notice (PMN) may be used. The need for risk
management steps4s based on the outcome of the assessment.
     Stressor Characteristics:  -The case examined a neutral organic compound:  The analysis began with a
consideration of the physical and chemical properties of the PMN substance along with data on the chemical
identity, structure, intended uses, and sites of use. This information was used to identify receptors potentially
at risk as well as the type of ecological effects that could occur. The evaluation focused on the parent
compound because investigators did not expect the PMN substance to degrade or be transformed into more
toxic metabolites.  The compound, which is expected to have low water solubility and sorb to sediments, has
a half-life for aerobic degradation of weeks; anaerobic degradation is slower.            '
     Ecological Receptors Potentially 'at Risk: .Because the chemical will be handled near water bodies and
discharged through publicly owned treatment works (POTWs), aquatic'biological communities were
considered the ecological receptors potentially at risk.                            '       .     '     •
     Ecological Effects:  Effects were inferred from the class of chemical to which the PMN substance
belonged.  Neutral organic compounds exert acute and chronic toxicity to aquatic biota through a narcotic or
nonspecific mode of action dependent on the molecular weight and octanol-partition coefficient.  Because of
the high Kow of the PMN substance, only chronic effects were expected at or below the chemical's solubility
limit.
                                                             '    —J           "      ,  -
     Endpoint Selection: The assessment endpoint is the survival, growth, and reproduction of aquatic fish,
invertebrates, and algae.  Measures of effect" included data on mortality, growth and development, and
reproduction derived from single species laboratory toxicity tests or quantitative structure-activity
relationships (QSARs).  Measures of exposure included data from laboratory-scale wastewater treatment
experiments and outputs from mathematical simulations of wastewater treatment plants.
                                                      A-15
                                                                                         i 0/13/95
-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
                      DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
benthic invertebrates using data from daphnids and modeled estimates for sediments; (4) collection of site-
specific data on use and disposal, and rerunning of the EXAMS II model; and (5) collection of actual toxicity
test data for benthic invertebrates. The emphasis on benthic invertebrates after the first few evaluations
reflected a growing awareness that the chemical would reside in sediment.
     Risk managers were involved at each iteration of the analysis. Communications between OPPT and the
manufacturer regarding the need'for additional information and the nature of ongoing analysis provided the
bases for subsequent iterations. The risk managers agreed that the PMN substance posed no unreasonable
risks. Because risks could be present at other sites not specifically evaluated, however, the final disposition
was a significant new use restriction (SNUR).
     Uncertainties:  The analysis identified uncertainties associated with each iteration as well as with the
overall assessment.  OPPT uses uncertainty "assessment factors" ranging from 1 to 1,000 to address: (1)
differences  in species sensitivity, (2) differences between acute and chronic effects, and (3) laboratory-to-field
extrapolations.
                                               A-17
                                                                                                  10/13/95
-------
                         '!»;' • • .' W i: '' i' ii: / '" •' '"; i'"'•'": i/ V ftk* *: "i iil|! „ ii^liil'l" '"<;i::!ii -Tui :> ^'§:' E'ftLl i i1 ilii;	^'JflffiWE
                                                     '	it	!'!i"l"''''! 'i	'	'
                                                        ..     ,
                                               i;"!1; i' , ,:!	I It'. 'I'	I'-SKj •!	fl. I,1"!	'ill' }' I]	!.ll ' TMUVWM	Blllil"'
            ..  •    ,   •	:	;L  „,;,„   DRAFT-DO MIT. Q,U9!E> CITE, | OR DISTRIBUTE
                      ,n        " i „ i|,|  .  >" „,'",<„  ,	•„ ' iji!1' ,:h, sill, „ . •,.]'' M',. |< 1.1:3,;, ;,s	•• v*";!1"1;1 i,,";1,, J""1 ••• IMII i"11 :Uiiiiu.;;.  '" 111, ni!n,| «•,.: t	, * ii;:/ ; ,:v  ;>';, •  iih,;, „ „ ' ". K',; i .n ';,, "iRir1 'iir'*if'"iL ' ; :i,;!!	:VHP'"",i,, • H' iirV" ''''^'N.'''!!!''^'!!1:..^!!!!''!!'.!!:!:;ii|;iH^:i:|iiiCf^iiiiii iil.iiiii|iijiP!|n  i]liiij£,,iiiiii;]ii!i'ili,iiiiiii!:!!iijyi:ii:tiTJl|;iij]B  ifjiiiinjiiiiiii.^'^^^^!'')^^^^^';^''?''!...!!:!'!;!!;!''!!	i;ta'i;Hi||fp«BjiH;jg          I
          8       i Problem Formulation
          9            Relation to Environmental Management Decisions: Information from this ecological assessment was
 ;:  :j	'•/;';"   a  ^^"'^'"^M^     	:ii:,:ii,,^       	if;"	!i!itl!I	                                   ,           ,                	
         10       considerecl ill deterrnining the need for cleanup of soils, ground water, and sediments. However, human

         11       health considerations ultimately drove the cleanup goals.
 i;,,    ••    , ,  ihr Y "•", :!li:;"':";li:",      .                     	        -,,,.-,	•'•	    ,       	  ,         ,,^        I
 -     1  25      fish populations in the Cochato River; (2) survival of local benthic invertebrate  populations in the river; (3)        |

 :,„.  T  26 „    survival and reproductive success of'songbirds in wetlands adjacent to the site; (4) survival a'n'3 rep'roSuctive

         27      success of small mammals in adjacent wetlands; (5) presence of a soil invertebrate community  that

         28      contributes to the functioning of the soil system and supports the local food chains; and (6) survival and

         29      growth of wetland vegetation.  Measures of effects included laboratory and field measurements of toxicity

  it     30       and direct observations of the presence of organisms.  Measures of exposure included analyses of chemical

   1      31      concentrations  in environmental media  and tissues as benchmarks for judging the poFe'ntial for toxic effects.
   !   '!' ,  	 	•, "",',„    'i	'   t&ll! ,, ,,<„,,	: J	',!,' li,,!!-  | i-1,!,,!!!!!):!,!1       "     '   	"i'MilUV?,.1	        .  	Illidl Illii	i	liliiiiiiilit	i	il':,	»[>,:>	,|,,-i!	,i	,	!:,:,(•	„	'	.'	ii	K/ii/	i	liiiini	i	•«	i	ii	iiiiiiiiiiiiiiii

 „' ,i i, '  , «' „    	   ;'   ",,,',   ',','" ,;;,ITHii  ii>,|  ,'  ' ',""  ','"','" T|,IHli, I"1 Sill1,: mi :T !'	 !, ,i;, 'I1 ',	i"1,?'!' ,,i,,: ! I'fl'nli1 i:i, ;T:ll;""i, ,:B itUi'^'firiM                           	I!K::II11IS::II!:II|IVI!'^I^^^^^
     '	'   ,„,!  /'    i 	   '   ,!,i	,,   iiii'iil'i , i'!,!!' , i;!1!, , 'i',,,/'	i "'„ ' !"\ !,i ^'f !l,|h , m ft , i!"',,"' P i',' ilii ,|!! il Sii,':,! ij'* -I il , iSiliiK       |:iiBM^


                           	
                                                                  	;,;;	'.	\	|	[	li,	:	|	;!	L;,,,,:	1	\	„,:	•	L	1	J	 .
                                                                                                                                'T;ii4::'!i!! I
-------
                                DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE'.
   1      Analysis           .      "
   2 .   -       Ecological Receptor Characterization: Field studies were conducted in the river, wetland, and upland
   3    '  habitats to describe the composition of invertebrate and vertebrate communities.  The studies focused
   4      specifically on benthic invertebrates (one-time survey), fish, nesting songbirds (detailed observations during
   5      breeding season), other bird species, amphibians (during breeding season), mammals (trapping over 2
   6      months), soil invertebrates (one-time quantitative survey), and wetland vegetation. A team with expertise in
   7      the particular groups of organisms made theseobservations.       :.                     .      •  _  '  "
   8       .    Characterization of Exposure:  Extensive measurements were made of contaminants in surface soils,
   9      ground water, surface water of the river, and sediments. Many of these observations were colocated with
  10      biological measures. Residue analyses were conducted onJish, soil invertebrates, juvenile songbirds (born at
  11      the site), small mammals, and wetland vegetation.  Dialysis bags filled with hexarie (surrogate fish) were also
  12      used to estimate exposure at various locations in the river. t Food chain models were constructed for wetland
  13.      songbirds and mammals based on a knowledge of feeding habits and tissue residues!
  14           Characterization of Ecological Effects: Effects were evaluated using a combination of literature
  15      values, laboratory and field toxicity studies, and field  observations. These included water quality and derived
  16    .  sediment quality values, sediment toxicity studies with three species, toxicity studies on ground water and
  17      aqueous soil extracts, laboratory and field toxicity studies of soil using earthworms and plants, a quantitative
  18  •    benthic survey, a quantitative soil invertebrate survey, and direct observations on the presence of birds and
  19      mammals with particular emphasis on juvenile birds as indicators of reproductive success.  Literature values
  20      were used to evaluate the toxicity of pesticides to birds and mammals.
  21  .    .-              •   .           .                                 .        ;
  22      Risk Characterization
•  23           Risks:.  A weight-of-evidence approach was used to evaluate risks in which various lines of evidence
  24      were compared. The quotient method was used as the primary basis for evaluating doses, exposure levels, or ,
  25   -  tissue levels against published or derived toxicity benchmarks. The case reveals the value of working with
  26      multiple lines of evidence when evaluating complex situations such as Superfund sites with multiple
  27      contaminants and natural variability. Some lines of evidence converged while others diverged, requiring a
  28      consideration of the weight that should be placed on each line of evidence in assessing the risks.
  29           The analyses revealed (1) risks to benthic invertebrates in some locations of the river, (2) risks  to soil.
  30.      invertebrate communities in swale areas where contaminants were transported through the wetlands, and (3)
                                                         A-19
10/13/95
-------
                                                                                                                  •
 2
 3
";4
 5
 6
 7
 8
 9
10
11
fa
13
      . i   I||N „ „  	,	,,  „	 .^..TO.^vJr0,0,	;y°J	§u,°,TBE,i	Sl?i,°Ji	ww£rram	
 possible risks to the reproductive success bit songbirds. The presence of contaminants did not appear to pose
I. ^     ,» '"',,'',„ '  lidililliiil  "i •': "•' " i '!	,,";,ii	ri|i .H, !:!!	if  f\'" ,1, '!" ' '"" I,"1"11" !!»„ I Mi!!! ,,'iii.,:!',1., 'M rJl'i '''ji I!111-!! In1'1! |j!:ll!ll;:l!!!1i:^il!!lt!hi<|l|l!»fli«i!!>i!lli!!!lii1^                   ll!!!!i!iiiiU!ili*!:ill!'1S;il1!!l:,mi!!IC!iilill!S!lTI!llia!iBiill!;U        IIEIilillliiillillliiiilliiEilllilillllilllllllllllllllllllllill'ITIlll
 a risk to wetland vegetation.
       Uncertainties:  The patchwork of confirmatory and contradictory findings highlights the limitations of
 available methods and cautions against reliance on any single method.  To some extent, uncertainties in the
 analyses can be addressed by a careful consideration of multiple lines of evidence.  To a limited degree,
 i,; •  , •"  ,•  J-. i1	ijiiiiiii • ;>;: ,.'v --i	!;> ••!3,	• n e ir  m-	i! .	IFI: :.,	11	',is-f*s':;(:«w    	IIB^^^^^^     	iiiiHiiiiii^        	syinaiM^^^^^^^^   	         :  HIv^^
 uncertainty wai evaluated quantitatively by considering the sensitivily of toxicity quotients to the range of
 values (e.g., toxicity benchmarks, exposure levels) obtained from the literature and/or derived or  measured  at
 the $tc. The resulting ranges in quotients were displayed graphically with respect to a hazard index scale.
 Logistics  precluded obtaining certain measurements that were more directly associated with assessment
 endpoints. The inability to obtain such data from the available resources contributes to the uncertainty in the
 analyses but also reflects the reality of field work at complex sites. These logistic constraints indicate the
, i i'  	• •)  ",;,; i, i; • i liillll1 • ::ilSi;i" iiiii 'in	1 i j"! *!!?'"J!!!!.!«',	,i!! jfftWffCM XT	W.W	! MWJfB	   ,   IIIIIII I "i I I       .                 ,       t  ,     .   ' '
 importance of well-designed but realistic plans for acquiring adequate data in order to conduct weight-of-
'! N,'I. ,   ,   ''' fiiki IIBIT' iiHIIi'l" i'1 !!,lll;' I'illii ii, rjiihiiif!lli;|ll!lllill!|llll||l!l;illE|ill!l|h|i|< illllll
                                                         i * <.,:.                                      •
                        il'iiK
                                                               i!	:	                                                                       I
                                            Rill< ir''I; "„ ,|	' *;i!.,1li ...'I	Xl1' «T1 n*."I '.!',' 'jM
                                                                                        lli'.iWilil!"! IT'llH •'!	'i illil,i''j	Illilll !' i|" ii ,i IIIII1 ii illillll iiiniHlHil II i1,1:' fcililJI ul tlPlililin'Klbii Ir iTftl	lii'liKIP+ilil!!!"!11™!1, Ijli!,,:!!!!!'!11:!!!!1!11!!!!!!!1!!!!1!!!!!!!!!!!!!11"!! I'11!1!!1!!!!!!!!!!!!!!!'!:!!!!!!1!!'::!!!!! j

                                                                                             lit.!! ''Pi;1!' ifl tiniii'lPf 111!1!!!:,!!l! !i|i|:i:i:|!::iii|lllllll!il|i:ii:i"lilli ^i' m,	i * lilllillllPlllilllliii'lHn'HIrflllllHllli1" 'i" lll'tlliiiMliililillllilllElnlllllirii.iiillillllllP    I
                        1 !TC1 i',,,1';1"

                                                                                                 iiMi' Si'i'iiiii'iiiiiiHii'i'1; iiinuiaiii ''iiiii||i|ii i N'liniii Jiiiiiiiiiiii1 i if" iiH'ij iini'iiiiLliiiiiii1' ii'i i'liiiiiiiiiini", ii tiiirniflii i '»ii
                                                                                               iiiiii'Nid
                          lili ' •?'. 1	11,1::'; (  ..'	't;:!', |,:|i|,! -,•,";	a •Tfjift >,nft'i,;:	(;  l.^.jtis Mf
                                                                                                    |,|	             -
                                                                                                    Hi
                                                                                            ISjiia^^^^^^^            	  • .
                                                                                            iii;i;"ii|ii	iih'ii'ill'i'llJliLluilliii'qliTThlPiiU'illll
                                                                                                              Ii'i! IBII'lUIIJIi Jilllliriidi. "'Illlipiilll!1' I'IIIPli lllilHiJI1' jljf
                                                            A-20
               1 Hi, ill'!1 , "i*" •!„;,' , »i,, • •!	''' I'll '„., Ii*",," '',  .' if •"!',»»''I1,:,1	Ini'.iLil'W1, "„ /JfllV""'1	i'1!"!1 ifyf '.JTIT	,.i,,-	
                                                                                                                         	[07T3795	
                         iSII ''•;•  • ''";:-:  "iiO,,1!!!:!"!: ;•, ,•;,• f,;;f,'!	£3it'^K^"K:r,	IttWm	KsJBIfoSI HHIiEfiiiaPf!!??	KSWBaS	ilill!	mm.

-------
  1
  2
  3
 .4
  5
  6
  7
  8   '
-9
 10
 11
 12
 13  '
 14
 15
 16
 17
 18
 19
 20
 21
•22
 23   -
 24
 25
 26
 27 "
 28
 29
 30
 31
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE             '
                                     Appendix A References

Brody, M.; Conner, W.; Pearlstine, L.; Kitchens, W. (1989) Modeling bottomland forest, and wildlife habitat
     changes in Louisiana's Atchafalaya Basin. In: Sharitz, R.R.; Gibbons, J.WV, eds. Freshwater wetlands
     and wildlife. Oak Ridge, TN: Office of Science and Technical Information, U.S. Department of Energy,
     U.S. Departmentof Energy Symposium Series, No. 61. CONF-8603101.

Brody, M.S.; Troyer, M.E.; Valette, Y. (1993) Ecological risk assessment case study: Modeling future losses
     of bottomland forest wetlands and changes in wildlife habitat within a Louisiana basin. In: A review of
     ecological assessment case studies from a risk assessment perspective. Washington, DC: Risk
     Assessment Forum, U.S. Environmental Protection Agency; pp. L2-l'to  12-39. EPA/630/R-92/005.

Burmaster, D.E.; Menzie, C.A., Freshman, J.S.; Bums, J.A.; Maxwell, N.I.; Drew, S.R  (1991) Assessment
     of methods for estimating aquatic hazards at Superfund-type sites: a cautionary tale. Environ. Toxicol.
    , Chem. 10:827-842.             '        '

Callahan, C.A.; Menzie, C.A.; Burmaster, D.E.; Wilborn, D.C.; Ernst, T.  (1991) On-site methods for
     assessing chemical impacts on the soil environment using earthworms: A case study at the Baird and
     McGuire Superfund site, Holbrook, Massachusetts. Environ. Toxicol. Chem. 10:817-826,

Conner, W.H.; Brody, M. (1989) Rising water levels and the future of southeastern Louisiana swamp
     forests. Estuaries  12(4):318-323.

Harris, H.J.; Wegner, R.B.;  Harris, V.A.; Devault,  D.S. (1994) A method  for assessing environmental risk: a
     case study of Green Bay, Lake Michigan. Environ. Manag.

Houseknecht, C.R. (1993) Ecological risk assessment case study: special review of the granular formulations -
     of carbofuran based on adverse effects on birds. In: U.S. Environmental Protection Agency. A review of
     ecological assessment case studies from a risk assessment perspective. Washington, DC: Risk
     Assessment Forum, U.S. Environmental Protection Agency; pp. 3-1 to 6-25. EPA/630/R-92/005.-
                                             A-2I
                                                                                                10/13/95
-------
'.,•	;;:
   j*','
                                            DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
                                                 I  II II 11 I        I III  II      |     *                j                    i             i
                   "                                         *                                                      "•
                   Lynch D.G,, Macek, G.J.; Nabhoiz, J.V.; Sherlock, S.M.; Wright, R.  (1994) Ecological risk assessment


                         case study; assessing the ecological  risks of a new chemical under the Toxic Substances Control Act.
                   ;           ,    I il  I    |      i ii  hii'i i  | (lull1   111 i | mi  i | 'j in  Mi Ml' III	Iliiililillill IIIlKll 1	,	1	,.	„     MI
                         In: U.S. Environmental Protection Agency. A review of ecological assessment case studies from a risk

                         assessment perspective, volume II. Washington,"DC: Risk Assessment Forum  U S.  Environmental

                         Protection Agency pp. 1-1 to 1-35.  EPA/630/R-94/003.
                                                                                       Ill 11 I 111111111 111
                                                                                       111 I III I III III!! Ill
                                                                                                 III 111 II 111 II 111
                                                                                                                  11  iiiiii 11 in in ill i iiiiiiiiiiiii
   7

   8
';  , ",i
:>•.,?
                   Menzie C.A, ^urmaster D.E.;  Freshman, J.S.; Callahan, C.A. (1992) Assessment of methods for

                         estimating ecological risk in the terrestrial component: a case study at the Baird & McGuire Superfund
•t:^j	i-ivpysii	.wis	$	M	w	M	*	.••	•	***	i	TI	-	«;	•	i	»	«»	«	5	'n^Sft	
  Site in Holbrook, Massachusetts.  Environ. Toxicol. Chem.  11:/.45-260.
    	i»',  !r,'iiiiiii,:«'::/i ,?™«t* i:• re" iia:• ^.r^fithMaKii'i^M
          13

          14

          15

          16

          I?

          «.
          19
         ,22

          23

         . W

          15
                                                                               Ullili'!1 •I1*1 lllllllllll|IIIIIIIIPIPIPIIIIinillllliPIIIIIIII i.iPl.ilii"!""''^^.^'!!!!^'!'!!^!!!!! !l!IPIi;!llllilllllll!lll!!llllliiilPPil!IIPIIii!i|i!!iPIII|lli!ipPP][!l:iiil!PIIIIPi|l'(PPPIillll'ii|'l I 'illBIIIIINl IIII I ill 111 111 ill III 111111
           ...       , , " , ,  ipi	, , •, •	 • •, ij.    y ,„ |'i"                 ,,,T	•,	,,'lli, |ir ,„: "ir „ „»•••, | "p "•¥!	ii 1	ii'VY"™'11™^                                   " I1""""" ""l llllliif
           U.S Department of Agricuiture Forest Service. '('f993)"1Pest	nsl'	assessment	of"tne	mportaiion'of'F'i'nus


                radiata, Hothofagus dombeyi, andLaiireliqphilippianalogs from Chile.  Washington, DC: Forest


                Service, U.S, Department of Agriculture. Nfisceffaneous	FiaT>licatIonTla	1517.	,	
           U.S. Environmental Protection Agency. (1989) CarboFuran specTaT review tec'ruilcal' support document.
                                                        PTograms"	UK	EnvTronrnental	PToTectlon	AgeiicyT
                 89/027.
                         .Ililliilllillllllll1" i1 • : i:: ": <»' i: 'ill » ...... ; r ..... imni.1; • "" . nin \\\;*\ ,\ . n i.|!'!ii||!: ; jj": :"":,. i'TiiUi,1;; i ' i"; | f,,
                                                           : »• : JA:» ..... tiilli 111 LniiiilliK.!!!: i HI I \/( ?'    '        'II
           y.S. Environmental IProtection Agency.  (1993) A Review of Ecological Ass^

                 Risk Assessment PerspecHve/W           DC: Risk Assessment Forum, U.S. Environmental
                             ,  ,
                         .'• .-^  ' ' • .....  ., illinium hi.i ......... ..................... '''"
                         Protection Agency.
                           Ill 11 . • I'1" ........ '' , ill :4 : ,;,„ ..... " !' •"' T:';, ;,< I' v" <\ KK : * ,II9; :f i!1.! til ' „ ' i !ip !P »lr ' i "Hii?:, ""lllll : II"'1!1! PilHIHinPt " 'llf'ilJulUJiil'l!1 VW Wft ' ..... , itmi. I ni1 "iirjailiuliiii.lii
                            |, i" P. •; |'| ,i|k '"f ii'; " Y\.;¥ 'llnrfif1"! l,ii IfSi 'Jpillil  m U I'lUII
                            hi.i ......... ..................... '''" -      * ~
                                                                                                           ii; PIHIil«ini'l j*'
                                                                                                              :I!E;
                                                                                                                " :;vli;iii|i|i;il ..... iPiLcllvii illillrllhlh, ' iiiil
                                                                                                                  11
                                                                                                              :I!E; fiifilllli:!1!;111!'!!!;,!!11:!!!11'1,!!!1"1!';,!, S!;1!!1!11'1!'1 'I!!!
           U.S. Environmental Protection Agency.  (1994) A Review of Ecological Assessment Case Studies from a
              • .'i    "I:'1	II!   UPPIPIIIIIIKI1 ..  f'JlS ::"i, •" „ ':	1'ir1' f" i ijh'1'i.i ,1' IIPillil'l I1 I	I	U 	I 'liH,,'"!!' ,i, ,.H|"". n'lllHiiiHI "I,,'',1111111111111111!l".i I'll	:i1li|ii||ii.ii|iiilii<'il,;;lilii|.|.iill:11ll;l:ll4ni|||ii|l!l|ii|ii|iilli         	I	I	JBI	d!|	,r	infill,	m,"I	il	i
                                             Agency
      !'	m,*mm\
      ..rfi:': )l-:  i'X'1!' -.i,:^: "'M
                                     ...U	':„	"X-• .;.?.>!*. K/li •!'«,"i'Ji'lHOTflJL'.at
                                     ,&* • i-w'-tiw'• Ki • •.It	Mil!	')EIfeJ
                                     >,)> t«	", 'i!	IS j [: i,'; j	•'!!!!!;:;:;:' £!' 11!":::1! «1 ,              •
                        " , r i , ''iliinIPP '  411111
                                                  i| Mi1:	Ill '' "!'/;,':  4 :"K!l|v "li,'"!'1 illEi'll'l1	nil.i'.Wm^^^  	'F>JAS\
                            *••• „  "• jwiiii'  ,;	i[.:"":;.,;"'".;,. i;»"i!i;i	ii'M""ii'i:: iin; ',„, i	*\ 'j:	'"M"1 .SiiiK \
                            " ii,'"" . Hilt • u'V ....i'"'"1!:"" :il:V:;i i-,»! WS"",  i1!' ^^J ,'^irrf3ti*'.?'V S
                              1 '  .mil . •"
                            1 '' .,"» •	iWllliil.'-ii.,'1
                                     , r	i ,,,n  ips,ii ji	 ', "ii, > i, iinf',«i, 'I1:'1'1 'n"" r ,;,' i >." Hi,"11 ii>	ii!ii:i:,!!!'!'i^ii,,:i!ii'i''!;:!:;pii!ii!!i,;iiii:iii!i!!;iiiiiii|iii!ii:i;iii(    iiiiiiiii|iii!!i|ip;iPiipi:^ii'iiBiiiiii,ii»ivi||i'iiiiiiii!i'ii'i'iiipi',',,iMpis'iPViLiii^     	iii.ii'i'iPi'iW'iiJiiiiiiiiiiiiiiiiiiiir'iiVi.hiii.iiliii'iiii'iiiiiii^
                                     HIP,.:  ' ;v./'' Ii ii," I!':,  '',; •;, fl "!!i II,;:,:!1,,: ..^ i/ n [>• r, %''! ^ 'As	:^:.i'li'il.il.'llii:^l;;l;|l'll:l;lillilllllili':!lil!l	niilliiliiillf;!;,,;;!;!       '	HI J!:!A."'il!!;ii Iliiitiliiiii'iliiSBnnnili.iiiintliiii.illllili       "     •    .  •
               i;, „, ,   •  '    , ,„; „ •. • t'.t;  i--1-:^ -^ •' ;• ;- •; 'ii'1'' 5, fl -f. 9*^ ^^v.;;,:^1^

                                                                             j'iixiirliMipjsiHpiiiaiiiiE^
                                                «,,!!.!l!lI!,1!'" Kill' in!!1" ii » In,,,,1!,1! 'if,I; :& 'H'1'!,;*,,:'!:: I'.1:,!!1;..'"!11/!1! illd
-------
                     DRAFT-DO NOT QUOTE, CITE, OR DISTRIBUTE
  1
  2;
  3
;4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
19
20
.21
22
23
24
25
26
27
28
29
30
                                  APPENDIX B - KEY TERMS
                    1   -'••     (Adapted in part from U.S. EPA, 1992a)
                                                                    ° ^
agent—Any physical, chemical, or biological entity that can induce an adverse response (synonymous with
     stressor).                            ,       ;
assessment endpoint—An explicit expression of the environmental value that is to be protected.
characterization of ecological effects-A portion of the analysis phase of ecological risk assessment that '
   '  evaluates the ability of a stressor to cause adverse effects under a particular set of circumstances.
characterization of exposure-A portion of the analysis phase of ecological risk assessment that evaluates the
     interaction of the stressor with one or more ecological components. Exposure can be expressed as co-  -
     occurrence, or contact depending on the stressor and ecological component involved.
community—An assemblage of populations of different species within a specified location in space and time.
comparafive risk assessment —A process that generally uses an expert judgment approach to-evaluate-the
     relative magnitude of effects and set priorities among a wide range of environmental problems
conceptual model—The conceptual model describes a series of working hypotheses of how the stressor might
     affect ecological components. The conceptual model also describes" the ecosystem potentially at risk,  the
     relationship between measurement and assessment endpoints, and exposure scenarios.
cumulative ecological risk assessment—A process that involves consideration of "the aggregate ecologic risk
     to the target entity caused by the accumulation of risk  from multiple stressors" (U.S. EPA, 1995b)
disturbance-Any event or series of events that disrupts ecosystem, community, or population structure and
     changes resources, substrate availability, or the physical environment (modified from  White and Pickett,
     1985).   •                                '  -              '   '
ecological component—Any part of an ecosystem, including individuals, populations, communities, and the
     ecosystem itself..                                                       .                    -
ecological risk assessment-The process that evaluates the likelihood that adverse ecological effects may
     occur or are occurring as a result of exposure to one or more" stressors.
ecosystem—The biotic community and abiotic environment within a specified location  in space and time.
                                               B-l
                                                                                                  10/13/95
-------
   1
   2.
  10
  11
  12
  13
  14
 ,15,,
  16
 19
~;2Q
 21
 22
 27
 28
":29:
1,30
                ,.	•" t:  iiiiiif. ' <;"«, f	,'" jiiKi!,;* "liiwrii ill 'Jill:',:•"«,";! •ifflww Til'i	VMU WKi.iM'KinE.M
              '< '<,;' ?ir	»i"  ' liiii',>"i*• v,.Vi.:; v-!-"-' i'JS^K: " • ;• i	Jliiii -iisiwKsiitcKfrifc3ifttiifS3K8FHaE.il	i	>
                                 DRAFT-PQ NOT QUOTE, CITE, OR DISTRIBUTE
 environmental impact assessment—Assessments required Under the National Environmental Policy Act
• "  ,  , (NEPA)lo, evaluate Mly environmental ..... ef&cts associated 'wiS'propbsed" major lederal ..... actions";1 ...... Like
               ecological risk assessments, environmental impact assessments typically require a "scoping process"
         •. : ] -, ;  . , ' ; i . '! : ! !.::.!•;: ;. , . . m :> ;« , A «»;•: ; ;;» is ..... ,••> : •• tf«JKy»                                            •_                '
          ,  ,   analogous to problem formulation, analysis by multidisciplinary teams, and an additional requirement
                                '! "ih-lu '„ I'11: SI IE	'
      that uncertainties be presented (CEQ, 1986 cited m'Suter, 1993 a).
" i"l»    •» ,   ', ' ' T ,! !,!" liiiiiiiiinii .' i r: •! " :«'•' ,. " "! ' " ''"!,: ill ',!, ill'" ..... ! u'lBi «•' "vijf "ii!11"1*}"*1"1 'yiiiii ISii    '11"1                            '                !''!•t^  1'i*!^:1*!!'1'!^     	                                                                              I
      that is used hi support of exposure or effects analysis.
:'"     ;, i ij1 „"  ; ui"" "<  iliiiii i, i-, 'ilui Li;'' ,j;"	i1';;,; i1;, I "in: iMiii: ,„: iwi11 fe,, :t,,:	r1»«s t;;,' !iiv",,i: n s iifli	is: y til1 nM ,l ""Si	I nJii!!!«;i	:IH	hi	a jiniinib: 1190	i
                                                                         llVfffiillliljiiilliliBilW
 measure of effect—A measurable ecological characteristic that is related to the valued characteristic chosen as
      the assessment endpoint.
 measure of exposure—A measurable stressor characteristic that is used to help quantify exposure.
 measurement endpoint—See ""measure of effect".
                                                                                                             iillllllXllIM       I
                                                                                                             HinEIH^        I
          ,j ,:.'  i 'vj[  " "	I'1 ", >,	i '  iillll'lllllllllll I ,l'il III! 11 ui1 in nil 'Slilii^  ! i1,;	,,T,	• • e	li!;,;, :fi!ifjifi,::1":!	i	11 ajiEif^          	IIBIIIM^^^^^^    IIIIIH^^^     	IIEH^^^^^^^^    	            -
                                                             sPW^
                                                                                 u1'iiiii, Mi:.j:i,; ™ '::: f.iK: :.iiii si ui;,1: 'BS:,::;;;	:;,,,:i; !„;,: 'nai:;1!	;::i:rii:: s;1 'uc;::!!;; s is :;I;BOO IILU^^^^^
                                 >• i;: ;•• i •^••rr.. •;•"•*.•?;!) ^3	.fi';:!fM,;!	•&	isii	m
                                 •'"''''" ,rf '''"
                                                                   !1'	                                                      I
-------
                              DRAFT--D.O NOT QUOTE, CITE, OR DISTRIBUTE
 1     > primary effect—An effect where the stressor acts on the ecological component of interest itself, not through
 2           effects on other components of the ecosystem (synonymous with direct effect; compare with definition  •
 3           for secondary effect).              .••"''                                                 .
, 4      receptor—The ecological component exposed to the stressor.
 5      recovery—The partial or full return of a population or community to a condition that existed before the
 6           introduction of the stressor.                   '  .     •        ,
 7      relative risk assessment—A process similar to comparative risk assessment. It involves estimating the risks
 8           associated with different stressors or management actions. To some, relative risk connotes the use of
 9           quantitative risk techniques, while comparative risk approaches more often rely on delphic approaches.
10           Others do not make this distinction.
11      risk characterization—A phase of ecological risk assessment that integrates the results of the exposure and
12           ecological effects analyses to evaluate the likelihood of adverse ecological effects associated with
13           exposure to a stressor. The ecological significance of the adverse effects is discussed, including
14           consideration of the types and magnitudes of the effects,  their spatial and temporal patterns, and the
15           likelihood of recovery.
16      secondary effect—An effect where the stressor acts on supporting components of the ecosystem, which in turn
17         -  have an effect on the ecological component of interest (synonymous with indirect effects; compare with
18           definition for primary effect).
19      source— An entity or action that releases to the environment or imposes on the environment a chemical,
20           physical, or biological stressor or stressors.                                        '  -  "   . .
21      'source term- As applied to chemical stressors, the type, magnitude, and patterns of chemical(s) released.
22      stress regime-The term stress regime has been used in at least three distinct ways.  (1) to characterize
23           exposure to multiple chemicals, or both chemical stressors (more clearly described as multiple exposure,
24           complex exposure, or exposure to mixtures,  (2) as a synonym for exposure that is intended to avoid
25           over-emphasis on chemical exposures (3)  to describe the series of interactions of exposures and effects
26           resulting in secondary exposures, secondary  effects, and, finally, ultimate effects (also known as risk
27           cascade [Lipton et al. 1993]) or causal chain, pathway, or network (Andrewartha and Birch, 1984).
28      stressor—Any physical, chemical, or biological entity that can induce an adverse response (synonymous with
29           agent).                                           •
                                                        'B-3
10/13/95
-------
1


2


3
          i ' ;    j   '  ", , •", i|,   "' il'j1 j!!,i ,  "',''" '«« ,„'"' • ' ", '  "„ ,i'  '' III11'1 MV	 '! ill., %#|V'''H'''^^             fli||n|^'*»'i!|iiiirt
                    ,-. •'  •: i '  11 Jiff i   i:''';-:"•  -'•" -i•  . "ti •	(;*';":<':' "'• :•<: wW .ft '    '                            	ijilil!^^^^^^^                                'I
                      „  .     . "Killl.,!   1 ;"   '    ,;'    *'. . ln 1.1,1	i;1;	;;	i. jijn ./I, nil.i|	 i;!;,.!.:;.!!!!	I	ii!i	M	iiiiil'...!!!':!.!"!!'.'.;!! rin:li;4.	It;'! Kigi[l:ili!in^^         lllllllillllM^^^^^^^^^	lilliiiil.UX11!1^^^^                            	

           stressor-response profile—The product of characterization of ecological effects in the analysis phase of
                      	i'' ' ^  , I'lluillfeii ''"' I,'*'1, "T. " ,. '"',„' ! * +'• ! 'ii C' "lit "Hi ii.,.1 ,v !«!!''' siili :i!l!..i::,n!!!. : vxilhiii, n,. !iii,.i!l!llii!l!n::'!'Ki:':i'ii:l"i:!il>i'i!!!i'!	aJiiillfi' iiPiiiJi'ipiJii'riiiiiiiiiiiiiiiipiipii'iii'ii'ji'iiiiiiiiiiiiiiifci \\f,\,	BIB!^^                          ,,,!"!i!!!B!!:i:i":;!.i!:l:i.,,l::!!>!B1, ii;,"

                  ecological  risk  assessment.  The stressor-response profile summarizes the data on the effects of



                  .sjressor-and the relationship of the data to the assessment endpoint.
4         trophic.levels—A functional classification of taxa within a community that is based on feeding relationships     ,
  ,: ",   "',	"   i  ii,   "i „„ 'i ",,:, i,,'   :' jiinairiki  .."if1;	, 'i» Jiii.i''"'^,'!!'!' jiiiiiHiiiiifi1!'!111!; ' jt ir «i riiBhiiii	iiiiS'fliiiiriJsiii'iiiiT'iKiiiii'i;"1!1:	iiiiiiiiic^iniPi's^iiip'Tii.H                             i ii  111 i mi in i HI iliniiiiiiin n in i nun i  iniiiii  in inn inn in nnniiinniiiiniiinnnllip innnnilinnnnnii iiiiinn inn

5'


6
                  (e.g., aquatic and terrestrial green plants comprise trie  first trophic level and herbivores comprise the


                  secdnd).
                                   ;i	MM	,1
                                   	>;:,;<*, 'f,'1:,  ;
                                                 •  i: f	i-:./	mm.%	mm
                                                             >iii: "i i	' ir toil, ,ik i;»:?;\;»I1; i>' iL/ri T ii!,:'.'	:	if L; 111: ;!:;';;,;?
                                                 •ife;^'?'
iiiiK^^^^^^^^                	iiiiiaifliiH^^^^^^^^
	iiiilPJptiin'viiiii MtllprPtiii'.iiiJ	
                               Ill III       "' V"'. f  I.P : '. \L\" I1"1'!::"!'"!:1",1  " ' ..,•'
                                                              ''''!,: ""^ :|!i";:ii:":" „!;' ,"''l!r"ll|!!,,;|1|ii;';!$.* :v!£S^!2!S'K.H?fr'$,!

                                                             'j ifi'tiHji'j^^^'ii'i^l^v^itjLi.ii	'lAMtNi.i^iAi.Kj
                                                                                                                                     -

                                                                                                                                liiiiiiCiiii^riiiii'iviii^LiiiiiiiiJii'ii'iiiJii
                                                                                                                                                i||;i;|||ii|iiiiniriniMiiiiiiijiniiiii«              I



                                                                                                                                               	',',",,'  ', |" *!, n'"	'' ,|	||nn]	''	]	|,'	I
                                                                                                                                              !!iiiiiiiiiii;;iiiiii:pi«iiii«!iiiii««f              I
                                                                        ',,iiii,,!i >iir:i,;.xiiVi«,,, .T**	,	,.,,„',	,    .,   • •   	
                                                                        ,>ni i vii1 "'f'n '<: i" /r' "i; 'iviPi,1:!v iiMSi liiiuH'i,' uniiiiiiiiJiiiiiiia	'"TiiikiifiPii1	iniiiini1" innnnnpnnniiiinnilniiiipniiiiniiisiiiiii.iiiii.iiiiiiiijjiiiiiiiiiiir.'iiiiiipi1::'
                                                                                                                                                  10/13/95
                                                                                                                                     !1; >i|i;iiiiiiiiiiiii»i;;iliPiiiiiiiiiinii,ii:pi!i:iiii:iiiiiiiinTil.iii
-------